Applied Mineralogy of Cement & Concrete 9780939950881

Citation preview

REVIEWS in MINERALOGY and GEOCHEMISTRY Volume 74

2012

Applied Mineralogy of Cement & Concrete EDITORS Maarten A.T.M. Broekmans Geological Survey of Norway (NGU) Trondheim, Norway

Herbert Pöllmann Martin Luther Universität Halle (MLU) HALLE (Saale), Germany

ON THE COVER: Background image: polymict aggregate in concrete, containing several types of sand-/siltstone in various colors, black lydite, chert/flint in various shades with and without cortex, white vein quartz and reddish quartzite, etc. Several lithologies behave alkali-reactive as revealed by internal cracking, presence of dark rims, etc. © Maarten ATM Broekmans. Front cover, top: Panorama of the Three Gorges Dam in China facing south, one of the largest concrete structures in the world, still under construction in October 2004. The dam with the Chinese characters on the right was temporary only, and was removed by blasting when the main structure was finished in 2009. © Maarten ATM Broekmans. Front cover, bottom: Rosette of layered doublehydrate (LDHt) phase in CAC. © Herbert Pöllmann.

Series Editor: Jodi J. Rosso MINERALOGICAL SOCIETY of AMERICA GEOCHEMICAL SOCIETY

Reviews in Mineralogy and Geochemistry, Volume 74 Applied Mineralogy of Cement & Concrete ISSN 1529-6466 ISBN 978-0-939950-88-1

Copyright 2012

The MINERALOGICAL SOCIETY of AMERICA 3635 Concorde Parkway, Suite 500 Chantilly, Virginia, 20151-1125, U.S.A. www.minsocam.org The appearance of the code at the bottom of the first page of each chapter in this volume indicates the copyright owner’s consent that copies of the article can be made for personal use or internal use or for the personal use or internal use of specific clients, provided the original publication is cited. The consent is given on the condition, however, that the copier pay the stated per-copy fee through the Copyright Clearance Center, Inc. for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other types of copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. For permission to reprint entire articles in these cases and the like, consult the Administrator of the Mineralogical Society of America as to the royalty due to the Society.

Applied Mineralogy of Cement & Concrete 74

Reviews in Mineralogy and Geochemistry

74

FROM THE SERIES EDITOR Several years in the making, Applied Mineralogy of Cement & Concrete was finally brought to completion in 2012 by the persistent efforts of the volume editors, Maarten Broekmans and Herbert Pöllmann. Their efforts are greatly appreciated. Any supplemental materials associated with this volume can be found at the MSA website, www.minsocam.org/MSA/RIM. Errata will also be posted there. The reader will also be able to find links to paper and electronic copies of this and other RiMG volumes. Jodi J. Rosso, Series Editor West Richland, Washington April 2012

PREFACE Since its inception in 1974, the “Reviews in Mineralogy” (“–and Geochemistry” from 2000) series has published over seventy volumes covering a diverse range of topics from theoretical to applied, and from very specific to generic. The idea for this RiMG volume was initially conceived in 2006, and the revised proposal was approved by the MSA Council in 2009. ‘Building materials’ as a generic term encompasses steel, aluminum, copper and a range of metal alloys, glass and glaze, particulate materials like sand, gravel, or crushed rock, and natural stone of sedimentary, igneous or metamorphic origin. Each of these materials sees a wide range of applications, from structural/bearing via functional to merely ornamental and decorative. The wide range of ‘building materials’ application is achieved through an equally wide range of processing, from use ‘as is’ (e.g., stacking boulders to make a retaining wall), through simple re-dimensioning and fitting (e.g., splitting and sizing of roofing slate) to purification and complex treatment in multi-stage processing (e.g., glass, Portland cement clinker, concreting). The use of building materials, their applications and processing has changed considerably with the development of civilization and technology. Consequently, comprehensive coverage of building materials, applications, processing and history would require multiple volumes. The present RiMG volume contains a selection of papers on the applied mineralogy of cement and concrete, by far the most popular modern building material by volume, with an annual production exceeding 9 billion cubic meters, and steadily growing. Not even all ‘concrete’ topics can be covered by a single volume, but an interesting assortment was finally obtained. The seven chapters deal with mineralogy and chemistry of (alumina) clinker production and hydration (Pöllmann), alternative raw clinkering materials to reduce CO2 1529-6466/12/0074-0000$05.00

DOI: 10.2138/rmg.2012.74.0

Applied Mineralogy of Cement & Concrete ‒ Preface emission (Justnes), assessment of clinker constituents by optical and electron microscopy (Stutzman), industrial assessment of raw materials, cement and concrete using X-ray methods in different applications (Meier et al.), in situ investigation of clinker and cement hydration based on quantitative crystallographic phase analysis (Aranda et al.), characterization and properties of supplementary cementitious materials (SCMs) to improve cement and concrete properties (Snellings et al.), and deleterious alkali-aggregate reaction (AAR) in concrete (Broekmans). Finding reliable volume contributors is never an easy task, and we are immensely grateful to all authors for their submissions and for keeping deadlines. We are also greatly indebted to RiMG Series Editor Jodi J. Rosso for editorial assistance and her responsiveness to quickly elucidate emerging copyright matters. Finally, we are thankful to our numerous ‘concrete colleagues’ in the field, for having inspired us unknowingly during the many meetings and gatherings we have had with them through the years. Trondheim / Halle (Saale) late April 2012 Maarten A.T.M. Broekmans Geological Survey of Norway (NGU) Herbert Pöllmann Martin Luther Universität Halle (MLU)

iv

Applied Mineralogy of Cement & Concrete 74

Reviews in Mineralogy and Geochemistry

74

TABLE OF CONTENTS

1

Calcium Aluminate Cements – Raw Materials, Differences, Hydration and Properties Herbert Pöllmann

INTRODUCTION ....................................................................................................................1 Raw materials ................................................................................................................1 MANUFACTURE OF CAC ....................................................................................................3 CHEMICAL AND MODAL MINERAL COMPOSITION .....................................................5 CALCIUM ALUMINATE CEMENT – PHASE DIAGRAMS ..............................................11 Mineralogical variability .............................................................................................13 CHEMICAL AND MINERALOGICAL PHASE COMPOSITION ......................................17 Calcium aluminates .....................................................................................................17 Calcium silicates..........................................................................................................20 Calcium aluminum silicates and aluminum silicates ..................................................23 Calcium aluminum ferrites ..........................................................................................27 Compounds containing magnesium and other species ................................................30 HYDRATION OF CALCIUM ALUMINATE CEMENTS ....................................................36 Description of crystalline hydration products of CAC ................................................39 Descriptions of amorphous, pseudocrystalline and crystalline aluminum hydroxides and sulfur-containing phases .............................................43 Hydration mechanisms and setting of CAC ................................................................51 Hydration of CAC under the influence of different admixtures .................................54 CRYO-SEM INVESTIGATIONS OF CAC HYDRATION: MICROSTRUCTURE DEVELOPMENT ........................................................................56 Effects of Li2CO3 accelerator on CAC hydration ........................................................59 Effect of Fe impurities on CAC hydration ..................................................................60 Hydration mixtures with different other materials ......................................................61 CAC - application and other cementitious mixtures ...................................................65 ACKNOWLEDGMENTS.......................................................................................................65 LITERATURE ........................................................................................................................65

v

Applied Mineralogy of Cement & Concrete ‒ Table of Contents

2

Alternative Low-CO2 “Green” Clinkering Processes Harald Justnes

ABSTRACT ............................................................................................................................83 INTRODUCTION ..................................................................................................................83 CEMENT CHEMISTRY BACKGROUND ...........................................................................84 Cement chemist’s short hand notation.........................................................................84 Clinker production for Portland cement ......................................................................84 High belite cement clinker...........................................................................................89 Calcium sulfoaluminate cement (CSA) .......................................................................89 Calcium aluminate cement (CAC) ..............................................................................91 MINERALS AS ALTERNATIVE TO LIMESTONE .............................................................93 General ........................................................................................................................93 Gypsum .......................................................................................................................93 Wollastonite .................................................................................................................94 Larnite, bredigite and calcio-olivine............................................................................95 Spurrite and associated minerals .................................................................................95 Hydrogrossular ............................................................................................................96 Anorthite and anorthosite ............................................................................................96 CONCLUSIONS.....................................................................................................................96 REFERENCES .......................................................................................................................97

3

Microscopy of Clinker and Hydraulic Cements Paul E. Stutzman

ABSTRACT ..........................................................................................................................101 INTRODUCTION ................................................................................................................101 PORTLAND CEMENT PHASE COMPOSITION ..............................................................102 CLINKER MICROSCOPY ..................................................................................................104 SPECIMEN PREPARATION FOR MICROSCOPY ...........................................................105 Materials for sample preparation ...............................................................................106 Preparation of clinker ................................................................................................106 Polished powder mounts of Portland cementitious materials....................................107 Cutting and grinding ..................................................................................................107 Polishing ....................................................................................................................108 Etching for light microscopy .....................................................................................113 SRM clinker...............................................................................................................113 Point count analysis ...................................................................................................114 SCANNING ELECTRON MICROSCOPY ANALYSIS .....................................................119 SEM imaging of microstructure ................................................................................120 Image processing .......................................................................................................131 Direct methods for development of standard reference materials ............................133 Phase estimates by microscopy and quantitative XRD .............................................137 Certified values by consensus means .......................................................................137 Application to cements ..............................................................................................137 vi

Applied Mineralogy of Cement & Concrete ‒ Table of Contents SEM imaging of fly ash .............................................................................................139 SUMMARY ..........................................................................................................................141 REFERENCES .....................................................................................................................143

4

Industrial X-ray Diffraction Analysis of Building Materials Roger Meier, Jennifer Anderson, Sabine Verryn

ABSTRACT ..........................................................................................................................147 INTRODUCTION ................................................................................................................147 METHODOLOGY ...............................................................................................................149 Phase identification with XRD ..................................................................................149 Phase quantification by using X-ray diffraction data ................................................149 Full pattern cluster analysis .......................................................................................152 Computed tomography ..............................................................................................152 APPLICATIONS...................................................................................................................154 Raw materials/quarry.................................................................................................154 Preheater/calciner ......................................................................................................156 Clinker/kiln ................................................................................................................156 Cement.......................................................................................................................158 Hydrated cement........................................................................................................159 Concrete.....................................................................................................................163 SAMPLE PREPARATION ...................................................................................................164 CONCLUSION .....................................................................................................................166 REFERENCES .....................................................................................................................166

5

Rietveld Quantitative Phase Analysis of OPC Clinkers, Cements and Hydration Products Miguel A. G. Aranda, Ángeles G. De la Torre, Laura León-Reina

BRIEF INTRODUCTION ....................................................................................................169 THE RIETVELD METHOD ................................................................................................170 General issues ............................................................................................................172 Structural description of the phases present in OPC materials..................................173 Whole-pattern quantitative phase analysis approaches .............................................177 SAMPLE PREPARATION AND DATA COLLECTION ....................................................179 SELECTED EXAMPLES OF RIETVELD QUANTITATIVE PHASE ANALYSIS ..........180 Clinkers......................................................................................................................181 Cements .....................................................................................................................185 Hydration products ....................................................................................................187 Durability studies ......................................................................................................190 Selective dissolution ..................................................................................................191 vii

Applied Mineralogy of Cement & Concrete ‒ Table of Contents INTERCOMPARISON AND COMPARISON WITH OTHER METHODS .......................192 Bogue and reverse Bogue calculation .......................................................................193 Optical and scanning electron microscopies .............................................................193 Thermodynamic modeling.........................................................................................193 Thermal measurements..............................................................................................194 Calorimetric data .......................................................................................................194 Nuclear Magnetic Resonance (NMR) spectroscopy .................................................194 GUIDELINES FOR RIETVELD QUANTITATIVE PHASE ANALYSES .........................195 Crystal structures .......................................................................................................195 Sample preparation and data collection.....................................................................196 Data analysis..............................................................................................................197 Final check.................................................................................................................198 FINAL REMARKS AND OUTLOOK .................................................................................198 ACKNOWLEDGMENTS.....................................................................................................201 REFERENCES .....................................................................................................................201

6

Supplementary Cementitious Materials Ruben Snellings, Gilles Mertens, Jan Elsen

INTRODUCTION ................................................................................................................211 DEFINITION AND CLASSIFICATION OF SUPPLEMENTARY CEMENTITIOUS MATERIALS.....................................................................................214 Definition ...................................................................................................................214 Classification .............................................................................................................214 MINERALOGY AND CHEMISTRY OF SCMS.................................................................216 Natural SCMs ............................................................................................................216 Thermally activated SCMs ........................................................................................224 By-product SCMs ......................................................................................................231 THE POZZOLANIC REACTION........................................................................................241 The pozzolanic reaction mechanism .........................................................................242 Pozzolanic activity.....................................................................................................248 Hydration mechanism and kinetics of blended cements ...........................................252 REACTION PRODUCTS .....................................................................................................253 Product assemblages..................................................................................................254 Hydration thermodynamics ......................................................................................256 PROPERTIES OF MORTAR AND CONCRETE CONTAINING SUPPLEMENTARY CEMENTITIOUS MATERIALS ..................................................260 Properties of uncured mortar and concrete containing SCMs ...................................260 Properties of hardened mortar and concrete containing SCMs .................................261 Durability of mortar and concrete containing SCMs ................................................264 CONCLUSIONS...................................................................................................................266 ACKNOWLEDGMENTS.....................................................................................................267 REFERENCES .....................................................................................................................268

viii

Applied Mineralogy of Cement & Concrete ‒ Table of Contents

7

Deleterious Reactions of Aggregate With Alkalis in Concrete Maarten A.T.M. Broekmans

INTRODUCTION ................................................................................................................279 Concrete in the built environment .............................................................................279 What is deleterious AAR? .........................................................................................280 Why is AAR important? ............................................................................................281 HISTORY AND BACKGROUND OF ASR .........................................................................282 First recognition ........................................................................................................282 Global and local acceptance of AAR.........................................................................282 Remediation and prevention ......................................................................................284 ORIGIN OF ALKALIS IN CONCRETE .............................................................................285 The Na2O-equivalent .................................................................................................285 Alkali from raw materials for Portland clinkering ....................................................285 Infiltrated alkali from seawater and deicers...............................................................287 Alkali released from aggregate ..................................................................................291 ALKALI-REACTIVITY POTENTIAL OF ‘SILICA’ .........................................................295 Quartz properties and its solubility under ASR conditions .......................................295 Moganite, chalcedony and opal .................................................................................300 Solubility of silica under ASR conditions .................................................................301 ALKALI-SILICA REACTION PRODUCTS.......................................................................303 Dissolution of silica under ASR conditions ..............................................................303 Appearance and chemical composition of ASR gel ..................................................305 Gel crystalline structure.............................................................................................308 ALTERNATIVE ALKALI-REACTIVE MINERAL SPECIES ...........................................312 Natural and industrial glass .......................................................................................312 Common rock-forming silicate minerals ...................................................................313 Alkali-reactivity of carbonate rocks and minerals.....................................................317 Other alkali-reactive species not being minerals .......................................................321 LABORATORY ASSESSMENT OF ASR CONCRETE .....................................................321 Sample acquisition and handling ...............................................................................321 Impregnation-fluorescence petrography ....................................................................324 Quantification of ASR damage and development over time .....................................326 Characterization of the aggregate in ASR concrete ...................................................331 In situ chemical analysis of ASR gel by SEM-EDS or EPMA .................................335 THE CRYSTALLINITY INDEX OF QUARTZ...................................................................339 SELECTED TOPICS FOR FUTURE RESEARCH .............................................................340 Reliable identification of quartz/silica properties governing alkali-reactivity ..........340 Extraction of alkali-reactive aggregate from field concrete ......................................341 Dissolution of quartz/silica under ASR conditions ...................................................341 Nano-structure of ASR gel ........................................................................................341 Effect of lithium on ASR ...........................................................................................342 SUMMARY AND CONCLUSIONS ....................................................................................342 ACKNOWLEDGMENTS.....................................................................................................343 REFERENCES .....................................................................................................................343

ix

Reviews in Mineralogy & Geochemistry Vol. 74 pp. 1-82, 2012 Copyright © Mineralogical Society of America

1

Calcium Aluminate Cements – Raw Materials, Differences, Hydration and Properties Herbert Pöllmann University of Halle-Wittenberg Von Seckendorffplatz 3 D-06120 Halle (Saale) Germany e-mail: [email protected]

INTRODUCTION High alumina cement was used widely in the UK after World War I, expressing its higher content of aluminum oxide in comparison to Portland cement. Several descriptions of investigations on calcium aluminate cements appeared, starting around 1850, with a first patent field in 1888 (Scrivener and Capmas, in Hewlett 1998). More widely known is the work of Bied (1909, 1926) filing a patent in 1909 for the fabrication of cement using bauxite or some similar aluminum or iron-rich material, with low SiO2-contents and limestone. In 1918, the trade name Ciment Lafarge Fondue was used for the first time. Meanwhile in the USA, Spackman (1908, 1910a,b) developed cementitious material marketed under the name of Alca natural cements. Several patents were applied and granted (Bates 1921). A description of non-Portland cements was given by Muzhen et al. (1992). The reason for looking into alternative cement materials was to develop cements with improved stability against sulfate corrosion. Nowadays, calcium aluminate cements are used specifically for their distinct properties (Brown and Cassel 1977), some of which are presented in Table 1. Calcium aluminate cements do have special applications and are therefore widely used despite the fact that worldwide fabrication is by no means comparable to OPCs (Höhl et al. 1936; Garcés et al. 1997; George 1976, 1980a,b, 1983, 1990, 1997; George and Montgomery 1992; George and Racher 1996; Gartner et al. 2002). Scrivener and Taylor (1990) and Scrivener et al. (1997a,b) described calcium aluminate cements and their use and microstructural developments. The use for experimental purposes was described by Auer et al. (1995). Thermal analyses for thermogravimetry of CAC-fraction and formation was discussed by Chudak et al. (1982, 1987). The CAC concretes and reactions were studied by Dunster et al. (1997, 2000) and Deloye et al. (1996) studied the so-called “Portland Fondu.”

Raw materials CACs are mainly produced out of limestone and bauxite (Bolger 1997). Sometimes hydrated lime, laterite, bauxite (Valeton 1986; Sehnke 1995), or alumina is also used as raw material. Reduced qualities are mainly obtained by increased content of silica (80

H2O (no additive) > K > Ca > Mg > Sr > NH4 2. anions: OH > H2O (no additive) > Cl >NO3>Br > acetate Overall effects on the addition of different salts to CAC are strongly dependent on the

Calcium Aluminate Cement

55

Figure 76. Heat-flow curves of CAC with the addition of different Li2CO3-solutions.

concentration and type of additive. Increasing temperature normally enhances hydration (Robson 1952, 1962, 1964, 1968; Pöllmann 1989a,b; Knöfel and Bier 1997; Schmidt and Pöllmann 2008, 2009). Besides Li-salts many different other salt solutions are applied as accelerator additives, but it has to be taken into account that in all cases the acceleration is strongly concentration and temperature dependent (Schmidt and Pöllmann 2008, 2009). For less active additives, temperature is even the dominant factor. Table 19 presents a summary on typical salts and their behavior as accelerators and retarders. The definite behavior must indeed be described as a function of type of salt, concentration, temperature and time and type of addition and method of addition (solid or liquid). The formation of a carbonate containing Li-LDH was studied by Beckermann et al. (1996), of others by Besserguenev et al. (1997). Organic acids and their salts do have a slight accelerating effect at low concentrations, but a strong retarding effect at higher concentrations. Table 19. Accelerating and retarding agents for CAC.

Accelerator Li-salts portlandite – Ca(OH)2 hydrated and reground CAC sodium carbonate – Na2CO3 sodium hydroxide – NaOH CaCO3 + CaCl2 CaSO4 + CaCl2 CaSO4 + CaCO3 dilute organic acids

Acceleration / Retardation Composition Determined magnesium salts calcium salts acetic acid formates acetates hemihydrate

Retarder borates carboxylic acids hydroxicarboxylic acids heavy metal salts sugars casein cellulose ethers glycerine citric acid

56

Pöllmann

Other typical additions for CACs and their mixtures are normally used for setting up definite properties like water reduction, fluidizing agents and air-entraining agents. The most effective salt for water reduction is NaNO3 (Bayoux et al. 1992a). Others are typically used for OPC applications as superplasticizers. Edmeades and Hewlett (1998), Currell et al. (1987), Sharp et al. (1990) describe the influence of (hydrated, reground) cement additives. Pöllmann (1990) gives information on (hydroxylated) carboxylic acids on the hydration of CAC, showing slight acceleration at low concentrations ~ 0.01 mmol/l, but strong retardation at higher concentrations ~ 1 mmol/l. The hydration products formed can be LDH-structures with carboxylic acid anions, calcium salts of carboxylic acids, but also amorphous gelous complex mixtures have been identified. Teoreanu et al. (1986) used organic admixtures on refractory CAC. Temperature influence on setting was described by Bushnell-Watson and Sharp (1986, 1990a,b) and Capmas et al. (1990). Other phases like ettringite (Moore and Taylor 1970; Kumarathasan et al. 1990; Pöllmann 2011), or LDHs are crystallizing in presence of other anions (Allmann 1968, 1970; Renaudin et al. 1999a,b; Pöllmann 2010a,b,c). The influence of foreign ions on the hydration was studied by Murat and Sadok (1990). The hydration in presence of sulfates was studied by Götz-Neunhöffer and Neubauer (2002) and of pure phases by Pöllmann (1989a,b) and Neubauer et al. (2002, 2003). Phase equilibria in the system CaOAl2O3-SiO2 were described by Quillin and Majumdar (1994). Kinetic mechanisms were investigated by Barret and Ménétrier (1974), Barret et al. (1974) and Barret and Bertrandie (1980). Nilforoushan and Sharp (1995) used alkali metal salts as additives.

CRYO-SEM INVESTIGATIONS OF CAC HYDRATION: MICROSTRUCTURE DEVELOPMENT The course of hydration can be imaged very nicely without destroying the watercontaining hydration products using cryo-SEM microscopy. The course of hydration and the formation of phases are given in Figures 77-88. Images were collected at variable intervals, to show the development of microstructure. Figure 77 shows typical microstructures of CAC after 1 hour of hydration. CAC particles have hydration products sprinkled spot wise on their surfaces, and space between individual particles is still wide open. Hydration has clearly initiated, but plenty of unhydrated CAC relics are still available. Figure 78 shows typical microstructure of CAC particles after 5 hours of hydration. The surface of the fragments is now covered with gelous hydrate. The microstructure reveals much more hydrates now partially filling the interstitial space between the CAC clinker fragments, consolidating the individual particles into a mass through adhesion. Figure 79 shows the microstructure of CAC fragments after 8 hours hydration, with hydrates now filling the space between the individual CAC fragments. The surface of these fragments is entirely covered by flakes and platelets 500-1000 nm of newly formed hydrates. These effectively armor their unhydrated clinker interiors against hydration, retarding further setting. Figure 80 shows typical microstructure of CAC fragments after 20 hours hydration. Calcium aluminate hydrates are now completely embedding the CAC fragments. The micrographs also show jagged edges on the CAC particles caused by dissolution. Unhydrated CAC clinker relics are firmly consolidated by newly formed hydrates, forming a compact microstructure. This hydration behavior can be affected using accelerating additives. The most widely used accelerator is dissolved lithium carbonate, which strongly accelerates the hydration even at low concentrations. Some images of hydration products obtained by Cryo-SEM are given in Figure 81. Alternatively, CAC hydration can be accelerated by increasing temperature.

Calcium Aluminate Cement

Figure 77. Cryo-SEM micrographs of hydrating CAC after 1 hour of hydration.

Figure 78. Cryo-SEM micrographs of hydrating CAC after 5 hours of hydration.

57

58

Pöllmann

Figure 79. Cryo-SEM micrographs of hydrating CAC after 8 hours.

Figure 80. Cryo-SEM micrographs of hydrating CAC after 20 hours.

Calcium Aluminate Cement

59

Figure 81 shows typical microstructures of CAC fragments after 10 min hydration under addition of Li2CO3. After few minutes only, 10 μm crystallites of C2AHx can be observed. A direct comparison of 10 min hydration of High-Alumina Cement CAC at 22 °C and 45 °C reveals the more advanced hydration at elevated temperature, as well as increased formation of calcium aluminum hydrates, visible even at lower magnification (Fig. 82).

Figure 81. Cryo-SEM micrographs of of CAC hydration without additives (left) and early hydration products (10 min) of CAC with addition of 0.01 molar Li2CO3 solution at 22 °C (right).

Figure 82. Cryo-SEM micrographs showing early hydration (10 min) of CAC at 22 °C (left) and 45 °C. (right)

Effects of Li2CO3 accelerator on CAC hydration The hydration and setting of CACs can be effectively influenced by additives. The most widely applied accelerating additive is dissolved lithium carbonate that is found effective even at low concentrations. Even after 20 min at a low 10 °C, newly formed platy hydrates are easily identified in hydrating CAC with added Li2CO3-solution accelerator. At 20 °C, much more yet finer grained hydrates are easy to identify after only half the time (10 min). Obviously, addition of Li2CO3solution accelerator to CAC effectively promotes the formation of CAH-hydrates. Figures 83 and 84 show the combined effect of temperature and Li2CO3-solution accelerator, both promoting the formation of hydrates from CAC hydration.

60

Pöllmann

Figure 83. Cryo-SEM micrographs showing very early hydration of CAC at (left) 10 °C (20 min) and (right) 20 °C (10 min) with 0.01 molar Li2CO3.

Figure 84. Cryo-SEM micrographs showing early hydration of CAC with addition of 0.01 molar Li2CO3 solution at (left) 22 °C (10 min) and (right) 45 °C (5 min).

Effect of Fe impurities on CAC hydration The presence of Fe impurities in clinker affects hydration behavior and formation of hydration phases of CACs in general. The effect of Fe on two specific CACs (C1 = iron-free CAC in Figs. 85-86, and C2 = iron-containing CAC in Figs. 87-88), was studied in detail by XRD, calorimetry and cryo-SEM microscopy (Pöllmann et al. 2008). Hydration conditions for both samples were water/cement ratio 0.4 at 23 °C and ~95 % relative humidity. For comparison, hydration of CAC C2 containing iron is given in Figures 87 and 88. The in situ measurement of hydration products only identified calcium aluminium hydrates of variable composition and AH3. Probably some incorporation of iron in these hydration products is possible; also some amorphous FH3 may have been formed. The many different in situ XRD diagrams can be ordered and simplified by using cluster analysis. The diagram in Figure 89 shows the PCA-development of CAC hydration by XRD using cluster analysis. It becomes obvious that all XRD diffractograms converge after ~30-40 h, implying that the hydration products have reached a similar intermediate stage (as hydration is not completed and will continue until long after).

Calcium Aluminate Cement

61

Figure 85. Hydration of iron-free CAC C1 clinker at 1 hour, 5 hours and 8 hours, by cryo-SEM and calorimetry.

Hydration mixtures with different other materials CACs are widely used in building material chemistry and are available blended with microsilica (Bentsen et al. 1990, 1996), blast furnace slag (Majumdar et al. 1990; Richardson et al. 1990; Osborn and Singh 1995), pozzolanic materials, (Collepardi et al. 1995), silica fume (Macdargent et al. 1992) and limestones or other fillers (Lagerblad and Vogt 2009). Phosphate modified CAC was described by Ma and Brown (1994). The addition of anhydrite, bassanite or gypsum leads to the formation of ettringite instead of calcium aluminate hydrates (Götz-Neunhöffer 2006; Westphal et al. 2009). The so-called ternary (cement) systems are based on mixtures of CAC with OPC and a calcium sulfate or sulfate-hydrate. These mortar systems show rapid hardening combined with rapid drying, and are occasionally applied in self-drying leveling compounds. Due to the fact that the formation of C2AH8 and CAH10 is limited and ettringite and LDH type phases are formed instead, the conversion to hydrogarnet may be partially inhibited. The

62

Pöllmann

Figure 86. Hydration of iron-free CAC C1 clinker at 15 hours, by in situ XRD (phase development determined by peak area analysis).

main phase found and reported in the presence of reactive silica is strätlingite, in the presence of sulfate and/or carbonate kuzelite, monocarbonate, hemicarbonate or other solid solutions. Kuzelite, TCAH and mono-carbonate can form extensive ternary solid solutions (Pöllmann 2006) of the composition 3CaO·Al2O3·(x)CaCO3·(y)Ca(OH)2·(z)CaSO4·nH2O, with modifiers x, y, z in the range 0.33 < x < 0.66, 0.17 < y < 0.50, and 0 < z < 0.17, provided x + y + z = 1. In addition, two new distinct ternary phases of AFm-type were identified 3CaO·Al2O3·1/6CaSO4·1/2Ca(OH)2·2/6CaCO3·nH2O 3CaO·Al2O3·1/6CaSO4·1/6Ca(OH)2·4/6CaCO3·nH2O Ettringite was first mentioned by Michaelis (1892), deteriorating the microstructure of these materials (whereas it normally positively contributes to hydration and early strength development). These complex LDH phases also occur in the presence of chlorine, when

Calcium Aluminate Cement

Figure 87. Hydration of iron-containing CAC C2 clinker at 1 hour, 5 hours, 8 hours and 15 hours by cryo-SEM and calorimetry.

63

64

Pöllmann

Figure 88. Hydration of iron-containing CAC C2 clinker, by in situ XRD (phase development determined by peak area analysis).

Figure 89. Cluster analysis of in situ XRD diagrams of the hydration of a typical CAC. The increasing numbers indicate hydration time up to 70 hours. After 30 hours, the XRD diagrams reveal differences are negligible.

Calcium Aluminate Cement

65

CAC-concretes are exposed to chlorine containing environments, in which LDH-type solid solutions of Friedel’s salt or hydrocalumites are crystallizing. The long-term stability is also influenced from the formation of these phases. The existence of these chlorine containing LDHs was reported by Kurdowski (1980) and Kurdowski et al. (1986, 1990) in presence of saline solutions. The solid solutions of ternary LDHs containing chlorine was reported by Pöllmann (1989a,b, 2010, 2011). Their composition can be represented as 3CaO·Al2O3·(x)CaCO3·(y) Ca(OH)2·(z)CaCl2·nH2O, in which x, y, z are 0 < x < 0.83, 0 < y < 0.66, 0 < z < 1, again provided x + y + z = 1. Many different LDH phases can be formed during these hydration reactions of these complex mixtures used in applications. The effect of different CAH’s on ettringite formation has been described by Yan Fu et al. (1996). Many new special cements are coming up nowadays, especially sulpho-aluminate cement to replace CAC. Li-LDH type materials were described by Mascolo (1986).

CAC - application and other cementitious mixtures Macdargent (1992) used CAC blended with silica fume. Properties, chemical resistance and studies of CAC and hydrated pastes were described by Midgley and Pettifer (1972), Neville (1975), Midgley (1976, 1980a,b, 1985), Mino et al. (1991), Mohan (1991), Muhamad et al. (1993). Parker (1954) described the constitution of CAC. Carbonation of phases in hydrated CAC was described Pérez Méndez et al. (1984). Osborn and Brecem (1994) used a rapid hardening cement based on CAC. Piasta et al. (1989) described the durability of CAC in sulfate water systems. Quillin (2001) gives details on belite-sulfoaluminate cements. Chemistry of strength retrogression is given by Rao (1980). A micro analytical study of CAC and slag mixtures is given by Rayment and Majumdar (1994). Risch et al. (1997) give the same kind of information on the use of CAC for immobilisation/solidification of waste. Activated CAC/ slag systems described Romero et al. (1997). Scrivener et al. (1997a,b) and Scrivener and Taylor (1990) described formulated CAC and the effect of CO2 on mechanical strength. Schuster et al. (1992) used a calciumaluminate method for precipitation of sulfate and heavy metals. Sudoh et al. (1980) described high strength cement in the system CaO-Al2O3-SiO2-SO3. Sugama and Carciello (1991) used phosphate-bonded CAC. Su et al. (1992) worked on non-Portland cement systems. Trivino (1986) worked on avoiding transformation reaction of “ciment fondu.” Walter and Odler (1996) also used phosphate CaO/Al2O3 mixtures. Xue et al. (1986) studied high self stress aluminate cement. The use of “Fondu” in concrete structures was described by French Standard (1991). Rapid hardening cement was presented by Unsin (1996).

ACKNOWLEDGMENTS High resolution cryo-micrographs were gratefully acquired at Centre for Material Analysis/ZWL Lauf. Th. Redtmann and D. Haberkorn helped in editing the manuscript. Many thanks to Dr. Maarten Broekmans, he was an extremely helpful and competent reviewer. The final manuscript was kindly improved with extreme competent and helpful effort by J. Rosso.

LITERATURE A literature summary is mainly based on properties and important CAC phases, the minor phases are not all discussed in detail in this contribution. Many interesting articles concerning CAC can be found in the proceedings of the calcium aluminate cement conferences in 1982, 1990, 2001 and 2008, comprising many contributions on CAC in the broadest sense.

66

Pöllmann

Abakumov AM, Rozova MG, Alekseeva AM, Kovba ML, Antipov EV, Lebedev OI, Tendeloo GV (2003) Synthesis and structure of Sr2MnGaO5+5 brownmillerites with variable oxygen content. Solid State Sci 5:871-882 Abe T, Sunagawa I (1995) Hexagonal CaAl2Si2O8 in a high temperature solution; metastable crystallization and transformation to anorthite. Mineral J 17(6):257-281 Abe T, Tsukamoto K, Sunagawa I (1991) Nucleation, growth and stability of CaAl2Si2O8 polymorphs. Phys Chem Miner 17:473-484 Abriel W (1983) Calciumsulfat-subhydrat CaS04·0.8H2O. Acta Crystallogr C 39:956-958 Abriel W, Nesper R (1993) Bestimmung der Kristallstruktur von CaSO4(H2O)0.5 mit Röntgenbeugungsmethoden und mit Potential-Profilrechnungen. Z Kristallogr 205:99-113 Abriel W, Reisdorf K, Pannetier J (1990) Dehydration reactions of gypsum a neutron and x-ray diffraction study. J. Solid State Chem 85:23-30 Adams LD (1997) Ettringite, the positive side. 19th International Conference on Cement Microscopy, Cincinnati/ OH, p 1-13 Aggarwal PS, Gard JA, Glasser FP (1972) Synthesis and properties of dicalcium aluminate 2CaO·Al2O3. Cem Concr Res 2:291-297 Akasaka M, Ohashi H (1985) Fe – Mössbauer study of synthetic Fe3+ - melilites. Phys Chem Miner 12:13-18 Ali MM, Agarwal SK, Handoo SK (1997) Diffusion studies in formation and sintering of CaAl2O4 and BaAl2O4 a comparative evaluation. Cem Concr Res 27:979-982 Ali MM, Raina SJ (1989) Kinetics and diffusion studies in CA2 formation. Cem Concr Res 19:47-52 Allmann R (1968) Die Doppelschichtstruktur der plättchenförmigen Calcium-Aluminium-H2O-Salze am Beispiel des (CaO)3·Al2O3·CaSO4·(H2O)12. Neues Jahrb Mineral Monatsh 140-144 Allmann R (1970) Doppelschichtstrukturen mit brucitähnlichen Schichtionen. Chimica Jhrg 24:99-108 Alonso C, Fernandez L (2004) Dehydration and rehydration processes of cement paste exposed to high temperature environments. J Mater Sci 39:3015-3024 Anderegg FO (1931) Industrial and Engineering Chemistry. 23(9):1058-64 Angel RJ (1992) Order-disorder and the high-pressure PT-IT transition in anorthite. Am Mineral 77:923-929 Angel RJ, Capenter MA, Finger LW (1990) Structural variation associated with compositional variation and order-disorder behavior in anorthite-rich feldspars. Am Mineral 75:150-162 Angel RJ, Prewitt CT (1986) Crystal structure of mullite: a re-examination of the average structure. Am Mineral 71:1476-1482 Anjos PNM, Pereira EC, Gobato YG (2005) Study of the structure and optical properties of rare earth-doped aluminate particles prepared by an amorphous citrate sol-gel process. J Alloys Compd 391:277-283 Aruja E (1956) The unit cell of orthorhombic pentacalcium trialuminate, 5CaO·3Al2O3. Acta Crystallogr 10:337-339 Aruja E (1960) Unit cell and space-group determination of tetra- and di-calcium aluminate hydrates. Acta Crystallogr 12(12):1018 Aruja E (1961) The unit cell and space group of 4CaO·Al2O3·19H2O polymorphs. Acta Crystallogr 14:12131216 Atoji M, Rundle RG (1958) Neutron diffraction study of gypsum. J Chem Phys 29:1306-1311 Auer S, Kuzel H-J, Pöllmann H, Sorrentino F (1995) Investigation on MSW fly ash treatment by reactive calcium aluminates and phases formed. Cem Concr Res 25:1347-1359 Bachiorrini A (1985) Polymorphism of monocalcium aluminate. Study by infrared spectroscopy. Cem Concr Res 15(1):167-173 Bacon GE, Roberts FF (1953) Neutron diffraction studies of magnesium ferrite-aluminate powders. Acta Crystallogr 6:57-62 Bacon GE, Welch AJE (1967) An x-ray examination of the spinel-type mixed oxide MgFeAlO4. Acta Crystallogr 7:36l-363 Bakalkin AP, Pitak NV, Shapiro YZ, Litvin LG, Gaodu AN, Sorin MN, Kupalova TP, Grebennykov PM, Shevchenko AT, Krutko GI, Mukhin AA, Karas GE, Elisova ZM (1986) Experience with producing high alumina cement by calcination. Refractories 27(9-10):592-594 Baker NC, Banfill PFG (1992) The use of admixtures in high alumina cement mortar for the marine environment. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 4:719-725 Baldock PJ, Parker A, Sladdin I (1970) X-ray powder diffraction data for calcium monoaluminate and calcium dialuminate. J Appl Cryst 3:188-191 Banfill PFG, Gill SM (1986) The rheology of aluminous cement pastes. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro 6:223-227 Banfill PFG, Gill SM (1993) Superplasticizers for cement fondu: effect on rheological properties of fresh paste and mortar. Adv Cem Res 5:131-138 Barnes P (1983) Structure and Performance of Cements. Applied Science Publishers, London

Calcium Aluminate Cement

67

Barret PP, Bertrandie D (1980) Coubre dInstabilité minerale dans une solution métastable de CA. Proceedings of the 7th International Congress on the Chemistry of Cement 3:134-139 Barret PP, Ménétrier D (1974) Contributions to the study of the kinetic mechanism of aluminous cement setting. II. Release of the factor responsible for breaking the latent period by the dissolution of a fraction of the initial cement. Cem Concr Res 4:723-733 Barret PP, Ménétrier D, Bertrandie D (1974) Contributions to the study of the kinetic mechanism of aluminous cement setting. I. Latent periods in heterogenous and homogenous milieus and the absence of heterogenous nucleation. Cem Concr Res 4:545-556 Bartl H (1969) Zur Struktur des 12CaO·7Al2O3. Neues Jahrb Mineral 91:397-404 Bartl H, Scheller T (1970) Zur Struktur des 12CaO·7Al2O3. Neues Jahrb Mineral 92:547-553 Bate SCC (1984) High Alumina Cement Concrete in Existing Building Superstructures. HMSO, Building Research Establishment Report SO4 Bates PH (1921) Cementing qualities of the calcium aluminates. Bureau of Standards 1921; Technology Papers, no. 197 Bayoux JP, Bonin A, Marcdargent S, Verschaeve M (1990a) Study of the hydration properties of aluminous cement and calcium sulphate mixes. In: Calcium Aluminate Cements. Mangabhai, RJ (ed) E and FN Spon, London. p 208-221 Bayoux JP, Letourneaux JP, Marcdargent S, Verschaeve M (1990b) Acidic corrosion of high alumina cement. In: Calcium Aluminate Cements. Mangabhai, RJ (ed) E and FN Spon, London. p 230-40 Bayoux JP, Sorrentino D, Falaschi JP, Capmas A (1992a) A study on calcium aluminate cements admixtures. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 4:705-711 Bayoux JP, Testud M, Esponosa B (1992b) Thermodynamic approach to understand the CaO-Al2O3-SO3 system, Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 4:164-1693 Beckerman SJ, Ford RB, Nemeth MT (1996) Conversion of gamma aluminate to lithium aluminium carbonate hydroxide hydrate. Powder Diffr 11(4):312-317 Bensted J (1976) The ferrite phase - an infrared spectroscopic study. II Cemento 1:45-51 Bensted J (1982) An infrared spectroscopic examination of high alumina cement and its hydration products. World Cem Technol 13:85-90 Bensted J (1983) comments on the cement chemical nomenclature system. II Cemento 4:193-200 Bensted J (1993) High alumina cement - present state of knowledge. Zement-Kalk-Gips 46(9):560-566 Bensted J (1996) A discussion of the paper “The influence of pozzolanic materials on the mechanical stability of aluminous cement” by M. Collepardi, S. Morosi and P. Picciolli. Cem Concr Res 26:649-650 Bensted J, Audley GJ, Aukett PN (1992) Applications of higher temperature conduction calorimetry to high alumina cement hydration. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 4:220-226 Bentsen S, Seltveit A, Sandberg B (1990) Effect of microsilica on conversion of high alumina cement. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London. p 294-319 Bentsen S, Selvod S, Sandberg B (1996) Effect of microsilica on conversion of high alumina cement. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London. p 294-298 Berastegui P, Erikson SG, Hull S (1999) A neutron diffraction study of the temperature dependence of Ca2Fe2O5. Mater Res Bull 34(2):303-314 Bertaut EF, Blum P, Sagnieres A (1959) Structure du ferrite bicalcique et de la brownmillerite. Acta Crystallogr 12:149-159 Besserguenev AV, Fogg AM, Francis RJ, Prince SJ, O’Hare D, Isupoc VP, Tolochko BP (1997) Synthesis and structure of gibbsite intercalation compounds (LiAl2(OH)6X (X=Cl, Br, NO3) and LiAl2(OH)6Cl(H2O) using synchrotron X-ray and neutron powder diffraction. Chem Mater 9:241-247 Bezou C, Nonat A, Mutin JC, Christensen AN, Lehmann MS (1995) Investigation of the crystal structure of γ-CaSO4, CaSO4·0.5H2O and CaSO4·0.6H2O by powder diffraction methods. J. Solid State Chem 117:165176 Bied J (1909) Aluminous cement. British Patent 8193 Bied J (1926) Recherches industrielles sur les chaux, ciments et mortiers. Paris: Dunod Bolger R (1997) Non-refractory bauxite. Industrial Minerals, March 21-29 Bowen NL, Greig JW, Zeiss EG (1924) Mullite, a new silicate of alumina. J Wash Acad Sci 14:183-191 Boysen H, Kaiser-Bischoff I, Lerch M (2008) Anion diffusion processes in O- and N-Mayenite investigated by neutron powder diffraction. Diffusion Fundamentals 8:2.1-2.7 Brisi C, Lucco-Borlera M, Montanaro L, Negro A (1986) Hydration of 5CaO·3Al2O3. Fig. No. 2492. Cem Concr Res 16:156-160 Brown RA, Cassel B (1977) High Alumina Cements - Background and Application of Thermal Analysis Methods. - Reprinted from American Laboratory, No. 1 Brunauer S, Emmett PH, Teller E (1938) Adsorption of gases in multimolecular layers. J Am Chem Soc 60(2):309-319

68

Pöllmann

Budnikov PP, Tcherkassova AF (1955) Gehlenite and its aluminous cement. Doklady Akademie Nauk SSSR 2(4):793-795 Bushnell-Watson SM, Sharp JH (1986) The effect of temperature upon the setting behavior of refractory calcium aluminate cements. Cem Concr Res 16:875-884 Bushnell-Watson SM, Sharp JH (1990a) Further studies of the effect of temperature upon the setting behavior of refractory calcium aluminate cements. Cem Concr Res 20(4):623-635 Bushnell-Watson SM, Sharp JH (1990b) On the cause of the anomalous setting behavior with respect to temperature of calcium aluminate cements. Cem Concr Res 20:677-686 Büssen W (1937) Die Struktur des Tetracalciumaluminatferrits. Fortschr Mineral 22:31-33 Büssen W, Eitel A (1936) Die Struktur des Pentacalciumtrialuminats. Z Kristallogr Kristallgeo Kristallphys Kristallchem 95:175-188 Buttner RH, Maslen EN (1992) Electron difference density and structural parameters in CaTiO3. Acta Crystallogr 48:644-649 Calleja J (1980) Calculation of hypothetically possible potential compositions of high alumina cements. Proceedings of the 7th International Congress on the Chemistry of Cement, Paris III:102-107 Capelle M, Capmas A, Chaix JM (1989) Characterization of phases in an aluminous cement by image analysis. Proceedings of the 11th International Conference on Cement Microscopy, New Orleans, USA, p 115-127 Capmas A, Metetrier-Sorrentino D, Damidot D (1990) Effect of temperature on setting time of calcium aluminate cements. In: Calcium Aluminate Cements. Mangabhai RJ (ed) London: E and FN Spon, p 65-80 Carlson ET (1964) J Res Natl Bureau Standards 68A:453 Carlson ET, Berman HA (1960) Some observations on the calcium aluminate carbonate hydrates. J Res Natl Bureau Standards 64A:333-341 Castellote M, Alonso C, Andrade C, Turrillas X, Campo J (2004) Composition and microstructural changes of cement pastes upon heating, as studied by neutron diffraction. Cem Concr Res 34:1633-1644 Causa MT, Zysler RD, Tovar M (1992) Magnetic properties of the CanFe2Tin-2O3n-1 perovskite related series: an EPR study. J Solid State Chem 98:23-32 Chan CJ, Kriven WM, Young JF (1992) Physical stabilization of the β→γ transformation in dicalcium silicate. J Am Ceram Soc 75(6):1621-1627 Chandran RG, Patil KC, Candrappa GT (1995) Combustion synthesis, characterization, sintering and microstructure of mullite-cordierite composites. J Mater Sci Lett 14:548-551 Chao GY, Baker J, Sabina AP, Roberts AC (1985) Doyleite, a new polymorph of Al(OH)3, and its relationship to bayerite, gibbsite and nordstrandite. Can Mineral 23:21-28 Chatterjee AK (1996) High belite cements – present status and future technological options. Part 1. Cem Concr Res 26(8):1213-1225 Chatterjee AK (2001) An update on the binary calcium aluminates appearing in aluminous cements. In: Calcium Aluminate Cements. Mangabhai RJ, Glasser FP (eds), IOM Communications Ltd., London, p 37-64 Chatterjee AK, Zhmoidin GI (1972) The phase equilibrium diagram of the system CaO-Al2O3-CaF2. J Mater Sci 7:93-97 Cheng GCH, Zussman J (1963) The crystal structure of anhydrite. Acta Crystallogr 16:767-769 Chesnokov B, Bashenova L, Byshmakin A, Kotl’rov V, Beloghyb E, Vilicov V, Klyuev A (1995) Caldekahydrate from dumps of the Chelyabinsk coal basin Ural. Mineral Sbornik 3 Chick LA, Pederson LR, Maupin GD, Bates JL, Thomas LE, Exarhos GJ (1990) Clycine-nitrate combustion synthesis of oxide ceramic powders. Mater Lett 10(1):6-12 Chou KS, Burnet G (1981a) Formation of calcium aluminates in the lime-sinter process. Part I: qualitative studies. Cem Concr Res 11:57-64 Chou KS, Burnet G (1981b) Formation of calcium aluminates in the lime-sinter process. Part II: kinetic studies. Cem Concr Res 11:167-174 Christensen AN, Scarlett NVY, Madsen IC, Jensen TR, Hanson JC (2003) Real time study of cement and clinker phases hydration. Dalton Trans 1529-1536 Chrysafi R, Perraki T, Kakali G (2007) Sol-gel preparation of 2CaO·SiO2. J Eur Ceram Soc 27:1707-1710 Chudak H-R, Bauer F, Oliew G, Geβner W (1982) Zum Einsatz der differentialkalorimetrischen Analyse (DCA) bei der Überwachung der Tonerdezement produktion. Silikattechnik 33(7):197-199 Chudak H-R, Bauer F, Oliew G, Möhmel S, Geβner W (1987) Hydratationseigenschaften handelsüblicher Tonerdezemente mit rund 70 % Al2O3. Silikattechnik 38(12):399-401 Cole WF, Lancucki CJ (1974) A refinement of the crystal structure of gypsum CaSO4(2H2O). Acta Crystallogr B 30:921-929 Collepardi M, Morosi S, Piccioli P (1995) The influence of pozzolanic materials on the mechanical stability of aluminous cement. Cem Concr Res 25:961-968 Colville AA (1970) The crystal structure of Ca2Fe2O5 and its relation to the nuclear electric field gradient at the iron sites. Acta Crystallogr 26:1469-1473 Colville AA, Geller S (1971) The crystal structure of brownmillerite, CaFeAlO5. Acta Crystallogr 27:2311-2315

Calcium Aluminate Cement

69

Colville AA, Geller S (1972) Crystal structure of Ca2Fe1.43Al0.57O5 and Ca2Fe1.28Al0.72O5. Acta Crystallogr 28:3196-3200 Cruickshank DWJ (1964) Refinements of structures containing bonds between Si, P. S or Cl and O or N. X. Beta-Ca2SiO4. Acta Crystallogr 17:685-686 Cuberos AJM, Torre ÀGDL, Martin-Sedeño MC, Moreno-Real L, Merlini M, Ordònez LM, Aranda MAG (2009) Phase development in conventional and active belite cement pastes by rietveld analysis and chemical constraints. Cem Concr Res 39:833-842 Currell BR, Grzeskowiak R, Midgley HG, Parsonage JR (1987) The acceleration and retardation of set high alumina cement by additives. Cem Concr Res 7:420-432 D’Ans J, Eick H (1953) Das System CaO-Al2O3-H2O bei 20 °C und das Erhärten der Tonerdezemente. ZementKalk-Gips 6:197-210 Damidot D, Rettel A, Capmas A (1996) Action of admixtures on fondue cement. I. Lithium and sodium salts compared. Adv Cem Res 8:111-119 Damidot D, Rettel A, Sorrentino D, Capmas A (1997) Action of admixtures on fondue cement. II. Addition of the effects of admixtures used to cancel the anomalous setting time observed with temperature. Adv Cem Res 9:127-133 Daniel P, Gillet ID, Gillet P, Mcmillan PF, Richet P (1995) An in-situ high temperature structural study of stable and metastable CaAl2Si2O8 polymorphs. Mineral Mag 59:25-33 Das S, Daspodda PK (1993) Some studies on the hydraulic activity of calcium aluminates. NML Technical J 35:29-38 [ref. CA 124/25756] Das SK, Mitra A, Poddar PKD (1996) Thermal analysis of hydrated calcium aluminates. J Therm Anal 47:765774 Dayal RR (1965) The System CaO-Al2O3-FeO-Fe2O3, Thesis, University of Aberdeen (1965), quoted in Robson – Chemistry of Calcium Aluminates and their relating Compounds – Proceedings of the 5th International Symposium on the Chemistry of Cement, Tokyo, Session 1-4, p 349-365 Dayal RR (1968) The System CaO-Al2O3-FeO-Fe2O3, PhD Thesis, Univ. of Aberdeen, (1965) Zitiert in ROBSON, 358-359 Deloye F-X, Lorang B, Montgomery R, Modercin I, Reymond A (1996) Comportment a long terme dune liason “Portland-Fondu”. Bull Lab Ponts Chaussées 202:51-59 Demichelis R, Noel Y, Ugliengo P, Zicovich-Wilson CM, Dovesis R (2011) Physicochemical features of aluminum hydroxides as modeled with the hybrid B3LYP functional and localized Basis functions. J Phys Chem C 115(27):13107-13134 Demichelis R, Noel Y, Zicovich-Wilson CM, Roetti C, Valenzano, Dovesi R (2008) Ab-initio quantum mechanical study of akdalaite (5Al2O3·H2O), Structure and vibrational spectrum. J Phys Conf Ser 117:012013-012013 Dittmeyer R, Keim W, Kreysa G, Oberholz A (2004) Nanomaterialien und Nanotechnologie. WILEY-VCH Verlag GmbH and Co. KGaA Band 2: Neue Technologie, Weinheim Dondony I, Buseck PR (2001) Polysomatism and modules of gehlenite composition and structure in the hanic phase (Ca5Al6MgSiO17). Phys Chem Miner 28:428-434 Douy A, Gervais M (2000) Crystallization of amorphous precursors in the calcia-alumina system: a differential scanning calorimetry study. J Am Ceram Soc 83(1):70-76 Dovàl M, Palou M, Mojumdar SC (2006) Hydration behavior of C2S and C2AS nanomaterials, synthesized by sol-gel method. J Therm Anal Calorim 86(3):595-599 Dukowicz A (1995) The dependence of the mineralogical composition of high alumina cements on raw material. Polish J Appl Chem 39(2):199-202 Dunster AM, Bigland DJ, Holton IR (2000) Rates of carbonation and reinforcement corrosion in high alumina cement concrete. Mag Concr Res 52:433-441 Dunster AM, Crammond NJ, Pettifer K, Rayment DL (1997) An assessment of ageing calcium aluminate concrete from marine structures. In: Mechanisms of chemical degradation of cement-based systems. Scrivener KL Young JF (Eds) F and FN Spon, London, p 263-271 Ecker M (1998) Diadochbeziehungen in Calciumaluminatferraten und deren Hydratations-produkte – Dissertation, Hallesches Jahrbuch für Geowissenschaften, Reihe B: Geologie, Paläontologie, Mineralogie, Beiheft 3, ISSN 1432-3702, 100 p Edmeades RM, Hewlett PC (1998) Cement admixtures. In: Lea’s Chemistry of Cement and Concrete. Hewitt PC (ed) Arnold Publishers Edmonds RN, Majumdar AJ (1988a) The hydration of 12CaO·7Al2O3 at different temperatures. Cem Concr Res 18:473-478 Edmonds RN, Majumdar AJ (1988b) The hydration of monocalcium aluminate at different temperatures. Cem Concr Res 18:311-320 Edmonds RN, Majumdar AJ (1989a) The hydration of mixtures of monocalcium aluminate and blast furnace slag. Cem Concr Res 19:779-782

70

Pöllmann

Edmonds RN, Majumdar AJ (1989b) The hydration of Secar 71 aluminous cement at different temperatures. Cem Concr Res 19:289-294 Emanuelson A, Hansen S, Henderson E, Landa-Cánovas A, Sjöstedt E (1997) Ferrite - microstructure in clinker and hydration of synthetic phases and sulfate resisting phases. Proceedings of the 10th International Congress on the Chemistry of Cement Göteborg, Schweden, 1i060 Enderle R, Götz-Neunhoeffer F, Neubauer J (2002) Der Einfluss des anorganischen additivs C12A7 auf das Hydratationsverhalten von Monocalciumaluminat. GDCH Monographie Bauchemie 25:179-181 Enderle R, Götz-Neunhoeffer F, Neubauer J (2004) Einfluss Ca-reicher Additive in Verbindung mit Li2CO3 auf das Hydratationsverhalten von Monocalciumaluminat. GDCH-Monographie Bauchemie 31:197-201 Evju C, Hansen S (2001) Expansive properties of ettringite in a mixture of calcium aluminate cement, Portland cement and β-calcium sulphate hemihydrate. Cem Concr Res 31:257-261 Feng QL, Glasser FP, Howie RA, Lachowski EE (1988) Chlorosilicate with the 12CaO·7Al2O3. Structure and its relationship to garnet. Acta Crystallogr C 44:589-592 Fentiman CH, Rashid S, Bayoux JP, Bonin A, Testud M (1990) The effect of curing conditions on the hydration and strength development in fondu: slag. In: Calcium Aluminate Cements. Proceedings International Symposium on Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London. p 272-281 Flörke OW (1952) Kristallographische und röntgenographische Untersuchungen im System CaSO4 CaSO4·2H2O. Neues Jahrb Mineral Abh 4:189-249 Foret J (1935) Recherches Sur Les Combinaisons Entre les sels de Calcium et les Aluminates de Calcium. Hermann and Cie, (eds) French Standard (1991) Use of “Fondu” Aluminous Cements In Concrete Structures. AFNOR P15-316 Freyer D, Voigt W (2003) Crystallization and phase stability of CaSO4 and CaSO4-based salts. Monatsh Chem 134:693-719 Fu Y, Ding J, Beaudoin JJ (1995) Mechanisms of strätlingite formation in high alumina cement-siliceous material systems. In: Symposium on Advances in Concrete Technology, p 461-471 Fujii K, Kondo W, Ueno H (1986) Kinetics of hydration of monocalcium aluminate. J Am Ceram 69:361-364 Fukuda K, Ando H (2002) Determination of the Pcmn/Ibm2 phase boundary at high temperatures in the system Ca2Fe2O5-Ca2Al2O5. J Am Ceram Soc 85(4):1300-1302 Füllmann T (1997) Quantitative Phasenanalyse von Tonerdezementen. Dissertation, Universität ErlangenNürnberg Gaki A, Chrysafi R, Kakali G (2007a) Chemical synthesis of hydraulic calcium aluminate compounds using the pechini technique. J Eur Ceram Soc 27:1781-1784 Gaki A, Perraki T, Kakali G (2007b) Wet chemical synthesis of monocalcium aluminate. J Eur Ceram Soc 27:1785-1789 Galasso FS (1969) Structure, Properties and Preparation of Perovskite-Type Compounds. Pergamon Press Ltd., Oxford Garcés P, Alcocel EG, Chinchón S, Andreu CG, Alcaide J (1997) Effect of curing temperature in some hydration characteristics of calcium aluminate cement compared with those of Portland cement. Cem Concr Res 27(9):1343-1355 Gartner EM, Young JF, Damidot DA, Jawed I (2002) Hydratation of Portland cement. In: Structure and Performance of Cements. 2nd Edition. Bensted J, Barnes P (eds) E and FN Spon Press, London, p 57-113 George CM (1976) The structural use of high alumina cement concrete. Revue des Matériaux et Construction. 201 George CM (1980a) Aluminous cements - a review of recent literature (1974 - 1979). Proceedings of the 7th International Congress on the Chemistry of Cement, Paris V1:4-26 George CM (1980b) Long term and accelerated tests of the resistance of cements to sea water, with special reference to aluminous cements. ACI SP65-19, 327-349 George CM (1983) Industrial aluminous cements. In: Structure and Performance of Cements. Barnes P (ed) Applied Science Publishers, London and New York, p 415-470 George CM (1990) Manufacture and performance of aluminous cement: a new perspective. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London, p 181-207 George CM (1997) Durability of CAC concrete: Understanding the evidence. In: Mechanisms of Chemical Degradation of Cement-based Systems. Scrivener KL, Young JF (eds) E and FN Spon, London, p 253-264 George CM, Cottin B (1982) Retard de prise du ciment fondu entre 26 et 30 °C. Calcium Aluminates Seminar. Turin, II-5, 131 George CM, Montgomery RGJ (1992) Calcium aluminate cement concrete: durability and conversion. A fresh look at an old subject. Materiales de Construccion (Madrid) 42(228):51-64 George CM, Racher RP (1996) Fondu and secar cements - refractory and non-refractory uses. Ind Min 224:2634 Ghosh NN, Pramanik P (1997) Synthesis of mullite powders using precipitated silica and tetraethoxy silane and aluminium-formate as precursor materials in aqueous medium. Bull Mater Sci 20(2):283-286

Calcium Aluminate Cement

71

Glasser FP, Marr J (1975) Quaternary phases in the system CaO-MgO-Al2O3-SiO2. Trans J Br Ceram Soc 74:113-119 Gloter A, Ingrin J, Bouchet D, Scrivener KL, Colliex C (2000) TEM-evidence of perovskite-brownmillerite coexistence on the Ca(AlxFe1-x)O2.5 system with minor amounts of titanium and silicon. Phys Chem Miner 27:504-513 Göbbels M (1997) Phasen der β-Tonerde- und Magnetoplumbit-Familien mit 2-wertigen charakteristischen Kationen (Ca, Sr, Ba) - Phasenbeziehungen, Kristallchemie und Struktur – Habilitation Thesis, Aachen Goktas AA, Weinberg MC (1991) Preparation and crystallization of sol-gel calcia-alumina compositions. J Am Ceram Soc 74(5):1066-1070 Goldschmidt HJ, Rait JR (1943) Silicates of the perovskite type of structure. Nature 152:356 Goodwin DW, Lindop AJ (1970) The crystal structure of CaO(Al2O3)2. Acta Crystallogr 26:1230-1238 Göske J, Pöllmann H, Dickschas D (1995) Hibschit, Perowskit und Ellestadit - Seltene Minerale in einem Xenolith im Basalt des Zeilbergs bei Maroldsweisach/Unterfranken. Der Aufschluss 46:69-74 Göske J, Pöllmann H, Pankau HG (2001) Hydration of high alumina cement investigations with low temperature SEM (cryo-technique). In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London, p 189-196 Götz-Neunhöffer F (1990) Untersuchungen zur Herstellung von Tonerdezement aus eisenreichem Bauxit nach dem Sinterverfahren im Drehrohrofen. Diploma thesis, Erlangen Götz-Neunhöffer F (1996) Synthese und mineralogische Charakterisierung von Calciumferrathydraten – Speicherminerale für die Anwendung in Reststoffdeponien. Dissertation, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen Götz-Neunhöffer F (2003) Der Einfluss Ca-reicher Verbindungen auf das Hydratationsverhalten von Monocalciumaluminat (CA). Tagungsbericht der 15. Internationalen Baustofftagung ibausil 1:467-478 Götz-Neunhöffer F (2003) Quantitative phase analysis of various Fe-rich calcium aluminate cement clinkers by selective dissolution and XRPD. Proceedings of the 11th International Congress on the Chemistry of Cement, Durban, p 125-133 Götz-Neunhöffer F (2005) Kinetik der Hydratation von Calciumaluminatzement mit Zusatzstoffen. ZementKalk-Gips International 58(4):65-72 Götz-Neunhöffer F (2006) Modelle zur Kinetik der Hydration von Calciumaluminatzementen mit Calciumsulfat aus kristallchemischer und mineralogischer Sicht. Habilitation, Friedrich-Alexander-Universität ErlangenNürnberg, Erlangen Götz-Neunhöffer F, Neubauer J (2002) Hydratationsverhalten von Monocalciumaluminat bei Zugabe von CaSO4·0,5H2O im Wärmeflusskalorimeter. GDCH-Monographie Bauchemie 25:20-26 Götz-Neunhöffer F, Neubauer J (2003) Effect in inorganic admixtures on the early hydration of Monocalciumaluminate (CA1-xFx). Proceedings of the 11th International Congress on the Chemistry of Cement, Durban, p 695-704 Götz-Neunhöffer F, Neubauer J (2004) Hydration of calcium aluminate in ettringite-based binder – investigation of model systems. Proceedings of the 26th International Conference on Cement Microscopy, San Antonio Texas, CD Guirado F, Gali S, Chinchon JS (1994) The crystallography of CA1-xFx using x-ray powder diffraction techniques. Cem Concr Res 24:923-930 Guirado F, Gali S, Chinchon JS (1998) Thermal decomposition of hydrated alumina cement (CAH10). Cem Concr Res 28:381-390 Gülgün MA, Popoola OO, Kriven WM (1984) Chemical synthesis and characterization of calcium aluminate powders. J Am Ceram Soc 77(2):531-539 Guse W, Saalfeld H (1990) X-ray characterization and structure refinement of a new cubic alumina phase (σ-Al2O3) with spinel-type structure. Neues Jahrb Mineral Monatsh 5:217-226 Hallstedt B (1990) Assessment of the CaO-Al2O3 system. J Am Ceram Soc 73(1):15-23 Halse Y, Pratt PL (1986) The development of microstructure in high alumina cement. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro 4:317-321 Hanic F, Handlovic M, Kapralic I (1980) The structure of a quaternary phase Ca20Al32-2vMgvSivO68. Acta Crystallogr B 36:2863-2869 Harchand KS, Chandra K (1983) A study of CaO(Fe2O3)1-x(Al2O3)x system. Cem Concr Res 13:465-469 Hauschild U (1963) Z Anorg Allg Chemie 324:15 Hazen RM (1988) Perowskite. Spektrum der Wissenschaft 8:42-50 Hentschel G (1964) Mayenit, 12CaO⋅7Al2O3, und Brownmillerit, 2CaO⋅(Al,Fe)2O3, zwei neue Minerale in den Kalksteineinschlüssen der Lava des Ettringer Bellerberges. Neues Jahrb Mineral Monatsh 1964:22-29 Hewlett PC, Desch CH (1956) The Chemistry of Cement and Concrete. 3rd Edition Hewlett PC (edr) (1998) Lea’s Chemistry of Cement and Concrete. 4th Edition. Arnold Publishers, London Höhl O, Weise C, Wentz B (1936) Der Tonerdezement. 4th Edition. Hochofenwerk Holterhoff AG (1993) Calcium aluminate cements. Am Ceram Soc Bull 72(6):89-90

72

Pöllmann

Hörkner W, Müller-Buschbaum HK (1976) Zur Kristallstruktur von CaO·Al2O3. J Inorg Nucl Chem 38:983-984 Hosono H, Abe Y (1987) Structure of C12A7. Inorg Chem 26:1192-1195 Houn S-H, Young JF (1999) Hydration kinetics and phase stability of dicalcium silicate synthesized by the Pechini process. J Am Ceram Soc 82:1681-1686 Huang X-H, Chang J (2007) Low-temperature synthesis of nanocrystalline beta-dicalcium silicate with high specific surface area. J Nanoparticle Res 9:1195-1200 Huckenholz HG, Ott WD (1978) Synthesis, Stability and aluminum-iron substitution in gehlenite-ferrigehlenite solid solutions. Neues Jarhb Mineral Monatsh 12:521-535 Imlach JA, Glasser FP (1971) Sub-solidus phase relations in the system CaO-Al2O3-Fe-Fe2O3. Trans Br Ceram Soc 70:227-234 Ingham J (2010) Geomaterials Under the Microscope, a Colour Guide. Manson Publishing, London Ishida H, Sasaki K, Mitsuda T (1992a) Highly reactive β-dicalcium silicate. I. Hydration behavior at room temperature. J Am Ceram Soc 75(2):353-358 Ishida H, Okada Y, Mitsuda T (1992b) Highly reactive β-dicalcium silicate. II. Hydration behavior at 25 °C followed by 29Si nuclear magnetic resonance. J Am Ceram Soc 75(2):359-363 Ishida H, Sasaki K, Okada Y, Mitsuda T (1992c) Highly reactive β-dicalcium silicate. III. Hydration behavior at 40-80 °C. J Am Ceram Soc 75(9):2541-2546 Ishida H, Sasaki K, Mizuno A, Okada Y, Mitsuda T (1992d) Highly reactive β-dicalcium silicate. IV. Ballmilling and static hydration at room temperature. J Am Ceram Soc 75(10):2779-2781 Isupov VP, Kotsupalo NP, Nemudry AP, Menzeres LT (1999) Aluminium hydroxide as selective sorbent of lithium salts from brines and technical solutions. In: Adsorption and its Applications in Industry and Environmental Protection. Vol. I. Applications in Industry 120(1):621-652 Ito J (1976) High Temperature solvent growth of anorthite on the join CaAl2Si2O8-SiO2. Contrib Mineral Petrol 59:187-194 Ito S, Ikai K, Suzuki K, Inagaki M (1975) Metastable orthorhombic CaO·Al2O3. J Am Ceram Soc 58(1-2):79-80 Iyatomi N, Aoki Y (1992) Crystallization behavior of P21 and C2 phases of anorthite composition. Neues Jahrb Mineral Monatsh 4:177-192 Janakova S, Salavcova L, Renaudin G, Filinchuk Y, Boyer D, Boutinaud P (2007) Preparation and structural investigations of sol-gel derived Eu3+-doped CaAl2O4. J Phys Chem Solids 68:1147-1151 Jeevaratnam J, Glasser FP, Glasser LS (1964) Anion substitution and structure of 12CaO·7Al2O3. J Am Ceram Soc 47:105-106 Jelenic I, Bezjak A (1982) Electron diffraction evidence for superstructures in δ-modification of dicalcium silicate. Cem Concr Res 12:785-788 Johnson RE, Muan A (1965) Phase equilibria in the system CaO-MgO-iron oxide at 1500 °C. J Am Ceram Soc 48(7):359-364 Jongejan A, Wilkins A (1963) Nat. Res. Council, Mineral Science Division, Internat. Report, cited in Sorrentino and Glasser Jost KH, Ziemer B, Seydel R (1977) Redetermination of the structure of β-dicalcium silicate. Acta Crystallogr 33:1696-1700 Kahlenberg V (2001) On the Al/Fe substitution in iron doped monocalcium aluminate – the crystal structure of CaAl1.8Fe0.2O4. Eur J Mineral 13:403-410 Kahlenberg V, Fischer RX, Shaw CSJ (2000a) High-pressure Ca4Al6O13. An example of a calcium aluminate with three different types of coordination polyhedra for aluminum. Am Mineral 85:1492-1496 Kahlenberg V, Fischer RX, Shaw CSJ (2000b) Rietveld analysis of dicalcium aluminate (Ca2Al2O5). A new high pressure phase with Brownmillerite structure. Am Mineral 85:1061-1065 Kaprálic I, Hanic F (1980) 23-studies of the system CaO-Al2O3-MgO-SiO2 in relation of the quaternary phase Q. Trans J Br Ceram Soc 79:128-133 Kaprálic I, Hanic F (1989) Phase relations in the subsystem C4A3S-CSH2-CH-H2O of the system CaO-Al2O3Cs-H2O referred to hydration of sulphoaluminate cement. Cem Concr Res 19:89-102 Kaprálic I, Stevula L, Hanic F (1989) Hydration and hydraulic properties of the Q-Phase in the system CaOAl2O3-MgO-H2O referred to high alumina cements. Cem Concr Res 19:519-526 Kay HF, Bailey PC (1957) Structure and properties of CaTiO3. Acta Crystallogr 10:219-226 Keil F (1971) Zement - Herstellung und Eigenschaften. Springer Verlag, Berlin-Heidelberg Kimata M, Li N (1982) The structural property of synthetic gehlenite, Ca2Al2SiO7. Neues Jahrb Mineral 143:254-267 Kirfel A, Will G (1980) Charge density in anhydrite from x-ray and neutron diffraction measurements. Acta Crystallogr B 36:2881-2890 Kirkpatrick RJ, Steele IM (1973) Hexagonal CaAl2SiO6. A new synthetic phase. Am Mineral 58:945-946 Kiskyras D (1960) Die mineralogische Zusammensetzung der griechischen Bauxite in Abhängigkeit von der Tektonik. Neues Jahrb Mineral Abh 94:662-680

Calcium Aluminate Cement

73

Klemm WA, Skalny J (1977) Selective dissolution of clinker minerals and its applications. – XII. Conf. on Silicate Science, Budapest, Hungary Knöfel D, Bier T (1997) Zement-Kalk-Gips 5(8):454-462 Koppmans HJA, Van Der Velde GMH, Gellings PJ (1983) Powder neutron diffraction study of the perovskite CaTiO3 and CaZrO3. Acta Crystallogr C 39:1323-1325 Kühl H (1958) Zement-Chemie. Volume 2. VEB Verlag Technik, Berlin Kühnel RA, Van der Gaast SJ (1993) Humidity controlled diffractometry and its applications. Adv X-Ray Anal 36:439-449 Kumar R, Vishwamittar, Chandra K (1985) Low temperature Mössbauer spectroscopic study of CaO(Fe2O3)1−x(Al2O3)x system. Cem Concr Res 15:520-524 Kumarathasan P, McCarthy GJ, Hassett DJ, Pflughoeft-Hassett DF (1990) Oxyanion substituted ettringites: Synthesis and characterization; and their potential role in immobilization of As, B, Cr, Se and V. Mater Res Soc Symp Proc 178:83-104 Kurdowski W (1980) Expansive cements. 7th International Congress on the Chemistry of Cement, Paris I:V2/1-V2/11 Kurdowski W, George CM, Sorrentino FP (1986) Special cements. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro, Brazil 1:292-318 Kurdowski W, Taczuk B (1990) Behavior of high alumina cement in chloride solutions. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London p 222-229 Kuzel H-J (1968) Neues Jahrb Mineral Monatsh 87-96 Kuzel H-J (1984) Ein leistungsfähiges Wärmeleitungskalorimeter. TIZ Fachberichte 108(1):46-51 Kuzel H-J (1987) Hydratationswärmen von α-Ca1ciumsulfaten. Neues Jahrb Mineral Abh 156:155-174 Kuzel H-J (1992) A study of the hydration of calcium aluminate cements with calorimetry, microscopy and x-ray techniques. Proceedings of the 14th International Conference on Cement Microscopy, Costa Mesa, USA, p 64-73 Kuzel H-J, Baier H (1996) Hydration of calcium aluminate cements in the presence of calcium carbonate. Eur J Mineral 8:129-141 Kuzel H-J, Hauner M (1987) Chemische und kristallographische Eigenschaften von CaSO4-Halbhydrat und Anhydrit III. Zement-Kalk-Gips 40:628 Kuzmenkov MI, Burak GA, Mazurenko VD (1991) Tonerdezement unter Verwendung von karbonatischen Reststoffen (Russisch). Tsement (3-4):20-21 Kvyatkovskii OV, Sokolov AN, Fleer SA, Ashkinadze GS, Osipo-Va LY, Orekhov PD, Krasnitskaya LA (1987) Burning clinker for high alumina cement in a rotary kiln. Refractories 28(5/6):333-334 Lacerda M, Irvine JTS, Glasser FP, West AR (1988) Nature 332:525 Lachowski EE, Glasser FP (1973) Synthetic gehlenite. Mineral Mag 39:412-419 Lager GA, Armbruster Th, Rotella FJ, Jorgensen JD and Hinks DG (1984) A crystallographic study of the lowtemperature dehydration products of gypsum. Am Mineral 69:910-918 Lagerblad B, Vogt C (2009) Fillers and ultrafine fillers for low cement concrete to reduce the carbon footprint. Tagung Bauchemie der GDCH-Fachgruppe Bauchemie, Freiberg/Deutschland, 1-9 Larsen ES, Foshag WF (1921) Merwinite, a new calcium magnesium orthosilicate from Crestmore, California. Am Mineral 6:143-148 Lee S-J, Kim G-S (2002) Characteristics and densification behavior of anorthite powder synthesized by a solution process employing a polymer carrier. J Ceram Processing Res 3(3):136-140 Lee S-J, Lee M-H (2003) Effect of degree of polymerization on synthesis of anorthite-type powder by wetchemical method employing PVA. J Mater Sci Lett 22:1291-1293 Lehmann H, Wuhrer J, Lahl W (1958) Das Brennverhalten von Kalkstein aus verschiedenen geologischen Formationen. TIZ International 82(20):434-440 Lucco BM, Brisi C (1986) Stability of the compound C5A3 in the system CaO-Al2O3. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro, Brazil IV:371-375 Ma C, Conolly HC, Beckett JR, Tschauner O, Rossmann GR, Kampf AR, Zega TJ, Sweeney-Smith SA, Schrader DL (2011) Brearleyite, a new alteration mineral from the NWA 1934 meteorite. Am Mineral 96:1199-1206 Ma C, Kampf AR, Conolly HC, Beckett JR, Rossman GR, Smith SA, Schrader DL (2010) Krotite, IMA 2010038, CNMNC newsletter, October 2010, page 861. Mineral Mag 74:859-862 Ma W, Brown PW (1994) Hydration of sodium phosphate modified high alumina cement. J Mater Res 9:12911297 Macdargent S (1992) Hydration and strengths of blends CAC – silica fume and stability of hydrates. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 4:651-657 Majumdar AJ (1964) The quaternary phase in high-alumina cement. Trans Br Ceram Soc 63(7):347-64 Majumdar AJ, Singh B, Edmonds RN (1990) Hydration of mixture of “Ciment Fondu” aluminous cement and granulated blastfurnace slag. Cem Concr Res 20:197-208

74

Pöllmann

Marciano E, Da Cunha Munos FA, Splettstoser J, Placido WF (1989) Characterization and quantitative determination of calcium aluminate clinker phases through reflected light microscopy. Proceedings of the 11th International Conference on Cement Microscopy, New Orleans, USA, p 1-15 Marinho MB, Glasser FP (1984) Polymorphism and phase changes in the ferrite phase of cement induced by titanium substitution. Cem Concr Res 14:360-368 Mascolo G (1986) Thermal stability of lithium aluminium hydroxyl salts. Thermochim Acta 102:67-73 Matusinovic T, Curlin D (1993) Lithium salts as set accelerators for high alumina cement. Cem Concr Res 23:885-895 Matusinovic T, Vrbos N (1992) Lithium salts as set accelerators for high alumina cement. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi III:425-430 Matusinovic T, Vrbos N (1993) Alkali metal salts as set accelerators for high alumina cements. Cem Concr Res 23:177-186 Mayerhofer W (1996) Rietveld Analyse der Ferratphase. Universität Erlangen-Nürnberg Mehta PK, Lesnikoff G (1971) Conversion of CaO·Al2O3·10H2O to 3CaO·Al2O3·6H2O. J Am Ceram Soc 56(4):210-212 Meier R, Pöllmann H (ICDD 45-0565) International Centre for Diffraction Data, Swarthmore, File-No.: 450565 Menetrier-Sorrentino D, George CM, Sorrentino FP (1986) The setting and hardening characteristics of calcium aluminate cements: studies of the system C12A7-C3A and C12A7-CA. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro 4:339- 343 Meyer HW, Pöllmann H, Kuzel H-J (1993) X-ray measurements of cement hydration products. Mater Sci Forum 133-136:645-650 Midgley HG (1967) The mineralogy of set high alumina cement. Trans Br Ceram Soc 66(4):161-187 Midgley HG (1968) The composition and possible structure of the quaternary phase in high-alumina cement and its relation to other phases in the system CaO-MgO-Al2O3. Trans Br Ceram Soc 67(1):1-14 Midgley HG (1976) Quantitative determination of phases in high alumina cement clinkers by x-ray diffraction. Cem Concr Res 6:217-224 Midgley HG (1979) The composition of pleochroite in high alumina cement clinker. Cem Concr Res 9:623-630 Midgley HG (1980a) The chemical resistance of high alumina cement concrete. Proceedings of the 7th International Congress on the Chemistry of Cement 3:85-87 Midgley HG (1980b) The relationship between cement clinker composition and strength recovery of hydrating high alumina cement during conversion. Proceedings of the 7th International Congress on the Chemistry of Cement 3:68-70 Midgley HG (1990) High alumina cement in construction - a future based on experience. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London, p 1-14 Midgley HG, Midgley A (1975) The conversion of HAC. Mag Concr Res 127:59-75 Midgley HG, Pettifer K (1972) Electron optical study of hydrated high-alumina cement pastes. Trans J Br Ceram Soc 71(2):55-59 Midgley HG, Rao PB (1978) Formation of strätlingite 2CaO·SiO2·Al2O3·8H2O in relation to the hydration of high alumina cement. Cem Concr Res 8:169-72 Midgley HG, Ryder JF (1977) The relationship between mineral composition and strength development of high alumina cement. Cem Concr Res 7:669-672 Mikouchi T, Zolensky M, Ivanova M, Tachikawa O, Komatsu M, Le L, Gounelle M (2009) Dmitryivanovite: A new high-pressure calcium aluminum oxide from the Northwest frica 470 CH3 chondrite characterized using electron backscatter diffraction analysis. Am Mineral 94:746-750 Mino I, Sakai E, Nishioka A (1991) The hydration and calcinations mechanism of calcium aluminate-based ultra-high strength cement with calcium sulfoaluminate compound. Mater Res Soc Symp Proc 179:193202 Mohan L (1991) Advances in some special and newer cements. In: Cement and Concrete Science and Technology. Gosh SN (ed) ABI Books, New Delhi I:253-312 Moore AE, Taylor HFW (1970) Crystal structure of Ettringite. Acta Crystallogr B 26:386-393 Moore PB, Araki T (1972) Atomic arrangement of merwinite, Ca3Mg(SiO4)2, an unusual dense packed structure of geophysical interest. Am Mineral 57:1355-1374 Motzet H (1998) Entwicklung eines Sinterverfahrens zur Herstellung von eisenreichem Tonerdezement im Drehrohrofen – Dissertation Universität Halle/Saale Motzet H, Pöllmann H (1998) Quantitative phase analysis of high alumina cements. Proceedings of the 20th International Conference on Cement Microscopy, Guadalajara, Mexico CD Muhamad MN, Barnes P, Fentimann CH, Hausermann D, Pöllmann H, Rashid S (1993) A time-resolved synchrotron energy dispersive diffraction study of the dynamic aspects of the synthesis of Ettringite during minepacking. Cem Concr Res 23:267-272 Müller D (1992) MISSING

Calcium Aluminate Cement

75

Müller D, Rettel A, Gessner W, Bayoux JP, Capmas A (1992) Progress in the 27Al MAS NMR spectroscopy: monitoring the hydration of calcium aluminate cements. In: Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi 6:148-154 Mumme WG, Hill RJ, Bushnell-Wye G, Segnit ER (1995) Rietveld crystal structure refinements, crystal chemistry and calculated powder diffraction data for the polymorphs of dicalcium silicate and related phases. Neues Jahrb Mineral Abh 169(1):35-68 Murat M, Sadok EH (1990) Role of foreign cations in solution in the hydration kinetics of high alumina cement. In: Calcium Aluminate Cements. Mangabhai RJ (ed) Chapman and Hall, London, p 155-166 Muzhen S, Kurdowski W, Sorrentino F (1992) Development in non-Portland cements. Proceedings of the 9th International Congress on the Chemistry of Cement, New Delhi I:317-354 Negro A, Stafferi L (1979) The hydration of calcium ferrites and calcium aluminoferrites. Zement-Kalk-Gips 2:83-88 (English Version) Neubauer J (1997) Application of Rietveld quantitative x-ray diffraction analysis on technically produced OPC clinkers, OPC’s and OPC/HAC blends. In: Tagungsbericht. 13th International Baustofftagung ibausil, Weimar, Volume 1, p 0745-0752 Neubauer J, Goetz-Neunhoeffer F, Lindner I (2002) Untersuchung des Hydratationsverhaltens synthetischer Kalziumaluminatferrate bei Anwesenheit von C3A, CA und Kalziumsulfat. GDCH-Monographie Bauchemie 25:13-19 Neubauer J, Goetz-Neunhoeffer F, Lindner I (2003) Investigation of hydration behavior of ferrite phase C6AxF3-x with different Al3+-content in mixes with C3A and gypsum using a revised highly efficient isothermal calorimeter. Proceedings of the 11th International Congress on the Chemistry of Cement Durban/South Africa Neumann A, Neubauer J, Goetz-Neunhoeffer F (2001) Hydratationsversuche an eisendotiertem Monocalciumaluminat, Bauchemie von der Forschung bis zur Praxis. GDCH-Monographie 24:141-145 Neville A (1975) High Alumina Cement Concrete. The Construction Press, Lancaster Nikushchenko VM, Khotimchenko VS, Rumyantsev PF, Kalinin AI (1973) Determination of the standard free energies of formation of calcium hydroaluminates. Cem Concr Res 3:625-32 Nilforoushan MR, Sharp JH (1995) The effect of additions of alkaline-earth metal chlorides on the setting behavior of a refractory calcium aluminate cement. Cem Concr Res 25(7):1523-1534 Nowack N (1994) CaO-activities and structure of CaO-Al2O3-melts. Steel Res 65(8):320-321 Nurse RW, Midgley HG (1973) The system CaO-Fe2O3-Al2O3-SiO2. J Iron Steel Inst 174:121-128 Oberste-Padtberg R (1995) Herstellung von eisenhaltigem Tonerdeschmelzzement. Offenlegungsschrift DE 43 30 596 A1 (patent application) Oberste-Padtberg R (1996) Hydraulisches Bindemittel und seine Verwendung. Patentschrift DE 43 42 407 C2 Obst K-H, Stradtmann J (1969) Das System Kalziumoxyd-Eisen(II)-oxyd als Grundlage für Untersuchungen zur Kalkauflösung bei der Stahlerzeugung. Archiv für das Eisenhüttenwesen 8(40):615-617 Oetzel M (1999) Untersuchungen der Phasenbeziehungen im System CaSO4-H2O – Dissertation Aachen Okada Y, Ishida H, Sasaki K, Young JF, Mitsuda T (1994) Characterization of C-S-H from highly reactive β-dicalcium silicate prepared from hillebrandite. J Am Ceram Soc 75(4):1313-1318 Omar AA, Salman SM, Mahmoud MY (1986) Phase relations in the diopside-anorthite-åkermanite system. Ceram Int 12:53-59 Onuma K, Yagi K (1967) The system diopside-åkermanite-nepheline. Am Mineral 52:227-243 Osborn EF, Muan A (1960) Phase equilibrium diagrams of oxide systems. The System CaO-Al2O3-SiO2. In: Phase Diagrams for Ceramists. Volume 1. Levin EM, Robbins CM, McMurdie HF (eds) The American Ceramic Society, p 219 Osborn G, Brecem J (1994) A rapid hardening cement based on high alumina cement. Proc Institution Civil Engineers Structures Buildings 104:93-100 Osborn GJ, Singh B (1995) The durability of concrete made with blends of high-alumina cement and ground granulated blast-furnace slag. In: Proceedings of the 5th CANMET/ACI International Conference on Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Milwaukee. ACI SP 153-47, 2:885-909 Packter A, Zaidi SA (1983) The heterogenous reaction of alumina with calcium-oxide-mineraliser system to form monocalciumaluminate. Analysis of Some Relevant Calcium Oxide-Metal Salt Phase Diagrams. Silicate Industriels 7-8:139-146 Page CH, Thombare CH, Kamat RD, Chatterjee AK (1991) Development of sol-gel technology for cement manufacture. Br Ceram Trans J 16:643-660 Palou M, Dovàl M, Dràbik M (2007) Applications of sol-gel technique to synthesize inorganic binder materials with upgraded hydration characteristics. Proceedings of the 12th International Congress on the Chemistry of Cement, Montreal/Canada Parker KM, Sharp JH (1982) Refractory calcium aluminate cements. Trans J Brit Ceram Soc 81(2):35-42 Parker TW (1954) The constitution of aluminous cement. In: Proceedings of the 3rd International Symposium on the Chemistry of Cement, London, 1952. Cement and Concrete Assoc., London, p 485-529

76

Pöllmann

Passaglia E, Galli E (1977) Vertumnite, a new natural silicate. Tscherm Mineral Petrol Mitt 24:57-66 Pati RK, Panda AB, Pramanik P (2002) Preparation of nanocrystalline calcium aluminate powders. J Mater Synth Process 10(4):157-161 Pechini MP (1967) US-Patent No: 3 33 0697, USA Pekarek D (1988) Rapid semiquantitative X-ray diffraction analysis of calcium aluminate cements. Silikat 32:267-271, (in Czech) Pérez Méndez M, Triviño Vázquez F (1984) Study of the strength developed by stable carbonated phases in high alumina cement. Cem Concr Res 14:161-169 Petrykin V, Kakihana M (2004) Chemistry and applications of polymeric gel precursors. In: Handbook of SolGel Science and Technology: Processing Characterisation and Applications. Volume I: Sol-Gel Processing. Sakka S (ed) Kluwer Academic Publishers, Norwell (USA) p 77-104 Petzold A, Riedel K (1976) Einfluss stark reduzierender Schmelzbedingungen auf Tonerdeschmelzzement. Baustoffindustrie A 1:12-14 Phillips B, Muan A (1959) Phase equilibria in the system CaO-iron oxide-SiO2 in air. J Am Ceram Soc 42(9):413-423 Piasta J, Sawicz Z, Piasta WG (1989) Durability of high alumina cement pastes with mineral additions in water sulfate environment. Cem Concr Res 19:103-113 Podwòrny J, Wala T, Sawkòw J (2007) Microstructural studies on the Q-phase in alumina cement clinker. Solid State Phenomena 130:225-228 Pöllmann H (1987) SEM-, X-ray and thermoanalytical studies on hydration products of tricalciumaluminate in the presence of sulfate, carbonate and hydroxide anions. Proceedings of the 9th International Conference on Cement Microscopy, Reno, USA, p 427-445 Pöllmann H (1989a) Mineralogisch-Kristallographische Untersuchungen an Hydratations-produkten hydraulischer Bindemittel, (Bd. I, 285 S.), (Bd. II 236 S.). Habil.-Schrift Pöllmann H (1989b) Study of the hydration mechanism and formation of new hydrates applying organic additives to the aluminate phase of cement. Proceedings of the 11th International Conference on Cement Microscopy, New Orleans, USA, p 287-305 Pöllmann H (1990) Study on the influence of citric acid on retardation and formation of new hydrates in cement. Proceedings of the 12th International Conference on Cement Microscopy, Vancouver, Canada, p 303-322 Pöllmann H (2001) Mineralogy and crystal chemistry of calcium aluminate cement. In: Calcium Aluminate Cements. Manghabai RJ, Glasser FP (eds) Edinburgh/UK, IOM Communications Ltd., p 79-119 Pöllmann H (2002) Composition of cement phases. In: Structure and Performance of Cements. 2nd Edition. Bensted J, Barnes P (eds) Spon Press, London, p 25-56 Pöllmann H (2006) Syntheses, properties and solid solution of ternary lamellar calcium aluminate hydroxi salts (AFm-phases) containing SO42–, CO32– and OH–. Neues Jarhb Mineral Abh 182/2:173-181 Pöllmann H (2007) Immobilisierung von Schadstoffen durch Speichermineralbildung. Shaker Verlag, Aachen Pöllmann H (2010a) Mineralisation of Wastes and Industrial Residues. Shaker Verlag, Aachen Pöllmann H (2010b) Synthesis and Characterization of Ettringites. 59th Annual Conference on Application of X-ray Analysis, Denver, 2.-6.8 Pöllmann H (2010c) The both sides of ettringites (AFt) - a mineral with different properties. Proceedings of the 32nd International Conference on Cement Microscopy, New Orleans, Louisiana, p 1-58, Pöllmann H (2011) AFm-phase in cementitious materials. Proceedings of the 33nd International Conference on Cement Microscopy, San Francisco California, USA, p 94 Pöllmann H and Oberste-Padtberg R (2001) Manganese in high alumina cement. Manghabai RJ, Glasser FP (eds) Edinburgh/UK, IOM Communications, p 139-148 Pöllmann H, Stöber S (1997) Hydration characteristics and new hydrates using organic additives (carboxylates and sulfonates). Proceedings of the 10th International Congress on the Chemistry of Cement, Göteborg, 3iii032 Pöllmann H, Wenda R, Fylak M, Göske J (2008) Cryo-SEM-FEG investigations on calcium aluminate cements. In: Calcium Aluminate Cements. Avignon/France, p 123-138 CD Pöllmann H, Winkler N, Oberste-Padtberg R, Meyer R, Goeske J, Raab B (2006) Quantitative mineralogical chemical and application investigations on some High Alumina Cements from different sources. Proceedings of the 28th International Conference on Cement Microscopy, Denver (USA), p 150-187 Ponomarev VI, Kheiker DM, Belov NV (1970, 1980) Crystal structure of calcium dialuminate. ICSD Collection Code 16191, Recorded Pozun A (1988) Betriebserfahrungen bei der Herstellung von Tonerdeschmelzzement. Zement-Kalk-Gips 41(7):344-346 Prince AT (1954) System CaO-MgO-Al2O3-SiO2, Phase Diagrams for Ceramists, No. 882, p 291 Puertas F, Blanco-Varela MT, Dominguez R (1990) Characterization of CaAl2MnO5. A comparative study between CaAlMnO5 and Ca2AlFeO5. Cem Concr Res 20:429-438

Calcium Aluminate Cement

77

Puertas F, Trivino F (1985) Examination by infra-red spectroscopy for the polymorphs of dicalcium silicate. Cem Concr Res 15:127-133 Puri A, Teoreanu I, Vrabete D (1985) Kinetik der Bildung von Calciumdialuminat im Temperaturbereich von 1200 bis 1-400 °C. TIZ-Fachberichte 109(9):663-665 Qijun Y, Xiuji F, Shuichi S (1997) Hydration characteristics of 11CaO·7Al2O3·CaF2. Proceedings of the 10th International Congress on the Chemistry of Cement, Gothenburg/Sweden, 2ii043-8 Quillin KC (2001) Performance of belite-sulfoaluminate cements. Cem Concr Res 31:1341-1349 Quillin KC, Majumdar AJ (1994) Phase equilibria in the CaO-Al2O3-SiO2-H2O system at 5 °C, 20 °C and 38 °C. Adv Cem Res 6:47-56 Raab B (2010) Synthese und Charakterisiserung nanoskaliger hydraulisch hochreaktiver Phasen des Tonerde und Portlandzementes. Dissertation Univ. Halle Raab B, Pöllmann H (2009a) Investigations of high reactive pure cement phases by low-temperature-synthesis methods and their hydration behavior. Proceedings of the 31st International Conference on Cement Microscopy, St. Petersburg, 17 pages, on CD, April Raab B, Pöllmann H (2009b) Metal formates used for production of reactive phases by spray drying. Z Kristallogr Suppl 29:100 Raab B, Pöllmann H (2009c) Synthese hochreaktiver Zementphasen und deren Hydratation. Tagung Bauchemie in Freiberg, October 2009. GDCH Monographie 41:233-236 Raab B, Pöllmann H (2009d) Synthese und Hydratation von Mayenit 12CaO·7Al2O3. Hall. Jahrbuch für Geowissenschaften, Halle 31:189 Raab B, Pöllmann H (2009e) Synthese von Monocalciumaluminat mittels Sprühtrocknung und Zersetzung von Formiaten. Hall. Jahrbuch für Geowissenschaften, Halle 31:190 Raab B, Pöllmann H (2010a) Hochreaktive Zementphasen. 9. Hallesche Diskussions- u. Arbeitstagung MTK CD Raab B, Pöllmann H (2010b) Investigations of the hydration behavior of high reactive pure cement phases. Proceedings of the 32nd International Conference on Cement Microscopy p 1-20 CD, New Orleans, Louisiana Raab B, Pöllmann H (2010c) Synthese und Hydratation hochreaktiver Brownmillerite. GDCH Tagung Bauchemie, p 214-221, Dortmund 7.-8.Oktober Radler MJ, Cohen JB, Faber JJR (1990) Point defect clusters in wuestite. J Phys Chem Solids 51:217-228 Raina SJ et al. (1992) Barium aluminate – its synthesis and formation kinetics. Proceedings of the 8th International Congress on the Chemistry of Cement, New Delhi, 3:245-249 Ramachandran VS, Beaudoin JJ (1976) Significance of w/c and temperature on the physico-mechanical characteristics of hydrating 4CaO·Al2O3·Fe2O3. J Mater Sci 11:1893-910 Ramachandran VS, Feldman RF (1973) Hydration characteristics of monocalcium aluminate at low water/solid ratio. Cem Concr Res 3:729-750 Rankin GA, Wright F E (1915) The ternary system CaO-Al2O3-SiO2. Am J Sci 39(229):1-79 Rao PB (1980) Chemistry of arresting strength retrogression in structural high alumina cements. Proceedings 7th International Congress on the Chemistry of Cement, Paris, 3(V):51-56 Rashid S, Barnes P, Turrillas X (1994) Conversion of calcium aluminate cement hydrates re-examined with synchrotron energy-dispersive diffraction. J Mater Sci Lett 13:1232-1234 Rayment DL, Majumdar AJ (1994) Microanalysis of high alumina cement clinker and hydrated HAC/slag mixtures. Cem Concr Res 24(2):335-342 Redhammer GJ, Trippelt G, Roth G, Amthauer G (2004) Structural variations in the brownmillerite series Ca2(Fe2-xAlx)O5; single-crystal x-ray diffraction at 25 °C and high-temperature x-ray powder diffraction (25 °C ≤ T ≤ 1000 °C). Am Mineral 89:405-420 Redler L (1991) Quantitative x-ray diffraction analysis of high alumina cements. Cem Concr Res 21:873-884 Remy C, Guyot F, Madon M (1995) High pressure polymorphism of dicalcium silicate Ca2SiO4. A transmission electron microscopy study. Phys Chem Miner 22:419-427 Renaudin G, Francois M, Evrard O (1999a) Order and disorder in the lamellar hydrated tetracalcium monocarboaluminate compound. Cem Concr Res 29:63-69 Renaudin G, Kubel F, Rivera J-P, Francois M (1999b) Structural phase transitino and high temperature phase structure of Friedels salt. Cem Concr Res 29:1937-1942 Rengade E (1937) International Association for Testing and Materials. Congress Book, London Rengade E, Lhopitallier P, Durand de Fontmagne P (1936) Recherche sur les causes de certains phénoménes daltération des bétons de ciment alumineux. Revue des Matériaux et Construction 22 Rettel A, Damidot D, Scrivener K (1997) Study of the hydration of mixtures of CA and α-Al2O3 at different temperatures. Proceedings 10th International Congress on the Chemistry of Cement, Göteborg, 2ii026 Rettel A, Gessner W (1988) Zur Frage der direkten Bildung von 3CaO·Al2O3·6H2O bei der Hydratation von CaO·Al2O3. Silikattechnik Jhrg 39(5):175-176

78

Pöllmann

Rettel A, Geßner W, Müller D, Scheler G (1985) On the hydration of CaAl2O4 at various temperatures. Br Ceram Trans J 84:25-28 Rettel A, Gessner W, Oliew G, Müller D (1983) Zur Abhängigkeit der Hydratation des CaO·Al2O3 von der Temperatur. Z Anorg Allg Chem 500:89-96 Rettel A, Seydel R, Gessner W (1993) Investigations on the influence of alumina on the hydration of monocalcium aluminate at different termperatures. Cem Concr Res 23(5):1056-64 Revay M (1992) A summary of experiences on estimating the expected changes of strength of concrete made with high alumina cement. Proceedings 8th International Congress on the Chemistry of Cement, Rio de Janeiro 1:351-356 Richard N, Lequeux N, Boch P (1995) Local environment of Al and Ca in CAH10 and C2AH8 by X-ray absorption spectroscopy. Eur J Inorg Chem 32(7/8):649-662 Richardson IG, Groves GW (1990) The microstructure of blast furnace slag/high alumina cement pastes. Proceedings International Symposium on Calcium Aluminate Cements, London, p 282-290 Rietveld HM (1969) A profile refinement method for nuclear and magnetic structures. J Appl Crystallogr 2:6571 Rinaldi R, Sacerdoti M, Passaglia E (1990) Straetlingite: crystal structure, chemistry, and a reexamination of its polytype vertumnite. Eur J Mineral 2:841-849 Risch A, Pöllmann H, Ecker M (1997) Application of Portland cement and high alumina cement for immobilization/solidification of a waste model composition. Proceedings of the 10th International Congress on the Chemistry of Cement, Göteborg, 4iv044 Roberts MH (1957) New calcium aluminate hydrates. J Appl Chem 7:543-546 Robson TD (1952) The characteristics and applications of mixtures of Portland and high-alumina cements. Chem Industry 1:2-7 Robson TD (1962) High Alumina Cement and Concrete. London: John Wiley and Sons Robson TD (1964) Aluminous cement and refractory castables. In: The Chemistry of Cements. Vol. 2. Taylor HFW (ed) Academic Press, London Robson TD (1968) Chemistry of calcium aluminates and their relating compounds. In: Proceedings of the 5th International Symposium on the Chemistry of Cement, Tokyo, Session 1-4, p 349-365 Rodger SA, Double DD (1984) The chemistry of hydration of high alumina cements in the presence of accelerating and retarding admixtures. Cem Concr Res 14:73-82 Rodrigues FA (2003a) Low-temperature synthesis of cements form rice hull ash. Cem Concr Res 33:1525-1529 Rodrigues FA (2003b) Synthesis of chemically and structurally modified dicalcium silicate. Cem Concr Res 33:823-827 Rodriguez-Carvajal J, Vallet-Regi M, Gonzalez-Calbet JM (1989) Perovskit threefold superlattices: a structure determination of the A3M3O8 phase. Mater Res Bull 24:423-430 Rogers DE, Aldridge LP (1977) Hydrates of calcium ferrites and calcium aluminoferrites. Cem Concr Res 7:399-419 Rogers DE, Aldridge LP (1983) Hydrates of calcium ferrite and calcium aluminoferrites. Cem Concr Res 13:696-702 Romero L et al. (1997) Activation of blast furnace slag/high alumina cement pastes: mechanical and microstructural evolution. In: Proceedings of the 10th International Congress on the Chemistry of Cement Göteborg, paper 3ii096 Roy DM, Oyefesobi SO (1977) Preparation of very reactive Ca2SiO4 powder. J Am Ceram Soc 3-4:178-180 Saalfeld H (1967) Beitrag zur Kristallchemie des Dikalziumsilikates, Ca2SiO4. Ber Dtsch Keram Ges 44:421429 Saalfeld H (1971) Kristallchemische Untersuchungen im System Ca2SiO4-Ca3(PO4)2, Beitrag zur Frage der Diadochie des Siliziums und Phosphors. Z Kristallogr 133:396-404 Saalfeld H (1975) X-ray investigation of single crystals of β-Ca2SiO4 (Larnite) at high temperatures. Am Mineral 60:824-827 Saini A, Giuliani N (1994) XRD Analysis of aluminate and aluminoferrate phases on the selective extraction clinker residue by using a profile fitting program. Comparison between experimental and calculation data in the light of optical microscopy examination. Proceedings of the 16th International Conference on Cement Microscopy, Richmond, USA, p 24-43 Saini A, Giuliani N (1995) Improvements in XDR analysis of clinker silikate phases by using optical microscopy and selective extraction techniques. Contribution and limits of calculation formulas. Proceedings of the 17th International Conference on Cement Microscopy, Calgary, Canada, p 162-181 Sasaki K, Ishida H, Okada Y, Mitsuda T (1993) Highly reactive β-dicalcium silicate: V. Influence of specific surface area on hydration. J Am Ceram Soc 76(4):870-874 Sasaki S, Prewit CT, Bass JD (1987) Orthorhombic perovskite CaTiO3 and CdTiO3 structure and space group. Acta Crystallogr C 43:1668-1674

Calcium Aluminate Cement

79

Sawkow J (1990) High alumina cements based on calcium aluminate clinker with different phase compositions and sintering degrees. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London, p 27-38 Sawkow J (1993) Untersuchung des Phasenbestandes und der Gebrauchseigenschaften von feuerfesten Tonerdezementen. Zement-Kalk-Gips 46(5):275-280 Scheller T, Kuzel HJ (1974) Studies on dicalcium aluminate hydrates. In: International Congress on the Chemistry of Cement, Moscow/Russia, Section II, 3-11 Schmid M, Götz-Neunhoeffer F, Neubauer J (2002) Hydratation von eisenfreiem Tonerdezement unter Einfluss von Portlandit und Li2CO3-Lösung. GDCH-Monographie Bauchemie 25:176-178 Schmid M, Götz-Neunhoeffer F, Neubauer J (2003) Effect of Ca(OH)2 on hydration behavior of synthetic calcium aluminate cements. In: Proceedings of the 11th International Congress on the Chemistry of Cement. Gieves G, Owens G (eds) Durban, p 645-654 Schmid M, Götz-Neunhoeffer F, Neubauer J (2004) Einfluss von ausgewählten Additiven auf das Hydratationsverhalten von Tonerdeschmelzzement. GDCH-Monographie Bauchemie 31:258-262 Schmid M, Götz-Neunhoeffer F, Neubauer J (2005) Quantitative phase analysis of synthetic CAC from melts in the system CaO-Al2O3-Fe2O3-SiO2. Proceedings of the 27th International Conference on Cement Microscopy, Victoria, Canada, p 15 Schmidt M, Pöllmann H (2008) Hydration behavior of CAC at various temperatures by isoperibolic heat flow calorimetry using acetates. In: Calcium Aluminate Cements. Avignon/France, 93-108 CD Schmidt M, Pöllmann H (2009) Investigations of the hydration behavior of CAC under the influence of acetate and chlorooacetates at different temperatures (10 °C, 25 °C, 45 °C and 60 °C) using cluster analysis. Proceedings of the 31st International Conference on Cement Microscopy, St. Petersburg, 47 pages, on CD, April Schneider H, Komarneni S (2005) Mullite. VCH-Verlag Schuster G, Lauterbach N, Hernitschek F, Hörold U (1992) Cal-ciumaluminat-Verfahren zur weitergehenden Fällung von Sulfat und Schwermetallen. WLB Wasser, Luft und Boden, 10:40-47 Scian AN, Porto-López JM, Pareira E (1987a) High alumina cement. Study of CaO·Al2O3 formation. I. Stochiometric mechanism. Cem Concr Res 17:198-204 Scian AN, Porto-López JM, Pareira E (1987b) High alumina cement. Study of CaO·Al2O3 formation. II. Kinetics. Cem Concr Res 17:525-531 Scian AN, Porto-López JM, Pareira E (1991) Mechanochemical activation of high alumina cements. Hydration behavior. I. Cem Concr Res 21:51-60 Scian AN, Porto-López JM, Pareira E (1994) Mechanical activation of high alumina cement. Cem Concr Res 24:937-947 Scrivener KL, Capmas A (1998) Calcium aluminate cements. In: Hewlett’s Chemistry of Cement and Concrete, 4th Edition. Hewlett PC (ed) Arnold Publishers, London, p 709-778 Scrivener KL, Lewis MC, Houghton SJ (1997a) Microstructural investigation of calcium aluminate cement concrete form structures. In: Proceedings of the 10th International Congress on the Chemistry of Cement. Gothenburg Göteborg, 4iv025 Scrivener KL, Rettel A, Beal L, Bier T (1997b) Effect of CO2 on the mechanical properties of a formulated product containing calcium aluminate cement. Tagungsbericht - Band 1 zur 13. Int. Baustofftagung ibausil, Weimar, S. 1-0745 - 1-0752 Sehnke ED (1995) Bauxite - a global review. Ind Miner 8:39-51 Seidel G, Huckauf H, Stark J (1978) Technologie der Bindebaustoffe, Bd. 3: Brennprozeβ und Brennanlagen. VEB Verlag für Bauwesen, Berlin, 232 Seiten Seufert S, Hesse C, Goetz-Neunhoeffer F, Neubauer J (2009) Quantitative determination of Anhydrite III from dehydrated gypsum by XRD. Cem Concr Res Sharp JH, Bushnell-Watson SM, Payne DR, Ward PA (1990) The effect of admixtures on the hydration of refractory calcium aluminate cements. In: Calcium Aluminate Cements. Mangabhai RJ (ed) E and FN Spon, London 127-41 Shpanov EP, Sidorenko GA, Stolyarova TI (1970) Akdalaite, a new hydrous modification of alumina. Zap Vses Min Obsha 99:333-339 Sokolov AN, Kvyatkovskii OV, Fleer SA, Ashkinadze GS, Krutko GI, Karas GE, Elisova ZM (1986) Preparation of the charge and firing in a shaft kiln of high alumina cement. Refractories 27(9-10):588-591 Sorrentino D, Sorrentino FP, George M (1995) Mechanism of hydration of calcium aluminate cement. In: Materials Science of Concrete IV. Skalny J (ed) American Ceramic Society, Westerville, p 41-90 Sorrentino FP (1973) Crystallization of CaO-Al2O3-Fe2O3-SiO2 Melts and the Phase Composition of High Alumina Cement, Ph. D. Thesis. University of Aberdeen Sorrentino FP, Glasser FP (1975) The system CaO-Fe2O3-Al2O3-SiO2 I. The pseudoternary section Ca2Fe2O5Ca2Al2O5-SiO2. Trans Br Ceram Soc 74:253-256

80

Pöllmann

Sorrentino FP, Glasser FP (1976) Phase relations in the system CaO-Al2O3-SiO2-Fe-O. Part II. The system CaOAl2O3-Fe2O3-SiO2. Trans Br Ceram Soc 76 95-103 Sorrentino FP, Glasser FP (1982) The phase composition of high alumina cement clinkers. Calcium Aluminates, Seminario Internationale, Turin, Tagungs-band, p 14-43 Sourie A, Glasser FP (1991) Studies on the mineralogy of high alumina cement clinkers. Br Ceram Trans J 90:71-76 Sourie A, Glasser FP (1993) Composition of melilithes crystallised from CaO-Al2O3-Fe2O3-SiO2 melts. Br Ceram Trans 92(4):152-154 Sourie A, Glasser FP, Lachowski EE (1994) Microstructural studies on pleochroite. Br Ceram Trans 93:41-48 Spackman HS (1908) Calcium Aluminum cement. British Patents 10110 and 1845 Spackman HS (1910) Aluminates: their properties and possibilities in cement manufacture. Proceedings of the American Society for Testing and Materials, p 315 Spackman HS (1910) Cement Age. 11(2):76 Spohn E (1954) Der Schachtofen von Morgen. Zement-Kalk-Gips 11:409-415 Spohn E (1961) Die Hauenschild’sche Reaktion beim Zementschachtofen. Zement-Kalk-Gips 14(3):105-108 Stark J, Müller A, Schrader R, Rümpler K (1981) Existenzbedingungen von hydraulisch aktivem Belit-Zement. Zement-Kalk-Gips 34(9):476-481 Stark J, Müller A, Seydel R, Jost K (1986) Conditions of the existence of hydraulically active belite cement. Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro, Brasilien, Communications Theme 1, II:306-309 Stemmermann P, Garbev K, Schweike U, Bäuchle G (2007) Belite-based binder produktion method. No: WO 2007/017142 A2 Stephan D, Plank J (2007) Influence of foreign oxides on lattice parameters and reactivity of pure cement phases of OPC. 12th International Congress on the Chemistry of Cement, Montreal/Canada Stephan D, Wilhelm P (2004) Synthesis of pure cementitious phases by sol-gel process as precursor. Z Anorgan Allg Chem 630:1477-1483 Stöber S (2000) Use of X-ray Diffraction for Hydration of Cementitious Systems. Halle/Saale Stöber S, Prokhnenko O, Döring T, Pöllmann H (2009a) Eigenschaften von Perovskiten Ca(Fe,Mn,Ti)O3-d. Hall. Jahrbuch Halle 31:245 Stöber S, Prokhnenko O, Döring T, Pöllmann H (2009b) Modulare perovskite structure type phases of manganese calcium aluminate cements. Z Kristallogr Suppl 29:102 Stöber S, Redhammer G, Prokhnenko O, Schorr S, Pöllmann H (2010a) Crystal chemistry of manganese containing perovskites related to calcium aluminate cements. 26th European Crystallographic Meeting, Darmstadt 29.8.-29.9 Stöber S, Redhammer G, Schorr S, Pöllmann H (2010b) Kristallchemie manganhaltiger Perovskite in Tonerdezementen. GDCH Tagung Bauchemie, p 230-237, Dortmund 7.-8. Oktober Su M, Kurdowski W, Sorrentino F (1992) Development in non-Portland cement. Proceedings 9th International Congress on the Chemistry of Cement, New Delhi 1:317-354 Sudoh G, Ohta T, Harada H (1980) High strength cement in the CaO-Al2O3-SiO2-SO3 system and its application. 7th Int. Congress on the Chem. of cement, Paris III:V152-V157 Sugama T, Carciello NR (1991) Strength and development in phosphate-bonded calcium aluminate cements. J Am Ceram Soc 74:1023-1030 Swainson IP, Dove MT, Schmahl WS, Putnis A (1992) Neutron powder diffraction study of the åkermanitegehlenite solid solution series. Phys Chem Miner 19:185-195 Takeuchi Y, Donnay G (1959) The crystal structure of hexagonal CaAl2Si2O8. Acta Crystallogr 12:465-470 Takeuchi Y, Haga N, Ito J (1973) The crystal structure of monoclinic CaAl2Si2O8; a case of monoclinic structure closely simulating orthorhombic symmetry. Z Kristallogr 137:380-398 Talaber J (1974) High Alumina Cements. A Principal Paper. Proceeding of the 6th International Congress on the Chemistry of Cement, Moscow, p 3-41 Teoreanu I, Angelescu N, Tuclea D (1986) The influence of some organic admitxtures on refractory aluminous cement hydration. II Cemento 3:191-198 Tilley CE (1929) On larnite and its associated minerals from the limestone contact zone of Scawt Hill. Mineral Mag 22:77-86 Tilley DB, Eggleton RA (1996) The natural occurrence of eta-aluminum in bauxite. Clays Clay Miner 44(5):658664 Torkar K, Egghart G, Krischner H, Worel H (1961) Untersuchungen über Aluminiumhydroxyde und-oxyde. 7. Mitt: Der thermische Abbau von reinsten Aluminiumhydroxyden. Monatsh Chem 92:512-524 Touché M (1926) Le Ciment. 31:240 Trettin R, Rettel A (1987) Zum Verlauf der Hydratation des CaO·Al2O3 in den Primärstadien. Z Anorg Allg Chemie 547:133-141

Calcium Aluminate Cement

81

Trivino E (1986) Aluminous cement: how to avoid degrading of mechanical resistance. In: Proceedings of the 8th International Congress on the Chemistry of Cement 4:417-422 Trolard F, Tardy Y (1987) The stabilities of gibbsite, boehmite, aluminous goethites and aluminous hematites in bauxites, ferricretes and laterites as a function of water activity, temperature and particle size. Geochim Cosmochim Acta 51:945-957 Turrillas X, Convert P, Hansen T, Aza AHD, Pena P, Rodriguez MA, Aza SS (2001) The dehydration of calcium aluminate hydrates investigated by neutron thermodiffractometry. In: Calcium Aluminate Cements Edinburgh/UK, 517-531 Überoi M, Risbud SH (1990) Processing of amorphous calcium aluminate powders at 1.2 % (Maki and Kato 1982). The maximum M content in C3S is 2%, while the limits for A and F are 1.0% and 1.1%, respectively. The name alite is actually used for impure C3S as found in Portland cement clinker. Alite is generally more hydraulic active than pure C3S and hydrates therefore substantially faster (Ono et al. 1966). The crystal lattices of the individually C3S modifications differ only slightly, which can be recognized from the very low transformation enthalpies of about 0.2-4 J/g (Regourd 1979), so the differences in hydraulic reactivity may well be due mainly to the lattice dislocations caused by the incorporation of foreign ions. Stephan and Plank (2007) studied the effect of multi-doping of alite with MgO, Al2O3 and Fe2O3 on the crystal lattice parameters by Rietveld analysis of the XRD-profiles as well as the on the setting and hydration rate of C3S by isothermal calorimetry. Stephan and Plank (2007) also used 29Si MAS (magic angle spinning) NMR (nuclear magnetic resonance) to study the microstructure of doped alite before hydration and thermal analysis (DTA/TG) for further information on hydration. Dicalcium silicate, C2S, is formed when the clinker is not fully saturated with C. C2S melts congruently at 2130 °C as can be seen from the C-S phase diagram. It is a very stable compound that is formed initially in lime-rich mixtures as a solid-state reaction. Dicalcium silicate occurs in five different crystal modifications known as α-, α′H-, α′L-, β- and γ-C2S. The subscripts H and L indicate high- and low-temperature modifications. A schematic overview of the transformation between the phases can be found in Niesel and Thormann (1967) and Lehmann et al. (1969) showing that the transformations β → γ during cooling and γ → αL′ during heating are not reversible. Furthermore, during the β → γ transformation there is a substantial change in lattice constants, recognizable but the fact that the density of the γ-modification is 2.94 g/ cm3 or about 10% lower than the β-modification with 3.20 g/cm3. As a result of this change in modification an originally compact burning product cracks and rapidly disintegrates into dust (so called “dusting”) as soon as the temperature during cooling falls below 500 °C. Furthermore, the β-modification is hydraulically active while the γ-modification is NOT. This is the second good reason to rapidly cool the clinker leaving the kiln. Dicalcium silicate modifications can take up more foreign ions in their crystal lattices than tricalcium silicate. Individual modifications can be stabilized in this way so that the modification which is stable at high temperatures is retained on cooling to normal temperatures. The impure β-modification of dicalcium silicate is called belite. Microprobe examinations of

Alternative Low-CO2 “Green” Clinkering Processes

87

industrial clinkers show that belite always contains Al as a foreign ion. Na, Mg, P, S, K, Ti, V, Cr, Mn and Fe may also be present (Barnes et al. 1978). The high temperature modifications of C2S can be stabilized by certain foreign constituents. Depending on the nature and the concentration of the foreign substances or foreign substances it is possible to produce certain modifications selectively. Examples of proven stabilizers are Ca3(PO4)2, Na4P2O7, V2O5, B2O3, SrO, BaO, K2O and combinations of Na2O and Fe2O3 (Regourd and Guinier 1974; Ghosh 1985; Suzuki et al. 1986). Tricalcium aluminate, C3A, is the component in the C-A binary system richest in C It melts incongruently at 1542 °C dissociating into solid C and a melt accordingly poorer in C. Pure C3A occurs only in one cubic crystal modification (Mondal and Jeffrey 1975). The crystal lattice of C3A can take up various foreign ions in solid solution, especially Fe3+, Mg2+, Si4+, Na+ and K+, as well as Cr3+, Ni2+ and Zn2+ (Stephan et al. 1999). The alkalis play a special role as their incorporation into the lattice changes its symmetry from cubic through orthorhombic to monoclinic. Stephan and Plank (2007) studied C3A doped with Fe2O3, SiO2 and Na2O in terms of changes to the crystal lattice as well as influence on the hydration. The change in reactivity by doping with foreign oxides was not found to be directly linked with the intensity of changes in the lattice parameters, but more a function of the kind and concentration of doped foreign oxide. Na2O was found to retard the reactivity of the C3A in short and medium terms, even when the dosage was kept below the concentration where the crystal structure changes from cubic to orthorhombic (< 2.4%). The retardation was, however, compensated when the hydration proceeded. The retarding effect of Na2O on C3A was amplified by the combined doping with Fe2O3. The Fe2O3 compound in the cement clinker was originally assigned a composition corresponding to the formula Ca4Al2Fe2O10 or C4AF in short hand notation (Hansen et al. 1928). It is therefore known as tetracalcium aluminoferrite or in natural rocks as brownmillerite after its discoverer (Hentschel 1964). C4AF is a phase from the incomplete solid solution series between dicalcium ferrite, C2F, which is stable under normal conditions and dicalcium aluminate, C2A, which only can be produced under high pressure (250 MPa at 1250 °C). This means that the Fe3+ and Al3+ ions are interchangeable in this compound within certain limits. The mixed crystals of this series should therefore have the formula C2(A,F). The end member of this solid solution series richest in A consists of 70% C2A and 30% C2F (Majumdar 1965). In the ternary phase diagram C-A-F by Newkirk and Thwaite (1958), it is shown that the calcium aluminoferrite crystal that is in stable equilibrium with C and C3A has a composition corresponding to 48 mol% C2A and 52% C2F and therefore very close to the composition C4AF. The calcium aluminoferrite mixed crystal can take up foreign ions in the crystal lattice (Hansen et al. 1928). The incorporation of Mg2+ is of particular industrial significance, but it can also take up manganese, Mn2+ (Guttmann and Gille 1929a,b,c; Goffin and Muβgnug 1933a,b; Parker 1952; Akatsu and Maeda 1967; Kondo et al. 1978), titanium, Ti4+ (Marinho and Glasser 1984) and chromium, Cr6+ (Sakurai and Sato 1968). Silicon, Si4+, is apparently not incorporated in the crystal lattice (Ono et al. 1985), but is present in the form of C2S inclusions (Neubauer et al. 1996). A SEM back-scattered electron image of cement grains cast in epoxy and plane polished to μm fineness is shown in Figure 1. The angular crystals of C3S and rounded crystals of C2S can clearly be seen embedded in a white mass of C4AF and C3A. The mix of C3A and C4AF is often referred to as the interstitials since they are surrounding C3S and C2S crystals as they were the melt from where the silicates where grown. When such a clinker mass is ground, each resulting grain may be composed of all 4 phases. Thus, one cannot simulate cement by blending the 4 pure phases together in their respective amounts.

88

Justnes

Figure 1. SEM image of a large cement grain from ground clinker showing that in consists of several phases; edgy C3S to the left and rounded C2S (striped) in the right part, both grown from the frozen melt of C3A and C4AF surrounding them (called “interstitials”).

Minor elements are often added deliberately to the raw meal as mineralizers. They may also be present as contaminations in the raw meal or in alternative fuels and raw materials (AFR), but they will all affect the clinkerization as if they were added as mineralizers. Thus, the effect of mineralizers is of importance to alternative raw materials. Mineralizers are compounds (mostly inorganic) which influence the process of reactions in the solid, liquid and solid-liquid interface face during burning of cement clinker. The possible effects can be summarized in changes of the chemical, mineralogical, structural, textural, mechanical and physical properties. The effects of the mineralizers can often be caused by specific elements in the compound added to the cement raw mix. Small additions of selected elements, often referred to as foreign ions, can alter the properties of the melt extensively. The modified properties can easily be seen in terms of reactivity and burnability of the raw mix. Mineralizers can be active at several stages in the burning process. The raw meal largely consists of ground limestone, clay and quartz. Some corrective (auxiliary) materials as iron oxide and bauxite (not shown) are also interground. The main reactions/mechanisms as the temperature increase are: •

calcinations (decarbonation) of the limestone at 700-900 °C: CaCO3 → CaO + CO2

•

transformation of quartz from low- to high-temperature modification

•

formation of calcium aluminates (C12A7), dicalcium silicate (belite = C2S) and ferrite phase [C2(A,F)] above 700 °C

•

formation of molten phase at 1300 °C and tricalcium silicate (Alite = C3S) is formed by the reaction: C2S + CaO → C3S

Mineralizers are reported to affect decarbonation, the formation of alite and properties of the melt and stabilize the hydraulic belite polymorphs (α- and β-forms). Substances that are affecting the melt properties are also called fluxes. Reaction temperatures are often decreased when applying mineralizer(s) which means increased burnability of the mix. The effects of adding a mineralizer to the cement raw mix can be the following:

Alternative Low-CO2 “Green” Clinkering Processes

89

•

decreased burning temperature due to changes in the reactivity and burnability

•

acceleration of the clinkerization reactions (higher activity of the clinker minerals) at lower burning temperatures, e.g., fluorosilicates

•

altered surface tension (e.g., SO42−) and viscosity of the melt

•

formation of new intermediate phases as well as more stable phases when approaching firing temperatures

•

controlled polymorphism of the clinker minerals

•

changed properties of the produced cement (e.g., hydraulic activity and strength development)

An extensive literature review on the effect of minor elements was carried out by Moir and Glasser (1992). They explored the effects of minor components in view of the periodic table of elements and divided them into the following groups; alkalis, transition metals, halogens and p-block elements.

High belite cement clinker High belite cement is essentially a Portland cement where the content belite, C2S, is much higher (45-60%) than that of alite, C3S (20-30%), or quite the opposite of an ordinary Portland cement (OPC). The lower calcium level should then give less CO2 emission providing the source is limestone. In addition, this cement requires about 100 °C lower kiln temperature than OPC, requiring less fuel and hence somewhat lower CO2 emission also for this reason. High belite cement is harder to grind than OPC and will require some extra energy in this respect. Gartner (2004) made an estimate of CO2 savings changing from a modern OPC with 65% C3S to a high belite cement with little or no C3S and found that the overall reduction in limestone consumption in total would not be more than 8%. Even allowing for the ensuing reductions in burning temperature, he pointed out that the likely maximum total CO2 emission savings only would be in the order of 10%. He also said that this has to be balanced against the fact that high belite clinkers are very hard to grind and thus require more energy. Very low rates of strength development are also considered unsatisfactory by most costumers. Chatterjee (2003) reviewed the status of high-belite cement and concluded that the interest in it has grown over the three last decades due to its anticipated multidimensional benefits like lower energy consumption, raw materials conservation and constructional durability of the resultant concrete. However, the product and its manufacturing technology are yet to be of extensive commercial significance as there are still no viable technologies to substantially enhance the intrinsic low reactivity of the belite phase and to generate large surface area for the cement at a reasonable energy input to achieve a higher degree of hydration in concrete (also see Popescu et al. 2002). It seems that high belite cement in practice is produced in Japan, India and China (Sui and Yao 2003) and that the application first and foremost is as low heat cement in massive structures like dams. However, low heat can also be obtained by for instance using large content of supplementary cementing materials as blast-furnace slag and fly ash in the concrete (also see Janotka and Krajci 1999). For the sake of cement with less CO2 emission it does not seem like high belite cement is worth pursuing with its low savings potential (about 10%), in particular when bearing in mind the low strength development rate that would have hampered building productivity.

Calcium sulfoaluminate cement (CSA) Calcium sulfoaluminate cement (CSA) has recently been promoted as the cement for sustainable development (Alaoui et al. 2007) as a typical cement composition is 53% C4A3 S ,

90

Justnes

18% C2S, 12% C and 15% C4AF (2% residual). According to their comparison with Ordinary Portland cement reproduced in Table 3, the CO2 emission is in theory not only less for calcium sulfoaluminate cement (−43%), but also the specific heat consumption during clinkering (−14%) due to lower temperature, as well as lower crushing energy (−40%) since the minerals are more friable. Table 3. Comparing CO2 emission (from raw materials) and energy of making ordinary Portland cement vs. sulfoaluminate cement (Alaoui et al. 2007).

Parameter CO2 emitted per tonne clinker specific heat consumption during clinkering1 energy for crushing3

Ordinary Portland Clinker

Sulfo-aluminate Clinker

535 kg/t 3.845 GJ/t2

305 kg/t 3.305 GJ/t

45 - 50 kWh/t

20 - 30 kWh/t

1

Popescu et al. 2002, 2BAT is 3.1 GJ/t, 3Janotka and Krajci 1999

Gartner (2004), however, pointed out the many practical problems with cementing systems based on ettringite (C6A S 3H32) as binder, such as sulfoaluminate cements, especially the problem of controlling the expansion associated with the reaction. Calcium sulfoaluminatebased cements are increasingly being used in special applications where high early strengths and self-stressing or shrinkage compensation are required (e.g., self-leveling screeds), but the more general application to concrete is limited to China, where a wide range of C4A3 S -based cements have been developed and normalized under the name of the “Third Cement Series” with the acronym “TCS” (Zhang et al. 1999). In the Chinese literature, it is stated that TCS, which are based on clinkers containing C4A3 S , belite and ferrite in various proportions as their major phases, can be used in a wide variety of applications depending on phase composition and on the amount of gypsum or anhydrite interground to make the final cement. The TCS technology practiced in China lately seems to mainly be based on clinkers with rather high C4A3 S contents (60-70%), utilized in the pre-stressed concrete sector in which the rapid strength development at moderate curing temperatures, plus self-stressing, are economic advantages. TCS is usually manufactured using bauxite as a principal raw material, making them relatively expensive compared to Portland cements. The TCS approach has been reinvestigated in some eastern European countries (Palou et al. 2003) to make energy-efficient sulfoaluminate-belite cements with lower C4A3 S contents and higher belite contents than TCS produced in China. However, available strength results are disappointing, probably due to the same problem of low belite reactivity in high belite Portland cements, so TCS cements may not offer very significant global CO2 savings if the usually construction productivity is to be maintained. Li et al. (2007) acknowledged the great potential of calcium sulfoaluminate cement in reducing CO2 emission by at least 20-30% compared to an OPC of equal performance providing the clinker is produced in a modern rotary kiln. They also added that a great deal further careful study will be required to fully understand the hydration of these interesting and novel cements in order to better optimize their compositions and thereby further decrease manufacturing CO2 emissions for equal concrete performance. Valenti et al. (2007) also pointed out a few other environmentally friendly aspects of calcium sulfoaluminate cements. Firstly, industrial wastes and by-products difficult to reuse and dispose can be used as raw materials for its clinker production, such as fluidized bed combustion waste, red mud, low-quality pulverized coal fly ash and chemical gypsum. Secondly, the intergrinding

Alternative Low-CO2 “Green” Clinkering Processes

91

of large amounts of gypsum with the clinker enables reduced clinker content and enhanced chemical gypsum utilization, in particular flue gas desulfurization gypsum generated worldwide in increasing amounts. Pèra and Ambroise (2004) listed the following advantageous applications of sulfoaluminate cements: •

development of concrete with high early strength: 40 MPa in 6 h and > 55 MPa at 24 h

•

design of self-leveling screed with limited curling when unbounded to its support

•

design of self-leveling topping mortar presenting the following properties: time of workability > 30 min, set within 75 min and low drying shrinkage (< 250 μm/m)

•

glass fiber reinforced cement (GFRC) composites that can be demolded 4 h after casting and present high ductility and durability after aging in different weathering conditions

Quillin (2001) acknowledged a very good sulfate resistance of sulfo-aluminate-belite cement, but the chloride diffusion was higher than Portland cement, and especially the carbonation rate. However, he admitted that the durability may have been improved using a suitable plasticizer to achieve a lower w/c (used w/c = 0.56). Glasser and Zhang (2001) evaluated the durability of 14 year old reinforced concrete pipes (w/c = 0.25) exposed to the tidal zone in China, and found that the mild steel mesh reinforcement was without corrosion. This may have been due to a dense matrix and rapid self-desiccation that is difficult to re-saturate. The aspects concerning durability has made applications of calcium sulfoaluminate cement outside China limited to for instance rapid repair mortars and self-leveling screeds. In China with its > 106 t/year production it is also used in construction, but apparently in low performance structures as in in-door housing etc. However, this makes quite a bit of the total concrete market in a society, so this cement may be worthwhile looking further into due to its large saving potential in CO2 emission and considering the fast strength development enabling faster building processes.

Calcium aluminate cement (CAC) Calcium aluminate cements (CACs) (Hewlett 1998) are cements consisting predominantly of hydraulic calcium aluminates. Alternative names are “aluminous cement,” “high-alumina cement,” and “Ciment Fondu” in French. They are used in a number of small-scale, specialized applications. An extensive review on calcium aluminate cements is given in Pöllmann (2012, this volume). The method of making CACs from limestone and low-silica bauxite was patented in France in 1908 by Bied, from the Pavin de Lafarge Company. The initial development was as a result of the search for cement offering sulfate resistance. Subsequently, its other special properties were discovered, and these guaranteed its future in niche applications. The cement is made by fusing together a mixture of a calcium-bearing material (normally limestone) and an aluminum-bearing material (normally bauxite for general purposes, or refined alumina for white and refractory cements). The liquefied mixture cools to a basalt-like clinker which is ground alone to produce the finished product. Because complete melting usually takes place, raw materials in lump-form can be used. A typical kiln arrangement comprises a reverberatory furnace provided with a shaft preheater in which the hot exhaust gases pass upward as the lump raw material mix passes downward. The preheater recuperates most of the heat in the combustion gases, dehydrates and de-hydroxylates the bauxite and de-carbonates the limestone. The calcined material drops into the cool end of the melt bath. The melt overflows the hot end of the furnace into molds in which it cools and solidifies. The system is fired with

92

Justnes

pulverized coal or oil. The cooled clinker ingots are crushed and ground in a ball-mill. In the case of high-alumina refractory cements, where the mix only sinters, a rotary kiln can be used. The main active constituent of calcium aluminate cements is monocalcium aluminate (CaAl2O4 or CA in the cement chemist short-hand notation). It usually contains other calcium aluminates as well as a number of less reactive phases deriving from impurities in the raw materials. Rather a wide range of compositions is encountered, depending on the application and the purity of aluminum source used (Taylor 1997). Constituents of some typical formulations are given in Table 4. The mineral phases all take the form of solid solutions with somewhat variable compositions. Because of their relatively high cost, calcium aluminate cements are used in a number of restricted applications: 1) in construction concretes, rapid strength development is achieved, even at low temperatures, 2) in construction concretes, high chemical resistance is possible, 3) in refractory concretes, strength is maintained at high temperatures and 4) as a component in blended cement formulations, various properties such as ultra-rapid strength development and controlled expansion can be obtained. Incorrect use of calcium aluminate cements has led to widespread construction problems, especially during the third quarter of the 20th century when this type of cement was used because of its faster hardening properties. After several years some of the buildings and structures collapsed due to degradation of the cement and many had to be torn down or reinforced. Heat and humidity accelerate the degradation process known as “conversion;” the roof of a swimming pool was one of the first structures to collapse in the UK (http://webs. demasiado.com/forjados/patologia/aluminoso/index.htm). In Madrid, Spain, a large housing block nicknamed Korea (because it was used to house Americans during the Korean war), built ~1951-54 was affected and had to be torn down in 2006. Also in Madrid, the Vicente Calderón soccer stadium was affected and had to be partially rebuilt and reinforced (www.elmundo.es/ papel/2007/02/07/madrid/2082060.html). Table 4. Oxide and mineral composition of some calcium aluminate cements (CAC).

Oxide/Mineral SiO2 (S) Al2O3 (A) Fe2O3 (F) CaO (C) MgO (M) Na2O (N) K2O (K) TiO2 (T) monocalcium aluminate (CA) dodecacalcium hepta-aluminate (C12A7) monocalcium dialuminate (CA2) belite (C2S) gehlenite (C2AS) ferrite (C4AF) pleochroite (~C6A4MS) wüstite (F) corundum (A)

General Purpose

Buff

White

Refractory

4.0 39.4 16.4 38.4 1.0 0.1 0.2 1.9 46 10 0 7 4 24 1 7 0

5.0 53.0 2.0 38.0 0.1 0.1 0 1.8 70 5 0 5 14 5 1 0 0

2.7 62.4 0.4 34.0 0.1 0 0 0.4 70 0 17 0 11 2 1 0 0

0.4 79.6 0 19.8 0 0 0 0.1 35 0 30 0 1 0 0 0 33

Alternative Low-CO2 “Green” Clinkering Processes

93

Calcium aluminate cements (CACs) are like calcium sulfoaluminate cements (TCS) usually manufactured using bauxite as a principal raw material, which is the main reason why they are relatively expensive compared to Portland cements. All CACs has monocalcium aluminate (CA) as the main cementing mineral (see Table 4) and thus they have a fairly low raw material derived CO2 emission. However, in addition to its higher raw materials cost, the general purpose clinker (“Ciment Fondu”) is made by a melt process, which is not very energy efficient compared with the Chinese TCS approach (which makes use of conventional rotary kilns).

MINERALS AS ALTERNATIVE TO LIMESTONE General Calcium sources other than limestone can, in theory, be used as raw materials, but according to Gartner (2004) there are not really any other sufficiently widespread and sufficiently concentrated sources of calcium available. However, it will also help to replace parts of the calcium with other minerals than limestone, not necessarily all of it. On the other hand, local deposits may play a role for new plants to be built as limestone may become scarce some places. Important developing countries like India, for instance, only have limestone sources available that can serve cement plants to be built until 2012 for their expected production time thereafter, according to Kulkarni (2011). It should be reinstated once more that the easiest short-time measure to reduce CO2 emission from Portland cement production, is to replace part of the clinker with supplementary cementing minerals (e.g., fly ash, blast furnace slag, calcined clay, impure limestone or calcined marl) as outlined by Justnes (2007a,b 2010). However, that is the topic of another contribution to this MSA Short Course (Snellings et al. 2012, this volume). Generally speaking, one should look for natural calcium silicates that also may contain hydroxyl groups or hydrate water without any problem. Minor content of carbonate may also be acceptable. The aluminate and ferrite content cannot be too large if it should significantly replace calcium and silicon in Portland cement clinker, since the maximum content in this clinker typical is 7% Al2O3 (A) and 3% Fe2O3 (F) as pointed out in Table 1. High A (≈ 39%) and F (≈ 16%) is found in CACs (see Table 4), but then the silica (SiO2 or S) content is low (≈ 4%). The CSA is also high in alumina (30% or 36% A depending on type), but also here silica is low (6-8% S) in addition to low ferrite ( 28 d) ages. Belite occurs in clinker as the second framework grain, typically appearing as rounded grains with a lameller structure (appears cross-hatched) that is brought out by chemical etching. Tricalcium aluminate (Ca3Al2O6) comprises from 1% to 18% by mass of a clinker, and may be referred to as C3A for the pure phase or aluminate for the industrial phase. It can occur in one or more of three crystallographic forms dependent upon the amount of chemical substitution. The aluminates react rapidly, but contribute little to the cement strength. Tetracalcium aluminoferrite (Ca2AlFeO5), or C4AF is found in concentrations between 5% to 15% by mass of clinker. The ferrite phase is a solid solution with variable Al2O3/Fe2O3 ratios and does incorporate substituted Mg+2, Si+4, Na+1, and K+1 ions (Hofmänner 1975). The solid solution phases are not distinguishable using light microscopy, and the ferrite phase is highly reflective and unaffected by common etching reagents. The aluminate and ferrite phases are part of a liquid phase in the clinkering process, forming a matrix linking the silicate framework grains. Their texture is influenced by the cooling rate, and may be coarse-grained for slowly cooled and fine-grained for rapidly cooled clinker. The fine-grained varieties can be difficult to distinguish using a light microscope and the terms “undifferentiated” and “differentiated” matrix is a commonly used descriptor. White cement is manufactured from clinker characteristically devoid of C4AF. The iron and minor amounts of manganese in the ferrite phase in Portland cement result in a strong coloration, imparting the typical gray color to the clinker nodules and to the cement. Periclase (MgO) is found in clinker as a result of MgO in the limestone feed material. Clinker with MgO contents in excess of 2% will generally exhibit some periclase. It is dispersed

104

Stutzman

throughout the clinker, within and between silicate and matrix phases as equant crystals, but may also assume a dendritic morphology. Periclase is relatively hard and resistant to etching reagents, making it stand out in topographic relief, and readily identified in low to mediummagnifications as pink grains when rocking the focus above and below the optimum. Periclase exhibits slow reactivity, forming brucite (Mg(OH)2) with a substantial volume increase. In a hardened concrete, this volume increase may not be accommodated and the strain may result in cracking. ASTM C 150 specification limits the total MgO content of cement to 6% in an effort to reduce the propensity for unsoundness. Some microscopy studies have questioned this limit, placing more emphasis on the size and distribution of periclase (Taylor 1997). Free lime (CaO) may occur in clinker where the CaO:SiO2 ratio of the raw feed material is too high, the raw feed material is not homogeneous, or from over-sized limestone fragments. Free lime, like periclase, may hydrate after the cement paste has hardened, with an increased potential for cracking. Free lime readily hydrates upon exposure to humid air, converting to Portlandite over a short time. The microscope and a qualitative scan of the clinker should yield clues to the source of free lime, which is an important part of monitoring the efficiency and effectiveness of the clinker production process (Hofmänner 1975). The alkali sulfates occur as a number of different mineral forms, the most common being arcanite (K2SO4) and aphthitalite ((K,Na)3Na(SO4)2). Less common are calcium langbeinite (K2Ca2(SO4)3), thenardite ((Na,K)SO4), and anhydrite (CaSO4) (Taylor 1997). In light microscopy, these phases are grouped as “alkali sulfates” and identifiable (as a group) with appropriate etching reagents. In SEM imaging, some distinctions may be made based upon X-ray microanalysis, especially if complementary X-ray powder diffraction data are available. Gypsum (CaSO4·2H2O) is added to the clinker during the cement grinding process. Although the term gypsum is used, either through impurities in the raw material or due to dehydration in the grinding process, the actual mineral composition of the gypsum may include bassanite (CaSO4·0.5H2O, often called hemihydrate or plaster), and anhydrite (CaSO4). The availability of sulfate ions directs the hydration reactions involving the aluminate phases to form ettringite preferentially to the AFm phases, which are responsible for the stiffening phenomena called flash set where, as the name indicates, the concrete mixture stiffens rapidly and becomes unworkable during placement and finishing. Another phenomenon, called false set, involves the hydration of bassanite to gypsum with a concomitant increase in the rheological properties. In contrast to flash set, this problem may be overcome with time and additional mixing. The clinker pore system may also be of interest, especially if an assessment of clinker grindability is being made. The pore system is a network of inter-connected and occluded voids that may vary significantly across a clinker nodule. Ideally, the embedding epoxy fills the voids to enhance their contrast with the other clinker phases. The discipline of clinker petrography has evolved to better understand the clinkering process, to evaluate process-related problems in clinker production, and to better relate clinker compositional and textural attributes to concrete performance.

CLINKER MICROSCOPY Microscopic examination is a direct means of analysis, in contrast to the indirect Bogue calculations that transform a bulk chemical analysis into phase estimates. While quantitative microscopy is an ideal technique for clinker analysis in cement manufacturing, it is not widely used today due to the time involved in a proper analysis. It remains very useful as a rapid screening tool to evaluate clinker production through semi-quantitative assessment of microstructural features and their association with potential processing problems. In addition,

Microscopy of Clinker and Hydraulic Cements

105

the microscope is used to evaluate the raw mix that is fed into the cement kiln as part of a quality control program (Fundal 1980; Miller 1981; Campbell 1999). Finally, the microscope has been invaluable in efforts to develop Standard Reference Materials for cement clinker, and to capture the compositional and textural aspects of hydraulic cements, fly ash, and slag for use in developing three dimensional microstructure development models of the hydration process. The microscope is a relatively rapid means to assess the clinker fabric for the cement manufacturer. Petrographic analyses of clinker utilize descriptive terminology for phase assemblages, texture and fabric, similar to that in igneous and metamorphic petrology. As described by Hofmänner (1975), clinker fabric is comprised of three features: phase association, texture, and structure. Phase association refers to the types and amounts of the constituent phases in the clinker. Texture stands for the size, shape, and distribution of the phases, and structure refers to potential grain orientation and the pore system. Textural terms commonly used are euhedral (or idiomorphic) for well-formed crystals, subhedral for moderately well formed crystals, and anhedral (xenomorphic) for poorly formed crystals, along with descriptive terms like dendritic, lath-shaped, and amoeboid. In addition to the individual phase descriptions, the phase distribution is an important reflection of the grinding, homogenization, and mixing of the source materials throughout the clinkering process. Phases may occur as clusters with distinct boundaries, in clusters with a ragged boundary, and as clusters forming streaks across the clinker nodule. These modes of occurrence may vary widely across a nodule so care must be taken to examine large areas in the examination of clinker. If the phase abundance is the primary goal for the examination, crushing the clinker should provide a more uniform specimen for examination. For cements, this is less of a problem as the grinding provides a degree of homogenization. Much effort has been directed toward relating clinker fabric with the manufacturing process, specifically anomalies in the processes of raw feed production and homogenization, and the clinkering processes Campbell (1999). Hofmänner (1975) provides an excellent chart that relates textural and structural aspects of the clinker constituents to the raw mix fineness, homogeneity, and chemical composition, the kiln burning conditions, and cooling conditions. Furthermore, according to Hofmänner (1975), the ideal clinker has a uniform distribution of all phases, euhedral to subhedral alite crystals, rounded belites, a fine-grained, differentiated matrix, and only minor, well-dispersed crystals of free lime. Semi-quantitative descriptions of clinker microstructure are useful in trying to evaluate the clinkering process. For example, if a clinker exhibits an abundance of belite, little alite, and no free lime, the raw mix may have had a low CaO:SiO2 ratio. Conversely, a clinker with abundant alite, little belite, and little free lime will have had a high CaO:SiO2 ratio of the raw mix. As the raw mix deviates from the ideal particle size distributions and homogeneity, a more heterogeneous clinker will result. Similarly, if the proportions of lime to silica or of aluminum to iron deviate from the ideal, changes in the phase proportions and microstructure will occur. The chart provided in Hofmänner (1975) on interpretation of clinker texture should be useful in unraveling the effects of confounding processes in clinker microstructure. Changes in cement manufacture toward greater energy efficiency, recycling of kiln dust, and use of waste materials as raw materials and fuels have added new variables in the clinkering process, and an update of this useful chart would be beneficial to the industry.

SPECIMEN PREPARATION FOR MICROSCOPY Obtaining a representative sample can be a challenge in a cement plant given the throughput of a few hundred tons per hour, or more. Often the objective of the study is to specifically analyze materials that are unusual, like fragments of material dissimilar to the clinker. For routine analyses, Hofmänner (1975) provides an example plan where three 2 kg

106

Stutzman

samples are takes at five-minute intervals from a selected sampling location. These samples are blended and split using a riffle splitter or by cone-and-quartering to a 500 g sample. The remaining 500 g sample is crushed to a 5 mm particle size, blended, and then further split to the desired sample size to produce two specimens. Proper specimen preparation methods facilitate examination and interpretation of microstructural features. Improper preparation methods, however, may obscure features, and even create artifacts that may be easily misinterpreted. Reflected light microscopy and scanning electron microscopy using backscattered electron and X-ray imaging require a highly polished surface. The polished surface has two distinct advantages to a fracture surface: 1) clear definition of the constituent phases, and 2) a planar surface amenable for quantitative analysis. Using a representative sample, specimens are potted in an epoxy resin to permeate the material’s pore system and to encapsulate the particles. The specimens are then cut or ground to expose a fresh cross section of particles, lapped to smooth the surface, and then polished using a series of successively finer grades of diamond paste. This polishing stage may be subdivided into a coarse polish where the grinding damage is removed to expose a blemish-free cross section, and a fine-polishing stage that removes the fine scratches that obscure the details of the microstructure. Each of these steps will be illustrated subsequently. Epoxy impregnation of the pore system serves three purposes: A) it fills the voids, B) it encapsulates the particles, creating a solid that is better able to resist plucking and spalling during the polishing process, and C) it enhances contrast between the pores and cementitious material. With relatively high permeability materials or powders such as clinker or Portland cement, an epoxy of low viscosity is necessary while for the less permeable cement pastes and concretes an ultra-low viscosity epoxy aids in rapid infiltration of the pore structure. The selection of epoxy depends upon the materials and the means for analysis. For polished sections, the ideal epoxy will exhibit good capillarity, wet and bond to pore walls and edges, fill voids, and will not leave any residue on the specimen surface that will adversely affect etching (for light microscopy) or backscattered electron contrast, will not soften (promoting particle plucking) from exposure to cleaning agents such as ethanol or acetone, and, particularly for SEM analyses, is relatively beam-stable. For clinker, a medium-viscosity epoxy1,2 is used to promote rapid infiltration of the accessible void spaces. Vacuum assist pulls the air out of the less accessible void spaces and forces the epoxy into these voids. While complete permeation may not be possible in many cases as a result of occluded voids, an additional step of back filling the open voids and a second epoxy cure after sectioning may prove useful. A higherviscosity epoxy is used for cement, fly ash, and slag powders. This epoxy cures harder and is a better match to the material when polishing.

Materials for sample preparation A list of equipment and materials necessary for preparation of polished specimens is given in Table 1. For some items, substitution may be possible if comparable supplies are available in the laboratory. The list is presented in order of use of the equipment or supplies.

Preparation of clinker Clinker nodules or crushed fragments are placed in a mold container and surrounded by epoxy leaving a top surface exposed to the laboratory air, allowing the epoxy to be drawn 1 Certain commercial materials and equipment are identified in order to adequately specify experimental procedures. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the items identified are necessarily the best available for the purpose. 2 Suitable materials include L.R. White, hard grade for ultra-low viscosity, and Epotek 301 for medium viscosity, and Epotek 353ND for powders.

Microscopy of Clinker and Hydraulic Cements

107

Table 1. Equipment and supplies for preparation of polished sections. Item

Purpose

Potting epoxies (medium and low viscosity) Mold cups Mold release Metal trays to hold specimens Vacuum chamber and pump Drying / curing oven Diamond blade wafering saw Propylene glycol Alcohol: 200 proof ethanol, or isopropyl Acetone Ultrasonic bath Lapidary wheel (minimum 200 mm) Diamond pen Abrasive papers (silicon carbide) Polishing cloths (low-relief) Diamond paste for polishing Lint-free cloths Compressed air Vacuum dessiccators

for powders and hardened pastes potting specimens facilitates removal of specimen / epoxy contains any leaking epoxy vacuum impregnation capable of at least 65 °C cutting after curing diamond saw cutting lubricant alternate cutting lubricant, cleaning aid final cleaning aid specimen cleaning grinding and polishing label engraving coarse to fine grinding, 240 to 1200 grit 6 mm and finer polishing 6, 3, 1, 0.25 mm in non-aqueous suspension specimen handling and cleaning specimen cleaning and drying specimen storage

into the microstructure by capillary suction. To speed the infiltration, the specimen may be completely immersed in epoxy, and a vacuum drawn to remove remaining air. When the bubbling stops the vacuum may be released, forcing the epoxy into the pore system. The epoxy is cured, and then is ready for the cutting and polishing. Figure 1 shows an example of a potted, polished set of clinker, cement, and fly ash specimens.

Polished powder mounts of Portland cementitious materials Powder mounts are prepared by suspending cement powder in epoxy, curing the epoxy, cutting and polishing a surface of the powder/epoxy composite (Fig. 2). To save preparation time, multiple specimens of powders may be mounted simultaneously by preparing sample disks that are drilled for each specimen; a 4 mm hole being suitable for obtaining a representative sampling of most powder specimens. A reference mark is then cut into the disk so it may be oriented in the microscope; a dry cut using the diamond wafering saw is suitable for this operation. The cement powder is mixed with a few drops of epoxy, adding powder to form a cohesive ball. The cement/epoxy mixture is placed in a drill cavity and pressed to fill the base of the cavity. The mixture is then consolidated in the sample mold by sharply tapping it on the laboratory bench top, and the epoxy is cured according to the manufacturer’s guidelines. After curing, the specimen is removed from the mold, labeled and a fresh surface is exposed using a wafering saw or by grinding.

Cutting and grinding The purpose of the cutting, grinding and polishing steps, which are common to all preparations, is to expose a fresh, flat surface. Diamond blade slab or wafering saws, lubricated using propylene glycol, are suitable for exposing a fresh surface. This surface needs to be smoothed by grinding. Abrasive papers of 400 and 600 grit (silicon carbide paper) used dry are also suitable for rapid removal of material by grinding. Coarser abrasive may be used if

108

Stutzman

Figure 1. Polished sections of clinker (upper-left), while cement and fly ash powder samples contain multiple samples that have been indexed for identification (mm scale at base).

necessary. Using successively finer grades of abrasive paper removes damage produced by the earlier grit. After the 600 grit grind, the surface is smooth enough for polishing with the diamond pastes. Visual examination of the specimen allows one to identify when the abrasive has cut the entire surface. Grinding striations on the specimen surface indicate that grit has completely removed a layer of material. By alternating grinding directions by 90 degrees one can insure that the entire surface has been ground. These operations damage the specimen surface necessitating a polishing step that is described next.

Polishing Polishing removes the damage imparted by the grinding operations. Using a sequence of successively finer polishing pastes composed of fine diamond particles (ranging from 15 mm

Microscopy of Clinker and Hydraulic Cements

109

Figure 2. Specimen mounts prepared using the same epoxy used to pot the powders. Cured disks (1) are pre-drilled (2) to accommodate multiple powder mixtures (3), are cured, cut and polished (4), and ready for the microscope.

down to 0.25 mm) and a lap wheel covered with a low-relief polishing cloth, an undisturbed microstructure is exposed and the fine scratches of the polishing operation are removed. The coarse polishing stage eliminates the grinding damage zone, exposing a relatively undamaged cross section, and the final polishing removes residual scratches from the coarse polishing stage, ideally leaving a damaged zone that is essentially invisible for the imaging method applied. Figures 3 and 4 illustrate the increased clarity of a clinker microstructure as the grinding damage is progressively removed with initial polishing stages using a 1200 grit silicon carbide paper or 9 mm diamond paste. These images are from a reflected light microscope from unetched specimens. Figure 3 (upper image) after sawing using a diamond blade wafering saw shows only the outline of the fragment, with no internal details discernable. In Figure 3 (lower image), the 6 mm polish is beginning to cut the topographically high portions off the specimen to reveal cross sections of the crystals of the specimen. This stage is perhaps the most important step, as incomplete exposure of undisturbed microstructure will result in the polishing of only a portion of the specimen, leaving an inordinately large apparent void system. A reflected light microscope is helpful to examine the clinker fragments, or a large cement grain to confirm that no grinding damage remains before moving on to the polishing steps. Figure 4 (upper image) illustrates a specimen with a minor amount of grinding damage, evidenced by the elongated, irregular-edged dark voids within the clinker crystals. Additional time spent coarse polishing almost completely removes the grinding damage, leaving a haze of fine scratches (Fig. 4, lower image). This specimen is now ready to move on to the final polishing stages. Subsequent polishing stages of 3 mm, 1 mm, and 0.25 mm pastes remove fine scratches from the 6 mm polish, further improving constituent definition.

110

Stutzman

Figure 3. Reflected light microscope images of a clinker surface after saw-cutting and grinding using 600grit silicon carbide (upper image) exhibits no discernible microstructure due to the rough surface. Increased polishing time (bottom image) using 6 mm diamond paste progressively removes grinding and cutting damage pits and begins to reveal the underlying microstructural features.

Microscopy of Clinker and Hydraulic Cements

111

Figure 4. Reflected light microscope images showing the effects of grinding and polishing. The upper image shows grinding pits due to incomplete grinding. The lower image shows a specimen where the coarse polishing has removed all the grinding pits, and the specimen is now ready for the 3 mm and finer polishing steps to remove any fine scratches.

Stutzman

112

The final polish must meet a number of criteria to ensure one achieves the best possible surface for analysis: a) few scratches on the specimen surface, b) sharp particle and pore perimeter edges (over-polishing will tend to round corners), c) well-defined phases and crystal boundaries, d) minimal surface relief, e) no etching due to the polishing process, f) epoxy completely fills all voids, g) no polishing media residue trapped within voids or on the surface, and h) the surface is cleaned using ethanol or isopropyl, followed by an acetone rinse. A thorough cleaning is necessary to provide a surface that will allow a uniform etch or carbon coating (for SEM examination). An example of a well-polished clinker fragment is shown in Figure 5. In this example, some of the pores appear unfilled and show reflection from the pore walls. The ferrite phase is the most distinct with its high reflectivity, but the remaining phases appear similar. The use of chemical etchants was devised as a means to enhance distinction between phases through enhanced coloration of reaction products and enhanced grain boundaries, facilitating examination by light microscopy.

Figure 5. SRM 2688 after polishing displays a surface suitable for etching for light microscopy or ready for SEM examination. Voids to the right are not filled with epoxy, resulting in some light reflection from the void walls.

Microscopy of Clinker and Hydraulic Cements

113

Etching for light microscopy Polished sections of clinker in the light microscope have little contrast aside from the more highly reflective ferrite phase. The use of etchants provides contrast between the phases due to selective reaction with the reagents and resulting precipitation of the reaction products, imparting color to each phase. Relief polishing produces enhanced grain boundaries, facilitating phase identification. Directions for making etch solutions and their applications are taken from Campbell (1999). Figure 6 shows the effect of a 0.1 mol/L aqueous solution of KOH applied to a polished clinker surface for 30 s, rinsed with isopropyl and dried. This etchant results in the blue coloration of the aluminate phase, darkening the alkali sulfates, and imparting coloration to free lime. This is typically followed by a nital (1.5 mL of nitric acid in 100 mL of ethanol) etch for about 10 s, then rinsed with isopropyl that results in a blue coloration for alite and brown for belite. The vapor of hydrofluoric acid makes an ideal etchant, leaving a uniform coloration to the silicates, slightly affecting the aluminates, darkening the alkali sulfates, and with little affect on periclase and the ferrite phase. The hazards in handling this acid, however, make it less popular.

Figure 6. SRM 2686 KOH etch provides some phase contrast by imparting a blue coloration on the aluminate matrix phase, the silicates alite and belite are uniformly affected, and the ferrite phase, unaffected, remains bright.

SRM clinker The Standard Reference Material (SRM) clinkers are used for developing and testing methods of quantitative phase analysis (Stutzman et al. 2008). These clinkers were selected as representative of the range of North American clinker production with respect to phase abundance, crystal size, and crystal distribution. The reference values represent consensus means and un-

114

Stutzman

certainties based upon three independent analytical methods: (1) quantitative XRD, (2) light microscope point counts, and (3) image analysis of scanning electron microscope image sets. Clinker 2686a is intermediate in crystal size relative to the other SRM clinkers, and exhibits heterogeneous phase distribution and little free lime (Fig. 7). Alite occurs as subhedral to anhedral crystals approximately 30 mm in size. Belite occurs in large clusters with an approximate crystal size of 20 mm. The matrix is differentiated with a medium- to fine-grained lath-like ferrite, and fine-grained aluminate filling the inter-lath voids. Equant periclase crystals up to 15 mm are common throughout the microstructure and the alkali sulfate phases aphthitolite and arcanite are disseminated throughout the microstructure. According to the Hofmänner (1975) diagnostic chart, the belite clusters may represent heterogeneity or over-sized silica grains in the raw mix. The fine-grained matrix indicates rapid cooling from 1300 ºC, and the slightly elevated belite content may reflect a slightly low CaO content. SRM 2687 is a fine-grained clinker with abundant subhedral alite with a grain size of about 40 µm, few belite clusters with rounded grains scattered throughout the clinker, and some diffuse free lime clusters (Fig. 8). This may be interpreted as a raw feed having a high CaO content or perhaps heterogeneous blending. The matrix is abundant, but poorly differentiated, indicating a rapid cooling. Alkali sulfates are common and dispersed throughout the clinker along grain boundaries. SRM 2688 is a coarse-grained, homogeneous clinker with euhedral to subhedral alite up to about 150 mm in size, and well-dispersed, rounded belite of size up to about 40 mm (Fig. 9). The matrix exhibits two distinct types, both well differentiated where ferrite occurs as either a coarse lath-like or a medium-grained dendritic form, aluminate occurs as fine-grained with the dendritic ferrite, and more lath-like with the coarse ferrite (Fig. 10). Periclase is a minor component and is fine-grained and dispersed within the matrix. Alkali sulfates are uncommon and reside within the matrix. The coarse-grained nature of the clinker silicates may reflect a long residence time in the kiln above 1300 °C, or a high maximum temperature. The coarse matrix may reflect a slow cooling. Clinker nodules may experience different thermal histories from the outer-portion relative to the core, especially if they are large. Examination of a wide cross section of particles may be necessary to be able to make definitive conclusions on a clinker.

Point count analysis Quantitative microscopy is based upon the relationship between area fraction and volume fraction. For composites consisting of randomly distributed, randomly oriented phases, the planar area fraction is an unbiased estimator of the volume fraction (Chayes 1956; Campbell and Galehouse 1991). The Glagolev-Chayes method, referred to as the point count method, is perhaps the most widely used technique in quantitative mineralogical analysis when using a microscope, and is the basis for ASTM C 1356, Standard Test Method for Quantitative Determination of Phases in Portland Cement Clinker by Microscopical Point-Count Procedure. This procedure utilizes a grid of points to sample the clinker polished section (Fig. 11). Selecting a combination of eyepiece reticule and magnification in conjunction with the clinker crystal size to achieve spacing so as adjacent points ideally do not fall on the same crystal is a general rule and sampling in a regular pattern across the sample. About 3000 to 4000 points on clinker phases are necessary to have reasonable counting statistics for the measurements and a duplicate sample is recommended Hofmänner (1973). Calculate mass percentages by multiplying the volume fractions by the density of the corresponding clinker phase and normalizing the totals to 100%. A number of sources of systematic and random error are present in any quantitative analyses, influenced by sample preparation, specimen polishing and etching, and operator experience in point counting and phase identification. A statement of uncertainty should accompany any measurement. Uncertainty analysis for each phase fraction estimate is based on counting statistics. Assume a representative sample

Microscopy of Clinker and Hydraulic Cements

115

Figure 7. SRM 2686a exhibits a heterogeneous silicate distribution and a fine-grained, differentiated matrix. Hydrofluoric acid vapor etching imparts color to the silicates, slightly darkens the aluminates and periclase, while the ferrite phase appears unaffected.

116

Stutzman

Figure 8. SRM 2687 with a hydrofluoric acid vapor etch shows the fine-grained character, the predominance of alite and the undifferentiated matrix.

Microscopy of Clinker and Hydraulic Cements

117

Figure 9. SRM 2688 with a hydrofluoric acid vapor etch shows uniform etching of the hexagonal alite rounded belite showing the lamellar structure, and a matrix of lath-like ferrite (bright white) and aluminate (off-white). Some belite dots occur in the matrix and some ferrite dots occur within the alite crystals.

118

Stutzman

Figure 10. SRM 2688, hydrofluoric acid vapor etch with the dendritic texture of ferrite (upper) and lathlike texture of ferrite (lower), very fine-grained belite within the matrix, fine-grained ferrite inclusions in alite, and decomposition to belite along alite grain boundaries.

Microscopy of Clinker and Hydraulic Cements

119

Figure. 11. Single sampling field for SRM 2688 clinker with a hydrofluoric acid vapor etch and 9-point grid where six points fall on alite, and one each on belite, aluminate, and ferrite.

was collected and that the sample size was reduced using proper practice. The measurement uncertainty of the phase fraction is related to the number of points falling on the phase and the total number of points counted, as described in Hofmänner (1973), Neilson and Brockman (1977), and Howarth (1998). The latter two references provide a more recent look at the sampling and measurement uncertainty associated with point counting, and also provide recommendations on sampling and methods to calculate confidence bounds on estimates from point count data. Using the clinker data provided in Table 2 on four specimens counted to about 3100 points each where N is the total number of counts, n the counts per phase, and f is the inverse of the F probability distribution for (1 − a, 2(N − n + 1), 2n), available from the F table or calculated through a spreadsheet function. The upper, p(n)u, and lower p(n)l, 95% confidence bounds are calculated as: p(n)u = 100 / 1 + ( N − n ) / {( n + 1) f } p(n)l = 100 / 1 + {( N − n + 1) f } / n 

SCANNING ELECTRON MICROSCOPY ANALYSIS The scanning electron microscope provides sets of images that are suitable for processing and analysis. The uniformity of the backscattered electron and X-ray images makes it possible to perform image processing (feature extraction) and analysis (measurements) for quantitative microscopy. SEM analysis is perhaps the only microscopic means to characterize fine-grained,

Stutzman

120

Table 2. Example point count data on four replicate SRM 2686 specimens, mean area fraction and 95% confidence bounds using a light microscope and point count analysis (Kanare 1987). Phase alite belite aluminate ferrite periclase alkali sulf. free lime Totals

Phase alite belite aluminate ferrite periclase alkali sulf. free lime

Counts (sample) 1

2

3

4

1892 708 39 354 87 12 8 3100

1809 781 38 357 107 3 3 3098

1936 637 34 387 105 0 1 3100

1865 730 42 384 63 6 10 3100

Area Fraction Mean

1

2

3

4

3.18 3.30 3.06 3.77 3.58 2.67 3.34

61.0 22.8 1.3 11.4 2.8 0.4 0.3

58.4 25.2 1.2 11.5 3.5 0.1 0.1

62.5 20.5 1.1 12.5 3.4 0.0 0.0

60.2 23.5 1.4 12.4 2.0 0.2 0.3

95% Confidence Bounds

60.5 23.0 1.2 12.0 2.9 0.2 0.2

Area Fraction

Density (g/cm3)

Upper

Lower

61.2 23.7 1.4 12.4 3.2 0.2 0.3

59.8 22.4 1.1 11.5 2.7 0.1 0.1

f 1.0307 1.0355 1.1372 1.0461 1.3664 1.0893 1.3751

multi-phase particles like cement, fly ash, and slag. The cement grinding operation reduces the clinker nodules to a size distribution that spans from about 45 mm to down to about 1 mm, destroying the clinker crystal arrangements that are so useful for phase identification. Light microscopy is of limited use because of the fine particle size and the potential for any etching operation to partially or completely dissolve the finer-sized particles. The multiple SEM phase and chemical imaging modes help overcome these limitations for qualitative and quantitative microscopy (Scrivener 1987; Stutzman 1994, 2007; Bentz and Stutzman 1994; Stutzman et al. 2008; Bullard et al. 2011). While the traditional point count analysis is readily accomplished using a SEM, image processing and analysis using the full image field is common and is amenable to automation in both data collection and analysis. Interestingly, this type of analysis essentially replicates that performed in the early days of microscopy when sketches were made of the fields of view to quantify phase fractions (Insley and Frechette 1955).

SEM imaging of microstructure The combined information from backscattered electron (BE) and X-ray (XR) imaging from the same field of view is used to quantify phase abundance. These two kinds of images are captured simultaneously and registered (same field of view), and present different perspectives (phase and chemistry) that will be used in conjunction with fabric characteristics (framework, matrix, or dispersed phase) to identify and quantify the constituent phases. In the BE compositional image, local brightness is proportional to the individual phase average atomic number ( Z ). The backscatter coefficient h is a measure of the backscattered electron fraction and, following Goldstein et al. (2003), is estimated using the mass fractions and h values for each constituent. Table 3 lists phases that are found in clinker, cement, and

Microscopy of Clinker and Hydraulic Cements

121

Table 3. Common phases in ordinary Portland cement clinker, blast furnace slag, and fly ash, ordered per material with decreasing brightness in SEM-BE imaging, with average atomic number Z and backscatterd electron coefficient h. Phase ferrite free lime alite belite arcanite aluminate-cubic aluminate-orth. aphthitalite anhydrite bassanite gypsum thenardite periclase

Z

η

16.65 16.58 15.06 14.56 14.41 14.34 13.87 13.69 13.42 13.03 12.12 10.77 10.41

0.186 0.188 0.172 0.166 0.165 0.164 0.159 0.159 0.154 0.149 0.138 0.125 0.121

Z

η

Slag merwinite average slag gehlenite melilite åkermanite

13.71 13.36 13.11 12.80 12.25

0.157 0.153 0.150 0.147 0.105

Fly Ash quartz mullite hematite magnetite

10.80 10.69 20.59 21.02

0.125 0.124 0.223 0.227

Phase

pozzolanic mineral additions in descending order of their backscattered coefficient and gray intensity. The contrast between alite ( Z = 15.06) and belite ( Z = 14.56) is relatively strong and their distinction is clear, while that between belite and cubic aluminate ( Z = 14.34) is generally too weak to distinguish these constituents. These values are estimates and the actual backscattered expression of a phase will depend upon any chemical substitution and the image collection dwell time. Longer image collection times may improve the distinction between phases. Typical SEM operating conditions for clinkers and cements are 10 kV accelerating voltage, about 3 nA probe current, which is adjusted to maximize count rates while keeping an X-ray dead time below 40% when collecting X-ray images, and 5 min per frame scan rate (1024 × 768 pixels) to minimize image noise. The magnification is adjusted to retain a 0.75 µm/pixel spatial resolution. Changes in the accelerating voltage affect both the size of the interaction volume, and as such, the spatial resolution, and it affects which peaks will appear, and the relative peak heights, in the X-ray spectral response. Practically, compromises are necessary and 10 kV represents a value that provides a reasonable spatial resolution, improves X-ray response for the lighter elements (Na, Mg, Al), and yet is sufficient to excite the heavier elements (Fe). A 10 keV beam (in BE mode) striking a calcium silicate has a signal interaction volume and an X-ray volume resolution that is about 0.75 µm (J. Davis, personal communication). X-rays are generated as a result of the interaction between the high-energy electron beam and the specimen. The X-ray spectrum consists of the characteristic lines for each element present, represented by peaks on the spectrum, and a background of white radiation. X-ray microanalysis may be used to identify and quantify chemical composition of phases that can be used to generate images of element spatial distribution. The latter capability is particularly useful for image processing as it enables a set of element spatial distribution images to be included in the image processing. XR imaging is necessary for distinguishing between phases that have the same BE coefficient yet are compositionally distinct and for identification of phases that are not easily detected in the BE image; periclase and some alkali sulfates will appear dark, like voids. Combining the XR and BE images allows a degree of phase discrimination that is not available from any single image. An example of the backscattered electron image and X-ray spectrum for aluminate from SRM 2686a is provided in Figure 12. While the SEM images lack the color information of the

122

Stutzman

Figure 12. SRM 2686a BE “compositional image” (upper) with 1) free lime, 2) ferrite, 3) alite, 4) belite, 5) aluminate, 6) periclase, 7) void, and 8) alkali sulfate and (lower) X-ray spectrum for aluminate indicates the presence of its constituent elements by their characteristic X-ray lines.

Microscopy of Clinker and Hydraulic Cements

123

light microscope images, the view is not unfamiliar to the microscopist, and is perhaps easier to view and understand due to the uniformity of appearance for each phase, the flexibility with the wide range of magnification, and the ancillary information provided by X-ray microanalysis. As with the light microscope, identification is made by a combination of phase morphology, association with other constituents (framework grain or matrix phase), BE brightness, and chemistry. For this clinker field of view, there is abundant alite, minor belite, a differentiated matrix with abundant ferrite and little aluminate, widely-dispersed periclase, a cluster of free lime, and some alkali sulfate along void and fracture walls. Figures 13 through 19 illustrate the microstructure of the three NIST SRM clinkers by SEM BE imaging from low to high magnifications. Descriptions of these clinkers were provided earlier with the light microscopy. SRM 2686a (Figs. 13-14) is a medium-grained, heterogeneous clinker with a fine-grained, differentiated matrix, abundant well-dispersed periclase as equant and occasionally dendritic crystals, and common alkali sulfates along pore walls and inter-grain boundaries. SRM 2687 (Figs. 15-17) is a fine-grained, moderately heterogeneous clinker with common belite nests, a fine-grained, poorly differentiated matrix, moderate alkali sulfate, and occasional free lime. The BE imaging mode shows a fine-scale differentiation of the aluminate and ferrite matrix phases in Figures 16 and 17, where the ferrite appears with a web-like texture within the aluminate. This feature was not observable in the light microscope. SRM 2688 (Figs. 18-19) is the coarsest-grained clinker of the three. It has a homogeneous distribution of phases and a well-differentiated matrix of ferrite and aluminate that ranges from very coarse, lath-like crystals to a finer-grained dendritic ferrite with fine-grained aluminate. A BE and X-ray image set from SRM clinker 2686a is presented in Figure 20, with the backscattered electron image labeled BE, and the X-ray images labeled by their respective elements. A common means of image collection stores the spectrum accumulated at each pixel into a file called a data cube. From this file the total spectrum may be calculated for any region and images of element spatial distribution (X-ray images, or maps) may be generated for subsequent analysis. The images displayed here were considered the most useful for phase identification. More element images may be collected or extracted from the data cube if thought necessary. Image analysis will use a subset of these images to reduce redundancies that can confound the mineral phase identification (segmentation) process. An example of redundancy is between the BE and Fe image. Ferrite is the second brightest (Table 3) phase and exhibits a unique grey level. The Fe image duplicates this with an image with inherently greater noise. The use of the Fe image for cements will not add any new information, and may confound the analysis due to the additional noise. Typically for clinker and cements, calcium, silicon, aluminum, magnesium, iron, potassium, sodium, and oxygen are the principal images selected from an initial screening of the data. A visual assessment is usually all that is needed to identify significant images necessary to extract the set of constituent phases. Fly ash and slag are inherently more complex and will require additional images. An example may be seen with aluminate and belite having distinct chemical compositions, yet exhibit a similar backscattered electron coefficient. Figure 20 illustrates this with relatively large, rounded belite grains appearing at the same grey level as the matrix-phase aluminate. However, belite contains appreciable silica while aluminate contains aluminum, so use of one or the other of the X-ray images will serve to distinguish these phases. Similarly, the calcium sulfate addition (for cement) and the alkali sulfates are usually difficult to see given the high brightness and contrast of the BE image. The X-ray images Ca, S, K, and Na may be used to aid in their distinction. The mineral phase identification process may be visualized through a process whereby the BE and XR images are merged into a red-green-blue composite image. A useful combination for cement color composites is BE = red, Mg = green and Al = blue to see the belite–aluminate

124

Stutzman

Figure 13. SRM 2686a at low magnifications illustrates distribution of phases and porosity. Abundant alite, clustering of belite, a differentiated matrix comprised mostly of ferrite and some fine-grained aluminate, and uniform distribution of periclase is typical for this clinker.

Microscopy of Clinker and Hydraulic Cements

125

Figure 14. SRM 2686a at higher magnification with belite inclusions in alite crystals, a differentiated matrix dominated by the ferrite phase, and both equant and dendritic (lower image) periclase.

126

Stutzman

Figure 15. SRM 2687 is a fine-grained clinker with abundant alite, some small nests of belite, variable porosity, and abundant, undifferentiated matrix.

Microscopy of Clinker and Hydraulic Cements

127

Figure 16. For SRM 2687, higher magnification images begin to reveal details of the matrix, which consists of both a medium and fine-grained ferrite.

Figure 17. SRM 2687 matrix appears to be a mixture of medium-grained ferrite, massive aluminate and a web-like textured ferrite intermixed with the aluminate.

128

Stutzman

Figure 18. SRM 2688 is a coarse-grained clinker with uniform phase distribution, no free lime, a mediumto coarse-grained well differentiated matrix, and occasional presence of periclase and alkali sulfates within the matrix and along grain boundaries.

Microscopy of Clinker and Hydraulic Cements

129

Figure 19. SRM 2688 at higher magnifications showing the belite and ferrite inclusions within alite, and a medium- to coarse-grained, well-differentiated matrix.

130

Stutzman

Figure 20. Backscattered electron (BE) and X-ray image set for SRM 2686a that will be used to segment the sample into the constituent phases.

Microscopy of Clinker and Hydraulic Cements

131

distinction, limestone, and the contribution of Mg in identification of periclase. In this case, the addition of the aluminum image makes the aluminate appears purple compared to the dark red of the belite, allowing their distinction. Another useful combination is BE= red, K=green and S=blue to highlight the locations of the gypsum addition to the cement, and alkali sulfates within and along cement grains. Processing the image set serves to enhance the distinction between the constituent phases by reducing noise and unwanted signals. This may be accomplished, as needed, by use of a median filter to reduce noise, yet retain edge definition, followed by clipping to establish a lower threshold to eliminate the background signal introduced by the continuous background of the X-ray spectrum. If difficulties are encountered in subsequent processing, the median filtered image may be useful in achieving a more successful segmentation. All these operations may be interactively applied using most image analysis codes, and are illustrated here using ImageJ (this processing and analysis code is available from the National Institutes of Health, http://rsbweb.nih.gov/ij/index.html).

Image processing Traditional image processing methods use Boolean logic and mathematical operators to threshold phases individually using one or more images from a set. These operations create binary images for each phase that are subsequently merged into a composite image indexed by phase. For example, ferrite is typically a bright phase and may be segmented using only the BE image on that basis. This may also be accomplished using thresholding tools where the upper and lower grey level bounds are interactively set prior to creating the binary image. A more difficult example may be found in the aluminate phase in Figure 21 where in the BE image, aluminate exhibits a gray level similar to that of belite, making these phases indistinguishable from the BE image alone. However, aluminate contains appreciable amounts of aluminum so we can make the distinction using a combination of the BE and aluminum images. This example makes use of a thresholding tool to set upper and lower gray level bounds for the BE and Al image to create the respective binary images, BE-t and Al-t. The Al-t is subtracted from BE-t to create a binary of belite distribution and the belite binary is subtracted from the BE-t to create a binary of aluminate distribution. Some general criteria for making phase distinctions for clinker and cement are presented below. Criteria for identification and segmentation of cement and clinker phases: • • • • • • • • • • •

Free lime – strong calcium, rounded grains, brightest BE Ferrite – high iron, prismatic matrix phase, bright BE Aluminate – matrix phase, high aluminum, low magnesium, intermediate BE Belite – rounded grains, low aluminum, intermediate BE Alite – medium-high BE, principal phase, hexagonal shape Periclase – equant to dendritic habit, may occur anywhere, high magnesium, low BE Alkali sulfate – along grain boundaries or voids, high sulfur, high K, medium Na Gypsum – high sulfur, calcium, low potassium, low BE Quartz – high silicon, intermediate BE Kaolin – high aluminum, high silicon Slag – sharp, angular grains, high silicon, magnesium, and aluminum, low BE

While this approach works well, it is tedious since constituents are segmented individually and then merged into a final image. Difficulties in reconciling the composite image are encountered with this process as areas of overlapping phase assignments and holes from incomplete segmentations will need to be resolved. In certain circumstances, however, it can be a rapid and useful means to isolate specific features for quantitative analysis.

132

Stutzman

Figure 21. Segmentation of belite and aluminate, which have similar BE grey levels, requires the additional Al image, thresholding the BE (BE-t) and Al (Al-t) images and finally image subtraction to generate binary images of belite and aluminate distribution. Field width = 400 mm.

Multi-spectral processing developed for analyzing hyperspectral remote sensing data provides an alternative method of image processing. Van Niekerk (2003) and Lydon (2005) used one such code, Multispec© (Landgrebe 2003; Landgrebe and Biehl 2011) for geological studies of rocks and the success of their efforts prompted application of this type of code to processing SEM image sets. Combinations of BE and XR images of cementitious materials are displayed to identify the constituent phases, to establish a user-defined training set of image regions typical of each phase (including voids), and to classify each pixel into the phase group to which it most likely belongs. This is in contrast to the sequential creation of binary images and eventual reconciling of image overlaps and dropouts. The suitability of the classifier and the operator-designated training set in segmentation may be assessed by a visual

Microscopy of Clinker and Hydraulic Cements

133

evaluation of the resulting segmented image and by evaluation of the accuracy in which the training set was properly classified. Multi-spectral processing begins with reading the subset of BE and XR images into a suitable image processing software application. Different combinations are displayed for phase identification and establishing the user-defined training set (Fig. 22). By merging individual images into a red-green-blue channel pseudo-colored image, the constituent phases may be highlighted and identified. Regions of each phase (including void space) are defined with a training set, with at least 100 pixels each to define each phase. The next step is to select a classifier and group like pixels and review the performance of that classifier on the training set. This may be accomplished graphically using an X-Y plot, demonstrated in Figure 23 with the BE image on the X-axis and aluminum channel on the Y-axis. The ferrite phase, being the brightest, is plotted high along the X-axis, reflecting it’s large BE signature. The distinction between belite (Class 2) and aluminate (Class 3), while having similar BE signatures, becomes clear with the addition of the aluminum image data. The classifier extrapolates the training set characteristics for each phase for the entire image to complete the classification. Reviewing the performance of the classifier on the designated training field pixels provides an initial sense of the success of the segmentation (Table 4). A successful classification will have high reference accuracies for each phase. The miss-classifications in the matrix to the right in Table 4 provide some clues to the nature of any miss-classifications. Often, the addition of a larger number of training pixels will reduce the classification error through an improved definition of the range of class attributes. The best assessment of the segmentation will be a comparison of the segmented image to the original BE image. Figure 24 shows the original BE and thresholded binary images for each phase in the field of view. Examination of the phase assignments provides a visual assessment on the success of the segmentation. The resulting image is typically displayed as a single image where each phase is uniquely identified by a color and an index value. These are important for visualizing the results and for subsequent measurements of phase abundance and surface area. Since clinker is a heterogeneous product, some measure of the uncertainty is required (Table 5). The results of a clinker SRM development by image analysis and X-ray powder diffraction are examined next.

Direct methods for development of standard reference materials The data set of SEM and XRD analyses for the development of SRM 2686a provides an opportunity to examine the results from two unique, direct methods for phase analysis. These data were used to establish consensus means and consensus uncertainty values for phase abundance for this SRM clinker. SRM certification requires at least two independent methods of analysis when no single method can provide the necessary level of accuracy and/or when there is no single method whose sources of uncertainty are well-understood and quantified. A common goal in the analysis of such data is to compute a consensus mean value and consensus uncertainty to that value (Stutzman and Leigh 2002). Error in microscopy may be due to incorrect classification as a result of operator misidentification of a phase, or due to irresolvable finely divided interstitial phases, or edge effects, which may preclude their proper classification. XRD data may be biased due to improper sample preparation, incomplete identification of phases as a result of difficulties in resolving weak diffraction peaks, the non-suitability of the structure models, effects of microabsorption (if significant), and the inability to identify and control correlations between variables. References on specimen preparation for XRD analysis include Bish and Reynolds (1989) and Bish and Plötze (2010). References specific to Rietveld analysis of XRD data include Young (1995) and McCusker et al. (1999). About 300 kg of clinker was obtained from a cement plant as sieved material in a size interval from 4 mm to 13 mm, representing a sampling of nodules that underwent a common

134

Stutzman

Figure 22. Selection of typical regions for each phase establishes the training set for classification. Field width = 250 mm.

Figure 23. The addition of the aluminum image (Y-axis) to the BE image (X-axis) shows the clustering facilitating the belite (2) and aluminate (3) distinction in the training set. Additional phases shown include alite (1), and ferrite (4).

Microscopy of Clinker and Hydraulic Cements

135

Table 4. Output from the classification provides a means to assess the re-classification of the training set pixels, the types of miss-classifications, and area fractions. TRAINING CLASS PERFORMANCE (Resubstitution Method) Number of Samples in Class Project Class Name

Class Reference Number Number Accuracy Samples (%)

alite belite aluminate ferrite periclase alkali sulfate free lime void

1 2 3 4 5 6 7 8

100.0 100.0 98.0 96.1 100.0 99.9 100.0 100.0

TOTAL Reliability Accuracy (%)

1

2

3

4

5

6.

7

8

1362 746 252 233 644 1536 1450 848

1362 0 0 0 0 0 0 0

0 746 2 0 0 1 0 0

0 0 247 0 0 0 0 0

0 0 1 224 0 0 0 0

0 0 0 0 644 0 0 0

0 0 0 0 0 1534 0 0

0 0 0 9 0 0 848 0

0 0 2 0 0 1 0 1450

7071

1362 749 100 99.6

247 100

225 99.6

644 100

1534 857 1453 100 98.9 99.8

CLASS DISTRIBUTION FOR SELECTED AREA Class 1 2 3 4 5 6 7 8

Number Samples

Percent

alite belite aluminate ferrite periclase alkali sulfate free lime void

116,731 12,084 4,294 11,639 6,670 6,181 7,492 26,909

60.8 6.3 2.2 6.1 3.5 3.2 3.9 14.0

TOTAL

192,000

100.0

Table 5. Phase abundance results from SEM / image analyses on n = 20 specimens, expressed as mass fractions. alite

belite

aluminate

ferrite

periclase

alkali sulfate

Average

64.2

19.7

2.2

10.1

3.2

0.7

1s

0.5

0.7

0.1

0.2

0.1

0.1

relative accuracy

0.3

1.6

2.7

0.8

0.7

5.2

95% CI

0.2

0.3

0.1

0.1

0.0

0.0

136

Stutzman

Figure 24. SEM BE image, individual phase binary images, and calculated mass fractions for this image field. This is typically rendered as a single, indexed image with pseudo-color to show phase assignments.

thermal history. This sample was not intended to be representative of the bulk clinker production, but was sampled and processed with the intent of creating a homogeneous lot of clinker fragments. The nodules were stage-crushed using a jaw crusher and sieved to capture the size fraction between 3 mm and 4 mm. The fragments were homogenized using a V-blender and packaged in containers each containing about 7 g of clinker. Sampling followed a random-stratified scheme, taking 16 of the containers for XRD analysis and 10 samples for SEM analyses. In preparation for the XRD analysis, each container was split into duplicates using a cone-and-quarter method, and the splits were

Microscopy of Clinker and Hydraulic Cements

137

ground individually to a particle size of less than 10 µm. The splits were analyzed in triplicate. Microscopy samples were potted and polished as described earlier. Between 7 and 12 fields of view were collected for each clinker specimen in the SEM and the images were processed as described above. Two individuals performed these analyses independently. The box plot provides a graphical comparison of the XRD and microscopy results through assessment of the alignment or misalignment of median values and differences in interquartile ranges. The phase abundance analysis characterized by important features of the box plot: 1. The width of each box is proportional to sample size, 2. The median value is used for its resistance to outliers, and is identified by the “X” inside the box, 3. The interquartile range (“middle half”) of the data are represented by the vertical extent of the box, 4. The extremes (minimum and maximum) are represented by the ends of the vertical lines projecting out of the box, and 5. Circles outside the extremes of each box represent outliers.

Phase estimates by microscopy and quantitative XRD The XRD data for alite (Fig. 25) and belite (Fig. 26) exhibit greater precision than the microscopy data and, while the boxes overlap, both phase median estimates by XRD are lower than those by microscopy. The heterogeneity of the phase distribution is responsible for the greater variability in the microscopy. Preferred orientation corrections were not made for alite, a phase that cleaves in a manner that is subject to orientation effects. Preliminary tests with orientation corrections where intensity and orientation were simultaneously refined indicated that they tend to increase bias, probably as a result of correlations between refined variables. Given that all the phase abundance values are correlated, any orientation corrections will influence all phase estimates, so they were not used. The XRD values for aluminate (Fig. 27) and ferrite (Fig. 28) are similar to these from microscopy but are consistently lower. Periclase estimates (Fig. 29) exhibit reasonably close agreement between XRD and microscopy. The alkali sulfate data (Fig. 30) represent a sum of arcanite and aphthitalite since the distinction was not made by SEM imaging.

Certified values by consensus means Certified values (Table 6) are unweighted averages of diffraction and microscopy mean values. The Type B on bias (BOB) approach (ISO GUM) to consensus means was used to establish the certified phase values (Levenson et al. 2000). The method is designed to handle cases where the number of analytical methods is small (2-5), and the ordinary sample standard deviation is an inadequate estimate of the uncertainty of the systematic effects (Levenson et al. 2000).

Application to cements The SEM provides an opportunity to quantitatively describe fine-grained multi-phase powders in a way not possible using light microscopy. Cement imaging poses a more complicated problem with material preparation, image interpretation and analysis. The use of a harder epoxy has proven useful in the polishing of an encapsulated powder by reducing particle plucking and having an epoxy with polishing characteristics closer to that of the powder. Data interpretation is complicated due to the loss of many phase associations due to the grinding. An advantage to powdered materials though is that the particle size reduction tends to produce a more homogeneous sampling compared to clinker. There are numerous multi-phase particles to assist in initial phase identification and the images are typically uniform such that criteria for identification made in one region carries through across the entire image field. Additional phases, such as cal-

(Fig. 29)

(Fig. 26)

Figures 25-30. Box plots showing the results of QXRD and microscopy estimates on phase abundance for alite (Fig. 25), belite (Fig. 26), aluminate (Fig. 27), ferrite (Fig. 28), periclase (Fig. 29), alkali sulfates, where the XRD determinations of arcanite and aphthitalite have been summed (Fig. 30).

(Fig. 28)

(Fig. 25)

(Fig. 30)

(Fig. 27)

138 Stutzman

Microscopy of Clinker and Hydraulic Cements

139

Table 6. Certified values for SRM 2686a; mass fractions (%) with Mean and k = 2 (95%) expanded uncertainty (2Uc) (Stutzman et al. 2008). SRM 2686a

Alite

Belite

Aluminate

Ferrite

Periclase

Alkali Sulfate

Mean 2Uc

63.53 1.04

18.80 1.10

2.46 0.39

10.80 0.84

3.40 0.23

0.86 0.17

cium sulfates, and pozzolanic additions, such as limestone, fly ash, or slag, must be considered. Previewing the image set generally provides for identification of these constituents (Table 3). Figure 31 shows the original BE image of a polished cement section and the resulting segmented image. From the segmented, indexed image, measurements of area fraction, surface perimeter fraction, and spatial distribution may be made for the constituent phases (Bentz et al. 1999; Bullard et al. 2011). Taking these data and combining them with X-ray computed tomography images of real cement particles has enabled the generation of virtual cement particles (Fig. 32) with the phase and textural characteristics of actual industrial cements, which has been invaluable in the development of virtual cement hydration models (Bullard et al. 2011).

SEM imaging of fly ash Since the 1950’s, supplementary mineral admixtures (SCM) have been used in conjunction with cements. These are typically waste products from other industrial processes, yet have the potential of adding value to concrete products because of reduced cost and increased durability. SCMs fall into two general classes, pozzolanic and hydraulic. Pozzolans react with the calcium hydroxide in Portland cements upon hydration to form additional hydration products, such as calcium-silicate-hydrate, one of the principal phases in hardened cement paste. Pozzolans such as fly ash and slag are generally high in SiO2 and Al2O3. Limestone is a special case as it too will react with the pore solution of hydrating cement, accelerating the reactions of some of the cement phases and the formation of monocarbonate (Taylor 1997). Hydraulic mineral admixtures are capable of reacting with water and setting without the addition of Portland cement. The particle size of these materials may be as fine as, or finer, than that of cements and may also affect the system as filler particles between the hydrating cement grains and as nucleation sites for hydration products (Taylor 1997). Fly ash is a waste byproduct from coal-burning power plants (Fig. 33). The mineral fraction of the coal, principally clays, calcite, quartz, and pyrite, becomes partially molten during the combustion process, cools post-combustion in the flue gasses, and is collected in precipitators designed to remove the ash. The ash is comprised of some crystalline phases, but is predominantly amorphous and contains some residual carbon. The mineral, glass, and carbon portions of the ash reflect a combination of the organic and mineral composition of the source coal, the burning and precipitation processes at the power plant, and any subsequent processing that may serve to beneficiate the ash. The glassy portion is thought to be the reactive portion of the ash and the approximately spherically shaped particles are a result of its formation in the flue gasses. Blast furnace slag (Fig. 34) is a waste product from the production of iron Snellings et al. (2012, this volume). The slag is skimmed from the molten iron and is quenched. The process produces an angular, glassy, uniform calcium alumino-silicate product that is crushed and graded based upon particle size, sulfur and sulfate content, and pozzolanic activity. The characteristic shapes of these materials make them relatively easy to identify in cements when examining powder mounts with a light microscope, and in polished section in the scanning electron microscope.

140

Stutzman

Figure 31. Segmentation of a polished section of hydraulic cement, following the same techniques used for clinkers, resulting in a segmented, indexed image. Field width = 500 mm.

Fly ash and slag phase characterization are difficult because of the lack of crystalline phases and limited knowledge of their glassy components (Chancey et al. 2010). The mineral constituents of these two latter materials can exert significant influence on their performance in a concrete, yet are not generally characterized. The glass phase(s) of these materials is not quantified and is typically considered as a single, homogeneous phase, which may not be the case. Spot X-ray analysis shows that slag is often comprised of calcium, silicon, and magnesium, with smaller amounts of aluminum, sulfur, and potassium. While the homogeneity and angularity of the slag is initially apparent in the backscattered electron image, local regions high in silicon and in aluminum indicate a heterogeneity resulting from very finely divided discrete phases.

Microscopy of Clinker and Hydraulic Cements

141

Figure 32. Virtual cement particle created by combining the results from analysis of SEM image data and X-ray computed tomography. Reconstruction of an actual cement particle is used in three-dimensional cement hydration modeling.

The SCMs are a case where the mixture of crystalline and glassy phases is ideally addressed through multiple analyses. Combined quantitative XRD and SEM with image analysis provide a means to explore the complicated microstructures of these materials to better understand their phase composition and textures and perhaps to explore potentially new methods for classification. The analytical methods are applied in the same manner as used for cements; a powder for XRD using an internal standard, and polished powder mounts for SEM.

Summary Microscopy has played a significant role in developing our understanding of cementitious materials compositions and their effects on cement and concrete performance. It continues to play an important role in the evaluation of cement clinker kiln operations and is a relatively straightforward quantitative tool for assessing phase compositions. Point-counting for quantitative phase abundance is a mature method, yet is one of the few direct methods to determine clinker phase compositions. The application of the scanning electron microscope with X-ray microanalysis complements light microscopy by not only providing analyses of clinker, but also of the more difficult fine-grained powders of Portland cement and pozzolans like fly ash, and slag. These images may also be point-counted, but image processing and analysis provides a means for full-field quantitative measurements on area fraction and surface perimeter fraction, and spatial distribution for the constituent phases. These data, coupled with X-ray computed tomography, is providing the means to generate three-dimensional particles that capture the characteristics of phase abundance and texture and are invaluable in the development of three-dimensional computer simulation models of cement hydration.

142

Stutzman

Figure 33. Fly ash BE and XR image set illustrate the chemical and compositional complexity of fly ash.

Microscopy of Clinker and Hydraulic Cements

143

Figure 34. Slag BE and XR images show the uniform, angular nature. Subtle differences in particle chemistry allow division of this slag into two distinct subgroups.

References ASTM C150 / C150M-11 (2011) Standard Specification for Portland Cement. American Society for Testing and Materials West Conshocken, PA, Annual Book of ASTM Standards, (4.01): Cement ASTM C595 / C595M-11 (2011) Standard Specification for Blended Hydraulic Cements. American Society for Testing and Materials West Conshocken, PA, Annual Book of ASTM Standards, (4.01): Cement ASTM C1157 / C1157M-11 (2011) Standard Performance Specification for Hydraulic Cement. American Society for Testing and Materials West Conshocken, PA, Annual Book of ASTM Standards, (4.01): Cement

144

Stutzman

ASTM C1356 (2011) Standard Test Method for Quantitative Determination of Phases in Portland Cement Clinker by Microscopical Point-Count Procedure American Society for Testing and Materials West Conshocken, PA, Annual Book of ASTM Standards, (4.01): Cement Bates PH, Klein AA (1917) Properties of the calcium silicates and calcium aluminates occurring in normal Portland cement. Technological Papers of the Bureau of Standards 78:1-38 Bentz DP, Stutzman PE (1994) SEM analysis and computer modeling of hydration of Portland cement particles. In: Petrography of Cementitious Materials. American Society for Testing and Materials, ASTM STP 1215:60-73 Bentz, DP, Stutzman PE, Haecker CJ, Remond S (1999) SEM/X-ray imaging of cement-based materials. Proceedings of the 7th Euroseminar on Microscopy Applied to Building Materials (EMABM), 457-466 Bhatty JI, Miller FM, Kosmatka S (eds) (2004) Innovations in Portland Cement Manufacturing. Portland Cement Association, Skokie/IL Bish DL, Plötze M (2010) X-ray powder diffraction with emphasis on qualitiative and quantitative analysis in industrial mineralogy. Advances in the characterization of industrial minerals. EMU Notes Mineral 9:35-76 Bish DL, Reynolds RC Jr (1989) Sample preparation for X-ray diffraction. Rev Mineral 20:73-99 Bogue RH (1955) The Chemistry of Portland Cement. 2nd ed, Reinhold, New York Bogue RH (1961) Origin of the special chemical symbols used by cement chemists. J PCA Res Develop Lab 3:20-21 Brown LS (1948) Microscopical study of clinkers in long-time study of cement performance in concrete. PCA Bull 26:877-933 Bullard JW, Lothenbach B, Stutzman PE, Snyder KA (2011) Coupling thermodynamics and digital image models to simulate hydration and microstructure development of Portland cement pastes. J Mater Res 26:609-626 Campbell DH (1999) Microscopical Examination and Interpretation of Portland Cement and Clinker. 2nd Edition. Portland Cement Association, Skokie/IL Campbell DH, Galehouse JS (1991) Quantitative clinker microscopy with the light microscope. Cem Concr Aggr 13(2):94-96 Chancey RT, Stutzman P, Juenger MCG, Fowler DW (2010) Comprehensive phase characterization of a class F fly ash. Cem Concr Res 40:146-156 Chayes F (1956) Petrographic Modal Analysis. An Elementary Statistical Appraisal. Wiley, New York EN197-1 (2011) Cement – part 1: Composition, specifications and conformity criteria for common cements. English version. CEN, Brussels: pp 29 Fundal E (1980) Microscopy of cement raw mix and clinker. F. L. S. Review 25:1-15 Goldstein J, Newbury DE, Joy DC, Lyman CE, Echlin P, Lifshin E, Sawyer L, Michael JR (2003) Scanning Electron Microscopy and X-Ray Microanalysis. A Text for Biologists, Materials Scientists, and Geologists. 3rd edition. Springer Verlag, Berlin Hewlett PC (ed) (1977) Lea’s Chemistry of Cement and Concrete. 4th Edition. John Wiley, New York Hofmänner F (1975) Microstructure of Portland Cement Clinker. Holderbank Management and Consulting, Ltd., Holderbank/CH Howarth RJ (1998) Improved estimators of uncertainty in proportions, point-counting and pass-fail test results. Am J Sci 298:594-607 Insley H, Frechette V (1955) Microscopy of Ceramics and Cements. Chapter 5, Special techniques. Academic Press, New York, p 177-207 Joint Committee for Guides in Metrology (2008) Evaluation of measurement data. Guide to the expression of uncertainty in measurement (GUM). 1st edition, corrected version 2010. BIPM, Sèvres, France. JCGM 100:1-134 Kanare HM (1987) Production of Portland cement clinker phase abundance standard reference materials. Final Report, CTL Project No. CRA012-840, Construction Technology Laboratories, Skokie/IL Kosmatka SH, Kerkhoff B, Panarese WC (2002) Design and Control of Concrete Mixtures, PCA Engineering Bulletin 001. Portland Cement Association, Skokie/IL Landgrebe D (2003) Signal Theory Methods in Multispectral Remote Sensing. Wiley Interscience, New York Landgrebe D, Biehl L (2011) An Introduction to Multispec©. http://dynamo.ecn.purdue.edu/~biehl/MultiSpec, retrieved 04 January 2012 Le Chatelier H (1905) Experimental Researches on the Constitution of Hydraulic Mortars. English translation by JL Mack, McGraw, New York Levenson MS, Banks DL, Eberhardt KR, Gill LM, Guthrie WF, Liu HK, Vangel MG, Yen JH, Zhang NF (2000) An approach to combining results from multiple methods motivated by ISO GUM. J Res NIST 105(4):571-579 Lydon JW (2005) The measurement of the modal mineralogy of rocks from SEM imagery: the use of Multispec© and ImageJ freeware. Geol Survey Canada Open File 4941:1-37

Microscopy of Clinker and Hydraulic Cements

145

McCusker LB, Von Dreele RB, Cox DE, Louer D, Scardi P (1999) Rietveld refinement guidelines. J Appl Crystallogr 32:36-50 Miller (1981) Microscopy as an aid in evaluation of mix burn ability and clinker formation. Proc. 3rd Int Conf Cem Micro, Duncanville, TX, 181-192 Neilson MJ, Brockman GF (1977) The error associated with point-counting. Am Mineral 62:1238-1244 Rankin GA (1915) The constituents of Portland cement clinker. J Ind Eng Chem 7:466-474 Rankin GA, Wright FE (1915) The ternary system CaO-Al2O3-SiO2; with optical study. Am J Sci 39:1-79 Scrivener KL (1987) The microstructure of anhydrous cement and its effect on hydration. Mater Res Soc Symp Proc 85:39-46 Snellings R, Mertens G, Elsen J (2012) Supplementary cementitious materials. Rev Mineral Geochem 74:211278 Stutzman PE (1994) Scanning electron microscopy imaging of hydraulic cement microstructure. Cem Concr Comp 26/8:957-966 Stutzman PE (2007) Multi-spectral SEM imaging of cementitious materials. Proc. 29th Int Conf Cement Micr, Québec City, Canada Stutzman PE, Bentz DP (1993) Imaging of cement and image-based simulation of hardened cement microstructure. Proc 15th Int Conf Cem Micr, Duncanville Texas, 312-323 Stutzman PE, Leigh S (2002) Phase composition analysis of the NIST reference clinkers by optical microscopy and X-ray powder diffraction. NIST Technical Note 1441:1-44 Stutzman PE, Lespinasse G, Leigh S (2008) Compositional analysis and certification of NIST reference material 2686a. NIST Tech Note 1602:1-49 Taylor HFW (1997) Cement Chemistry. 2nd Edition. Thomas Telford, New York Van Niekerk D (2003) Modal analysis and phase identification in meteorite thin sections using freeware for PC. Annual Lunar and Planetary Sci Conf, League City, TX http://www.lpi.usra.edu/meetings/lpsc2003/ pdf/2015.pdf Young RA (ed) (1995) The Rietveld Method. Int Union of Cryst Monographs on Crystallography 5:1-308

4

Reviews in Mineralogy & Geochemistry Vol. 74 pp. 147-165, 2012 Copyright © Mineralogical Society of America

Industrial X-ray Diffraction Analysis of Building Materials Roger Meier PANalytical Lelyweg 1, PO Box 13 7600 AA Almelo, The Netherlands e-mail: [email protected]

Jennifer Anderson PANalytical 117 Flanders Road Westborough, Massachusetts 01581, U.S.A.

Sabine Verryn PANalytical (Pty) Ltd 363 Oak Avenue Ferndale 2194, South Africa

XRD Analytical and Consulting cc 75 Kafue Street Lynnwood Glen 0081, South Africa

e-mail: [email protected]

ABSTRACT X-ray analysis of polycrystalline powder samples has grown beyond its roots in the world of laboratory research and is regarded as one of the most powerful industrial process-control tools in the field of building materials and minerals. This is the key to characterize the element and the phase composition of the material. This is largely due to the development of industrial X-ray analytical systems, which have transformed these advanced analytical techniques born in the laboratory into a robust, workmanlike and easy-to-use tool for today’s heavy industries. X-ray diffraction (phase analysis) opens enormous possibilities for process and quality control. Moreover, the recent development of ultra-high-speed X-ray detectors allows for “on the fly” quantitative X-ray diffraction analysis and truly interactive process control. Hydration of cements can be studied relative ease. Additionally Computed X-ray Tomography (CT) can yield valuable information in the study of mortars and concrete.

INTRODUCTION Heavy duty environments like those found in the building materials and minerals industry are notoriously difficult for sensitive test and process-control equipment. The need to often work in dusty conditions, high humidity and extreme temperatures places special demands on such equipment, which must combine ruggedness, ease of use and reliability with the precision and sensitivity of a laboratory instrument. This transition from laboratory instrument to industrial system is not always simple but is nevertheless essential if the industry is to benefit from the latest analytical tools. One of the most recent analytical tools to make this transition is X-ray diffraction (XRD). Although the power of XRD as the only analytical technique capable of dis1529-6466/12/0074-0004$05.00

DOI: 10.2138/rmg.2012.74.4

Meier, Anderson, Verryn

148

tinguishing crystalline phases is well recognized, it has long been considered a tool best suited to the laboratory, because of its low measurement speed and need for specialist knowledge. All this had changed, however, with the introduction of the latest generation of industrial X-ray equipment. These instruments are not only able to withstand the rigors imposed by an industrial regime; they are also designed to be exceptionally easy to use even by non-specialist operators. Moreover, even speed is no longer an issue with the introduction of ultra-fast linear and two dimensional XRD detectors. Figure 1 shows the differences between a point detector and linear detector setup. The linear detector can be regarded as a multiple detector array that collect the data in parallel and gain therefore in the overall measurement speed. The introduction of this silicon-based technology gives X-ray detectors up to 150-fold increase in data acquisition speed. This means that a scan formerly requiring three hours of data collection time is now recorded in less than two minutes with, moreover, no compromise on resolution. Various sample chambers can be used to study hydration in different environments and under different conditions. Modern X-ray diffraction equipment allows for fast exchange of sample stages as well as optics. This paper describes the industrial applications, rather than the detailed description of methods such as Rietveld analysis, which is covered in great detail by Aranda et al. (2012, this volume). Different systems and software packages are available and in use. A volume of Reviews in Mineralogy, dedicated to Modern Powder Diffraction (Bish and Post 1989), comprised a series of key articles on the basics of powder diffraction, sample preparation and synchrotron and neutron powder diffraction. In that volume, quantitative phase   analysis was   discussed in detail by Snyder and Bish (1989). There the Reference Intensity Ratio

a  a     

b  b  Figure 1. Schematic of a Bragg-Brentano diffractometer with (a) point detector and (b) linear detector.

Industrial X-ray Diffraction Analysis of Building Materials

149

approach (Davis 1990) (also known as Chung method – see Chung 1974a,b, 1975), the method of standard additions (also known as spiking method) and the full pattern-fitting approach using both the Rietveld method (Rietveld 1969; Young 1993) and the observed patterns method (Smith et al. 1987) are presented and are therefore not discussed much further here. Although XRF analysis is a valuable tool and often quoted here, in this paper the authors want to show that quantitative XRD and statistical evaluation can be taken from the laboratory environment into the industrial environment for plant control. Often XRF results are used to calculate the Bogue values on the basis of equilibrium conditions. In the past the results obtained by the Bogue method were mostly close to the real phase composition, because the process followed conditions close to the modeled equilibrium by Bogue. Nowadays, due to the use of alternative fuels, which introduce many “new” elements into the process and can change the burning conditions the trend towards a higher material throughput, where the materials have not enough time to reach the equilibrium condition, the results obtained by Bogue may significantly different compared with the “true” results by microscopy and X-ray diffraction.

METHODOLOGY Phase identification with XRD In the production of minerals and building materials, accurate information about the phase composition is essential for determining the quality of the semi-finished and finished products and allowing manufacturers to optimize the efficiency of their process. Although many nondestructive analytical techniques exist to measure the concentrations of elements within a sample, in almost all cases the properties of the material are not determined by the relative amounts of elements alone, but more importantly by the crystalline phases present. Qualitative phase analysis is one goal of an X-ray diffraction experiment. Establishing which phases are present in a sample is usually the first step of a whole series of analyses and form the basis of investigations on how much of each phase is present (quantitative phase analysis). All crystalline materials have their own unique, characteristic X-ray finger print (or stick pattern), based on their crystal structure. When diffraction data for a particular sample is compared against a database of known materials, the crystalline phases within the sample can be identified (Fig. 2). This classic X-ray diffraction analysis approach takes into account the peak positions and net profile data of every peak in a single diffraction pattern. The identification of mineral phases in a sample is a powerful tool for characterization of materials. Modern XRD software provides search-match routines to aid in phase identification.

Phase quantification by using X-ray diffraction data Possible phase composition of Portland cement in cement plants is traditionally estimated using optical microscopy by either point counting or image analysis, which is highly operator dependent. The preparation and analysis time limits its usefulness for process control. Alternatively, it is derived from chemical analysis, usually by X-Ray Fluorescence employing the Bogue method (Bogue 1929; Taylor 1989). However, these methods have sources of error, such as sampling difficulties for microscopic investigations and the use of inappropriate phase compositions using the Bogue method (Aranda et al. 2012, this volume). Quantitative XRD gives a direct measurement of the phase content of building materials, and, as shown in the introduction, fast XRD systems for quantitative XRD analysis of building materials are becoming more widespread. XRD can even be used for plant and kiln control. The classical calibration methods. Quantitative phase analysis using diffraction data can be done with a number of methods. These methods are described in various classical works

150

Meier, Anderson, Verryn

Figure 2. Principle of the phase identification of multiphase diffraction diagrams.

(Nuffield 1966; Klug and Alexander 1974; Cullity 1978; Snyder and Bish 1989; Zevin and Kimmel 1995; Jenkins and Snyder 1996). The calibration method assumes that the phase composition of the sample is known. The intensity information of one peak (or a group of peaks), belonging to the phase to be analyzed, is used to quantify the abundance of that phase in the sample. The measured peak area or peak intensity is compared against a calibration curve built from samples with known concentrations of the phase (or phases) in question. There are several drawbacks to these methods for the building materials industry. Standards used for this method are often expensive, unstable or otherwise difficult to obtain. Additionally, the method cannot be extended to include all phases relevant to the cement industry. There are extensive overlapping peaks in the diffraction patterns of cementitious materials and several phases exhibit preferred orientation making the calibration of many phases impossible. Only very few phases can be calibrated for, these include free lime in clinker and in some cases calcite in cement. Aldridge (1982) also states that these XRD methods were generally unsatisfactory. Rietveld analysis. Nowadays, quantitative XRD analysis from powder diffraction data is mainly based on the Rietveld method (Rietveld 1969; Hill and Howard 1987; Bish and Howard 1988; Bish and Post 1993; Madsen and Scarlett 2000, 2008). The Rietveld method was first described in 1966 by Hugo M. Rietveld (Rietveld 1966, 1969) to refine crystal structures from powder data measured on neutron diffractometers. The method was first reported at the seventh Congress of the IUCr in Moscow in 1966 (Rietveld 1966). Aranda et al. (2012, this volume) address in detail the description of the Rietveld method, therefore here only a brief summary, as this paper deals with the actual application in industry. The Rietveld method is based on the idea that the overall scale of a phase is proportional to its abundance in a phase mixture. The intensity relationship between peaks in a diffraction pattern is governed by the crystal structure of each of the phase present in the mixture. A calculated model of the diffraction data is optimized to match that of the observed data and to determine the quantities of the phases present. Sophisticated Rietveld analysis software packages perform a comparison of measured and calculated profiles and minimize the differences between the profiles by a least-squares fit of many parameters. This is a profile fitting method that takes into account the whole diffraction pattern. Figure 3 shows a graphical representation of a Rietveld refinement of an industrial Portland cement clinker.

Industrial X-ray Diffraction Analysis of Building Materials

151

Figure 3. Graphical output of a Rietveld refinement of a Portland cement clinker: lower graphics = difference between calculated and observed pattern.

Rietveld analysis can be performed automatically with robotic software modules. Specifically designed for unattended, industrial Rietveld analysis, this application also contains the option to create tailor-made reporting, putting Rietveld analysis into the reach of the nonspecialist while providing extended possibilities for more advanced investigations. The traditional Rietveld method uses a refinement strategy, in which the full diffraction profile of high quality is fitted according to a model with known crystal structures. Within the constraints of the application, the traditional Rietveld method is the only method for analysis of complex mixtures of several crystalline phases that show peak overlaps. That yields precise, relatively correct results—enough to verify any changes and to maintain the status quo. It is recommended to compare the precise Rietveld quantitative data to other results obtained by established techniques like optical microscopy. This referenced Rietveld method enables an absolutely correct XRD quantification, accurate enough to push the limits of process control. For example a high quality Portland cement clinker allows you to use more additions in place of valuable clinker. It also allows for a tighter control of the different sulfate phases in OPC (ordinary Portland cement), resulting in more constant hydraulic properties, which facilitates the use of the final product in building projects. For more details on using the Rietveld analysis of cements and the optimization of the method for cement analysis the reader is referred to Aranda et al. (2012, this volume). Apart from the described methodologies in analysis of Building materials, it must be mentioned that it is also possible to quantify amorphous components present in building materials. This is often achieved by the addition of a standard as described by Walenta and Füllmann (2004) in their papers relating to slag and fly ash, which are used as alternative raw materials for cement and concrete production, as well as hydrated cement analysis.

Meier, Anderson, Verryn

152 Full pattern cluster analysis

Modern X-ray diffraction equipment as described above allows the rapid collection of hundreds of scans in a short time. This can be useful in application areas such as process and quality control in the building materials industries. For such large amounts of data, it is very time consuming and often not practical to analyze every individual measurement. This implies that a data reduction tool is required. Cluster analysis starts with a set of scans that are compared and sorted into classes based on their similarity. The similarity comparison is based on the peak and intensity information, but in this case the statistical contribution counts as well. The theory of this method is presented in various literature citations (Lance and Williams 1966; Mardia et al. 1979; Rousseew 1987; Kelley et al. 1996; Lohninger 1999). In summary, cluster analysis is a method that uses statistical methods to simplify the analysis of large amounts of data by: •

Automatically sorting all scans of one or more experiments into classes of closely related scans

•

Identifying the most representative scan of each class

•

Identifying the two most different scans of each class

•

Identifying outliers not fitting into any class (non-members).

This drastically reduces the amount of data that has to be processed, because only representative scans, outliers and sometimes the most different scans are analyzed in more detail. Further cluster analysis can be used to discover hidden features/structures in the data as it is sensitive to small, otherwise unnoticeable, changes in the obtained information. Comparisons of the full peak and profile of every powder diffraction pattern in a set of n patterns with every other pattern can be presented as a correlation matrix. The correlation matrix is used as input to a hierarchical agglomerative cluster analysis, which puts the patterns into classes defined by their similarity. This method starts with each data-set representing a distinct cluster. At each step of the analysis, two clusters with the highest degree of similarity are merged into a single cluster. The process stops with the final step, when only one cluster containing all data-sets remains. The result of this analysis step is usually displayed as a dendrogram (Fig. 4). A well-known and in principle unsolved problem is to find the “right” number of clusters (Kelley et al. 1996). This means cutting the dendrogram at a given dissimilarity and retaining a meaningful set of clusters, where the scans inside a cluster are closely related while the different clusters are different enough to keep them apart. Principal Components Analysis (PCA) can be carried out as an independent method to visualize the quality of the clustering. The correlation matrix is used as input. The method can handle enormous amounts of data and additionally can also deliver process relevant information in the form of a “yes” or “no” result. The cluster approach uses all relevant information in a measurement and included in the analysis is subtle information which is covered in the noise. The method is not based on crystallographic principals and therefore the analysis is not directly related to the Bragg peak information. It is a non-biased method and no mistake can be made based on wrong assumptions. While the cluster analysis can be performed automatically by modern software, some parameters can be changed by the user to optimize the result.

Computed tomography X-ray CT (computed tomography) is a non-destructive technique for visualizing features in the interior of solid objects, and for obtaining digital information on their 3D geometries and properties. Although this is a technique more suitable for a central research laboratory, it gives valuable information and is therefore included here (Desrues et al. 2006).

Industrial X-ray Diffraction Analysis of Building Materials

153

Figure 4. The dendrogram is a graphical display of the result of an agglomerative hierarchical cluster analysis (actual cut-off indicated by a stippled line).

The fundamental principle behind computed tomography is to acquire multiple views of an object over a range of angular orientations. In this way, additional dimensional data are obtained in comparison to conventional X-radiography, in which there is only one view. In our experiments we use the so-called volume CT method, where a cone beam or highly-collimated, thick, parallel beam is used in combination with a 2D (area) detector. The radiation transmitted through the object at each angle is measured and the detector data is stored as 2D X-ray images. The series of 2D X-ray projections, used to generate 3D images, is a collection of images acquired while progressively rotating the sample step-by-step through a full 360-degree rotation within the field of view at increments of less than 1 degree per step. These projections, effectively X-ray attenuation data, represent a measure of the reduction in X-ray intensity that result from absorption and scattering by the sample and contain information on the position and density of absorbing object features within the sample. The accumulated 2D projections data is then used for the numerical reconstruction of the volumetric data (volume rendering). This volume data is compiled as a visualization of the reconstructed layers in a 3D view by CT reconstruction software, which provides these 3D volume results using a Filtered Back-Projection algorithm, the co-called “Feldkamp” algorithm (Feldkamp et al. 1989). The 3D CT data are rendered as voxels (volume element) with threedimensional resolution depending on the X-ray detector pixel size. In general, any sample that fits entirely within the field of view and is completely penetrated through all directions perpendicular to the rotational axis can be imaged in this way. The energy spectrum of the X-ray source defines the penetrative ability of the X-rays, as well as their expected relative attenuation as they pass through materials of different density. Higher-energy X-rays penetrate more effectively than lower-energy ones, but are less sensitive to changes in material density and composition. The X-ray intensity directly affects the signalto-noise ratio and thus image clarity. More detailed description can be found in Bowen and Tanner (1998) and Stock (2008).

154

Meier, Anderson, Verryn

The usefulness of this method in building materials investigations has been described in various papers (Masad et al. 2002; Gopalakrishnan et al. 2007; Schmidt et al. 2010).

APPLICATIONS The use of X-ray techniques to characterize the materials involved in the production process of cementitious materials results in most of the cases in benefits for the process and/or the product. The X-ray diffraction methodologies shown here range from phase identification to full pattern phase quantification or even statistical interpretation of the collected data and can be applied during the different process steps. The degree of utilization of the obtained results varies in terms of the process details, the available equipment and the process management. The examples mentioned below should provide a rough idea of the working principle and corresponding achievable benefits. The sequence of the applications is presented to follow the standard workflow of the cement manufacturing process (Fig. 5).

Figure 5. The three main steps during the cement manufacturing process are the preparation of the raw mixture, the production of clinker (thermal process) and the grinding/blending of the cement.

Raw materials/quarry The overall goal of the raw mixture preparation is to achieve a homogeneous mixture of limestone, clay and other needed compounds, such as bauxite, iron ore and sand, to produce a proper kiln feed with the correct chemical composition. The corresponding main application is the determination of the chemical composition, where X-ray fluorescence in the laboratory and neutron activation based methods on the belt are mandatory on a routine bases to run the process. In order to make cement it is essential to know the raw materials. An oxide analysis of the raw materials is the first step, which provides input for a wide variety of ratios and moduli that relate oxide compositions to one another. These include: LSF (lime saturation factor), SM (silica modulus or ratio), AR (alumina-to-iron ratio), and other lesser-used formulas like the hydraulic modulus. The silica ratio represents the burnability of a raw mix. The burnability impacts how much energy is put into the system. As the ratio of silica to alumina plus iron increas-

Industrial X-ray Diffraction Analysis of Building Materials

155

es, it becomes harder to “burn” and more difficult to combine the raw materials into the phases needed. As the ratio decreases, the tendency for fluxing (the ability of the solid materials to become liquid) increases, and the combining reactions become easier. Another consideration is that silica present as quartz is generally more difficult to combine than silica present as silicates. The alumina-to-iron ratio is important because it controls the potential C3A/C4AF ratio in the finished cement, which is important because of sulfate resistance, heat generation, and admixture compatibility issues. The lime saturation factor controls the potential C3S to C2S ratio in the finished cement. C3S governs the early age strength development while C2S hardens slowly and contributes largely to strength increases at ages beyond 7 days (Moore 1982; Taylor 1990). In addition, X-ray diffraction can contribute in many cases to improve the overall performance. Especially important is the detection of crystalline phases which contribute in a negative way to the process. The detection of coarse quartz in the limestone or high amounts of chlorine-containing clay varieties can be applied during the exploration phase or at different steps of the raw mixture preparation. The location of the sampling as well as the timing has to be adapted to the process to maximize the resulting benefits. Quartz in the raw mix can influence the clinkerization process in a negative way. Coarse quartz grains need more energy (time and temperature) to fully react with the other involved minerals of the raw mix to form C3S and C2S. Furthermore, it is most likely that coarse quartz pieces will increase the wear of the raw mill and therefore reduce the mill life. The presence of quartz can be monitored by automatic Rietveld analysis (Fig. 6). Certain elements, like chlorine, will cause the formation and precipitation of ‘unwanted’ minerals during the calcinations and/or sintering process. These “sticky” minerals, for example paraspurrite, can cause clogging in the pre-heater/calciner or even in the cement kiln. A continuous monitoring of the incoming materials or the corresponding XRF analysis during the exploration phase can prevent the clogging events by a proper raw material selection and/ or additional blending steps. Using XRD (and XRF) quality control of gypsum is important as natural or artificial gypsum are often mixtures of gypsum, hemihydrate and anhydrite and other phases, which do have different reaction behavior, but can be easily distinguished by XRD. Purity of limestone

Figure 6. Rietveld refinement of a raw mix.

156

Meier, Anderson, Verryn

and other raw materials such as blast furnace slag and fly ash can be assessed. Early detection of detrimental phases, such as pyrite, is possible.

Preheater/calciner The introduction of pre-heaters around 1930 and pre-calciners in 1970 (Peray 1998) increased the efficiency of the clinker sintering significantly. However, next to the intended advantages, any additional step in the production process carries the risk of possible problems. Hence, one of the most critical process steps during Portland Cement Clinker production is the operation of the preheater system. The major risk in case of the preheater/calciner is the blockage by clogging as described before, which increases the risk of blockages and downtimes. The origin of this potential problem is the formation of detrimental phases, which influence the flow of the hotmeal and lead to the build-up of cloggings. A complete phase analysis of the material before and after the preheater/calciner can provide data to recognize the imminence of problems. The detrimental phase formations can be monitored directly by X-ray diffraction using Rietveld phase quantification in real time. This not only helps to completely understand the process, but also allows proactive strategies, when clogging precursor phases occur, for reducing and solving blockage problems. The monitoring of the material flow in the preheater/calciner is also gaining more and more importance due to the fact that alternative fuels bring critical elements, such as chlorine, into the process The range of fuels is extremely wide. Traditional kiln fuels are gas, oil or coal. Materials like waste oils, plastics, auto shredded residues, waste tyres and sewage sludge are often proposed as alternative fuels for the cement industry. Also all kinds of slaughterhouse residues are offered as fuel nowadays (Kääntee et al. 2000). In many cases, spot checks of the fuels alone are not sufficient, due to the inconsistency of the alternative fuel composition, which shows huge variations and low sampling frequency is by far not sufficient to handle the possible variations. The regular analysis using automated Rietveld procedures of the involved material streams delivers the key information for a smooth process. The tracing of the efficiency of the heating/calcinations process presents also valuable information for process improvement. The commonly used parameter is the degree of calcinations, which can be directly determined by measuring the ratio of free lime vs. carbonate+free lime (Fig. 7).

Clinker/kiln The chemical reactions that occur in the kiln are described in detail by Hewlett (1998). The temperature is increased when going from the meal feed to the rotary kiln. The most important oxides that participate in the reactions are CaCO3, SiO2, Al2O3 and Fe2O3. Up to about 700 °C water is removed from the meal. In the preheating section (700-900 °C ) calcination as well as an initial combination of alumina, ferric oxide and silica with lime takes place. Between 900 °C and 1200 °C belite, C2S (= 2CaO·SiO2), forms. Above 1250 °C a liquid phase appears and this promotes the reaction between belite and free lime to form alite, C3S (= 3CaO·SiO2). During the cooling stage the molten phase forms C3A, tri calcium aluminate, (= 3CaO·Al2O3) and if the cooling is slow alite may dissolve back into the liquid phase and appear as secondary belite. Calculations of clinkering reactions are described by Barry and Glasser (2000). Usually the production of clinker is done so that one type of clinker allows the plant to manufacture several well-defined types of cement that comply with the physical demands as specified by cement standards. Figure 8 shows the schematic view of clinker formation reactions. Already for a long time the clinker material is tested on its free lime content. The usual target values for the free lime concentration are around 1%. Significant lower values for the free lime concentration indicate a too high temperature during the sintering process. On the one hand this causes wastage of fuel and on the other hand it can influence the clinker properties, especially the grindability, in a negative way. Higher contents of free lime appear if the temperature during the sinter process has been too low and/or the raw meal has been too coarse

Industrial X-ray Diffraction Analysis of Building Materials

157

   

Raw Meal Proportions 

Figure 7. Rietveld refinement of hot meal.

CaCO3  Free Lime  Alite 

Belite  Low Quartz 

High Quartz

Liquid 

Clay Minerals 

C12A7

Fe2O3  H 2O

C2(A,F)

C3 A

Liquid 

C4AF 

0             200            400            600          800            1000           1200           1400                                                             Temperature (°C) 

Figure 8. Schematic view of clinker reactions.

and/or the reaction time has been too short. Free lime concentrations above 2% can generate problems in durability of the concrete. The free lime will react with water and form the phase portlandite. The portlandite phase requires more space than the free lime and causes, if present in higher amounts, cracks in the concrete. High amounts of sodium and potassium lead to the formation of orthorhombic instead of cubic C3A and both modification show differences in reactivity and hydration. The ratio C3A-

Meier, Anderson, Verryn

158

cubic/C3A-orthorhombic influences the water consumption of the cement, as there is higher water consumption by orthorhombic C3A. In the case that Na and K are fixed in alkali-sulfates, these elements are not available for the incorporation in C3A and by changing the degree of sulfatization the ratio of C3A-cubic/C3A-orthorhombic can be controlled. The increasing use of alternative fuels introduces more variations in the clinker composition and in the burning conditions. The subsequent potential implications on the product properties require a full quantitative phase analysis of the clinker to be able to produce a constant product with the desired properties. Table 1 gives an overview of the main phases—property relations in clinker and cement. A more complete list of phases can be obtained from Aranda et al. (2012, this volume). Bahtty (1995) describes the role of minor elements in cement manufacture in more detail. Pöllmann (2002) describes the composition of cement phases as well as influences of different phases and elements. Another related opportunity to improve the consistency of the product and to influence the cement manufacturing process in a positive way is the monitoring of the bypass dust composition. The composition of bypass dust is based on the elements, which cannot be incorporated in the clinker minerals. The elemental as well as the phase composition of the bypass dust provides important information of the process. The statistical data analysis by using cluster analysis, as described above, can be applied as well to monitor effects of the raw feed changes and fuel composition on the process. The subsequent interaction on the process is based on the obtained cross correlations of historical data. The described cluster analysis can be also applied for the clinker to get a visually easy readable presentation of the data. Often the principal component analysis plot is used to display the results of the cluster analysis. Differences in the phase composition and therefore also in the material properties are seen as distances of the displayed positions (Fig. 9).

Cement The full crystalline phase quantification of the cement can contribute to the prediction of the product properties. The usual strength tests take a relatively long time before the results become available. By that time, the corresponding material is often already used and is it very difficult and expensive to correct possible discrepancies afterwards. The results of the X-ray analysis are available much faster, allowing corrective actions to be taken much earlier in the process. Table 1. Phase property relations in clinker and cement. Cement Phase

Trivial Name

Relevance

CaO Ca(OH)2 C 3S C2S C3A C4AF MgO CaSO4·2H2O CaSO4·0.5H2O CaSO4 CaCO3

Free Lime Calcium hydroxide/ Portlandite Alite Belite Aluminate Ferrite Periclase Gypsum Hemihydrate Anhydrite Calcite

Kiln Temperature Control Kiln Temperature - Cooling Conditions early strength, hardness later strength, hardness Setting Time Color expansion hydration control hydration control hydration control

SiO2

Quartz Glass

insufficient grinding temperature

Industrial X-ray Diffraction Analysis of Building Materials

159

Figure 9. Principal Component Analysis Plot of the Cluster analysis of clinker samples and indicating   different properties.

 

ASTM C 150, Standard Specification for Portland Cement, recognize eight basic types of Portland cement concrete (see Table 2). There are also many other types of blended and proprietary cements that are not mentioned here. The permissible additions of slag, fly ash and other extenders are shown in Bye (1999). Examples of Rietveld quantifications of different cement types are shown in Figures 10-12. The quantitative phase analysis also provides information about various process steps. For example are there phase changes between the different sulfate phases (gypsum, hemihydrates and anhydrite) which would be an indicator for the milling conditions and can be used in many cases to increase the mill efficiency. At the same time, the quantitative ratio of the sulfate phases together with the abundance of the calciumaluminate polymorphs (cubic and orthorhombic C3A) is important for the workability/setting time of the concrete. Another important benefit of the X-ray analysis is the quantitative compound determination of cement additions (Bye 1999). Nowadays, the Rietveld quantification includes not only the crystalline phases, but the Rietveld analysis can also provide information about non-crystalline compounds of the cement, like blast furnace slags, fly ashes and other pozzolanic materials (Walenta and Füllmann 2004; Westphal et al. 2009, 2010). See also Figures 11 and 12. The statistical cluster data analysis gives a user-friendly presentation of process trends and Figure 13 shows the principle component analysis plot of some different cement types as listed in Table 2 and again the material properties are seen as distances of the displayed position.

Hydrated cement X-ray diffraction analysis has also been widely used to determine the hydration of hardened cement paste (Scrivener et al. 2004; Hoshino et al. 2006; Hesse et al. 2009). In earlier days, the results were mainly qualitative, but using modern equipment together with Rietveld analysis

Meier, Anderson, Verryn

160

Table 2. ASTM types of Portland cement. Type I

Name

Purpose

Normal

General-purpose cement suitable for most purposes.

IA

Normal - Air-Entraining

An air-entraining modification of Type I.

II

Moderate Sulfate Resistance

Used as a precaution against moderate sulfate attack. It will usually generate less heat at a slower rate than Type I cement.

IIA

Moderate Sulfate Resistance Air-Entraining

An air-entraining modification of Type II.

III

High Early Strength

Used when high early strength is needed. It is has more C3S than Type I cement and has been ground finer to provide a higher surface-to-volume ratio, both of which speed hydration. Strength gain is double that of Type I cement in the first 24 hours.

IIIA

High Early Strength Air-Entraining

An air-entraining modification of Type III.

IV

Low Heat of Hydration

Used when hydration heat must be minimized in large volume applications such as gravity dams. Contains about half the C3S and C3A and double the C2S of Type I cement.

V

High Sulfate Resistance

Used as a precaution against severe sulfate action principally where soils or groundwaters have high sulfate content. It gains strength at a slower rate than Type I cement. High sulfate resistance is attributable to low C3A content.

Figure 10. Rietveld refinement of CEM I.

Industrial X-ray Diffraction Analysis of Building Materials

161

Figure 11. Rietveld refinement of CEMM II AM (10 wt% fly ash, 8 wt% slag).

Figure 12. Rietveld refinement of CEM II B (65 wt% slag).

an accurate analysis of hydration products can be achieved. The methods are, again, described in more detail by Aranda et al. (2012, this volume). With the reduced measurement times as described above, the quantification can be done “online” and the hydration process does not have to be stopped as the sample can be analyzed repeatedly at short intervals. Cluster analysis as described above can then be employed to evaluate the large data sets produced (Fig. 14). After identification of phases present, subsequent data sets can be evaluated using automated Rietveld analysis and scans can be visualized using 3D functions supplied by most modern XRD software (Fig. 15).

162

Meier, Anderson, Verryn

 

Figure 13. Principal Component Analysis Plot of the Cluster analysis of different cement type samples black: CEM I 32.5, dark grey: CEM I 42.5, light grey: other).

Figure 14. Principal Component Analysis Plot of the Cluster analysis of a cement hydration study.

Industrial X-ray Diffraction Analysis of Building Materials

163

Figure 15. 3D plot of a cement hydration study.

Concrete Apart from the phase composition, more features are relevant for the properties of the final cement and/or concrete product. Porosity and permeability are especially important for the later strength and durability of the concrete. The determination of the porosity and the permeability by using classical methods, like macroscopic (Fig. 16) and microscopy investigations descriptions (Fig. 17) or mercury intrusion test, are very time and labor intensive next to the difficulties to handle hazardous chemicals. Computed tomography (CT) based analysis offers an attractive alternative to these classical methods because of the user friendliness and the shorter analysis time. Furthermore CT delivers

Figure 16. Macroscopic slice of a norm bar for concrete testing (40 mm × 40 mm × 2 mm).

164

Meier, Anderson, Verryn

Figure 17. Photomicrograph of a concrete samples showing pores, sand grains and hydrated cement.

information about the fabric of the concrete and the understanding of the related properties than the classical methods. In this example, a few CT images are shown from a study, which was made of a glassreinforced cement concrete sample (Pöllmann et al. 2010, 2011). This type of concrete material is characterized by a distinctly more “rigid” structure than standard concrete, which consists of cement and sand (mainly quartz) in a roughly 1:3 ratio. CT imaging with the PIXcel 3D detector reveals a very distinctive porosity and cementation of the constituents. In Figure 18, the quartz grains (dark grey) are embedded in the cement clinker matrix (medium grey/yellow in the highlighted center of the picture) with a distinct narrow reaction rim (lighter grey). The densely intergrown matrix of this material underlines its rigidity. Even more important is the porosity and the pore shape. The porosity is known to be gas-bubble related and shows the typical shape of non-interconnected, spherical gas inclusion remnants. This explains the rigidity of the material, as well as the low permeability of his material, compared to standard concretes. This is highlighted in Figure 19 in an inverted 3D view of the concrete sample showing the pore distribution and the pore shape and size. The pore volume and size distribution can be quantified. This figure is another example of the magnificent displaying power of the X-ray CT technique, when using a very high resolution pixel detector.

SAMPLE PREPARATION In most of the above, XRD and XRF methods were mentioned. In the process control environment samples can be prepared either individually for each technique or the same sample can be analyzed in sequence. A good sample preparation guide is Buhrke et al. (1998). The Rietveld method does take preferred orientation, possibly introduced by the samples preparation method into account during refinements. Various systems for automated sample preparation are available. A manual back-loading sample preparation in many cases results in lesser preferred orientation than the automated front-pressing technique. Nevertheless an automated sample preparation is preferable, as the errors are more reproducible and user independent and therefore can be corrected during the data analysis.

Industrial X-ray Diffraction Analysis of Building Materials

165

Figure 18. CT slice of glass-reinforced concrete, highlighting the mode of intergrowth between the sand grains (quartz, dark grey) and the cement matrix (medium grey/yellow).

Figure 19. 3D CT image in an “inverted” view, highlighting the pores and their spatial distribution in the concrete sample.

166

Meier, Anderson, Verryn CONCLUSION

From the above it is clear that the analyses as described above are a valuable tool in the process and product control environment. With automated XRD and Rietveld analysis (and XRF analysis), the various systems for automation of sample preparation, processing and saving of data, allow these methods to be time and cost saving. Microscopic methods can give many more details but do need appropriate personnel and are more time consuming.

REFERENCES Aldridge VW (1982) Accuracy and precision of phase analysis in Portland Cement by Bogue, microscopy and X-Ray diffraction methods. Cem Concr Res 12:381-398 Aranda MAG, De la Torre AG, León-Reina L (2012) Rietveld quantitative phase analysis of OPC clinkers, cements and hydration products. Rev Mineral Geochem 74:169-209 ASTM Standard C150. ASTM C150 / C150M - 11 Standard Specification for Portland Cement. ASTM International, West Conshohocken, PA. http://www.astm.org/Standards/C150.htm, DOI: 10.1520/C0150_ C0150M-11 Barry TI, Glasser FP (2000) Calculations of Portland cement clinkering. Adv Cem Res 12(1):19-28 Bish DL, Howard SA (1988) Quantitative phase analysis using the Rietveld method. J Appl Crystallogr 21:86-91 Bish DL, Post JE (eds) (1989) Modern Powder Diffraction. Reviews in Mineralogy. Volume 20. Mineralogical Society of America, Washington Bish DL, Post JE (1993) Quantitative mineralogical analysis using the Rietveld full-pattern fitting method. Am Mineral 78:932-940 Bogue RH (1929) Calculation of the compounds in Portland cement. Ind Eng Chem 1:192-197 Bowen DK, Tanner BK (1998) High Resolution X-Ray Diffractometry and Topography. CRC Press Buhrke VE, Jenkins R, Smith DK (1998) A Practical Guide for the Preparation of Specimens for X-Ray Fluorescence and X-Ray Diffraction Analysis. Wiley-VCH, New York Bye GC (1999) Portland Cement. 2nd ed. Thomas Telford Publishing, London Chung FH (1974a) Quantitative interpretation of x-ray diffraction patterns of mixtures. I. Matrix flushing method for quantitative multicomponent analysis. J Appl Cryst 7:519-525 Chung FH (1974b) Quantitative interpretation of x-ray diffraction patterns of mixtures. II. Adiabatic principle of x-ray diffraction analysis of mixtures. J Appl Cryst 7:526-553 Chung FH (1975) Quantitative interpretation of x-ray diffraction patterns of mixtures. III. Simultaneous determination of a set of reference intensities. J Appl Cryst 8:17-19 Cullity BD (1978) Elements of X-Ray Diffraction. Addison-Wesley, Reading, Massachusetts Davis BL, Kath R, Spilde, M (1990) The reference intensity ratio: its measurement and significance. Powder Diffr 5(2):76-78 Desrues J, Viggiani G, Besuelle P (2006) Advances in X-ray Tomography for Geomaterials. ISTE Ltd Feldkamp LA, Goldstein SA, Parfitt AM, Jesion G, Kleerekoper M (1989) The direct examination of threedimensional bone architectures in vitro by computed tomography. J Bone Mineral Res 4:3-11 Gopalakrishnan, K, Ceylan H, Inanc F (2007) Using X-ray computed tomography to study paving materials. Proceedings of the ICE - Construction Materials 160(1):15-23 Hill RJ, Howard CJ (1987) Quantitative phase analysis from neutron powder diffraction data using the Rietveld method. J Appl Cryst 20:467-474 Hesse C, Goetz-Neunhoeffer F, Neubauer J, Braeu M, Gaeberlein P (2009) Quantitative in situ X-ray diffraction analysis of early hydration of Portland cement at defined temperatures. Powder Diffr 24:112-115 Hewlett PC (1998) Lea’s Chemistry of Cement and Concrete. 4th ed. Arnold Publishers. London Hoshino S, Yamada K, Hirao H (2006) XRD/Rietveld analysis of the hydration and strength development of slag and limestone blended cement. J Adv Concr Tech 4:357-367 Jenkins R, Snyder RL (1996) Introduction to X-Ray Powder Diffractometry. Wiley-Interscience, New York Kelley LA, Gardner SP, Sutcliffe MJ (1996), An automated approach for clustering an ensemble of NMRderived protein structures into conformationally-related subfamilies. Protein Eng 9:1063-1065 Kääntee U, Zevenhoven R, Backman R, Hupa M (2000) The impact of alternative fuels on the cement manufacturing process. Proceedings of R’2000 Recovery-Recycling-Reintegration, Toronto, Canada, June 1070-1075 (CD-ROM) Klug HP, Alexander LE (1974) X-Ray Diffraction Procedures. Wiley-Interscience, New York Lance GN, Williams WT (1966) A general theory of classification sorting strategies 1. Hierarchical system. Comp J 9:373-380 Lohninger H (1999) Teach/Me Data Analysis. Springer-Verlag, Berlin-New York-Tokyo

Industrial X-ray Diffraction Analysis of Building Materials

167

Madsen IC, Scarlett NVI (2000) Cement: Quantitative Phase Analysis of Portland Cement Clinker. In: Industrial Applications of X-Ray Diffraction. Chung FH, Smith DK (eds) Taylor & Francis, New York, p 415-440 Madsen IC, Scarlett NVI (2008) Quantitative Phase Analysis. In: Powder Diffraction Theory and Practice. Dinnebier RE, Billinge SJL (eds) Royal Society of Chemistry Publishing, Cambridge, p 298-331 Mardia KV, Kent JT, Bibby JM (1979) Multivariate analysis. Academic Press, London Masad E, Jandhyala, VK, Dasgupta N, Somadevan N, Shashidhar N (2002) Characterization of air void distribution in asphalt mixes using x-ray computed tomography. J Mater Civil Eng 14:122-130 Moore CW (1982) Chemical control of Portland cement clinker. Am Cer Soc Bull 61(4):511-515 Nuffield EW (1966) X-ray Diffraction Methods. Wiley, New York Peray KE (1998) The Rotary Cement Kiln, 2nd ed. Chemical Publishing Co, Inc Pöllmann H (2002) Composition of cement phases. In: Structure and Performance of Cements. Bendted J, Barnes P (eds) Routledge Chapman & Hall, p 25-56 Pöllmann H, Meier R, Riedl U, Blaj G (2010) A multidimensional investigation using X-ray diffraction and Computer tomography. Adv X-ray Anal 54(13):101-107 Pöllmann H, Meier R, Riedl U, Blaj G (2011) Multi-dimensionale Röntgenuntersuchung an Zementhydratprodukten – Von der klassischen Bragg-Brentano Röntgenphasenanalyse zur 3-dimensionalen CT-Mikrostruktur und Porenanalyse. GDCH Bauchemie, Hamburg p 101-106 Rietveld HM (1966) A method for including the line profiles of neutron powder diffraction peaks in the determination of crystal structures. Acta Crystallogr 21:A228 Rietveld HM (1969) A profile refinement method for nuclear and magnetic structures. J Appl Cryst 2:65-71 Rousseeuw JP (1987) Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. J Comp Appl Math 20:53-65 Schmidt M, Pöllmann H, Egersdoerfer A, Goeske J, Winter S (2010) Investigations on the puzzolanic reactivity of a special glass meal in a cementitious system. Proc of the 32nd Int Conf on Cement Microscopy, New Orleans, Louisiana, p 1-33 Scrivener KL, Fullmann T, Gallucci E, Walenta G, Bermejo E (2004) Quantitative study of Portland cement hydration by X-ray diffraction/Rietveld analysis and independent methods. Cem Concr Res 34:1541-1547 Smith DK, Johnson Jr GG, Scheible A, Wims AM, Johnson JL, Ullmann G (1987) Quantitative x-ray diffraction method using the full diffraction pattern. Powder Diffr 2(2):73-77 Snyder RL, Bish DL (1989) Quantitative analysis. Rev Mineral 20:101-144 Stock R (2008) MicroComputed Tomography: Methodology and Applications. CRC Press Taylor HFW (1989) Modification of the Bogue calculation. Adv Cem Res 2:73-77 Taylor HFW (1990) Cement Chemistry. Academic Press, London Walenta G, Füllmann T (2004) Advances in quantitative XRD analysis for clinker, cements, and cementitious additions. Adv X-ray Anal 47:287-296 Westphal T, Füllmann T, Pöllmann H (2009) Rietveld quantification of amorphous portions with an internal standard − mathematical consequences of the experimental approach. Powder Diffr 24:239-243 Westphal T, Füllmann T, Pöllmann H (2010) Quantifizierung amorpher Anteile mittels Rietveldmethode. GDCH Tagung Bauchemie, Dortmund, Oktober 2010 Young RA (ed) (1993) The Rietveld Method. IUCr Monographs on Crystallography, 5, Oxford Science Publications Zevin LS, Kimmel G (1995) Quantitative X-Ray Diffractometry. Springer-Verlag, New York

5

Reviews in Mineralogy & Geochemistry Vol. 74 pp. 169-209, 2012 Copyright © Mineralogical Society of America

Rietveld Quantitative Phase Analysis of OPC Clinkers, Cements and Hydration Products Miguel A. G. Aranda,* Ángeles G. De la Torre Departamento de Química Inorgánica, Cristalografía y Mineralogía Universidad de Málaga 29071 Málaga, Spain email: [email protected]

Laura León-Reina Servicios Centrales de Investigación Universidad de Málaga 29071 Málaga, Spain BRIEF INTRODUCTION It has been more than twenty years since the excellent volume of Reviews in Mineralogy dedicated to Modern Powder Diffraction was published (Bish and Post 1989). That volume contained a series of key articles ranging from the basic of powder diffraction to sample preparation and synchrotron and neutron powder diffraction. Within that volume, quantitative phase analysis was extensively discussed in a specific chapter (Snyder and Bish 1989). Snyder and Bish (1989) discussed the Reference Intensity Ratio approach (also known as Chung method), the method of standard additions (also known as spiking method) and the full patternfitting approach using both the Rietveld method and the observed patterns method. The reader is referred to that volume for the basics of powder diffraction, and to the specific chapter by Snyder and Bish (1989) for the history of quantitative phase analysis from powder diffraction and for a discussion or the early findings. Quantitative phase analysis by X-ray powder diffraction dates back to 1925 (Navias 1925). In this work, the amount of mullite obtained by firing selected clays (and a feldspar) was determined by the direct comparison of the intensities of two diffraction lines of the fired samples with those of pure mullite. The patterns were recorded on photographic negatives after an X-ray exposure of 165 hours (almost a week). Quantitative phase analysis (QPA) from diffraction data can be obtained from a number of methods explained in classical books (Klug and Alexander 1974; Cullity 1978; Snyder and Bish 1989; Zevin and Kimmel 1995; Jenkins and Snyder 1996). However, it is now safe to say that QPA from powder diffraction data is nowadays mainly based on the Rietveld methodology (Rietveld 1969; Hill and Howard 1987; Bish and Howard 1988; Bish and Post 1993; Madsen and Scarlett 2008). Hence, the term Rietveld quantitative phase analysis (RQPA) has been coined. The advantages of using a full pattern-fitting over single peak(s) approaches are discussed in those papers and they will not be further addressed here. Hence, we will focus on RQPA. The readers should also be aware of two excellent articles which were the outcome of a round robin study on quantitative phase analysis by powder diffraction mainly using the Rietveld method (Madsen et al. 2001; Scarlett et al. 2002). These interlaboratory comparison studies made some indispensable recommendations for carrying out an accurate RQPA and included results concerning the influence of sample-related effects such as preferred orientation and microabsorption. However, it is also worth mentioning, that the key book entitled “The Rietveld Method” (Young 1993) does not contain a chapter on quantitative phase 1529-6466/12/0074-0005$05.00

DOI: 10.2138/rmg.2012.74.5

170

Aranda, De la Torre, León-Reina

analysis. Furthermore, the key paper “Rietveld refinement guidelines” (McCusker et al. 1999), supported by the International Union of Crystallography, did not dedicate a section to RQPA. On the other hand, building materials, such as ordinary Portland cements (OPCs), are very complex samples of worldwide importance, hence quantitative knowledge of their mineralogical composition is necessary to predict performances (Taylor 1997). Ultimately, the assemblage of crystalline phases, and not the bulk chemistry, determines cement features. In fact, the hydraulic properties of a mortar/concrete mainly depend on the cement mineralogical composition and its texture (Bentz 2008; Skibsted and Hall 2008; Scrivener and Nonat 2011). The most widely used method of estimating the potential phase composition of Portland cement in cement plants is the Bogue calculation from its elemental analysis usually determined by X-Ray Fluorescence, XRF (Bogue 1929; Taylor 1989). However, it is well known that the phase abundance calculated by this indirect approach may be quite far from the true values. This is due mainly to three reasons: i) the four main clinker phases are solid solutions with compositions significantly different from the stoichiometric pure-phases; ii) there is no certainty of attaining equilibrium conditions (both in the kiln and for sure in the cooling process, Hong et al. 2001); iii) the presence of minor phases. As an alternative approach, RQPA allows a direct measurement of the phase content of cements. So, on-line systems for RQPA of clinkers and cements are becoming widespread. However, we will not deal here with on-line RQPA at cement plants. The reader is addressed to specific publications dealing with this subject including reproducible sample preparation, fast data acquisition and fast-and-robust data analysis (Moller 1998; Manias et al. 2001; Scarlett et al. 2001; Fullmann and Walenta 2003; Paul et al. 2004; Enders and Berger 2007; De la Torre et al. 2011). Therefore, this review focuses on the use of RQPA for understanding OPC clinkers, cements and pastes, at central laboratories. It must be highlighted that a very recent review article has been dedicated to the application of the Rietveld method to the analysis of anhydrous cements (Le Saout et al. 2011). Furthermore, a second broad article dealing with direct determination of phases in OPCs by quantitative X-ray powder diffraction has also been reported (Stutzman 2011). Hereafter, cement nomenclature will be used, i.e., C = CaO, S = SiO2, A = Al2O3, F = Fe2O3, M = MgO, S = SO3, C = CO2, H = H2O, K = K2O and N = Na2O. Hence, typical cement phase such as alite (Ca3SiO5), belite (Ca2SiO4), tricalcium aluminate (Ca3Al2O6), calcium aluminoferrate (Ca4Al2Fe2O10), gypsum (Ca2SO4·2H2O) or calcite (CaCO3) are abbreviated as C3S, C2S, C3A, C4AF, C S H2 and C C, respectively. Most cement compounds are not pure stoichiometric phases but they may incorporate many ions as extensively discussed (Taylor 1997; Le Saout et al. 2011). In addition, AFm and AFt set of phases should be defined. AFm stands for the abbreviation for “alumina, ferric oxide, mono-sulfate” or “Al2O3-Fe2O3-mono”, in the same way AFt stands for a similar abbreviation but “tri-sulfate” or “Al2O3-Fe2O3-tri”. The AFm phase refers to a family of hydrated calcium aluminates based on the hydrocalumite structure, Ca4Al2(OH)12·[Cl(OH)]·6H2O. The archetype AFm phase is C3A·CaSO4·12H2O or Ca4Al2(OH)12·[SO4]·6H2O, known as Kuzelite, but Al can be partly replaced by Fe and SO42− can be partly or fully replaced by OH−, CO32− and other anions (Pollmann 2006, Matschei et al. 2007a). The different hydration stages strongly influence the X-ray powder diffraction patterns (Pollmann 2007; Balonis and Glasser, 2009). By far the most common AFt phase is ettringite which has the following stoichiometry C3A·3CaSO4·32H2O that can also be written as Ca6Al2(SO4)3(OH)12·26H2O.

THE RIETVELD METHOD Crystalline compounds have long range periodic order and its interaction with X-rays (or neutrons) yield powder diffraction patterns plenty of peaks/reflections. The position, height and even width of these reflections may be used to determine many aspects of the sample structure/

Rietveld Quantitative Phase Analysis of OPC

171

microstructure. The reader is referred to the two most recent books in “Powder Diffraction” to expand on that (Pecharsky and Zavalij 2005; Dinnebier and Billinge 2008). The Rietveld method is a technique devised by Hugo Rietveld in the late sixties (Rietveld 1967, 1969) for a deeper characterization of polycrystalline compounds by treating in a “better/different” way the powder diffraction patterns. The very original contribution of the Rietveld method to powder diffraction was its conceptual breakthrough: “To use measured powder pattern intensities instead of reflection (peak) intensities.” This conceptual breakthrough, together with the coming of computers, allowed to properly dealing with strongly overlapping reflections. The introduction of this technique was a significant step forward in the diffraction analysis of crystalline powder samples. The powder diffraction analysis at that moment worked with extracted reflection intensities which was a very serious difficulty in the case of overlapping reflections. The Rietveld (whole-profile) method uses a least squares approach to optimize a theoretical line profile until it matches in the best possible way the measured sample powder diffraction profile (see Eqn. 1); where Sy is the function to be minimized, wi is the statistical weight, and yi(obs) and yi(cal) are the observed and calculated powder diffraction intensities for the i-point of the powder pattern, respectively. = Sy

∑w

i

yi (obs) − yi (cal)

2

(1)

i

The calculated intensity, yi(cal), for each point of the powder pattern, 2qi, is obtained as the sum of the contribution of all reflections (k) which give intensity to that i-point above the background, yb(2qi), see Equation (2): yi (cal) =

yb (2qi ) + Sa ∑ mk Fk h(2qi − 2qk ) Lp(2qi ) Pk 2

(2)

k

where Sa is the scale factor for the pure crystalline a-phase to be studied, k stands for the reflections which contribute to that point of the pattern, mk is the multiplicity of that reflection, Fk is the structure factor of that reflection, h(2qi−2qk) is the function that distributes the intensity of that reflection in a given 2q range, Lp(2qi) stands for the Lorentz and polarization correction, and Pk stands for additional corrections that may be needed (preferred orientation, extinction, etc.). Equation (2) may be extended to a sample containing m-crystalline phases, see Equation (3), by summing up the contribution of every crystalline phase. m

yi (cal) = yb (2qi ) + ∑ Si ∑ mk Fk h(2qi − 2qk ) Lp(2qi ) Pk i =1

2

(3)

k

The Rietveld method was originally devised for the refinement of crystal and magnetic structures from powder neutron data. However, today the uses of the Rietveld method are numerous and help extracting the maximum information already present in a powder diffraction pattern. This information is listed just below and it must be highlighted that the Rietveld method is making a profound impact in every listed use as it increases the accuracy and precision of the extracted results. It can also be said that Hugo Rietveld did not envisage some of these uses including quantitative phase analysis (Rietveld 2010). From the position of the diffraction peaks: i) lattice parameters, ii) space group determination, iii) qualitative phase analysis (phase identification), and iv) macro-strain; from the intensities of the diffraction peaks: v) crystal (nuclear) structure (including atomic positions, occupation factors and atomic displacement parameters), vi) magnetic structure, vii) texture (preferred orientation), and viii) quantitative phase analysis; from the widths/shapes of the diffraction peaks: ix) micro-strains (mainly in solid solutions), and x) coherent diffraction domain size(s).

172

Aranda, De la Torre, León-Reina

In this review article, we will focus on the viiith-application, quantitative phase analysis. We address the reader to some selected books and papers for a deeper insight of the role of the Rietveld method in the remaining applications (Young 1993; McCusker et al. 1999; David et al. 2002; Shin et al. 2002; Fitzpatrick and Lodini 2003; Balzar et al. 2004; Pecharsky and Zavalij 2005; Kaduk 2007a,b; Dinnebier and Billinge 2008; David and Shankland 2008; Soleimanian and Aghdaee 2008; Markvardsen et al. 2008). Finally, new ways of learning are being explored. In addition to books, papers and training in laboratories with proven experience, other methods are emerging. If the reader is keen of using internet-distributed audio/visual material, then the following references are a good starting point (Toby 2007, 2010).

General issues For a successful Rietveld quantitative phase analysis, several steps have to be fulfilled. Initially, sample has to be properly prepared and this will depend upon the nature of the sample itself and the diffractometer where the data will be taken. This will be briefly discussed in a later section. Secondly, the diffractometer should be well aligned and maintained. Different optical configurations are possible in modern diffractometers and the optimal set-up should be used. Under these two pre-requisites, a good powder diffraction pattern may be taken and RQPA can be carried out. It must be highlighted that the results of any analysis cannot be better than the raw data. Therefore, a lot of care is needed in these two initial steps in order not to spend/waste a lot of time in data analysis/evaluation without conclusive results. Once good powder diffraction data have been taken, a third step follows: qualitative phase analysis. Every crystalline phase in the sample should be identified. This is easy to say but sometimes quite complex to fulfill. The strong peak overlapping in the diffraction patterns does not allow to conclusively determining all phases present in some cases. Then a possible strategy follows, we compute the RQPA with the phases which are clearly present in the pattern and, from the net intensity in the difference curve, the remaining low-content phases are determined. Alternatively, a trial-and-error method can be used. In addition to the main, clearly-observed phases, dubious phases are added to the Rietveld calculation and its absence/ presence can be individually estimated. In any case, when the main (all) phases are identified, the fourth step is to carry out the RQPA with the appropriate software. A full list, and discussion, of the programs available for this task is out of the scope of this paper. However, we can suggest the ccp14 web site for a list of Rietveld programs (http://www.ccp14.ac.uk/solution/rietveld_software/index.html) and mention that GSAS (Larson and von Dreele 1994; Toby 2001) and FULLPROF (RodriguezCarvajal 1993; Roisnel and Rodriguez-Carvajal 2001) are the widest used packages. Many other packages can also be used for RQPA and we underline just a few: BGMN, SIROQUANT, TOPAS and HighScore Plus. In any case, in addition to the raw data, any Rietveld program needs a control file to execute the refinements. In this control file, the crystal structures of the different components must be included. The fit is carried out by optimizing all appropriate variables such as: i) scale factor of every crystalline phase; ii) background parameters for the chosen function; iii) unit cell parameters for every crystalline phase; iv) peak shape parameters for every computed phase; and finally, v) correction parameters which may be phase-dependent (such as preferred orientation, extinction, etc.) or pattern-dependent (zero-shift, absorption correction when working in transmission geometry, etc.). Usually, for RQPA the structural descriptions (atomic positional parameters, atomic displacement parameters and occupation factors) are not optimized but kept as reported in the structural studies. The RQPA method does not require calibration curve or internal standard. However, the crystal structures of all crystalline constituents must be known. This is a prerequisite as the process consists of the comparison between the measured and the calculated patterns (the calculated pattern being computed from the crystal structures).

Rietveld Quantitative Phase Analysis of OPC

173

The application of RQPA to clinkers/cements/pastes is not straightforward for the following reasons: i) there are many phases, usually more than five, which increases the diffraction peak overlapping and so the correlations; ii) each phase has its own mass absorption coefficient which may yield the microabsorption problem (see below); iii) the small mean penetration depth of X-rays (~30 mm for Cu Ka, 1.54 Å) implies that only a thin layer is analyzed in the Bragg-Brentano q/2q geometry which may lead to poor particle statistics; iv) some phases, for instance alite or gypsum crystallize as plaques which show preferred orientation and so increases the errors; v) phases can crystallize as several polymorphs that must be identified a priori; vi) the diffraction peak broadening for some phases may be anisotropic and it must be properly model; and vii) the atomic impurities inside each phase are not known and their scale factors are computed for ideal/stoichiometric phases. In any case, this method has several advantages over other methods based on powder diffraction and other technologies (microscopy, thermal analysis, etc.). A comparison between RQPA results and those obtained with alternative/complementary technologies will be given below. The direct output of the RQPA is a set of scale factors, one for every crystalline phase within the mixture with computed crystal structures. In addition, several other parameters may be of interest: i) unit cell parameters (to have an idea of the existence of solid solutions); ii) preferred orientation of some phases (to have an insight of the particle shapes of these phases); iii) peak shape widths to know about the microstructures. The key transformation of the phase scale factors to phase contents is discussed in a section below. However, it must be highlighted that the best available structural description should be used in order to extract the best possible scale factor value for every phase. This is evaluated by the flatness of the difference curve and also by obtaining low Rietveld-disagreement indices (R-factors). However, the lowest R-factors may also be obtained for wrong analysis. So, it is important to optimize those parameters that allow their minimization without excessive correlations. It is also important to understand the definitions for R-factors which are discussed in standard papers (Young 1993; McCusker et al. 1999). However, a much personal view has been reported (Toby 2006) which was mainly dedicated to the refinement of crystal structures but some concerns and discussions are of overall interest.

Structural description of the phases present in OPC materials The main clinker phases are alite, belite and aluminate in white Portland clinkers and also ferrite in grey Portland clinkers. In addition several other minor phases may be present: lime, periclase, arcanite, aphthitalite and others, see Table 1. An excellent review about the phases present in OPC clinkers has been very recently published (Le Saout et al. 2011). We want to draw the attention that several issues are discussed in that paper including the selection of the alite polymorph, the implications of the Fe/Al ratio in ferrite quantification, and the consequences of the solid solution existence in most clinker phases (small change in the structure factors and densities of the analyzed phases). Portland cements are fabricated by adding a setting regulator to the clinker. The sulfate phases added for this purpose may be gypsum (CaSO4·2H2O), hemihydrate (CaSO4·½H2O) also known as bassanite or anhydrite (CaSO4). Different Portland cements may have different additions, for instance calcite (CaCO3) or quartz (SiO2). Additionally, other highly-amorphous materials may be added for blended cement and concrete production such as blast furnace slag (BFS), pulverized fly ash (PFA) or pozzonale minerals. It must be kept in mind that these additions also contain some crystalline phases (for instance mullite or hematite) that must be quantified in blended cements. Table 2 gives a list of additional phases that may be present in (blended) OPC cements. New phases appear during hydration of OPC cements. Table 3 lists additional (hydrated) phases that may be present in Portland cement pastes. It should be noted that a review of

Ca3SiO5-Mg,Al Ca3SiO5-Mg Ca3SiO5 Ca2SiO4 Ca2SiO4 Ca2SiO4 Ca3Al2O6 Ca8.5NaAl6O18 Ca8.25Na1.5Al6O18 Ca2AlFeO5 CaO MgO K2SO4 K3Na(SO4)2 Na2SO4 Ca2K2(SO4)3 Ca5(SiO4)2(SO4) Ca10(SiO4)3(SO4)3Cl2 Ca10(SiO4)3(SO4)3F2 Ca12Al14O33

Alite

Monoclinic/M3 Triclinic/T3 Triclinic/T1 Monoclinic/b Orthorhombic/a′ Orthorhombic/g Cubic Orthorhombic Monoclinic Orthorhombic Cubic Cubic Orthorhombic Rhombohedral Orthorhombic Orthorhombic Orthorhombic Hexagonal Hexagonal Cubic

Crystal system/ notation 310.4 308.5 304.6 299.9 298.1 268.2 277.1 278.9 246.5 502.8 401.2 99.5 219.9 195.4 73.0 209.1 255.6 272.7 261.8 198.6

m (cm−1) 94742 162744 4331 81096 81097 81095 1841 100220 100221 9197 52783 9863 79777 26018 81506 040989 085123 154205 97203 241243

ICSD codes

01-070-1846 01-086-0398 01-086-0399 01-086-0397 01-070-0839 01-083-1359 01-083-1360 01-071-0667 01-071-4121 01-071-1176 01-083-0681 01-074-0398 00-037-1465 01-074-0404 01-088-0812 00-041-0479 01-072-7301 70-2144

01-070-8632

PDF codes [1] [2] [3] [4] [4] [4] [5] [6] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

Ref#

# References: [1] De la Torre et al. 2002; [2] De la Torre et al. 2008; [3] Golovastikov et al. 1975; [4] Mumme et al. 1995; [5] Mondal and Jeffry 1975; [6] Takeuchi et al. 1980; [7] Colville and Geller 1971; [8] Smith and Leider 1968; [9] Sasaki et al. 1979; [10] Ojima et al. 1995; [11] Okada and Ossaka 1980; [12] Rasmussen et al. 1996; [13] Speer and Salje 1986; [14] Irran et al. 1997; [15] Saint-Jean and Hansen 2005; [16] Pajares et al. 2002; [17] Palacios et al. 2007.

Ferrite Lime Periclase Arcanite Aphthitalite Thenardite Ca-Langbenite Sulfate-spurrite Ellesteadite Fluorellesteadite Mayenite

Aluminate

Belite

Formula

Phase

Table 1. Structural details for phases that may be present in OPC clinkers.

174 Aranda, De la Torre, León-Reina

141.0 193.4 188.7 220.1 194.3 193.8 134.7 92.3 205.9 169.9 169.9 99.5 153.8 1163.5 1183.4 222.4 253.5 258.3 203.0 193.8

m (cm−1) 151692 79528 24473 16382 157072 80869 31277 41414 87144 80361 9560 23867 69380 82904 49549 30884 2282 43078 158177 34591

ICSD codes 33-0311 01-083-0438 01-073-1942 01-072-0916 28-0739 01-086-0174 01-075-1711 46-1045 01-089-5917 42-1478 01-071-0969 01-073-1389 01-080-1547 01-087-1166 01-077-1545 00-043-1460 01-070-1138 01-089-2432 00-035-0592 00-035-0590

PDF codes [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

Ref#

References: [1] De la Torre et al. 2004; [2] Bezou et al. 1995; [3] Floerke 1952; [4] Kirfel and Will 1980; [5] Ballirano et al. 2005; [6] Maslen et al. 1995; [7] Effenberger et al. 1983a; [8] Will et al. 1988; [9] Louisnathan 1971; [10] Calos et al. 1995; [11] Saalfeld and Depmeier 1972; [12] Sadanaga et al. 1962; [13] Steele and Pluth 1990; [14] Sawada 1996; [15] Fleet 1984; [16] Hesse 1984; [17] Saburi et al. 1976; [18] Yamaguchi and Suzuki 1967; [19] Gemmi et al. 2007; [20] Onken 1965.

#

Monoclinic Monoclinic Hexagonal Orthorhombic Monoclinic Rhombohedral Rhombohedral Rhombohedral Tetragonal Orthorhombic Cubic Orthorhombic Rhombohedral Rhombohedral Cubic Monoclinic Monoclinic Monoclinic Tetragonal Orthorhombic

CaSO4.2H2O CaSO4.0.5H2O CaSO4 CaSO4 K2Ca(SO4)2.H2O CaCO3 CaMg(CO3)2 SiO2 Ca2Al2SiO7 Ca4Al6SO16 Ca4Al6SO16 Al4SiO8 Ca5.5Al11Si5O32 Fe2O3 Fe3O4 CaSiO3 Ca3Si2O7 Ca3Mg(SiO4)2 Ca2Mg(Si2O7) CaMg(SiO4)

Gypsum Hemihydrate Anhydrite-III Anhydrite-II Syngenite Calcite Dolomite Quartz Gehlenite Yeelemite

Mullite Yoshiokaite Hematite Magnetite Wollastonite Rankinite Merwinite Akermanite Monticellite

Crystal system/ notation

Formula

Phase

Table 2. Structural details for additional phases that may be present in OPC (blended) cements.

Rietveld Quantitative Phase Analysis of OPC 175

Rhombohedral Structure not reported Hexagonal Rhombohedral Rhombohedral Rhombohedral Triclinic Monoclinic Structure not reported Structure not reported Rhombohedral

Ca(OH)2 Al(OH)3 Ca5Si6O16(OH)2·7H2O Ca9Si6O18(OH)6·8H2O Ca3Al2(OH)12 Ca3Al2(OH)7.6(SiO4)1.1

Ca6Al2(OH)12(SO4)3·26H2O Ca6Al2(OH)12(CO3)3·26H2O Ca6Si2(OH)12(CO3)2(SO4)2·24H2O

Ca4Al2(OH)12[SO4]·6H2O

Ca4Al2(OH)12[Cl]2·4H2O Ca4Al2(OH)12[(SO4)0.5Cl]·5H2O Ca4Al2(OH)12[CO3]·5H2O Ca4Al2(OH)12[Cl(CO3)0.5]·4.8H2O Ca4Al2(OH)12[OH(CO3)0.5]·5.5H2O Ca4Al2(OH)12[Al(OH)4]2·6H2O Ca4Al2(OH)12[AlSi(OH)8]2·2H2O (Gehlenite hydrate)

Portlandite Gibbsite Tobermorite Jennite Hydrogarnet or C3AH6 Katoite AFt Ettringite Ettr-CO3 Thaumasite AFm

59327 63250

69413

98.8

88617

146.7 124.0 124.7 130.1

100138

31247

85.7 115.8

155395

15471 6162 152489 151413 202316 172077

ICSD codes

84.4

211.4 57.1 145.0 164.3 163.7 185.5

m (cm-1)

01-089-8294 00-019-0203 01-087-0493 01-078-2050 00-041-0221 00-011-0205 01-080-1579

01-083-1289

00-041-1451 00-036-1465 01-075-1688

01-072-0156 01-070-2038 00-029-0331 00-018-1206 01-084-1354 00-038-0368

PDF codes

[15]

[10, 11] [12] [13] [14]

[9]

[8]

[7]

[1] [2] [3] [4] [5] [6]

Ref#

References: [1] Petch 1961; [2] Saalfeld and Wedde 1974; [3] Bonaccorsi et al. 2005; [4] Bonaccorsi et al. 2004; [5] Lager et al. 1987; [6] Ferro et al. 2003; [7] Goetz-Neunhoeffer and Neubauer 2006; [8] Effenberger et al. 1983b; [9] Allmann 1977; [10] Renaudin et al. 1999; [11] Rousselot et al. 2002; [12] Mesbah et al. 2011; [13] François et al. 1998; [14] Sacerdoti and Passaglia 1988; [15] Rinaldi et al. 1990.

#

Kuzelite or C4A SH12 Friedel’s salt Kuzel’s salt Monocarbo-aluminate Hydrocalumite Hemicarbo-aluminate C2AH8 Strätlingite or C2ASH8

Rhombohedral Monoclinic Monoclinic Triclinic Cubic Cubic

Formula

Phase

Crystal system/ notation

Table 3. Structural details for additional phases that may be present in OPC hydration products.

176 Aranda, De la Torre, León-Reina

Rietveld Quantitative Phase Analysis of OPC

177

calcium silicate hydrates has been reported (Richardson 2008). Furthermore, the reader is also referred to an excellent review dealing with the density of cement phases including those of AFm and other hydrates (Balonis and Glasser 2009). The mineralogy (phases) of the hydrates compounds may depend upon the additions in blended cements (Matschei and Glasser 2010) and this should be taken into account.

Whole-pattern quantitative phase analysis approaches As stated above, conventional RQPA requires all crystal structures to be known. There are alternative whole-pattern quantitative phase analysis methods for crystalline phases with unknown structures (Smith et al. 1987; Taylor and Zhu 1992; Scarlett and Madsen 2006) however these approaches will not be discussed/reviewed here. The output of a RQPA study is a set of m-crystalline phase scale factors, ΣmSa, for a sample with m-crystalline phases. A phase scale factor, Sa, is related to the phase weight content, Wa, by Equation (4) (Hill and Howard 1987; Bish and Howard 1988)  Wa  Sa = K e   2  raVa m s 

(4)

where Ke is a constant which depends on the diffractometer operation conditions, ra is the crystallographic density of the a-phase, Va is the unit cell volume of a-phase, and ms is the sample mass absorption coefficient. Equation (4) can be rewritten as given in Equation (5):   Wa Sa = K e    ( ZMV )a m s 

(5)

where instead of using ra, the relation between Sa and Wa is based on the “ZMV” term with Z being the number of chemical units/formulas within the unit cell, M being the molecular mass of the chemical formula, and V the unit cell volume. Once the crystal structure is known, the “ZMV” term is known. In any case, the parameter to be extracted, Wa, depends on the phase scale factor, Sa, but also on Ke and ms. Unfortunately, these two variables are not known and they cannot be derived from the powder diffraction pattern of the sample under study. Currently, there are three main ways to derive the phase content, Wa, from the Rietveld refined scale factor, Sa. These three methods are based on different mathematical approaches and they have different experimental complexities. They will be treated in detail just below. I) Normalization to full crystalline phase content method. The simplest approach is the approximation that the sample is composed only of crystalline phases with known structures. These crystal structures are incorporated into the control file, and it was already shown (Hill and Howard 1987) that the weight fraction of a-phase, for a m-crystalline phase mixture, may be given by Equation (6):

Wa =

Sa ( ZMV )a m

∑ Si (ZMV )i

(6)

i =1

The use of Equation (6) in RQPA eliminates the need to measure the instrument calibration constant, Ke, and the sample mass absorption coefficient, ms. However, the method normalizes the sum of the analyzed weight fractions to 1.0. Thus, if the sample contains amorphous phases, and/or some amounts of unaccounted for crystalline phases, the analyzed weight fractions will be overestimated. This approach is by far the most widely used method in RQPA and also in RQPA of OPCs. However, it must be highlighted that the resulting weight fractions are only accurate if the

178

Aranda, De la Torre, León-Reina

amount of unaccounted crystalline phases and amorphous content are very small (negligible) which may be not the case in anhydrous OPCs and for sure is not the case in OPC pastes. II) Internal standard method. A second, more experimentally-demanding, approach is to mix the sample with a crystalline standard in a known amount, Wst (also known as spiking method). This standard must be free of amorphous content or at least its non-diffracting content must be known. This (artificial) mixture must be homogenized as the particles should be randomly arranged. Under these considerations, Equation (7) is fulfilled, where Sst is the Rietveld scale factor of the standard in the artificial mixture (Hill and Howard 1987): Wa = Wst

Sa ( ZMV )a Sst ( ZMV )st

(7)

The addition of the standard dilutes the crystalline phases in the sample. This may be quite problematic for low-content phases. A procedure for Rietveld quantitative amorphous content analysis was outlined, in which the effects of systematic errors in the powder patterns were studied (De la Torre et al. 2001a). The approach requires two powder patterns to be collected but it minimizes the errors. Firstly, a powder pattern of the sample of interest is collected and the crystalline phase contents are determined as stated in the previous section (normalized to a 100% of crystalline phases). A second pattern is collected for the sample mixed with the internal standard. The weight fraction added of the internal standard is known, Wst, as it was weighed. The RQPA gives an overestimated value of this content, Rst, due to the presence of amorphous phase(s), misfitting problems of the analyzed crystalline phases, and because some crystalline phases may not be included in the control file due to several reasons (its crystal structure is not known, the phase was not identified, etc.). This overall content is hereinafter named ACn which stands for Amorphous and Crystalline not-quantified, to highlight that not only an amorphous fraction but any not-computed crystalline phase and any misfit problem (for instance the lack of an adequate structural description for a given phase) may contribute to this number. The method derives the (overall) ACn content of the sample from the small overestimation of an internal crystalline standard, Equation (8) (De la Torre et al. 2001a): = ACn

1 − ( Wst Rst )

(100 − Wst )

× 10 4 (wt%)

(8)

For this second refinement, the phase ratios between the crystalline phases in the sample may be kept fixed to the values obtained in the first fit and only the fraction of the added standard is optimized. Under these conditions, the overall ACn content is derived and this is applied to the weight fractions determined in the first fit to place them in an absolute scale. The errors associated to this approach and the optimum amount of standard has been recently discussed (Westphal et al. 2009). Finally, NIST standard reference material (SRM) 676a, corundum (a-Al2O3) powder, has been certified to have a phase purity of 99.02% ± 1.11% (95% confidence interval) by RQPA against a suitable primary standard (powder silicon carefully prepared from a single crystal). This novel certification method permits quantification of amorphous content for any sample of interest by this spiking method (Cline et al. 2011). This methodology has been applied to anhydrous OPC and also to pastes. However, the addition of an internal standard may alter the OPC hydration reactions and dilutes the phases in the pastes. III) External standard method (G-factor approach). To avoid complications that may arise from mixing an internal standard with the sample, it is possible to use an external standard method. This approach requires the recording of two patterns in identical diffractometer configuration/conditions for Bragg-Brentano q/2q reflection geometry. The method was proposed some time ago (O’Connor and Raven 1988) and very recently applied to anhydrous cements (Jansen et al. 2011a) and organic mixtures (Schreyer et al. 2011). This method

Rietveld Quantitative Phase Analysis of OPC

179

consists in determining the diffractometer constant, Ke, with an appropriate standard (for instance silicon powder from Si-single crystal), see Equation (9), derived from Equation (4): G K= = Sst e

rstVst2m st Wst

(9)

where Sst is the Rietveld scale factor of the (external) standard, rst corresponds to the density of the standard, Vst is the unit cell volume of the standard, Wst is the weight fraction of the standard (ideally 100 wt%) in the external standard pattern, and mst is the mass attenuation coefficient of the standard. This method is also known as G-method as the standard allows to calculate the G-factor of the diffractometer in the operating conditions. The calculated G-factor represents a calibration factor for the whole experimental setup and comprises the diffractometer used, radiation, optics, and all data acquisition conditions, (e.g., detector configuration, integration time, etc.). This G-factor is used to determine the mass concentration of each phase of the sample under study using Equation (10):

Wa = Sa

raVa2m s G

(10)

This method allows determining the absolute weight fractions by using a diffractometer constant that must be previously determined. However, the mass attenuation coefficient of the sample is needed, ms. This must be independently determined, and the most common way to obtain it, in cements, it is by X-ray fluorescence spectrometry (Jansen et al. 2011a). This methodology has been applied to anhydrous OPCs (Jansen et al. 2011a) and to OPC pastes (Jansen et al. 2012a,b). From Equation (10), it is clear that the weight fractions of all computed phases within the control file do not need to add up to 100 wt%. From the difference between 100 wt% and the sum of the crystalline phase contents, an overall unaccounted/left-out weight percentage, ACn, can be derived. In any case, the external standard method is experimentally more demanding but it may have the brightest future for the study of hydration reactions. This sentence is based on the lack of need of spiking an internal standard. The addition of an internal standard is a very important drawback in hydration reaction studies since not only dilutes the phases in the sample but it may also interfere with the hydration reactions.

SAMPLE PREPARATION AND DATA COLLECTION The preparation of samples for laboratory X-ray powder diffraction (LXRPD) is the crucial first step for a good quantitative analysis (Jenkins et al. 1996; Bish and Reynolds 1989; Buhrke et al. 1998). In order to extract the maximum (accurate) information from a powder pattern, two parameters have to be carefully controlled: i) the positions of the diffraction maxima should be at the right place, which means good diffractometer alignment/maintenance and good sample mounting practice; and ii) the peak intensities should be those expected from the crystal structure and phase assemblage, which means that a sufficiently large number of crystallites contributing to each reflection of every phase should be bathed by the X-rays. It is necessary to obtain diffraction peaks of reproducible intensity without altering the sample or inducing too much preferred orientation. The particle statistics is of the outmost importance in quantitative studies (Elton and Salt 1996; Whitfield and Mitchell 2009). For CuKa radiation in Brag-Brentano configuration, the relative standard deviation of the measured intensity due to particle statistics is less than a few percent when the particle size is smaller than 5 mm; however, the statistical error increases rapidly as the particle size is bigger than about 10 mm (Smith 1992; Jenkins et al. 1996). For OPC samples where alite has usually particles much larger than that, this is a very serious problem. Severe grinding of the

180

Aranda, De la Torre, León-Reina

sample can be performed in order to reduce the average alite particle sizes to values close to 5 mm, however this treatment may alter the sample as partial amorphization may take place. Alternatively, rotating a flat-plate sample in its own diffracting plane does not help to reduce the prefer orientation effect but increases the reproducibility of the recorded intensities. Hence, samples with large particle sizes should be measured in the rotating stage and this recording setup is strongly recommended for Rietveld studies (McCusker et al. 1999). Furthermore, rotating the sample should be compulsory for RQPA of OPC samples. If milling is necessary, then the addition of a milling agent (e.g., n-propyl alcohol or acetone) is recommended as this significantly increases the efficiency of grinding by uniformly reducing the particle size of the hard materials, without overgrinding the softer ones (Bish and Reynolds 1989). Extensive grinding may lead to the transformation of gypsum into bassanite (or even anhydrite) and even the amorphization of sulfate phases (Fullmann and Walenta 2003; Enders 2005). Milling OPC materials usually does not help to avoid microabsorption problems which may be an important source of errors in RQPA. This problem has been extensively discussed for cements (Le Saout et al. 2011) so it is not further treated here. We want just to mention that this effect may be corrected (Brindely 1945; Taylor and Matulis 1991) but its implementation in practice is extremely difficult. On the other hand, OPC samples contain phases with different hardness’s. The necessary clinker grinding may lead to higher concentrations of brittle and softer phases (gypsum, alite) in the finer fraction and harder phases (e.g., belite) in the coarser fraction. This was already demonstrated by Gutteridge in 1984 and therefore, ground samples should not be sieved. Some OPC samples contain unstable phase(s) in room conditions. Sample preparation should be carried out to avoid/minimize sample alteration. In this case, there are differences between anhydrous OPCs and hydrated pastes. For anhydrous OPCs, the main unstable phase is free lime that is hydrated to portlandite and later carbonated to calcite within days under laboratory atmosphere (Fullmann and Walenta 2003). Therefore, if CaO is going to be analyzed, data should be recorded as soon as possible in flat-plate reflection (Bragg-Brentano) geometry, or sealed in an appropriate capillary when working in transmission (Debye-Scherrer) geometry. These considerations are even more important for OPC pastes as recently formed portlandite is very reactive and it is easily carbonated. Therefore, the flat-surface pastes should be covered with Mylar or Kapton thin tapes (see for instance, Jansen et al. 2011b). Finally, a lot of care in the experimental setup has to be devoted in order to avoid self-desiccation problems. A very recent study compared reflection versus transmission (both capillary and flat-sample) geometries for in-situ laboratory X-ray powder diffraction of Portland cement hydration at early ages (Dalconi et al. 2011). Unfortunately, the flat-sample held between two Kapton films underwent hydration stop due to self-desiccation very likely because a perfect sealing was not achieved. It was also observed that higher amounts of ettringite and portlandite were quantified analyzing Bragg-Brentano data with respect to capillary data. These discrepancies were related to sample segregation effects in reflection geometry, which is always a concern in this type of configuration.

SELECTED EXAMPLES OF RIETVELD QUANTITATIVE PHASE ANALYSIS The initial work on RQPA of Portland clinker was published back on 1993 (Taylor and Aldrige 1993). Soon after, other works showed the advantages of this method which does not require standards or calibration curves (Moller 1995, 1998; Meyer et al. 1998; Taylor et al. 2000). Initially, the studies were focused on OPC clinkers and later expanded to different types of cements. The accuracy of such type of analysis was investigated (De la Torre and Aranda

Rietveld Quantitative Phase Analysis of OPC

181

2003). Subsequently, RQPA has been employed to investigate the hydration products of several types of OPC cements. The main uses of Rietveld quantitative analysis for extracting information of OPC materials are given in Figure 1. The next subsections review some papers on these uses but the discussed list is not intended to be exhaustive.

Clinkers A RQPA of a commercial OPC clinker was studied by high-energy (l = 0.442 Å) synchrotron powder diffraction in Debye-Scherrer (transmission) configuration (De la Torre et al. 2001b). Data for the same sample were also collected in a laboratory diffractometer (CuKa1,2, l = 1.54 Å) in Bragg-Brentano (reflection) configuration. The high-resolution nature of the synchrotron data allowed establishing the phase’s polymorphism without selective dissolution. This information is important not only for accurate RQPA, but also because the polymorphs may influence the strength development during hydration (Stanek and Sulovsky 2002). Furthermore, the use of capillaries of 1.0 mm of diameter in the synchrotron study allowed a large amount of sample to be tested which yield very reproducible diffraction intensities and so accurate RQPA results. However, the expensive nature of synchrotron measurement only allows its use for very special problems or for proof-of-principle experiments. The use of RQPA for analyzing OPC clinkers was nicely exemplified in the study of a commercial white Portland clinker (De la Torre et al. 2003). The RQPA results of the white clinker normalized to 100 wt% of crystalline phases gave 77.3 wt% of C3S, 19.8 wt% of C2S and 2.9 wt% of C3A. However, Bogue calculations from XRF data gave: 71.3, 12.2 and

Main uses of RQPA for OPC materials Clinkers

Cements

Modifications in the kiln Use of alternative mineraliser/flux agents Addition of (industrial) wastes to the raw meals

Hydration products

Single phases Analysis of blended cements (modelling) Hydration of OPCs Alteration during storing Durability Influence of w/b Mineralogy ratio, T & P Role of of sulfates

Impact of the change of raw materials Laboratory quality control On-line control of production

This know-how is also being used for new eco-cements !

superplasticizers

Hydration of blended cements Binary (f.i. OPC + FA)

Special cements

Ternary (f.i. OPC + FA + CC or: OPC + BFS + FA)

Figure 1. Main uses of Rietveld quantitative phase analysis for studying ordinary Portland clinkers, cements and hydration products.

Aranda, De la Torre, León-Reina

182

13.1 wt%, respectively. It is clear that there is a strong disagreement between RQPA and Bogue calculations for C3A which is necessary for understanding cement sulfate-resistance properties. Annealing the commercial white clinker at 1500 °C for 1 hour allowed obtaining a sample with obviously the same elemental analysis, but quite different crystalline phase content: 76.8 wt% of C3S, 14.1 wt% of C2S and 9.1 wt% of C3A, see Figure 2, demonstrating the advantage of RQPA over XRF.

C3S C3A

C3S

C3S

C3A

C3S

º/2

C2S

C3S

C3S

C3S

C2S

C3S

C3S

C3S

The use of synchrotron X-ray powder diffraction (SXRPD) for grey clinker analysis gave better results since the data have higher resolution (peak widths are narrower) and the intensities are much more accurate (working in transmission with a short wavelength allows to test a much larger amount of crystallites). An initial study with commercial grey OPC clinkers showed alite phase coexistence for a range of magnesium contents (De la Torre et al. 2005), see Figure 3. The MgO and SO3 contents, for a clinker showing alite coexistence, were 1.24 and 1.15 wt%, respectively (measured by XRF). Furthermore, the ACn content was determined in several samples using the internal standard approach and it was shown that white OPC clinkers have much higher contents than grey OPC clinkers, ~20 and ~8 wt%, respectively. The study of NIST reference material RM8488 by the internal standard method gave an overall

º/2 Figure 2. (top) Selected region of a Rietveld plot (CuKa1,2) for a commercial white Portland clinker with main peaks labeled. (bottom) Selected region of the Rietveld plot for the same clinker annealed at 1500 °C for 1 hour. Note the increase of the C3A main diffraction peak (modified after De la Torre et al. 2003).

C3S1

C3S1

C4AF

C4AF

C 3A

C3S2

C3S2

C3S2 C 2S

NaK3(SO4)2

C2S

NaK3(SO4)2

C3S2

C3S1

C3S1

C3S2

183

Counts

-1.0

0.0

1.0

2.0

X10E 4 3.0

4.0

V/Z (C3S1) = 120.4 Å3, Mg poor V/Z (C3S2) = 119.9 Å3, Mg rich

C3S1

Rietveld Quantitative Phase Analysis of OPC

8.0 2-Theta, deg

8.2

8.4

8.6

8.8

9.0

9.2

9.4

Figure 3. Selected region of a high-resolution synchrotron Rietveld plot (l = 0.43 Å) for a commercial grey Portland clinker with main peaks labeled. Note the alite phase coexistence (modified after De la Torre et al. 2005).

ACn content of 8 wt% (De la Torre et al. 2006). It was also possible to use synchrotron X-ray powder diffraction to characterize the quality of laboratory powder diffraction data. One study analyzed the same OPC by SXRPD data, LXRPD-CuKa1 and LXRPD-CuKa1,2 (see Fig. 4). It was concluded that LXRPD-CuKa1 diffraction peaks were only slightly broader than those in the SXRPD pattern (De la Torre et al. 2006). The quantification of the ACn content in Portland clinkers has been a matter of deep research for a number of years. Suherman et al. (2002) already carried out a thorough study using SXRPD and LXRPD as well as internal and external standard approaches. Furthermore, selective dissolution and point-counting microscopy were also used to have the best possible results. The overall ACn contents for NIST reference clinker materials RM8486, RM8487 and RM8488 were ~6 wt%, ~11 wt% and ~7 wt%, respectively. Two commercial OPC grey clinkers were also analyzed and the determined ACn contents were 14 and 8 wt%. These NIST reference materials were also studied by RQPA and the importance of correct peak shape fits was stressed in order to have good RQPA results (Pritula et al. 2003). A subsequent analysis was dedicated to the study of the overall non-diffracting content in several commercial grey OPC clinkers and cements. The used approach was the internal standard (spiking) method and ACn contents close to 20 wt% were determined (Whitfield and Mitchell 2003). Neutron powder diffraction (NPD) has also been used for RQPA of clinkers. NPD refinements were found to be more stable than the X-ray ones, but there was not any notable difference in the final determined phase assemblage (Pritula et al. 2004a). Furthermore, NPD, SXRPD and LXRPD were used to study the NIST reference clinker material RM8488. The study gave quite satisfactory results although it seems that the quantification of C3A by NPD was somewhat unstable (Peterson et al. 2002, 2006). CO2 emission reduction can be obtained in Portland clinkers manufacturing by partly replacing calcite by calcium oxide-bearing waste materials such as blast furnace slag or fly ash. RQPA was used to derive the phase assemblages of these laboratory-prepared clinkers

Aranda, De la Torre, León-Reina

184 y

C3 A C2S 8.2

32.0

C4AF C4AF

C3A

31.0

34.0

35.0

C3S

33.0

33.0

C4AF

C 3A

C 2S 32.0

C4AF

NaK3(SO4)2

C2S 31.0

C3S

C3 S 30.0

NaK3(SO4)2

C3 S

C3S

I (a.u)

8.8

C3 C3S2 C2S

NaK3(SO4)2

C2 S

NaK3(SO4)2

C3S1 C3S2 y 30.0

 = CuK1,2

29.0

8.6

S1

C3S1

C3S1 C3S2 29.0

8.4

C3S1

8.0

C4AF C4AF

C3S1 C3S2 C3S1 C3S2

7.8

 = CuK1

C3S2

C2S

NaK3(SO4)2

C3S1 C3S2 7.6

NaK3(SO4)2

C3S1 C3S2

I (a.u) I (a.u)

7.4

C3S1

 = 0.40 Å (synchrotron data)

34.0

º/2

35.0

Figure 4. Selected region of the Rietveld plots for the same commercial Portland clinker. Data were collected at: ID31 diffractometer of ESRF synchrotron (top), a laboratory diffractometer with a Ge(111) primary monochromator (middle), and a laboratory diffractometer with a graphite secondary monochromator (bottom). Note the similar resolution of synchrotron and CuKa1 patterns (modified after De la Torre et al. 2006).

Rietveld Quantitative Phase Analysis of OPC

185

using these special raw meals (Chen and Juenger 2009a,b). However, it is now recognized that these materials (blast furnace slag or fly ash) are more eco-friendly as additions to ordinary Portland clinkers to produce blended cements in Europe or directly in the fabrication of concretes (North America). Therefore, the use of RQPA for studying these additions will treat in the next section dedicated to OPC cements. Another general use of RQPA is to study the role of mineralizing/flux agents in the kiln or the processing of industrial wastes in the kilns by analyzing the resulting clinker phase assemblage. As it has been described (Herfort et al. 2010), the mechanism of action of the trace elements incorporated into the clinker can be divided in two main steps depending on the amount of these elements: i) at low concentration, trace elements enter into the structure of the initial phases of the clinker as solid solutions (C3Sss, C2Sss, C3Ass, C4AFss); ii) at higher concentrations, the presence of new phases may be detected and quantified (Gineys et al. 2011). RQPA may help in both cases, initially by following the unit cell variations of the solid solution(s) and secondly, by quantification of the new phases. The implications of processing industrial waste(s) in the Portland kilns have been studied by RQPA. For instance, this approach has been used to evaluate the effects of chromium and nickel additions to Portland clinker raw meals. The final goal was to better characterize the processing of galvanic sludge which is the main hazardous solid waste produced by some metallurgical industries. A small decrease in the C3S contents were measured as the amount of Ni and Cr in the raw meals increases (Ferreira et al. 2008). For the production of Portland clinkers, mineralizers and/or fluxes are added to the raw mixes to accelerate reactions and enhance burnability. The traditional fluxes (Fe2O3 and Al2O3) may be partially substituted by the mineralizing pair CaF2/CaSO4 to produce clinkers with low aluminate contents at temperatures close 1400 °C. This is particularly useful for manufacturing white Portland clinkers because of the potential for energy conservation and seawater resistance. However, new phases may appear and in order to carry out RQPA, the crystal structures must be known. A work (Pajares et al. 2002) identified fluorellestadite in the mineralized white Portland clinker. Its crystal structure was determined and a satisfactory RQPA of the mineralized white Portland clinker was obtained. Figure 5 displays the Rietveld plot of this mineralized white Portland clinker, and a similar Rietveld plot for an ordinary white Portland clinker is also given for the sake of comparison.

Cements RQPA in OPC cements can be used for a number of applications including: i) quantification of the crystalline phases in OPCs, including the ACn content if needed; ii) to analyze the amount and role of sulfate-containing phases; iii) to study the mineralogical phase assemblage in the materials used as addition(s), including their ACn contents; iv) to quantify all phases in blended OPC cements. The overall amorphous content has been studied in OPC clinkers as previously discussed. This type of study has also been carried out in cements and a thorough study by the external standard method (Jansen et al. 2011a) concluded that no significant amorphous content could be proven in that particular analyzed OPC cement. Commercial OPCs contain sulfate carriers in variable amounts. Gypsum (or other calcium sulfates) is added to the clinker during the milling process, where it may partially dehydrates to bassanite or even to soluble anhydrite-III. Due to different hydration kinetics of these phases, it is necessary to characterize the mineralogical composition of sulfate in a cement system in order to reach an optimal and reproducible setting and cement hydration (Seufert et al. 2009a). The formations and transformations of the five different phases in the CaSO4-H2O system have been studied in detail (Christensen et al. 2008). The use of RQPA for determining the amounts of gypsum, bassanite and anhydrite in OPC was already demonstrated several years ago (Fullmann and Walenta 2003; Walenta and Fullmann 2004; Seufert et al. 2009b). The

C3S C11A7.CaF2 FLELL

FLELL

C2S

C3S

FLELL

C2S, FLELL

C3S C3S

X10E 3 3.0 2.0

C3S

C3S C3A

C2S

C2S

C2S

Counts

.0

2.0

C3S

C3S

X10E 3 4.0

C3S

.0

1.0

C3S

Aranda, De la Torre, León-Reina

186

28.0 2-Theta, deg

29.0

30.0

31.0

32.0

33.0

34.0

35.0

Figure 5. (top) Selected region of a Rietveld plot (CuKa1,2) for a mineralized white Portland clinker. (bottom) Selected region of the Rietveld plot for an ordinary commercial white Portland clinker. The main peaks are labeled and FLELL stands for fluorellestadite (modified after Pajares et al. 2002).

mineralogy of the sulfate source is very important for the fluidity, setting and hydration of mortars and concretes (Tang and Gartner 1988; Rossler et al. 2008). This technique has been very recently used for quantitative determination of the hydration products formed within minutes of mixing (e.g., ettringite, syngenite and secondary gypsum), to help identify the cause(s) of early stiffening (Ramlochan and Hooton 2011). Today supplementary cementitious materials (SCMs) are widely used in concrete either in blended cements or added separately in the concrete mixer. The use of silica rich materials influences the amount and kind of hydrates formed and thus the volume, the porosity and finally the durability of these materials (Lothenbach et al. 2011). Therefore, it is easy to understand that the mineralogical phase assemblage in different common additions has been deeply studied including PFA, bottom ash, metakaolin and BFS (Kumar et al. 2008; Korpa et al. 2009; Gonçalves et al. 2009; De Weerdt et al. 2011; Narmluk and Nawa 2011). Furthermore, RQPA has also been used to characterize other less common additions such as drinking water treatment plant sludge (Husillos-Rodriguez et al. 2010) or natural zeolites (Snellings et al. 2010). Table 4 reports a RQPA study of a fly ash carried out in our laboratory by the internal standard method described above. Figure 6 (top) shows the Rietveld plot for

Rietveld Quantitative Phase Analysis of OPC

187

Table 4. Rietveld quantitative phase analysis (wt%) of a fly ash. Phase Mullite Hematite Quartz Wollastonite Corundum (WS = 47.7 wt%) ACn content#

Fly ash + internal standard RQPA direct result

Fly ash analysis Final result

9.5(2) 3.9(2) 7.4(2) 0.8(2) (RS = ) 78.3(6) -----

11.1(4) 4.6(6) 8.7(6) 1.0(6) ----74.6

# ACn stands for amorphous plus not-quantified crystalline phase(s) which includes misfitting problems and not-computed phase(s).

this type of analysis with the main peaks labeled. An overall ACn content of ~75 wt% was obtained with mullite and quartz being the main crystalline phases (see Table 4). Finally, the analysis of blended cements containing BFS and PFA additions by RQPA (with internal standard) was already reported several years ago (Westphal et al. 2002; Walenta Fullmann and 2004). Many more studies have been reported in this subject mainly linked to the hydration characterization. Hence, some of these papers will be discussed in the next subsection. As an example of this type of analysis, Figure 6 (bottom) shows the Rietveld plot of a RQPA for a blended OPC cement obtained with the fly ash described in the previous paragraph. It was possible even to quantify the mullite content, in the cement. This number may be used to track down the approximate amount of fly ash added to the cement in the industry. Quartz is not a suitable compound to carry out this type of calculations as it may be present in the gypsum and/or in the additions like limestone, chert, etc.

Hydration products RQPA has been employed for a number of applications related to the hydration reactions of OPC materials. The uses have been expanded from the hydration of model systems (for instance a single phase or an artificial mixture, Bellman et al. 2010) to blended cements and the role of admixtures and superplasticizers. It must be highlighted that the impact of admixtures on the hydration kinetics of Portland cement has been recently reviewed (Cheung et al. 2011) although that paper focused on the materials and not on the techniques to be used. RQPA was used to study the hydration reactions of commercial OPC in reflection geometry with laboratory data (Scrivener et al. 2004). The results were satisfactorily compared to those obtained from thermal analysis and electron microscopy. RQPA, using the internal standard approach in transmission geometry, was employed for studying OPC hydration products (Mitchell et al. 2006). The data obtained from capillary measurements showed little preferential orientation, and produced the progression of phase contents expected from the reaction. This study highlighted the benefits of the transmission geometry as more particles are measured which yields more reliable quantitative results. The early hydration of white Portland cement was also studied by in situ X-ray powder diffraction (reflection geometry) at defined temperatures and with different water/cement ratios (Hesse et al. 2008, 2009). The hydration of an OPC cement at 28 days was studied by RQPA using the external and internal standard methods including the role of isopropanol to stop the hydration reaction (Le Saout et al. 2007). Figure 7 displays the typical Rietveld plot for that material where the relevant peaks are labeled. RQPA based on in-situ synchrotron powder diffraction was used to monitor the evolution of hydrous phases during early hydration with a time resolution of 10 seconds (Weyer et al. 2005). A related work but with lower time resolution (minutes) studied the hydration process

Al2O Al32O3

Al2O Al32O3 Al2O Al32O3

Mullite Mullite

Mullite Mullite

Blended OPC cement Blended OPC cement

Quartz Quartz

Calcite Calcite

Al2O Al32O3 Quartz Quartz

Quartz Quartz

Quartz Quartz

Mullite Mullite

Gypsum Gypsum

Mullite Mullite

Al2O3 (spiking method)

Al2O Al32O3

Al2O Al32O3

Al2O Al32O3

Fly ash + Fly + Al2Oash 3 (spiking method)

Al2O Al32O3

Aranda, De la Torre, León-Reina

188

Figure 6. (top) Full region of a Rietveld plot (CuKa1) for a fly ash mixed with Al2O3 as internal standard (47.7 wt%). The main peaks are labeled. (bottom) Full region of the Rietveld plot (CuKa1) for an industrial blended OPC obtained from that fly ash. The main diffraction peaks, which are not due to clinker phases, are highlighted.

Rietveld Quantitative Phase Analysis of OPC

CH

CH

189

CSH2

Figure 7. Rietveld plot (CuKa1,2) for a 28 days hydrated OPC cement. The upper right inset shows the influence of isopropanol for the cement hydrated at 28 days (modified after Le Saout et al. 2007).

on synthetic clinker phases (C3A and C4AF) and on commercial OPC cements (Merlini et al. 2007a,b). Furthermore, synchrotron powder diffraction was also used to monitor the evolution of ettringite in C3A-gypsum synthetic mixture and in commercial OPC cement systems during the first hours of the hydration process (Merlini et al. 2008). In-situ synchrotron RQPA was also carried out for studying the very early hydration of Class A and H oil well Portland cements with different amounts of CaCl2 at 25 and 50 °C (Jupe et al. 2007). On the other hand, synchrotron radiation may be used in more sophisticated types of characterization. For instance, high-energy synchrotron X-ray microdiffraction was used to quantify the orientation distribution of ettringite crystals. Diffraction images were analyzed using the Rietveld method to obtain information on textures within thin slabs of mortars (Wenk et al. 2009). The hydration reactions of an alkali-poor and an alkali-rich OPC were followed by RQPA. Significant differences during the early hydration were measured due to the presence of syngenite, K2Ca(SO4)2·H2O, when the alkali content is high and secondary gypsum when the alkali content is low. Furthermore, the pore solutions of the hydrated cements were analyzed and super- or undersaturations for relevant minerals were calculated (Stark et al. 2008). In situ X-ray diffraction for monitoring cement hydration was used to study well defined Portland cement clinkers consisting of alite and aluminate doped with different amounts of Na2O. Other techniques such as isothermal conduction calorimetry and differential scanning calorimetry, scanning electron microscopy and 27Al NMR technique were also used (Wistuba et al. 2007). Finally, flash setting accelerators (both alkali-rich and alkali-free) are a class of admixtures commonly used for sprayed concrete during tunnel excavation. RQPA was also used to studying the setting behavior in this special application (Maltese et al. 2007). RQPA has been used in many works to study the hydration reactions of blended cements. Initially, this technique was used to analyze the hydration progress of cement pastes prepared by adding BFS and limestone powder (Hoshino et al. 2006). Selective dissolution was also employed to distinguish between the amorphous contents coming from BFS and newly-formed

190

Aranda, De la Torre, León-Reina

CSH gel (the amorphous Calcium-Silicate-Hydrated gel formed in the hydration of the calcium silicates). It was concluded that BFS accelerates the hydration of C3S, C3A and especially C4AF. The early age hydration and pozzolanic reaction in natural zeolite blended Portland cements has been studied by in situ synchrotron RQPA to determine the reaction kinetics and products (Snellings et al. 2010). A deep study was also carried out to quantitatively explain the effect of water curing condition on compressive strengths of fly ash cement pastes (Termkhajornkit et al. 2006). Replacement ratios of fly ash were 0%, 25% and 50% of total powders and the water to binder ratio was relatively low, 0.80 and 1.00 by volume. The time evolution of every OPC initial phase was worked out by RQPA. Unfortunately, the time evolution of portlandite was not quantitatively reported in that paper. The hydration degree of belite was the most affected parameter by the fly ash. On the other hand, a very recent and complete study used RQPA, together with thermogravimetry, scanning electron microscopy and isothermal calorimetry, in order to understand the hydration mechanisms of blended Portland cements containing fly ash and limestone powder (De Weerdt et al. 2011). In addition, pore solution analysis and thermodynamic modeling techniques were also employed. The time-evolution of all phases during hydration were studied (including portlandite from TGA and RQPA), and not only the pozzolanic effect was studied, but the variations in chemical shrinkages were also understood. Furthermore, the effect of fly ash on the kinetics of Portland cement hydration at different curing temperatures has also been investigated by RQPA (Narmluk and Nawa 2011). The hydration reactions of OPC were quantified by RQPA and the overall degree of fly ash hydration was determined from a selective dissolution method. Ternary binders composed of OPC, calcium sulfoaluminate clinker (CSA) and anhydrite were examined in order to study the impact of variations of the OPC:CSA:C S ratio on the hydration process and related mortar properties. RQPA was used to determine the mineralogical composition of the starting cementitious materials. Thermodynamic modeling was used to establish the phase assemblage which was also studied by calorimetry, DTA-TGA, SEM and X-ray powder diffraction for phase identification (Pelletier et al. 2010). RQPA and thermal methods were used to determine the phase development up to 28 days of hydration in normal and ultra-high performance cementitious systems (UHPC) that contains silica fume and fly ash (Korpa et al. 2009). For the calculation of the ACn content, the vacuum dried powdered specimens were mixed with ZnO as internal standard. For both formulations the most remarkable changes of the phase contents were measured in the first few days of hydration. To finish this section we would like to highlight that many hydrated phases may be present in a given hydrated/hydrating sample. Figure 8 displays the Rietveld plot of a sample showing one of the richest phase assemblage found in our laboratory. The peaks are labeled to easily identify each crystalline phases. The crystalline phases of this hydrated cement are mainly arising from the calcium sulfoaluminate cement fraction but we choose this sample to illustrate the amount of phases that can coexist in a paste (Fig. 8, Table 3).

Durability studies Rietveld quantitative phase analysis may also help to understand/characterize the durability of the resulting mortars and concrete. Deterioration of cementitious building materials is often caused by sulfate attack at moderate temperatures due to delayed ettringite formation. Hence, several studies addressed this issue using RQPA (Katsioti et al. 2011). For instance, four cements were used to address the effect of tricalcium silicate content on external sulfate attack in sodium sulfate solution (Shanahan and Zayed 2007). Durability was studied by using linear expansion and compressive strength. Phases associated with deterioration were examined using scanning electron microscopy and RQPA. The resistance to sulfate attack of mixtures accelerated with alkali-free and alkaline accelerators was also studied by a number of techniques including RQPA which enabled the quantification of ettringite and gypsum over time (Paglia et al. 2003). On the other hand, thaumasite is mostly observed at low temperatures

191

C3AH6 AFm-C

AFt

C3AH6 CSH0.5 AFm-C

katoite CSH0.5 C3AH6 CC

CSH0.5

C3AH6

AFm-C

AFt AFm-C

AFt AH3

C3AH6 AH3

C AH katoite 3 6

CSH0.5 AFt

AFm-C

AFt

Rietveld Quantitative Phase Analysis of OPC

Figure 8. Rietveld plot (CuKa1) of a fully hydrated blended cement paste mainly composed of calcium sulfoaluminate cement. The main crystalline peaks are labeled.

(usually lower than 15 °C). Hence, the effect of ettringite on thaumasite formation was studied in synthetic OPC materials using this methodology (Kholer et al. 2006). It must be highlighted that ettringite and thaumasite form a solid solution that has been extensively studied (Barnett et al. 2002; Torres et al. 2004). Another durability concern in OPC concretes is the alkali-aggregate/alkali-silica reaction (see for instance: Thomas et al. 2006; and references therein). Some aggregates, mainly silica but also carbonate, may provoke the expansion with failure of OPC concretes. RQPA has been employed to study the aggregates with the final aim to understand these reactions (Grattan-Bellew et al. 2010). Furthermore, this technique has also been employed, plus other characterization tools, to study the alkali-aggregate reaction in OPC and waterglass-alkaliactivated slag mortars (Puertas et al. 2009). Finally, composition and microstructure changes of cement pastes under a heating and cooling cycle were monitored by neutron powder diffraction. The parameters involved in the study were the heating ramp, the state of the sample (in block or ground) and the type of cement. Unfortunately, the Rietveld method was not applied to quantify the phases in the mixtures (Castellote et al. 2004). Neutron powder diffraction was subsequently used for studying the phase composition changes of cement pastes during accelerated carbonation experiments (Castellote et al. 2008).

Selective dissolution Selective dissolution may be applied to OPC clinkers (or cements) or to blended cements with different methodologies (Gutteridge 1979; Luke and Glasser 1987). For OPC clinkers, both aluminates and silicates residues can be obtained. This is very useful for ensuring the polymorph present in the samples as the enriched phase may be identified much more easily. On the other hand, selective dissolution of blended cements is also used to determine the hydration degree of the addition (for instance fly ash or blast furnace slag). In these cases,

192

Aranda, De la Torre, León-Reina

special selective dissolution methodologies have been developed (see for instance Ben-Haha et al. 2010). In 2003, Lundgaard and Jons described the application of RQPA to the aluminate residues of grey Portland clinkers. Several NIST reference clinkers were analyzed. The salicylic acid/ methanol extraction method (SAM) was used to dissolve silicate phases (alite, belite) and free lime. So, the minor content phases, aluminate, ferrite and sulfates were enriched. Furthermore, residues obtained from the chemical treatment of three NIST reference materials RM8486, RM8487 and RM8488 were analyzed by SEM and RQPA. The main finding was that chemical treatment was not fully selective/quantitative (Pritula et al. 2004b). So, this methodology is appropriate for enriching the low content phases. However, selective dissolution and RQPA of the residues does not directly improve the accuracy of the analytical results. On the other hand, special selective dissolutions for studying the hydration of blended cements have already been discussed above (Hoshino et al. 2006; Termkhajornkit et al. 2006; Brunet et al. 2010).

INTERCOMPARISON AND COMPARISON WITH OTHER METHODS Before comparing the results of RQPA of cements with other techniques for quantitative mineralogical analyses, RQPA results of different laboratories should be compared. We highlight that this is possible for clinkers and cements but not for pastes as the evolving/ unstable nature of the samples do not allow to easily carry out an inter-laboratory study. Round robin (inter-laboratory) studies of RQPA were initially carried out for several types of samples but not OPCs (Toraya et al. 1999; Madsen et al. 2001; Scarlett et al. 2002). More recently, a partial round robin on RQPA of cement samples was published (Stutzman 2005; Stutzman and Leigh 2007). Unfortunately, the accuracy and uncertainty of the OPCs RQPA were not tested. Four cement reference specimens were prepared using NIST SRM clinkers compounded with known amounts of gypsum, bassanite, anhydrite, and/or calcite. The results of this study were used to estimate the inter- and intra-laboratory precision and bias of phase abundance determinations. Values of repeatability and reproducibility were given, but the statistical study was only based on the precision of the measurements/analyses. To evaluate the accuracy of the results, the use of these samples is not fully adequate since the “true” mineralogical compositions were not known. In a later study (Leon-Reina et al. 2009), a round-robin was conducted with two sets of samples, artificial mixtures and commercial OPCs. Artificial mixtures were prepared by mixing (weighing) synthesized single-crystalline phases in the appropriate proportions. These two samples were used to assess the accuracy and uncertainty of the procedure, as an expected mineralogical phase fraction—the “true mineralogical percentage”—is available under the assumption of negligible ACn contents. For a level of confidence of 95%, the general uncertainties were in the range 4.1-6.5% for C3S, 2.8-5.5% for C2S, 0.9-2.5% for C3A, 1.3-2.4% for C4AF, 1.0-1.6% for gypsum and 1.5-3.8% for calcite. The obtained precision values were much better. On the other hand, comparison of RQPA results can be carried out, with due care, with other analytical techniques. For clinkers and cements, point-counting microscopy techniques (optical and electronic) also directly measure the phase contents. In addition to these techniques, XRF directly measures the elemental compositions and, under some assumptions, a potential phase content can be established. For pastes, scanning electron microscopy, coupled to EDX microanalysis, is being utilized. However, a much straight forward method is the thermal decompositions, with their associated weight losses, as the decomposition temperature range may indicate the phase and the associate weight loss may allow measuring

Rietveld Quantitative Phase Analysis of OPC

193

its content. Furthermore, other techniques such as calorimetry and nuclear magnetic resonance spectroscopy are emerging as quantitative tools. All these techniques are briefly reviewed next.

Bogue and reverse Bogue calculation The most widely used method for estimating the potential phase composition of OPCs from the oxide analysis was developed long time ago (Bogue 1929; Taylor 1989) and generalized by the routine use of XRF analysis. An extensive work in this topic has been very recently reported/discussed (Le Saout et al. 2011) showing the associated errors to this methodology, so it is not further discussed here. We only want to highlight an excellent work (Crumbie et al. 2006) that employed four analytical techniques: RQPA, Bogue, optical microscopy and scanning electron microscopy coupled with energy dispersive spectroscopy (SEM-EDS); for the quantitative study of eight clinkers. That paper shows that the use of standard Bogue calculation to predict the phase composition of Portland clinkers can give serious errors.

Optical and scanning electron microscopies After the appropriate sample treatment, optical microscopy coupled with point counting techniques can produce very reliable phase quantification especially for the silicate phases, alite and belite (Campbell and Galehouse 1991; Taylor 1997; Campbell 1999; Fullmann and Walenta 2003). However, the quantification of the aluminate and the ferrite phases is often quite difficult, due to the very small crystal size of these interstitial phases. Furthermore, the different crystallographic forms of C3A (cubic, orthorhombic) cannot be differentiated (Walenta and Fullmann 2004). There is no need to emphasize that sample preparation, data acquisition, and data analysis are more demanding than for RQPA. It must be mentioned that the International Cement Microscopy Association http://www.cemmicro.org/ which began in 1981, develops several activities including the organization of annual meetings. The published proceedings serve as a working tool and source of informative references in the areas of clinker, cement, concrete, and other building materials. On the other hand, SEM-EDS studies are carried out to quantify the chemical (elemental) composition of selected phases within OPCs (see for instance: Gobbo et al. 2004; Crumbie et al. 2006). Both optical and electron microscopies are good complementary techniques to RQPA of clinkers and cements (Campbell and Galehouse 1991; Stutzman and Leigh 2002; Suherman et al. 2002; Stutzman 2011 and references therein). Other electron microscopy studies can be carried out. For instance, high-resolution cold field emission-scanning electron microscopy, in addition to isothermal conduction calorimetry and RQPA of the initial cements, was used to understand the end of the induction period of OPCs (Makar and Chan 2008). Finally, it should be kept in mind that microscopy techniques directly give volume fractions but Rietveld software usually gives weight fractions. The comparison is straightforward for anhydrous cements by using the crystallographic densities (already within the software calculations) to renormalize one of the results. However, this comparison is not straightforward for hydrating samples as the densities of some (amorphous) hydrates are not well known. The volume variation during hydration is necessary for understanding chemical shrinkages of pastes, mortars and concretes.

Thermodynamic modeling In this approach, the calculated hydration rates of the individual clinker phases are used as the (time-dependent) input under the relevant conditions once the appropriate database is developed (Matschei et al. 2007b). The modeled data can be compared with the measured composition of pore solutions as well as with any experimental quantitative phase analysis technique (Lothenbach and Winnefeld 2006). This approach was extended to variabletemperature hydration studies (Lothenbach et al. 2008a) and it has been very recently reviewed (Damidot et al. 2011).

194

Aranda, De la Torre, León-Reina

For instance, this methodology (in combination with RQPA and DTA-TGA, see below) was used for characterizing the role of limestone on the hydration of Portland cement (Lothenbach et al. 2008b). Another example is its use (in combination with X-ray powder diffraction) to establish the impact of chloride on the mineralogy of hydrated Portland cement systems (Balonis et al. 2010).

Thermal measurements Differential thermal analysis ‑ thermogravimetric analysis (DTA-TGA) measurements may be used to quantify phases in cements and pastes. To the best of our knowledge, there are no many examples of using DTA-TGA studies for corroborating the RQPA results in clinkers and cements. However, the use of DTA-TGA is straightforward as it allows determination of the overall calcite content from its decomposition temperature, between 700 and 800 °C, in OPCs. Furthermore, if portlandite has been formed by hydration of free lime, its overall content can also be determined from the DTA-TGA study, as its decomposition temperature is close to 450 °C. Finally, DTA-TGA also allows distinguishing between gypsum and bassanite under the appropriate experimental conditions (Blaine 1995), and this was used to verify the gypsum/bassanite ratio obtained from RQPA (Leon-Reina et al. 2009). DTA-TGA technique is widely used for studying cement pastes as it allows having an insight about the overall degree of reaction by measuring the overall water content (after subtraction of the free water). Furthermore, the evaluation of some phases may be carried out independently from their thermal decomposition temperatures. A key use is to determine the time-evolution of portlandite from RQPA and DTA-TGA measurements (Scrivener et al. 2004; Puertas et al. 2010). Furthermore, DTA-TGA and RQPA have been carried out to monitor the long-term hydration behavior of cement monoliths containing organic waste (Leoni et al. 2007), and to investigate long-term leaching in concretes (Marinoni et al. 2008).

Calorimetric data A new approach is being developed which consists in correlating the heat flow curves with the results of RQPA. This methodology allows understanding the chemical origin of different regions/features in the calorimetric curves (hydration reactions) and also supports the accuracy of RQPA (Hesse et al. 2011). This type of study was initially carried out for synthetic cement and later used to characterize the hydration of alite including the ACn content which was determined by the external standard, G-method, approach (Jansen et al. 2011b). Finally, the power of this methodology (including the G-method for measuring the ACn content) has been demonstrated by characterizing the early-age hydration reactions of an OPC cement by RQPA and isothermal calorimetry (Jansen et al. 2012a). Furthermore, the same group used this methodology to understand the changes in reaction kinetics of an OPC caused by a superplasticizer (Jansen et al. 2012b).

Nuclear Magnetic Resonance (NMR) spectroscopy NMR techniques have been increasingly employed in studies of cementitious materials mainly for understanding hydration reactions (Skibsted et al. 2002) including blended OPC cements (Dyson et al. 2007). Its full application to anhydrous OPCs samples has only been recently introduced (Poulsen et al. 2009). These authors used NMR techniques to determine the alite-to-belite Si-ratio which was compared to the alite-belite ratio determined by RQPA. Good agreement was found, although belite may be slightly underestimated under some circumstances. The (independent) use of RQPA and NMR methods for characterizing anhydrous OPCs and the hydration reaction products has been reviewed (Skibsted and Hall 2008; Stark 2011). It should be noted that the combined use of these two techniques for quantitative analysis is not frequently used. We can highlight that 29Si and 27Al solid-state NMR spectroscopy with

Rietveld Quantitative Phase Analysis of OPC

195

complementary X-ray diffraction analysis and selective dissolution techniques have been used to study the hydration of a CEM V blended cement containing pulverized fly ash and blast furnace slag in order to understand hydration processes which influence the paste microstructure (Brunet et al. 2010).

GUIDELINES FOR RIETVELD QUANTITATIVE PHASE ANALYSES There are reported guidelines for Rietveld studies that should be known (McCusker et al. 1999). This section gives a methodology to carry out a RQPA. However, it is not a universal guideline, not being the optimum way for all cases. As the analyses are becoming more and more complex (for instance hydration reactions of blended cements), extreme care has to be exercised to carry out the best possible analysis. So, it is advisable to learn the methodology in a well-trained laboratory with previous experience in the system or at least in related-systems. Some general rules are given in the next subsections that should be cautiously taken when applied to the specific problem.

Crystal structures Tables 1-3 give a selected compilation of crystal structures that can be used in the Rietveld control file. Many crystal structures have been already used in our laboratory but not all of them. Furthermore, sometimes there is more than one structural description available in the literature for a given phase. The methodology that we use in our laboratory to select a phase is to carry out two (or more) RQPA with the two (or more) reported structural descriptions. We select the structure that gives the best fit. This is easier to say than to fulfill. As a general rule, our approach is to select the phase which gives a lower RF factor (see McCusker et al. 1999; Toby 2006) hopefully linked to a higher quantification value. Additionally, cross-checking with other method(s) to determine the chemical compositions of the modifications is appropriate, but it is time-demanding and expensive. If needed, transmission electron microscopy coupled with chemical analysis may be used to measure average elemental substitutions in some phases. Tables 1-3 are quite exhaustive but more phases may be present in a given powder pattern. It is not possible to deal here with every phase that may be present in a Portland cement. For instance, the authors found ZnO in a clinker produced with used tyres which were employed as alternative fuel. Therefore, the methodology to finish the RQPA of an OPC material is to identify the phase(s) responsible for any additional diffraction peak in the powder pattern, if needed. This is usually carried out by comparing the extra diffraction peaks with those reported in the Powder Diffraction Database (www.icdd.com). However, a good chemical knowledge of the system is very important as the powder patterns may be very complex and the identification may not be conclusive. The second step is to find the structural description for that phase in order to be inserted in the Rietveld control file. One can check in the bibliography and on the internet. The fastest way is to search that phase in three structural data bases: i) AMCSD “American Mineralogist Crystal Structure Database” (http://rruff.geo.arizona.edu/AMS/amcsd. php); ii) COD “Crystallography Open Database” (http://www.crystallography.net); and iii) ICSD “Inorganic Crystal Structure Database” (http://www.fiz-karlsruhe.de/icsd.html). If more than one phase may explain a set of extra peaks, then the approach described in the previous paragraph applies. Finally, the crystal structure to be used in the quantification has sometimes to be adapted to describe a given stoichiometry. This is illustrated with the Katoite structure which is a solid solution, Ca3Al2(OH)12−4x(SiO4)x (0 ≤ x ≤ 1.5), that may go from the end member hydrogarnet, Ca3Al2(OH)12 (x = 0), up to Ca3Al2(OH)6(SiO4)1.5. Another examples is brownmillerite, Ca4(Fe4−xAlx)O10, where the iron/aluminum ratio is variable (Redhammer et al. 2004) but close to x = 2 in Portland cements. In addition to the variable iron content, for a fixed Fe/

196

Aranda, De la Torre, León-Reina

Al ratio, the occupations of Al and Fe on tetrahedral and octahedral sites influence the peak intensities (Neubauer et al. 1996; Neubauer and Mayerhofer 2000). It is known that very fast cooling leads to lower intensity of the (020) peak of C4AF (located at 12.2° 2q CuKa) due to a random distribution of Al and Fe over the two sites, when compared to the main reflection (141) located at 33.7° 2q CuKa. Conversely, slower coolings allow iron to locate preferably on the octahedral site, and so aluminum would mainly located at the tetrahedral site. This metal distribution results in higher intensity of the (020) peak when compared to that of (141). Additionally, other transition metals can be incorporated into the brownmillerite structure and its role on the diffraction intensities may be significant (Zotzl and Pollmann 2006).

Sample preparation and data collection This subsection is dedicated to two (related) issues. Sample preparation will depend on the data collection strategy as it is not the same to prepare the sample for a q/2q (reflection) experiment that for a transmission measurement. In general, transmission measurements yield more reliable results since more particles are tested. However, the diffracted intensity may be low due to absorption. Furthermore, absorption varies with the diffracted angle and this must be corrected (which is not the case for reflection geometry). For reflection geometry experiments, the sample surface should be as flat as possible to avoid surface roughness problems in the pattern. However, strong pressing of the sample is not advised as this may lead to preferred orientation. This may be corrected through the March-Dollase ellipsoidal correction (Dollase 1986) but in any case it decreases the amount of information present in a powder pattern. So, preferred orientation should be minimized during sample preparation and corrected during the data analysis stage (to be discussed below). Flat-samples should be rotated during data collection to enhance particle statistics. Reactive samples should be covered with a film (usually Mylar or Kapton) to avoid reaction (for instance carbonation of portlandite) during data collection. Finally, it must be kept in mind that most interiors of X-ray powder diffractometers have very low humidity to avoid/minimize external corrosion of the X-ray tube (for instance we use P2O5). Thus, the surfaces of early-age hydration pastes may be (partially) dehydrated before/during the measurement when kept within the diffractometer. So, we advise to collect the data on pastes as soon as possible or alternatively it is much better to use a reaction/hydration chamber. For transmission geometry experiments, absorption correction should be carried out for flat samples and capillaries. Additionally, counting time should be large enough to have good statistics in the intensities. Furthermore, the size of the samples (thickness of the flat sample or diameter of the capillary) should be small to have good resolution in the resulting powder pattern. However, very small samples may result in diffraction peak intensity problems as the amount of powder tested by the X-ray may be not fully representative of the sample. Our experience shows that capillaries of 0.3 mm of diameter (or larger) give good patterns. However, diluted samples in capillaries of 0.3 mm may start to display problems. Similarly, transmission data for flat samples with thicknesses of the order of 0.2-0.3 mm yield good RQPA results. Finally, some other issues deserve a brief comment. The data collection range for OPC samples is usually 8-70° (2q) for CuKa radiation. The high angle limit is defined by the strong overlapping in the powder pattern and diffraction data above 60° (2q) add very little extra information to the analysis. However, the low angle region is important mainly in cement analysis as the peaks of C4AF, gypsum and bassanite should be collected. It is also important to use an optic setup that ensures the X-ray beam is fully within the sample at 10° (2q). On the other hand it is advisable to have the highest resolution data. This may allow distinguishing between related phases and it also minimizes correlations during the refinement stage. So, strictly monochromatic laboratory data are the best option for a good in-house RQPA. A good alignment of the diffractometer is a pre-requisite and the sample should be mounted exactly on the diffractometer axis to avoid problems in the recorded pattern due to sample displacement.

Rietveld Quantitative Phase Analysis of OPC

197

Data analysis Once the best possible powder pattern has been recorded, Rietveld analysis must be executed with the appropriate program and the control file. Let’s us assume that all crystalline phases have structural descriptions and they have been correctly inserted within the control file. This last point is not a trivial statement as sometimes temperature factors are not correctly imported when using automatic structure insertion procedures. Hence, it is a very good practice to check that all structures within the control file have been properly inserted. Furthermore, some space groups have two possible origins but some Rietveld programs only allow the use of one setting. The Rietveld procedure consists in optimizing several parameters in order to get the best possible agreement between the experimental and the calculated patterns. As a general rule, the structural parameters (atomic positional parameters, atomic temperature factors—also known as atomic displacement parameters since they may have a spatial disorder contribution—and occupation factors) are not generally optimized but they are kept fixed to the reported values. However, the overall parameters (those that affect to the full pattern) are usually optimized. These overall parameters are grouped in sets: i) phase scale factors; ii) background coefficients; iii) zero-shift error or sample displacement parameter; iv) phase unit cell parameters; v) phase peak shape parameters; vi) phase preferred orientation, if needed; vii) sample absorption coefficient (only for capillary or flat-sample transmission data, but the equation correction may depend on the geometry used). The order in which the parameters should be optimized is always a concern. As a rule of thumb the order should be to start with the parameters which are quite far from the final values, and so proceed adding new parameters which are not that far. We next suggest a sequence which does not need to be always the correct one. Firstly, the phase scale factors and the background coefficients. Secondly, the sample displacement coefficient and the peak shape parameters for the high-content phases. Then, the preferred orientation correction for the phases that display this effect. So, we may continue with the unit cell parameters. Finally, the phase peak shape parameters for low-content phases. Unfortunately, correlations may be very high for complex refinements and not all parameters may be optimized. Then several tricks may be applied which are somewhat arbitrary but, if they are always applied in the same way, they may give consistent results that allow to establish trends, etc. Now, we must highlight that the overlapping of the main diffraction peaks is a key issue. So, we initially define high-, intermediate- and low-content phases as those present in the sample at contents higher than 20 wt%, between 5-20 wt% and lower than 5 wt%. However, these “labels” will certainly depend upon the degree of peak overlapping. For instance, a phase present in a sample at 2 wt% but with narrow peaks and little overlapping may be considered as intermediate-content phase, and conversely, a phase at 9 wt% level but with very strong overlapping may be considered as a low-content one. For high-content phases every parameter may be refined included at least two coefficients for the peak shape (both Gaussian and Lorentzian contributions when using a pseudo-Voigt peak shape function). For intermediate-content phase, dumping (cushioning) values should be quite high. Furthermore, overlapping may not allow optimizing two peak shape parameters. Then one value should be kept fix to a reasonable value and just refine one (and check that it does not get a very large, unreasonable, value). Finally, unit cell parameters and peak shape coefficients may be quite difficult to optimize for low-content phases. They should be refined with extreme care or kept fixed to goods values. For instance, they may be kept to the refined values in selective dissolution residues where their contents were much higher. If these residues are not available, the experience in previous fits uses to be a very good asset. Preferred orientation (a partial microparticle aligned arrangement) deserves a final comment. Working in reflection geometry, sample rotation increases the particle statistics

198

Aranda, De la Torre, León-Reina

but it does not minimize preferred orientation. This effect is shown up in the Rietveld fit of the powder pattern as some peaks having smaller calculated intensities meanwhile other reflections having larger calculated intensities. In order to correct this effect with the MarchDollase algorithm, the preferred orientation axis (corresponding to the cleavage plane or a main growth axis) must be inserted in the Rietveld control file. We next give the axes for some phases showing this effect. However, it must be highlighted that preferred orientation strongly depends on the microparticle habits that may vary according to the synthetic conditions. The axes for some phases are: gypsum [010]; portlandite [001]; AFm [001]; calcite [104]; alite-M3 [ 1 01]; and ettringite [100].

Final check After a RQPA is carried out two final checks are advisable. The first is to ensure that the refined parameters have not converged to unreasonable values. This may be avoided if a higher and lower limit is set up for each adjustable parameter. This is the strategy for on-line fully automatic RQPA in many cement plants where robustness is ensured even although some parameters are not allowed to vary totally free. Phase peak shape parameters are the most critical variables and they should be kept under permanent surveillance. The final values of the unit cell parameters for low-content phases should also be examined with care. Secondly, it is very important to compare the RQPA results with others obtained from complementary techniques, if possible. A previous section was devoted to this issue, so it is not further treated here. However, a lot of confidence is gained when RQPA results are successfully compared with other analytical results (for instance portlandite content determined from thermal analysis).

FINAL REMARKS AND OUTLOOK Rietveld quantitative phase analysis is being increasingly used for characterizing OPC materials. Initially, it was mainly devoted to measure anhydrous cements but it is nowadays being used a lot for analyzing hydration products. It is clear that RQPA will be used for studying the hydration reactions of many blended cements. The number of residues, by-products, etc. (with different chemical and mineralogical compositions) that can be blended with Portland cements is quite large, and expanding. In addition, new types of eco-cements, those releasing less CO2 in their fabrication processes, are being deeply investigated. Here, RQPA has a bright future as most of the knowhow developed for characterizing OPC samples can be directly exported to analyze these new cements (calcium sulfoaluminate, sufobelite, alkali-activated fly ash or slag, etc.). In addition to the (new) cements to be studied, some theoretical/methodological advances are expected. Firstly, some advances concerning the crystal structures will be very welcomed: i) Reporting new crystal structures as some phases have not reported structures that must be developed for RQPA (e.g., the crystal structure of C2AH8 is lacking); ii) Obtaining better structural descriptions for some phases, which have approximate crystal structures, in order to have more accurate phase analyses (e.g., a good structural description at room temperature for a-C2S is lacking); iii) when several structural descriptions for one phase are available, some (ideally inter-laboratory) comparative studies should be carried out to select the most adequate crystal structure to carry out the RQPA. Another field where progress is expected is increasing the reliability of the diffraction data. This improvement can be obtained at least in two ways. One approach is to carry out a better microparticle sampling by using higher-energy Mo-radiation. To the best of our knowledge, there is not a single reported RQPA of OPC material with Mo-radiation. This is likely due to the lower resolution from existing powder diffractometers. Figure 9 shows the Rietveld fits of

Rietveld Quantitative Phase Analysis of OPC

0.5

X10E 4 1.0

CuK1

Counts -0.5

0.0

1.5406Å

X10E 4 1.0

199

29.0

30.0

31.0

32.0

33.0

34.0

35.0

Counts -0.5

0.0

0.5

2-Theta, deg

10.0

20.0

30.0

40.0

50.0

60.0

70.0

2-Theta, deg

3.0

4.0

X10E 4

5.0

MoK1,2

2.0 1.0 0.0 Counts

4.0

X10E 4

0.7093Å 0.7136Å

14.0

15.0

16.0

Counts

0.0

2.0

13.0 2-Theta, deg

5.0

10.0

15.0

20.0

25.0

30.0

2-Theta, deg

Figure 9. (top) Full region of a Rietveld plot of high-resolution CuKa1 laboratory data in reflection geometry for a commercial OPC grey clinker with 6 phases, from bottom to top: belite, alite, aluminate, ferrite, aphthitalite and periclase. (bottom) Full region of a Rietveld plot of medium-resolution MoKa1,2 laboratory data in reflection geometry for the same clinker. The insets show enlarged views of the common 2q range, 29-35° for the Cu-pattern and 12.9-16.2° for the Mo-pattern.

a grey OPC clinker where the phase analysis has been carried out fitting simultaneously a high resolution CuKa1 pattern collected in reflection geometry and a medium-resolution MoKa1,2 pattern also collected in reflection geometry. The phase analysis obtained fitting both patterns simultaneously was: C3S-M3, 69.6(7) wt%; b-C2S, 15.3(1) wt%; C3A-cubic, 4.4(2) wt%; C4AF, 9.0(1) wt%; MgO, 0.86(1) wt%; and K3Na(SO4)2, 0.84(2) wt%. The large penetration depth of Mo-Ka radiation has two consequences on the alite quantification. Firstly, many more particles are analyzed when irradiated with high energy wavelengths 0.7093 and 0.7136 Å. So, alite peaks are nicely fitted in the Mo-Ka pattern, meanwhile the fit is much worse for the Cu-Ka pattern (due to a poorer particle statistics). Furthermore, preferred orientation (along [ 1 01] axis of alite-M3) is also less important in the Mo-Ka pattern. The March-Dollase

Aranda, De la Torre, León-Reina

200

optimized coefficients were 0.855(3) and 0.961(2) for the Cu-Ka1 and Mo-Ka1,2 patterns, respectively. However, the resolution of the Mo-pattern is not very good (see inset of Fig. 9, bottom) and better equipment is expected in the near future, including the key development of a primary monochromator for laboratory Mo-Ka radiation. A second approach is to increase the accuracy of the RQPA results by combining at least two data sets. For instance, it is possible to carry out a RQPA for a single sample but from two data sets, e.g., one high-resolution powder data collected in reflection and a second pattern collected in transmission in order to have a better powder averaging although the resolution of the diffraction peaks may be lower. Figure 10 shows the Rietveld fits of calcium

y

1.5

Reflection

X10E 4

2.0

CuK1

2.0

Counts

0.0

0.5

3.0

1.0

CSH2

X10E 4

4.0

y

10.0

15.0

20.0

25.0

30.0

35.0

Counts

0.0

1.0

2-Theta, deg

10.0

20.0

30.0

40.0

2-Theta, deg

60.0

70.0

X10E 3

CuK1,2

Counts

0.0

1.0

2.0

2.0

Transmission

X10E 3

50.0 y

15.0

20.0

25.0

30.0

35.0

Counts

0.0

1.0

CSH2

10.0 2-Theta, deg

10.0

20.0

30.0

40.0

50.0

60.0

70.0

2-Theta, deg

Figure 10. (top) Full region of a Rietveld plot of high-resolution CuKa1 laboratory data collected in reflection geometry for a calcium sulfoaluminate cement containing nine crystalline phases. (bottom) Full region of a Rietveld plot of medium-resolution CuKa1,2 laboratory data collected in transmission geometry for the same calcium sulfoaluminate. The insets show enlarged views of the low-angle regions to highlight the effects of the gypsum preferred orientation.

Rietveld Quantitative Phase Analysis of OPC

201

sulfoaluminate cement where the phase analysis has been carried out fitting simultaneously a high resolution pattern collected in reflection and a medium-resolution pattern collected in transmission. We must highlight that both fits are carried out with a final gypsum optimized content of 13.5(1) wt% but the March-Dollase preferred orientation correction coefficients were 0.499(8) and 1.37(2) for the reflection and transmission patterns, respectively. This cement has a very complex phase assemblage formed by: cubic-Yeelimite, Ca4Al6O12SO4, 23.6(2) wt%; orthorhombic-Yeelimite, 15.9(2) wt%; ternesite, Ca5(SiO4)2SO4, 16.7(3) wt%; gypsum, 13.5(1) wt%; b-belite, 9.9(1) wt%; anhydrite-II, 8.3(1) wt%; alite-M3, 6.1(1) wt%; calcium titanium perovkskite, CaTiO3, 4.7(1) wt%; and dolomite, (Mg,Ca)CO3, 1.2(1) wt%. RQPA results obtained from the simultaneous two-histogram refinement is more accurate that using just one single dataset. Finally, we expect to see a large increase of the use of the external standard method in hydration products analyses. This approach which still is not very common has two clear advantages over the internal standard (spiking) method as i) no dilution of low content phases takes place, and ii) the (possible) interference of the standard with the hydration reactions is avoided.

ACKNOWLEDGMENTS This work has been supported by Spanish Ministry of Science and Innovation through the MAT2010-16213 research grant which is co-funded by FEDER. We also thank Junta de Andalucía (Spain) for funding through the FQM-113 project.

REFERENCES Allmann R (1977) Refinement of the hybrid layer structure (Ca2Al(OH)6)+(0.5SO4·3H2O). Neues Jahrb Mineral Monatsh 136-144 Ballirano P, Belardi G, Maras A (2005) Refinement of the structure of synthetic syngenite K2Ca(SO4)2(H2O) from X-ray powder diffraction data. Neues Jahrb Mineral Abh 182:15-21 Balonis M, Glasser FP (2009) The density of cement phases. Cem Concr Res 39:733-739 Balonis M, Lothenbach B, Le Saout G, Glasser FP (2010) Impact of chloride on the mineralogy of hydrated Portland cement systems. Cem Concr Res 40:1009-1022 Balzar D, Audebrand N, Daymond MR, Fitch A, Hewat A, Langford JI, Le Bail A, Louer D, Masson O, McCowan CN, Popa NC, Stephens PW, Toby BH (2004) Size-strain line-broadening analysis of the ceria round-robin sample. J Appl Crystallogr 37:911-924 Barnett SJ, Macphee DE, Lachowski EE, Crammond NJ (2002) XRD, EDX and IR analysis of solid solutions between thaumasite and ettringite. Cem Concr Res 32:719-730 Bellmann F, Damidot D, Moser B, Skibsted J (2010) Improved evidence for the existence of an intermediate phase during hydration of tricalcium silicate. Cem Concr Res 40:875-884 Ben-Haha M, De Weerdt K, Lothenbach B (2010) Quantification of the degree of reaction of fly ash. Cem Concr Res 40:1620-1629 Bentz DP (2008) A review of early-age properties of cement-based materials. Cem Concr Res 38:196-204 Bezou C, Nonat A, Mutin JC, Christensen AN, Lehmann MS (1995) Investigation of the crystal structure of gamma-CaSO4, CaSO4·0.5(H2O) and CaSO4·0.6(H2O) by powder diffraction methods. J Solid State Chem 117:165-176 Bish DL, Howard SA (1988) Quantitative phase analysis using the Rietveld method. J Appl Crystallogr 21:8691 Bish DL, Post JE (1993) Quantitative mineralogical analysis using the Rietveld full-pattern fitting method. Am Mineral 78:932-940 Bish DL, Post JE (eds) (1989) Modern Powder Diffraction. Reviews in Mineralogy. Vol 20. Mineralogical Society of America, Washington Bish DL, Reynolds RC Jr (1989) Sample preparation for X-ray diffraction. Rev Mineral 20:73-99 Blaine R (1995) Determination of calcium-sulfate hydrates in building-materials using thermal-analysis. Am Lab 27:24-28 Bogue RH (1929) Calculation of the compounds in Portland cement. Ind Eng Chem 1:192-197

202

Aranda, De la Torre, León-Reina

Bonaccorsi E, Merlino S, Kampf AR (2005) The crystal structure of Tobermorite 14 Å (Plombierite), a C-S-H phase. J Am Ceram Soc 88:505-512 Bonaccorsi E, Merlino S, Taylor HFW (2004) The crystal structure of jennite Ca9Si6O18(OH)6·8H2O. Cem Concr Res 34:1481-1488 Brindley GW (1945) The effect of grain or particle size on X-ray reflections from mixed powders and alloys, considered in relation to the quantitative determination of crystalline substances by X-ray methods. Philos Mag 36:347-369 Brunet F, Charpentier T, Chao CN, Peycelon H, Nonat A (2010) Characterization by solid-state NMR and selective dissolution techniques of anhydrous and hydrated CEM V cement pastes. Cem Concr Res 40:208-219 Buhrke VE, Jenkins R, Smith DK (1998) A Practical Guide for the Preparation of Specimens for X-Ray Fluorescence and X-Ray Diffraction Analysis. Wiley-VCH, New York Calos NJ, Kennard CHL, Whittaker AK, Davis RL (1995) Structure of calcium aluminate sulfate Ca4Al6O16S. J Solid State Chem 119:1-7 Campbell DH (1999) Microscopical Examination and Interpretation of Portland Cement and Clinker. Second Edition. Portland Cement Association. Skokie Campbell DH, Galehouse JS (1991) Quantitative clinker microscopy with the light microscope. Cem Concr Aggr 13:94-96 Castellote M, Alonso C, Andrade C, Turrillas X, Campo J (2004) Composition and microstructural changes of cement pastes upon heating, as studied by neutron diffraction. Cem Concr Res 34:1633-1644 Castellote M, Andrade C, Turrillas X, Campo J, Cuello G (2008) Accelerated carbonation of cement pastes in situ monitored by neutron diffraction. Cem Concr Res 38:1365-1373 Chen IA, Juenger MCG (2009a) Incorporation of waste materials into Portland cement clinker synthesized from natural raw materials. J Mater Sci 44:2617-2627 Chen IA, Juenger MCG (2009b) Incorporation of waste materials into Portland cement clinker synthesized from reagent-grade chemicals. Int J Appl Ceram Technol 6:270-278 Cheung J, Jeknavorian A, Roberts L, Silva D (2011) Impact of admixtures on the hydration kinetics of Portland cement. Cem Concr Res 41:1289-1309 Christensen AN, Olesen M, Cerenius Y, Jensen TR (2008) Formation and transformation of five different phases in the CaSO4-H2O system: crystal structure of the subhydrate b-CaSO4·0.5H2O and soluble anhydrite CaSO4. Chem Mater 20:2124-2132 Cline JP, Von Dree RB, Winburn R, Stephens PW, Filliben JJ (2011) Addressing the amorphous content issue in quantitative phase analysis: the certification of NIST standard reference material 676a. Acta Crystallogr Sect A 67:357-367 Colville AA, Géller S (1971) The crystal structure of brownmillerite, Ca2FeAlO5. Acta Crystallogr B27:23112315 Crumbie A, Walenta G, Fuellmann T (2006) Where is the iron? Clinker microanalysis with XRD Rietveld, optical microscopy/point counting, Bogue and SEM-EDS techniques. Cem Concr Res 36:1542-1547 Cullity BD (1978) Elements of X-Ray Diffraction. Addison-Wesley, Reading, Massachusetts Dalconi MC, Favero M, Aritoli G (2011) In-situ XRPD of hydrating cement with lab instrument: reflection vs. transmission measurements. Z Kristallogr Proc 1:155-161 Damidot D, Lothenbach B, Herfort D, Glasser FP (2011) Thermodynamics and cement science. Cem Concr Res 41:679-695 David WIF, Shankland K (2008) Structure determination from powder diffraction data. Acta Crystallogr 64:5264 David WIF, Shankland K, McCusker LB, Baerlocher Ch (2002) Structure Determination from Powder Diffraction Data. International Union of Crystallography Monographs on Crystallography. Oxford University Press, Oxford De la Torre AG, Alonso-Llorente H, Aranda MAG, León-Reina L (2011) Últimas novedades en la aplicación del método de Rietveld en el control de calidad de cementos. Cemento Hormigón 944:16-27 De la Torre AG, Aranda MAG (2003) Accuracy in Rietveld quantitative phase analysis of Portland cements. J Appl Crystallogr 36:1169-1176 De la Torre AG, Bruque S, Aranda MAG (2001a) Rietveld quantitative amorphous content analysis. J Appl Crystallogr 34:196-202 De la Torre AG, Bruque S, Aranda MAG (2003) Cuantificación mineralógica directa de cementos Portland por el método de Rietveld. Cemento Hormigón 850:16-27 De la Torre AG, Bruque S, Campo J, Aranda MAG (2002) The superstructure of C3S from synchrotron and neutron powder diffraction and its role in quantitative phase analyses. Cem Concr Res 32:1347-1356 De la Torre AG, Cabeza A, Calvente A, Bruque S, Aranda MAG (2001b) Full phase analysis of Portland clinker by penetrating synchrotron powder diffraction. Anal Chem 73:151-156

Rietveld Quantitative Phase Analysis of OPC

203

De la Torre AG, Cabeza A, Losilla ER, Aranda MAG (2006) Quantitative phase analysis of ordinary Portland cements using synchrotron radiation powder diffraction. Z Kristallogr Sup 23:587-592 De la Torre AG, de Vera RN, Cuberos AJM, Aranda MAG (2008) Crystal structure of low-magnesium content alite: application to Rietveld quantitative phase analysis. Cem Concr Res 38:1261-1269 De la Torre AG, Lopez-Olmo MG, Alvarez-Rua C, Garcia-Granda S, Aranda MAG (2004) Structure and microstructure of gypsum and its relevance to Rietveld quantitative phase analyses. Powder Diffr 19:240246 De la Torre AG, Losilla ER, Cabeza A, Aranda MAG (2005) High-resolution synchrotron powder diffraction analysis of ordinary Portland cements: Phase coexistence of alite. Nucl Instrum Methods Phys ResB 238:87-91 De Weerdt K, Haha MB, Le Saout G, Kjellsen KO, Justnes H, Lothenbach B (2011) Hydration mechanisms of ternary Portland cements containing limestone powder and fly ash. Cem Concr Res 41:279-291 Dinnebier RE, Billinge SJL (eds) (2008) Powder Diffraction Theory and Practice. Royal Society of Chemistry Publishing, Cambridge Dollase WA (1986) Correction of intensities for preferred orientation in powder diffractometry: application of the March model. J Appl Crystallogr 19:267-272 Dyson HM, Richardson IG, Brough AR (2007) A combined 29Si MAS NMR and selective dissolution technique for the quantitative evaluation of hydrated blast furnace slag cement blends. J Am Ceram Soc 90:598-602 Effenberger H, Kirfel A, Will G (1983a) Studies of the electron-density distribution of dolomite, CaMg(CO3)2. Tscher Mineral Petrogr Mitteil 31:151-164 Effenberger H, Kirfel A, Will G, Zobetz E (1983b) A further refinement of the crystal structure of thaumasite, Ca3Si(OH)6(CO3)(SO4)(H2O)12. Neues Jb Mineral Monatsh 60-68 Elton NJ, Salt PD (1996) Particle statistics in quantitative X-ray diffractometry. Powder Diffr 11:218-229 Enders M (2005) Sample preparation for quantitative X-ray diffraction in cement plants: sources of errors and solutions. ZKG Int 58:28-37 Enders M, Berger M (2007) Quantitative XRD (Rietveld-method) in cement plants: quality control of clinker production. ZKG Int 5:50-59 Ferreira AV, Righi A, Araujo FGS, Espinosa DCR, Tenório JAS (2008) Applications of the Rietveld method to quantify the crystalline phases of Portland cement clinker doped with nickel and chromium. Powder Diffr Suppl 23:S42-S45 Ferro O, Galli E, Papp G, Quartieri S, Szakall S, Vezzalini G (2003) A new occurrence of katoite and reexamination of the hydrogrossular group. Eur J Mineral 15:419-426 Fitzpatrick ME, Lodini A (eds) (2003) Analysis of Residual Stress by Diffraction using Neutron and Synchrotron Radiation. Taylor & Francis, London Fleet ME (1984) The structure of magnetite: two annealed natural magnetites, Fe3.005O4 and Fe2.96Mg0.04O4. Acta Crystallogr C40:1491-1493 Floerke OW (1952) Kristallographische und rontgenometrische Untersuchungen im System CaSO4 CaSO4·2H2O. Neues Jahrb Miner Abh 84:189-240 Francois M, Renaudin G, Evrard O (1998) A cementitious compound with composition 3CaOAl2O3CaCO311H2O. Acta Crystallogr C 54:1214-1217 Fullmann T, Walenta G (2003) The quantitative Rietveld phase analysis in industrial applications. ZKG Int 56:45-53 Gemmi M, Merlini M, Cruciani G, Artioli G (2007) Non-ideality and defectivity of the akermanite - gehlenite solid solution: an X-ray diffraction and TEM study. Am Mineral 92:1685-1694 Gineys N, Aouad G, Sorrentino F, Damidot D (2011) Incorporation of trace elements in Portland cement clinker: Thresholds limits for Cu, Ni, Sn or Zn. Cem Concr Res 41:1177-1184 Gobbo L, Sant’Agostino L, Garcez L (2004) C3A polymorphs related to industrial clinker alkalies content. Cem Concr Res 34:657-664 Goetz-Neunhoeffer F, Neubauer J (2006) Refined ettringite (Ca6Al2(SO4)3(OH)12·26(H2O)) structure for quantitative X-ray diffraction analysis. Powder Diffr 21:4-11 Golovastikov NI, Matveeva RG, Belov NV (1975) Crystal structure of the tricalcium silicate (CaOSi2)3 = C3S. Kristallografiya 20:721-729 Gonçalves JP, Tavares LM, Toledo-Filho RD, Fairbairn EMR (2009) Performance evaluation of cement mortars modified with metakaolin or ground brick. Const Build Mater 23:1971-1979 Grattan-Bellew PE, Mitchell LD, Margeson J, Minb D (2010) Is alkali-carbonate reaction just a variant of alkali-silica reaction ACR = ASR? Cem Concr Res 40:556-562 Gutteridge WA (1979) On the dissolution of the interstitial phases in Portland cement. Cem Concr Res 9:319324 Gutteridge WA (1984) Quantitative X-ray powder diffraction in the study of some cementive materials. Brit Ceram Proc 35:11-23

204

Aranda, De la Torre, León-Reina

Herfort D, Moir GK, Johansen V, Sorrentino F, BolioArceo H (2010) The chemistry of Portland cement clinker. Adv Cem Res 22:187-194 Hesse C, Degenkolb M, Gaeberlein P, Goetz-Neunhoeffer F, Neubauer J, Schwarz V (2008) Investigations into the influence of temperature and w/c ratio on the early hydration of white cement. Cem Int 6:68-78 Hesse C, Goetz-Neunhoeffer F, Neubauer J (2011) A new approach in quantitative in-situ XRD of cement pastes: Correlation of heat flow curves with early hydration reactions. Cem Concr Res 41:123-128 Hesse C, Goetz-Neunhoeffer F, Neubauer J, Braeu M, Gaeberlein P (2009) Quantitative in situ X-ray diffraction analysis of early hydration of Portland cement at defined temperatures. Powder Diffr 24:112-115 Hesse KF (1984) Refinement of the crystal structure of wollastonite-2M (parawollastonite). Zeit Kristallogr 168:93-98 Hill RJ, Howard CJ (1987) Quantitative phase analysis from neutron powder diffraction data using the Rietveld method. J Appl Crystallogr 20:467-474 Hong H, Fu Z, Min X (2001) Effect of cooling performance on the mineralogical character of Portland cement clinker. Cem Concr Res 31:287-290 Hoshino S, Yamada K, Hirao H (2006) XRD/Rietveld Analysis of the Hydration and Strength Development of Slag and Limestone Blended Cement. J Adv Concr Tech 4:357-367 Husillos-Rodriguez N, Martinez-Ramirez S, Blanco-Varela MT, Guillem M, Puig J, Larrotcha E, Flores J (2010) Re-use of drinking water treatment plant (DWTP) sludge: Characterization and technological behaviour of cement mortars with atomized sludge additions. Cem Concr Res 40:778-786 Irran E, Tillmanns E, Hentschel G (1997) Ternesite, Ca5(SiO4)2SO4, a new mineral from the Ettringer Bellerberg/Eifel, Germany. Mineral Petrol 60:121-132 Jansen D, Bergold ST, Goetz-Neunhoeffer F, Neubauer J (2011b) The hydration of alite: a time-resolved quantitative X-ray diffraction approach using the G-factor method compared with heat release. J Appl Crystallogr 44:895-901 Jansen D, Goetz-Neunhoeffer F, Lothenbach B, Neubauer J (2012a) The early hydration of Ordinary Portland Cement (OPC): An approach comparing measured heat flow with calculated heat flow from QXRD. Cem Concr Res 42:134-138 Jansen D, Neubauer J, Goetz-Neunhoeffer F, Haerzschel R, Hergeth WD (2012b) Change in reaction kinetics of a Portland cement caused by a superplasticizer — Calculation of heat flow curves from XRD data. Cem Concr Res 42:327–332 Jansen D, Stabler CH, Goetz-Neunhoeffer F, Dittrich S, Neubauer J (2011a) Does Ordinary Portland Cement contain amorphous phase? A quantitative study using an external standard method. Powder Diffr 26:3138 Jenkins R, Fawcett TG, Smith DK, Visser JW, Morris MC, Frevel LK (1986) JCPDS-international centre for diffraction data sample preparation methods in X-Ray Powder Diffraction. Powder Diffr 1:51-63 Jenkins R, Snyder RL (1996) Introduction to X-Ray Powder Diffractometry. Wiley-Interscience, New York Jupe AC, Wilkinson AP, Luke K, Funkhouser GP (2007) Slurry consistency and in situ synchrotron x-ray diffraction during the early hydration of portland cements with calcium chloride. J Am Ceram Soc 90:2595-2602 Kaduk JA (2007a) Chemical reasonableness in Rietveld analysis: organics. Powder Diffr 22:74-82 Kaduk JA (2007b) Chemical reasonableness in Rietveld analysis: inorganics. Powder Diffr 22:268-278 Katsioti M, Patsikas N, Pipilikaki P, Katsiotis N, Mikedi K, Chaniotakis M (2011) Delayed ettringite formation (DEF) in mortars of white cement. Constr Build Mater 25:900-905 Kholer S, Heinz D, Urbonas L (2006) Effect of ettringite on thaumasite formation. Cem Concr Res 36:697-706 Kirfel A, Will G (1980) Charge density in anhydrite, CaSO4, from X-ray and neutron diffraction. Acta Crystallogr B36:2881-2890 Klug HP, Alexander LE (1974) X-Ray Diffraction Procedures. Wiley-Interscience, New York Korpa A, Kowald T, Trettin R (2009) Phase development in normal and ultra-high performance cementitious systems by quantitative X-ray analysis and thermoanalytical methods. Cem Concr Res 39:69-76 Kumar S, Kumar R, Bandopadhyay A, Alex TC, Ravi-Kumar B, Das SK, Mehrotra SP (2008) Mechanical activation of granulated blast furnace slag and its effect on the properties and structure of Portland slag cement. Cem Concr Res 30:679-685 Lager GA, Armbruster Th, Faber J Jr (1987) Neutron and X-ray diffraction study of hydrogarnet Ca3Al2(O4H4)3. Am Mineral 72:756-765 Larson AC, Von Dreele RB (1994) Los Alamos National Lab. Rep. No. LA-UR-86-748, Los Alamos Le Saout, Fullmann T, Kocaba V, Scrivener K (2007) Quantitative study of cementitious materials by X-ray diffraction/Rietveld analysis using an external standard. In: Proceedings of 12th International Congress on the Chemistry of Cement (ICCC2007), July 8-13, 2007, Montréal, Canada. 12 pages Le Saout, Kocaba V, Scrivener K (2011) Application of the Rietveld method to the analysis of anhydrous cements. Cem Concr Res 41:133-148

Rietveld Quantitative Phase Analysis of OPC

205

Leoni M, Scardi P, Pelosato R, Sora IN, Dotelli G, Stampino PG, Presti AL (2007) Phase composition gradient in leached polluted cement monoliths. Cem Concr Res 37:1483-1495 León-Reina L, De la Torre AG, Porras-Vázquez JM, Cruz M, Ordonez LM, Alcobé X, Gispert-Guirado F, Larranaga-Varga A, Paul M, Fuellmann T, Schmidt R, Aranda MAG (2009) Round robin on Rietveld quantitative phase analysis of Portland cements. J Appl Crystallogr 42:906-916 Lothenbach B, Matschei T, Möschner G, Glasser FP (2008a) Thermodynamic modelling of the effect of temperature on the hydration and porosity of Portland cement. Cem Concr Res 38:1-18 Lothenbach B, Le Saout G, Gallucci E, Scrivener K (2008b) Influence of limestone on the hydration of Portland cements. Cem Concr Res 38:848-860 Lothenbach B, Scrivener K, Hooton RD (2011) Supplementary cementitious materials. Cem Concr Res 41:217-229 Lothenbach B, Winnefeld F (2006) Thermodynamic modelling of the hydration of Portland cement. Cem Concr Res 36:209-226 Louisnathan SJ (1971) Refinement of the crystal structure of a natural gehlenite, Ca2Al(Al,Si)O7. Can Mineral 10:822-837 Luke K, Glasser FP (1987) Selective dissolution of hydrated blast furnace slag cements. Cem Concr Res 17:273-282 Lundgaard N, Jons ES (2003) Quantitative Rietveld XRD Analysis. World Cem 34:59-63 Madsen IC, Scarlett NVI (2008) Quantitative Phase Analysis. In: Powder Diffraction Theory and Practice. Dinnebier RE, Billinge SJL (eds). Royal Society of Chemistry Publishing, Cambridge, p 298-331 Madsen IC, Scarlett NVY, Cranswick LMD, Lwin T (2001) Outcomes of the International Union of Crystallography Commission on powder diffraction round robin on quantitative phase analysis: samples 1a to 1h. J Appl Crystallogr 34:409-426 Makar JM, Chan GW (2008) End of the induction period in ordinary Portland cement as examined by highresolution scanning electron microscopy. J Am Ceram Soc 91:1292-1299 Maltese C, Pistolesi C, Bravo A, Cella F, Cerulli T, Salvioni D (2007) A case history: effect of moisture on the setting behaviour of a Portland cement reacting with an alkali-free accelerator. Cem Concr Res 37:856865 Manias C, Madsen IC, Retallack D (2001) Plant optimisation and control using continuous on-line XRD for mineral phase analysis. ZKG Int 54:138-145 Marinoni N, Pavese A, Voltolini M, Merlini M (2008) Long-term leaching test in concretes: an X-ray powder diffraction study. Cem Concr Comp 30:700-705 Markvardsen AJ, Shankland K, David, WIF, Johnston JC, Ibberson, RM, Tucker M, Nowell, H, Griffin T (2008) ExtSym: a program to aid space-group determination from powder diffraction data. J Appl Crystallogr 41:1177-1181 Maslen EN, Streltsov VA, Streltsova NR, Ishizawa N (1995) Electron density and optical anisotropy in rhombohedral carbonates. III. Synchrotron X-ray studies of CaCO3, MgCO3 and MnCO3. Acta Crystallogr B51:929-939 Matschei T, Glasser FP (2010) Temperature dependence, 0 to 40 °C, of the mineralogy of Portland cement paste in the presence of calcium carbonate. Cem Concr Res 40:763-777 Matschei T, Lothenbach B, Glasser FP (2007a) The AFm phase in Portland cement. Cem Concr Res 37:118130 Matschei T, Lothenbach B, Glasser FP (2007b) Thermodynamic properties of Portland cement hydrates in the system CaO-Al2O3-SiO2-CaSO4-CaCO3-H2O. Cem Concr Res 37:1379-1410 McCusker LB, Von Dreele RB, Cox DE, Louer D, Scardi P (1999) Rietveld refinement guidelines. J Appl Crystallogr 32:36-50 Merlini M, Artioli G, Cerulli T, Cella F, Bravo A (2008) Tricalcium aluminate hydration in additivated systems. A crystallographic study by SR-XRPD. Cem Concr Res 38:477-486 Merlini M, Artioli G, Meneghini C, Cerulli T, Bravo A, Cella F (2007b) The early hydration and the set of Portland cements: in situ x-ray powder diffraction studies. Powder Diffr 22:201-208 Merlini M, Meneghini C, Artioli G, Cerulli T (2007a) Synchrotron radiation XRPD study on the early hydration of cements. Z Kristallogr Sup 26:411-416 Mesbah A, François M, Cau-dit-Coumes C, Frizon F, Filinchuk Y, Leroux F, Ravaux J, Renaudin G (2011) Crystal structure of Kuzel’s salt 3CaO·Al2O3·1/2CaSO4·1/2CaCl2·11H2O determined by synchrotron powder diffraction. Cem Concr Res 41:504-509 Meyer KW, Neubauer J, Malovrh S (1998) New quality control with standardless clinker phase determination using the Rietveld refinement. ZKG Int 3:152-162 Mitchell LD, Margeson JC, Whitfield PS (2006) Quantitative Rietveld analysis of hydrated cementitious systems. Powder Diffr 21:111-113 Moller H (1995) Standardless quantitative phase analysis of Portland cement clinkers. World Cem 26:75-84

206

Aranda, De la Torre, León-Reina

Moller H (1998) Automatic profile investigation by the Rietveld method for standardless quantitative phase analysis. ZKG Int 1:40-50 Mondal P, Jeffery JW (1975) The crystal structure of tricalcium aluminate, Ca3Al2O6. Acta Crystallogr B31:689-697 Mumme WG, Hill RJ, Bushnell-Wye G, Segnit ER (1995) Rietveld crystal structure refinements, crystal chemistry and calculated powder diffraction data for the polymorphs of dicalcium silicate and related phases. Neues Jahrb Mineral Abh 169:35-68 Narmluk M, Nawa T (2011) Effect of fly ash on the kinetics of Portland cement hydration at different curing temperatures. Cem Concr Res 41:579-589 Navias AL (1925) Quantitative determination of the development of mullite in fired clays by an X-ray method. J Am Ceram Soc 8:296-302 Neubauer J, Kuzel HJ, Sieber R (1996) Rietveld quantitative XRD analysis of Portland cement: II. Quantification of synthetic and technical Portland cement clinkers. In: Proceedings of the 18th International Conference on Cement Microscopy, 100-111, Houston, USA Neubauer J, Mayerhofer W (2000) Solid solution series of ferrate and aluminate phases in OPC: Part I. The ferrate phase. In: Proceedings of the 22nd International Conference on Cement Microscopy, 54-64, Montreal, Canada O’Connor BH, Raven MD (1988) Application of the Rietveld refinement procedure in assaying powdered mixtures. Powder Diffr 3:2-6 Ojima K, Nishihata Y, Sawada A (1995) Structure of potassium sulfate at temperatures from 296 K down to 15K. Acta Crystallogr B51:287-293 Okada K, Ossaka J (1980) Structures of potassium sodium sulfate and tripotassium sodium disulfate. Acta Crystallogr B36:919-921 Onken H (1965) Verfeinerung der Kristallstruktur von Monticellit. Tsch Mineral Petrol Mitt 10:34-44 Paglia C, Wombacher F, Bohni H (2003) The influence of alkali-free and alkaline shotcrete accelerators within cement systems: Influence of the temperature on the sulfate attack mechanisms and damage. Cem Concr Res 33:387-395 Pajares I, De la Torre A, Martinez-Ramirez S, Puertas S, Blanco-Varela MT, Aranda MAG (2002) Quantitative analysis of mineralized white Portland clinkers: The structure of Fluorellestadite. Powder Diffr 17:281286 Palacios L, De la Torre AG, Bruque S, Garcia-Munoz JL, Garcia-Granda S, Sheptyakov D, Aranda MAG (2007) Crystal structures and in-situ formation study of Mayenite electrides. Inorg Chem 46:4167-4176 Paul M, Hornung D, Enders M, Schmidt R, (2004) Process monitoring in a cement plant: the combination of optimized preparation procedures for clinker and cement and Rietveld analysis. Cem Int 2:98-105 Pecharsky V, Zavalij P (2005) Fundamentals of Powder Diffraction and Structural Characterization of Materials. Springer, New York Pelletier L, Winnefeld F, Lothenbach B (2010) The ternary system Portland cement-calcium sulphoaluminate clinker-anhydrite: Hydration mechanism and mortar properties. Cem Concr Comp 32:497-507 Petch HE (1961) The hydrogen positions in portlandite, Ca(OH)2, as indicated by the electron distribution. Acta Crystallogr 14:950-957 Peterson VK, Hunter BA, Ray A, Aldridge LP (2002) Rietveld refinement of neutron, synchrotron and combined powder diffraction data of cement clinker. Appl Phys A Suppl S1409-S1411 Peterson VK, Ray AS, Hunter BA (2006) A comparative study of Rietveld phase analysis of cement clinker using neutron, laboratory X-ray, and synchrotron data. Powder Diffr 21:12-18 Pollmann H (2006) Syntheses, properties and solid solution of ternary lamellar calcium aluminate hydroxi salts (AFm-phases) containing SO42−, CO32− and OH−. Neues Jahrb Mineral Abh 182:173-181 Pollmann H (2007) Characterization of Different Water Contents of Ettringite and Kuzelite. In: Proceedings of 12th International Congress on the Chemistry of Cement, Montreal, Canada Poulsen SL, Kocaba V, Le Saout G, Jakobsen HJ, Scrivener KL, Skibsted J (2009) Improved quantification of alite and belite in anhydrous Portland cements by 29Si MAS NMR: effects of paramagnetic ions. Solid State NMR 36:32-44 Pritula O, Smrcok L, Baumgartner B (2003) On reproducibility of Rietveld analysis of reference Portland cement clinkers. Powder Diffr 18:16-22 Pritula O, Smrcok L, Ivan J, Izdinsky K (2004b) X-ray quantitative phase analysis of residues of the reference Portland clinkers. Ceram Silik 48:34-39 Pritula O, Smrcok L, Tobbens DM, Langer V (2004a) X-ray and neutron Rietveld quantitative phase analysis of industrial Portland cement clinkers. Powder Diffr 19:232-239 Puertas F, García-Díaz I, Palacios M, Gazulla MF, Gómez MP, Orduña M (2010) Clinkers and cements obtained from raw mix containing ceramic waste as a raw material. Characterization, hydration and leaching studies. Cem Concr Comp 32:175-186

Rietveld Quantitative Phase Analysis of OPC

207

Puertas F, Palacios M, Gil-Maroto A, Vazquez T (2009) Alkali-aggregate behaviour of alkali-activated slag mortars: Effect of aggregate type. Cem Concr Comp 31:277-284 Ramlochan T, Hooton RD (2011) Use of X-ray Diffraction to Investigate the Early Stiffening of Cement Pastes. In: Proceedings of 13th International Congress on the Chemistry of Cement, 7 pages, Madrid, Spain Rasmussen SE, Jorgensen JE, Lundtoft B (1996) Structures and phase transitions of Na2SO4. J Appl Crystallogr 29:42-47 Redhammer GJ, Tippelt G, Roth G, Amthauer G (2004) Structural variations in the brownmillerite series Ca2(Fe2−xAlx)O5: Single-crystal X-ray diffraction at 25 °C and high-temperature X-ray powder diffraction (25 °C ≤ T ≤ 1000 °C). Am Mineral 89:405-420 Renaudin G, Kubel F, Rivera JP, Francois M (1999) Structural phase transition and high temperature phase structure of Friedels salt, 3(CaO)(Al2O3)(CaCl2)·10(H2O). Cem Concr Res 29:1937-1942 Richardson IG (2008) The calcium silicate hydrates. Cem Concr Res 38:137-158 Rietveld HM (1967) Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Crystallogr 22:151-152 Rietveld HM (1969) A profile refinement method for nuclear and magnetic structures. J Appl Crystallogr 2:65-71 Rietveld HM (2010) The Rietveld method. Conference at the 10th Spanish Summer School on the Rietveld Method. Castellon (Spain) Rinaldi R, Sacerdoti M, Passaglia E (1990) Straetlingite: crystal structure, chemistry, and a reexamination of its polytype vertumnite. Eur J Mineral 2:841-849 Rodríguez-Carvajal J (1993) Recent advances in magnetic structure determination by neutron powder diffraction. Phys B 192:55-69 Roisnel T, Rodríguez-Carvajal J (2001) WinPLOTR: a Windows tool for powder diffraction patterns analysis. Mater Sci Forum 378-381:118-123 Rossler C, Eberhardt A, Kucerova H, Moser B (2008) Influence of hydration on the fluidity of normal Portland cement pastes. Cem Concr Res 38:897-906 Rousselot I, Taviot-Gueho C, Leroux F, Leone P, Palvadeau P, Besse JP (2002) Insights on the structural chemistry of hydrocalumite and hydrotalcite-like materials: investigation of the series Ca2 M(3+) (OH)6Cl2(H2O) (M(3+): Al, Ga, Fe, and Sc) by x-ray powder diffraction. J Solid State Chem 167:137-144 Saalfeld H, Depmeier W (1972) Silicon-free compounds with sodalite structure. Kristallogr Tech 7:229-233 Saalfeld H, Wedde M (1974) Refinement of the crystal structure of gibbsite, Al(OH)3. Kristallogr Kristallgeome Kristallphys Kristallchem 139:129-135 Saburi S, Kusachi I, Henmi C, Kawahara A, Henmi K, Kawada I (1976) Refinement of the structure of rankinite. Mineral J 8:240-246 Sacerdoti M, Passaglia E (1988) Hydrocalumite from Latium, Italy: its crystal structure and relationship with related synthetic phases. Neues Jahrb Mineral Monatsh 462-475 Sadanaga R, Tokonami M, Takeuchi Y (1962) The structure of mullite, (Al2O3)2SiO2, and relationship with the structures of sillimanite and andalusite. Acta Crystallogr 15:65-68 Saint-Jean SJ, Hansen S (2005) Nonstoichiometry in chlorellestadite. Solid State Sci 7:97-102 Sasaki S, Fujino K, Takeuchi Y (1979) X-ray determination of electron-density distributions in oxides, MgO, MnO, CoO, and NiO, and atomic scattering factors of their constituent atoms. Proc Japan Acad 55:43-48 Sawada H (1996) An electron density residual study of alpha-ferric oxide. Mater Res Bull 31(2):141-146 Scarlett NVY, Madsen IC (2006) Quantification of phases with partial or no known crystal structures. Powder Diffr 21:278-284 Scarlett NVY, Madsen IC, Cranswick LMD, Lwin T, Groleau E, Stephenson G, Aylmore M, Agron-Olshina N (2002) Outcomes of the International Union of Crystallography Commission on Powder Diffraction Round Robin on Quantitative Phase Analysis: samples 2, 3 and 4, synthetic bauxite, natural granodiorite and pharmaceuticals. J Appl Crystallogr 35:383-400 Scarlett NVY, Madsen IC, Manias C, Retallack D (2001) On-line X-ray diffraction for quantitative phase analysis: application in the Portland cement industry. Powder Diffr 16:71-80 Schreyer M, Guo L, Tjahjono M, Garland M (2011) Three approaches to total quantitative phase analysis of organic mixtures using an external standard. J Appl Crystallogr 44:17-24 Scrivener KL, Fullmann T, Gallucci E, Walenta G, Bermejo E (2004) Quantitative study of Portland cement hydration by X-ray diffraction/Rietveld analysis and independent methods. Cem Concr Res 34:15411547 Scrivener KL, Nonat A (2011) Hydration of cementitious materials, present and future. Cem Concr Res 41:651-665 Seufert S, Hesse C, Goetz-Neunhoeffer F, Neubauer J (2009a) Quantitative determination of anhydrite III from dehydrated gypsum by XRD. Cem Concr Res 39:936-941 Seufert S, Hesse C, Goetz-Neunhoeffer F, Neubauer J (2009b) Discrimination of bassanite and anhydrite III dehydrated from gypsum at different temperatures. Z Kristallogr Suppl 30:447-452

208

Aranda, De la Torre, León-Reina

Shanahan N, Zayed A (2007) Cement composition and sulfate attack: Part I. Cem Concr Res 37:618-623 Shin EJ, Seong BS, Lee CH, Huh MY (2002) Quantitative phase analysis of strongly textured alloy mixtures using neutron diffraction. J Appl Crystallogr 35:571-576 Skibsted J, Hall C (2008) Characterization of cement minerals, cements and their reaction products at the atomic and nano scale. Cem Concr Res 38:205-255 Skibsted J, Hall C, Jakobsen HJ (2002) Nuclear magnetic resonance spectroscopy and magnetic resonance imaging of cements and cementbased materials. In: Structure and Performance of Cements. 2nd Edition, Bensted J, Barnes P (ed), Spon Press, London, p 457-476 Smith DK (1992) Particle statistics and whole-pattern methods in quantitative X-ray powder diffraction analysis. Adv X-Ray Anal 35A:1-15 Smith DK, Johnson GG, Scheible A, Wims AM, Johnson JL, Ullrnann G (1987) Quantitative X-ray powder diffraction method using the full diffraction pattern. Powder Diffr 2:73-77 Smith DK, Leider HR (1968) Low-temperature thermal expansion of LiH, MgO and CaO. J Appl Cryst 1:246249 Snellings R, Mertens G, Cizer O, Elsen J (2010) Early age hydration and pozzolanic reaction in natural zeolite blended cements: Reaction kinetics and products by in situ synchrotron X-ray powder diffraction. Cem Concr Res 40:1704-1713 Snyder RL, Bish DL (1989) Quantitative analysis. Rev Mineral 20:101-144 Soleimanian V, Aghdaee SR (2008) Comparison methods of variance and line profile analysis for the evaluation of microstructures of materials. Powder Diffr 23:41-51 Speer D, Salje E (1986) Phase transitions in langbeinites I: Crystal chemistry and structures of K-double sulfates of the langbeinite type M2(2+)K2(SO4)3, M(2+) = Mg, Ni, Co, Zn, Ca. Phys Chem Mineral 13:17-24 Stanek T, Sulovsky P (2002) The influence of the alite polymorphism on the strength of the Portland cement. Cem Concr Res 32:1169-1175 Stark J (2011) Recent advances in the field of cement hydration and microstructure analysis. Cem Concr Res 41:666-678 Stark J, Möser B, Bellmann F (2008) Multi-method investigation of the ordinary Portland cement hydration. Cem Wapno Beton 13:57-65 Steele IM, Pluth JJ (1990) Crystal structure of synthetic yoshiokaite, a stuffed derivative of the tridymite structure. Am Mineral 75:1186-1191 Stutzman P (2005) Powder diffraction analysis of hydraulic cements: ASTM Rietveld round-robin results on precision. Powder Diffr 20:97-100 Stutzman P (2011) Direct Determination of Phases in Portland Cements by Quantitative X-Ray Powder Diffraction. NIST Technical Note 1692, 59 pages Stutzman P, Leigh S (2002) Phase Composition Analysis of the NIST Reference Clinkers by Optical Microscopy and X-ray Powder Diffraction. NIST Technical Note 1441, 44 pages Stutzman P, S. Leigh S, (2007) Phase analysis of hydraulic cements by X-ray powder diffraction: precision, bias, and qualification. J ASTM Int 4:1-11 Suherman PM, van Riessen A, O’Connor B, Li D, Bolton D, Fairhurst H (2002) Determination of amorphous phase levels in Portland cement clinker. Powder Diffr 17:178-185 Takeuchi Y, Nishi F, Maki I (1980) Crystal-chemical characterization of the 3(CaO)Al2O3-Na2O solid-solution series. Z Krist 152:259-307 Tang FJ, Gartner EM (1988) Influence of sulfate source on Portland cement hydration. Adv Cem Res 1:67-74 Taylor HFW (1989) Modification of the Bogue calculation. Adv Cem Res 2:73-77 Taylor HFW (1997) Cement Chemistry. Telford, London Taylor JC, Aldridge LP (1993) Full-profile Rietveld quantitative XRD analysis of Portland cement: Standard XRD profiles for the major phase tricalcium silicate (C3S: 3CaO.SiO2). Powder Diffr 8:138-144 Taylor JC, Hinczak I, Matulis CE (2000) Rietveld full-profile quantification of Portland cement clinker: The importance of including a full crystallography of the major phase polymorphs. Powder Diffr 15:7-18 Taylor JC, Matulis CE (1991) Absorption contrast effects in the quantitative XRD analysis of powders by full multi-phase profile refinement. J Appl Crystallogr 24:14-17 Taylor JC, Zhu R (1992) Simultaneous use of observed and calculated standard profiles in quantitative XRD analysis of minerals by the multiphase Rietveld method: the determination of pseudorutile in mineral sands products. Powder Diffr 7:152-161 Termkhajornkit P, Nawa T, Kurumisawa K (2006) Effect of water curing conditions on the hydration degree and compressive strengths of fly ash-cement paste. Cem Concr Comp 28:781-789 Thomas M, Fournier B, Folliard K, Ideker J, Shehata M (2006) Test methods for evaluating preventive measures for controlling expansion due to alkali-silica reaction in concrete. Cem Concr Res 36:1842-1856 Toby BH (2001) EXPGUI, a graphical user interface for GSAS. J Appl Crystallogr 34:210-213 Toby BH (2006) R factors in Rietveld analysis: How good is good enough? Powder Diffr 21:67-70 Toby BH (2007) A new approach for instruction in powder crystallography. Powder Diffr 22:83-84

Rietveld Quantitative Phase Analysis of OPC

209

Toby BH (2010) Observations on online educational materials for powder diffraction crystallography software. J Appl Crystallogr 43:1271-1275 Toraya H, Hayashi S, Nakayasu T (1999) Quantitative phase analysis of a- and b-silicon nitrides. II. Round robins. J Appl Crystallogr 32:716-729 Torres SM, Kirk CA, Lysndale CJ, Swamy RN, Sharp JH (2004) Thaumasite-ettringite solid solutions in degraded mortars. Cem Concr Res 34:1297-1305 Walenta G, Fullmann T (2004) Advances in quantitative XRD analysis for clinker, cements, and cementitious additions. Powder Diffr 19:40-44 Wenk HR, Monteiro PJM, Kunz M, Chen K, Tamura N, Lutterotti L, Del Arroz J (2009) Preferred orientation of ettringite in concrete fractures. J Appl Crystallogr 42:429-432 Westphal T, Füllmann T, Pöllmann H (2009) Rietveld quantification of amorphous portions with an internal standard-Mathematical consequences of the experimental approach. Powder Diffr 24:239-243 Westphal T, Walenta G. Fullmann T, Gimenez M, Bermejo E, Scrivener K, Pollmann H (2002) Characterization of cementitious materials – Part III. Int Cem Rev July:47-51 Weyer HJ, Muller I, Scmitt B, Bosbach D, Putnis A (2005) Time-resolved monitoring of cement hydration: Influence of cellulose ethers on hydration kinetics. Nucl Instr Meth Phys Res B 238:102-106 Whitfield PS, Mitchell LD (2003) Quantitative Rietveld analysis of the amorphous content in cements and clinkers. J Mater Sci 38: 4415-4421 Whitfield PS, Mitchell LD (2009) The effects of particle statistics on Rietveld analysis of cement. Z Kristallogr Suppl 30:53-59 Will G, Bellotto M, Parrish W, Hart M (1988) Crystal structures of quartz and magnesium germanate by profile analysis of synchrotron-radiation high-resolution powder data. J Appl Crystallogr 21:182-191 Wistuba S, Stephan D, Raudaschl-Sieber G, Plank J (2007) Hydration and hydration products of two-phase Portland cement clinker doped with Na2O. Adv Cem Res 19:125-131 Yamaguchi G, Suzuki K (1967) Structural analysis of merwinite. J Cer Soc Japan 75:220-229 Young RA (ed) (1993) The Rietveld Method. International Union of Crystallography monographs on Crystallography. Oxford University Press, Oxford Zevin LS, Kimmel G (1995) Quantitative X-Ray Diffractometry. Springer-Verlag, New York Zotzl M, Pollmann H (2006) Stability and properties of brownmillerites Ca2(Al, Mn,Fe)2O5 and perovskites Ca(Mn,Fe)O3−x in the system Ca2Fe2O5-“Ca2Mn2O5”-“Ca2Al2O5”. J Am Ceram Soc 89:3491-3497

6

Reviews in Mineralogy & Geochemistry Vol. 74 pp. 211-278, 2012 Copyright © Mineralogical Society of America

Supplementary Cementitious Materials Ruben Snellings, Gilles Mertens and Jan Elsen Department of Earth and Environmental Sciences Katholieke Universiteit Leuven B-3001 Leuven, Belgium e-mail: [email protected]

INTRODUCTION The current widespread use of calcium silicate or aluminate hydrate binder systems in the construction industry finds its roots in the Antique world where mixtures of calcined lime and finely ground reactive (alumino-)silicate materials were pioneered and developed as competent inorganic binders. Architectural remains of the Minoan civilization (2000-1500 BC) on Crete have shown evidence of the combined use of slaked lime and additions of finely ground potsherds to produce stronger and more durable lime mortars suitable for water-proof renderings in baths, cisterns and aqueducts (Spence and Cook 1983). It is not clear when and where mortar technology evolved to incorporate volcanic pumice and ashes as a functional supplement. A plausible site would be the Akrotiri settlement at Santorin (Greece), where archeological indications of strong ties with the Minoan culture were found and large quantities of suitable highly siliceous volcanic ash were present. This so-called Santorin earth has been used as a pozzolan in the Eastern Mediterranean until recently (Kitsopoulos and Dunham 1996). Evidence of the deliberate use of this and other volcanic materials by the ancient Greeks dates back to at least 500-400 BC, as uncovered at the ancient city of Kamiros, Rhodes (Efstathiadis 1978; Idorn 1997). In the subsequent centuries the technological knowledge was spread to the mainland (Papayianni and Stefanidou 2007) and was eventually adopted and improved by the Romans (Mehta 1987). The Roman alternatives for Santorin earth were volcanic pumices or tuff found in neighboring territories, the most famous ones found in Pozzuoli (Naples), hence the name pozzolan, and in Segni (Latium). Preference was given to natural pozzolan sources, but crushed ceramic waste was frequently used when natural deposits were not locally available. The exceptional lifetime and preservation condition of some of the most famous Roman buildings such as the Pantheon or the Pont du Gard constructed with the aid of pozzolan-lime mortars and concrete testify to both the excellent workmanship reached by Roman “engineers” and to the durable properties of the utilized binder materials. In designing the binder mixes Romans seem to have paid much attention to a very thorough grinding and mixing of components and to the granulometric gradation of the aggregate with an increased proportion of fines (Day 1990). This principle of extending the range of particle sizes to improve space filling is based on the so called Apollonian concept and has more recently been applied to the binder phase in modern concrete to achieve ultrahigh strength concretes (Scrivener and Kirkpatrick 2008). The utilization of lime-pozzolan binders recovered gradually after the decline of the Roman Empire, particularly due to their hydraulic capability of hardening underwater. However, pozzolan-lime binders were gradually replaced by Portland cement based binders over the course of the 19th and early 20th century (Blezard 2001).

1529-6466/12/0074-0006$10.00

DOI: 10.2138/rmg.2012.74.6

212

Snellings, Mertens, Elsen

More recently, especially since the 2nd half of the 20th century, the addition to Portland cement of natural or artificial materials able to react with lime to produce a cementitious product has received renewed attention. The production of Portland clinker being an energy-intensive process, in which raw materials are typically burned at 1450 °C, renders the economic advantages of replacing a substantial part of the clinker by cheap naturally available pozzolans or industrial by-products obvious. The replacement of a part of the cement clinker does not need to have negative effects on the mechanical and durability performance of the so-called blended cement. To the contrary, many studies have indicated that particularly the durability of blended cements is substantially improved. An overview of the technological effects of using blended cements is included in the present chapter. A more recent important incentive to increase and optimize the incorporation of supplementary cementitious materials (SCMs) in blended cements is the problem of climate change associated with the anthropogenic emission of greenhouse gasses. Additional to the need of reaching high temperatures in the cement kilns, the release of CO2 by decomposition of limestone results in an average ratio of 0.87 kg CO2 emitted per kg of Portland cement produced (Damtoft et al. 2008). It is estimated that with a current annual cement production of 2.8 billion tonnes the cement industry alone is responsible for 5-7% of the anthropogenic emissions. Future prospects foresee a drastic increase in Portland cement production in developing countries. Therefore, the contemporary cement industry is faced with the challenges of producing more sustainable, less energy intensive products without sacrificing the mechanical or durability performance of the end product. The societal incentives and associated emission quotas in order to reduce global CO2 emissions are rapidly evolving into an all-important issue for the cement industry (Damtoft et al. 2008; Habert et al. 2010). In response, the currently most common development with limited interference in the conventional production process is the increased blending of supplementary cementitious materials or pozzolans with Portland cement (Gartner 2004). The utilization of industrial by-products available in large and regular quantities of suitable consistent composition, i.e., ground granulated blast furnace slags from iron smelting and coal or lignite fly ashes from electricity production, has been firmly established in many countries. However, the supply of high quality SCM by-products is limited and many local sources are already fully exploited. In addition, a decline in production of blast furnace slags and fly ashes is expected due to future developments in steel and electricity production (Scrivener and Kirkpatrick 2008). An illustration of the evolution of the clinker substitution in France is given by Habert et al. (2010). Figure 1 shows that since the early 70’s the total replacement percentage has remained stable at 20%, while the main cement replacement materials have shifted from slag and fly ash to limestone. Alternatives to the traditional industrial by-product SCMs are to be found on the one hand in an increased usage of naturally occurring SCMs and on the other hand in the expansion of the range of industrial by-products or societal waste to substitute for clinker. The development of cement and concrete prescriptions and standards towards more performance based conditions highlights the generally accepted view that the utilization of a wider array of SCMs at higher replacement percentages should be allowed and judged on the eventual performance of the end product (Hooton 2008; Kaid et al. 2009). Natural SCMs or pozzolans are abundant in certain locations and are extensively used as an addition to Portland cement in countries such as for example Italy, Germany, Greece and China. Compared to traditional industrial SCMs they are characterized by a larger range in composition and a larger variability in physical properties. The application of natural pozzolans in Portland cement is mainly controlled by the local availability of suitable deposits and the competition with the accessible traditional industrial by-product SCMs. In part due to the exhaustion of the latter sources and the extensive reserves of natural pozzolans available, partly because of the proven technical advantages of an intelligent use of natural pozzolans, their use is expected to be strongly expanded in the future (Mehta 1987; Damtoft et al. 2008).

Supplementary Cementitious Materials

213

Figure 1. Evolution of clinker substitution materials in France from 1973 to 2007 (Habert et al. 2010).

A substantial part of the by-products and waste materials generated by present-day society have the potential of being utilized in construction materials. Immobilization of harmful metals in waste materials by incorporation in the high pH environment of hydrated cement offers interesting perspectives for application. Apart from avoiding costs of waste disposal and environmental pollution, lowering cement costs and increasing the product sustainability, some materials may also beneficially affect the microstructure and the mechanical and durability properties of mortars and concrete (Meyer 2009). Recently, a substantial amount of research has been dedicated to improve the understanding of the behavior of a broadened range of waste materials in cement and concrete products and to explore the potential applications. This review will be limited to waste materials or by-products which are already extensively employed as reactive binder components, e.g., fly ashes and blast furnace slags, and have successfully passed the development and testing stage. For a recent review of the usage of by-product or waste materials as aggregate, the reader is referred to the comprehensive survey of Siddique (2008). The potential of clinker substitution by SCMs to decrease production costs and to increase sustainability and durability of the end-products is reflected in the large and steadily growing numbers of peer-reviewed papers published on the subject. Excellent general literature reviews which have been remarkable sources of information for the present paper have been published earlier by Massazza (1974, 2001), Sersale (1980, 1993), Takemoto and Uchikawa (1980), Swamy (1986), Malhotra and Mehta (1996) and Siddique (2008). This literature review will mainly consider more recent insights and research findings and put them into perspective with the previously existing knowledge on supplementary cementitious materials. Though the combined use of chemical admixtures and SCMs is common, a thorough treatment on the effect of chemical admixtures on cement properties would require a separate review. Also, no account is made of the different methodologies for testing the reactivity of SCMs or pozzolans. Instead, here, a general definition and classification of SCMs is presented first, followed by sections treating the physico-chemical properties of specific SCM groups. A detailed account of recent developments in the understanding of the pozzolanic reaction mechanism and kinetics has been provided, together with a general overview of the reaction products encountered. Finally, a concise outline of the properties of SCM-blended cement binders is offered.

214

Snellings, Mertens, Elsen Definition and classification of SUPPLEMENTARY CEMENTITIOUS MATERIALS

Definition The group of supplementary cementitious materials comprises materials that show either hydraulic or pozzolanic behavior. A hydraulic binder is a material that can set and harden submerged in water by forming cementitious products in a hydration reaction. Iron blast furnace slags commonly show a “latent hydraulicity,” i.e., their hydraulic activity is relatively low compared to Portland cement and activation by chemical or physical means is needed to initiate and accelerate the hydration reaction (Regourd 1986; Lang 2002). Blast furnace slags can be chemically activated by addition of alkali-hydroxides, sulfates in the form of gypsum or anhydrite or more frequently by the addition of lime or lime-producing materials such as Portland cement. It should be noted that hydraulic materials can replace Portland cement up to a much larger extent than materials showing pozzolanic behavior. A pozzolan is generally defined in ASTM C618 as a siliceous or siliceous and aluminous material which, in itself, possesses little or no cementitious value but which will, in finely divided form and in the presence of moisture, react chemically with calcium hydroxide (lime) at ordinary temperature to form compounds possessing cementitious properties (Mehta 1987). It should be remarked that the definition of a pozzolan imparts no bearing on the origin of the material, only on its capability of reacting with lime and water. A quantification of this capability is comprised in the term pozzolanic activity. The pozzolanic activity is by convention described as a measure for the degree of reaction over time between a pozzolan and Ca2+ or Ca(OH)2 in the presence of water. Physical surface adsorption is not considered as being part of the pozzolanic activity, because no irreversible molecular bonds are formed in the process (Takemoto and Uchikawa 1980). The driving force underlying the pozzolanic activity is the difference in Gibbs free energy between the initial and final reaction stages, while the reaction kinetics are governed by the activation energy barrier which needs to be surmounted to proceed in the reaction (Felipe et al. 2001). It should be remarked that the bulk properties of the end product (i.e., mortar, concrete etc.) are not directly related to the SCM inherent pozzolanic activity. Physical bulk properties such as permeability and mechanical strength are more strongly dependent on the type, shape, dimensions and distribution of reaction products and pores than on the extent of the lime-pozzolan reaction (Takemoto and Uchikawa 1980; Massazza 2001). The former factors are mainly affected by the mix design and curing conditions and can thus be controlled. However at equal binder preparation conditions, the pozzolanic activity remains a primary factor that controls the capability of a material to engage in the formation of cementitious compounds and in consequence also the contribution of the material to the binder performance.

Classification The general definition of supplementary cementitious material embraces a large number of materials which vary widely in terms of origin, chemical and mineralogical composition, and typical particle characteristics. Although it is generally accepted that the hydraulic or pozzolanic activity of SCMs depends largely on their physico-chemical properties rather than their origin (Sersale 1993), classifications of SCMs according to their activity or their performance in concrete (e.g., Mehta 1989) have known little success. Still more commonly accepted is the classification based on the origin of the SCM (Massazza 2001) and this will be followed here. Two broad categories can be distinguished, on the one hand materials of a natural origin and on the other hand materials of man-made or artificial origin. The former group consists of materials that can be used as SCM in their naturally occurring form. In most cases they only need conditioning of particle characteristics by sieving and grinding processes. Typical natu-

Supplementary Cementitious Materials

215

ral SCMs are pyroclastic rocks, either diagenetically altered or not, and highly-siliceous sedimentary rocks such as diatomaceous earths. The group of artificial SCMs includes materials which have undergone structural modifications as a consequence of manufacturing or production processes. Artificial SCMs can be produced deliberately, for instance by thermal activation of kaolin-clays to obtain metakaolin, or can be obtained as waste or by-products from hightemperature processes such as blast furnace slags, fly ashes or silica fume. A general genetic classification scheme is presented in Figure 2. Natural SCMs are subdivided into materials of primary volcanic origin and materials of sedimentary origin. The volcanic materials utilized are generally pyroclastics and can be altered by diagenetic processes to zeolite-rich tuffs. The sedimentary rocks comprise chemical and detrital sediments. Both biochemically deposited SCMs such as diatomaceous earths and deposits resulting from the circulation of hydrothermal fluids are included in the former category. Naturally burned clays, such as gliezh, are an example of the use of detrital sediments. Some individual materials, e.g., Danish moler and French gaize, cannot be distinctively categorized as either a natural or artificial SCM because their natural pozzolanic activity is commonly enhanced by thermal treatment. Here, they are included together with the artificial SCMs. Other materials, such as Sacrofano earth, containing components of mixed natural origins (volcanic, detritic and biogenic) are classified under the category of materials of sedimentary origin. The classification of artificial SCMs is conveniently based either on the industrial processes producing the SCMs or on the original materials that are thermally treated to intentionally manufacture SCMs. Some waste materials with the potential to become more widely used as SCM in the future, but which need further experimental evaluation are also mentioned in the classification scheme.

Figure 2. General classification scheme of supplementary cementitious materials. * denotes materials which can present hydraulic activity, all other materials display pozzolanic behavior. A selection of promising SCMs still largely under development are positioned below the dashed line (modified after Massazza 2001).

216

Snellings, Mertens, Elsen

Within a group of SCMs of the same origin there can be considerable variability in the physico-chemical properties. The extent of variability depends on the SCM origin. The large compositional variability of volcanic extrusive rocks is a reflection of the natural variability in magmatic and diagenetic processes leading to their present condition. In contrast, the characteristics of silica fume produced in the controlled manufacturing of silicon metal and ferrosilicon alloys show much more consistency. Apart from differences in variability, separate groups show characteristic ranges of chemical composition. Figure 3 illustrates the chemical composition and typical variability for the most commonly used groups of SCMs in a CaO-SiO2Al2O3 ternary variation diagram. Alkalis, MgO and Fe2O3 content are ignored in this diagram. The figure can also be instrumental in estimating the impact of SCM incorporation on the blended cement chemistry and can eventually be used to predict reaction product assemblages (cf. infra).

Figure 3. Ternary CaO-SiO2-Al2O3 diagram (wt% based) situating the chemical constitution of the major SCM groups (modified after Glasser et al. 1987).

MINERALOGY AND CHEMISTRY OF SCMs Natural SCMs The great majority of natural pozzolans in use today is of volcanic origin, mainly owing to the widespread availability of volcanic rocks in many countries. In Figure 4 the global distribution of volcanic rocks can be compared with the reported occurrences of natural pozzolan deposits. It is apparent that the overwhelming majority of these deposits are located in areas of Cenozoic volcanic activity. However, not all volcanic rocks are suitable as pozzolanic material. Pyroclastic materials resulting from explosive eruptions such as ashes or pumices show higher pozzolanic activity owing to their higher glass content and highly porous or vesicular nature. The eruption type largely depends on the magma viscosity which is related to the “acidity” (i.e., SiO2 content) of the magma. In general, more siliceous magma produces more explosive volcanism and products with better pozzolanic properties. Coarser highly vesicular pyroclastic

Supplementary Cementitious Materials

217

Figure 4. Global distribution of volcanic rocks (grey areas) and deposits of reported natural supplementary cementitious materials (black dots).

material forms pumice-type deposits. Finely divided materials are transported further away from the volcanic source and are deposited as ash layers. Subsequent to the deposition of the pyroclastic material, diagenetic alteration of the vitreous material to crystalline zeolites can occur under certain circumstances. The resulting zeolite-bearing rocks are often coherent tuffs and present, when ground to sufficient fineness, high pozzolan activities. As the likelihood of a material having undergone diagenetic alteration increases with its age, zeolitized tuffs tend to become more abundant in rocks of increasing geologic age throughout the Neogene (Gottardi and Obradovic 1978). The type of zeolites formed and the quality of the zeolitized rocks is largely related to the composition of the original vitreous material. Regional differences are therefore common and are ultimately related to the type of volcanism and the corresponding geological situation. The utilization of materials of sedimentary origin is scarcer, which is evidently related to the general stability at ambient conditions of the mineral assemblages deposited as sediments. However, in some specific conditions sediments rich in pozzolanically active components can be formed during deposition, e.g., diatomaceous earths, or due to subsequent alteration, e.g., naturally burned clays. Topical reviews on natural pozzolans and their applications can be found in Cook (1986a), Day (1990), Malhotra and Mehta (1996), Colella et al. (2001) and Massazza (2001, 2002). Unaltered pyroclastic materials. The major pozzolanically active component of unaltered pyroclastic pumices and ashes is a highly porous glass (Ludwig and Schwiete 1963). The easily alterable, or highly reactive, nature of these ashes and pumices limits their occurrence largely to recently active volcanic areas. Most of the traditionally used natural pozzolans belong to this group, i.e., volcanic pumice from Pozzuoli, Santorin earth and the incoherent parts of the German trass. The international (IUGS) classification of glassy or aphanitic rock types based on the chemical composition has been applied to natural pozzolans of volcanic origin. The reported chemical data of 150 unaltered pyroclastic materials and 83 zeolitized rocks used as natural pozzolanic material were plotted in a total alkali over SiO2 diagram on a recalculated 100% volatile-free basis in Figure 5. A large spread of data indicates the variability in composition. SiO2 being the major component, most natural unaltered pumices and ashes fall in the intermediate (52-66 wt% SiO2) to acid (> 66 wt% SiO2) composition range. The

218

Snellings, Mertens, Elsen

Figure 5. IUGS classification (Le Maître et al. 1989) of vitreous rock types based on chemical composition on an anhydrous basis.

predominant rock types are dacite and rhyolite, representing respectively 29% and 21% of the reported analyses. Basic (45-52 wt% SiO2) and ultrabasic pyroclastics (< 45 wt% SiO2) are less commonly used as natural pozzolans, and represent only 15% of the reported analyses. The total alkali content is variable and linked to the regional type of volcanism. It can reach levels higher than 11 wt% on an anhydrous basis in Neapolitan pozzolans (e.g., Battaglino and Schippa 1968), Moroccan “leucitite” (Hilali et al. 1981 in Day 1990) or pumicite from Idaho (Asher 1965 in Day 1990). A synthesis of the collected chemical data of the unaltered pyroclastics is shown as a series of box plots in Figure 6. Apart from SiO2 as the main component, Al2O3 is present in substantial amounts in most reported pozzolans. Most samples contain Fe2O3 and MgO in minor proportions, which is typical of more acid rock types. CaO and alkali concentrations are usually modest, but can vary substantially depending on for instance the presence of calcite as a secondary phase. Loss on ignition (LOI) is generally low but can reach values over 10 wt% in some trasses and tuffs which probably contain substantial amounts of unreported zeolites and/ or clay minerals. A summarizing representation of the collected chemical data of the unaltered pyroclastics is shown as a series of box plots in Figure 6. Apart from SiO2 as the main component, Al2O3 is present in substantial amounts in most reported pozzolans. Most samples contain Fe2O3 and MgO in minor proportions, which is typical for more acid rock types. CaO and alkali concentrations are usually modest, but can vary substantially depending on for instance the presence of calcite as a secondary phase. Loss on ignition (LOI) is generally low but can reach values over 10 wt% in some trasses and tuffs which probably contain substantial amounts of unreported zeolites and/or clay minerals.

Supplementary Cementitious Materials

219

Figure 6. Box plots of the distribution of collected chemical data (N = 150) of unaltered pyroclastic materials investigated as pozzolanic material. Crosses are outliers.

The mineralogical composition of unaltered pyroclastic rocks is mainly determined by the presence of phenocrysts and the chemical composition of the parent magma. Additionally, the pyroclastic material can become intermixed with the constituents of detrital or biochemical sediments during deposition. Finally, some limited alteration can have occurred after deposition. The major component is volcanic aluminosilicate glass typically present in quantities over 50 wt%. Unaltered pyroclastics containing significantly less volcanic glass, such as the trachyandesite from Volvic (France) with only 25 wt% are reported to be less reactive (Mortureux et al. 1980). Within a group of unaltered natural pozzolans with similar volcanic origin and composition, a correlation between amorphous phase content and pozzolanic activity has been observed (Mehta 1981). The glass SiO2 content ranges between 45 and 75 wt% and its chemical composition can differ slightly from the bulk composition through the incorporation of lithophile elements in high-temperature phenocrysts (e.g., Ca in anorthite). Mielenz et al. (1950) reported that basaltic glass appeared to be inferior to more acid glasses in terms of pozzolanic activity. Akman et al. (1992) confirmed the better performance of natural pozzolans of rhyolitic and trachytic composition over more basaltic pozzolans. Apart from the glass content and its morphology associated with the specific surface, also defects and the degree of strain in the glass phase appear to be important for the pozzolanic activity (Mehta 1981). The impact of the latter parameters remains unclear due to difficulties in their experimental quantification. Typical associated minerals present as large phenocrysts are members of the plagioclase feldspar solid solution series varying between albite and anorthite depending on the parent magma composition. In pyroclastic materials in which alkalis predominate over Ca, K-feldspar such as sanidine or albite Na-feldspar were identified (Neapolitan pozzolans). Leucite is present as phenocrysts in the K-rich, silica-poor Latium pozzolans (Costa and Massazza 1974). Quartz is usually present in minor quantities in acidic pozzolans, while pyroxenes and/or olivine phenocrysts are often found in more basic materials. Xenocrysts or rock fragments incorporated during the violent eruptional and depositional events are also encountered. Zeolite and clay minerals are often present in minor quantities as alteration products of the volcanic glass. While zeolitization in general is beneficial for the pozzolanic activity, clay formation or argillitization

220

Snellings, Mertens, Elsen

has adverse effects on the performance of lime-pozzolan blends or blended cements (Massazza 1974; Sersale 1980; Akman et al. 1992; Türkmenoğlu and Tankut 2002). Other early diagenetic products reported are opal CT (a more or less disordered interlayering of the SiO2 polymorphs cristobalite and tridymite) and secondary calcite. The former shows elevated pozzolanic activity, the latter is not a pozzolan but can contribute to the material performance of Portland cement by reacting with calcium-aluminate-hydrate reaction products (cf. infra; Matschei et al. 2007a; Matschei and Glasser 2010) or acting as a filler material. Another type of unaltered volcanic pozzolanic material is perlite, a volcanic glass containing relatively high amounts of water. Perlite is typically rhyolitic, containing around 75 wt% SiO2, 10-15% Al2O3 and additional alkalis. The pozzolanic activity of ground perlite from deposits in Turkey (Erdem et al. 2007) and China (Yu et al. 2003) was evaluated to be high. Altered pyroclastic materials. Diagenetic alteration of pyroclastic deposits by alkaline fluids results in the formation of micrometer size zeolite minerals (Fig. 7). Alteration by less alkaline fluids usually leads to the development of clay minerals. The recrystallization of the volcanic glass often results in a more compact and coherent tuffaceous rock. Zeolites are members of the tectosilicate group and possess open, porous framework structures of cornersharing AlO4 and SiO4 tetrahedra. The aluminosilicate tetrahedra are arranged in rings which are three-dimensionally connected to form open cages and channels running through the crystal. The substitution of Al3+ for Si4+ imposes a net negative framework charge, which is compensated by exchangeable cations in cages and channels. Water molecules are sequestered in the cages due to charge-dipole interactions. Zeolitization can occur in a range of geological environments. The most important zeolitization processes in terms of deposit volume are due to circulation of saline-alkaline lacustrine waters, of alkaline ground water or of hydrothermal fluids and due to low-grade burial diagenesis metamorphism (Hay and Sheppard 2001). The range in chemical composition of the zeolitized pyroclastic materials used as pozzolans broadly coincides with that of their unaltered counterparts (Fig. 8). The most prominent difference is the higher loss on ignition due to the formation of zeolite and/or clay minerals containing substantial amounts of water. On an anhydrous base, the altered pyroclastics used as

Figure 7. SEM picture of typical micrometer-size “coffin”-shaped clinoptilolite crystals formed as alteration product of volcanic glass in a tuff sample from Karacaderbent, Turkey.

Supplementary Cementitious Materials

221

Figure 8. Box plots of the distribution of collected chemical data (N = 83) of zeolitized pyroclastic materials investigated as pozzolanic material.

pozzolans tend to be more siliceous and contain somewhat more alkalis. A comparison of the unaltered and altered pyroclastics in a ternary diagram in Figure 9 corroborated this observation in that a larger proportion of the unaltered pumices and ashes are shifted towards the CaO and Al2O3 apices. This can be linked with the fact that zeolitized rocks used as pozzolans are mostly reported from regions with products of more siliceous or alkali-rich volcanism. More siliceous zeolitized rocks are encountered in the Balkan region (e.g., Držaj et al. 1980;

Figure 9. Ternary CaO-SiO2-Al2O3 diagram (wt% based) comparing the chemical composition of unaltered and zeolitized pyroclastics used as pozzolanic material.

222

Snellings, Mertens, Elsen

Naidenov 1991; Janotka et al. 2003), Greece (e.g., Kitsopoulos and Dunham 1996; Fragoulis et al. 1997; Stamatakis et al. 1998; Perraki et al. 2003) and in the Turkey-Iran volcanic chain (e.g., Canpolat et al. 2004; Yilmaz et al. 2007; Uzal et al. 2010; Ahmadi and Shekarchi 2010). In the Circumpacific siliceous zeolite deposits utilized or evaluated as pozzolans occur for instance in New Zealand, the Philippines (Mertens et al. 2009), Japan (Takemoto and Uchikawa 1980), the western United States (Mielenz et al. 1950), Mexico (Rodriguez-Camacho and Uribe-Afif 2002) and Ecuador (Mertens et al. 2009). Zeolite deposits of trachytic to phonolitic composition are situated on the Italian peninsula (e.g., Sersale and Frigione 1987; Liguori et al. 2003) and in the Eifel region in Germany (e.g., Liebig and Althaus 1998). The type of zeolite assemblage formed during alteration depends on a number of factors. Most important are the temperature of formation, the chemical composition of the volcanic glass and the zeolitizing fluid (Chipera and Apps 2001). Zeolites are not the thermodynamically most stable phases and form because of faster reaction kinetics. The most widespread siliceous zeolites are clinoptilolite, mordenite and erionite. These zeolites are observed to form in the alteration of siliceous glasses, whereas the common more aluminous zeolites, phillipsite, chabazite, analcime and heulandite (alumina-rich polymorph of clinoptilolite) form from the alteration of more basic glasses. Heulandite-clinoptilolite is by far the most frequently (up to 60%) identified zeolite mineral in natural zeolite-rich pozzolans. The compositional flexibility of this zeolite in terms of Si-Al and exchangeable cation composition promotes its formation. Mordenite was identified as a major constituent in about 10% of the altered natural pozzolans and is a silica-rich zeolite with a predominance of Na over Ca and K as exchangeable cation. Most of the identified chabazite and phillipsite pozzolans are derived from alkali-rich, intermediate to basic pyroclastics occurring in Italy or Germany and together make out 25% of the reported occurrences. Phillipsite is usually formed in K-rich volcaniclastics such as the Neapolitan Yellow Tuff. Analcime is a Na-rich zeolite which can be formed by the reaction of volcanic glass or zeolite precursors with Na-rich fluids and commonly occurs together with quartz. The eventual zeolite content of the tuff is a function of the original amount of volcanic glass and the extent of the zeolitization process. Altered pyroclastic materials can consist of several types of zeolite and zeolites are often associated with other alteration products such as opal A (amorphous), opal CT, clay minerals and authigenic feldspar and the original pyrogenic phenocrysts (cf. supra). The main pozzolanically active phases are considered to be the zeolite and silicapolymorphs together with relict volcanic glass. Zeolite is reported to be more reactive than its unaltered vitreous counterpart (Sersale and Frigione 1987). The pozzolanic activity of a zeolitized pozzolan seems to be a function of a number of variables. First of all, the content of zeolite and other active phases is important (Liguori et al. 2003), next the reactivity of the zeolite itself depends on its crystallinity (Snellings et al. 2010b), and its framework and exchangeable cation composition (Caputo et al. 2008; Mertens et al. 2009; Snellings et al. 2009). Silica-rich zeolites have been found to render the blended cements more performant in terms of strength and durability than aluminous zeolites (Huizhen 1992; Fragoulis et al. 1997; Rodriguez-Camacho and Uribe-Afif 2002; Caputo et al. 2008), which can be linked with larger amounts of calcium-silicate-hydrate reaction products (C-S-H) being formed. Furthermore, clinoptilolite tuffs exchanged to Na or K were more reactive than tuffs exchanged to Ca (Luke 2007; Mertens et al. 2009). A classification solely based on the presence of a particular zeolite type does not seem to be meaningful due to the variability in zeolite content, composition and crystallinity in tuffs. (Bio)chemical sediments. This category comprises both biogenic sediments, resulting from the deposition of (micro-) organism skeletons, as well as chemical precipitates resulting from the circulation of hydrothermal waters. Diatomaceous earths are the principal biogenic materials that show pozzolanic activity. They mainly consist of siliceous skeletons or frustrules

Supplementary Cementitious Materials

223

of diatom micro-organisms together with variable amounts of calcareous biogenic material and detrital sediment such as clay minerals. The diatoms can be deposited in lacustrine and marine waters. The depositional environment influences the size of the diatoms, those of fresh-water origin generally being smaller and more pozzolanically reactive (Stamatakis et al. 2003). The distribution of chemical data of 22 diatomaceous earths is presented in Figure 10. The silica content is usually high but can drop significantly because of mixing of siliceous diatoms with calcareous material. In reported impure diatomaceous earths calcite contents may reach 50 wt% or more. The presence of feldspars, clay minerals and finely divided iron(hydr)oxides is evidenced by significant amounts of Al2O3 and Fe2O3. The levels of alkalis and MgO are usually very low.

Figure 10. Box plots of the distribution of collected chemical data of diatomaceous earths (N = 22) materials investigated for application as pozzolanic material.

Diatom frustrules are composed of opal A. The higher the opal A content of the diatomaceous earth, the higher the pozzolanic activity and the better the performance. The pozzolanic activity of diatomaceous earths containing large amounts of clay minerals, such as Danish “moler,” can be improved by thermal decomposition of the clay minerals (Johansson and Andersen 1990). In calcite-rich diatomites the reduced amount of opal A can be partially compensated by a filler effect of the calcareous compounds (Stamatakis et al. 2003). The intricate surface morphology and high porosity of the disk-shaped diatom frustrules (Fig. 11) results in typically elevated specific surface areas and water demand, limiting the incorporation of diatomite earths to about 15% without usage of water-reducing agents such as superplasticizers (Stamatakis et al. 2003; Degirmenci and Yilmaz 2009). Interest in some hydrothermal siliceous sinters has been raised principally because of their high silica content. Deposits considered for utilization as pozzolanic material are located in New Zealand, Japan (Takemoto and Uchikawa 1980) and Turkey (Davraz and Gunduz 2005, 2008). Opal A is the main constituent together with some quartz and the material is described as a pumiceous, porous rock demonstrating a high specific surface area. Materials of detrital and mixed origin. Detrital sediments are usually largely composed of stable mineral compounds derived from the erosion and weathering of other rocks. In these

224

Snellings, Mertens, Elsen

Figure 11. SEM picture of disk-shaped diatom frustules from Giannota, Greece. [Used by permission of Elsevier, from Fragoulis et al. (2005), Cement and Concrete Composites, Vol. 27, Fig. 1, p. 206.]

sediments, it is rather uncommon to find pozzolanically reactive compounds such as reworked volcanic glass fragments or diatoms in sufficient quantities to be suitable for application. Some exceptional materials of detrital or mixed origin have nevertheless been used as a pozzolan. One is the Sacrofano earth of mixed origin which can be found near Viterbo (Italy) and shows a very high SiO2 content around 85-90 wt%. Volcanic particles, diatoms and some crystalline minerals present were severely altered by acid fluids infiltrating the uppermost layers of the deposit and leaving essentially a dessicated pozzolanically reactive silica gel (Massazza 2001). Other reactive materials deriving from detrital rocks are naturally burned clays such as porcellanite from Trinidad (Day 1990) and gliezh from Central Asia (Kantsepolsky et al. 1969). The former is thought to have formed by spontaneous combustion of bituminous or lignitic clays and the latter are shales burned by natural subsurface coal fires (Massazza 2001). Gaize is a sedimentary rock occurring in the French Ardennes and Meuse valley. It has a high proportion of quartz and biogenic siliceous material and a substantial amount of clay. Gaize is generally calcined before use as a pozzolan (Cook 1986a). Table 1 shows an overview of the chemical composition of several particular natural pozzolans. In Figure 12 natural pozzolans of a sedimentary origin can be compared in a CaO-SiO2-Al2O3 ternary plot. The presence of calcite in some diatomaceous earths is illustrated by the mixing trend occurring between the CaO and SiO2 apices. The majority of other pozzolans are characterized by their high SiO2 content.

Thermally activated SCMs The use of natural pozzolans was traditionally limited to the regions where they were locally available. To produce water-proof lime-based binders in other regions, people traditionally resorted to blends of thermally activated clays or soils and lime. Usually waste streams of clay pottery, bricks or tiles were recycled as pozzolanic additive. The widespread use of these thermally activated pozzolans before the advent of Portland cement can be illustrated by their presence in e.g., the flooring fundaments of the San Marco basilica in Venice or the Hagia Sophia in Istanbul (Zendri et al. 2004). Crushed brick pozzolans remains continued as a traditional building material in some countries, e.g., under the name of surkhi in India, homra in Egypt and sarooj in Oman (Cook 1986b; Al-Rawas et al. 1998). In modern construction, large volumes of purposely burned clays have been used in the construction of large scale structures such as dams (Jones 2002). The excellent pozzolanic properties of kaolinite-rich materials

Supplementary Cementitious Materials

225

Table 1. Overview of the chemical composition of natural pozzolans of sedimentary origin. pozzolana

origin Al2O3 SiO2 Fe2O3 CaO Na2O K2O MgO SO3 LOI SUM ref.

Sacrofano earth Gliezh Gaize Silica sinter Silica sinter

Italy USSR France Japan Turkey

3.1 12.4 7.1 2.4 2.6

89.2 72.8 79.6 87.7 92.5

0.8 5.1 3.2 0.5 0.1

2.3 2.8 2.4 0.2 0.3

0.4 0.1 1.1

1.7 0.1 0

1.1 1.0 0.2 0

1.5 0.9 0.1 0.1

4.7 5.9 4.1 1.9

100.1 97.8 100.1 95.4 98.6

1 2 3 4 5

1: Battaglino and Schippa (1968); 2: Kantsepolsky et al. (1969); 3: Lea (1970); 4: Takemoto and Uchikawa 1980; 5: Davraz and Gunduz 2005

Figure 12. Ternary CaO-SiO2-Al2O3 diagram (wt% based) illustrating the distribution in chemical composition of a selection of natural pozzolans from a sedimentary origin.

burned under controlled conditions, better known as metakaolin, has drawn renewed attention towards deliberately thermally activated clays. The pozzolanic properties of burned clays and shales have been reviewed by Cook (1986b), more specific reviews on metakaolin can be found in Sabir et al. (2001), Jones (2002) and Siddique (2008). Next to burned clays, shales and soils, burned agricultural residues consisting predominantly of amorphous silica have received considerable attention as pozzolan for construction purposes in rural areas deprived of other SCMs. General reviews are presented by Cook (1986c) and Siddique (2008). Burned clays and shales. Clays are defined as sediments which consist primarily of particles smaller than 2 mm. The constituting particles are mostly phyllosilicates incorporating a considerable amount of water. These minerals are commonly termed clay minerals and are composed of sheets of tetrahedrally (T) coordinated SiO4 and AlO4 connected to sheets of octahedrally (O) coordinated cations such as Al3+ or Mg2+ to form T-O or T-O-T layers. Depending on the layer charge, exchangeable interlayer cations can be present between the compound layers. Water is present as H2O in the interlayer and as (OH)- in the octahedral

Snellings, Mertens, Elsen

226

layer. During burial, clays compact to coherent mud rock and eventually shale when cleavage is developed. Clay minerals are the common product of chemical weathering of primary igneous minerals such as feldspars or form during low-temperature diagenetic alteration. Depending on the weathering conditions and the chemical composition of the altered material, various clay minerals form. Commonly encountered clay minerals are kaolinite, smectites, illite, chlorite and palygorskite-sepiolite. General chemical compositions of common clay minerals, lateritic soils and red mud waste are given in Table 2. Untreated clays show unsatisfactory pozzolanic activity and low technological performance. According to He et al. (2000), this is due to the stability of their crystal structures, their high specific surface requiring a high water to binder ratio for a desired workability, and the platy particle morphology and preferential cleavage. Non-pozzolanic micro- and nanoscale clay particles were observed to change the cement paste structure by serving as preferential substrates for C-S-H nucleation. Addition of smectite and palygorskite resulted in a more open, interconnected cement pore structure compared to the addition of silica fume (Lindgreen et al. 2008). Only kaolinite and smectite are reported to show good pozzolanic activity when fired at appropriate temperatures (Mielenz et al. 1950; Ambroise et al. 1985; He et al. 1995; Liebig and Althaus 1997). The pozzolanic activity of thermally activated clay minerals is closely linked with their behavior during heating. Optimal activation is achieved through amorphization of the clay mineral upon complete dehydroxylation of the octahedral sheet. Overheating results in particle agglomeration and crystallization of inactive high-temperature phases. The temperatures at which dehydration, dehydroxylation and recrystallization occur are determined by the clay mineral structure and composition. Most important in the formation of a pozzolanically active material is the temperature range of dehydroxylation which depends on the bonding of the hydroxyl groups in the octahedral sheet. The thermal strength of these bonds increases from Fe-OH over Al-OH to Mg-OH. The formation of recrystallization products depends mainly on the chemical composition of the raw materials. The commonly formed minerals are mullite, cordierite, enstatite and cristobalite (Emmerich 2010). The thermal behavior of the various common clay minerals is illustrated in Figure 13. The T-O clay mineral kaolinite, Al2Si2O5(OH)4, does not contain exchangeable cations or interlayer water. The temperature of dehydroxylation (endothermal) of kaolinite depends on the structural layer stacking order. Disordered kaolin dehydroxylates between 530 and 570 °C, ordered between 570 and 630 °C. Dehydroxylated disordered kaolinite shows higher pozzolanic activity than ordered (Kakali et al. 2001; Bich et al. 2009). Upon dehydroxylation kaolinite transforms into metakaolin, a complex amorphous structure which retains some longrange order due to layer stacking (Bellotto et al. 1995). Much of the aluminum contained in the octahedral layer becomes tetrahedrally and pentahedrally coordinated (Justnes et al. 1990; Rocha and Klinowski 1990; Fernandez et al. 2011). Upon further heating a defect Al-Si spinel and Table 2. Characteristic chemical compositions for a selection of clay minerals, a laterite soil and red mud waste. Material

SiO2 Al2O3 Fe2O3 TiO2

Kaolinite Illite Montmorillonite Palygorskite Chlorite Laterite Red mud

46.6 54.0 43.8 58.4 30.3 3 5.0

39.5 17.0 18.6 6.2 17.1 10 15

CaO MgO Na2O K2O

1.9

3.1 1.0

15.1 75 26.6

7.3 1.1

14.7 25.4 15.8

22.2

1.0

1.0

0.0

H2O

SUM

14.0 12.0 36.1 19.7 12.1 10 12.1

100.1 95.3 100.6 99.0 100.0 98.0 98.7

Supplementary Cementitious Materials

227

Figure 13. Representative differential thermal analysis (DTA) curves for common clay minerals and gibbsite (modified after Cook 1986b).

eventually mullite are formed. Reported optimal activation temperatures vary between 550 °C (He et al. 1994) and 850 °C (Ayub et al. 1988) for varying durations, however most values cluster around 650-750 °C (Murat 1983; Ambroise et al. 1987; De Silva and Glasser 1992; Kakali et al. 2001; Chakchouk et al. 2009). In comparison with other clay minerals kaolinite shows a broad temperature gap between dehydroxylation and recrystallization, this much favors the formation of metakaolin. Also, because the Al-OH groups of the kaolinite octahedral sheets are directly exposed to the interlayer, structural disorder can be attained more easily upon dehydroxylation of kaolinite than in T-O-T clay minerals (Fernandez et al. 2011). If the raw clay material is sufficiently pure, metakaolin is a highly active SCM which can be used in high-performance, high-strength binders to improve the compressive and flexural strength and increase durability and resistance to chemical attack (Siddique 2008). For low-cost applications kaolinite deposits or tropical soils of lower purity can be used (e.g., Cara et al. 2006; Kakali et al. 2001; Badogiannis et al. 2005). Highly active metakaolin can also be produced from paper sludge waste or oil sand fine tailings. Paper sludge waste is mainly composed of cellulose fibers, kaolinite and calcite. Calcination at 700 °C for 2-5 hours decomposes and volatilizes the organic material and activates the kaolinite while no calcite decarbonation occurs (Péra and Amrouz 1998; Frias et al. 2008). In some cases the clay fraction of oil and tar sands consists mainly of kaolinite. The separated fine tailings can be valorized by controlled firing and production of metakaolin (Wong et al. 2004). Alternatively, kaolinite can be activated by mechanochemical (Vizcayno et al. 2009) or acid treatment. Aluminum extraction from calcined kaolin by sulfuric acid results in a dealuminated kaolin with high pozzolanic activity (Mostafa et al. 2001).

228

Snellings, Mertens, Elsen

Smectites are a group of T-O-T clay minerals containing exchangeable cations and showing the remarkable capability to absorb water between the layers and swell. Isomorphous substitutions in the tetrahedral and octahedral sheets as well as the interlayer cation type influence the dehydration and dehydroxylation behavior. Dehydration occurs below 300 °C and is associated with the release of water bound to outer surfaces and coordinated to interlayer cations. Dehydroxylation temperature depends on the type of cations contained in the octahedral sheet (Fe < Al < Mg) and their arrangement. When the cations contained in the octahedral sheet are mainly trivalent (Al3+, Fe3+) then only two out of three cation positions are filled and the clay minerals are termed dioctahedral. The disposition of the hydroxyl groups with respect to the resulting vacancy defines two kinds of configurations: cis-vacant when both hydroxyls are located on one side of the vacancy and trans-vacant when they are located on either side (Tsipursky and Drits 1984). Dehydroxylation of smectites which show mainly cis-vacant octahedral sheets occurs at temperatures around 700 °C, 150-200 °C above the dehydroxylation of trans-vacant octahedral sheets. Smectites contain predominantly cis-vacancies while illite clay minerals are mostly trans-vacant (Drits et al. 1995). Recrystallization starts above 850 °C, depending on the chemical composition of the clay. The reported optimal activation temperature ranges for montmorillonite-rich clays are typically higher and narrower defined when compared to kaolinite: 800-830 °C for 1-5 hours (Mielenz et al. 1950; He et al. 1996; Liebig and Althaus 1997; Habert et al. 2009). Overheating to recrystallization temperatures is easily attained and is together with the variable chemical composition a complicating factor in thermal activation. Activated montmorillonite has been observed to be a good pozzolanic material, incorporation into lime or Portland cement based binders contributed significantly to the compressive strength (He et al. 1995; Liebig and Althaus 1997). Other clays are reported to be less pozzolanically active upon thermal activation. The dehydroxylation of illite, a T-O-T clay mineral with a composition intermediate between mica and smectite, occurs at relatively low temperatures (580 °C). However dehydroxylation does not result into a collapse of the structure into a largely amorphous state before recrystallization into spinel and corundum occurs (He et al. 1995). In consequence, the resulting calcined material shows much lower pozzolanic activity than that formed from less stable smectites or kaolinite. Mg-rich T-O-T clays of the palygorskite-sepiolite group contain both interlayer and zeolite water and show complex dehydration reactions. The release of zeolite water and adsorbed water proceeds simultaneously at low temperature (Fig. 13) and is succeeded by the release of interlayer water. Dehydroxylation of palygorskite occurs below 500 °C, sepiolite dehydroxylates at 820 °C, simultaneously with decomposition and recrystallization into enstatite and cristobalite. The pozzolanic activity was observed to be low (He et al. 1996). The T-O-T octahedral sheet in chlorite clay minerals tends to dehydroxylate at rather high temperatures due to the elevated Mg content. Recrystallization into spinel and olivine occurs at low temperatures leaving a narrow window for thermal activation. Natural clay deposits often contain more than one type of clay mineral. Habert et al. (2009) observed that ideal activation of all different clay minerals in a mixture is difficult, because recrystallization of previously activated clay can occur when a second one is activated. Furthermore they found that optimal activation and recrystallization temperatures were lowered in the impure mixtures with respect to the purified materials. In tropical climates, intense chemical weathering leaches out silica, alkaline earths and alkalis from soils to leave a residue of aluminum and ferric (hydr)oxides. The resulting soils are bauxites and laterites, respectively. Calcination of laterites at 750 °C has been reported to produce a pozzolanically active material which improved the compressive strength of blended cements (Péra et al. 1998). Red mud waste obtained from the extraction of aluminum from bauxites can be transformed into a pozzolanically active material by calcination in the range

Supplementary Cementitious Materials

229

of 600-800 °C (Péra et al. 1997). Dehydroxylated aluminum and ferric hydroxides tend to produce cementitious reaction products when contacted with lime. Oil shales are fine-grained sedimentary rocks containing 5-65% of organic material and detrital minerals such as clay minerals, quartz and calcite. If the level of organic material is sufficient, oil shales can be combusted and utilized for electricity production or as alternative raw material in the production of Portland cement. Depending on the type of organic material, the oil can also be extracted by pyrolysis. The resulting ashes can show hydraulic or pozzolanic activity. In ashes abundant in calcium, hydraulic calcium silicates (b-C2S) and aluminates were observed to form (Smadi and Haddad 2003). The pozzolanic activity of the low calcium ashes is strongly linked to the original mineralogy of the shale and the burning conditions (Ish-Shalom et al. 1980). Inactive phases such as quartz and mica can represent an important fraction of the ash and lower the pozzolanic activity (Feng et al. 1997). Although waste ceramic material has been utilized intensively as a pozzolanic additive in the past, contemporary ceramic waste does not find the same application. Modern ceramic production usually involves controlled firing at elevated temperatures between 900-1050 °C, which are invariably above the optimal activation temperature of the original clay minerals (Baronio and Binda 1997). In most cases significant high-temperature recrystallization occurs and in consequence the pozzolanic activity is relatively low, resulting in lowered compressive strength values upon increasing replacement of Portland cement (Wild et al. 1997; Lavat et al. 2009). Burned organic matter residues. Ashes of burned agricultural residues can be used as an inexpensive alternative material for partial replacement of Portland cement or lime. While a large proportion of agricultural residues are recycled for fertilization or stock feed purposes, some is utilized as fuel or left as waste. Firing leads to the decomposition of the organic carbon and a concentration of the silica in the residue. Under controlled firing conditions a highly active pozzolan material can be produced (Mehta 1977). The elevated specific surface area and considerable surface roughness are often associated with a high water demand of the ashes and necessitate the use of superplasticizers when the replacement percentage of optimally fired ashes is more than 10-20%. The yield of production of a pozzolanic material from organic residues is determined by two factors: the inorganic residue after ignition or ash content and the silica content of the inorganic residue. Typical inorganic residue fractions and silica contents are enlisted in Table 3 for common agricultural residues. The greatest yield can be expected for the calcination of rice husks. Upon firing, a highly siliceous material with a very large surface area can be obtained. Some of the enlisted materials such as corn or rice straw face strong competition from the utilization as stock feed (Cook 1986c).

Table 3. An overview of ash content of plant parts and silica content of the ash (modified after Cook 1986c). Plant Sorghum Wheat Corn Sugar cane Sugar cane Sunflower Rice Rice Palm

Part of plant

Ash (%)

Silica (%)

Leaf sheath Leaf sheath Leaf blade Bagasse Straw Leaf and stem Husk Straw Fibers and shells

12.55 10.48 12.15 14.71

88.7 90.56 64.32 73 62 25.32 93 82 65

11.53 22.15 14.65

230

Snellings, Mertens, Elsen

A ternary plot of reported chemical compositions of various organic material ashes in Figure 14 demonstrates the highly siliceous nature of rice husk ash compared to other residues. The level of soil derived impurities can often reach significant levels and can become concentrated in the inorganic residue upon firing. This can lead to a dilution of the active components in the ash (Cordeiro et al. 2008). Also soil and climate conditions can influence the ash and silica contents of the residue (Biricik et al. 1999). The pozzolanic activity of the inorganic residue is strongly affected by the firing temperature. Optimal activation is achieved when the cellulose and other combustibles are removed and the pore structure and associated large surface area (25-40 m²/g) of the silica-rich skeleton are preserved (Jauberthie et al. 2000) (see Fig. 15). The dissolution of the silica is strongly linked with the concentration of silanol groups at the surface (Nair et al. 2008). Over-heating leads to reductions in specific surface and to transformation of the amorphous silica to crystalline hightemperature silica polymorphs such as cristobalite and tridymite. Firing should therefore be performed in an oxidizing atmosphere at temperatures above 400 °C to decompose the organic material (James and Rao 1986) but, for rice husk ash, below 700 °C to avoid the formation of cristobalite and tridymite (Hamdan et al. 1997). Prolonged exposure to elevated temperatures leads to a collapse of the pore structure and to decreased recrystallization temperatures (Cook 1986c). The sensitivity of the activated material to the firing conditions and the utilization of agricultural residues as fuel for domestic or industrial purposes have been primary obstacles to the widespread application of organic material ashes as a pozzolanic material. Therefore, recently, research focus shifted to the beneficiation of ashes of organic residues used as fuel. The sensitivity of the activated material to the firing conditions and the utilization of agricultural residues as fuel for domestic or industrial purposes have been primary obstacles to the widespread application of organic material ashes as a pozzolanic material. Therefore, recently, research focus shifted to the beneficiation of ashes of organic residues used as fuel. Sugar cane bagasse and straw ashes resulting from uncontrolled combustion in boilers of

Figure 14. CaO-SiO2-Al2O3 ternary diagram (wt% based) of chemical composition of reported organic residue ashes.

Supplementary Cementitious Materials

231

Figure 15. SEM picture of fragments of the highly-porous silica-rich skeletons of rice husks after uncontrolled burning in a field furnace.

sugar and alcohol factories at 700-900 °C were found to behave pozzolanically when finely ground (Cordeiro et al. 2008, 2009a; Chusilp et al. 2009; Morales et al. 2009). Also the ashes of uncontrolled firing of rice husks (De Sensale 2006; Cordeiro et al. 2009b), rice husks and eucalyptus bark (Tangchirapat et al. 2008) and palm oil residue (Tangchirapat et al. 2009) show pozzolanic activity when finely ground.

By-product SCMs The valorization of industrial and societal waste in construction materials is one of the main routes for progress in increasing the sustainability of present-day society. A large diversity of waste materials can be considered and applications can range from low-value products such as aggregates to high-value products such as some SCMs. In this section, the properties of the most widely used industrial by-product SCMs are reviewed. The widespread success of blast furnace slags, coal and lignite fly ash and silica fume can serve to corroborate the view that construction materials will increasingly become a primary target for disposal of industrial and societal waste. The extensive experience and knowledge built up over years of using these established by-product SCMs should be considered as valuable when proceeding into a future expansion of the range of waste materials suitable for construction purposes. Blast furnace slags. The most widely utilized slags in cements are obtained as a byproduct in the extraction of pig iron in blast furnaces. Blast furnace slags commonly present latent hydraulic behavior if they are quenched sufficiently rapidly as a vitreous phase from the melt at 1350-1550 °C to below 800 °C and finely ground. The first hydraulic cements incorporating blast furnace slags were pioneered in the 2nd half of the 19th century in Germany, first as Ca(OH)2-slag blends, later on as SCM in Portland cement. Nowadays, blast furnace slags are mostly used in combination with Portland cement and, being latent hydraulic, can constitute a much larger proportion of the blended cement than pozzolanic materials. In addition, blast furnace slags present marked cementitious properties when activated by either lime, alkali hydroxides, sodium carbonates or sodium silicates (water glass) and calcium or magnesium sulfates (e.g., supersulfated slag cements). Over recent years the activation of various types of slags has received much attention as potential alternative for Portland cement, a detailed

232

Snellings, Mertens, Elsen

account of the progress made is however out of the scope of this chapter. Topical reviews on blast furnace slags in cements are available (Smolczyk 1980; Regourd 1986, 2001; Lang 2002). The chemical composition of a slag varies widely depending on the composition of the raw materials in the iron production process. Silicate and aluminate impurities from the ore and coke are combined in the blast furnace with a flux which lowers the viscosity of the slag. In the case of pig iron production the flux consists mostly of a mixture of limestone and forsterite or in some cases of dolomite, the latter for reasons of economy. In the blast furnace the slag floats on top of the iron and is decanted for separation. Slow cooling of slag melts results in an unreactive crystalline material consisting of an assemblage of Ca-Al-Mg silicates. To obtain the slag hydraulicity, the slag melt needs to be rapidly cooled or quenched below 800 °C in order to prevent the crystallization of merwinite and melilite (Regourd 1986). To cool and fragment the slag a granulation process can be applied in which the molten slag is subjected to jet streams of water or air under pressure. Alternatively, in the pelletization process the liquid slag is partially cooled with water and subsequently projected into the air by a rotating drum. Pumiceous, porous pellets are produced in which the mostly glassy fraction finer than 4 mm can be used as a supplementary cementitious material. The coarser, typically more crystalline fraction can be used as lightweight aggregate. The pelletization process is more economical in terms of water consumption during the cooling process and energy needed to dry the slurry of quenched slag fragments but is usually less effective in obtaining high glass contents (Taylor 1990) compared to the granulation process. In order to obtain a suitable reactivity, the slag fragments are ground to reach the same fineness as the Portland cement. The main components of blast furnace slags are CaO (30-50%), SiO2 (28-38%), Al2O3 (824%) and MgO (1-18%) (Regourd 2001). In a CaO-SiO2-Al2O3 ternary diagram (Fig 16, Fig. 3) the blast furnace slags can be situated somewhere in between typically pozzolanic materials such as natural SCMs or silica fume and Portland cements, indicating the “latent hydraulicity” of the slags. In general, increasing the CaO content results in raised slag basicity and an

Figure 16. Ternary diagram (wt% based) of the compositional distribution of a selection of blast furnace slags.

Supplementary Cementitious Materials

233

increase in compressive strength of the binder, the MgO and Al2O3 contents show the same effect up to respectively 10-12% and 14% beyond which no further improvement can be observed. Also higher levels of alkali lead to strength improvement, especially at early ages. Other minor components such as MnO and TiO2 exert a negative effect on the hydraulic character of the slag. Several compositional ratios or so-called hydraulic indices have been used to correlate slag composition with hydraulic activity; the latter being mostly expressed as the binder compressive strength. Normative compositional requirements have also been formulated based on a number of hydraulic indices, for instance the EN-197-1 norm requires a (CaO+MgO)/ SiO2 larger than 1, the German and Japanese standard demand a (CaO+MgO+Al2O3)/SiO2 larger than respectively 1 and 1.4. However, the relationship between the hydraulic indices and compressive strength are very roughly defined and do not allow an accurate prediction of compressive strength solely based on the slag chemical composition for a series of slags originating from different production sites (Lang 2002). Other parameters which have been often included in multiple regression analyses for strength prediction are the slag fineness and glass content (e.g., Douglas et al. 1990; Escalante et al. 2001; Pal et al. 2003; Bougara et al. 2010). The glass content of slags suitable for blending with Portland cement typically varies between 90-100% and depends on the cooling method and the temperature at which cooling is initiated. The glass structure of the quenched slag largely depends on the proportions of networkforming elements such as Si and Al over network-modifiers such as Ca, Mg and to a lesser extent Al. The network-forming atoms are tetrahedrally coordinated by oxygen atoms and show a varying degree of polymerization or connectivity depending on the ratio of network-forming to network-modifying elements. Increased amounts of network modifiers lead to higher degrees of network depolymerization and reactivity (Goto et al. 2007). The rate of cooling influences the amount of structural defects in the glass phase, the higher the cooling rate, the more defects and the higher the reactivity. The presence of small amounts of finely dispersed crystalline material has been observed to improve the reactivity of the slag (Demoulian et al. 1980; Frearson and Uren 1986), especially when dendritic crystallization of merwinite (Ca3Mg(SiO4)2) occurs. The Al-enrichment of the glass near the merwinite crystallites and the mechanical stress and presence of nucleation sites introduced by the phase separation are suggested to enhance the slag reactivity. Another common crystalline constituent of blast furnace slags is melilite, a solid solution between gehlenite (Ca2Al2SiO7) and åkermanite (Ca2MgSi2O7). The CaO-SiO2-Al2O3 phase diagram at 10% MgO (Fig. 17) shows that most blast furnace slags will initially crystallize melilite, other minor components which might form during progressive crystallization are belite (Ca2SiO4), monticellite (CaMgSiO4), rankinite (Ca3Si2O7), (pseudo-) wollastonite (CaSiO3) and forsterite (Mg2SiO4). The a′ and b-polymorphs of belite and possibly also melilite show hydraulic activity and can contribute to the compressive strength. Minor amounts of reduced sulfur are commonly encountered as oldhamite (CaS) (Scott et al. 1986). The reducing environment that blast furnace cements can provide are often advantageously used for waste stabilization (Roy 2009). Deleterious free lime (CaO) or periclase (MgO) are usually not present in blast furnace slags. Fly ash. Coal fly ash is a gigascale material (Scheetz and Earle 1998). Over one billion tons of by-products are generated annually during the combustion of coal (Kutchko and Kim 2006). These by-products include mainly dry bottom ash, wet bottom boiler slag, economizer ash, fly ash and flue gas desulphurization or scrubber sludge. Of these by-products, the annual production of fly ash is estimated around 500 millions of tons (Joshi and Lothia 1997). Coal is not exclusively composed of organic matter, which produces the energy during coal firing in power plants. It also contains a variable amount of inorganic material. This intermixed inorganic material may remain unaffected or will be transformed during combustion. It will then be concentrated in the by-products. Besides coal fly ash, other types of fly ash; for instance

234

Snellings, Mertens, Elsen

Figure 17. Phase diagram of the CaO-SiO2-Al2O3 system (wt% based) at 10 wt% MgO, the blast furnace chemical composition distribution largely falls into the melilite field (modified after Satarin 1974).

Municipal Solid Waste Incineration (MSWI) fly ash exist. However, coal fly ash is by far the most widely used and investigated type of fly ash. Therefore most attention will go to coal fly ash in what follows. It will simply be designated as “fly ash” hereafter. Much of the information in the following paragraphs is taken from the valuable work of Joshi and Lothia (1997) and Sear (2001) who made extensive reviews of the production, the use and the properties of fly ash. From all coal combustion products (CCP), coal fly ash (CFA), also designated as fly ash (FA) or pulverized fuel ash (PFA) is most widely used as a supplementary cementitious material. Fly ash is produced when C-rich sediments are burned at temperatures reaching 1450 °C or higher. Whereas the carbon-based materials are mainly transformed to gaseous compounds as CO2, H2O, NOX and SO2, most of the minerals that are present in the burned sediments will not be volatilized. Instead, they may undergo various chemical, physical and mineralogical changes. Whereas some of the primary minerals remain unaffected during the burning process, others will melt, become amorphous and/or recrystallize to form secondary minerals. These materials are then recovered at the bottom of the furnace/boiler or from the flue gases by electrostatic or mechanical precipitators or bag houses (Joshi and Lothia 1997). Fly ash is the material recaptured from the exhaust gases of power plants using pulverized coal as a combustion product. Fly ash is generally subdivided in categories following the national or international standards. It is a versatile, heterogeneous material that is extensively researched regarding all its intrinsic properties, its use as a supplementary cementitious material or for multiple other applications. Since the beginning of the 20th century, fly ash has been recognized as a pozzolanic constituent (Joshi and Lothia 1997). Initially, only the Ca-poor fly ashes resulting from the burning of bituminous coal were considered as supplementary cementitious materials. However, from about the middle of the 20th century, also the more Ca-rich fly ashes were considered as a cement replacement material. The actual use of fly ash dates back to more than 50 years ago (Sear et al. 2003). Besides scientific research dealing with fly ash characterization, many studies were done on the various applications of fly ash. A considerable portion of this research focuses on its use as a supplementary cementitious material in cement or concrete, even though other uses increasingly attract attention. Most of the scientific work focuses on the physical/

Supplementary Cementitious Materials

235

mechanical properties of cement/fly ash mixtures (e.g.: Papadakis 1999; Papadakis 2000; Maltais and Marchand 1997; Erdoğdu and Türker 1998; Grzeszczyk and Lipowski 1997; Payá et al. 1997), investigates the durability of fly ash use (e.g.: Bijen 1996; Lilkov et al. 1997; Papadakis 2000; Roy et al. 2001), or studies the hydration reactions more in general (e.g., Sharma and Pandey 1999; Lilkov et al. 1997; Papadakis 1999; Maltais and Marchand 1997). Although the suitability of some types of fly ash may still be a matter of debate (Papadakis 2000; Sear et al. 2003), the use of fly ash in concrete is generally considered as beneficial from an ecological, economical and technological point of view. In addition to the use as a SCM in traditional calcium silicate or aluminate hydrate binders, fly ash has many other applications. Amounts of 5-15% are generally cited as the proportion of fly ash used in concrete compared to the global amount of fly ash produced. Some European countries consume up to 100% of their production (Queralt et al. 1997), whereas other countries as Israel seem to have already reached a much lower maximum for the addition of fly ash to cement clinker or concrete (Nathan et al. 1999). Mainly in Europe, other applications account for the use of increasingly large amounts of this secondary raw material. Due to its bulk mineralogy and chemistry, it can serve as a source of raw materials for large-volume, low-tech applications (Scheetz and Earle 1998). As fly ash is a complicated heterogeneous material (Vempatie et al. 1994) of variable quality (Sakai et al. 2005), a detailed knowledge of the physical and chemical characteristics is generally required (Vempatie et al. 1994) before use. The numerous cementitious applications of fly ash include grouts, block manufacture and road sub-base and base construction (Sear et al. 2003). High proportions of fly ash can also be incorporated in dams, walls, girders, roller-compacted concrete pavements and parking areas (Manz 1998; Sakai et al. 2005). However, large amounts of fly ash can also be used for land reclamation (Mondragon et al. 1990; Sear et al. 2003). Moreover, increasing efforts are done to use the material for technologically valuable applications as for the manufacture of monolithic ceramics (Mondragon et al. 1990; Queralt et al. 1997; Ilic et al. 2003) or for alkali-activated materials (Puertas and Fernández,-Jiménez 2003; Bakharev 2005) with a significant compressive strength. Other high-tech applications include the production of zeolites (Mondragon et al. 1990; Murayama et al. 2002) or its use for phosphate immobilization from agricultural activities (Grubb et al. 2000). Instead of using fly ash as a bulk resource, some authors (Cheriaf et al. 1999; Vassilev et al. 2003) suggest to separate several fractions for different uses. The ultrafine fraction (0.1-1 mm) could for instance be used in high performance concretes (Sear 2001), cenospheres as a strong lightweight inert filler material (Sear 2001) or for specific aerospace applications (Mondragon et al. 1990). The magnetic fraction for instance could be competitive as ferro-pozzolan for the production of dense concretes (Vassilev et al. 2004). Bottom ash, another secondary raw material formed during coal combustion is also used in concrete as a low-cost replacement material for sand or as a base in road construction. Bottom ash is also used as a fertilizer (Cheriaf et al. 1999). Future research will be required to ensure the continuous proper application of coal combustion waste products as fly ash, considering for instance a changing fly ash composition. The types of fuel used in the future will change and diversify and blends of fuels will be used (Steenari and Lindqvist 1999; Koukouzas et al. 2007). Thereby, the nature of the fly ash will inevitably be affected, requiring an adapted utilization. Moreover, other coal combustion by-products such as bottom ash will probably gain more interest in the future, as they are potentially appropriate as secondary raw material for various applications. Other types of fly ash such as MSWI fly ash could similarly be used more often in the future (Kirby and Rimstidt 1993; Ferreira et al. 2003). The quality of the fly ash varies widely (Sakai et al. 2005). Depending on the burning temperature, the coal type, the processing and many other factors, fly ash exhibits different physical, chemical and mineralogical properties (Erdoğdu and Türker 1998; Vassilev et al. 2003). These will in turn affect the properties of the concrete or other products in which they are used. The first standards were introduced to classify fly ashes in order to reduce this variability

236

Snellings, Mertens, Elsen

for fly ash users. In 1953-1954, the ASTM standard C350 for fly ash as an admixture was introduced (Malhotra 1993; Manz 1998). In 1960 it was extended to the use of fly ash as a pozzolan. The ASTM standard C618 was introduced in 1968 and covers both natural pozzolans and fly ashes for their use as mineral admixtures in Portland cement concrete. A first British standard BS 3892 was introduced in 1965 in which fly ash was treated as a fine aggregate with three classes of fineness for use in concrete. It was revised afterwards. In 1995, a common specification for EU countries, EN 450, was introduced for fly ash used in concrete. Particular for this common European norm compared to the BS 3892, is the definition of the Activity Index, imposing a minimum strength for fly ash/Portland Cement mixtures. Also the ASTM standard, C618 requires both fly ash/Portland cement and fly ash/lime mixtures to achieve a minimum strength after 28 or 7 days respectively. The ASTM C618 defines three categories of mineral admixtures; Class N, Class F and Class C. Class N includes mainly the group of natural pozzolans. Class F fly ash is a siliceous type of fly ash mainly obtained from the combustion of bituminous or hard coals. The main part of the fly ash used in concrete is of this type. Class C fly ash is richer in calcium and results primarily from the combustion of lignite or brown coal. Whereas the classification in ASTM C618 of mineral admixtures in Portland cement concrete derives from a genetic differentiation, some authors (Mehta 1983, 1989; Joshi and Lothia 1995, 1997; Manz 1998) repeatedly suggested making a distinction based on the properties of the admixtures. Joshi and Lothia therefore proposed to define one class of “pozzolanic but non-self cementitious” materials and another class of “pozzolanic and self cementitious” or “hydraulic” materials. The two classes would in that case be distinguished by their difference in loss on ignition. It is generally believed that there is a close relationship between the properties of supplementary cementitious materials (as fly ashes) and their mineralogy (Vassilev and Vassileva 1996; Manz 1998). The mineralogy of fly ashes is in general not considered in the norms. This is due to a lack of quantitative analytical data (Ward and French 2006) and/or knowledge of mineralogical analysis techniques (Manz 1998). However, reliable methodologies for mineral quantification have been developed (Winborn et al. 2000) and can be used to establish correlations with the Activity Index and other relevant parameters. It is therefore reassuring that an increasing number of studies systematically use quantitative mineralogical information for interpreting data. Pulverized Fuel Ash may also be classified as “low-lime” and “high-lime” fly ashes (Dhir 1986). This classification roughly corresponds to the ASTM class F and C fly ashes respectively. An alternative classification was proposed by Roy et al. (1981) and has been adopted by others (Goodarzi 2006). Three major groups of oxides are defined; SiO2 + Al2O3 + TiO2 (sialic); CaO + MgO + Na2O + K2O + (BaO) (calcic) and Fe2O3 + MnO + P2O5 + SO3 (ferric). The fly ash compositions are plotted in a ternary diagram where the three major classes sialic, calcic and ferric, are defined next to the intermediate classes ferrocalsialic, ferrosialic, calsialic and ferrocalcic. Fly ash is composed of mainly spherical particles ranging in size from less than 1 to about 300 mm (Fig. 18). These particles form upon rapid cooling of droplets of viscous, molten or even vaporized mineral matter that was initially present in the combustion product. The major consequence of the rapid cooling is that only few minerals will have time to crystallize and that mainly amorphous, quenched glass remains. Nevertheless, some refractory minerals in the pulverized coal will not melt (entirely) and remain crystalline. In the following sections, the physical, chemical and mineralogical properties of fly ashes in general and of their discrete constituents in particular will be discussed. The average specific gravity of fly ash is estimated around 2.2 with a standard deviation of about 0.3. The BET specific surface area may range from less than 0.5 to more than 10 m²/g.

Supplementary Cementitious Materials

237

Figure 18. SEM picture of a typical fly ash with its mostly spherical and some angular particles.

Average Blaine surface areas are in the order of 0.35 m²/g and are relatively less divergent compared to tabulated BET surface areas. When pulverized, the specific gravity of the fly ash increases to about 2.7 as hollow particles break up (Dhir 1986). There also seems to be a direct relation between specific gravity of the bulk fly ash and its mineralogy. Higher quartz and mullite contents account for a lower specific gravity (Joshi and Lothia 1997). The particle size distribution appears to be strongly dependent on the method of collection (Dhir 1986). Fly ashes collected from electrostatic precipitators are two to five times finer compared to fly ashes from mechanical separators. The former type is therefore more generally used as a SCM. The fineness of fly ash is often expressed as the fraction passing a 45 mm wet sieve. It also serves as a routine measure for quality control and to assure uniformity in fly ash supply. A substantial correlation exists between the fineness of the fly ash (% retained on a 45 mm sieve) and its Activity Index (Dhir 1986). However, fly ash is a heterogeneous material. Physically distinct particles can be distinguished. 1 to 2 wt% (Sear 2001) or 15 to 20 vol% (Vassilev and Vassileva 1996) of the fly ash consist of hollow spherical particles known as cenospheres. Cenospheres have diameters ranging from 50 to 200 mm, with a wall thickness of approximately 10% of their radius (Sear 2001). Cenospheres form when trapped organics, carbonates, sulfides, sulfates or hydrosilicates decompose or water evaporates and induces an expansion while the particle is still in a viscous state (Kolay and Singh 2001; Kutchko and Kim 2006). They appear to be more characteristic for fly ashes obtained from coals enriched in finely dispersed illite and quartz (Vassilev and Vassileva 1996). Their density is very low and ranges from 250 to 800 kg/m³. If the fly ash is stored in a lagoon, the cenospheres will float at the surface. The Blaine surface area of cenospheres collected from such a lagoon in Dahuna, India, is about 0.05 m²/g (Kolay and Singh 2001) and thereby much lower than the surface area of the bulk fly ash. This is explained by their nearly perfect spherical shape and their hollow structure. Because of their particular properties, cenospheres are used for specific applications as in lightweight constructions. As cenospheres are hollow, they have good isolating properties. Plerospheres are like cenospheres, but instead of being empty, they contain smaller spherical or other particles. Dermaspheres are defined as plerospheres that have crystal nuclei of mullite, hematite and other minerals,

238

Snellings, Mertens, Elsen

covered with amorphous aluminosilicate envelopes (Vassilev and Vassileva 1996). Ferrospheres are spherical particles enriched in iron, be it amorphous or crystalline. Spheroïds are a specific type of spheres that look more porous or vesicular particles, with sizes ranging from 10 to 80 mm (Vassilev and Vassileva 1996). Ramsden and Shibaoka (1982) made their own classification of fly ash particles and defined seven categories; (1) unfused detrital minerals, (2) irregularspongy particles, (3) vesicular colorless glass, (4) solid glass, (5) dendritic iron oxide particles, (6) crystalline iron oxide particles and (7) unburned char particles. Like in most other supplementary cementitious materials, SiO2, Al2O3, Fe2O3 and occasionally CaO are the main components present in fly ashes. A selection of fly ash compositions is plotted in Figure 19. Following the ASTM classification, a main distinction can be made between Class F and Class C fly ashes. The relatively more CaO-rich Class C fly ashes plot further away from the SiO2-Al2O3 border of the diagram compared to the Class F fly ashes. The ASTM standard requires the sum of SiO2, Al2O3 and Fe2O3 to be greater than 70% for Class F and greater than 50% for Class C fly ashes. The mineralogy of fly ashes is very diverse. The main phases encountered are a glass phase, together with quartz, mullite and the iron oxides hematite, magnetite and/or maghemite. Other phases often identified are among others; cristobalite, anhydrite, free lime, periclase, calcite, sylvite, halite, portlandite, rutile and anatase. The Ca-bearing minerals anorthite (feldspar), gehlenite, åkermanite and various calcium silicates and calcium aluminates identical to those found in cement clinker can be identified in Ca-rich fly ashes. These, mainly type C, fly ashes may have hydraulic properties, as they contain minerals that react with water to form calciumsilicate/aluminate hydrates with binding properties. A similar mineralogical composition to that of coal fly ash is found in Municipal Solid Waste Incinerator ash (Kirby and Rimstidt 1993). For fly ashes in general, the Ca-minerals, free lime, periclase and sulfate minerals are probably most critical towards their properties (Sear et al. 2003). There appears to be a strong relation between the mineralogy of the fly ashes and the mineralogy of the feed coals (Ward and French

Figure 19. The reported chemical compositions of fly ashes plotted into a CaO-SiO2-Al2O3 ternary diagram (wt% based).

Supplementary Cementitious Materials

239

2006; Koukouzas et al. 2007) as well as their combustion temperature (Koukouzas et al. 2007). This is also true for alternative fuel sources. Wood chips, being rich in Ca, will generate Ca-rich fly ashes, whereas the combustion of biomass rich in alkalis will generate fly ashes containing minerals alkali-bearing minerals. Three large groups of components are identified in fly ashes; (1) an organic char fraction, (2) an inorganic amorphous and (3) an inorganic crystalline fraction. 1) Char particles are concentrated in the larger grain size fractions (Kutchko and Kim 2006). The char fraction mainly consists of carbon and correlates with the Loss on Ignition (LOI). Whereas the relative abundances of the main elements in the fly ash are chiefly dependent on the source of the pulverized fuel, the LOI is strongly dependent on the burning process. During the “boosting period,” i.e., at the start-up of the power plant, an increase of the LOI is generally observed (Sear 2001). About 90% of this LOI is due to unoxidized elemental carbon. Older power stations also tend to yield high LOI fly ashes. The same goes for modern low NOx burners. In general, lower temperatures correspond to lower NOx values, but higher unburned carbon. LOI values may range from less than 1 to more than 20%, although values of maximum 6 or 7% are accepted by most standards for fly ash in Portland cement concrete. 2) The glass or inorganic amorphous phase in fly ashes may represent up to 90% of the total weight. An average value is probably in the order of about 60 to 80 wt%. Its quantity can be accurately determined from Quantitative X-Ray Diffraction (QXRD) measurements using an internal standard. Microscopic methods are often not suitable for quantifying fly ashes as the glass and crystalline phases are generally intimately intermixed (Ward and French 2006). The glass phase is principally composed of silica and alumina, although many other constituents are present. The silica is present as cross-linked tetrahedra, thereby showing a short- but no long-range ordering (Bijen 1996). The basicity of the glass phase can be calculated through the same formula as that used for blast furnace slag employed in German and Japanese concrete (Sakai et al. 2005): basicity =



( CaO + MgO + Al2O3 ) SiO2

(1)

Sakai et al. (2005) found that when the glass content of the fly ash is low, its basicity tends to decrease. Moreover, there appears to be an inverse correlation between the mullite content and the amount of amorphous material in fly ashes (Sakai et al. 2005). Obviously, higher mullite contents result in lower quantities of amorphous material with a lower average basicity. Compared to granulated blast furnace slags, fly ashes have low basicities. Some authors (Bijen 1996; Manz 1998; Ward and French 2006) consider the amorphous part as the “active part.” It is likely that the amorphous fraction is not composed of a single glass phase, but that it consists of different phases with a dissimilar composition (Nathan et al. 1999).

3) A tremendous work was done by Vassilev and Vassileva (1996) and Vassilev et al. (2003), who discussed the origin and occurrence of all minerals and groups of minerals found in 11 Bulgarian fly ashes. Their data include an exhaustive list of minerals, which is relevant for all studies on fly ash mineralogy. Individual fly ash particles, with the exception of plerospheres, are chemically fairly homogeneous (Gieré et al. 2003). Nevertheless, bulk fly ash is heterogeneous and differences in chemistry are observed between size fractions and between the categories of fly ash particles defined earlier. Cenospheres for instance are more silica-, alumina- and potassium-rich and poorer in calcium compared to the bulk fly ash (Vassilev and Vassileva 1996; Sear 2001). Ferrospheres are obviously rich in Fe and Fe-bearing minerals. Element partitioning in fly ashes induces important differences in chemistry and mineralogy among the

240

Snellings, Mertens, Elsen size fractions. Many comprehensive papers deal with this partitioning (Filipidis and Georgakopulos 1992; Vassilev and Vassileva 1996; Erdoğdu and Türker 1998; Hower et al. 1999; Gieré et al. 2003; Vassilev et al. 2004; Chen et al. 2005) and the environmental/toxicological issues related with the finest fly ash particles (Coles et al. 1979; Tazaki et al. 1989). Magnetic or other separable fractions have also been studied separately (Hower et al. 1999; Vassilev et al. 2004).

The majority of trace elements in fly ash is found in the fines (Vassilev and Vassileva 1996). Moreover, there is a clear correlation with the mineralogy, as accessory minerals are mainly identified in the smaller grain size fractions. In these smaller fractions, trace elements may form discrete mineral phases (Vassilev and Vassileva 1996). Crystalline particles as small as 20 nm have been observed by Transmission Electron Microscopy (TEM) (Chen et al. 2005). Toxic elements that are volatilized during the combustion can also be absorbed on the surface of very small particles (Gieré et al. 2003). More in particular, glass particles tend to attract many trace elements, due to their reactive surface (Dudas and Warren 1987). It has been mentioned that the composition of the glass fraction is probably not uniform throughout the fly ash. The same is true for the crystalline components. In a TEM-study dedicated to the composition of the mullite phase, Gomes and François (2000) discovered that its composition is very heterogeneous. However, from the determination of the lattice parameters by X-ray diffraction, an average mullite composition can be obtained (Cameron 1977). Similarly, the composition of magnetite can also vary between individual fly ash particles (Gomes et al. 1999). Silica fume. Silica fume is a by-product of the silicon metal and ferro-silicon alloy industries, the terms “condensed silica fume” and “microsilica” are also used. It is an excellent supplementary cementitious material with a high pozzolanic activity due to a high content of SiO2 in amorphous form and a very fine particle size distribution (0.1-0.2 mm average diameter). Major reviews on silica fume and its applications in concrete can be found in Mehta (1986), Kjellsen et al. (1999), Fidjestol and Lewis (2001), Justnes (2002), Chung (2002) and in Justnes (2007). Silica fume is produced during the reduction of quartz at high temperatures in electric arc furnaces. High purity quartz is heated to 2000 °C together with coal, coke or wooden chips to remove the oxygen. One of the reactions involves the formation of SiO vapor which oxidizes and condenses in the form of very small amorphous silica spheres. The first experiments on the use of silica fume in concrete were carried out at the Norwegian Institute of Technology in Trondheim (1950) but extensive research started only in the 1970’s and widespread commercial use of silica fume in concrete started in the 1980’s. Important production countries at this moment are China, Norway, South Africa, USA, Canada, Spain, Russia and France. Unlike other thermally activated supplementary cementitious materials such as for example fly ash, silica fume from one production source has nearly no variation in chemical composition over time, because of the use of relatively pure raw materials. The SiO2 content varies with the silica content of the alloy being produced and should be higher than 85% for silica fumes suitable for use as pozzolan (ASTM C 1240). In general, the chemical composition of silica fume is not complex and consists usually of more than 90% of SiO2 (85-99%). Other oxides such as Fe2O3, Al2O3, CaO, MgO, Na2O and K2O are normally below 1.0%, the value for the loss on ignition varies between 1.0% and 2.0%. Silica fume consists essentially of an amorphous silica structure and a very large peak can be observed centered at about 4.4 Å using X-ray powder diffraction. The silica fume particles are spherical in shape and using electron microscopy it was shown that they have an average diameter size between 0.1 and 0.2 mm (Fig. 20). Silica fume has an approximate value of 2.2 g/ cm³ for the specific gravity and a very high surface area value of about 20-22 m²/g as measured with BET N2-adsorption.

Supplementary Cementitious Materials

241

Figure 20. SEM picture of a typical silica fume with average particle diameter of 0.1 mm. [Used by permission of Elsevier, from Jo et al. (2007), Construction and Building Materials, Vol. 21, Fig. 1, p. 1352.]

THE POZZOLANIC REACTION The recombination of an (alumino-)silicate or aluminate material and Ca2+ or Ca(OH)2 in the presence of water into hydrated reaction products with binding properties can be schematically formulated in cement chemistry notation (A = Al2O3, C = CaO, H = H2O, S = SiO2; hyphenation denotes variable stoichiometry) as:

AS + CH + H → C-S-H + C-A-H (2)

The driving force behind this simplified reaction is the difference in Gibbs energy between the reactants and the eventual products. The reaction rate is however, determined by the individual elementary steps or processes in the reaction. The reaction step with the slowest rate of conversion is the rate-controlling process and is typically the one with the highest activation energy barrier. It is generally accepted that the initial rate-controlling process consists of the release or dissolution of silica from the pozzolan. The increasing pore solution saturation degree eventually gives rise to heterogeneous nucleation and growth of the C-S-H reaction products at the pozzolan surface. Subsequent to the formation of a layer of reaction products enveloping the reactant grains, the reaction rate is commonly assumed to be limited by the diffusion of ions through the growing and densifying layer of products (Kondo et al. 1976; Držaj et al. 1978; Takemoto and Uchikawa 1980). Most published reviews on supplementary cementitious materials do not treat the subject of the mechanism of the pozzolanic reaction in detail, but tend to focus on material properties defining the activity and performance of pozzolans. However, difficulties are met when trying to compare and relate the activity controlling properties among SCMs of different origin (e.g., Mehta 1987; Sersale 1993). In this respect, empirically established relationships between pozzolan properties and activity remain limited to materials of a similar origin. To determine which material properties are of importance and when they become essential in the pozzolanic reaction, a detailed knowledge of the pozzolanic reaction mechanism is needed. In this particular area much work remains to be done. Unraveling the impact of the heterogeneous group of SCMs on cement hydration will constitute one of the major challenges in cement science for years to come. As a starting point, in the following sections, the traditional views on the pozzolanic reaction are combined with recent fundamental insights into the dissolution of minerals developed in geochemistry (Lasaga 1998; Dove et al. 2005).

242

Snellings, Mertens, Elsen

The pozzolanic reaction mechanism A general overview of the different stages in the pozzolanic reaction can be obtained by monitoring the rate at which heat is evolved during the reaction. The rate of heat evolution is closely related to the actual rate of reaction provided that the reactions are exothermal (Gartner et al. 2002). In Figure 21 an idealized heat release curve for the pozzolanic reaction between silica and Ca(OH)2 is presented. Three main stages can be recognized. The duration of each of the stages and also the shape of the heat release curve vary significantly depending on the experimental conditions and the pozzolanic activity of the material. The initial sharp peak occurring directly after mixing (stage I) lasts only for several minutes and is followed by a period of low activity designated as the induction period (stage II). Renewed activity defines the initiation of stage III, which can be divided tentatively in parts of an accelerating and a decelerating reaction rate. The induction period usually lasts only for several hours to some days. The duration of stage III is indefinite. In some cases the reaction can proceed at very low rates for extended periods of years to decades (e.g., Taylor et al. 2010).

Figure 21. Schematic overview of the rate of heat release during the pozzolanic reaction of silica and Ca(OH)2. Stage I, represents the initial dissolution period. Stage II corresponds to the induction period and stage III is the phase in which the main reaction occurs.

I. Initial dissolution period and aluminosilicate dissolution. Both lime-based and Portland cement based binders have in common that initial fast dissolution of Ca(OH)2 or clinker minerals in water rapidly results in an alkaline solution saturated in Ca(OH)2. In Ca(OH)2based systems the solution pH is usually around 12.4, while in Portland cement based systems the solution may reach values of 13.7 when abundant soluble alkalis were released from the clinker minerals. At alkaline pH above 10.7 the solubility of silica and silicates (i.e., amount of silica in solution) increases continuously with pH (Iler 1979; Knauss and Wolery 1988) and pozzolans will be subjected to dissolution. The silicate dissolution rate at high pH is governed by processes of hydration, deprotonation, ion-adsorption and hydrolysis at the mineral-water interface. Greenberg (1961) concluded that the rate-controlling step in the pozzolanic reaction was the hydrolysis of surface silica groups. To understand and predict the kinetics of dissolution, the molecular details and energetics of the actual pathway that links reactants and products via the activated complex need to be known. The activation energy is the energy difference between the activated complex or transition state and the reactants. Ab initio quantum mechanical calculations of finite molecular clusters clarified the pathway to dissolution of silicates and aluminosilicates at high pH in recent years (cf. Xiao and Lasaga 1994; Lasaga 1998). The generally accepted reaction pathway for the hydrolysis of the bridging bond (Obr) in a

Supplementary Cementitious Materials

243

(HO)3-Si-Obr-Si-(OH)3 (Q1) cluster is illustrated in Figure 22 and the corresponding energetics are given in Figure 23 (Morrow et al. 2009). In alkaline conditions the silanol groups at the silicate surface are partially deprotonated. Xiao and Lasaga (1996) have shown that nucleophilic hydroxyl attack on a neutral Si-OH surface group is equivalent to hydrolysis of a deprotonated Si-O- surface group. Both situations result in a H-bonded H2O adsorption onto the negatively charged Si-O− site (RC or reaction precursor complex). The key step in the reaction is the formation of a negatively charged fivefold coordinated trigonal bipyramidal Si species. To form this reaction intermediate (INT) the reaction has to pass through a transition state (TS1) and surmount the associated large energy barrier (calculated to be 110 kJ/mol in the gas phase; Morrow et al. 2009). The Si-Obr bond in the formed intermediate state is significantly weakened and the energy barrier associated with the final step of bond breakage is much smaller (22 kJ/ mol in gas phase; Morrow et al. 2009).

Figure 22 (above). Reaction stages along the reaction profile for the hydrolysis of a deprotonated Si-O-Si species (Morrow et al. 2009). RC is the reaction precursor complex, TS1 and TS2 are transition states, INT is an intermediary state and PC the product complex. Figure 23 (to the left). Energy profile for the Si-O-Si hydrolysis reaction shown in Figure 22 (modified after Morrow et al. 2009).

The activation energy of Si-Obr-Si hydrolysis in larger clusters where the considered Si atoms are doubly (Q2) or triply (Q3) connected to neighbor Si atoms via Si-Obr-Si bonds (e.g., Fig. 24) was investigated by Pelmenschikov and co-workers for dissolution by H2O attack (Pelmenschikov et al. 2000, 2001) and by Criscenti et al. (2006) for dissolution by H3O+ attack on a Q3 cluster. The calculations indicated that the activation energy increased with connectivity (up to 205 kJ/mol for a Q4 cluster), this was attributed to resistance of the remaining bridging bonds against relaxation of the partially uncoupled Si species (Pelmenschikov et al. 2000). In the neutral silanol state the breakage of the Si-Obr-Si bond is expected to be followed by a very fast condensation or rehealing reaction. The experimentally determined activation energy of silicate dissolution (67-92 kJ/mol; Knauss and Wolery 1988; Brady and Walther 1990; Dove and Crerar 1990; Walther 1996) would then be associated with the hydrolysis of the last Si-ObrSi bonds (Q1 and/or Q2) (Criscenti et al. 2006). In an alkaline medium the rehealing reaction is suggested to be partially prevented by the deprotonation of the formed Si-OH HO-Si defect (Pelmenschikov et al. 2001).

244

Snellings, Mertens, Elsen

Figure 24. Reaction profile for the hydrolysis of a Q3 connected Si-Obr-Si bond at the (111) b-cristobalite surface (Pelmenshikov et al. 2000). Obr stands for bridging oxygen. Ow was originally part of the attacking water molecule.

Morrow et al. (2009) compared the pathways of dissolution of aluminosilicate Q1 clusters with the earlier reported silicate dissolution mechanism. At high pH dissolved Al is fourfold coordinated (Swaddle et al. 2005), only one transition state was found for the dissolution and release of Al(OH)4−. This step corresponded with the formation of a fivefold coordinated almost trigonal bipyramidal Al species with a much lengthened Al-Obr bond interaction (Fig. 25). At acid and neutral pH the barrier heights of Si-Obr-Si hydrolysis (resp. 63 and 146 kJ/mol) were considerably higher than that of Al-Obr-Si linkages (resp. 38 and 39 kJ/mol). This is supported by the experimental observation of leaching of Al from the surface of feldspars at low pH (Stillings and Brantley 1995). However, at high pH there was no significant difference (79 kJ/ mol for both), which is in agreement with the observation that little or no Al depletion occurs on aluminosilicate surfaces at high pH (Hamilton et al. 2001). The dissolution rates of silicate minerals were observed to be affected by the solution pH or, equivalently, by the distribution of protonated, neutral or deprotonated groups at the mineral surface. Based on such a surface speciation model and transition state theory (cf. Lasaga 1998) the pH dependence of the dissolution rate of quartz can be predicted (Dove 1994; Nangia and Garrison 2008). However, for mixed oxides containing Si-Obr-Si linkages such as aluminosilicate feldspars, the activities of the leached cations in solution should be included into the model (Oelkers 2001). For orthosilicates and minerals without silicate linkages, dissolution rates can be correlated with the solvation-water exchange constant of the non-silicate cations (Westrich et al. 1993). Background cations in solution change the aluminosilicate surface speciation by adsorption and proton exchange reactions. At high pH, alkali and alkaline-earth cations are drawn to the negatively charged surface by electrostatic attraction (Iler 1979). Both

Figure 25. Reaction stages along the reaction profile for hydrolysis of a deprotonated Al-Obr-Si (Morrow et al. 2009). RC is the reactant complex, TS the transition state and PC the product complex.

Supplementary Cementitious Materials

245

experimental (Sjöberg 1989; Dove and Crerrar 1990) and computational (Strandh et al. 1997) results show that the adsorption of cations generally leads to enhanced dissolution rates and weakened bridging bonds of the Si or Al species with the lattice. In general, alkaline-earth cations in solution bind more strongly to the mineral surface and remain only partially hydrated. Alkali cations retain significant water shielding upon interaction and would increase dissolution rates by increasing the reaction frequency compared to alkaline-earth cations which show slower exchange of solvation water (Dove 1999). When water is added to dry blended cement or lime-SCM binders, dissolution of clinker constituents, SCMs or lime occurs at first in far-from-equilibrium, highly undersaturated conditions. The initial fast dissolution rate subsequently slows down significantly until a Ca(OH)2 saturated solution is reached. In terms of silicate dissolution theory, this parabolic dissolution rate behavior has been attributed to the rapid initial dissolution of fine particles or sites with high surface energy (Brantley 2008). Other models explain the parabolic rate behavior by the formation of a leached surface layer due to non-stoichiometric dissolution or by the precipitation of a protective membrane at the mineral surface through which dissolved ions must diffuse (Schott and Petit 1987). The latter two models are very similar to the mechanisms invoked to explain the occurrence of the induction period during alite hydration (see below; Gartner et al. 2002; Bullard et al. 2011). More recently a mechanistic model for mineral dissolution was formulated to enable the prediction of dissolution rates depending on the saturation state of the solution (Dove et al. 2005; Lasaga and Lüttge 2005). This model is based on concepts borrowed from crystal growth theory where dissolution can be regarded as the inverse of crystal growth. Dissolution occurs as the result of horizontal step retreat at incomplete surface layers and of vertical removal of atoms at plain surfaces or at the intersection of dislocation or point defects with the surface (Fig. 26). Three different regimes of mineral dissolution can be distinguished depending on the dominant mechanism of dissolution. At very high undersaturation, twodimensional pits or vacancy islands can nucleate at perfect surfaces without any dislocations present. However, the activation energy barrier for this mechanism is high and is expected to occur for only a very short time in cementitious systems (Juilland et al. 2010). Closer to equilibrium, two-dimensional pitting on plain surfaces will end. However, step nucleation can proceed at dislocation defects due to the associated strain field. Eventually, when the saturation degree reaches near-equilibrium, only step retreat is possible. No more steps can form at the surface or near dislocation defects. Step nucleation only occurs at crystal edges and the crystal becomes progressively smoother and edges more rounded (Brantley 2008). In this model the dissolution rate depends on the density of steps present or nucleated at the surface and the velocity of step retreat at the surface. Both parameters are a function of the degree of solution undersaturation, thus the dissolution rate will decrease sharply when the undersaturation degree

Figure 26. The dominant dissolution mechanism depends on the level of undersaturation. At very high undersaturations two-dimensional pits or vacancy islands can nucleate at perfect surfaces, at lower levels of undersaturation step nucleation can still proceed at dislocation defects. Eventually near equilibrium step nucleation ends and step retreat becomes the dominant dissolution mechanism. W identifies with the saturation degree (modified after Dove et al. 2005 and Juilland et al. 2010).

246

Snellings, Mertens, Elsen

approaches equilibrium. However, the step nucleation mechanism and rate depend also on the nature of the crystal and the crystal face, especially in terms of the liquid-solid interface energy (Dove et al. 2005). Furthermore, dissolution rates near equilibrium increase with increasing defect density as exemplified for alite by Juilland et al. (2010) or for quartz by Blum et al. (1990). The suggestion by ab initio cluster studies that the precursor molecules of silicate dissolution should generally have a connectedness lower than 3 is consistent with a steady-state silicate dissolution mechanism dominated by step retreat. II. Induction period. In the pozzolanic reaction the reaction rate decreases rapidly during period I and remains low during the ensuing induction period (Fig. 21). In most models this evolution is linked to the formation of a protective barrier layer on the reacting pozzolan particles (Takemoto and Uchikawa 1980; Glasser et al. 1987; Fraay et al. 1989). This barrier layer is expected to shield the pozzolan from the surrounding basic solution and to hamper its dissolution. Two distinct hypotheses are put forward regarding the nature of the protective layer. 1) The leached layer hypothesis is based on the incongruent dissolution of the pozzolan (cf. Livingston et al. 2001 for alite). As alkalis are leached from the surface an amorphous layer consisting of Si and Al remains, meanwhile Ca2+ is adsorbed at the surface and a double layer is created, inhibiting further dissolution. Eventually Si and Al dissolve and recombine with the adsorbed Ca2+ to form C-S-H and C-A-H phases (Takemoto and Uchikawa 1980). 2) The protective precipitate hypothesis is based on a dissolution-reprecipitation process, in which the pozzolan first releases Si and Al which then subsequently reprecipitate at the pozzolan surface as a coating of stable or metastable C-S-H and C-A-H reaction products (Greenberg 1961; Glasser et al. 1987). Alternatively, the reaction rate can also decrease sharply because of a decrease in the degree of undersaturation of the solution as suggested by the mechanistic model for mineral dissolution. If the degree of undersaturation would drop below the threshold for step nucleation at dislocation edges, the dissolution rate would be considerably lowered. The latter mechanism could of course also take place together with the formation of a barrier layer at the pozzolan-liquid interface. The end of the induction period is marked by the massive nucleation and subsequent growth of reaction products. Several theories have been proposed to explain the sudden transition from the induction to the main period of reaction. The suggested theories were mainly derived from concepts developed to explain the hydration behavior of alite (cf. Gartner et al. 2002; Bullard et al. 2011). Takemoto and Uchikawa (1980) have suggested that the protective membrane would be semi-permeable, allowing osmosis of water from the outer solution to the concentrated inner solution. The rising osmotic pressure due to the ongoing dissolution of the pozzolan would eventually result in a rupture of the membrane, release of dissolved silica and alumina and the start of the main period of C-S-H and C-A-H precipitation. This theory was based on the osmotic pump model for the hydration of alite by Double et al. (1978) and the observation of relict hollow shells of reaction products by electron microscopy. Also more recent Nuclear Resonance Reaction Analysis results that indicate the development of a depleted silica gel layer at the surface of alite grains during the induction period have been linked to the existence of a semi-permeable layer (Livingston et al. 2001). A different, more widely accepted mechanism was first suggested by Stein and Stevels (1964) for the hydration of alite in the presence of silica. They proposed that the conversion of the protective layer of metastable C-S-H(m) or disordered gel type phases to a more stable C-S-H form would result in the acceleration of the hydration reaction of alite. The same mechanism can be applied to the pozzolanic reaction, the transformation would be triggered by the reaching of a certain degree of supersaturation with respect to the stable C-S-H assemblage. However, detailed studies on the evolution of the fluid saturation degree during the first stages of the pozzolanic reaction are still lacking. III. Main reaction period. At the end of the induction period, typically only a very limited amount of reaction products has been formed and the pozzolan has barely reacted (Snellings

Supplementary Cementitious Materials

247

et al. 2009). The initiation of the main reaction period can be observed by an exponentially increasing heat release rate (Fig. 21) related to the large-scale nucleation and growth of the reaction products as the rate-controlling steps. The exponential reaction rate increase is shortlived, and eventually the reaction rate starts decreasing again. It has been widely considered that the decrease in rate is due to the onset of a diffusion-controlled process. Many observations of the formation of an enveloping layer of reaction products on the pozzolan grains supported the interpretation that the reaction rate becomes limited by the diffusion of reactants through the reaction product layer (Držaj et al. 1978; Türker and Yeginobali 2003; Mertens et al. 2009). Other processes may also cause a deceleration of the reaction rate. Consumption of the smallest particles will leave coarse particles that react more slowly. Also a lack of space or densification of the C-S-H rim can hinder the growth of C-S-H particles and thus decrease the reaction rate (Bishnoi and Scrivener 2009). Finally, the growing competition for a dwindling supply of water can lead to the deceleration of the reaction (Bullard et al. 2011). In the traditional view of reaction rate deceleration due to diffusion control, several mathematical models have been used to fit the evolution of the degree of reaction a. In Ca(OH)2pozzolan systems the degree of reaction is usually quantified by the direct determination by X-ray diffraction or thermogravimetry of the amount of Ca(OH)2 reacted (Kondo et al. 1976; Takemoto and Uchikawa 1980; Shi and Day 2000) or by the evolution of the overall heat release curve measured by calorimetry. Most kinetic models are based on the Jander equation developed to describe three-dimensional diffusion control during solid-state sintering of a contracting reactant sphere (Jander 1927): 2 2 kt 1 − (1 − a )1/3=  = Kt   r2

(3)

Where a identifies with the fractional reaction, k with the rate constant for the diffusion process, t equals to the time since the onset of the diffusion controlled reaction process, r is the initial radius of the spherical reactant and K represents a constant proportional to k. In the original Jander equation the thickness of the interface layer or equivalently, its diffusion coefficient is taken not to change over time. However, in cementitious systems the interface layer is expected to grow or densify gradually with the proceeding reaction. Therefore, a modified, more general form of this equation has been used more frequently (Kondo et al. 1976; Shi and Day 2000; Mertens et al. 2009). N

1 − (1 − a )1/3  = Kt  

(4)

This equation allows the classification of the ongoing reaction based on the value of the exponent of reaction N. If N = 1, then dissolution or nucleation/precipitation processes at the surface of the grains are the rate-limiting step. With N > 1, three-dimensional diffusion through a layer of reaction products is considered to be the rate-limiting step. A decreasing permeability of the interface layer induced by thickening or densification by reaction product precipitation would correspond with N < 2. In all studies a conspicuous increase in N was observed over the course of hydration, implying a decreasing permeability of the reaction product interface layer (Takemoto and Uchikawa 1980; Shi and Day 2000; Mertens et al. 2009). Cabrera and coworkers have successfully applied both the original Jander equation on metakaolin-Ca(OH)2 systems (Cabrera and Rojas 2001) as well as the modification developed by Ginstling and Brounshtein (1950) to allow for the decreasing permeability of the interface layer on silica fume-Ca(OH)2 and trass-Ca(OH)2 systems. The Ginstling-Brounhstein equation is formulated as follows: 2 3 1 − (1 − a )1/3  − 2 1 − (1 − a )1/3  = Kt   3 

(5)

Villar-Cociña et al. (2003) have applied a decreasing nucleus model to model the decrease in

248

Snellings, Mertens, Elsen

electrical conductivity of a saturated Ca(OH)2 solution upon the addition of a pozzolan. The model was used to determine diffusion coefficients of the interlayer and overall reaction rate constants to indicate the activity of the added pozzolans. It should be noted that the fitting of the presented kinetic models can only give general information of the reaction mechanism and that the results should be interpreted cautiously. Variations in reaction product morphology, thickening or densification of the interface layer will influence the permeability and diffusion coefficient of the barrier layer. Additionally, obtained reaction rate constants should be considered as overall apparent values because the apparent rate constant will encompass a series of processes which are considered to be combined in a pseudo-first order reaction. Averaging occurs also over the reaction rates of inner and outer product formation. Additionally, finer fractions of a pozzolan will react more rapidly than the coarser ones and an overall value will be obtained.

Pozzolanic activity The reaction rate during and the timing and duration of the described stages of the pozzolanic reaction are highly dependent on the intrinsic activity and characteristics of the pozzolan. Highly active pozzolans such as metakaolin will increase heat release during the initial dissolution period and significantly shorten the duration of the induction period (Massazza 2001). Additional to intrinsic pozzolan properties such as specific surface area, chemical composition or active phase content, the consumption of Ca(OH)2 over time is also depending on external factors such as mix design and curing conditions. Intrinsic pozzolan properties. It is generally accepted that a first-order relationship exists between the rate of dissolution and the total or reactive mineral surface area under far-fromequilibrium conditions (Brantley 2008). However, the determination of the specific surface area is not entirely straightforward. It can be measured by geometric calculations departing form the particle size distribution curve or by the BET N2 adsorption technique. Both techniques have drawbacks. Geometric calculations demand the assumption of particle shapes and in the BET technique the internal particle porosity is often included (White and Brantley 2003). Both mineral dissolution and the initial stages of the pozzolanic reaction consist of processes taking place at the liquid-solid interface. Therefore, at least initially, a linear relationship is expected between the activity of a pozzolan and the available surface area for reaction. Correlations between the Blaine fineness or the BET specific surface of a specific pozzolan and its activity or even the evolution of compressive strength have been reported frequently for the early reaction period, varying in duration from 7 days up to 3 months (Ludwig and Schwiete 1963; Costa and Massazza 1974; Takemoto and Uchikawa 1980; Day and Shi 1994, Mertens et al. 2009). However, inconsistencies arise when the activity dependence on specific surface of materials of different origins are compared. The absence of a single encompassing relationship points to the fact that the pozzolanic activity is dependent on more factors. To illustrate the effect of particle shape and internal porosity, for similar particle size distributions or fineness the complex structure and highly porous nature of rice husk ash or diatomite constituents may result in very different activities compared to that of spherical non-porous fly ash particles. To enhance the specific surface area and thus the activity of a pozzolan, grinding is a commonly used technique. Even materials which are commonly not regarded to behave pozzolanically, such as quartz, can become reactive when ground below a certain critical particle diameter (Benezet and Benhassaine 1999). The particle diameter or the curvature of the surface of a pozzolan particle also affects the surface energy or surface tension through the Gibbs-Thomson effect and results in a higher solubility of smaller particles (cf. Ostwald ripening) (Iler 1979). Additional to increasing the total surface area, grinding also results in the creation of surface defects, i.e., sites experiencing lattice strain or sites partially disconnected from the underlying undisturbed material (Alexander 1960). The lowered activation energy of hydrolysis of the Si-

Supplementary Cementitious Materials

249

Obr-Si bonds at these sites has a definite positive effect on the pozzolanic activity. Activation of pozzolans by acid treatment results in an etched surface and the enhanced activity relies on a similar creation of abundant surface pits and steps. More generally, the bulk density of defects in a material exerts a significant influence on the activity of both minerals and vitreous materials (Shi 2001; Nair et al. 2008). Highly-crystalline materials showing very few linear or planar defects are generally observed to be much more stable in Ca(OH)2 saturated solutions than amorphous materials with large concentrations of bulk and surface defects. This can be illustrated by the higher pozzolanic activity of metakaolin produced from paper sludge ash compared to regularly produced metakaolin of higher purity and larger specific surface area. The increased reactivity of the former was related to an increased concentration of surface defects (Péra and Amrouz 1998). Bich et al. (2009) characterized the presence of surface defects in kaolinite based on the asymmetry of the DTA peak of dehydroxylation. Thermally activated disordered kaolinite with many surface defects was shown to be more reactive than burned ordered kaolinite showing few defects. Activation of kaolinite by prolonged grinding was related to surface defect creation. This is illustrated by the disappearance of the DTA peak of dehydroxylation and the remarkable increase in weight loss over the lower temperature range of 100-500 °C in Figure 27, indicating the presence of disordered weakly bonded hydroxyl and water groups (Vizcayno et al. 2009). In determining the pozzolanic and cementitious properties of SCMs also the mineralogical composition plays a prominent role. The proneness to reaction of the amorphous and crystalline phases is linked to their structural stability in an alkaline Ca(OH)2 saturated solution. The effect of crystal structural properties on the mineral stability can be illustrated by variations in

Figure 27. DTA (top) and TG (bottom) curves for natural kaolinite and kaolinite ground for 15, 30, and 60 min. in a Herzog-mill (modified after Vizcayno et al. 2009).

250

Snellings, Mertens, Elsen

enthalpy of formation of pure-silica polymorph structures. Besides the naturally occurring silica polymorphs, a- and b-quartz, cristobalite, tridymite, coesite and stishovite, a large diversity of synthesized all-silica molecular sieves or zeolites exists and is used in industrial processes. Petrovic et al. (1993) and Piccione et al. (2000) experimentally determined the enthalpies of transition from the stable a-quartz structure to the metastable polymorph structures. The range of energies is quite narrow at only 6.8-14.4 kJ/mol SiO2 above quartz, and a strong linear correlation between enthalpy and framework density (the number of tetrahedral framework atoms per nm³) was observed (Fig. 28). Both experimental and theoretical (Henson et al. 1994; Sastre and Corma 2006; Zwijnenburg et al. 2007) results indicate that for the pure silica polymorphs the quality of packing of the SiO4 tetrahedra is the most important parameter controlling silica stability. Furthermore, the formation enthalpy of an amorphous silica glass was only 7 kJ/mol higher than quartz and lower by 0-7 kJ/mol than the zeolitic structures (Petrovic et al. 1993). On purely crystal structural grounds, this difference may serve as an explanation of the observed higher activity of natural zeolites compared to volcanic glass of similar composition and origin (Mortureux et al. 1980; Sersale 1980). The higher enthalpy of formation of less dense structures is expected to increase the difference in Gibbs energy between reactants and reaction products. Nevertheless, the effect of framework density on dissolution activation energy and thus on reaction kinetics remains unclear. The situation becomes more complicated when typical SCMs consisting of several mixed oxide phases are considered. In general, the pozzolanic activity of minerals thermodynamically stable at ambient conditions is low when compared on an equal specific surface basis to less stable mineral assemblages. Volcanogenic deposits containing large amounts of volcanic glass or zeolites are more reactive than quartz sands or detrital clays. In this respect, the thermodynamic driving force behind the pozzolanic reaction may serve as a rough indicator of the potential reactivity of a specific crystalline or non-crystalline phase (Takemoto and Uchikawa 1980). Obviously, the content of active phases in a specific SCM is a factor of primary importance (Millet and Hommey 1974; Sersale 1993). Many studies have suggested that a correlation exists between the long term performance in terms of compressive strength or durability and the bulk chemistry of the SCM without reference to the mineralogical composition (e.g., Watt and Thorne 1965; Costa and Massazza 1974;

Figure 28. Experimental transition enthalpy from quartz compared to framework density of pure-silica polymorphs (Piccione et al. 2000). Qtz stands for quartz, Co for coesite, Cr for cristobalite, Tr for tridymite, CHA for chabazite and FAU for faujasite.

Supplementary Cementitious Materials

251

Hanna and Afify 1974; Cavdar and Yetgin 2006). The pozzolanic activities and blended cement or pozzolan-lime binder performance showed positive linear correlations of varying statistical significance with the sum of bulk SiO2 and Al2O3, in some cases also Fe2O3 was added. The sum SiO2 + Al2O3 + Fe2O3 ≥ 70 wt% remains one of the fulfillment criteria for pozzolans in ASTM C618. In case of a one phase material the chemical composition can be considered as a meaningful parameter. However most natural and artificial SCMs consist of a heterogeneous mixture of phases and then a direct relationship between overall chemical composition and pozzolanic activity becomes less obvious. The fact that all correlations were reported for long term pozzolanic activity and/or performance indicates that the characteristics and the distribution of the reaction products should relate SCM chemistry and long term performance. The eventual, long term reaction product assemblage is controlled by the overall chemistry of the active phases (Massazza 2001). Addition of blast furnace slag or metakaolin has been observed to change the Ca/Si ratio, the silicate polymerization and the morphology of the main C-S-H reaction products and can thus alter the permeability of the reaction product barrier layer and the eventual performance of the binder (Richardson 1999, 2004). As the main reaction products are calcium-silicate-hydrates (with some Al incorporation) and calcium-aluminate-hydrates (containing additional Si and Fe), the total SiO2 + Al2O3 + Fe2O3 content of the active phases may be considered as an indication of the Ca(OH)2 binding potential of an SCM. In practice, it is very difficult to separate the contributions to the SCM activity of physical particle characteristics and mineralogical properties. This is considered one of the primary reasons of contradictory findings in literature concerning the relative activities of SCM phases. Furthermore, because SCMs of comparable origin often show broad similarities in physical particle properties if not in mineralogical and chemical composition, the widespread adoption of the genetic classification scheme of SCMs can be considered to remain sensible. External factors. The rate of the pozzolanic reaction also depends on the mix design, larger water/binder ratios will result in increased pozzolanic activity but will inevitably decrease the performance of the binder due to the increased overall porosity. The ratio of SCM over Ca(OH)2, or equivalently the ratio of SCM over Portland cement obviously affects the pozzolanic activity in increasing or decreasing the frequency of fulfillment of the reaction configuration. The Ca(OH)2:pozzolan ratio for optimal performance and activity is depending on the overall content, composition and activity of the constituent phases of the SCM, but is usually situated in between 1:1 (Murat 1983; Bakolas et al. 2006) and 2:1 (Takemoto and Uchikawa 1980; Costa and Massazza 1974). Pozzolans rich in Al2O3 generally need higher Ca(OH)2:SCM ratios for optimal reactivity, and SCMs displaying hydraulic activity usually need much less Ca(OH)2 to activate the hydration reactions (Lang 2002). In terms of Portland cement over SCM ratio optimal replacement percentages are often defined based on the desired properties of the hardened cement. In general, the optimal replacement ratio depends on the water demand, i.e., surface roughness and specific surface, and the activity of the SCM. The higher the water demand and activity, the lower the optimal replacement ratio is, typically 10-15 wt% for silica fume or metakaolin. Optimal replacement ratios defined for durability properties tend to be somewhat higher than ratios for optimal strength performance. At excessive replacement levels the pozzolanic activity is lowered because of the premature depletion of the solution alkalinity by the reacting SCM. A significant drop in solution pH below 10 may not only effectuate a decrease in pozzolanic reaction rate, but can also lead to the destabilization of AFm and AFt reaction products in the blended cement (Lothenbach et al. 2011). To increase the pozzolanic reaction rate, curing at elevated temperatures can be applied. The temperature dependence of the pozzolanic reaction on the short term can be described by the Arrhenius equation, implying an exponential dependence of the reaction rate on temperature (Snellings et al. 2009). The curing temperature should however not exceed the stability field of the C-S-H phase and result in the precipitation of more crystalline phases (typically above

252

Snellings, Mertens, Elsen

80-100 °C; Taylor 1990). The acceleration induced by elevated curing temperatures is much more pronounced for the pozzolanic reaction compared to the hydration of Portland cement constituents due to the higher activation energy of the pozzolanic reaction (Shi and Day 1993).

Hydration mechanism and kinetics of blended cements In blended Portland cements the hydration reactions of the clinker phases are complemented by the pozzolanic or hydraulic reactions of the added SCM. Although the hydration processes of the clinker phases follow different mechanisms and rates than the pozzolanic or hydraulic reactions of the SCM, the clinker hydration in blended cements is influenced by the presence of SCMs. Reaction kinetics, products and the properties of fresh and hardened pastes can be manipulated by the replacement of a fraction of the Portland cement by SCMs. Both the properties of the SCM as the mix design are determining factors with the potential to affect all stages of the hydration and pozzolanic reactions in the blended cement. Influence of SCMs on the hydration of clinker phases. To eliminate the interference of simultaneously occurring reactions in a blended cement, the effect of SCM addition on the hydration kinetics of single clinker compounds has been investigated by numerous researchers. The hydration mechanisms of the individual clinker components have been recently reviewed by Gartner et al. (2002) and Bullard et al. (2011). Similar to the pozzolanic reaction mechanism, the hydration of C3A in the presence of gypsum and the hydration of C3S experience both a brief initial phase of high reactivity followed by a dormant period and an eventual main reaction stage. C3S. The main component of Portland cement is C3S, constituting 60-70 wt% of the cement. In general, the addition of an SCM has an accelerating effect on the hydration of C3S (Ogawa et al. 1980). The heat evolution rate during the main reaction period and the cumulative amount of heat released over the complete reaction are increased, especially when recalculated to the C3S content in the samples (Massazza 2001). The effect of the SCM on the early reaction is mainly governed by its fineness, as illustrated in Figure 29. The initial dissolution period is

Figure 29. The effect of the silica fume specific surface area on the heat evolution rate curves in hydrating pastes of C3S with 20 wt% silica fume (Beedle et al. 1989).

Supplementary Cementitious Materials

253

lengthened and the induction stage shortened, when SCMs of increasing fineness are introduced (Stein and Stevels 1964; Kurdowski and Nocun-Wczelik 1983; Beedle et al. 1989; Korpa et al. 2008). This phenomenon is usually termed the filler effect. The addition of extremely fine particles results however in a decreased hydration rate because the high water demand of the particles limits the amount of water that can participate in the hydration and pozzolanic reactions (Beedle et al. 1989; Korpa et al. 2008). SCMs with low specific surface area such as certain fly ashes have been observed to lengthen the induction period, but increase the cumulative heat released during the main reaction stage (Watt and Thorne 1965, 1966). This dependence on the SCM specific surface has been attributed to Ca2+ adsorption and C-S-H nucleation at the SCM surface. In consequence, the layer of reaction products on the C3S particles would be thinner and C3S dissolution would thus last longer (Wu and Young 1984). In addition, the lowered Ca2+ and hydroxyl concentrations at the C3S surface may accelerate the conversion of an initially precipitated impermeable metastable C-S-H(m) to a more stable and permeable C-S-H form (Stein and Stevels 1964). Eventually Ca(OH)2 formed as a result of the C3S hydration, will be partially or completely consumed by the pozzolanic reaction. C3A. The hydration of C3A in the presence of gypsum is an important regulator of the setting of the blended cement. Contrary to the effect of SCMs on C3S hydration, the heat evolution rate of the main C3A hydration stage is lowered in the presence of SCMs (Collepardi et al 1978; Uchikawa and Uchida 1980). The reasons underlying the decreased C3A hydration rate remain unclear. A correlation seems to exist with the SCM specific surface; the higher the specific surface, the lower the hydration rate (Collepardi et al. 1978). Other factors could be the alteration of solution composition by SCM addition, the adsorption of sulfate at the pozzolan surface or the altered hydration product assemblage. Blended cements. The hydration behavior of blended cements is dominated by the hydration of its main component. In the case of pozzolan containing blended cements, C3S is the most prominent constituent. In cements consisting of a large fraction (50-90 wt%) of a hydraulic SCM, e.g., granulated blast furnace slag, the long term behavior is governed by the hydration of the SCM. In consequence, for pozzolanic SCMs, the early hydration behavior is mainly affected by the specific surface of the SCM. On the condition that sufficient water remains available for hydration, an increase in specific surface results in an accelerated cement hydration, lengthened early dissolution stage and shortened induction period (Takemoto and Uchikawa 1980). To the contrary, slag dominated cements are known to hydrate more slowly and expel less heat than Portland cement and are therefore suitable for applications in mass concrete structures such as dams, reservoir, quays etc. (Lang 2002). However, when recalculated to the actual clinker content, the heat liberated by the slag-cement over a longer time period is higher than that released by the Portland cement (Hooton and Emery 1983). Obviously, the added SCMs consume Ca(OH)2 released by clinker hydration to form supplementary cementitious reaction products. In addition the reaction product composition, structure, morphology and in some cases also the assemblages are considerably changed.

REACTION PRODUCTS Compared to the wide variability in supplementary cementitious materials, there exists only a relatively small range of compounds formed in the pozzolanic, hydraulic or hydration reactions. This is mainly due to the similarity in overall chemical composition of the hydrating mixtures of pozzolans and lime or blended cements, regardless of the type of SCM. In consequence, only a limited number of phases will be thermodynamically stable or metastable at ambient conditions. This observation enabled the application of thermodynamic models to calculate the ultimate hydrate mineralogy from chemistry by Lothenbach and coworkers (e.g., Lothenbach and Winnefeld 2006; Matschei et al. 2007b). Combined with quantitative kinetic

254

Snellings, Mertens, Elsen

data or models on reactant consumption or product formation, thermodynamic calculations have also been used to study evolving hydration processes (e.g., Lothenbach et al. 2007, 2008a; Winnefeld and Lothenbach 2010). Although kinetic barriers may impede the attainment of ultimate thermodynamic equilibrium, or formation of products may not conform to the predicted assemblage due to locally prevailing chemical conditions. The thermodynamic approach has shown to be successful in predicting the behavior of many product assemblages in function of variable pH, sulfate, carbonate or temperature conditions (e.g., Lothenbach and Gruskovnjak 2007; Pelletier et al. 2010; Matschei and Glasser 2010). Hence, reported observations on product assemblages of the pozzolan-lime reaction or the hydration of blended cements may be compared at least in a qualitative way with thermodynamic predictions.

Product assemblages The chemical composition of the mix, more specifically the amount and composition of the pozzolanically or hydraulically active phases together with the pozzolan - lime or pozzolan - Portland cement ratio are primary factors in the determination of the reaction product assemblage. In addition, the presence of soluble sulfate, carbonate or chloride may provoke the formation of AFt and AFm phases incorporating the corresponding anion groups (Matschei et al. 2007c; Balonis et al. 2010). Curing conditions are equally important, partial CO2 pressure and curing temperature may affect the product assemblage considerably (Lothenbach et al. 2008a). Evolving reaction conditions, e.g., the depletion of Ca(OH)2 by the pozzolanic reaction or decreasing released heat of hydration, may result in changes in the product assemblage during ageing. Pozzolan-lime reaction products. In Ca(OH)2 saturated solutions silica released from the pozzolan in combination with Ca2+ primarily forms C-S-H. Dissolved alumina can be incorporated into the C-S-H phase (Richardson et al. 1993) or can be precipitated in combination with Ca2+ as calcium-aluminate-hydrates. The Ca/Si ratio of the C-S-H phase is variable both in space as in time and also depends on the activity and composition of the pozzolan and the mix design. Increasing polymerization of the silicate groups in the C-S-H is observed upon prolonged curing—the silicate monomer and dimer contents decrease while the polymer content increases (Massazza and Testolin 1983; Brough et al. 1995). In the absence of sulfate, carbonate or chlorides C4AH13-19 forms as main calcium aluminate hydrate (Taylor 1990). If general or local deficiency of Ca(OH)2 occurs, often encountered when metakaolin is added as SCM, the recombination of alumina and silica with Ca2+ to form strätlingite (hydrated gehlenite), C2ASH8, can be observed (Serry et al. 1984; Ambroise et al. 1994). Strätlingite allows structural substitution of Na+ and K+ for Ca2+ (Jones 2002). At later stages of reaction, higher water/solid ratio or in the presence of abundant alkalis, hydrogarnet or katoite, C3AH6, is stabilized (Takemoto and Uchikawa 1980; Serry et al. 1984; Sersale 1993). Hydrogarnet can incorporate Si in a solid solution series between katoite Ca3Al2(OH)12 and grossularite Ca3Al2Si3O12. In metakaolin-lime pastes the formation of hydrogarnet is mostly observed at elevated curing temperatures above 40 °C (De Silva and Glasser 1992). Although C4AH13 and strätlingite are considered to be thermodynamically unstable towards hydrogarnet and Ca(OH)2, no evidence of the transformation of strätlingite and C4AH13 to hydrogarnet was observed to occur at 60 °C (Rojas and Cabrera 2002; Rojas 2006). At temperatures below 40 °C, typically only traces of hydrogarnet can be observed. Hydrogarnet forms very slowly at low temperatures, probably due to kinetic reasons (Takemoto and Uchikawa 1980; Massazza 2001). C4AH13 can possibly be stabilized by the incorporation of small amounts of carbonate or sulfate (Pöllmann 2006). SCMs containing soluble sulfate such as fly ashes or blast furnace slags will react with Ca(OH)2 to form ettringite, C6A S 3H32, or monosulfoaluminate (kuzelite), C4A S H12, or a combination of both depending on the sulfate over alumina ratio (Mortureux et al. 1980). A

Supplementary Cementitious Materials

255

limited solid solution of sulfate into C4AH13 up to SO4/2OH of 0.5 was observed at 25 °C (Matschei et al. 2007c). Addition of Na2SO4 accelerates the pozzolanic reaction. Removal of dissolved sulfate by ettringite precipitation is compensated by an increase in hydroxide concentration and pH to maintain electroneutrality (Shi and Day 2000). Ettringite initially formed transforms partially or completely into kuzelite when all soluble sulfate is consumed. If excess alumina is present also other hexagonal calcium aluminate hydrates or hydrogarnet may be encountered (De Silva and Glasser 1992). The considerable amount of MgO present in blast furnace slags can also give rise to hydrotalcite-type phases with an ideal composition of Mg4Al2(OH)14∙3H2O but allowing extensive cationic and anionic substitution. Virtually no Mg2+ was observed to enter the C-S-H phase (Richardson et al. 1994; Richardson and Groves 1997). Upon exposure to air or in binders with very low carbonate contents, carbonation of the C4AH13 to hemicarboaluminate, C4AC0.5H12, will occur. More extensive carbonation due to prolonged exposure or due to substantial amounts of carbonates such as calcite present in the binder results in the formation of monocarboaluminate, C4ACH11 (Matschei et al. 2007c). In Figure 30 the phase assemblages of the AFm-type structure are presented in a ternary diagram with at the apices OH-AFm (C4AH13), SO4-AFm (kuzelite) and CO3-AFm (monocarboaluminate). Except from the limited solid solution between C4AH13 and kuzelite, the end members behave as separate phases from a mineralogical point of view. The carbonation of C4AH13 prevents the conversion of ettringite into kuzelite (Kuzel 1996), therefore at high levels of carbonation often ettringite can be found, while at low levels kuzelite is present (Massazza and Daimon 1992; Atkins et al. 1993). In the presence of calcite and at temperatures below 20 °C, also ettringite was observed to show substantial solid solution between the ideal SO4-ettringite and the CO3-ettringite end members (Barnett et al. 2001). Blended cement reaction products. The reaction product assemblages formed in the hydration of blended cements are similar to the compounds formed in the reaction between SCMs

Figure 30. Calculated phase assemblage between different AFm phases at 25 °C. A possible range of stoichiometry of hemicarboaluminate and the limited formation of ternary solid solutions are not shown (Pöllmann 2006) (modified after Matschei et al. 2007c).

256

Snellings, Mertens, Elsen

and lime. The underlying reason is that in most cases the overall bulk chemical composition of the binders are comparable, with the exception that Portland cement usually contains calcium sulfates and commonly incorporates some carbonate from adsorption of atmospheric CO2 or interblending with a limited amount of limestone. Compared to the hydration product assemblage of ordinary Portland cement, SCM addition mostly results in a variation in the relative proportions of the reaction products. At ambient conditions the hydration products in blended cements commonly comprise C-(A)-S-H and AFm-type phases, ettringite and when substantial MgO is available, hydrotalcite-type phases. The Ca(OH)2 content depends mainly on the blending ratio of SCM over Portland cement and the composition and activity of the SCM (Massazza 2001). When all Ca(OH)2 is consumed in the pozzolanic reaction, strätlingite may develop. The C-S-H phase in blended cements usually displays Ca/Si ratios in the range of lower than 1 to 1.8, i.e., lower than the observed C-S-H Ca/Si range in ordinary Portland cement of 1.2 to 2.1 (Richardson 1999). This is generally accepted to be due to the larger availability of Si and Al originating from the dissolving SCMs (Richardson 2008). Conjointly the mean (alumino)silicate chain length is increased from values between 2 to 5 in respectively young and mature PC pastes to 10 or more in some blended cements (Richardson 1999). The extent to which mean chain lengths are affected depends on the activity and composition of the SCM and the curing conditions. The C-S-H phase nanostructure in blended cements is considered to be most compatible with a defect-tobermorite structure, rather than a jennite-based structure (Cong and Kirkpatrick 1996a, 1996b; Richardson 2004). Tetrahedral substituent ions such as Al3+ can be incorporated into C-S-H only in the bridging tetrahedron (Richardson and Groves 1993; Andersen et al. 2006). Sorption or incorporation of alkali and sulfate ions at the C-(A)S-H surface is related to its Ca/Si ratio and Al content. Alkali sorption increases with decreasing Ca/Si ratio and increasing Al content (Hong and Glasser 1999), while sulfate adsorption decreases with decreasing Ca/Si (Matschei et al. 2007b). In addition to sorption behavior, also the morphology of the C-S-H phase is dependent on its composition (Richardson 2004). As Ca/ Si decreases and Al/Ca increases in C-A-S-H in GGBFS and MK blended cements a transition occurs from fibrillar, thin particles to sheet-like two-dimensional foils (Richardson and Groves 1997; Richardson 1999). Due to the presence of easily soluble calcium sulfates, ettringite is able to form in the initial hydration stages. At more advanced stages of hydration ettringite can be partially or completely transformed to kuzelite. This transformation is obviously controlled by the overall bulk SO3/Al2O3 ratio, but also, as mentioned before, by the CO2/Al2O3 ratio. Calculated phase assemblage variations in function of changing sulfate and carbonate levels as expected to occur in Portland cement hydrated at 25 °C are given in Figure 31. At low carbonate and sulfate levels hydrogarnet was calculated to be present in the product assemblage, at higher carbonate and sulfate levels hydrogarnet is destabilized (Matschei and Glasser 2010). In blended cements, strätlingite can be encountered when Ca(OH)2 is not present and SO3/Al2O3 is low (Grutzeck et al. 1981).

Hydration thermodynamics The recent development of a comprehensive and internally consistent thermodynamic database for cement hydrate compounds allows clarifying and eventually predicting the response of product assemblages on compositional changes of the binder (Matschei et al. 2007b). To illustrate the ultimate binder mineralogy, the product assemblages in the ternary CaO-SiO2Al2O3 system at 25 °C and water over solid ratio of 1:1 were explored following the methodology outlined by Lothenbach and Winnefeld (2006) based on a Gibbs energy minimization. For illustration, two different cases were considered. In one case calcium sulfates were absent, in the other 5 wt% of gypsum was added to the system. To visualize the phase relationships between the hexagonal hydrates and the C-S-H, the precipitation of the hydrogarnet solid solution series

Supplementary Cementitious Materials

257

Figure 31. Calculated phase assemblage of a hydrated mixture consisting of C3A, Ca(OH)2 and varying initial sulfate and carbonate ratios at 25 °C (modified after Matschei and Glasser 2010).

needed to be suppressed. Ideal solid solution between C-S-H of tobermorite type (Ca/Si = 0.83) and jennite (Ca/Si = 1.6) was assumed. No incorporation of Al into C-S-H was accounted for. Obviously, this model is a simplification and serves mainly to highlight the expected changes in product assemblages. Incorporation of sulfate, carbonate, chloride, alkali or magnesium or iron compounds would definitely alter the nature and extent of the product stability fields. CaO-SiO2-Al2O3. Upon inspection of the ternary diagram in Figure 32, the phases containing silica are C-S-H and strätlingite. Alumina is distributed over the C4AH13 and strätlingite phases. C-S-H is the most stable phase at low Ca levels and excess silica and alumina can be considered as unreacted SCM material. Increasing the Ca content from the field where C-S-H is the sole reaction product, a tieline is crossed connecting C-S-H of tobermorite composition (Ca/Si = 0.83) with the Al2O3 apex. In the corresponding field the C/S ratio of the mix is sufficiently high to allow the precipitation of strätlingite in the presence of alumina. When augmenting the Ca proportion eventually excess alumina will combine with Ca to form C4AH13, in addition strätlingite will start decomposing into C-S-H and C4AH13. Finally, at high Ca levels strätlingite disappears and the ultimate product assemblage is predicted to consist of C-S-H, C4AH13 and excess Ca(OH)2. In the rare case that alumina is prevalent over silica and sufficient Ca is available, C-S-H is destabilized in favor of strätlingite, C4AH13 and excess alumina. The presented ternary diagram confirms the observed incompatibility of Ca(OH)2 and strätlingite. Takemoto and Uchikawa (1980) indicated that the Ca2+ concentration needed for the precipitation of C-A-H phases is in general higher than for C-S-H phases, possibly explaining why C-A-H phases can usually be found outside the C-S-H layer enveloping the reacting SCM or clinker particles. CaO-SiO2-Al2O3-SO4. In the presence of sulfate the topology of the ternary diagram remains essentially unchanged (Fig. 33), only ettringite and kuzelite are stabilized over C4AH13. In the presence of C4AH13, a solid-solution member of the kuzelite-C4AH13 series

258

Snellings, Mertens, Elsen

Figure 32. Reaction product assemblages in the CaO-SiO2-Al2O3 ternary system (wt% based) at 25 °C and w/s ratio of 1:1. Str stands for strätlingite, Hc for C4AH13, CH and C-S-H are cement shorthand for Ca(OH)2 and calcium-silicate-hydrates. In a) excess silica and alumina is present respectively as quartz and gibbsite, in b) excess alumina is present as gibbsite.

Figure 33. Reaction product assemblages in the CaO-SiO2-Al2O3 ternary system (wt% based) at 25 °C and w/s ratio of 1:1, 5 wt% gypsum was added to the system. Str stands for strätlingite, Hc for C4AH13 and Ett for ettringite. Kuz corresponds with kuzelite and Kuz-ss with a solid solution member of the kuzelite-C4AH13 solid solution series. CH and C-S-H are cement shorthand for Ca(OH)2 and calciumsilicate-hydrates. In a) excess silica, alumina and gypsum are present, in b) excess alumina and gypsum, in c) excess alumina and d) excess gypsum.

Supplementary Cementitious Materials

259

is predicted to be present. Strätlingite is predicted to precipitate in the absence of Ca(OH)2. Higher CaO/SO3 and Al2O3/SO3 favor the formation of kuzelite over ettringite as corroborated by Figure 31. In general, at higher curing temperatures denser and more heterogeneously distributed hydrates are formed resulting in a coarser porosity (Kjellsen et al. 1991). Above 50 °C kuzelite is increasingly stabilized over ettringite and monocarboaluminate (Thomas et al. 2003; Christensen et al. 2004; Lothenbach et al. 2008b). The product assemblage formed in concrete cured at elevated temperature may thus change when temperatures are lowered under service. The stabilization of ettringite over kuzelite at lower temperatures may thus be a primary cause of delayed ettringite formation and the associated concrete deterioration (Famy et al. 2002). Matschei and Glasser (2010) reported that above 25 °C kuzelite formation is favored over C4AH13 or a solid solution member of the kuzelite-C4AH13 series. One of the benefits of thermodynamic calculations of the hydration processes is the prediction of the volume of solids present in the cement. Coupled with kinetic data on the consumption of reactants an evolving picture of the hydrate assemblage in a binder can be established. Assuming that the total volume of the binder paste remains constant, the increase in volume of the solid phases due to the formation of hydrated compounds allows evaluation of the total porosity of the system (Lothenbach et al. 2008a). Properties known to be roughly correlated with total porosity such as compressive strength or permeability can as such be estimated. The calculated solid phase evolution of a hydrating slag-silica fume cement as displayed in Figure 34 was shown to correspond well with experimental observations on the product assemblage (Lothenbach et al. 2009).

Figure 34. Modeled volume changes (cm³/100 g of unhydrated binder) during the hydration of low alkali blended cement containing 66.6% ground granulated blast furnace slag, 10% of silica fume and less than 23.3% of Portland cement (Lothenbach et al. 2009).

260

Snellings, Mertens, Elsen PROPERTIES OF MORTAR AND CONCRETE CONTAINING SUPPLEMENTARY CEMENTITIOUS MATERIALS

The physical bulk properties of SCM blended lime or cement binders depend both on mix design and curing conditions and on the inherent characteristics of the SCM. Therefore, the literature on performance shows a myriad of studies evaluating the effect of one or more of these factors. Variations in the inherent SCM characteristics often preclude direct comparisons between different studies, but some general trends can be observed. The emerging view is that the utilization of SCMs is regarded as beneficial in terms of performance, durability and sustainability. Instead of going into detail on these aspects for all separate SCMs, this section rather presents a brief generic overview of the effect of SCM addition on lime and Portland cement binders illustrated by specific examples. More comprehensive reviews centered on the effect of addition of different SCMs on the performance and durability of SCM-lime or SCM blended cements binders can be found in Malhotra and Mehta (1996), Hewlett (2001), Bensted and Barnes (2002) and Siddique (2008).

Properties of uncured mortar and concrete containing SCMs The properties of the freshly mixed and early cured mortar and concrete are important because they can exert a significant control on the ultimate performance and durability of the binder. In this respect the particle characteristics and the initial reactivity of SCMs can influence the water demand and setting time of the mortar or concrete. This is for instance of importance when low water/binder ratios are needed in the preparation of high-performance concrete and the use of superplasticizers becomes necessary. Water demand. The effect of SCMs on the amount of water needed for the binder to reach a specified workability or fluidity is largely depending on the particle characteristics of the SCM and its proportion in the mortar or concrete. Both the particle size distribution and the particle shape and porosity determine the specific surface and fineness of the SCM. In general, the higher the fineness or specific surface area and the more irregular the particle shape of SCM, the higher the water requirement of the blended binder is. SCMs such as silica fume, metakaolin or diatomite earths that show large specific surface areas typically increase the water demand substantially. Obviously, the higher the proportion of fine SCM particles added to the mix, the higher the water demand as illustrated in Table 4 for metakaolin blended cements (Badogiannis et al. 2005). To the contrary, the spherical particles and relatively low specific surface area encountered in fly ashes may result in lowered water requirements. Since the water/binder ratio is directly related to the binder porosity, increased binder water requirement commonly results in lowered strength performance and vice versa. Especially at higher replacement levels it may thus be essential to introduce superplasticizers to reduce the water/binder ratio. Setting time. SCMs can both increase and decrease the setting in blended cements. The end of setting is generally conceived to be related to the end of the induction period and the start of hardening due to C-S-H reaction product formation. In this respect the acceleration of alite hydration can decrease setting times in blended cements. However, this acceleration effect is neutralized when at higher cement replacement ratios the overall content in clinker hydration products is lowered due to the dilution effect of the SCM. Also increased water requirements at higher replacement ratios may extend the setting time. In metakaolin blended cement the initial and final setting times at low replacement ratios and low water demand were observed to be slightly reduced, while at higher replacement ratios and water demand the rate of setting was slowed down (Table 4, Badogiannis et al. 2005). Heat evolution. The heat evolved during hydration of blended cements strongly depends on the early reactivity of SCMs. Highly reactive SCMs such as silica fume or metakaolin increase the heat released during the initial dissolution period and the subsequent main hydration period (cf. Fig. 21). This has been attributed in part to the acceleration of clinker hydration and in part

Supplementary Cementitious Materials

261

Table 4. Variation of water demand and initial and final setting times for cements blended with metakaolin from differing sources. MK identifies with metakaolin in the sample index, followed by the respective replacement percentages of 10 or 20 wt% (Badogiannis et al. 2005). Setting time (min)

Sample 

Metakaolin (wt%)

Water demand (water/solid)

Initial

Final

PC MK1-10 MK2-10 MK3-10 MK4-10 MKC-10 MK1-20 MK2-20 MK3-20 MK4-20 MKC-20

– 10 10 10 10 10 20 20 20 20 20

27.5 29 29 32 32.5 31 32 31.5 38.5 41 37.5

105 75 85 105 155 95 105 110 120 205 140

140 130 130 160 180 130 160 165 160 230 170

to the extensive formation of supplementary products of the pozzolanic reaction. However, most SCMs of lower activity can be effectively used to reduce the hydration heat in large structures. The combined effect of clinker dilution and a more sluggish pozzolanic reaction lowers the maximum temperature reached and minimizes the risk of cracking as exemplified in Figure 35 for ground granulated blast furnace slag cement.

Properties of hardened mortar and concrete containing SCMs The properties of hardened SCM blended binders are strongly related to the development of the binder microstructure, i.e., to the distribution, type, shape and dimensions of both reaction products and pores. The general beneficial effects of SCM addition in terms of both strength performance and durability are mostly attributed to the pozzolanic reaction in which Ca(OH)2 is consumed to produce additional C-S-H and C-A-H reaction products. The formation of

Figure 35. Influence of ground granulated blast furnace slag replacement ratio on adiabatic temperature rise (modified after Wainwright and Tolloczko 1986). PC stands for Portland cement. PC stands for Portland cement.

Snellings, Mertens, Elsen

262

pozzolanic reaction products results in infilling of interstitial porosity and a refining of the pore size distribution or pore structure. Pore structure. The pore structure is one of the most important factors governing the durability in terms of attack by aggressive agents such as CO2, sulfates and chlorides (Luke 2002). It is generally observed that the addition of SCMs results in an increase of total porosity, but a consistent decrease in mean pore sizes. Capillary porosity, larger than 30-40 nm, is generally reduced and “gel porosity” increased, the latter being smaller than 10 nm and related to the typical distances between C-S-H particles (Takemoto and Uchikawa 1980). This can be explained by the infilling of coarse and capillary pores by C-S-H phases. In Table 5 the total pore volume and volume of pores smaller than 20 mm are compared for ordinary Portland cement and Portland cement blended with metakaolin. An overall increase of both pore volume and proportion of fine pores can be observed. Also the threshold pore radius, the radius below which the porosity sharply increases, is lowered considerably (Fig. 36) (Khatib and Wild 1996). The refinement of the pore structure generally results in lowered permeability and ionic diffusion coefficients, thus effectively improving the durability of the binder. Moreover, the coarse and capillary porosity have been observed to be closely correlated with the compressive strength, increasing porosities leading to decreased binder strength

Table 5. The effect of blending Portland cement with metakaolin on the pore volume and proportion of small pores in a blended cement (Khatib and Wild 1996). % of fine pores (radii < 20 mm)

Pore volume (mm³/g) Metakaolin (%) Age (days)  3 7 14 28 90

Metakaolin (%)

0

5

10

15

0

5

10

15

262 229.6 209.9 189.1 181.4

257.6 261.7 203.4 205.3 180.8

284.1 268.8 221 237.1 219.6

277.6 251.6 212.1 222.7 198.9

22.2 26.5 30.3 33.7 37.3

28.3 32.1 43 43.5 44.7

31 41 53.9 48.7 49.9

39.9 50.4 55.7 54.9 57.6

Figure 36. The introduction of metakaolin reduces the threshold pore radius significantly. The threshold radius is the pore radius below which the porosity sharply increases (Khatib and Wild 1996).

Supplementary Cementitious Materials

263

(Takemoto and Uchikawa 1980; Papayianni and Stefanidou 2006). The improvement of the cement microstructure is also visible at the interface zone between aggregate and binder in mortars and concrete. The interface zone in ordinary Portland cement is a region of low C-S-H and high ettringite and Ca(OH)2 concentrations between 25 and 100 mm in thickness with decreased microhardness. In blended cements both the amount of Ca(OH)2 and its preferential orientation is reduced (Larbi and Bijen 1990). Instead additional C-S-H replaces Ca(OH)2, fills porosity and reduces the thickness of the interface zone (Shannag 2000). The microhardness of the interface zone is increased (Asbridge et al. 2002) and the adhesion between binder and aggregate improved. Strength. The water/binder ratio is a factor of primary importance in governing strength development and ultimate strength of both SCM-lime and SCM-Portland cement based binders. As mentioned earlier, a close relationship exists between water/binder ratio and binder porosity on the one hand and binder porosity and strength on the other hand. The high water demand of very fine pozzolans such as silica fume may necessitate the use of water-reducing agents to lower the water requirement of the blend and improve the eventual binder strength. Strength development in SCM-lime binders depends on the one hand on the mix design, the SCM/lime and water/binder ratio, and the curing conditions, elevated temperature curing increases the rate of strength development but often lowers final strength, and on the other hand on the SCM reactivity. The latter can widely differ from SCM-type to SCM-type as exemplified in Figure 37 (Shi and Day 1995), depending on the particle characteristics, active phase content and chemical and mineralogical composition of the SCM. The ultimate strength of SCM-lime binders may exceed 20 MPa, which is high enough to serve for many common applications (Massazza 2002). In blended cement pastes, mortars and concrete the result of SCM incorporation on the development of strength is conceived to be controlled by the dilution effect, the filler effect, the hydration acceleration effect and the pozzolanic or hydraulic reaction. The first three

Figure 37. Compressive strength development of blends of lime (20%) and a variety of SCMs (80%). SCMs with hydraulic properties develop the highest strengths over time (Shi and Day 1995). NPB stands for natural vitreous pozzolan from Bolivia, NPG for natural vitreous pozzolan from Guatemala, LFA for low-lime fly ash, HFA for high-lime fly ash, the slag used was a typical ground granulated blast furnace slag.

264

Snellings, Mertens, Elsen

factors have similar effects as an inert filler and dominate the initial strength development. The contribution of the pozzolanic or hydraulic reactions to cement strength is usually developed in a later curing stage, depending on the SCM reactivity. In the large majority of blended cements initial lower strengths can be observed compared to the parent Portland cement. However, especially in the case of SCMs of higher fineness than the Portland cement, the decrease in early strength is usually less than what can be expected based on the dilution factor. This can be explained on the one hand by the contribution of the filler effect, in which small SCM grains fill in the interstitial space between the cement particles, resulting in a much denser binder matrix. On the other hand, the acceleration of the clinker hydration reactions (cf. Fig. 29) can also at least partially accommodate the loss of early strength. At later curing ages blended cements typically show a higher rate of strength development due to the supplementary formation of products of the pozzolanic or hydraulic reactions. Depending on the mix design and the SCM activity, in many cases the ultimate strength can become higher than that of the parent Portland cement as illustrated in Figure 38 for ground granulated slag cement. The rate of the pozzolanic and hydraulic reactions principally determines the moment when the blended cement strength exceeds the parent Portland cement strength. Highly reactive SCMs such as metakaolin or silica fume were observed to enhance even early strengths within one day (Sabir et al. 2001), while slags of much lower reactivity typically present positive strength contributions only after 14 to 28 days of curing (Lang 2002). It should be noted that because hydrated ordinary Portland cement only contains about 20 wt% of Ca(OH)2, at high cement replacement percentages of 40% or more by pozzolans, strength development may be hampered because of a general lack of Ca(OH)2 (e.g., Yilmaz et al. 2007).

Durability of mortar and concrete containing SCMs One of the main advantages of SCMs, and one of the early incentives to introduce SCMs into blended cements, is the significantly increased chemical resistance of the binder to the ingress and deleterious action of aggressive solutions. The improved durability of SCMblended binders enables to lengthen the service life of structures and reduces the costly and inconvenient need to replace deteriorated constructions. In general, one of the principal reasons of increased durability in SCM-blended cements is the lowered Ca(OH)2 content available to take part in deleterious reactions. Furthermore, the higher content in C-S-H binder phase with a reduced Ca/Si ratio results in finer pore size distributions and lower permeabilities

Figure 38. Development of compressive strength in concrete blended with various amounts of ground granulated blast furnace slag (Khatib and Hibbert 2005).

Supplementary Cementitious Materials

265

and in a thermodynamically more stable and chemically more resistant C-S-H phase with increased bonding capability of chlorine and alkaline ions. It is apparent that the replacement ratios of Portland cement by SCMs required for optimal durability are usually higher than the replacement ratios needed for optimal strength. Sulfate attack. The interaction of sulfate containing solutions with concrete structures can result in swelling, cracking and eventual structural failure. Sulfate attack involves the formation of expansive compounds in the reaction of sulfate with Ca(OH)2 to form gypsum or in combination with aluminates to form ettringite. The intensity of sulfate attack depends on the associated cation, increasing in the order Ca2+ < Na+ < Mg2+. In addition of ettringite formation induced by interaction with Ca-sulfate solutions, Na-sulfate leads to the additional formation of gypsum when reacted with Ca(OH)2. MgSO4 is very aggressive, not only leading to the formation of gypsum and ettringite but also to the disintegration of the C-S-H phase into brucite, gypsum and silica. SCM addition can reduce or eliminate the deleterious formation of expansive compounds by lowering the overall amount of Ca(OH)2, by reducing the diffusion rate of sulfate in the pore solution and possibly by increasing the chemical resistance of the C-S-H phase. Replacement percentages of 30-40% of Portland cement by vitreous natural pozzolans were observed to be very effective in reducing expansion upon sulfate attack (Fig. 39) (Massazza and Costa 1979). Lower additions do not completely consume Ca(OH)2, leaving a potential for expansion. Resistance to MgSO4 attack can only be improved at low (2%) concentrations of MgSO4. Chloride attack. Exposure to de-icing salts, seawater or salt-bearing groundwater may result in increased leaching of Ca(OH)2, higher binder porosity and lower strength. In addition, crystallization of salts in pores may cause expansion. In reinforced concrete, increased Cl−/ (OH)− ratios can lead to steel passivation and rebar corrosion. The lowered permeability of SCM-blended cements strongly reduces the diffusion rate of chloride ions into the binder matrix. Moreover, relatively large amounts of Cl− can be bound to C-S-H and AFm phases giving blended cements that contain higher amounts of these reaction products a larger chloride binder capacity (Balonis et al. 2010). Lowered Cl− concentrations in the depth profiles from a surface exposed to a chlorine bearing solutions in Figure 40 illustrate the improved resistance of blended concrete to chlorine attack (Chan and Ji 1999).

Figure 39. Reduction of mortar expansion due to attack of a 1% MgSO4 solution decreases with increasing Portland cement replacement by a vitreous natural pozzolan (modified after Massazza and Costa 1979).

266

Snellings, Mertens, Elsen

Carbonation. Reaction of hydration products with carbonate bearing solutions results in the formation of CaCO3 and silica and/or alumina gel. In porous binders, intense carbonation can result in a decreased pore solution pH, possibly leading to steel passivation in reinforced concrete. However, in relatively impermeable binders, carbonation is usually confined to the upper surface and does not progress into the binder matrix. Although a depleted reserve of Ca(OH)2 might render blended cements more susceptible to carbonation, in general the reduced permeability effectively counteracts the loss of buffering Ca(OH)2. Alkali-silica reaction. Alkali-silica reactions cause severe damage to concrete structures. Expansion is generally caused by the formation of an alkali-calcium-silica gel due to the interaction of alkalis present in the concrete pore solution with Ca(OH)2 and reactive silica from the aggregate. The expansion risk is especially high when cements containing high levels of alkali are used in combination with reactive aggregates. The utilization of SCMs can prevent alkali-silica reactions by reducing the availability of alkalis, lowering the pH and depleting Ca(OH)2. Although SCMs can contain relatively large amounts of alkali, in most cases the effective soluble amount of alkalis is rather low. Furthermore, the increased alkali-binding capability of the hydration products (i.e., C-A-S-H) enables to lower the alkalinity of the pore solution. However, it should be noted that at low replacement ratios below 20% the addition of SCMs can increase alkali-silica reactions with respect to the parent Portland cement. This pessimum behavior can be related to the supplementary release of alkalis from SCMs combined with an incomplete consumption of Ca(OH)2 (Hobbs 2002). More than 20% replacement is usually sufficient to avoid expansion. Figure 41 illustrates the expansion abatement in zeolite tuff blended cements, showing that expansion is strongly reduced when 20% or more zeolite tuff is added to the high-alkali cement (Feng and Peng 2005). In general, low-alkali SCMs containing both silica and alumina are observed to be most effective in minimizing expansion. Alumina-poor SCMs have less potential to reduce the alkali-silica reaction.

CONCLUSIONS The benefits of supplementary cementitious materials utilization in the cement and construction industry are threefold. First is the economic gain obtained by replacing a substantial part of the Portland cement by cheap natural pozzolans or industrial by-products. Second is the lowering of the blended cement environmental cost associated with the greenhouse gases emitted during Portland cement production. A third advantage is the durability improvement of the end product. Additionally, the increased blending of SCMs with Portland cement is of limited interference in the conventional production process and offers the opportunity to valorize and immobilize vast amounts of industrial and societal waste into construction materials. Supplementary cementitious materials can be conveniently classified according to a genetic classification scheme. A distinction is made between naturally occurring and artificial SCMs. The latter category is subdivided into intentionally thermally activated materials and byproducts of industrial processes. Detailed accounts on the physical, chemical and mineralogical characteristics for the various SCM groups and subgroups corroborate the view that there exists a close relationship between the material properties and their hydraulic and pozzolanic reactivity that promises to be quantified and predicted. To identify the material properties of importance and when they become essential in the pozzolanic reaction, a detailed knowledge on the pozzolanic reaction mechanism is needed. Recent developments regarding the dissolution kinetics and mechanisms of aluminosilicates in high pH environments are illustrated to be instrumental in obtaining more fundamental insights into the pozzolanic reaction. Improved knowledge on the kinetics and mechanism of the pozzolanic reaction will allow making better thermodynamic predictions of evolving reaction product assemblages over the course of hydration.

Supplementary Cementitious Materials

267

Figure 40. Chloride concentrations in function of depth from the concrete surface exposed to a Cl- rich solution after 30 days of exposure. The concrete water to binder ratio was 0.33 (modified after Chan and Ji 1999).

Figure 41. Alkali-silica expansion abatement in natural zeolite blended concrete at various Portland cement replacement percentages (Feng and Peng 2005).

The reaction product assemblages are observed to be mainly a function of the chemical composition of the reactive components in the reactant mixture. In contrast to the wide variability in SCMs, there exists only a relatively small range of thermodynamically stable or metastable compounds formed in the pozzolanic, hydraulic or hydration reactions. Recent developments in thermodynamic modeling allow an improved, more quantitative understanding of the parameters controlling the assemblage of reaction products to be obtained. A synopsis of the technological effects of using blended cements is included in the present paper. Properties of fresh and hardened mortar and concrete incorporating SCMs are compared. The effects of SCM incorporation are obviously strongly linked to the specific physical characteristics and can differ widely from slowly reacting blast furnace slag to highly reactive metakaolin. The emerging image is that the appropriate use of SCMs should be regarded as beneficial in terms of performance, durability and sustainability of the end product.

ACKNOWLEDGMENTS R. Snellings extends his gratitude towards the Research Foundation - Flanders (FWO) for financial support.

268

Snellings, Mertens, Elsen References

Ahmadi B, Shekarchi M (2010) Use of natural zeolite as a supplementary cementitious material. Cem Concr Comp 32:134-141 Akman MS, Mazlum F, Esenli F (1992) Comparative study of natural pozzolans used in blended cement production. ACI Special Publication 132:471-494 Alexander KM (1960) Reactivity of ultrafine powders produced from siliceous rocks. J Am Concr I 57:557-569 Al-Rawas AA, Hago AW, Corcoran TC, Al-Ghafri KM (1998) Properties of Omani artificial pozzolana (sarooj). Appl Clay Sci 13:275-292 Ambroise J, Gniewek J, Dejean J, Péra J (1987) Hydration of synthetic pozzolanic binders obtained by thermal activation of montmorillonite. Ceram Bull 66:1731-1733 Ambroise J, Maximilien S, Péra J (1994) Properties of metakaolin blended cements. Adv Cem Based Mater 1:161-168 Ambroise J, Murat M, Péra J (1985) Hydration reaction and hardening of calcined clays and related minerals. V. Extension of the research and general conclusions. Cem Concr Res 15:261-268 Andersen MD, Jakobsen HJ, Skibsted J (2006) A new aluminium-hydrate species in hydrated Portland cements characterized by 27Al and 29Si MAS NMR spectroscopy. Cem Concr Res 36:3-17 Asbridge AH, Page CL, Page MM (2002) Effects of metakaolin, water/binder ratio and interfacial transition zones on the microhardness of cement mortars. Cem Concr Res 32:1365-1369 Asher RR (1965) Volcanic construction materials in Idaho. Idaho Bureau of Mines and Geology 135:1-94 Atkins M, Lachowski EE, Glasser FP (1993) Investigation of solid and aqueous chemistry of 10-year-old Portland cement pastes; with and without silica modifier. Adv Cem Res 5:97-102 Ayub M, Yusuf M, Beg A, Faruqi FA (1988) Pozzolanic properties of burnt clays. Pakistan J Sci Ind R 31:1-5 Badogiannis E, Kakali G, Dimopoulou G, Chaniotakis E, Tsivilis S (2005) Metakaolin as a main cement constituent. Exploitation of poor Greek kaolins. Cem Concr Comp 27:197-203 Bakharev T (2005) Geopolymeric materials prepared using Class F fly ash and elevated temperature curing. Cem Concr Res 35:1224-1232 Bakolas A, Aggelakopoulou E, Moropoulou A, Anagnostopoulou S (2006) Evaluation of pozzolanic activity and physico-mechanical characteristics in metakaolin-lime pastes. J Therm Anal Calorim 84:157-163 Balonis M, Lothenbach B, Le Saout G, Glasser FP (2010) Impact of chloride on the mineralogy of hydrated Portland cement systems. Cem Concr Res 40:1009-1022 Baris YL, Simonato L, Saracci R, Skidmore JW, Artvinli M (1981) Malignant mesothelioma and radiological chest abnormalities in two villages in central Turkey: An epidemiological and environmental investigation. Lancet 317:984-987 Barnett SJ, Adam CD, Jackson A (2001) XRPD profile fitting investigation of the solid solution between ettringite, Ca6Al2(SO4)3(OH)12·26H2O, and carbonate ettringite, Ca6Al2(CO3)3(OH)12·26H2O. Cem Concr Res 31:13-17 Baronio G, Binda L (1997) Study of the pozzolanicity of some bricks and clays. Constr Build Mater 11:41-46 Battaglino G, Schippa G (1968) On a new method for evaluating the aptitude of a material to be employed as pozzolanic addition. Ind Ital Cem 38:175-178 Beedle SS, Groves GW, Rodger SA (1989) The effect of fine pozzolanic and other particles on hydration of C3S. Adv Cem Res 2:3-8 Bellotto M, Gualtieri A, Artioli A, Clark SM (1995) Kinetic study of the kaolinite-mullite reaction sequence. Part I: Kaolinite dehydroxylation. Phys Chem Miner 22:207-217 Benezet JC, Benhassaine A (1999) Grinding and pozzolanic reactivity of quartz powders. Powder Technol 105:167-171 Bensted J, Barnes P (2002) Structure and Performance of Cements. 2nd edition. Spon Press, London Bich C, Ambroise J, Péra J (2009) Influence of degree of dehydroxylation on the pozzolanic activity of metakaolin. Appl Clay Sci 44:194-200 Bijen J (1996) Benefits of slag and fly ash. Constr Build Mater 10:309-314 Biricik H, Aköz F, Berktay I, Tulgar AN (1999) Study of pozzolanic properties of wheat straw ash. Cem Concr Res 29:637-643 Blezard RG (2001) The history of calcareous cements. In: Lea’s Chemistry of Cement and Concrete. 4th edition. Hewlett P (ed) Butterworth-Heinemann, Oxford, p 1-24 Blum AE, Yund RA, Lasaga AC (1990) The effect of dislocation density on the dissolution rate of quartz. Geochim Cosmochim Acta 54:283-297 Bougara A, Lynsdale C, Milestone NB (2010) Reactivity and performance of blast furnace slags of different origin. Cem Concr Comp 32:319-324 Brady PV, Walther JV (1990). Kinetics of quartz dissolution at low temperatures. Chem Geol 82:253-264 Brantley SL (2008) Kinetics of mineral dissolution. In: Kinetics of Water-Rock Interaction. Brantley SL, Kubicki JD, White AF (eds) Springer, New York, p 151-210

Supplementary Cementitious Materials

269

Brough AR, Dobson CM, Richardson IG, Groves GW (1995) A study of the pozzolanic reaction by solid-state 29 Si nuclear magnetic resonance using selective isotopic enrichment. J Mater Sci 30:1671-1678 Bullard JW, Jennings HM, Livingston RA, Nonat A, Scherer GW, Schweitzer JS, Scrivener KL, Thomas JJ (2011) Mechanisms of cement hydration. Cem Concr Res 41:1208-1223 Cabrera J, Rojas MF (2001) Mechanism of hydration of the metakaolin-lime-water system. Cem Concr Res 31:177-182 Cameron WE (1977) Mullite: a substituted alumina. Am Mineral 62:747-755 Canpolat F, Yilmaz K, Köse MM, Sümer M, Yurdusev MA (2004) Use of zeolite, coal bottom ash and fly ash as replacement materials in cement production. Cem Concr Res 34:731-735 Caputo D, Liguori B, Colella C (2008) Some advances in understanding the pozzolanic activity of zeolites: The effect of zeolite structure. Cem Concr Comp 30:455-462 Cara S, Carcangiu G, Massidda L, Meloni P, Sanna U, Tamanini M (2006) Assessment of pozzolanic potential in lime-water systems of raw and calcined kaolinic clays from the Donnigazza Mine (Sardinia-Italy). J Appl Clay Sci 33:66-72 Çavdar A, Yetgin Ş (2007) Availability of tuffs from northeast of Turkey as natural pozzolan on cement, some chemical and mechanical relationships. Constr Build Mater 21:2066-2071 Chakchouk A, Trifi L, Samet B, Bouaziz S (2009) Formulation of blended cement: effect of process variables on clay pozzolanic activity. Constr Build Mater 23:1365-1373 Chan SYN, Ji X (1999) Comparative study of the initial surface absorption and chloride diffusion of high performance zeolite, silica fume and PFA concretes. Cem Concr Comp 21:293-300 Chen Y, Shah N, Huggins FE, Huffman GP (2005) Transmission electron microscopy investigation of ultrafine coal fly ash particles. Environ Sci Technol 39:1144-1151 Cheriaf M, Cavalcante Rocha J, Péra J (1999) Pozzolanic properties of pulverized coal combustion bottom ash. Cem Concr Res 29:1387-1391 Chipera SJ, Apps JA (2001) Geochemical stability of natural zeolites. Rev Mineral Geochem 45:117-161 Christensen AN, Jensen TR, Hanson JC (2004) Formation of ettringite, Ca6Al2(SO4)3(OH)12·26H2O, AFt, and monosulphate, Ca4Al2O6(SO4)·14H2O, AFm-14, in hydrothermal hydration of Portland cement and of calcium aluminum oxide—calcium sulphate dihydrate mixtures studied by in situ synchrotron X-ray powder diffraction. J Solid State Chem 177:1944-1951 Chung DDL (2002) Review: Improving cement-based materials by using silica fume. J Mater Sci 37:673-682 Chusilp N, Jaturapitakkul C, Kiattikomol K (2009) Utilization of bagasse ash as a pozzolanic material in concrete. Constr Build Mater 23:3352-3358 Colella C, de Gennaro M, Aiello R (2001) Use of zeolitic tuff in the building industry. Rev Mineral Geochem 45:551-587 Coles DG, Ragaini RC, Ondov JM, Fisher GL, Silberman D, Prentice BA (1979) Chemical studies of stack fly ash from a coal-fired power plant. Environ Sci Technol 13: 455-459 Collepardi M, Baldini G, Pauri M, Corradi M (1978) The effect of pozzolanas on the tricalcium aluminate hydration. Cem Concr Res 8:741-751 Cong X, Kirkpatrick RJ (1996b) 29Si MAS NMR study of the structure of calcium silicate hydrate. Adv Cem Based Mater 3:144-156 Cong X, Kirkpatrick RJ (1996a) 29Si and 17O NMR investigation of the structure of some crystalline calcium silicate hydrates. Adv Cem Based Mater 3:133-143 Cook DJ (1986a) Natural pozzolanas. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 1-39 Cook DJ (1986b) Calcined clay, shale and other soils. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 40-72 Cook DJ (1986c) Rice husk ash. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 171-196 Cordeiro GC, Filho RDT, Fairbairn EMR (2009b) Use of ultrafine rice husk ash with high-carbon content as pozzolan in high performance concrete. Mater Struct 42:983-992 Cordeiro GC, Filho RDT, Tavares LM, Fairbairn EMR (2008) Pozzolanic activity and filler effect of sugar cane bagasse ash in Portland cement and lime mortars. Cem Concr Res 30:410-418 Cordeiro GC, Filho RDT, Tavares LM, Fairbairn EMR (2009a) Ultrafine grinding of sugar cane bagasse ash for application as pozzolanic admixture in concrete. Cem Concr Res 39:110-115 Costa U, Massazza F (1974) Factors affecting the reaction with lime of pozzolans. Proceedings of the 6th International Congress on the Chemistry of Cement III-6:1-17 Criscenti LJ, Kubicki JD, Brantley SL (2006) Silicate glass and mineral dissolution: calculated reaction paths and activation energies for hydrolysis of a Q3 Si by H3O+ using ab initio methods. J Phys Chem A 110:198206 Damtoft JS, Lukasik J, Herfort D, Sorrentino D, Gartner EM (2008) Sustainable development and climate change initiatives. Cem Concr Res 38:115-127

270

Snellings, Mertens, Elsen

Davraz M, Gündüz L (2005) Engineering properties of amorphous silica as a new natural pozzolan for use in concrete. Cem Concr Res 35:1251-1261 Davraz M, Gündüz L (2008) Reduction of alkali silica reaction risk in concrete by natural (micronized) amorphous silica. Constr Build Mater 22:1093-1099 Day RL (1990) Pozzolans for use in low-cost housing. International Development Research Centre, Ottawa, Canada Day RL, Shi C (1994) Influence of the fineness of pozzolan on the strength of lime natural-pozzolan cement pastes. Cem Concr Res 24:1485-1491 De Sensale GR (2006) Strength development of concrete with rice-husk ash. Cem Concr Comp 28:158-160 De Silva PS, Glasser FP (1992) Pozzolanic activation of metakaolin. Adv Cement Res 4:167-178 Degirmenci N, Yilmaz A (2009) Use of diatomite as partial replacement for Portland cement in cement mortars. Constr Build Mater 23:284-288 Demoulian E, Gourdin P, Hawthorn, F, Vernet C (1980) Influence de la composition chimique et de la texture des laitiers sur leur hydraulicité. Proceedings of the 7th International Congress on the Chemistry of Cement II-III:105-111 Dihr RK (1986) Pulverised-fuel ash. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 197-255 Divet L, Randriambololona R (1998) Delayed ettringite formation: the effect of temperature and basicity on the interaction of sulphate and C-S-H phase. Cem Concr Res 28:357-363 Double DD, Hellawell A, Perry SJ (1978) The hydration of Portland cement. Proc R Soc London A 359:435-451 Douglas E, Elola A, Malhotra VM (1990) Characterisation of ground granulated blast furnace slag and fly ashes and their hydration in Portland cement blends. Cem Concr Aggr 12:38-46 Dove PM (1999) Dissolution kinetics of quartz in aqueous mixed cation solutions. Geochim Cosmochim Acta 63:3715-3727 Dove PM, Crerar DA (1990) Kinetics of quartz dissolution in electrolyte solutions using a hydrothermal mixed flow reactor. Geochim Cosmochim Acta 54:955-969 Dove PM, Han N, De Yoreo JJ (2005) Mechanisms of classical crystal growth theory explain quartz and silicate dissolution behavior. P Natl Acad Sci USA 102:15357-15362 Drits VA, Besson G, Muller F (1995) An improved model for structural transformations of heat-treated aluminous dioctahedral 2:1 layer silicates. Clays Clay Miner 43:718-731 Držaj B, Hočevar S, Slokan M, Zajc A (1978) Kinetics and mechanism of reaction in the zeolitic tuff-CaO-H2O systems at increased temperature. Cem Concr Res 8:711-720 Dudas MJ, Warren CJ (1987) Submicroscopic model of fly ash particles. Geoderma 40:101-114 Efstathiadis E (1978) Greek Concrete of Three Millenniums. Hellenic Ministry of Public Works, Athens. Emmerich K (2010) Thermal analysis for characterization and processing of industrial minerals. In: Industrial Mineralogy. EMU Notes in Mineralogy, 9. Christidis G (ed) Eötvös University Press, Budapest, p 129-170 Erdem TK, Meral Ç, Tokyay M, Erdogan TY (2007) Use of perlite as a pozzolanic addition in producing blended cements. Cem Concr Comp 29:13-21 Erdoğdu K, Türker P (1998) Effects of fly ash particle size on strength of Portland cement fly ash mortars. Cem Concr Res 28: 1217-1222 Escalante JI, Gomez LY, Johal KK, Mendoza G, Mancha H, Mendez J (2001). Reactivity of blast furnace slag in Portland cement blends hydrated under different conditions. Cem Concr Res 31:1403-1409 Famy C, Scrivener KL, Atkinson A, Brough AR (2002) Effects of an early or a late heat treatment on the microstructure and composition of inner C-S-H products of Portland cement mortars. Cem Concr Res 32:269-278 Felipe MA, Xiao Y, Kubicki JD (2001) Molecular orbital modeling and transition state theory in geochemistry. Rev Miner Geochem 42:485-531 Feng NQ, Chan SYN, He ZS, Tsang MKC (1997) Shale ash concrete. Cem Concr Res 27:279-291 Feng NQ, Peng GF (2005) Applications of natural zeolite to construction and building materials in China. Constr Build Mater 19:579-584 Fernandez R, Martirena F, Scrivener KL (2011) The origin of the pozzolanic activity of calcined clay minerals: A comparison between kaolinite, illite and montmorillonite. Cem Concr Res 41:113-122 Ferreira C, Ribeiro A, Ottosen L (2003) Possible applications for municipal solid waste fly ash. J Hazard Mater B96:201-216 Fidjestol P, Lewis R (2001) Microsilica as an addition. In: Lea’s Chemistry of Cement and Concrete. Hewlett PC (ed) Butterworth-Heinemann, Oxford, p 679-712 Filippidis A, Georgakopoulos A (1992) Mineralogical and chemical investigation of fly ash from the Main and Northern lignite fields in Ptolemais, Greece. Fuel 1992:373-376 Fraay ALA, Bijen JM, De Haan YM (1989) The reaction of fly ash in concrete a critical examination. Cem Concr Res 19:235-246

Supplementary Cementitious Materials

271

Fragoulis D, Chaniotakis E, Stamatakis MG (1997) Zeolitic tuffs of Kimolos island, Aegean Sea, Greece and their industrial potential. Cem Concr Res 27:889-905 Fragoulis D, Stamatakis MG, Papageorgiou D, Chaniotakis E (2005) The physical and mechanical properties of composite cements manufactured with calcareous and clayey Greek diatomite mixtures. Cem Concr Res 27:205-209 Frearson JPH, Uren JM (1986) Investigation of a granulated blast furnace slag containing merwinite crystallisation. Proc 2nd Int Conf Use of Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete II:1401-1421 Frias M, Rodriguez O, Vegas I, Vigil R (2008) Properties of calcined clay waste and its influence on blended cement behavior. J Am Ceram Soc 91:1226-1230 Gartner EM (2004) Industrially interesting approaches to ‘low-CO2’ cements. Cem Concr Res 34:1489-1498 Gartner EM, Young JF, Damidot DA, Jawed I (2002) Hydration of Portland cement. In: Structure and Performance of Cements. Bensted J, Barnes P (eds) Spon Press, London, p 57-113 Gieré R, Carleton LE, Gregory RL (2003) Micro- and nanochemistry of fly ash from a coal-fired power plant. Am Mineral 88:1853-1865 Ginstling AM, Brounhstein BI (1950) On diffusion kinetics in chemical reactions taking place in spherical powder grains. Zhur Priklad Khim 23:1249-1259 (in Russian) Glasser FP, Diamond S, Roy DM (1987) Hydration reactions in cement pastes incorporating fly ash and other pozzolanic materials. Proc Mater Res Soc Symp 86:167-186 Gomes S, François M (2000) Characterization of mullite in silicoaluminous fly ash by XRD, TEM, and 29Si MAS NMR. Cem Concr Res 30:175-181 Gomes S, François M, Abdelmoula M, Refait Ph., Pellisier C, Evrard O (1999) Characterization of magnetite in silico-aluminous fly ash by SEM, TEM, XRD, magnetic susceptibility, and Mössbauer spectroscopy. Cem Concr Res 29:1705-1711 Goodarzi F (2006) Characteristics and composition of fly ash from Canadian coal-fired power plants. Fuel 85:1418-1427 Goto S, Fujimori H, Tsuda T, Ooshiro K, Yamamoto S, Ioku K (2007) Structure analysis of glass of CaO-Al2O3SiO2 system by means of NMR. Proc 12th Int Congr Chem Cement Concrete Gottardi G, Obradovic J (1978) Sedimentary zeolites in Europe. Fortschr Mineral 56:316-366 Greenberg SA (1961) Reaction between silica and calcium hydroxide solutions, I. Kinetics in the temperature range 30 to 85 °C. J Phys Chem 65:12-16 Grubb DG, Guimaraes MS, Valencia R (2000) Phosphate immobilization using an acidic type F fly ash. J Hazard Mater 76:217-236 Grutzeck MW, Roy DM, Scheetz BE (1981) Hydration mechanisms of high-lime fly ash in Portland-cement composites. Proc Mater Res Symp MRS-N 92–101 Grzeszczyk S, Lipowski G (1997) Effect of content and particle size distribution of high-calcium fly ash on the rheological properties of cement pastes. Cem Concr Res 27:907-916 Habert G, Billard C, Rossi P, Chen C, Roussel N (2010) Cement production technology improvement compared to factor 4 objectives. Cem Concr Res 40:820-826 Habert G, Choupay N, Escadeillas G, Guillaume D, Montel JM (2009) Clay content of argillites: Influence on cement based mortars. Appl Clay Sci 43:322-330 Hamdan H, Muhid MNM, Endud S, Listiorini E, Ramli Z (1997) 29Si MAS NMR, XRD and FESEM studies of rice husk silica for the synthesis of zeolites. J Non-cryst Solids 211:126-131 Hamilton JP, Brantley SL, Pantano CG, Criscenti LJ, Kubicki JD (2001) Dissolution of nepheline, jadeite and albite glasses: Toward better models for aluminosilicate dissolution. Geochim Cosmochim Acta 65:36833702 Hanna KM, Afify A (1974) Evaluation of the activity of pozzolanic materials. J Appl Chem Biotechn 24:751757 Hay RA, Sheppard RL (2001) Formation of zeolites in open hydrologic systems. Rev Mineral Geochem 45:261275 He C, Mackovicky E, Osbaeck B (1996) Thermal treatment and pozzolanic activity of Na- and Camontmorillonite. Appl Clay Sci 10:351-368 He C, Mackovicky E, Osbaeck B (2000) Thermal stability and pozzolanic activity of raw and calcined mixedlayer mica/smectite. Appl Clay Sci 17:141-161 He C, Makovicky E, Osbaeck B (1994) Thermal stability and pozzolanic activity of calcined kaolin. Appl Clay Sci 9:165-187 He C, Makovicky E, Osbaeck B (1995) Thermal stability and pozzolanic activity of raw and calcined illite. Appl Clay Sci 9:337-354 He C, Osbaeck B, Makovicky E (1995) Pozzolanic reactions of six principal clay minerals: activation, reactivity assessment and technological effects. Cem Concr Res 25:1691-1702

272

Snellings, Mertens, Elsen

Henson NJ, Cheetham AK, Gale JD (1994) Theoretical calculations on silica frameworks and their correlation with experiment. Chem Mater 6:1647-1650 Hewlett PC (2001) Lea’s Chemistry of Cement and Concrete. 4th edition. Butterworth-Heinemann, Oxford Hilali EA, Jeannette A, Nataf M (1981) Pozzolanas. Mines, Geologie et Energie, Maroc 49:195-208 Hobbs DW (2002) Alkali-silica reaction in concrete. In: Structure and Performance of Cements. Bensted J, Barnes P (eds) Spon Press, London, p 265-281 Hong SY, Glasser FP (1999) Alkali binding in cement pastes: Part I. The C-S-H phase. Cem Concr Res 29:18931903 Hooton RD (2008) Bridging the gap between research and standards. Cem Concr Res 38:247-258 Hooton RD, Emery JJ (1983) Glass content determination and strength development predictions for vitrified blast furnace slag. ACI Special Publication 79:943-962 Hower JC, Rathbone RF, Robertson JD, Peterson G, Trimble AS (1999) Petrology, mineralogy, and chemistry of magnetically-separated sized fly ash. Fuel 78:197-203 Hower JC, Trimble AS, Eble CF, Palmer CA, Kolker A (2000) Characterization of fly ash from low-sulfur coal sources: Partitioning of carbon and trace elements with particle size. Energ Source 21:511-525 Huizhen L (1992) Effect of structure and composition on reactivity of zeolite-tuff used as blending material of Portland cement. Proceedings of the 9th International Congress on the Chemistry of Cement 128-134 Idorn MG (1997) Concrete Progress from Antiquity to the Third Millenium. Telford, London Iler RK (1979) The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties, and Biochemistry. John Wiley and Sons, New York Ilic M, Cheeseman Ch, Sollars Ch, Knight J (2003) Mineralogy and microstructure of sintered lignite coal fly ash. Fuel 82:331-336 Ish-Shalom M, Bentur A, Grinberg T (1980) Cementing properties of oil-shale ash: I. Effect of burning method and temperature. Cem Concr Res 10:799-807 James J, Rao MS (1986) Reactivity of rice husk ash. Cem Concr Res 16:296-302 Jander W (1927) Reaktionen im festen Zustande bei höheren Temperaturen. Reaktionsgeschwindigkeiten endotherm verlaufender Umsetzungen. Z Anorg Allg Chem 163:1-30 Janotka I, Krajči L, Dzivak M (2003) Properties and utilization of zeolite-blended Portland Cements. Clays Clay Miner 51:616-624 Jauberthie R, Rendell F, Tamba S, Cisse I (2000) Origin of the pozzolanic effect of rice husks. Constr Build Mater 14:419-423 Jo BW, Kim CH, Tae GH, Park JB (2007) Characteristics of cement with nano-SiO2 particles. Constr Build Mater 21:1351-1355 Johansson S, Andersen PJ (1990) Pozzolanic activity of calcined moler clay. Cem Concr Res 20:447-452 Jones TR (2002) Metakaolin as a pozzolanic addition to concrete. In: Structure and Performance of Cements. Bensted J, Barnes P (eds) Spon Press, London, p 372-398 Joshi RC, Lohtia RP (1997) Fly ash in concrete production, properties and uses. Advances in Concrete Technology 2. Gordon and Breach Science Publishers, Amsterdam Juilland P, Gallucci E, Flatt R, Scrivener K (2010) Dissolution theory applied to the induction period in alite hydration. Cem Concr Res 40:831-844 Justnes H (2002) Condensed silica fume as a cement extender. In: Structure and Performance of Cements. Bensted J, Barnes P (eds) Spon Press, London, p 399-408 Justnes H (2007) Silica fume in high-quality concrete - A review of mechanism and performance. Proc 9th CANMET/ACI Int Conf on Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, p 63-78 Justnes H, Meland I, Bjoergum JO (1990) Nuclear magnetic resonance (NMR) - a powerful tool in cement and concrete research. Adv Cem Res 3:105-110 Kaid N, Cyr M, Julien S, Khelafi H (2009) Durability of concrete containing a natural pozzolan as defined by a performance-based approach. Constr Build Mater 23:3457-3467 Kakali G, Perraki T, Tsivilis S, Badogiannis E (2001) Thermal treatment of kaolin: the effect of mineralogy on the pozzolanic activity. Appl Clay Sci 20:73-80 Kantsepolsky IS, Terekhovich SV, Khlebo AP Ronchinsky EM (1969) Frost and atmosphere resistance of solutions in pozzolan Portland cements. Tr Alma-At Nauk Issled Proekt Inst Stroit Mater 9:119-125 (in Russian) Khatib JM, Hibbert JJ (2005) Selected engineering properties of concrete incorporating slag and metakaolin. Constr Build Mater 19:460-472 Khatib JM, Wild S (1996) Pore size distribution of metakaolin paste. Cem Concr Res 26:1545-1553 Kirby CS, Rimstidt JD (1993) Mineralogy and surface properties of municipal solid waste ash. Environ Sci Technol 27:652-660 Kitsopoulos KP, Dunham AC (1996) Heulandite and mordenite-rich tuffs from Greece: a potential source for pozzolanic materials. Mineral Deposita 31:576-583

Supplementary Cementitious Materials

273

Kjellsen KO, Detwiler RJ, Gjørv OE (1991) Development of microstructures in plain cement pastes hydrated at different temperatures. Cem Concr Res 21:179-189 Kjellsen KO, Wallevik OH, Hallgren M (1999) On the compressive strength development of high performance concrete and paste - effect of silica fume. Mater Struct 32:63-69 Knauss KG, Wolery TJ (1988) The dissolution kinetics of quartz as a function of pH and time at 70 °C. Geochim Cosmochim Acta 52:43-53 Kolay PK, Singh DN (2001) Physical, chemical, mineralogical, and thermal properties of cenospheres from an ash lagoon. Cem Concr Res 31:539-542 Kondo R, Lee K, Diamon M (1976) Kinetics and mechanisms of hydrothermal reaction in lime-quartz-water systems. J Ceram Soc Jpn 84:573-578 Korpa A, Kowald T, Tretting R (2008) Hydration behavior, structure and morphology of hydration phases in advanced cement-based systems containing micro and nanoscale pozzolanic additives. Cem Concr Res 38:955-962 Koukouzas N, Hämäläinen J, Papanikolaou D, Tourunen A, Jäntti T (2007) Mineralogical and elemental composition of fly ash from pilot scale fluidised bed combustion of lignite, bituminous coal, wood chips and their blends. Fuel 86:2186-2193 Kurdowski W, Nocun-Wczelik W (1983) The tricalcium silicate hydration in the presence of active silica. Cem Concr Res 13:341-348 Kutchko BG, Kim AG (2006) Fly ash characterization by SEM-EDS. Fuel 85:2537-2544 Kuzel HJ (1996) Initial hydration reactions and mechanisms of delayed ettringite formation in Portland cements. Cem Concr Comp 18:195-203 Lang E (2002) Blast furnace cements. In: Structure and Performance of Cements. 2nd edition. Bensted J, Barnes P (eds) Spon Press, London, 310-325 Larbi JA, Bijen JM (1990) Orientation of calcium hydroxide at the Portland cement paste-aggregate interface in mortars in the presence of silica fume: A contribution. Cem Concr Res 20:461-470 Lasaga AC (1998) Kinetic theory in the earth sciences. Princeton University Press, Chichester. Lasaga AC, Lüttge A (2005) Kinetic justification of the solubility product: application of a general kinetic dissolution model. J Phys Chem B 109:1635-1642 Lavat AE, Trezza MA, Poggi M (2009) Characterization of ceramic roof tile wastes as pozzolanic admixture. Waste Manage 29:1666-1674 Le Maître RW, Bateman P, Dudek A, Keller J, Lameyre J, Le Bas MJ, Sabine PA, Schmid R, Sorensen H, Streckeisen A, Woolley AR, Zanettin B (1989). A Classification of Igneous Rocks and Glossary of Terms, Recommendations of the International Union of Geological Sciences, Subcommission on the Systematics of Igneous Rocks. Blackwell Scientific, Oxford Lea FM (1970) The Chemistry of Cement and Concrete. 3rd Edition. Chemical Publishing, New York Liebig E, Althaus E (1997) Kaolinite and montmorillonite as pozzolanic components in lime mortars: untreated and after thermal activation. ZKG Int 50:282-290 Liebig E, Althaus E (1998) Pozzolanic activity of volcanic tuff and suevite: effects of calcination. Cem Concr Res 28:567-575 Liguori B, Caputo D, Marroccoli M, Colella C (2003) Evaluation of zeolite-bearing tuffs as pozzolanic addition for blended cements. ACI Special Publications 221:319-333 Lilkov V, Dimitrova E, Petrov OE (1997) Hydration process of cement containing fly ash and silica fume: the first 24 hours. Cem Concr Res 1997:577-588 Lindgreen H, Geiker M, Krøyer H, Springer N, Skibsted J (2008) Microstructure engineering of Portland cement pastes and mortars through addition of ultrafine layer silicates. Cem Concr Comp 30:686-699 Livingston RA, Schweitzer JS, Rolfs C, Becker HW, Kubsky S (2001) Characterization of the induction period in tricalcium silicate hydration by nuclear resonance reaction analysis. J Mater Res 16:687-693 Lothenbach B, Gruskovnjak A (2007) Hydration of alkali-activated slag: thermodynamic modelling. Adv Cem Res 19:81-92 Lothenbach B, Le Saout G, Gallucci E, Scrivener K (2008b) Influence of limestone on the hydration of Portland cements. Cem Concr Res 38:848-860 Lothenbach B, Matschei T, Möschner G, Glasser FP (2008a) Thermodynamic modelling of the effect of temperature on the hydration and porosity of Portland cement. Cem Concr Res 38:1-18 Lothenbach B, Scrivener K, Hooton RD (2011) Supplementary cementitious materials. Cem Concr Res 41:1244-1256 Lothenbach B, Winnefeld F (2006) Thermodynamic modelling of the hydration of Portland cement. Cem Concr Res 36:209-226 Lothenbach B, Winnefeld F, Alder C, Wieland E, Lunk P (2007) Effect of temperature on the pore solution, microstructure and hydration products of Portland cement pastes. Cem Concr Res 37:483-491 Ludwig U, Schwiete HE (1963) Lime combination and new formations in the trass-lime reactions. Zem-KalkGips 10:421-431

274

Snellings, Mertens, Elsen

Luke K (2002) Pulverized fuel ash as a cement extender. In: Structure and Performance of Cements, 2nd edition. Bensted J, Barnes P (eds.) Spon Press, London, p 353-371 Luke K (2007) The effect of natural zeolites on the composition of cement pore fluids at early ages. Proceedings of the 11th International Congress on the Chemistry of Cement Malhotra VM (1993) Fly ash, slag, silica fume, and rice-husk ash in concrete: a review. Concr Int 15:23-28 Malhotra VM, Mehta PK (1996) Pozzolanic and Cementitious Materials. Overseas Publishers Association, Amsterdam Maltais Y, Marchand J (1997) Influence of curing temperature on cement hydration and mechanical strength development of fly ash mortars. Cem Concr Res 27:1009-1020 Manz OE (1998) Coal fly ash: A retrospective and future look. Energeia 9:1-3 http://www.caer.uky.edu/energeia/ PDF/vol9-2.pdf Massazza F (1974) Chemistry of pozzolanic additions and mixed cements. Proceedings of the 6th International Congress on the Chemistry of Cement, 1-65 Massazza F (2001) Pozzolana and pozzolanic cements. In: Lea’s Chemistry of Cement and Concrete. Hewlett PC (ed) Butterworth-Heinemann, Oxford, p 471-636 Massazza F (2002) Properties and applications of natural pozzolanas. In: Structure and Performance of Cements, 2nd edition. Bensted J, Barnes P (eds.) Spon Press, London, p 326-352 Massazza F, Costa U (1979) Aspects of the pozzolanic activity and properties of pozzolanic cements. Il Cemento 76:3-18 Massazza F, Daimon M (1992) Conference report: Chemistry of hydration of cements and cementitious systems. Proceedings of the 9th International Congress on the Chemistry of Cement p 383-446 Massazza F, Testolin M (1983) Trimethylsilylation in the study of pozzolana-containing pastes. Il Cemento 80:49-62 Matschei T, Glasser FP (2010) Temperature dependence, 0 to 40 °C, of the mineralogy of Portland cement paste in the presence of calcium carbonate. Cem Concr Res 40:763-777 Matschei T, Lothenbach B, Glasser FP (2007a) The role of calcium carbonate in cement hydration. Cem Concr Res 37:551-558 Matschei T, Lothenbach B, Glasser FP (2007b) Thermodynamic properties of Portland cement hydrates in the system CaO-Al2O3-SiO2-CaSO4-CaCO3-H2O. Cem Concr Res 37:1379-1410 Matschei T, Lothenbach B, Glasser FP (2007c) The AFm phase in Portland cement. Cem Concr Res 37:118-130 Mehta PK (1977) Properties of blended cements, cements made from rice-husk ash. J Am Concrete I 74:440-442 Mehta PK (1981) Studies on blended Portland cements containing Santorin earth. Cem Concr Res 11:507-518 Mehta PK (1986) Condensed silica fume. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 134-170 Mehta PK (1987) Natural pozzolans: Supplementary cementing materials for concrete. CANMET Special Publication 86:1-33 Mehta PK (1989) Pozzolanic and cementitious by-products in concrete. Another look. ACI Special Publication 114:1-44 Mertens G, Snellings R, Van Balen K, Bicer-Simsir B, Verlooy P, Elsen J (2009) Pozzolanic reactions of common natural zeolites with lime and parameters affecting their reactivity. Cem Concr Res 39:233-240 Meyer C (2009) The greening of the concrete industry. Cem Concr Comp 31:601-605 Mielenz RC, White LP, Glantz OJ (1950) Effect of calcination on natural pozzolanas. Symp on Use of Pozzolanic Materials in Mortars and Concrete. ASTM Special Technical Publication 99:43-92 Millet J, Hommey R (1974) Etude minéralogique des pâtes pouzzolanes-chaux. Bull Liaison Ponts et Chaussées 74:59-63 Mondragon F, Rincon F, Sierra L, Escobar J, Ramirez J, Fernandez J (1990) New perspectives for coal ash utilization: synthesis of zeolitic materials. Fuel 69:263-266 Morales EV, Villar-Cociña E, Frias M, Santos SF, Savastano H (2009) Effects of calcining conditions on the microstructure of sugar cane waste ashes (SCWA): Influence in the pozzolanic activation. Cem Concr Comp 31:22-28 Morrow CP, Nangia S, Garrison J (2009) Ab initio investigation of dissolution mechanisms in aluminosilicate minerals., J Phys Chem A 113:1343-1352 Mortureux B, Hornain H, Gautier E, Regourd M (1980) Comparaison de la réactivité de differentes pouzzolanes. Proceedings of the 7th International Congress on the Chemistry of Cement IV:110-115 Mostafa NY, El-Helamy SAS, Al-Wakeel EI, El-Korashy SA, Brown PW (2001) Characterization and evaluation of the pozzolanic activity of Egyptian industrial by-products. I: Silica fume and dealuminated kaolin. Cem Concr Res 31:467-474 Murat M (1983) Hydration reaction and hardening of calcined clays and related minerals. I.Preliminary investigation on metakaolinite. Cem Concr Res 13:259-266 Murayama N, Yamamoto H, Shibata J (2002) Mechanism of zeolite synthesis from coal fly ash by alkali hydrothermal reaction. Int J Miner Process 64:1-17

Supplementary Cementitious Materials

275

Naidenov V (1991) Rapid hardening cement-gypsum composites for shotcreting on the base of Bulgarian raw materials: Part I. Introduction, materials, design of the compositions, strength and deformability. Cem Concr Res 21:896-904 Nair DG, Fraaij A, Klaassen AAK, Kentgens APM (2008) A structural investigation relating pozzolanic activity of rice husk ashes. Cem Concr Res 38:861-869 Nangia S, Garrison BJ (2008) Reaction rates and dissolution mechanisms of quartz as a function of pH. J Phys Chem A 112:2027-2033 Nathan Y, Dvorachek M, Pelly I, Mimran U (1999) Characterization of coal fly ash from Israel. Fuel 78:205-213 Oelkers EH (2001) General kinetic description of multioxide silicate mineral and glass dissolution. Geochim Cosmochim Acta 65:3703-3719 Ogawa K, Uchikawa H, Takemoto K, Yasui I (1980) The mechanism of the hydration in the system C3S pozzolanas. Cem Concr Res 10:683-696 Pal SC, Mukherjee A, Pathak SR (2003) Investigation of hydraulic activity of ground granulated blast furnace slag in concrete. Cem Concr Res 33:1481-1486 Papadakis VG (1999) Effect of fly ash on Portland cement systems Part I. Low-calcium fly ash. Cem Concr Res 29:1727-1736 Papadakis VG (2000) Effect of fly ash on Portland cement systems Part II. High-calcium fly ash. Cem Concr Res 30:1647-1654 Papayianni I, Stefanidou M (2006) Strength-porosity relationships in lime-pozzolan mortars. Constr Build Mater 20:700-705 Papayianni I, Stefanidou M (2007) Durability aspects of ancient mortars of the archeaological site of Olynthos. J Cult Herit 8:193-196 Payá J, Monzó J, Borrachero MV, Peris E, González-López E (1997) Mechanical treatments of fly ashes. Part III: studies on strength development of ground fly ashes (GFA) - cement mortars. Cem Concr Res 27:13651377 Pelletier L, Winnefeld F, Lothenbach B (2010) The ternary system Portland cement-calcium sulphoaluminate clinker-anhydrite: Hydration mechanism and mortar properties. Cem Concr Comp 32:497-507 Pelmenschikov A, Leszczynski J, Petterson LGM (2001) Mechanism of dissolution of neutral silica surfaces: Including effect of self-healing. J Phys Chem A 105:9528-9532 Pelmenschikov A, Strandh H, Petterson LGM, Leszczynski J (2000) Lattice resistance to hydrolysis of Si-OSi bonds of silicate minerals: Ab initio calculations of a single water attack onto the (001) and (111) b cristobalite surfaces. J Phys Chem B 104:5779-5783 Péra J, Ambroise J, Messi A (1998) Pozzolanic activity of calcined laterite. Silicates Industriels 7-8:87-106 Péra J, Amrouz A (1998) Development of highly reactive metakaolin from paper sludge. Adv Cem Based Mater 7:49-56 Péra J, Boumaza R, Ambroise J (1997) Development of a pozzolanic pigment from red mud. Cem Concr Res 27:1513-1522 Perraki Th, Kakali G, Kontoleon F (2003) The effect of natural zeolites on the early hydration of Portland cement. Microporous Mesoporous Mater 61:205-212 Petrovic I, Navrotsky A, Davis ME, Zones SI (1993) Thermochemical study of the stability of frameworks in high silica zeolites. Chem Mater 5:1805-1813 Piccione PM, Laberty C, Sanyuan Y, Camblor MA, Navrotsky A (2000) Thermochemistry of pure silica zeolites. J Phys Chem B 104:10001-10011 Pöllmann H (2006) Syntheses, properties and solid solution of ternary lamellar calcium aluminate hydroxy salts (AFm-phases) containing SO42−, CO32− and OH−. Neues JB Miner Abh 128:173-181. Puertas F, Fernández-Jiménez A (2003) Mineralogical and microstructural characterization of alkali-activated fly ash/slag pastes. Cem Concr Res 25:287-292 Queralt I, Querol X, López-Soler A, Plana F (1997) Use of coal fly ash for ceramics: a case study for large Spanish power station. Fuel 76:787-791 Ramsden AR, Shibaoka M (1982) Characterization and analysis of individual fly-ash particles fom coal-fired power stations by a combination of optical microscopy, electron microscopy and quantitative electron microprobe analysis. Atmos Environ 16: 2191-2206 Regourd M (1986) Slags and slag cements. In: Cement Replacement Materials. Swamy RN (ed) Surrey University Press, London, p 73-99 Regourd M (2001) Cements made from blast furnace slag. In: Lea’s Chemistry of Cement and Concrete. 4th edition. Hewlett PC (ed) Butterworth-Heinemann, Oxford, p 637-678 Richardson IG (1999) The nature of C-S-H in hardened cements. Cem Concr Res 29:1131-1147 Richardson IG (2004) Tobermorite/jennite and tobermorite/calcium hydroxide-based models for the structure of C-S-H: applicability to hardened pastes of tricalcium silicate, b-dicalcium silicate, Portland cement, and blends of Portland cement with blast furnace slag, metakaolin, or silica fume. Cem Concr Res 34:17331777

276

Snellings, Mertens, Elsen

Richardson IG (2008) The calcium silicate hydrates. Cem Concr Res 38:137-158 Richardson IG, Brough AR, Brydson R, Groves GW, Dobson CM (1993) Substituted calcium silicate hydrate (C-S-H) gels as determined by 29Si and 27Al NMR and EELS. J Am Ceram Soc 76:2285-2288 Richardson IG, Brough AR, Groves GW, Dobson CM (1994) The characterization of hardened alkali-activated blast furnace slag pastes and the nature of the calcium silicate hydrate (C-S-H) phase. Cem Concr Res 24:813-829. Richardson IG, Groves GW (1992) Models for the composition and structure of calcium silicate hydrate (C-S-H) gel in hardened tricalcium silicate pastes. Cem Concr Res 22:1001-1010 Richardson IG, Groves GW (1993) The incorporation of minor and trace elements into calcium silicate hydrate (C-S-H) gel in hardened cement pastes. Cem Concr Res 23:131-138 Richardson IG, Groves GW (1997) The structure of the calcium silicate hydrate phases present in hardened pastes of white Portland cement/blast furnace slag blends. J Mater Sci 32:4793-4802 Rocha J, Klinowski J (1990) 29Si and 27Al magic-angle-spinning NMR studies of the thermal transformation of kaolinite. Phys Chem Miner 17:179-186 Rodriguez-Camacho RE, Uribe-Afif R (2002) Importance of using the natural pozzolans on concrete durability. Cem Concr Res 32:1851-1858 Rojas MF (2006) Study of hydrated phases present in a MK-lime system cured at 60 °C and 60 months of reaction. Cem Concr Res 36:827-831 Rojas MF, Cabrera J (2002) The effect of temperature on the hydration rate and stability of the hydration phases of metakaolin-lime-water systems. Cem Concr Res 32:133-138 Roy A (2009) Sulfur speciation in granulated blast furnace slag: an X-ray absorption spectroscopic investigation. Cem Concr Res 39:659-663 Roy DM, Arjunan P, Silsbee MR (2001) Effect of silica fume, metakaolin, and low-calcium fly ash on chemical resistance of concrete. Cem Concr Res 31:1809-1813 Roy WR, Thiery RG, Schuller R, Suloway J (1981) Coal fly ash: a review of the literature and proposed classification system with emphasis on environental impacts. Illinois Geol Surv Environ Geol Notes EGN 96:1-69 Sabir BB, Wild S, Bai J (2001) Metakaolin and calcined clays as pozzolans for concrete: a review. Cem Concr Comp 23:441-454 Sakai E, Miyahara S, Ohsawa S, Lee S-H, Daimon M (2005) Hydration of fly ash cement. Cem Concr Res 35:1135-1140 Sastre G, Corma A (2006) Rings and strain in pure silica zeolites. J Phys Chem B 110:17949-17959 Satarin VI (1974) Slag portland cement. Proeedings of the 6th International Congress on the Chemistry of Cement, principal paper 1-51 Scheetz BE, Earle R (1998) Utilization of fly ash. Curr Opin Solid State Mater Sci 3:510-520 Schott J, Petit JC (1987) New evidence for the mechanisms of dissolution of silicate minerals. In: Aquatic Surface Chemistry: Chemical Processes at the Particle-Water Interface. Stumm W (ed) John Wiley and Sons, New York, p 293-315 Scott PW, Critchley SR, Wilkinson FCF (1986) The chemistry and mineralogy of some granulated and pelletized blast furnace slags. Mineral Mag 50:141-147 Scrivener KL, Kirkpatrick RJ (2008) Innovation in use and research in cementitious material. Cem Concr Res 38:128-136 Sear LKA (2001) Properties and Use of Coal Fly Ash: A Valuable Industrial By-Product. Thomas Telford, London Sear LKA, Weatherley AJ, Dawson A (2003) The environmental impacts of using fly ash - the UK producers’ perspective. International Ash Utilization Symposium, Center for applied Energy Research , University of Kentucky. http://www.ukqaa.org.uk/Papers/KentuckyAshSymposiumLKASearEtAl2003.pdf Serry MA, Taha AS, El-Hemaly SAS, El-Didamony H (1984) Metakaolin-lime hydration products. Thermochim Acta 79:103-110 Sersale R (1980) Structure et charactérisation des pouzzolanes et des cendres volantes. Proceedings of the 7th International Congress on the Chemistry of Cement IV-1:3-21 Sersale R (1993) Advances in Portland and blended cements. Proceedings of the 9th International Congress on the Chemistry of Cement 261-302 Sersale R, Frigione G (1987) Portland-zeolite cement for minimizing alkali-aggregate expansion. Cem Concr Res 17:404-410 Shannag MJ (2000) High strength concrete containing natural pozzolan and silica fume. Cem Concr Comp 22:399-406 Sharma RL, Pandey SP (1999) Influence of mineral additives on the hydration characteristics of ordinary Portland cement. Cem Concr Res 29:1525-1529 Shi C (2001) An overview on the activation of reactivity of natural pozzolans. Can J Civil Eng 28:778-786

Supplementary Cementitious Materials

277

Shi C, Day RL (1995) Microstructure and reactivity of natural pozzolans, fly ash and blast furnace slag, Proceedings of the 17th International Conference on Cement Microscopy 150-161 Shi C, Day RL (2000) Pozzolanic reaction in the presence of chemical activators. Part I. Reaction kinetics. Cem Concr Res 30:51-58 Siddique R (2008) Waste Materials and By-products in Concrete. Springer, Berlin Sjöberg L (1989) Kinetics and non-stoichiometry of labradorite dissolution. In: Proceedings of the 6th International Symposium on Water-Rock Interaction. Miles DL (ed) p 639-642 Smadi MM, Haddad RH (2003) The use of oil shale ash in Portland cement concrete. Cem Concr Comp 25:4350 Smolczyk HG (1980) Slag structure and identification of slag. Proceedings of the 7th International Congress on the Chemistry of Cement I:III1-17 Snellings R, Mertens G, Cizer Ö, Elsen J (2010a) Early age hydration and pozzolanic reaction in natural zeolite blended cements: Reaction kinetics and products by in situ synchrotron X-ray powder diffraction. Cem Concr Res 40:1704-1713 Snellings R, Mertens G, Elsen J (2010b) Calorimetric evolution of the pozzolanic reaction of zeolites. J Therm Anal Calorim 101:97-105 Snellings R, Mertens G, Hertsens S, Elsen J (2009) The zeolite-lime pozzolanic reaction: Reaction kinetics and products by in situ synchrotron X-ray powder diffraction. Microporous Mesoporous Mater 126:40-49 Spence RJS, Cook DJ (1983) Building Materials in Developing Countries. Wiley and Sons, London Stamatakis MG, Fragoulis D, Csirik G, Bedelean I, Pedersen S (2003) The influence of biogenic micro-silicarich rocks on the properties of blended cements. Cem Concr Comp 25:177-184 Stamatakis MG, Fragoulis D, Papageorgiou A, Chaniotakis E (1998) Zeolitic tuffs from Greece and their commercial potential in the cement industry. World Cem 29:98-102 Steenari B-M, Lindqvist O (1999) Fly ash characteristics in co-combustion of wood with coal, oil or peat. Fuel 78:479-488 Stein HN, Stevels JM (1964) Influence of silica on the hydration of 3CaO.SiO2. J Appl Chem 14:338-346 Stillings LL, Brantley SL (1995) Feldspar dissolution at 25 °C and pH 3: Reaction stoichiometry and the effect of cations. Geochim Cosmochim Acta 59:1483-1496 Strandh H, Petterson LGM, Sjöberg L, Wahlgren U (1997) Quantum chemical studies of the effects on silicate mineral dissolution rates by adsorption of alkali metals. Geochim Cosmochim Acta 61:2577-2587 Swaddle TW, Rosenqvist J, Yu P, Bylaska E, Phillips BL, Casey WH (2005) Kinetic evidence for fivecoordination in AlOH(aq)2+ ion. Science 308:1450-1453 Swamy RN (1986) Cement Replacement Materials. Surrey University Press, London Takemoto K, Uchikawa H (1980) Hydratation des ciments pouzzolaniques. Proceedings of the 7th International Congress on the Chemistry of Cement IV-2:1-29 Tangchirapat W, Buranasing R, Jaturapitakkul C, Chindaprasirt P (2008) Influence of rice husk-bark ash on mechanical properties of concrete containing high amount of recycled aggregates. Constr Build Mater 22:1812-1819 Tangchirapat W, Jaturapitakkul C, Chindaprasirt P (2009) Use of palm oil fuel ash as a supplementary cementitious material for producing high-strength concrete. Constr Build Mater 23:2641-2646 Taylor HFW (1990) The Chemistry of Cement. Academic Press, London Taylor R, Richardson IG, Brydson RMD (2010) Composition and microstructure of 20-year-old ordinary Portland cement-ground granulated blast furnace slag blends containing 0 to 100% slag. Cem Concr Res 40:971-983 Tazaki K, Fyfe WS, Sahu KC, Powell M (1989) Observations on the nature of fly ash particels. Fuel 68:727-734 Thomas JJ, Rothstein D, Jennings HM, Christensen BJ (2003) Effect of hydration temperature on the solubility behavior of Ca-, S-, Al-, and Si-bearing solid phases in Portland cement pastes. Cem Concr Res 33:20372047 Tsipursky SI, Drits VA (1984) The distribution of octahedral cations in the 2:1 layers of dioctahedral smectites by oblique electron diffraction. Clay Miner 19:177-193 Türker P, Yeginobali A (2003) Comparison of hydration products of different pozzolanic systems. Proceedings of the 25th International Conference on Cement Microscopy 1-9 Türkmenoğlu AG, Tankut A (2002) Use of tuffs from central Turkey as admixture in pozzolanic cements Assessment of their petrographical properties. Cem Concr Res 32:629-637 Uchikawa H, Uchida S (1980) Influence of pozzolana on the hydration of C3A. Proceedings of the 7th International Congress on the Chemistry of Cement IV:24-29 Uzal B, Turanli L, Yücel H, Göncüoğlu MC, Culfaz A (2010) Pozzolanic activity of clinoptilolite; A comparative study with silica fume, fly ash and a non-zeolitic natural pozzolan. Cem Concr Res 40:398-404 Vassilev SV, Mendez R, Borrego AG, Diaz-Somoano M, Martinez-Tarazona MR (2004) Phase-mineral and chemical composition of coal fly ashes as a basis for their multicomponent utilization. 3. Characterization of magnetic and char concentrates. Fuel 83:1563-1583

278

Snellings, Mertens, Elsen

Vassilev SV, Menendez R, Alvarez D, Diaz-Somoano M, Martinez-Tarazona MR (2003): Phase-mineral and chemical composition of coal fly ashes as a basis for their multicomponent utilization. 1. Characterization of feed coals and fly ashes. Fuel 82:1793-1811 Vassilev SV, Menendez R, Diaz-Somoano M, Martinez-Tarazona MR (2004): Phase-mineral and chemical composition of coal fly ashes as a basis for their multicomponent utilization. 2. Characterization of ceramic cenosphere and salt concentrates. Fuel 83:585-603 Vassilev SV, Vassileva CH (1996) Mineralogy of combustion wastes from coal-fired power stations. Fuel Process Technol 47:261-280 Vempati RK, Rao A, Hess TR, Cocke DL, Lauer Jr. HV (1994) Fractionation and characterization of Texas lignite class ‘F’ fly ash by XRD, TGA, FTIR and SEM. Cem Concr Res 24:1153-1164 Villar-Cociña E, Valencia-Morales E, Gonzalez-Rodriguez R, Hernandez-Ruiz J (2003) Kinetics of the pozzolanic reaction between lime and sugar cane straw ash by electrical conductivity measurement: A kinetic-diffusive model. Cem Concr Res 33:517-524 Villar-Cociña E, Valencia-Morales E, Gonzalez-Rodriguez R, Hernandez-Ruiz J (2003) Kinetics of the pozzolanic reaction between lime and sugar cane straw ash by electrical conductivity measurement: A kinetic-diffusive model. Cem Concr Res 33:517-524 Vizcayno C, De Gutierrez RM, Castello R, Rodriguez E, Guerrero CE (2009) Pozzolan obtained by mechanochemical and thermal treatments of kaolin. Appl Clay Sci 49:405-413 Walther JV (1996) Relation between rates of aluminosilicate mineral dissolution, pH, temperature, and surface charge. Am J Sci 296:693-728 Ward CR, French D (2006) Determination of glass content and estimation of glass composition in fly ash using quantitative X-ray diffractometry. Fuel 85:2268-2277 Watt JD, Thorne DJ (1965) Composition and pozzolanic properties of pulverised-fuel ashes, Part 1-2. J Appl Chem 15:585-604 Watt JD, Thorne DJ (1966) Composition and pozzolanic properties of pulverised-fuel ashes, Part 3. J Appl Chem 16:33-39 Westrich HR, Cygan RT, Casey WH, Zemitis C, Arnold GW (1993) The dissolution kinetics of mixed-cation orthosilicate minerals. Am J Sci 293:869-893 White AF, Brantley SL (2003) The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and the field? Chem Geol 202:479-506 Wild S, Gailius A, Hansen H, Pederson L, Szwabowski J (1997) Pozzolanic properties of a variety of European clay bricks: Comparative study of pozzolanic, chemical and physical properties of clay bricks in four European countries for utilization of pulverized waste clay brick in production of mortar and concrete. Build Res Inf 25:170-175 Winburn RS, Grier DG, McCarthy GJ, Peterson RB (2000) Rietveld quantitative X-ray diffraction analysis of NIST fly ash standard reference materials. Powder Diffraction 15:163-172 Winnefeld F, Lothenbach B (2010) Hydration of calcium sulfoaluminate cements - Experimental findings and thermodynamic modelling. Cem Concr Res 40:1239-1247 Wong RCK, Gillott JE, Law S, Thomas MJ, Poon CS (2004) Calcined oil sands fine tailings as a supplementary cementing material for concrete. Cem Concr Res 34:1235-1242 Wu ZQ, Young JF (1984) The hydration of tricalcium silicate in the presence of colloidal silica. J Mater Sci 19:3477-3486 Xiao Y, Lasaga AC (1994) Ab initio quantum mechanical studies of the kinetics and mechanisms of silicate dissolution: H+(H3O+) catalysis. Geochim Cosmochim Acta 58:5379-5400 Xiao Y, Lasaga AC (1996) Ab initio quantum mechanical studies of the kinetics and mechanisms of quartz dissolution: OH- catalysis. Geochim Cosmochim Acta 60:2283-2295 Yilmaz B, Uçar A, Öteyaka B, Uz V (2007) Properties of zeolitic tuff (clinoptilolite) blended Portland cement. Build Environ 42:3808-3815 Yu LH, Ou H, Lee LL (2003) Investigation on pozzolanic effect of perlite in concrete. Cement Concr Res 33:73-76 Zain MFM, Islam MN, Mahmud F, Jamil M (2010) Production of rice husk ash for use in concrete as a supplementary cementitious material. Constr Build Mater, doi:10.1016/j.conbuildmat.2010.07.003 Zendri E, Lucchini V, Biscontin G, Morabito ZM (2004) Interaction between clay and lime in “cocciopesto” mortars: a study by 29Si MAS spectroscopy. Appl Clay Sci 25:1-7 Zwijnenburg MA, Corá F, Bell RG (2007) On the performance of DFT and interatomic potentials in predicting the energetic of (three-membered ring-containing) siliceous materials. J Phys Chem B 111:6156-6160

7

Reviews in Mineralogy & Geochemistry Vol. 74 pp. 279-364, 2012 Copyright © Mineralogical Society of America

Deleterious Reactions of Aggregate With Alkalis in Concrete Maarten A.T.M. Broekmans Department of Industrial Minerals and Metals Geological Survey of Norway (NGU) PO Box 6315 Sluppen, N-7491 TRONDHEIM, Norway e-mail: [email protected]

INTRODUCTION Concrete in the built environment The word “concrete” is derived from the Latin concretus (compact, condensed), representing a conjunction of con (together) and the past participle of cresco (to grow; compare: crescendo). Thus, concrete could be liberally translated as ‘grown solid together,’ alluding to the consolidation of a particulate aggregate material with a cement binder of some sort. Concrete containing aggregate has been used in construction by the ancient Greek and Romans, possibly as a further development of clay initially used by the Assyrians and Babylonians as a binder, later superseded by burnt lime and gypsum by the Egyptians. As a construction material, concrete allows architects and engineers to design a structure with only minimal constraints to its form. A complicated shape requiring great effort to chisel out from a piece of stone can simply be poured in a mold and reproduced as often as desired, also ex situ. Both the invention of Ordinary Portland Cement (OPC), first patented by British bricklayer Joseph Aspdin in 1824, and of reinforced concrete first patented by Parisian gardener Joseph Monier in 1867 (for making durable flower pots), contributed to the development of mechanically stronger concrete allowing yet slimmer and taller structures to be built. Today, concrete is the most popular building and construction material with an annual production volume exceeding 7.5 km3, or about 20 billion tonne. Concrete is the prime ingredient in the world’s largest and most prominent structures and landmark edifices, including hydropower dams (e.g., Three Gorges Dam, Yichang/CN; also see Charlwood and Solymar 1995), coastal defense works (e.g., Delta Works/NL), telecommunication (e.g., CN Tower, Toronto/CA), office skyscrapers (e.g., Burj Khalifa, Dubai/UAE), theatres (e.g., Opera House, Sydney/AU), hotels and casinos (e.g., Marina Bay Area, Singapore/SG), nuclear power plants, oil and gas drilling and productions rigs (e.g., Sakhalin/RU), sea ports and harbors (e.g., Rotterdam/NL), airports and runways, as well as in numerous but perhaps less glamorous motorway and railroad structures, industrial floors, utility structures and facilities, residential dwellings, even bath and shower tubs, kitchen sinks and—historia repentur—flower pots, garden ornaments and home decorations. Elsen et al. (2010) present a more detailed historical overview; Scrivener and Kirkpatrick (2008) present inspiration for future research. Without concrete, modern society would not be able to function as we have gotten used to. A broken concrete flower pot is a mere trifle, a leaking concrete tub may represent considerable discomfort, but untimely failure of virtually any concrete structure beyond household scale inevitably leads to disaster and loss of life. For reasons of civic safety and security as well as cost, society simply cannot accept its citizens to live with e.g., reduced functionality of infrastructural works, interrupted or discontinued power production, potential failure of coastal 1529-6466/12/0074-0007$10.00

DOI: 10.2138/rmg.2012.74.7

Broekmans

280

defense works. The tax payer should be burdened least possible with increased expenses for maintenance, rehabilitation, remediation and replacement beyond normal schedule of concrete structures, if such can be avoided by correct design engineering and selection of suitable materials. Modern society literally builds on the reliability of concrete as a durable and energy efficient material. Modern concrete engineering is able to design structures fulfilling society’s anticipations regarding functionality, projected service-life and longevity. Modern concrete materials science is essential to specify material properties and qualities required to comply with the above expectations and prospects, and to warrant sustainable development of scarce natural resources.

What is deleterious AAR? The two main benefits of mixing in aggregate with cement to make concrete are: 1. saving cost: by far the most expensive constituent of bulk concrete is the OPC binder, which can be diluted by adding 80-85 vol% total aggregate coarse plus fine, and 2. increasing strength: correct grading of coarse and fine aggregate with the even finer grained cement essentially produces a packstone, in which the interstices between coarser particles are filled with ever finer fractions. Thus, mechanical forces are dissipated from a handful of coarse particles to a great many fine particles with a positive effect on compressive and tensile strength, as well as fracture behavior. Further reduction of cement content can be achieved by replacing OPC with pozzolanic materials, occasionally even up to 95 wt% (Snellings et al. 2012, this volume; Justnes 2012, this volume). Consequently, modern concrete relies as much on the proper functioning of the cement binder as it does on the soundness of the aggregate it contains. Ideally, concrete aggregate is: 1. of high cubicity (i.e., equi-dimensionality) favoring rheological properties of the wet mix during placement, and it is 2. isotropic, lacking preferred orientation and/or a pronounced cleavage to minimize splitting under mechanical load, and it is 3. ever-lasting chemically inert for exposure to a Portland cement environment in concrete. While the first two requirements can be met relatively easily by selecting suitable resources (e.g., moraine gravel, dolerite) combined with adequate processing (e.g., screening, comminution), the third quality represented by chemical inertia is far less conspicuous. The interior of a Portland cement concrete structure presents a highly alkaline environment with ~pH 13 after pouring and curing, gradually increasing over time to ~pH 14. Reinforcement steel embedded in concrete is prevented from corroding by an impermeable passive layer between the iron metal and the hydrous paste environment, stable from ~pH 10.5 and over—provided absence of chloride (Leek 1991). Whereas the high alkalinity in the concrete interior protects the reinforcement, it is detrimental to certain aggregate constituents at the same time. What happens is that ‘susceptible minerals’ react under moisture saturated conditions with available Na and K to form a hygroscopic and hydraulic alkali-silica gel. Upon further reaction with water, the gel expands and cracks up the surrounding concrete, ultimately deteriorating the structure as a whole. The damage attributed by one single alkali-reactive aggregate particle is negligible and does not affect projected lifetime of the structure. However, if a sufficient proportion of the aggregate behaves alkali-reactive, structural and material integrity will decay and

Deleterious Reactions of Aggregate With Alkalis in Concrete

281

consequently, user safety and security will be compromised, eventually to an unacceptably low level if no proper precautions are taken.

Why is AAR important? After first local recognition in a concrete structure, deleterious AAR seems to change status a couple of times before AAR is officially acknowledged as a damage mechanism with countrywide importance. In a typical scenario, the first damaged structure is initially regarded as ‘the exception confirming the rule’ that there is no deleterious AAR elsewhere. The second stage can perhaps be described as ‘living in denial’: more and more structures are identified with deleterious AAR, but as the only known case has officially been declared an exception, obviously: ‘this has to be something else’. Popular sophistries to explain the extensive cracking include damage from frost or freeze-thaw cycling, surface exsiccation due to inadequate curing, poor workmanship, or simply putting the cracking off as ‘irrelevant’. Consequently, preventive guidelines for new structures and/or remedial strategies for existing ones are not considered necessary. As the number of damage observations increases with time, however, evidence that ‘something AAR-like’ may be going on gradually stacks up. This may take considerable time if the evidence is spread over separate official bodies and instances, impeding full overview. Ultimately, when deleterious AAR is finally acknowledged as a relevant and serious damage mechanism and receives ‘official problem’ status for the whole country, then suddenly everybody wants to hop on the bandwagon, get their structures appraised and many are indeed diagnosed with deleterious AAR, further expanding the problem. Whereas the presence of deleterious AAR in the concrete plate bridge to Farmer Giles’ land hardly affects society as a whole, its presence in large public works certainly does. For an important part, society lends its security, stability and daily-life predictability from e.g., reliable coastal defense works, functioning infrastructure, safe energy sources, and dependable utilities. Periodic maintenance varying from local repair to full overhaul, and ultimately complete replacement after ended service-life are virtually inevitable for concrete structures, but are rather predictable and can be scheduled at regular intervals. That includes ‘common damage’ like e.g., corroding reinforcement (carbonation of the cement paste, infiltration of chloride), biogenic sulfuric acid/sulfate attack, and other surface-controlled deterioration mechanisms. In contrast, deleterious AAR is an intrinsic bulk concrete property, as a result of the combination of that particular aggregate with that particular cement. Therefore, deleterious AAR is not just limited to the exposed concrete surface, but occurs through and through, rendering repair of AAR damaged structures extremely complicated and not rarely highly impracticable. Consequently, the only genuinely long-term durable solution is complete replacement of the whole structure, which is time consuming, an inconvenience for structure’s owners and users, and expensive for society. Before the new structure is fully operational, user safety and security has to be sustained for the existing structure, further adding cost. The combination of factors mentioned above renders deleterious AAR a genuine problem, not the mere ‘fringe issue’ it is sometimes said to be. Including planning ahead, building and construction of mega projects like e.g., a hydroelectric power dam, a high speed railroad connecting major cities, a storm surge barrier, easily takes 2-3 decades and costs multiple billions. Therefore, and for additional safety, such large and important structures are designed to last for 200-300 y, i.e., substantially longer than the 50-80 y service life common in road construction. Society (i.e., as a collective of faithful tax payers) can hardly afford the time and money for untimely replacement of such mega projects, and is not served by the additional insecurity of a potentially malfunctioning or failing structure.

Broekmans

282

HISTORY AND BACKGROUND OF ASR First recognition Any attempt to expand the annotated bibliography on “Alkali aggregate reactions in concrete […] 1939-1991” of Diamond (1992) is feeble, though an update to include papers on AAR published post-1991 could be interesting to have. According to its abstract on page ix, the bibliography lists nearly 1300 citations, including from literature in Japanese, French, Chinese and German language. The paper by Stanton (1940b) is generally considered to be the first to unambiguously identify alkali-aggregate reactions in concrete as such. This paper was preceded by eight earlier papers, six of American origin (including a prelude by Stanton 1940a), and two of German origin. Numbers of publications per year as collated in Diamond (1992) are plotted in Figure 1 below. Stanton (1940) is indicated by a solid lozenge. Data on journal publications on AAR are not accessible after 1991, thus, the data have been supplemented with numbers of contributions published in the Proceedings of the International Conference on Alkali-Aggregate Reactions in Concrete (ICAAR), marked with solid triangles (Poole 1992; Shayan 1996; Bérubé et al. 2000; Tang and Deng 2004; Broekmans and Wigum 2008). Peaks appear to coincide with abundant national-institutional production (1948, 1950: America, 1958: Denmark, 1973: Germany, Canada), and/or with the International Conference on Alkali-Aggregate Reactions in Concrete (acronymed ICAAR in English, CIRAG in French) held in 1974, 1975, 1976, 1978, 1981, 1983, 1986, 1992, 1996, 2000, 2004, 2008, and its latest episode in May 2012. The valleys defining between-ICAAR years might imply authors were saving for the next, at least until 1992. In total, Figure 1 comprises 1997 publications on AAR. If journal papers from 1992 until present are included, and publications in languages other than Japanese, French, Chinese or German (e.g., Spanish, Portuguese), as well as theses published at a number of universities around the world, the total must be much higher.

Global and local acceptance of AAR Since its first recognition in 1940, deleterious AAR has been identified in structures on all continents. The map in Figure 2 shows countries in which deleterious AAR has been identified as a relevant deterioration mechanism as per the end of 2010, based upon Diamond (1992) and personal communication with members of RILEM TC-219 ACS (Alkali-silica reaction in Concrete Structures) committee. The map should be read with care and interpreted with moderation. Countries and/or confederations with a large area, notably Canada, Unites States of America, Russia, Australia, and Brazil, all appear as ‘AAR territory,’ whereas deleterious AAR does not pose a problem of the same extent in all parts of each nation, but rather is limited to certain geological regions where AAR susceptible rock types are abundant (see for instance Shrimer et al. 2008). Meanwhile, the map in Figure 2 is discrete in that it does not disclose individual AAR damaged structures, which would be rather easy to identify especially in little populated areas with only few concrete structures around. The map could be refined considerably in much greater detail if official data on AAR would be available per country, rather than collating scarce data from the published literature and/or experts active in the field. Yet further refinement would be possible by attributing different weights to structure types, i.e., AAR damage in a major hydropower dam should weigh heavier than the odd backroad viaduct (or in other words, ‘plotting tons of AAR concrete,’ instead of the number of AAR structures). Despite above limitations and reservations, the map in Figure 2 nevertheless demonstrates that deleterious AAR is indeed a worldwide occurring problem. According to the European Cement Association Cembureau in Brussels, the European Union plus Norway and Switzerland (but minus Cyprus, Greece, Luxemburg and Monte Negro) in 2010 spent nearly 645 billion

Deleterious Reactions of Aggregate With Alkalis in Concrete

283

Figure 1. Numbers of journal and conference publications on AAR from 1923-1991 from Diamond (1992). Data from 1992 and onward represent numbers of contributions in ICAAR Proceedings, for 2012 tentative (triangles).

Figure 2. World map showing countrie with known cases of deleterious AAR, imcomplete and still ‘expanding’.

Euro on ‘modernization and renovation’ of existing structures (Cembureau 2012). So, even a mere 1%-fringe of the total expenditure amounts to 6.5 billion EUR, annually recurring, and only in the EU. Thus, the global annual amount spent on repair and maintenance must inevitably be a multiple of that. Whereas nations tend to regard territorial boundaries, geology and mineral resources do not. In general, the compositions of clinker, Portland cement, and/or aggregate reflect to a large extent the local/regional availability of geological resources. For instance, Cretaceous chert and/or opaline limestone occurring in locally extracted aggregates has damaged numerous structures with deleterious AAR in a region comprising parts of northern France (Parisian Basin), western Germany (Rhine Graben), Belgium, southern The Netherlands, Luxemburg

Broekmans

284

(Brabant Massif and Meuse valley), and Denmark. Thus, guidance to prevent AAR in concrete structures on a national basis is required (and present) in all of these countries, even when other parts of the country would not be affected by AAR. Norway has gone one step further and has mapped AAR-susceptible rock types on its mainland. Through petrographic assessment of concrete from a large number of damaged structures as well as expansion testing of virgin material in the laboratory, aggregate deposits producing alkali-reactive rock types could be identified (Jensen 1993; Jensen and Haugen 1996; Normin 1996; Normin 1999; Lindgård and Wigum 2003). This work has contributed to a national database hosted at the Geological Survey of Norway (NGU), free accessible after registration, containing petrographic data of about 6000 large to small deposits, including deposits currently producing as well as those no longer in operation. In addition, a list has been composed describing Norwegian potentially alkali-reactive rock types, ambiguous rock types that only react occasionally, and innocuous rock types that never have been observed to react deleteriously (Wigum et al. 2004). This list is now part of the official Norwegian guideline for AAR prevention (Norwegian Concrete Association 2008). It is tempting to apply this list (or a similar one) to aggregate materials used in other countries, especially for lithologies carrying the same name. For instance, sandstone poses a potential danger for deleterious AAR in Norwegian concrete, as demonstrated by many damaged field structures. Consequently, its content in bulk virgin aggregate material is strictly limited to 20.0 vol% (twenty). However, Dutch concrete aggregate of domestic origin typically contains 85-95 vol% sandstone, of various types. Yet, sandstone is only rarely observed to react deleteriously in Dutch structures, in which most AAR damage is caused by >2.00 vol% (two) of porous chert, chalcedony and/or opal (Broekmans and Jansen 1998, Broekmans 2002). Several reasons for this observed discrepancy can be due to, for example, differences in aggregate composition and detailed mineral content despite identical lithological name, differences in cement composition, different building and construction standards. The map in Figure 2 is conservative, as more countries will be added when structures appearing healthy and sound today get through the incubation period (a characteristic feature of AAR) after when damage becomes manifest. In addition, more countries will discover deleterious AAR on their territory because of ever improving diagnostic methods, now correctly identifying as deleterious AAR what previously had been diagnosed as another type of deterioration. However, acceptance by local authorities that deleterious AAR is indeed a damage mechanism to reckon with in the design, maintenance, replacement and recycling of concrete structures is often burdened with political, monetary and other issues, not the in the last place loss of prestige.

Remediation and prevention In order to prevent deleterious AAR to incur damage to future structures, many countries (but not all those marked in Fig. 2) have instated a set of rules, guidelines, codes of conduct or the like. Though most of these share the same characteristics, they do differ in applicable values for accept/reject criteria, primarily reflecting local experiences with e.g., the amount of concrete expansion still considered acceptable, available aggregate materials, feasible and reliable assessment methods. Many methods have been attempted to rehabilitate existing structures damaged by AAR, or to prevent its progress. Examples from practice include but aren’t by any means limited to: • local application of repair mortar by hand to cover up the cracking (a cosmetic measure with negligible effect on service life); •

applying Li-solution to the concrete surface to immobilize incipient alkali-silica gel (but Li hardly penetrates via the pore system due to its large hydrated size, unless the

Deleterious Reactions of Aggregate With Alkalis in Concrete

285

concrete is cracked—at which point it gets rather useless to prevent AAR); •

covering the structure’s top surface with a water tight impermeable membrane and the down-facing surface with a repellant coating open for vapor, to exsiccate the structure by ‘natural breathing’ (but the ASR reaction product silica gel is highly hygroscopic and doesn’t dry out under ambient conditions);

•

injecting the fracture pattern with epoxy or polyurethane resin (which turns into a never ending story as AAR damage progresses further);

•

installing additional reinforcement by gluing metal strips or carbon fiber mats onto the surface (which introduces additional shear against non-reinforced parts);

• driving hundreds of rock bolts through an already heavily cracked bridge deck (punching holes through interlocking pieces of concrete rubble held together by a cage of rebar); •

peeling the concrete cover to behind the rebar and restore to original profile with new concrete (to see that crack again by the still expanding original concrete core).

• slot cutting with a wire saw to release stress and accommodate further expansion (thereby accepting that it will continue to expand). A broad diversity of repair and rehabilitation methods have been tried on a range of AAR damaged field structures (see e.g., Bakker 1999; Jensen 2004; Thomas et al. 2008), with an inconsistent success rate. More often than not, renovation has only delayed complete demolition and replacement of the whole structure, rather than genuinely curing AAR and restoring the structure’s original status.

ORIGIN OF ALKALIS IN CONCRETE The Na2O-equivalent Already since early papers on AAR, alkalis Na and K have been presented in literature as “Na2O-equivalents” merging Na and K into one single parameter denoted as “Na2Oeq,” “eq-Na2O,” or similar. The Na2Oeq is calculated straight-forwardly as [Na2O + 0.658 K2O] expressed in weight percent. Alternatively, Na2Oeq values are commonly given in kg·m−3, especially when referring to bulk concrete or mortar, as e.g., in standards for AAR prevention. The factor 0.658 simply represents the ratio in molar weights between Na2O (61.98 g) and K2O (94.20 g). Clearly, use of the Na2Oeq assumes identical behavior in (pore) solution, in mineral constituents composing cement and aggregate, and their interaction. However, it has been long known that ‘Na = K’ comprises an oversimplification, arguably acceptable as a coarse first approximation, but not in science with today’s level of knowledge (see e.g., Leemann and Lothenbach 2008). Yet, the Na2Oeq is the de facto standard in virtually all recent papers on deleterious AAR, despite many papers demonstrating different behavior of Na versus K in quartz/silica dissolution, e.g., Iler (1955), Marshall and Warakomski (1980), Wijnen (1990), Heaney et al. (1994), Dove (1999), Dove et al. (2005), and many, many others. The fact that the behavior of Na is different from K in minerals resulting in markedly different properties and qualities of Na and K varieties hardly needs mention.

Alkali from raw materials for Portland clinkering In a typical process for the manufacture of ordinary Portland Cement (OPC) clinker, ground calcareous material (e.g., limestone, marl) is burnt together with argillaceous material (e.g., clay, shale) high in alumina and iron, possibly with the minor addition of iron ore (e.g.,

286

Broekmans

roosted pyrite), bauxite or quartz sand to arrive at a suitable bulk composition. Local variations to this typical production recipe may apply, depending on availability of mineral resources. Details on raw materials for OPC clinkering are given in for instance Taylor (1997) and Hewlett (1998). To reduce CO2-emission, OPC may be prepared from alternative materials including e.g., anorthosite rock rich in Ca, Si and Al as mainly composed of anorthite (also see Gartner 2004; Justnes 2012, this volume). Some or all the raw materials used in OPC clinkering may contain alkalis Na and K as impurities, consolidated in common rock-forming minerals like e.g., albite, K-feldspar, muscovite, biotite, clay minerals. Depending on origin and nature of the alkali minerals in the resource material, type of comminution and processing (wet, dry) prior to burning, and burning conditions, part of the Na/K present is removed during the process, whereas another part may behave refractory and be retained in the clinker product. The incorporation of alkalis in main clinker minerals alite (C3S), belite (C2S), and ferrite (C4AF) is very limited. Only main clinker constituent aluminate (C3A) is able to host up to 5.7 wt% of Na2O replacing CaO. Alkalis mainly build their own minerals in the cooling clinker, notably thenardite (Na2SO4), arcanite (K2SO4), aphthitalite ([K,Na]3Na[SO4]2), the calcium analogue of langbeinite (Ca2K2[SO4]2), and some rare alkali carbonates or potassium aluminate occurring in trace amounts only in certain clinkers (Taylor 1997, and references therein). The subordinate importance of alkalis in clinker production is illustrated to some extent by the fact that they are not included in the original Bogue model calculation to ‘determine’ modal composition of clinker from bulk chemical composition, a method analogous to a CIPW model calculation for igneous rocks (Bogue 1929; Barry and Glasser 2000; also see Meier et al. 2012, this volume; Stutzman 2012, this volume). The alkalis comprised in the common clinker constituents mentioned above are released into the solution upon addition of water to initiate setting and hardening of the Portland cement in a mortar or concrete mixture. (In contrast, the small amounts of alkalis contained in alite, belite, aluminate and/or ferrite are gradually released as hydration of clinker progresses over time until completion (see Pöllmann 2012, this volume). Part of the water will be used for hydration of clinker minerals, whereas the excess water ends up as ‘interstitial pore fluid’ suspending the hydrating cement particles, improving workability of the concrete mixture during placement. Modern concrete mix designs make extensive use of (super-) plasticizers to optimize workability, which property in traditional concrete made several decades ago or more was controlled through the water/cement ratio, and frequently extra water was added on the job if the mix was still considered too dry. After setting, excess water defines to a large extent the capillary porosity of the hardened concrete. The hydrate minerals are able to accommodate limited alkali in their structure (Richardson and Groves 1993), and some of the Na and K in the pore solution are indeed consolidated by the cement paste upon initial setting and hardening of the concrete mixture. Immobilization by paste silicate-hydrates is considered to render alkalis unavailable to induce deleterious AAR, and is therefore actively sought in AAR prevention through replacement of OPC with ground-granulated blast-furnace slag, fly ash or other supplementary cementitious materials (see Snellings et al. 2012, this volume). Alkalis are partially recycled by recrystallization of the cement hydrates in the paste upon aging, and consequently, the composition and pH of the pore water evolves over time. Whereas the above applies to the alkali household in common field concrete, the situation may be radically different in mortars and concretes specially prepared for expansion testing in a laboratory. To achieve a certain alkali concentration with minimal effort, the alkali content of the (mortar, concrete) mix is adjusted by pre-dissolving NaOH in the water often up to 1 N, presupposing the alkalis to eventually show up in the pore water as dissolved hydroxides anyway. Today, this is a widely accepted practice prescribed in the test procedures of many

Deleterious Reactions of Aggregate With Alkalis in Concrete

287

national guidelines for AAR prevention, e.g., ASTM C227-10 (2010), ASTM C1260-07 (2007), RILEM methods AAR-2, AAR-3 (2000), and the still forthcoming RILEM AAR-4 method (2007). Adding alkalis to the mix water is much more convenient than the alternative of stocking a collection of cements with varying alkali contents (but with concomitant additional variations in other respects than alkali content, potentially undesirable), and is the preferred method for reason of practicality. But the convenience comes at a price: the highly alkaline mix water with pH around 13-14 instantly dissolves Ca-sulfate (added to the Portland clinker to regulate aluminate hydration; see Pöllmann 2012, this volume). While Na+ and SO42− remain in solution, portlandite (Ca[OH]2) precipitates removing OH− and reducing alkalinity. The additional SO42− gets incorporated in Ca-silicate hydrates, from which it may be released again upon aging, not unlike another damage mechanism known as DEF (delayed ettringite formation). More elaborate descriptions of experiments illustrating above observations are given in Ong and Diamond (1993), and Diamond (1997). In summary, both short and long term behavior of alkalis in paste constituents and pore water as well as pH are changed when alkalis are pre-dissolved in the mix water. This does affect the development of expansive AAR under testing conditions, but exactly to what extent is difficult to estimate. Using one standardized cement with well-documented composition from a single manufacturer for expansion testing (as indeed recommended in RILEM procedures), the error may be systemized but cannot be circumnavigated. The alkali content of concrete and its evolution over time is a complicated net sum of short and long term effects, in laboratory mixes specifically prepared for testing, but also in regular field concrete.

Infiltrated alkali from seawater and deicers Concrete structures exposed to marine or brackish water, sodium salts applied as deicers, or less common in certain industrial applications involving (spillage of) alkali solutions, may see an increase in alkali content from the surface towards the interior. Penetration depth of solution and solutes primarily depends on the quality and condition of the exposed surface and exposure duration, but can be greatly enhanced by the presence of cracks and ‘interconnected porosity’ in general, often representing construction imperfections, by inadequate structural or material design, or by implementation errors on the work floor. On average, seawater contains ~1 wt% dissolved Na+ and ~0.04 wt% K+, or about 30× less. Other main dissolved constituents include chloride Cl− (~1.94 wt%), magnesium Mg2+ (~0.13 wt%), and sulfate SO42− (~0.27 wt%); many other dissolved species are present in (far) lesser amounts. In general, total salinity is higher in (sub-) tropical climates where evaporation rates are higher, influx of fresh water is limited, and/or recirculation is confined due to restricted communication with other basins, as for instance the Red Sea and the Mediterranean. Conversely, salinity is lower where mixing and dilution with fresh water runoff from rivers or glaciers is more abundant, notably in more moderate and polar climate regions, but also where monsoon rains reign in the tropics. Consequently, the total amount of dissolved solids approximately ranges from 35-45 g·L−1 (of which ~10 g Na+K), and seawater alkalinity ranges from pH 7.5-8.4, i.e., well below that of pore water in fully hardened concrete at ~pH 14. During winter, various salts are applied to road surfaces as deicers, to maintain safe driving conditions. Salt may be applied dry to the road surface, or mixed with hot sand or hot water to increase the thawing effect. By far the most popular type of deicer for sheer reason of cost is ordinary NaCl as sea, rock or kitchen salt, which is effective at temperatures down to −7 °C rendering it rather useless in large parts of Scandinavia and other northern territories on the globe. Chlorides of magnesium MgCl2 and calcium CaCl2 are effective to −15 °C and −31 °C,

288

Broekmans

respectively, but also release chloride into the environment, which is undesirable. Alkali-free isopropyl alcohol H3C-HCOH-CH3 and/or propylene glycol H2COH-HCOH-CH3 are often used on airport runways and taxiways as these do not contain chloride which is corrosive for the metallic parts in aircraft, as well as because they are effective down to −50 °C or even lower. Alternatives for this type of deicing application urea H2N-CO-NH2 and ethylene glycol H2COH-H2COH have been banned for their toxicity and detrimental effect on the environment. Since about two decades, solutions of potassium formate HCOOK or acetate CH3COOK (occasionally abbreviated as KAc) have gained popularity enabling effective deicing at temperatures from −30 °C down to −60 °C. The sodium equivalents are also used. The concentration of carboxylic acid solutions as applied on the surface to be deiced is typically around 50-55 wt%, or over 6 M for K-acetate, with pH around 11, generally considered safe and harmless for concrete, and too low to cause deleterious AAR (typically > pH 13.5). However, such deicers represent concentrated solutions of weak acids, with consequences for solution behavior and their interaction with concrete (see below). Alkali contents of the solutions as applied are high, with ~23 wt% K for potassium formate, a little lower (~20 wt%) for an acetate solution. Commercial solutions usually also contain a dye to help uniform application additives to prevent steel reinforcement corrosion, and others. Sodium from ordinary salt dissolved in sea water or applied as deicer is the dead-painted external source of alkali potentially contributing to deleterious AAR, and has been subject of many studies until recent (e.g., Chatterji et al. 1987; Nixon et al. 1987; Sibbick and Page 1987; Kawamura et al. 1996; Duchesne and Bérubé 1996; Shayan 1998; Bérubé and Dorion 2000; Katayama 2004b; Shayan et al. 2010), including both field structures and concrete/mortar prepared for laboratory testing. Infiltration of NaCl—irrespective its origin—is the primary cause of chloride-induced reinforcement corrosion, a mechanism that is essentially confined from the concrete surface to the outermost steel reinforcement mesh, i.e., the cover. Removal of the concrete cover by demolition until behind the reinforcement and restoring to original profile with fresh concrete (often sprayed) is dusty, noisy, laborious and hence costly, and therefore often undesirable or impractical. Alternatively, electro-chemical extraction of infiltrated chloride is possible by embedding a temporary external Ti-wire mesh in an alkaline waste-paper or wood saw-dust pulp to the structure surface, and applying an electrical current (DC). The electrochemical repair process produces additional alkalinity as OH− around the internal reinforcement by electrolysis of the concrete pore fluid, whereas charged species migrate in the electrical field created between the internal reinforcement and temporary external mesh, direction depending on charge sign. Thus, negatively charged chloride Cl− is driven towards the electrically positive outer mesh, as is the newly created OH− (re-) saturating the concrete cover. In contrast, positively charged species like the alkalis Na+ and K+ are drawn towards the electrically negative inner reinforcement. Electrochemical chloride-extraction is often less successful than anticipated (or hoped for) as chloride tends to get immobilized in the paste by alumino-ferrite present in the paste, through precipitation in compositions near ~hydrocalumite Ca2Al[OH]6[Cl1−xOHx]∙3H2O. These represent complex calcium-aluminate hydrate compounds closely related to “Friedel’s salt” (e.g., Fischer et al. 1980), but still rather poorly characterized in the cement paste (e.g., BirninYauri and Glasser 1998). Arya and Xu (1995) noticed different extent of chloride bonding by cement pastes of different composition, whereas Suryavanshi et al. (1995) demonstrate the complexity of chloride binding by the paste, which mechanism also involves sodium. Whereas infiltration of aggressive solutions is mostly limited to the exposed concrete surface, penetration to greater depths is strongly facilitated by an existing crack fabric. Broekmans (2002) studied complete top-bottom cross sections of the decks in two ASRdamaged viaducts scheduled for demolition and replacement. The fabric and spatial distribution of cracks in the decks’ cross sections were determined on fluorescence-impregnated plane

Deleterious Reactions of Aggregate With Alkalis in Concrete

289

sections cut lengthwise from Ø100 mm cores cf. Danish Standard 423.39 (2002). Results were presented as a “damage rating index” (DRI) with arbitrary units 0-5, ranging from ‘no visible cracks’ to ‘incoherent disintegrated concrete.’ Attributed DRI values were interpreted in terms of ‘system openness’ rather than ‘amount of ASR damage,’ as genuine-ASR cracks might have been enhanced by mechanical load during service. Whole-rock geochemistry on bulk concrete and total chloride by ISE after dissolution in excess nitric acid were determined on separate cores Ø170 mm drilled in closest possible conjunction with the cores for petrography. Each Ø170 mm core was divided in sixteen equally high sections, and each section was analyzed separately (Broekmans 2002; also see Broekmans 2006). Figure 3 above shows cores BB243/4/5 from one viaduct, cores BB246/7 from another. Values equal to or below LLD (0.005 wt%) are plotted as 1/10 that value and occur in all cores, implying a ‘chloride-free’ initial background. Both structures had been exposed to NaCl deicer during service life, but not to marine water. The coincidence between chloride content and DRI is striking, with the exception of the central portion in core BB246, in which the cracks in sections 6, 8-9, and 11 represent artifacts from previous strength testing in the laboratory (see Den Uijl et al. 2000), thus were never exposed to chloride in the field. Figure 3 shows that the interior of a 600-1000 mm thick concrete deck (tapering) is open to communicate with the outside world through cracks present. Apparently, these provide a convenient pathway for the infiltration of fluids carrying dissolved species like chloride, as well as for leaching of other species. The concentration of chloride (or of any other infiltrated

Figure 3. Distribution of infiltrated chloride (solid lines) vs. cracking DRI (dashed lines) across viaduct decks. Adapted from Broekmans (2002).

290

Broekmans

species, for that matter) in the crack itself and its direct vicinity will be (substantially) higher than the plotted values, as it is in fact averaged out over the entire core section by the wholerock analytical procedure. In situ assessment of chloride contents by EPMA in thin sections across the crack fabric will give more realistic concentrations. While above excursion to chloride illustrates the openness of cracked concrete for hydrous solutions, the situation may be different for the infiltration of Na and/or K along the same crack fabric. Diamond (1997) reports that drying and rewetting of concrete cores tends to reduce the alkali content of expressed pore fluid, and argues this is due to consolidation in the paste. Many if not most field concretes damaged by deleterious AAR have been exposed to drying and rewetting cycles during their lifetime, though conditions may range widely from arctic to tropical, with wetting from seasonal to permanently humid. Hong and Glasser (1999, 2002) did laboratory experiments with synthetically prepared amorphous CSH and CASH ‘gels,’ and exposed these to Na and/or K solutions at water:solid ratios much higher than in field concrete. Combining observations from these papers seems to suggest that alkalis are partly absorbed to the paste minerals’ surface, partly consolidated in the paste through exchange and more difficult to remobilize upon rewetting: Diamond (1997) found that four months of continuous moisture saturation only remobilized 10-20% of the fixed alkali. Thus, exsiccation could be effective to halt the (further) development of deleterious AAR. However, any cracks present would facilitate infiltration of fresh alkalis gradually over time increasing the total amount of alkali present, or replenishing the amount fixed with unfixed alkalis from the concrete interior. Without cracks, redistribution of alkalis would be predominantly diffusion controlled, several orders of magnitude slower than via cracks serving as convenient fluid pathways. An excellent investigation of field concrete damaged by infiltrated alkalis combining multiple analyses is Katayama et al. (2004), and includes a calculated estimate of the alkali-balance. Another type of AAR damage from alkali infiltration has gained increasing attention recently. Since about 1990, K-acetate solutions have been applied as non-corrosive deicers on (mainly military) airfields in the United States of America as well as on the German Autobahn. Damage patterns are typical for infiltration-controlled mechanisms, with extensive peripheral cracking, and much less farther away from exposed surfaces or edges (Giebson et al. 2010). Even when the alkalinity of solutions as applied is around pH 11 and hence ‘safe for concrete,’ practice proves different. Diamond et al. (2006) added portlandite Ca[OH]2 (an abundant paste constituent in common OPC concrete) to laboratory prepared KAc solutions, and measured their pH to jump to >15! Laboratory concretes produced with low-alkali cement and several types of aggregate did expand rapidly upon exposure to this kind of solutions cf. ASTM C1293-08b (2008), whereas the same mixes behave non-expansive when exposed to 1 N NaOH in the same procedure. In the same experiments, the pH of K-acetate deicer increased from initially ~pH 11 for virgin solutions until >pH 14 after exposure to concrete, coinciding with earlier observations. In concentrated solutions, however, activities of the dissolved species have to be taken into account, so that the actual increase of [OH−] in this situation is less than 105 = 100,000-fold than suggested by the increase in pH. Extensive research by Giebson et al. (2010) on synthetic and real pore solutions, model calculations and performance tests on concrete and mortar specimen, concludes that two mechanisms seem to enhance alkali reactivity of the aggregate: 1) the additional supply of alkalis, and 2) the dissolution of portlandite Ca[OH]2 into the pore solution increasing [OH−] and rising pH. In summary, infiltration of alkalis through solution ingress is able to induce deleterious AAR in concrete, especially in structures with a large exposure area and a small ‘body volume,’ for instance pavements, airfield runways, viaducts, and the like. Depending on exposure conditions, and the nature of the infiltrating alkali solutions, the degree of potential enhancement of deleterious AAR may vary from ‘negligible’ to ‘most effectively.’

Deleterious Reactions of Aggregate With Alkalis in Concrete

291

Alkali released from aggregate Release of alkalis Na and/or K is a matter of intensive dispute, in concrete structures suffering from deleterious ASR as well as in testing and assessment of aggregate materials for use in concrete. Alkali release from rock forming minerals under natural conditions has been studied extensively. The Reviews in Mineralogy volumes by Hochella and White (1990) and White and Brantley (1995) present excellent reviews and a wealth of literature references. Alkali release from common rock forming minerals exposed to ‘concrete conditions’ (either laboratory simulated, or in real life) has been investigated by Diamond et al. (1964), Van Aardt and Visser (1977a,b, 1978), Way and Cole (1982), Bérubé et al. (1988, 1990, 2002), Choquette et al. (1987, 1991), Kawamura et al. (1989), Goguel (1995), Bérubé and Duchesne (1996), Constantiner and Diamond (2003), Bérubé and Fournier (2004), Leemann and Holzer (2005), Lu et al. (2006), Wang et al. (2008), Locati et al. (2010), and others, in both sialic and mafic concrete aggregate lithologies. In practice, K-feldspar and sodic plagioclase appear to be the most susceptible for alkali release especially with increasing degree of alteration to kaolinite or saussurite, together with micas and other alkali-phyllosilicates due to their perfect cleavage and loosely bonded alkalis. In existing concrete, a number of aggregate lithologies have been indicated to release poorly bonded alkalis from certain minerals, in particular K-feldspar partly altered to sericite or kaolinite, but also muscovite, illite, and others. Supplementary cementitious materials (SCMs; also see Snellings et al. 2012, this volume) like e.g., blast furnace slag, fly ash, microsilica or others that are added to OPC to incorporate and immobilize alkali in the paste during their hydration and thus inhibit deleterious AAR, are feared ineffective against internally released alkalis, for at least two main reasons. First, there is timing: alkalis in the paste are released instantly with the addition of mix water, and most of it is again consolidated in the paste during setting and hardening, essentially a few months in traditional OPC concrete (though hydration continues as long as unhydrated clinker is available and has access to moisture). The internal pore volume in individual particles is generally low for most natural rock types used as concrete aggregate (typical range 1.0-0.1 vol%, but often lower). In addition, interstitial pores between mineral grains are generally small (large voids are undesirable as they reduce compressive strength) and with low permeability due to small and/or necked interconnections. Porosity (and permeability) of aggregate compares as ‘low,’ relative to bulk traditional concrete with 15-20 vol%. Thus, concrete aggregate stored and stockpiled outside is moisture saturated, so that infiltration of dissolves species through the existing pore system into the interior of individual particles (whether inhibiting or enhancing deleterious AAR) will be predominantly diffusion-controlled, and hence slow. Conversely, ‘exfiltration’ of soluble species through leaching via the aggregate’s pore system will be slow as well. Therefore, whereas paste alkalis are immobilized at an early age, alkalis released by aggregate become available only after some delay, the extent of which depends on existing porosity and permeability of the actual aggregate lithology, as well as particle size, grain size and fabric parameters together determining the effective length of the diffusion pathway. By the time the alkali concentration in the releasing particle’s interior has reached a critical level, the immobilizing action of the SCMs added may not be adequate to prevent deleterious AAR. Second, there is the issue of location: alkalis released in the interior of an aggregate particle cannot be immobilized by SCMs outside that particle, unless both reservoirs are able to communicate chemically (see Fig. 4). Again, this will be diffusion-dominated, and communication would only be effective if ‘alkalis-to-be-immobilized’ were indeed present, which only happens delayed.

292

Broekmans

Figure 4. Optical micrograph showing ASR gel coating an air void, in a slag cement concrete with characteristic blue-green paste. As paste alkalis had been immobilized by the slag, the gel was produced with alkalis released in the particle’s interior (top right corner, micaceous siltstone). While a curiosity in Dutch (slag) concrete, what would happen if a given aggregate contains a substantial proportion of such particles? The circle marked D refers to a detail shown in Figure 8.

Alkalis Na and/or K are present in the most common rock forming minerals, e.g., acidic plagioclase (albite-oligoclase, Na1−xCaxAlSi3−xAlxO8, with x = 0-0.3), K-feldspar (KAlSi3O8), muscovite (KAl2Si3AlO10[OH]2), illite (~K0.65Al2Si3.35Al0.65O10[OH]2), and other micas or phyllosilicates, alkali amphiboles, etc., all rather common and occasionally abundant in typical concrete aggregate lithologies, except limestone. The weak bonding of Na+ (and to a somewhat lesser extent of K+) is largely due to its monovalent nature, combined with a small radius of ~100 pm (K+ ~140 pm) in an oxide or silicate matrix. This renders Na (and K) prone to removal from the crystal structure relative to other species in the mineral structure, for instance by dissolution in aqueous solution, by evaporation upon exposure to an electron probe beam (see below), or otherwise. The structure of micas and a number of clay minerals can be described in a simplified way as consisting of a single layer of AlO4[OH]2-octahedra sharing edges (representing gibbsite), flanked on either side by layers of SiO4-tetrahedra arranged in six-rings, with all tetrahedraapices sharing O with the gibbsite layer. This Si-Al-Si-oxyhydroxide sandwich layer is often abbreviated as “TOT layer,” for tetrahedral-octahedral-tetrahedral. In muscovite, TOT layer stacks are kept together by K+ ions positioned above the centers of two opposite silica sixrings, defining a TOT-TOT interlayer. Muscovite easily parts into flexible and elastic thin flakes, defining the perfect basal cleavage characteristic for micas. Instead of K+, the interlayer may contain other ionic species to create other minerals, e.g., by Na+ (paragonite), Ca2+ (margarite, a brittle mica), whereas partial filling with K+, vacancies and/or hydronium H3O+ gives illite. Part of the Si4+ in the sandwich layer can be replaced by additional Al3+ to maintain charge balance. Instead of Al3+, the octahedral layer in the sandwich may also be constructed of Mg2+ and/or Fe2+ in variable ratios (together with a range of other species usually in minor quantities), thus making common micas like e.g., phlogopite (K-Mg with subordinate Fe), biotite (K-Fe

Deleterious Reactions of Aggregate With Alkalis in Concrete

293

with subordinate Mg), celadonite, glauconite, and many more (also see Hazen and Wones 1972, 1973). The presence of less common chemical species replacing Al3+ in the tetrahedraloctahedral-tetrahedral sandwich combined with substantial consequently makes less common mica minerals that are rare to exotic in concrete aggregate. The efficacy of alkali release into solution is in principle related straight-forwardly to available surface area, or inversely to mineral grain size. Thus, a reduction of mineral grain size due to tectonic milling (here under subgraining) or otherwise facilitates alkali release, as does the presence of cleavage trails (notably in feldspars and micas) or microcracks by providing access for the pore solution to the grain’s interior. In addition, alteration through retrograde metamorphism or weathering may render the alkali-bearing mineral more susceptible to alkali release. Known examples include e.g., sericitized or kaolinitized K-feldspar, saussuritized plagioclase, muscovite altered to illite. In effect, both latter features (i.e., fissuring, alteration) represent net grain size reduction as well. Alkali release from other minerals than mica and phyllosilicates is similar along broad lines. Planes of weakness (as e.g., cleavage planes at atomic scale, grain/twin/exsolution boundaries at crystal lattice scale, interstitial pore space at particle interior scale, total accessible surface area at particle exterior scale) facilitate fluid communication in and out. Common rock forming minerals containing alkalis and with pronounced cleavage include micas (e.g., muscovite, biotite, phlogopite, celadonite, glauconite, phengite) and intermediates towards clays, and to a lesser extent also alkali-amphiboles. Though (acidic) plagioclase and K-feldspar have three independent cleavages, their higher mechanical tenacity would render them less susceptible to leaching. On the other hand, altered plagioclase and K-feldspar may be more prone to alkali release due to the fine-grained nature of the alteration products (saussurite and sericite/kaolinite, respectively), often poorly crystalline. Grattan-Bellew and Beaudoin (1980) describe the release of K+ from natural phlogopite branded as “Suzorite” (composition reported as K2Mg4.32Fel.16Al0.35[Si5.75Al2.25O20][OH]4) added to OPC matrices with different initial Na2O-equivalents. In all combinations investigated, concrete with phlogopite expanded more than the equivalent mixes without, though some remained under the expansion limit. Phlogopite exposed to lime water (saturated Ca[OH]2 solution) with or without various amounts of added NaOH (representing simulated concrete pore fluid) was found to release K+ into solution, whereas the exposed phlogopite was demonstrated by SEM-EDS to have a lower Mg/K ratio, however exchange of Mg for Ca could not be confirmed. Also see Figure 5, showing alkali release from alkali-reactive sand-/siltstone in Dutch concrete from the mid-1960’s. It is interesting to take a look at the observations of Boles and Johnson (1983). They added 1 g of chopped biotite and/or muscovite to solutions initially alkalized with 0.1 M NaOH to pH 11.5, then acidified by titration with 0.1 M HCl to pH 4. They found that biotite acts as a base raising pH by absorbing H+, throughout the entire pH range investigated. In contrast, below ~pH 5.5 muscovite acts as an acid shedding H+. The net capacity to increase pH under alkaline conditions is attributed to differences in mineral-structural chemistry, muscovite being di-trigonal, biotite tri-trigonal, interlayer species present (K+, Na+, hydronium) and the resulting surface charge and charge density. Tentative calculations suggest that pH increase may be up to several orders at close range to the mineral surface. Above experiments seem confirmed by observations on biotite in certain oil-bearing sandstones, in which calcite or prehnite exclusively occur intercalated within biotite cleavage, which is attributed to local stabilizing through H-capture by the biotite host. Similar observations have been reported by Claeys and Mount (1991, and references therein), Nijland et al. (1994) for hydrogarnet lenses in biotite, and Nijland (pers. comm. 1997) who reported portlandite precipitated within and around biotite flakes in OPC concrete.

294

Broekmans

Figure 5. Element maps by EPMA showing distribution of K (left) and Na (right) in an alkali- reactive sandstone, extruding alkali-silica gel into the surrounding paste through a crack. Detrital grains of elongate muscovite and more rounded K-feldspar and albite are easily recognized in the particle’s interior.

Along the same line of thought, Bjørkum (1996) identified detrital muscovite flakes penetrating into quartz grains, without bending or buckling. A simple calculation using elasticity properties of muscovite showed that the flakes are mechanically too weak to penetrate. Furthermore, SEM-cathodoluminescence unequivocally demonstrated the detrital origin of the penetrated quartz grains, as opposed to precipitated quartz of neogenic origin. In summary, the observed penetration is interpreted as a result of quartz dissolution enhanced by muscovite through an increase of the local pore fluid pH. This mechanism has been successfully applied to describe the compaction of North Sea sandstones, and is supported by petrographic observations (Oelkers et al. 1996). Whether mica-enhanced quartz dissolution is genuinely catalytic, or alternatively involves the interaction of K+ at some stage, requires more research. Some alkali-reactive sandstone in Dutch concrete reveal droplet-shaped volumes with increased interstitial porosity between quartz grains at the ends of detrital muscovite flakes (Broekmans and Jansen 1998). Element maps by EPMA reveal that the alkali-reactive sandstone particles are substantially richer in K (and to a lesser extent also in Na) than the embedding cement, and that alkali-rich gel is extruded into the surrounding paste through AAR-induced cracks (see e.g., Fig. 6). Lithologies with exceptionally high alkali contents include peralkaline rock types like syenite and albitite which are often near monomineralic. Thus, a hypothetical syenite rock consisting of 100 vol% K-feldspar KAlSi3O8 would contain up to 16.9 wt% K2O (= 11.1 wt% Na2Oeq), and an albitite with 100 vol% albite NaAlSi3O8 in the same manner 11.1 wt% Na2O (of course = 11.1 wt% Na2Oeq). Yet higher alkali contents are certainly possible (e.g., kalsilite KAlSiO4 with 29.8 wt% K2O), but would be highly unusual for concrete aggregate, and descriptions of such cases seem absent from the literature (e.g., Diamond 1992). Gillott and Rogers (1994, 2003) describe alkali release from the uncommon carbonate mineral dawsonite (NaAl[CO3][OH]2) in concrete made with silico-carbonatite aggregate (see e.g., Stevenson and Stevenson 1965). Dawsonite occurs widespread in a broad range of lithologies as an authigenous cement mineral in carboniferous sandstones and/or alkaline shales (e.g., Anthony et al. 1990-2003, and references therein). The stability of dawsonite s.s. and ion-exchanged relatives (NH4+, K+) at pH 2-14 and ambient P and T in various solutes has been

Deleterious Reactions of Aggregate With Alkalis in Concrete

295

Figure 6. Optical micrographs showing detrital muscovite in a similar sandstone as in Figure 5, in crossed polarized light (XPL), and in fluorescence (FL). The increased fluorescence at the tip ends of the mica flake may indicate catalytic action. Intergranular cracks visible in FL are connected to a larger network of cracks ultimately extruding ASR gel into the paste.

investigated by Stoica and Pérez-Ramirez (2010). At pH 14, they found Na-dawsonite and its potassium and ammonium relatives to dissolve completely, whereas at lower pH, dawsonites decompose into a number of phases including alumina hydrates, depending on dissolved species present. Moreover, SEM observations revealed extensive recrystallization of dawsonite in carbonate solutions of ~pH 12, demonstrating easy cation exchange. Both dissolution and cation exchange may have contributed to the deleterious ASR described by Gillot and Rogers (1994, 2003).

ALKALI-REACTIVITY POTENTIAL OF ‘SILICA’ Quartz properties and its solubility under ASR conditions By far the most abundant mineral in concrete aggregate worldwide is quartz. The name quartz seems derived from Old German Querklufterz (Lüschen 1979), describing reject material unfit for further processing as queer vein ore (i.e., “gangue”; also compare querulant and quarrel). The abundance of quartz in chert as well as in most other crustal rocks justifies an elaborate description of its ideal and real structures and chemical compositions. Structure and composition determine the mechanical, physical and chemical properties of quartz, essential for understanding behavior in real life applications, including its use for concrete aggregate and its undesirable dissolution in structures under ASR conditions (Dove and Rimstidt 1994). Eventually, this may provide essential clues for optimizing test and assessment methods for aggregates and constituents, as well as criteria for how to distinguish the good from the bad and ugly. In the following, we will use “quartz” as a synonym for α-quartz. Further specification will be supplied as appropriate. Crystal structure of quartz, silica polymorphism and phase transitions. Ideally, the chemical composition of quartz is SiO2, silicon dioxide. The crystal structure can be described as composed of SiO4 tetrahedra, with O located on each apex and Si in its center of gravity. Each oxygen atom is shared with an adjacent tetrahedron, reducing the contribution of each O to any tetrahedron to only half, and thereby reducing the composition of bulk quartz to SiO2. The structure of the α-quartz polymorph stable at temperatures below 573 °C is best explained as a distorted version with lower symmetry of β-quartz that is stable above that

296

Broekmans

temperature (Heaney 1994). A view down the c-axis (i.e., viewing a high-quartz crystal on its pyramid top down its length axis) reveals silica tetrahedra arranged in six rings. Each ring is composed of two intertwined helical chains // c-axis of silica tetrahedra with six-fold symmetry, each chain extending over three ‘ring storeys’ (and that pattern being repeated indefinite). All chains have the same chirality, i.e., the same sense of rotation + translation = screw direction, resulting in left- or right-handed lattice. The chains are linked by independent silica tetrahedral not belonging to either of the helices. Thus, the whole quartz structure in fact represents one single polymer molecule, with six-sided di-trigonal channels parallel to its c-axis. Upon cooling through the β-α transition temperature at 573 °C, the Si-O bond length changes only marginally. In contrast, bond angles change substantially. Due to rotation of silica tetrahedra around the a-axis (oriented perpendicular to the c-axis), the helical chains buckle and contract along the c-axis. The tetrahedral rotation occurring at 573 °C reduces the initial hexagonal symmetry of the six rings in β-quartz with all equal angles, into six rings with ditrigonal symmetry with three wider angles alternated by three tighter angles. This α-quartz structure is stable from ambient temperatures (and far below) up to 573 °C. Looking at thermal behavior of α-quartz, linear expansion is ‘normal’ but highly anisotropic: quartz expands approximately only half as much along its c-axis compared to perpendicular equatorial directions (Rosenholtz and Smith 1941). At 573 °C, the c-axis shows a stepwise length increase induced by the α-β transition, after which further thermal expansion continues as ‘normal’ and again linear. The large thermal expansion anisotropy of α-quartz results in substantial internal stress even over relatively small temperature intervals, making it an important yet often underestimated factor in rock weathering and alteration (Siegesmund et al. 2003). The α-β transition of quartz at 573 °C merely involves change of bond angles by rotation of SiO4 tetrahedra and is hence called displacive. Thus, the transition is instantaneous and reversible. Upon further heating, further transitions may occur, from β-quartz to cristobalite at 870 °C, at 1470 °C from cristobalite to tridymite, and melting at 1705 °C (e.g., Heaney 1994). However, in practice, heating of dry α-quartz at atmospheric pressure results in transition to β-quartz at 573 °C, and a transition directly to tridymite at ~1470 °C, entirely skipping the intermediate phase cristobalite. Upon cooling, (high-) cristobalite transforms displacively to low-cristobalite at 230 °C, whereas (high-) tridymite transforms to low-tridymite in the same fashion (Putnis and McConnell 1980). Depending on heating-cooling history of the silica, heating-cooling trajectories may see alternative transitions to different polymorphs (Heaney 1994). All these transitions are very slow and require exceedingly long dwell times near the transition temperature (i.e., months rather than weeks) in pure silica, some even to the extent that effectively, the transition is undetectable. Transition speeds are greatly increased by a large factor by the addition of fluxing agents or mineralizers (e.g., Na-tungstate) to pure silica, so that heating-cooling experiments can be conducted over the course of hours, and no phases are being skipped (e.g., Kühnel et al. 1987). The main reason for skipping intermediate phases when transitioning from one polymorph to the other, as well as the sluggish reaction rates, is that all these phase transitions require a lot of energy to break Si-O bonds and rearrange all silica tetrahedra into the new structure. Therefore, this type of phase transition is called ‘reconstructive:’ complete reconstruction is required as the new structure cannot be formed through mere displacement (e.g., rotation, translation) of tetrahedra in the precursor polymorph. Note that both cristobalite and tridymite also have low- and high-modifications that can transform through displacive transformation into the other, instantaneously and reversibly. As a result, polymorphs may persist and/or can even be formed meta-stable outside their formal stability region. An example is volcanic glass, a common matrix component in basalt rock, or making its own rock type known as obsidian. Upon aging (no laboratory has the luxury

Deleterious Reactions of Aggregate With Alkalis in Concrete

297

of geological time) under ambient conditions, the unstable random glass structure crystallizes to form cristobalite (e.g., Smith et al. 2001). That variety of obsidian is known as snowflake obsidian, indeed with cristobalite snowflakes. The cristobalite has a denser tetrahedra packing than the glass precursor, thus creating pore space. The process is known as “devitrification” and has also been identified in medieval stained glass windows and artifacts (e.g., Garcia-Vallès et al. 2003). The newly created pore space provides access channels for further fluid infiltration (e.g., Çopuroğlu et al. 2009). A similar devitrification process in opaline silica ages amorphous silica (opal-A, opal-AN) into cristobalite (opal-C, opal-CT). Meta-stability of silica is a delicate and potentially dangerous state in an AAR-context. Transformation to the stable polymorph can be triggered by only a minor event, but may then occur catastrophically. The catastrophic transformation of meta-stable olivine to the spinel structure under earth mantle conditions has been suggested as a cause for deep-seated earthquakes (Putnis and McConnell 1980, p 192). Similarly, one could imagine some meta-stable silica polymorph to go into solution very quickly when exposed to highly alkaline ASR conditions. Summarizing the above, quartz has a di-trigonal symmetry consisting of interconnected SiO4 tetrahedra sharing apical O. The tetrahedra are arranged in two intertwined helical chains linked by independent tetrahedra. The resulting crystal structure has six-sided di-trigonal channels running along the c-axis and parallel to the helical chains. As all tetrahedra are interconnected, a single quartz crystal in fact represents one single silica polymer molecule. The α-β transition in quartz at 573 °C occurs instantly and is reversible due to its displacive nature, and similar displacive low-high transformations in cristobalite and tridymite are known to occur. In contrast, phase transitions between different polymorphs (i.e., quartz, cristobalite, tridymite, glass) are reconstructive, and hence sluggish. This latter quality provides the main reason for the meta-stable persistence of silica polymorphs well outside their stability region, prone to catastrophic transformation. Quartz crystal chemistry and substitutions. The idealized chemical composition of quartz is SiO2, with linked SiO4 tetrahedra forming a crystal structure with di-trigonal channels running parallel to the c-axis. Though certain qualities of quartz are among the purest minerals found in nature (i.e., ultrapure quartz known as lascas), rock quartz does normally contain a number of chemical impurities. These can be incorporated in the crystal structure in a number of ways. Si4+ can be conveniently substituted by Ti4+ or Ge4+, which bears little to negligible effect on the crystal structure of the quartz host. Alternatively, Si4+ can be substituted by Al3+ or Fe3+ that both fit well in a tetrahedral position provided by the surrounding four oxygens. However, to maintain charge balance, trivalent replicants must be supplied with additional monovalent species, for instance Na+, K+ or Li+, which all can be conveniently hosted in the di-trigonal channels. The Al3+ + Li+ coupled substitution is occasionally dubbed “spodumene substitution,” and indeed, the spodumene (LiAlSi2O6) pyroxene structure is a distorted derivative of the quartz structure through complete substitution of Si4+ with Al3+ + Li+ in Si4+Si2O6, just an alternative notation of ordinary quartz (Palmer 1994). Another option for charge compensation involving Al3+ is coupled substitution of yet another Si4+ with P5+. Thus, ordinary quartz noted alternatively as Si4+Si4+O4 then becomes Al3+P5+O4. The mineral AlPO4 is isostructural with quartz, and is known from nature as berlinite with (optical and other) properties virtually identical to quartz (e.g., Motchany and Chvanski 2001), and the 2Si → Al+P replacement is therefore known as “berlinite substitution.” Upon exposure to natural irradiation, Al-substituted quartz turns brownish to smoky quartz, whereas Fe-substituted quartz turns purple into amethyst. Without irradiation, Al- and/or Fesubstituted quartz remains colorless (Nassau 1983). Other colors may be caused by mineral

298

Broekmans

impurities, for instance avanturine green is colored by green fuchsite mica flakes, chrysoprase by green pimelite flakes, and some blue or rose-colored quartzes by micro-fine fibers of dumortierite (Applin and Hicks 1987). Another common substitution in quartz is the replacement of Si4+ by 4H+, changing one SiO4 tetrahedron into (OH)4, which arrangement in quartz is known as a silanol group. Especially quartz formed under hydrothermal conditions (natural or laboratory) contains a considerable amount of silanol, and is hence described as “wet quartz,” as opposed to “dry quartz” with little or no silanol. Both varieties behave markedly different upon stress, wet quartz being softer by a factor ~10 (Kekulawala et al. 1978). Remarkably, amethyst (i.e., Fe-doped quartz) behaves about as soft as wet quartz, both being about half as hard as heat-treated dry quartz. Obviously, the presence of water as silanol in the quartz structure has a very significant effect on lattice rigidity. Many more element substitutions have been described in quartz (Müller et al. 2003; Götze et al. 2004), but the above are the most common and widespread. Which are present in quartz from a given lithology depends on geological conditions under which the quartz was deposited, e.g., P, T, stress and strain, fluid present (amount and speciation), and species dissolved in the fluid including mineralizers and solvents. Thus, the composition of quartz reflects its geological background and history to a large extent. Some substitutions have been calibrated and can now be used as an instrument, for instance Ti in quartz as a very accurate geothermometer (Wark and Watson 2004). Apart from bonded impurities as the lattice substitutions described above, the channelized quartz structure is able to accommodate virtually the entire periodic system (Kats 1962a,b). Using an electrical field, chemical species can be introduced from an aqueous solution into quartz in its structural channels (e.g., Kronenberg and Kirby 1987). This is by no means a novel technique: the phenomenon was already described by Curie (1889), the brother of Mme Curie who discovered radioactivity. Changing polarity will again remove most of the introduced chemical species from the structure. Impurities are very common in natural quartz, which is why only the cleanest qualities of highest purity are suitable for production of Si-metal for electronics and solar cells. While chemical impurities incorporated in the quartz lattice are very common, their concentrations are generally low, rather on the order of a few hundreds to a few thousands of ppm [mg/kg] altogether, ranking quartz among the purest minerals occurring in nature. On the other hand, even such low impurity levels do have a marked and measurable effect on its properties, for instance compressive strength (Kekulawala et al. 1978), as well as solubility (Dove and Rimstidt 1994). Microstructure of quartz. The Si-O bond character in quartz is predominantly covalent, with only ~25% ionicity (Stewart et al. 1980; compare halite NaCl with 63%), explaining the generally poor solubility of quartz in aqueous solutions. Obviously, bonds surrounding chemical substitutions (i.e., foreign species) differ from bulk quartz structure, due to valence differences between the original Si4+ and the substituting species (e.g., Al3+, Fe3+, 4H+, Ti4+, Ge4+, P5+), atomic/ionic radius and resulting field strength, as well as the presence of balancing charges (e.g., Li+, Na+, K+) at positions not normally occupied. For instance, 4H+ substituting 1Si4+ to form a silanol group replaces a single Si bonding its four surrounding O’s with four mutually repelling -OH hydroxyl groups. Consequently, the crystal structure around impurities is distorted relative to the pure ideal material. Thus, impure quartz is under constant mechanical stress, which increases its susceptibility to stress corrosion (e.g., Hadizadeh and Law 1991). The above discusses the effects of various impurities on the mechanical and chemical stability or solubility of quartz at the atomic scale, assuming an otherwise contiguous structure. At a scale of observation of a few hundred up to several thousand unit cell axis lengths (i.e.,

Deleterious Reactions of Aggregate With Alkalis in Concrete

299

a few tens of nanometers up to several microns, the structure of natural quartz can usually be regarded as truly contiguous. Larger contiguous volumes (‘domains’) are known to exist but are not common in natural quartz; they are more typical of synthetic quartz grown under carefully controlled conditions not exposed to mechanical deformation. Domain size is in its own right important in quartz dissolution as inversely related to surface area, in the same fashion as grain and particle sizes at even larger scales. The definition of a domain is a net sum of the various definitions of its borders, representing some type of discontinuity terminating a contiguous crystal structure. Discontinuities include low- or highangle dislocation walls, twin interfaces, or grain boundaries. In general, a crystal structure is distorted by any foreign species present, as discussed extensively above. In addition, the structure can also be distorted by errors that render repetition of crystal-structural build units less than perfect. Such defects in the structure may for instance be an additional atom or ion at a position that normally is unoccupied, or a vacancy at a location that normally would be occupied. Stacking faults result from dislocated atomic bonds latching on to where they wouldn’t in the ideal structure. Imagine a growing crystal plane on which ‘fundamental crystal building blocks’ are being deposited, not unlike Lego bricks of uniform size and shape being stacked in a completely regular pattern. Accidentally, a bond is formed between the low edge of one Lego brick just getting attached, with the top edge of an adjacent brick so that the landing brick’s position is tilted and a step is created. In reality, the misfit of the brick is minimized by redistributing the tilted bond angles over a larger number of adjacent bricks (real-life crystal building blocks are much more flexible than the macroscopic Lego bricks in above rationalization). The result is the formation of a so-called screw dislocation, where the amount of dislocation is largest at the core (one single step) and evens out with increasing distance away. Many more types of dislocations do exist, for which the reader is referred to Putnis and McConnell (1980) and particularly Poirier (1985). Depending on conditions, dislocations can promote crystal growth and/or dissolution (rates), due to interatomic bonds being under stress at the dislocation core. In minerals under natural/geological conditions, dislocations occur together in so-called “dislocation walls,” forming planar arrangements not predicted by laws of crystal symmetry, often curved and/or highly irregular. The dislocation angle is variable, i.e., the degree of misorientation between the individual lattices on either side of the dislocation wall, and a distinction is made between low- and high-angle boundaries. Lattice misorientations 13) in decreasing order as: opal-A, opal-C, opal-CT (Gutteridge and Hobbs 1980), moganite, chalcedony, fine grained quartz (i.e., chert and flint, but also siltstone), and finally, coarse grained quartz. Probably, as data on the solubility of opal-AN and moganite under ASR conditions in concrete are scarce. However, moganite is absent in rocks older than 100-150 Ma, whereas opaline silica is absent in younger rocks of only 60 Ma, which seems to confirm the greater stability of moganite relative to opal. By comparison, quartz is present in the very oldest rocks currently known, with ages up to 4.3 Ga, even including some chert beds. To the best of the author’s knowledge, alkalireactivity of moganite has never been specifically reported, though it must represent a common constituent in Danish, Dutch, German, Belgian, French and UK concrete aggregate containing potentially alkali-reactive chert from the Cretaceous (Heaney and Post 1992). The solubility of silica is furthermore affected by presence of other dissolved species in the concrete pore solution, which may go either way. This effect is known as “salting-in” or “salting-out,” respectively, depending whether solubility increases or decreases (e.g., Marshall and Warakomski 1980; Chen and Marshall 1982; Marshall and Chen 1982; Curtil et al. 1992; Dove and Rimstidt 1994; Broekmans 2002, 2004a). Diamond (1989) reported that solubility of opaline silica in alkaline NaOH solution decreased substantially in the presence of Ca(OH)2, confirming salting-out behavior (also see Curtil et al. 1992). Opaline silica. Alkali-reactivity of opaline silica has been extensively investigated, e.g., Vivian (1950), Tang and Xue (1962), Diamond and Thaulow (1974), Diamond (1976), Ludwig and Bauer (1976), Midgley (1976), Poole (1976), Diamond (1978), Baker and Poole (1980), Gutteridge and Hobbs (1980), Figg (1981), Kawamura et al. (1983), Giovambattista et al. (1986), Curtil et al. (1992), Scrivener and Monteiro (1994), Tajing et al. (1995), Rodrigues et al. (2001), Mitchell et al. (2004), Ponce and Batic (2006), Bulteel et al. (2010), and as part of the general review on ASR and performance testing by Lindgård et al. (2012). The violent alkali-reactivity potential of opal s.l. poses problems with reliable testing of aggregate materials. Around the turning of the millennium, some Danish laboratories were experimenting with further acceleration of mortar bar tests to decrease duration, as desired by building contractors. Thus, the experiment temperature of a given procedure was increased

302

Broekmans

from 38 °C to 60 °C. Despite using a well-known highly alkali-reactive opaline limestone of local origin, initially no expansion was found. Thus, the experiment was repeated after double checking all materials, equipment, instrumentation and settings, to find the same result: no expansion at all, whereas structures containing the same limestone are known to develop damage only short time after. Post-mortem thin section petrography on the exposed test specimen revealed a spongy fabric of ‘empty’ cement paste, the voids containing only scarce relics of dissolved and disintegrated opaline limestone. Such outcome is false and potentially poses a great danger: if expansion were the only criterion used to accept or reject a material, then poorly performing aggregate might be falsely approved for use in concrete. Some countries have included post-mortem thin section petrography as a yet unofficial addition to the standard test procedure, to verify the nature of the expansion data, irrespective whether or not local accept/reject criteria based on expansion behavior have been fulfilled. The opposite situation has been encountered as well, where aggregate performing well seems to have been incorrectly labeled as potentially alkali-reactive. Some sea dredged Cretaceous cherts reveal internal porosity in thin section petrography and therefore classified potentially alkali-reactive and hence unsuitable for use as concrete aggregate, conform Dutch guideline CUR-Recommendation 89 (2008). However, expansion testing has not been able to confirm the alleged alkali-reactivity potential: the material behaves essentially non-expansive and stays within defined limits. The material has resided at the seabed for hundreds to possibly thousands of years, during which soluble silica species (i.e., opal, moganite) may have been dissolved and removed by the saline seawater. However, to the knowledge of the authors, this hypothesis has never been verified by mineralogical or geochemical analysis. Opaline silica often occurs in the interstitial pore space between detrital grains associated with neogenic clay minerals. This is difficult or even impossible to detect in thin section petrography, but may render a given lithology potentially alkali-reactive. Moganite and chalcedony. The presence of moganite and/or opaline silica in silt- and sandstones may well explain the reactivity observed in some varieties. Moganite has been demonstrated identical with lutécite in Parisian Basin sandstones, and there seems no reason why moganite could not also be present in silt-/sandstone from other localities. However, its problematic detection (and reluctant official acknowledgement as a mineral) may have prevented its identification as a common constituent in lithologies frequently used for concrete aggregate, including chert, flint, siltstone and sandstone. To the best knowledge of the present authors, the alkali-reactivity potential of moganite in terms of ASR-susceptibility has never been specifically investigated. Chalcedony has been suggested a fibrous silica composite with α-quartz and moganite in variable proportions. Depending on moganite content and internal porosity, the solubility of chalcedony may differ over a wide range as reported in the existing literature. Comparatively, the solubility of chalcedony is intermediate between that of moganite and pure α-quartz. The variable solubility under natural geological conditions might possibly coincide with the different behavior of (lithologies containing) chalcedony as observed in concrete: some varieties are alkali-reactive, others behave innocuous. Uniform and fully reliable petrographic criteria to distinguish reactive from innocuous chalcedony are currently lacking, though internal porosity is most typically taken as it is perhaps the easiest property to assess. Fine and coarse grained α-quartz. Dissolution behavior of α-quartz is also determined by grain size. Depending on the material’s nature, grain size may either represent “domain size” or “particle size,” both of which may apply to the silica polymorphs occurring in chert and flint. There are fundamental differences between convex (i.e., spherical) and concave surfaces (i.e., hollow, enclosed). Individual α-quartz grains smaller than ~0.1 mm have a measurably increased solubility relative to bulk quartz (Dove 1995), which can be attributed to a

Deleterious Reactions of Aggregate With Alkalis in Concrete

303

shifting ratio between surface vs. bulk thermodynamic free energies towards smaller grain/ domain/particle sizes. Meanwhile, concave surfaces of equally small radius like for instance fractures from brittle deformation/compaction, interstitial pores between fine sedimentary detritus, nanopores in sponge spiculae, contact points between adjacent quartz grains, have a markedly decreased solubility, for the same thermodynamic background (Brantley 1992). As a consequence, α-quartz particles 17 vol%. Thus, neither Reaction (4) nor (5) can contribute to the expansion observed in ACR-damaged concrete. However, a net shrinkage of >22 vol% quite likely enhances paste permeability by increased microporosity as a consequence of reduced solids volume, as correctly implied in Katayama (2004a). Meanwhile, Reaction (5) also reduces the pH-buffering capacity of the concrete bulk by paste carbonation. Prince et al. (1994) identified brucite + pirssonite Na2Ca[CO3]2∙2H2O as reaction products (alternative to brucite + calcite) in experiments exposing powdered dolomite to alkali solutions (as opposed to field or lab concrete), attributed to the relative activities of Ca2+ and Na+. Sherwood and Newlon (1964) also identified gaylussite Na2Ca[CO3]2∙5H2O and anhydrous bütschliite K2Ca[CO3]2 (reported as Ca2K6[CO3]5∙6H2O) as accessory phases. Katayama also mentions pirssonite, and in addition thermonatrite Na[CO3]∙H2O and trona Na3H[CO3]2∙2H2O occurring in cracks in close proximity to alkali-reactive dolomite. It is unknown whether these phases are stable or represent intermediate products that eventually transform to brucite + calcite. The alkali-reactions occurring in non-dolomitic limestone are less well established, which may be due to poor or inadequate characterization of reactive constituents, and/or due to the intricate differences in crystal structure and chemistry among carbonate minerals. The most common rock-forming carbonate minerals for concrete aggregate are calcite CaCO3 (including magnesian varieties) and dolomite CaMg[CO3]2, followed at some distance by magnesite MgCO3, ankerite CaFe[CO3]2, siderite FeCO3, and less common ones like rhodochrosite MnCO3, kutnahorite CaMn[CO3]2, etc. Apart from above idealized end-member compositions,

Deleterious Reactions of Aggregate With Alkalis in Concrete

319

natural calcite almost always contains subordinate quantities of Mg, Fe and Mn, in addition to traces of Sr, Ba, and others (e.g., Wolf et al. 1976). The presence of foreign species causes local disorder in the calcite structure affecting its stability, which has been studied extensively in particular for Mg (see e.g., Mackenzie et al. 1983). There is a general consensus that high-Mg calcites see decreasing aqueous solubility with increasing near-ambient temperature (Plummer and Busenberg 1982), and that high-Mg has a tendency to recrystallize to dolomite + low-Mg calcite (Bertram et al. 1991; Loste et al. 2003), also at ambient temperatures. The impurities appear to build a superstructure, consisting of layers of very pure CaCO3 intercalated with layers of ‘impure calcite,’ their respective volume proportions depending on total impurity content (Reksten 1990). Upon recrystallization, such superstructure layers in high-Mg calcite may transform into low-Mg calcite, the excess Mg precipitating as dolomite micro-inclusions, as observed by Althoff (1977), and Lohmann and Meyers (1977) in natural rocks. Stipp et al. (1996) studied the dynamics of freshly cleaved calcite surfaces exposed to air, using several scanning-probe microscopy (SPM) techniques. They observed hillocks and holes at nanometer scale spontaneously emerge and disappear again over the course of minutes to hours at room temperature, which effects appeared promoted at increased relative humidity. How these observations relate to a calcite (super-) structure was not further addressed in that study. Using atomic-force microscopy (an SPM variant), Harstad and Stipp (2007) identified an inverse correlation of trace element impurity content in apfu relative to Ca of very pure natural calcite (known as Iceland Spar among collectors) versus dissolution rate in (also very pure) MilliQ water. They conclude that the inhibiting effect of common impure calcite in natural marble, limestone or chalk, may be substantially greater. Finally, a recent study by Lakshtanov and Stipp (2010) shows spontaneous precipitation of calcium carbonate in the presence of dissolved silica. However, studies extending above findings to undesirable dissolution of nondolomitic carbonate aggregate in concrete are still lacking to the best knowledge of the author. In field concrete, alkali-reactivity of non-dolomitic carbonate has a similar appearance as alkali-dolomite reaction: the reaction gradually proceeds inward, does not cause expansion cracks, and reduces the mechanical integrity of the concrete. The surrounding cement paste may show cloudy carbonation, suggesting a similar alkali-recycling mechanism as for alkalidolomite, conform Reaction (5). Detailed routine characterization of deleterious ACR in general is hindered in practice by the particular properties of carbonate minerals. For instance, due to extreme birefringence, a thin section of a carbonate rock displays ‘cream-higher order’ in crossed polarized light, obliterating the presence (let alone identification) of any but strong-colored minor constituents. The moderate abrasion hardness of common carbonates combined with rhombohedral cleavage causes the mottled view known as ‘bird’s eye effect’ in thin sections, which may even be used diagnostically to identify trace carbonate in a siliceous rock type, but which again is impracticable in the identification of finely dispersed, fine-grained accessory constituents in a carbonate-dominated rock. Visibility of minor and trace minerals in carbonate rock is greatly enhanced in double-polished thin sections, prepared with bonded-diamond abrasives to minimize undercutting and artificial specimen topography (Humphries 1992; Vandervoort 1999). Distinction between various carbonate minerals is furthermore facilitated by convenient etching and staining protocols (Friedman 1959; Warne 1962; Hitzman 1999). Another popular method to determine mineral content in bulk whole-rock is by XRD on powdered material. A computer program will break down the diffractogram into its constituent components and faithfully list all minerals found. However, unless proper care is taken, powder XRD analysis of bulk whole-rock is prone to artifacts from both sample and specimen preparation. In carbonate rocks containing silicate impurities, intergrinding (= comminution of soft minerals by harder species) which in—impure—carbonate rocks may lead to overgrinding of carbonate and undergrinding of silicate or other impurities, deflecting grain

320

Broekmans

size (-distribution) away from optimum for both (carbonate: too fine, silicates: too coarse), depending on total impurity content as well as comminution equipment (e.g., ball mill, mortar and pestle, micronizer) (Bish and Reynolds 1989; Buhrke et al. 2002; Bish and Plötze 2010). Furthermore, carbonate rhombs as well as mica or clay mineral flakes are prone to preferred orientation upon mounting of the comminuted sample material in the specimen holder, leading to overexposure of certain diffraction peaks, and underexposure of others. Together, sample and specimen preparation artifacts complicate simple correlation of peak height or surface area with quantity, also affecting background noise level and detection limits (Klug and Alexander 1974; Bish and Plötze 2010). Nevertheless, above issues seem generally not considered important enough to report, as apparently few authors only care to specify preparation and analytical procedures in adequate detail. The alkali-reaction of carbonate rock types not containing dolomite are typically attributed to their content in argillaceous (clay) minerals, notoriously known for their swelling capacities (see preceding section), which formally is not considered as deleterious ASR. In thin section, alkali-reactive non-dolomitic limestone particles show cracking or spalling and often develop a halo of carbonated paste, particularly along emanating cracks. The cracks appear filled with siliceous ASR gel, amorphous or crystalline depending on aging. Chemical composition usually includes substantial Mg, reflecting its different origin compared to ASR gel in concrete with siliceous aggregate. Details on ACR-gel chemistry are elaborated extensively by Katayama (2004, 2010a, also see references therein), based upon analysis by SEM-EDS. He identified silica gels in ACR concrete resembling sepiolite Mg4Si6O15[OH]2∙6H2O, clinochlore (Mg,Al)6(Si,Al)4O10[OH]8, and nearby compositions with variable Ca and minor amounts of Al, SO4 and Cl. However, crystal structure is unconfirmed as of yet, so mineral names are based solely on chemistry and should hence be applied with suitable reluctance. Bachiorrini and Mantanaro (1988) convincingly identified ASR-gel containing Si and Na (by SEM-EDS) in alkali-reactive fossiliferous carbonate aggregates liberated from seven different ACR-damaged field structures, presumably all in Italy. They also confirmed the presence of silicates (feldspars, micas) by XRD on powdered sample material, and determined the degree of disorder in the silica present (strained quartz, chalcedony) by IR (Bachiorrini et al. 1986; Bachiorrini 1987). Development of deleterious ASR appeared unrelated to internal porosity of the aggregate by MIP. Two decades later, the exercise was independently repeated by Katayama in Japan and Grattan-Bellew and co-workers in Canada, using sample materials from Austria, Bohemia/ Germany, Canada, China, Nevada, United Kingdom and elsewhere (Katayama and Sommer 2008; Grattan-Bellew et al. 2008). Insoluble residues were liberated from the carbonate by phosphoric acid according to the procedure described in Japanese Standard JCAS I-31 (1996), in turn derived from the method by Talvitie (1951); also see McCrea (1950) and Ray et al. (1957). Apart from sedimentary detritus, the insoluble residues were also found to contain illite clay and silica of authigenic/diagenetic origin, in variable proportions. Figure 7 on p 560 in Grattan-Bellew et al. (2010) and Figure 6DE on p 654 in Katayama (2010a) reveal the presence of idiomorphic quartz crystals, from dissolution of detrital silica (and/or silicates) and reprecipitation as neogenic quartz. In addition, opal-AN and opal-CT were both identified. A clear correlation exists between free silica content (i.e., excluding clay minerals) and amount of expansion in laboratory testing, which earlier research failed to identify (Grattan-Bellew et al. 2010, and references therein). Kreshkov et al. (1965) describe a rapid wet-chemical procedure to determine the content of free (amorphous) silica in clays by selective digestion, which also could be helpful in this context. Due to the different nature of the alkali-carbonate reaction compared to the alkali-silica reaction, it expansion behavior, and the nature and composition of the mineral constituents

Deleterious Reactions of Aggregate With Alkalis in Concrete

321

involved, testing of aggregate materials for their ACR-potential requires a different approach. After its carbonate nature has been confirmed, carbonate concrete aggregate shall also be assessed by whole-rock XRF to determine content of Ca, Mg, Al and Si, which are then used as criteria to classify risk-level as low, moderate or high, conform RILEM AAR-5 (2005), CSA A23.2-09 (2009a); also see Qian et al. (2002), Lu et al. (2004), Sommer et al. (2004). However, element species of interest like Mn, Fe, Sr and others (as referred to in the above) should perhaps be included too, as well as the volatile content (ignition loss LOI as the sum of CO2+H2O; also see Broekmans 2009). With current XRF instrumentation and standard laboratory equipment, all can be included in a single analytical procedure, specimen preparation by fusion in excess flux being part of that.

Other alkali-reactive species not being minerals Eglinton et al. (1994) described a highly uncommon example of alkali-aggregate reaction in concrete from Trinidad, that shortly after placement revealed reddish-brownish stains and blemishes on its surface. Assessment of the concrete with a range of analytical techniques including XRD, wet chemistry and thin section petrography eventually identified an ironorganic coating of aggregate particles (or impregnation of its natural porosity) as the cause of the trouble. The high alkalinity of the concrete remobilized these compounds which work as a retarder inhibiting proper hydration and setting of the paste. Adhesion of aggregates particles with embedding cement paste is compromised, which has resulted in a pock-marked surface where particles had become completely detached as well as limited loss of compressive strength, and substantial loss of tensile strength (also see Broekmans 2009). The mineral content of the aggregate lithologies present did not reveal signs of deleterious alkali-reactivity, and the aggregate particles did not disintegrate either. To some extent, the damage described by Eglinton et al. (1994) resembles that caused by sparingly soluble salts deposited on the surface of certain sea-dredged aggregates. Of course, sea-dredged aggregates are thoroughly washed and rinsed in plenty of water to remove all salt, but some sparingly soluble species will nevertheless retain in some rare occasions. Slow dissolution of these salts in concrete over time leads to detachment of the particles and limited strength loss, whereas the chloride introduced to the concrete interior causes corrosion of the steel reinforcement. In both above examples, the minerals composing the aggregate are not alkali-reactive themselves and do not contribute to the deterioration. Rather, the aggregate serves as a medium hosting the deleterious compounds. Thus, above examples do not represent ‘alkali-aggregate reaction’ in the conventional perception, and would be more appropriately named as ‘rapid concrete damage from soluble (organic) constituents in naturally contaminated aggregate,’ or something similar.

LABORATORY ASSESSMENT OF ASR CONCRETE In order to determine a cost-effective and technologically feasible repair and rehabilitation strategy, thorough assessment in adequate detail of the materials composing a given AARdamaged structure, the way they have been applied and used to present a structure, as well as their current status and prospective development, is required (Wood 2007; CURRecommendation 89 2008).

Sample acquisition and handling In geological fieldwork practice, rock samples for analysis by thin section petrography or whole rock geochemistry ‘at home’ are typically collected through punitive martelation to subjugate even the most defiant rock types. Away from the point of impact, rock structure and

322

Broekmans

mineral texture are little affected, and as crack (micro-) fabric is hardly ever point of attention in traditional geological research, this is an effective way and analytically acceptable of collecting sample materials for research. In contrast, samples of hardened concrete are nearly always extracted by core drilling. The total cost of a concrete sample per kilogram easily exceeds its weight in gold, if all expenses made for core extraction, packaging, shipping, storage, and sample/specimen preparation are taken into consideration, as well as indirect expenses for items like scaffolding, road blocks and signposting, safety measures, which in densely populated areas are minor compared to the burden society pays for productivity loss, travel delays, increased pollution and the like incurred by traffic disruption or reduced structure service. Identification of deleterious AAR as cause of damage observed in a concrete structure, as well as its quantification and progress over time rely to an important extent on crack fabric and spatial distribution through petrographic assessment of plane and thin sections. Obviously, results will be biased by any artifact cracks introduced during the course from sample extraction until finished specimen. Poor past experience with assessment of AAR in field structures worldwide has taught the hard way about what is ‘bon ton’ and ‘not done.’ Applying rigorous procedures for sample extraction, handling and preparation, and the use of modern equipment and instrumentation, introduction of artifacts are reduced to a minimum and can be recognized as such by an aware observer. Nevertheless, only the fewest recent guidelines do actually prescribe sample acquisition and handling strategies as an essential part of the whole assessment procedure, like e.g., Dutch CUR-Recommendation 102 (2008). In daily practice, structure concrete is perceived as a robust material tough enough to take rough handling whether by impact or otherwise, and practically insensitive to wet, dry, heat or cold. The typical coring contractor has a primary interest in producing ‘holes’ in concrete, and in their typical perception, cores coming out represent “construction and demolition waste” (CDW) and may thus be handled without respect. Contrary to popular belief, this could hardly be farther from reality for damaged concrete to be subjected to a kind of forensic assessment: AAR damaged concrete is weakened by internal cracking and therefore fragile, whereas AAR gel is sensitive to changes in humidity conditions and consequently temperature. Thus, AAR damaged concrete is a delicate material, and to preserve its status as in the structure it has been extracted from, does require special attention. Sample acquisition starts with core drilling. First, a sturdy drill stand is firmly affixed to the structure surface using an anchor bolt or similar, to minimize off-axis momentum during drilling. A fresh diamond crown is mounted on a drill with tight fitting bearings to ensure the drill tube runs free from the cut faces, and to allow the coolant to flush the debris out. Minimum force is applied to sink the drill into the concrete, to avoid off-axis deflection and excessive heating. The core is to be drilled using the least possible amount of water necessary to prevent any significant temperature increase of the core, while keeping leaching of soluble constituents from the concrete at a minimum. When the desired depth is reached, the core is carefully dislodged using least possible force. Arguably the most delicate way to liberate a core is through wet swelling of a wedge of peach wood inserted in the cut, a technique already known by the ancient Egyptians. A coherent core is then lifted out using one or several brass strips with a small perpendicular edge at one end, grappling the core down hole. A core consisting of disintegrated concrete can be retrieved using a vacuum cleaner with a piece of steel mesh over its mouth. The core is immediately and quickly rinsed with little water to remove adhering drill slurry, then tightly wrapped in moist (not soaking wet) plain cotton rag, if necessary piecing loose parts back together in their original positions. The cotton wrapped cores are then firmly rolled in cling foil also covering the ends, adequately securing the cores against exsiccation during transport.

Deleterious Reactions of Aggregate With Alkalis in Concrete

323

Next, wrapped-up cores are placed individually in thick-walled PVC tubing. Including all wrapping, cores Ø100 mm fit snugly in standard Ø110 mm sewer tubes. Finally, tubed cores are boxed parallel and the remaining space filled with polystyrene balls or chips to fix the tubes firmly against moving inside the box, securing the cores against impact and mechanical damage during transport. Deleterious AAR negatively affects tensile strength of the concrete (ultimately to the point of disintegration), and the weakened concrete has to be protected against tensile stress during transport. Likewise: concrete foundation piles are reinforced to prevent cracking during transport and handling in sideways position, but hardly need any while being driven or in use in upright position. Thus, tall and slender cores (e.g., Ø70×400 mm, with aspect ratio ~1:5.7) require heavier (temporal, external) reinforcement than short and stubby ones (e.g., Ø170×400 mm, aspect ratio ~1:2.4) (Fig. 10). Taken together, above steps define a procedure that has been proven effective to prevent exsiccation as well as to minimize mechanical damage during (international) transport and handling. Obviously, many parts of above procedure can be neglected if the extracted cores are not intended for petrographic assessment, but for instance to determine compressive/tensile strength, or for bulk chemical analysis.

Figure 10. Core drilling and handling artifacts; a) cracks splaying from a rebar dislodged by excessive force and wedging between drill crown and core wall, in PL (left) and FL (right); b) cracks from fierce hammering to remove the core from the hole destructing the top surface and rough handling in PL (left; note smears from cutting) and FL (right); c) cracks along core periphery (top left, bottom right) from stress relaxation, a very telling artifact in expansive but in the structure still not cracked ASR concrete, in FL.

324

Broekmans

Impregnation-fluorescence petrography Crack fabric and spatial distribution are essential to determine whether or not deleterious AAR is a contributing damage mechanism. Depending on width and scale of observation, cracks can be exceedingly hard to visualize, let alone distinguish from artifacts. An effective way to enhance visibility of pores and cracks in thin section is by adding a blue dye to the epoxy resin, a technique commonly applied in sedimentary petrography for (oil, gas) reservoir analysis (e.g., Humphries 1992; Ingham 2011). However, contrast is not very great so that thin intragranular cracks and/or small interstitial voids are still easily overlooked, and the application of blue dyed epoxy is only effective in thin sections studied in transmitted illumination, not in reflected light, nor on plane sections. Moreover, the method’s applicability for quantitative determination of porosity or other features is limited as the thin sections are typically produced following standard techniques not specifically designed to minimize the chance for preparation artifacts. A far superior method to visualize any cracks present in both plane and thin sections is impregnation with epoxy carrying a fluorescent dye known by its generic name Hudson Yellow, which is also marketed under several brand names. It is non-toxic, and dissolved in a sterile solution for ophthalmic inspection of cornea damage. Dissolved in epoxy, Hudson Yellow has an excitation wavelength at 485 nm in the blue part of the spectrum (i.e., not UV!), and an emission wavelength of 515 nm in the green. Upon irradiation with light ≤ 485 nm (e.g., blue, violet or UV) below the excitation wavelength, the dissolved pigment/dye produces a bright greenishyellow fluorescence. High intensity excitation or prolonged exposure to less intense sources quenches fluorescence beyond self-restoration in a process known as photo-bleaching. Therefore, fluorescence-impregnated specimen must be stored in the dark at cool room temperature. The Hudson Yellow pigment granules are (scarcely) soluble in organic solvents. To prepare dyed epoxy, an amount of pigment is premixed for at least 24 h with the resin at a predefined weight ratio, the hardener is added during preparation. The type of epoxy is selected to have low viscosity as well as a high capillary affinity (i.e., good wetting of geomaterials), which properties in combination with correctly applied preparation procedures guarantees uniform impregnation. Procedures are laid down in detail in Danish national guideline DS 423.39 and DS 423.40 (2002) for plane and thin sections respectively, and are outlined in brief below. To prepare an impregnated plane sections, cores are carefully dried at 35 °C to remove water from the pores and make these accessible for epoxy. This may sound contradictory to the measures taken to prevent exsiccation as described in the previous section but really isn’t, as in the laboratory exsiccation can be accurately controlled which it can’t be during transport. Concrete shrinks upon exsiccation, and if uncontrolled (e.g., single-sided axial/peripheral, incremental, temperature gradient) it may induce crack artifacts, which can be avoided. Temperature is limited to 35-50 °C, i.e., well outside the range at which some constituents start to decompose, e.g., ettringite (also see Pöllmann 2012, this volume). Depending on core dimensions, coherence, paste porosity and constituents, exsiccation to constant weight may take over two weeks’ time. Next, the dry core is tightly wrapped in a sturdy PE bag together with a length of PE hose running up from the core bottom. Wrapped cores are placed in a chamber, individual tubes put through an air-tight duct and closed with a clamp, after which the chamber is evacuated. When constant pressure (485 nm and only allows rays ≤485 nm (which an incandescent bulb has no problem with delivering) to illuminate and excite the specimen. The emitted luminescence is passed through a long-wave pass (LWP) band filter that cuts off all light 99.5 wt% and have once been extracted for TV tube glass, but were never used in concrete. To prepare for analysis, bulk concrete was comminuted using a jaw crusher and a spring loaded roller mill to pass a 0.5 mm screen. Next, a ~35 g split was pulverized in a vibratory disk mill with agate lining. A powdered subsample was mixed with excess Limetaborate and digested at 1050 °C in a platinum crucible. The resulting tablet was analyzed by XRF in a Philips PW1480 instrument for Na, K, Mg, Ca, Mn, Al, Fe, Si, Ti, and P. Ignition loss LOI (comprising both CO2 and H2O) was determined gravimetrically in the same procedure.

Broekmans

332

A separate subsample of powdered concrete was digested at room temperature in excess aqua regia, a mixture of concentrated hydrochloric and nitric acid in a ratio 3HCl:1HNO3. While the mixture is able to dissolve gold (and in a 9:1 ratio platinum), it hardly dissolves silica, but on the other hand leaches some silicates. The solution was then analyzed by ICPAES for the same main element species as by XRF, minus Si which isn’t stable in this solution (Varma 1991). Additional subsamples were analyzed for total-C and total-S by Leco, for which sample powder is incinerated at 1450 °C and the gaseous emission analyzed spectroscopically. Detection limits by XRF are on the order of 0.01 wt% for most species except Na2O (0.10 wt%), SiO2 (0.50 wt%), and LOI (also 0.50 wt%). Results from XRF, ICP-AES and Leco were recalculated as oxides Na2O, K2O, MgO, CaO, MnO, Al2O3, Fe2O3-total, TiO2, SiO2, P2O5, CO2, and SO3, respectively. Water content H2O was calculated as LOI minus CO2. Oxide contents listed above by XRF were summarized together with LOI and Cl, with net sum totals ranging between 100.0±1.50 wt%. For oxides also determined by ICP-AES, relative recoveries were calculated as the ratio of ICP-AES over XRF (in wt%) ×100%. Two typical examples are given in Table 3 below. Concrete thin section petrography prior to analysis revealed that the mature fluviatile aggregate mainly consisted of quartz in various types of sand- and siltstones, (milky) vein quartz, with a few vol% of alkali-reactive chert containing chalcedony and minor opal. Particles are well rounded, and even when the material is polymict, differences in mineral content among the sand- and siltstones are only small, first and foremost pertaining to the ‘mineral pigment’ interstitial between the quartz grains, e.g., limonite, organic matter, glauconite, pumpellyite, responsible for their different colors. The fine aggregate fraction contains the occasional white mica flake, a few K-feldspar sand grains, even more rarely a carbonate particle. Other trace minerals like e.g., apatite, rutile, zircon, monazite are even rarer and may be virtually absent in a given thin section. This material is considered characteristic for mature Dutch concrete aggregate as mentioned before. Table 3. Whole-rock geochemical analysis of two ASR concretes with aggregates of different maturity, by XRF, gravimetry (LOI), Leco (total-S, total-C) and ICPAES, in wt%. Further explanation in text. mature aggregate

containing species

XRF

Na2O K2O CaO MgO MnO Fe2O3-total Al2O3 SiO2 TiO2 P2O5 SO3 LOI

0.19 0.69 10.80 0.25 0.03 1.67 2.65 77.90 0.15 0.09 0.30 5.15

0.05 0.24 8.76 0.20 0.01 1.24 1.27 — 0.03 0.07 — —

sum total sum ICP-set

99.90 16.50

CO2 H2O (calc.)

1.14 4.01

ICP

immature aggregate recov.

XRF

ICP

recov.

26 35 81 80 33 74 48 — 20 78 — —

2.15 1.15 15.70 3.42 0.08 4.61 10.50 55.40 0.59 0.09 0.67 6.35

0.12 0.28 11.00 1.52 0.03 1.92 2.10 — 0.17 0.07 — —

6 24 70 44 38 42 20 — 29 78 — —

11.90 11.90

12 72

100.70 38.30

17.20 17.20

17 45

— —

— —

0.92 5.43

— —

— —

Deleterious Reactions of Aggregate With Alkalis in Concrete

333

In contrast, the (ASR concrete containing) immature polymict sedimentary aggregate material of an undisclosed resource, consisting of angular to sub-rounded particles of granite/ gneiss, greywacke, various sand- and siltstones, minor retrograded gabbroic rock, and mineral fragments. Thin section petrography identified abundant calcic clinopyroxene partially altered to clinoamphibole, oligoclase, K-feldspar, chlorite with sagenite, epidote, minor calcite/ ankerite carbonate, traces of relic altered biotite, apatite, zircon, monazite, and rare cleavage particles of deep purple fluorite in the fine fraction. In general, the mineral assemblage of the bulk immature aggregate is inferior to quartz regarding resistance to mechanical abrasion and chemical alteration. Upon sedimentary reworking, most other minerals would be pummeled and/or dissolved away, enriching quartz in the residual material. While it is difficult to indicate a typical provenance area, immature polymict sedimentary material is very commonly used for concrete aggregate. Table 3 shows clear differences in recovery ratio of ICP-AES over XRF between the two aggregate materials with different degrees of maturity. The poor recovery ratios for Na, K, Mn and Ti illustrate the rather limited partial digestion of common (but not necessarily abundant) aggregate minerals in aqua regia. The composition of the non-soluble residue from the mature aggregate by XRF+LOI is chemically indistinguishable within analytical error from virgin material, though its mineral content will have changed after the aqua regia extraction (unpublished data Broekmans). As the origin of the immature material is unknown, this cannot be verified, but seems rather unlikely based on mineral content of the material embedded in concrete. In summary, liberation of alkali-reactive aggregate from its ASR-concrete host by selective digestion in aqua regia is only possible for aggregates very rich in quartz and poor in any other minerals, e.g., mature sediments that have been extensively reworked during their geological history. These may include polymict materials composed of several types of quartzrich lithologies. Use of selective dissolution using excess concentrated strong acids (blends) for immature aggregate materials is rather limited, due to contamination by dissolving mineral phases from the aggregate. However, the method is comparatively fast and uncomplicated (if the aggregate’s mineral content permits), and uses standard off-the-shelf chemicals and equipment. Inorganic acids are unsuitable for a range of volcanic rocks including basalts, as they would readily dissolve any olivine Mg2SiO4 and nepheline [Na,K]AlSiO4 present. Demoulian et al. (1980), developed a method for the assessment of unhydrated slag in cement paste by selective dissolution in EDTA (ethylene-diamine tetra-acetic acid, or [HOOC-CH2]2NN[CH2-COOH]2), a common and harmless industrial chemical available in large quantity. For their excellent metal ion-scavenging properties, EDTA and its derivatives are widely used in medical chelation therapy to treat (heavy) metal poisoning. St. John and Goguel (1992) applied solutions of 1M nitric acid HNO3 and/or several EDTA mixtures to a number of basalts and andesites used for concrete aggregate. Whereas the nitric acid did attack and partially dissolve the rocks, in particular the glassy matrix, the EDTA sensu lato was found not to, apparently only attacking the paste. More refined further experiments on additional rock types and nepheline mineral separate and detailed analysis of solutions and solid residue confirmed that the EDTA-solutions used efficiently dissolved the cement paste (save hydrotalcite, if present), but left the investigated aggregate lithologies and minerals unaffected (Goguel 1995, 1996). However, preparation of the necessary solutions was rather complicated due to required prior purification of the EDTA, demanding access to special laboratory equipment as well as skilled personnel. Implementation of the EDTA-solution method as standard practice is greatly promoted by solutions being available as regular inexpensive stock-ware, and with a practicable shelf-life. While the methods described above for chemical separation of (deleterious) aggregate from concrete allow for chemical separation of paste from aggregate and their chemical char-

334

Broekmans

acterization, they are destructive as they require comminution of the materials prior to digestion. Physico-mechanical liberation of aggregate particles. Instead of applied chemistry to liberate deleterious aggregate from ASR concrete, aggregate and paste can be separated using mechanical methods. In its simplest and most low-tech form, this comprises hammering between sheets of thick paper, handpicking of liberated particles, and reiteration until a satisfactory result is achieved. Essentially the same result can be achieved with mechanized comminution equipment, like a jaw crusher, a spring-loaded roller mill, a cross-impact mill, or other. Liberation is generally mediocre to poor with paste adhering to aggregate particles, which is difficult to remove completely. Individual fragments thus extracted may have suitable size for thin section petrography, but are likely to show fracturing artifacts from the merciless and uncontrolled comminution and extraction procedure. These artifacts may vary from minor cracks in the particle interior to complete breakage and size reduction. The latter may also separate layered rock types (e.g., banded gneiss) into their composing lithologies (i.e., granite and amphibolite), artificially flawing the modal composition of the bulk. Alternatively, aggregate has been attempted liberated from cement paste by a freeze-thaw procedure (Haugen et al. 2003). The concrete sample (a drilled core) is saturated with water first, then immersed in liquid nitrogen (−196 °C) until completely cooled. Next, the sample is heated in a microwave oven until thawed, after which the procedure is reiterated until the sample disintegrates, if necessary with the gentle aid of a hammer and a steel prod. Separated aggregate particles are incompletely liberated with paste remnants adhering (notably for the finer fraction) and/or breakage artifacts (coarse fraction), though arguably to a lesser extent than with ordinary mechanical comminution. Finally, adhering paste was tried removed by rinsing with concentrated hydrochloric acid, which however also dissolves any carbonates present. Thus, in summary, this method to liberate aggregate from hardened concrete too introduces a number of artifacts difficult to recognize or control. Above methods using traditional equipment to liberate aggregate from (AAR) concrete have limitations of instrumental, procedural or artifact-related nature, effectively inhibiting or at least complicating routine extraction of coarse and fine particles for assessment and analysis. However, recent advances in crushing technology for the mineral extraction industry proffer interesting new opportunities. Rock samples can be shattered by applying a highvoltage electrical discharge through contact with electrodes. The degree of shattering can be controlled by customizing the pulse profile, including rise and decay times, pulse height and duration, total delivered charge, etc. (Andres and Bialecki 1986; Andres 1989, 1995; Bialecki et al. 1992). The pulse creates a plasma along existing grain boundaries and builds up internal charge, rapidly disintegrating the rock. The technique is known as electric-pulse disintegration (EPD), and has been successfully applied for the extraction of zircons and apatite for rock age dating, for the liberation of gems (e.g., sapphire, diamond, amethyst, demantoid, alexandrite) from host rock (Andres and Bialecki 1986), for the recovery of tiny fossils (Saini-Eidukat and Weiblen 1996), for the liberation of precious metals (e.g., Pt, Au, Ag) in ore deposit prospects (McDonald et al. 2010; Chernet 2012), and for the recovery of finely intergrown minerals difficult to separate otherwise (Malarkey et al. 2010). The material emerging from EPD processing is exceptionally well liberated with negligible adhering phases (including cement paste), yet leaving patina—if present—intact. In addition, the product mostly retains its original grain size as it is separated along existing intermineral boundaries, reducing production of ‘crusher dust’ to a minimum. Consequently, in mineral processing (e.g., ores, industrial minerals, gems), EPD vastly increases recovery rates (i.e., the extractable percentage of the total content) and through that the value of a deposit.

Deleterious Reactions of Aggregate With Alkalis in Concrete

335

Ongoing research focuses on development of EPD as a contiguous process, required for large scale implementation at mining sites. An overview and comparison to traditional comminution and separation methods is given in Cabri et al. (2008), further detail on technical background in Rudashevsky et al. (1995, and references therein). Electric-pulse disintegration has also been applied to concrete in a context of reuse and recycling of (sound, non-deleterious) aggregate (Zinovyev et al. 1988; Zinovyev 1989), also in large scale pilot plants for processing of “construction and demolition waste” (CDW) (e.g., Bluhm et al. 2000). Coarse aggregate extracted from concrete with EPD has virtually no adhering paste and displays negligible size reduction from breakage. Particles often have a frosted-lustrous patina, which is attributed to chemical etching in the alkaline concrete paste. Laboratory equipment and/or services for EPD batch processing are commercially available from a number of suppliers, however, to the best knowledge of the author, EPD has not yet been tested on ASR concrete.

In situ chemical analysis of ASR gel by SEM-EDS or EPMA The reliability of chemical composition data is greatly affected by the way an SEM-EDS or EPMA instrument is set up and operated. A properly calibrated and stable instrument run by a cunning operator is able to produce better quality data than a more recently installed instrument operated with incorrect settings and conditions. With today’s sophisticated software, operation of even the most complex instruments has almost become child’s play. Moreover, regardless settings and conditions, the instrument will diligently spit out a list of numbers allegedly representing the ‘chemical composition’ of a sample. However, this does not discharge the instrument user from critically reviewing initial results and adjust settings and conditions as required so as to elevate reliability of the obtained results to an acceptable level. The section about ‘operator awareness’ in Diamond et al. (1974, p 912) has by no means become obsolete! Effective and straight-forward comparison of chemical ASR gel compositions reported in the literature is hampered by the fact that their acquisition sensu lato is rarely specified in adequate detail, rendering it nearly impossible to evaluate data reliability. Undeniably, there are differences between instruments at various facilities regarding their technical capabilities and analytical options, how they are set up, operated and maintained, as well as operator background and experience, with repercussions for the reliability of the reported data. Thus, adequate specification of relevant settings and conditions for analysis by SEM-EDS or EPMA provides the critical reader with essential details to assess reliability of the presented data. The below assumes thin section petrography using an optical microscope prior to EPMA or SEMEDS assessment, to identify areas of interest and to verify that spots for prospective analysis are not underlain by or clouded with other minerals, which is visible in transmitted light, but not in an electron microscope of some sort. Sample and specimen preparation. A modern SEM instrument does not per sé require sample preparation to be able to render an image, a key advantage for the study the micromorphology of small crystals, ASR gel, or fracture surfaces in concrete or mortar. However, an unprepared mineral surface is unsuitable for reliable quantitative analysis, for various reasons (Kjellsen and Monsøy 1996; Detwiler et al. 2001; Kjellsen et al. 2003). The placement of energy-dispersive and/or wavelength dispersive X-ray detectors (EDS and WDS, respectively) on SEM and EPMA instruments is restricted to coincide with the “take-off angle” at which characteristic X-rays are emitted. Multiple detectors on a single instrument share a common focus, essentially located at the point of incidence of the probe beam on the sample. By adjusting sample elevation, a flat polished specimen with negligible topography can be positioned to optimally comply with the take-off angle at the actual acceleration voltage. This is simply not possible with non-prepared specimen: the random surface topography scatters emitted X-rays in an uncontrollable manner, which may again be partially absorbed

336

Broekmans

by protruding parts of the sample in the line of sight to the detector. Both affect peak heights in an unpredictable way, occasionally to the extent that normally rather robust element ratios are inverted (Diamond et al. 1974; Goldstein et al. 2003, §10.4), rendering the analytical result unreliable or even completely invalid. In addition, the use of (polished) planar specimen allows reliable quantitative assessment of phase modal content in volume percent, or of textural features (Kühnel 1995; Hart et al. 2010), similar to optical petrography and point counting. Another preparation issue is specimen coating with a conductive material to prevent charging, as most minerals in cement and aggregate are oxides and hence poor electrical conductors. Whereas SEM imaging is generally improved by coating with Au or Pd or other metals, these attenuate emitted X-rays in SEM-EDS or microprobe analysis more than a carbon coating does. Therefore, carbon coating is preferred for microchemical analysis (Goldstein et al. 2003, §11.2, chapter 15). In summary, optimal results are obtained using a polished plane specimen with negligible relief at the actual scale of observation, sputter coated with carbon. Instrument settings, use of standards, and limits of detection. Thus, for the reader to be able to comprehend how a given instrument was operated and set up to produce the chemical data as reported, authors (-s) should specify in adequate detail all of instrument make and model, vacuum conditions of the sample chamber (Torr, Pa), acceleration voltage (kV), probe beam current (pA), spot diameter (mm), acquisition time (s), EDS dead time (percentage of real time), WDS crystals (e.g., LIF, TAP, PET), all species analyzed (for minerals in concrete and aggregate typically the rock-forming elements Na, K, Ca, Mg, Mn, Fe, Al, Si, Ti, P, S, F, and Cl, normally presented as their respective oxides), for multiple detector systems also whether each element was measured on EDS or WDS, and in which order per WDS detector. Reliable quantitative analysis of minerals sensu lato requires use of internal (mineral) standards for instrument calibration. Complete sets comprising multiple certified mineral standards are available from a number of suppliers. A given element species is usually present in several mineral standards in the same set, allowing the user to select the matrix that matches best the actual sample material. Raw analytical data are pre-processed using complex algorithms, e.g., Bence-Albee, PAP, ZAF, j-r-Z (phi-rho-zee), to correct for various analytical side-effects from absorption, backscatter, fluorescence, etc. by the matrix. A detailed account on relevant effects and how to correct these in a range of materials is given in Goldstein et al. (2003, chapter 9). Which element species have been calibrated against which mineral standards, and which correction algorithm was used, should both by default be part of the specification of instrument settings and operating conditions. In addition to instrumental hardware capabilities, instrument settings and operating conditions including calibration against known standards determine lower detection limits, i.e., the lowest concentrations that can be reliably detected, usually abbreviated as “LLD” or “LoD” (NB: ‘detected,’ as different from ‘measured’). For a typical EPMA instrument properly calibrated against internal standards, LLD for Na2O (i.e., oxide) and/or F (i.e., element) is ~0.01 wt%, for heavier oxides possibly somewhat lower. In contrast, LLD for a typical SEM-EDS not using standards is rather 0.1-0.2 wt% for the same species, provided use of properly prepared polished and coated specimen. It should be noted however that measured values below ~50× LLD are generally regarded as qualitative only. Consequently, for analysis by EPMA, values >0.50 wt% are generally fit for quantitative calculations and chemical modeling, whereas SEMEDS data require concentrations >5.0 wt%. Thus, it is important to specify actual LLD values to enable correct interpretation of composition data, for ASR gel as much as for other minerals. Data processing and presentation. Regardless instrumentation, composition data are presented in the literature in many different ways. Many data are presented with as many digits or decimals as available, suggesting a level of accuracy unachievable with the actual instrumental

Deleterious Reactions of Aggregate With Alkalis in Concrete

337

hardware at hand. However, with typical analysis by EPMA or SEM-EDS, three significant digits (i.e., 0.XX to XX.X) is more realistic, even when the data output lists more. Note that in this context, ‘typical analysis’ comprises the net sum of sample/specimen preparation, instrument make and model, settings and conditions, analysis, and pre-processing using correction algorithms. Another issue is the tendency to normalize data to 100 wt%, either by the analytical instrument control software, or by post-processing in a desktop application. As an example, consider analysis of pure calcite CaCO3 (in oxide notation: CaO∙CO2) by EPMA. Its analytical sum total will be very near 56 wt% all of it representing CaO, the complementary ~44 wt% is missing as CO2 is not analyzed. The analytical sum total confirms that the analysis is reliable and representative for calcite, even when it is low and nowhere near 100 wt% (also see Lane and Dalton 1994). In contrast, normalization to 100 wt% would suggest the mineral be composed of pure CaO, i.e., lime, which is incorrect. Thus, the analytical sum total does convey a message, which is destructed upon normalization. According to the chemical compositions collated in Table 1, analytical sum totals for ‘ASR gel minerals’ range from 92-66 wt%. Even when only theoretical, an important bit of information is destroyed if their sum totals would be simply normalized to 100 wt% (also see Goldstein et al. 2003, §9.7.5 on p421; Reed 2005, §7.9 on p 125). Where possible, mineral composition data in weight percentage oxides should be recast into atoms per formula unit (apfu). Procedures for recalculation are described in detail in Papike (1987, 1988), Deer et al. (1992), and elsewhere. Recalculation of large number of analyses is easy in a spreadsheet, and a number of dedicated software applications is available. The number of oxygen atoms (a.k.a. O-base) required to complete recalculation is determined through prior identification of the mineral using optical thin section petrography (as argued for above), which at the same time allows to anticipate whether a deficit in the analytical sum total represents hydrous species, carbon dioxide, or both. While this calculation also comprises some sort of normalization, it preserves the information content of the analytical data, including a deficit sum total if present, rather than destructing. The calculated composition in apfu can now be held against published compositions, also to compare e.g., site occupancies of certain species in the mineral structure. Even when the true identity of a mineral remains unresolved by optical petrography (and hence the O-base), normalization by recasting weight percentages into apfu (i.e., in a standardized manner) offers some advantages by allowing direct comparison with earlier results. If the true O-base is unknown, then recalculation can still be standardized using 12 or 24 O, which are dividable by 1, 2, 3, 4, and 6. The calculated result will most likely not represent a true mineral composition in apfu, but it does enable comparison with other compositions. If results are unsatisfactory, then the recalculation can be repeated with another O-base. In that way, recalculation into apfu is somewhat comparable to recalculation of whole-rock chemical compositions according to the CIPW-norm as originally described by Cross et al. (1902) for igneous and volcanic rock types. Not because the calculated mineral modal content represents an accurate rendition of reality—as in many cases thin section petrography reveals a substantially different mineral assemblage—but because it offers a standard way of comparison between largely different rocks not easily achieved otherwise. Another standard recalculation recipe popular in clinker and cement production is the norm by Bogue (1929), whose results rarely resemble the modal mineral content of an actual clinker, but which nevertheless does enable straight-forward comparison between clinkers of different origin (also see Taylor 1989; Crumbie et al. 2006). Recommendations regarding error propagation in composition calculations are given in Giaramita and Day (1990). Evaporation of Na upon exposure to the electron beam. As elaborated earlier above, the crystal structure of ASR gel is rather open and composed of generally light elemental species including a substantial amount of various species of ‘water.’ The sodium present is rather loosely

338

Broekmans

bonded, for one due to its rather small ionic radius, for another due to its monovalent character. The net sum of mechanical and crystal-chemical properties renders ASR gel a challenging material to analyze reliably. In the high vacuum of the SEM-EDS or EMPA sample chamber (typically at ~10−5 Torr or lower), the gel loses water through exsiccation, in addition to water already lost by evaporation during sampling and specimen preparation under ambient conditions. The gel shrinks and cracks (further), possibly creating artificial surface topography in an otherwise planar polished specimen. Under the incident electron beam, the gel decomposes releasing its most volatile components, obviously including remaining hydrous species, but also loses Na by evaporation. The latter effect has been recognized for about half a century, see e.g., Varshneya et al. (1966), Siivola (1969), Walker and Howitt (1989), Jbara et al. (1995). The evaporation is attributed to excessive (beam) current density per unit surface in nano-Ampère per square micrometer (Morgan and London 1996, 2005). The content of weakly bonded Na in ASR gel gradually diminishes upon exposure to the electron beam. Analysis by SEM-EDS integrates the X-ray signal over time, and thus represents some sort of an average of relatively high initial Na-concentration, diluted with low Na-concentrations acquired after the Na evaporated. Thus, the overall Na content stated by SEM-EDS analysis is too low. A simple test known as ‘burn in’ by repeated analysis by EPMA of the exact same spot gives very interesting results, confirming evaporation of Na from the ASR gel. Up to 4-5 iterations, each next analysis shows a lower Na content until a stable level near the LLD is reached. When still the same spot is analyzed once more the next day, the Na content has increased slightly, which is attributed to back-migration by diffusion (Broekmans and Fernandes, unpublished data). Evaporation of Na can be circumnavigated in a fairly simple manner by optimizing instrument settings and conditions to compensate for the inevitable effect. To obtain reliable results, Na should be allocated to a WDS detector to be measured directly at the start of each analytical run. Thus, Na is measured at a stage when evaporation is still negligible, by a detector with high sensitivity specifically for Na. furthermore, probe beam diameter should be defocused to 20 mm (to spread the incident energy over a sufficiently large area), beam current should not exceed 2 nA, and counting time limited to 30-40 s. This practically limits current density per unit surface to ~6.4 pA·mm−2. If the instrument’s capabilities permit, then K, Al, and Si should be measured simultaneously, to compensate for the opposite of loss known as grow-in. If instrumental facilities are limited, then priority should be given to Na, Al, K, and Si, in that particular order. Additional elements should be subsequently measured in a separate run, with a higher current, resulting in a stronger signal and higher detector count rates. Most modern instrument control software supports adjustment of conditions during a single analytical run, allowing setup of multi-condition analysis. Katayama (2010a) uses 15 kV, 0.12 nA, beam diameter 0.4 mm as verified by the imprint left on the 15 nm thick carbon coating, with a resulting current density of ~950 pA·mm−2, i.e., about 150× greater than recommended by Morgan and London (1996). With the actual beam surface area (0.126 mm2), maximum beam current should rather have been 8 pA, which is still 100× lower than the lowest setting (0.08 nA) used in that study. Indeed, the author reports a heat-induced halo about five times larger than the beam imprint on the gel surface (Katayama 2010a, §3.2.1, p644). Even with the use of internal mineral standards and ZAF correction, the overall error range is estimated as up to 30%, which seems optimistic as all elements were measured using an EDS detector.

Deleterious Reactions of Aggregate With Alkalis in Concrete

339

Detailed further recommendations are given in Nielsen and Sigurdsson (1981), Morgan and London (1996, 2005) and Spray and Rae (1995). An alternative of measuring under cryogenic conditions within the sample chamber gives good results as well, but is technically more complicated to implement as it requires special equipment.

THE CRYSTALLINITY INDEX OF QUARTZ To prevent deleterious ASR to develop in future structures, the materials to be included in the concrete mixture are subject to prior assessment, usually comprising a cascade of procedures and tests of increasing rigorousness, to maximize reliability and preclude false results to the extent possible (e.g., CUR-Recommendation  89 2008). For concrete aggregate, assessment typically starts with petrography (e.g., RILEM AAR-1), which—if inconclusive—is followed by additional ultra-accelerated expansion testing on mortar bars and/or concrete prisms as required (e.g., RILEM AAR-2, AAR-3, AAR-4). However, expansion testing is expensive and time demanding (AAR-3 expansion testing on concrete prisms requires one full year of exposure), which is for certain projects impracticable. Another recurring argument is that petrography is subjective and operator-dependent, which is to some extent true regarding identification of lithologies and their nomenclature (Broekmans et al. 2009), thereby affecting the quantitative part of point counting to determine its modal lithological composition. A third issue is that the various properties and qualities making quartz alkali-reactive are difficult to quantify, and often a matter of discussion among operators. In summary, an operator-independent, instrumental method of adequate accuracy to determine and quantify the alkali-reactivity potential of silica sensu lato. Arguably the most popular and widely applied instrumental method to assess the ASRpotential of quartz is by determining its “crystallinity index” by X-ray powder diffraction. The method was originally described by Murata and Norman (1976), and is based on the peak-over-background ratio of the first peak and the subsequent trough in the quintuplet centered around 67.74° 2θ (non-monochromated CuKα with λ = 1.541840 Å). No arguments whatsoever were given by the authors as to why this particular quintuplet was considered most suitable. In reality, the quintuplet represents a triplet, doubled by CuKa1 and CuKa2 diffracting increasingly different towards higher angles, and one of the 3×2 peaks being overlapped by the others. Removal of CuKa2 by monochromation (McNally et al. 2004) immediately invalidates determination of QCI as originally defined, as it deconstructs the quintuplet into a triplet. The raw peak ratio is rescaled 1-10 using an instrumental correction factor obtained from the assessment of ‘clear euhedral quartz’ of supposedly unflawed (yet unverified) crystallinity, which as an internal standard is attributed the maximum index of 10. However, the quality of the clear euhedral quartz is left unverified by independent methods, and a perfect outer habit may hide a less perfect interior (e.g., Frondel 1945; Friedlaender 1951, 1952; Bambauer 1961; Bambauer et al. 1961; Rykart 1995). Thus, a QCI of 10 does not per sé represent the highest possible score for quartz in general. The method of Murata and Norman (1976) has a number of technological and procedural shortcomings, as discussed in Broekmans (2002, 2004c), and in more extensive detail in Marinoni and Broekmans (2012). The key issue is that the QCI is not attributed to any particular crystal-structural property of quartz, in contrast to the Hinckley, Kübler, or Árkai indices for clay minerals that each are linked to such property of clays (Guggenheim et al. 2002). Furthermore, the transition from amorphous silica to crystalline quartz is not sharp but gradual, from totally random to local short-range order comprising directly adjacent silica tetrahedral only, via intermediate-range order extending over a few tetrahedra away, to long-range order over large enough distances to be picked up by XRD. The QCI (or any other type of crystallinity

340

Broekmans

index) is a sum total of a number of crystal-structural aspects including crystalline perfection (degree of distortion of the crystalline structure from ideal, by chemical impurities, mechanical deformation, dislocations, or otherwise), and crystallite size/shape, and domain size/shape. Consequently, a given QCI value represents a composite of multiple variables whose individual contributions are unknown, and the same QCI value can be obtained by different combinations. Assessment of a full diffraction profile without attributing some sort of index (as opposed to a seemingly arbitrary quintuplet) is a more reliable way to assess crystallinity and its various aspects (Marinoni and Broekmans 2012). Alternative methods to determine a crystallinity index of quartz use DTA on the reversible α-β transition at 573 °C (Deutsch et al. 1989; also see Smykatz-Kloss 1970, 1972; Moore 1986), or FT-IR (Shoval et al. 1991; also compare Plyusnina 1978; Bachiorrini et al. 1986; Bachiorrini 1987). Both methods use clear euhedral quartz of unverified quality as an internal standard to rescale raw measured values. Instead, Gregg et al. (1977) use fluorite for an internal standard, and calculate their crystallinity index from peak areas and weight percentages of both quartz and fluorite in the sample as analyzed. However, none of these alternative QCI methods has become as popular as the one by XRD cf. Murata and Norman (1976), probably because of the much easier access to XRD instrumentation compared to DTA and FT-IR, and deceiving simplicity of sample preparation only requiring powdering. Despite its shortcomings, QCI by has been widely applied to assess or even predict the alkali-reactivity potential of quartz, e.g., Morino (1989), Strogen (1993), Lee and Yu (1994), Wigum (1995), Liang et al. (1997), Morino et al. (1997), Wakizaka (2000), McNally et al. (2004), McNally and Richardson (2005). Broekmans (2002) found no consistent correlation between QCI and expansion data for Norwegian mylonites (Wigum 1995), low-grade metamorphic Ohioan cherts previously investigated by Kneller (1967) and Kneller et al. (1968), and non-metamorphic Dutch cherts, together spanning most of the QCI range from 1-10. In summary, the crystallinity index for quartz as originally described by Murata and Norman (1976) is unfit as such, for a number of reasons mentioned above. The correlations that are occasionally found may be attributable to the application of QCI to a single lithology, e.g., chert (Morino 1989; Strogen 1993; McNally et al. 2004, McNally and Richardson 2005), rhyolite (Wakizaka 2000), siliceous limestone (Liang et al. 1997), quartzite (Korovkin et al. 2012) which possibly limits the variability of some crystal-structural aspects together determining QCI. With the resulting uncertainty, the QCI-XRD method as currently applied should be abandoned.

SELECTED TOPICS FOR FUTURE RESEARCH While the amount of published data on the diverse aspects of deleterious ASR in concrete structures is huge, many topics are still left open for further investigation. A number of suggestions is specified below, but the list is by no means exhaustive nor complete, and undoubtedly, new issues will emerge over time.

Reliable identification of quartz/silica properties governing alkali-reactivity Why do certain sandstone particles react deleterious in concrete, whereas directly adjacent sandstone particles of another type do not? Particle and quartz grain size, texture/fabric, main mineral (modal) content, provenance, etc. may all be the same, yet they do react very differently. In what way is the quartz in the two rock species different, and which properties govern its alkali-reactivity? Criteria beyond application of rock nomenclature are essentially lacking (Broekmans et al. 2009), and false assessment results are highly undesirable. For instance, a concrete

Deleterious Reactions of Aggregate With Alkalis in Concrete

341

structure made with falsely approved aggregate will have a shorter service life than designed, with increased cost of maintenance and repair, and could possibly even compromise public safety. On the other hand, a potentially valuable deposit could be wasted if falsely identified as unsuitable for use as concrete aggregate. Thus, reliable identification of quartz/silica is an important issue in the petrographic assessment of existing concrete and/or virgin aggregate prior to further testing and/or use. Standard optical petrography could be extended with optical cathodoluminescence (Št’astná et al. 2012), EBSD mapping to reveal lattice strain, (sub-) grain and twin boundaries, LA-ICPMS to assess detailed crystal chemistry, TEM to measure dislocation density (see e.g., Blum et al. 1990), or any other technique suitable to distinguish alkali-reactive from innocuous silica.

Extraction of alkali-reactive aggregate from field concrete Related to the above is the study of aggregate in field concrete. While plane and thin section petrography sensu lato allow assessment of selected particles, access to bulk ‘heritage material ‘from field concrete of ‘accredited reactivity’ with minimal chemical or physical artifacts would be of great help. In principle, electrical pulse disintegration (EPD) offers excellent opportunity for effective liberation of bulk aggregate from field concrete or post-mortem expansion testing specimen. The cleanliness of the liberated particles with minimal adhering relict paste enables direct comparison with virgin material from the same deposit by determining modal content of its constituent lithologies or minerals (or bulk whole-rock geochemistry) to check intra-deposit variation, for micro-mechanical assessment, or for preparation for further in situ assessment.

Dissolution of quartz/silica under ASR conditions In a way, deleterious ASR comprises the undesired dissolution of quartz under ‘ASR conditions.’ This can be investigated on prepared samples under controlled laboratory conditions, as done previously for instance on quartz dissolution under geological conditions (Dove et al. 2005), and for hydration of alite as a Portland clinker constituent (Juilland et al. 2010). One possible setup could be use of double polished thick sections and characterize the surface using a range of techniques (e.g., optical-CL, SEM, EBSD, atomic force microscopy AFM). Subsequently, the characterized surface is exposed to a solution of choice under controlled conditions and re-assessed to study any changes in surface morphology due to dissolution etching (see e.g., Leemann and Holzer 2005). These findings could then be compared with dissolution phenomena observed in actually alkali-reactive particles liberated from field concrete, or post-mortem expansion test specimen. Individual particles could be extracted by careful drilling, bulk aggregate could be liberated by EPD as described above.

Nano-structure of ASR gel There seems to be general consensus about kanemite representing a suitable model for crystalline structure of ASR gel. Reported results are based on a range of synthetic kanemites prepared in the laboratory and one single type of exuded ASR gel from Furnas Dam (Kurtis et al. 1998, 1999, 2002, 2003; Hou et al. 2005, Kirkpatrick et al. 2006, Tambelli et al. 2006, Benmore and Monteiro 2010). Only recently, another exuded ASR gel from Moxotó Dam was assessed (Meral et al. 2011), corroborating earlier results. Field ASR gel appears to consist of local domains ~10 Å in size with a kanemite structure, with longer-range order to adjacent domains poor or absent. The interstices between the domains are assumed to be able to accommodate water thus contributing to expansion, but their true nature is currently unresolved. However, exuded gels are known to have a different composition low in Ca compared to internal ASR gels (Knudsen and Thaulow 1975), and the latter type may have a different crystalline structure.

Broekmans

342

Transmission X-ray microscopy (TXM) as recently described by Brisard et al. (2012) is a most promising novel method for imaging in the nano- to micrometer region. TXM has already been successfully applied to flocculated clays, hydrated and hydrating Portland cement paste forming CSH minerals, and is even capable of tomography and 3D imaging. Resolving the currently unknown microstructure of the crystalline kanemite domains and their gelatinous superstructure of field ASR gels, both exuded and internal, could come within reach with this instrument.

Effect of lithium on ASR The inhibiting effect of Li-salts blended into the concrete mix has been demonstrated (Feng et al. 2010; Tremblay et al. 2010). Yet, certain reactive aggregates appear insensitive to Li, even when added in much higher doses than normally considered effective. Whether this can be attributed to mineral content of the aggregate, mineral composition, the spatial distribution of Li within the concrete, or to the presence on a yet unidentified constituent inactivating Li as an inhibitor, as a few possible options, is unknown. In situ analysis of Li is still not possible with EPMA, but feasible with SIMS (McRae 1995) or LA-ICP-MS (Oberti et al. 2003; Tiepolo et al. 2007).

SUMMARY AND CONCLUSIONS More than seventy years have gone since the first recognition of deleterious alkali-aggregate reaction by Stanton in 1940, and a tremendous lot of research has been committed since, a small part of which is described above, summarized in below list: 1. deleterious AAR is a worldwide problem occurring on all continents, and new countries/regions are still added to the list as additional structures are being diagnosed with the damage mechanism. The irreparable character of AAR as a concrete property inherited from its main constituents, renders it an expensive one especially for large infrastructural works or any other type of structure designed with a long service life in mind; 2. alkalis Na and K are a natural part of the concrete composition, either inherited from the raw materials used for clinkering of the Portland cement, or infiltrated from seawater or deicers or other alkali-containing chemicals, or occasionally released from the aggregate upon exposure to the concrete interior; 3. several types of deleterious AAR are known. The most common and widespread is alkali-silica reaction ASR, less abundant is the alkali-carbonate reaction ACR that may either be ‘cryptic ASR’ from finely dispersed silica in carbonate rock which is deleterious, or a true reaction with carbonate which is non-deleterious. Both ACR and cryptic-ASR may occur in a given aggregate material; 4. the most abundant silica polymorph in the supracrustal rocks used for concrete aggregate is quartz, a-SiO2. Additional alkali-reactive silica species include chert/ flint, chalcedony and opal in sedimentary rocks, in volcanic rocks also cristobalite or tridymite, that may be important constituents in certain regions. Alkali-reactivity is affected by silica grain size, accessibility (e.g., internal porosity, permeability), and a range of properties and qualities known to govern the dissolution of quartz under geological conditions; 5. the alkali-silica reaction mechanism is very complicated and affected by the type of alkali-reactive silica and pore solution chemistry. Alkalis are regenerated by subsequent reactions rather than consolidated with the reaction products, and are thus available for further reaction;

Deleterious Reactions of Aggregate With Alkalis in Concrete

343

6. alkali-silica reaction products have variable optical properties in thin section, morphology and chemical composition, related to position (inside the alkali-reactive particle, or in the paste), age, and natural variation in locally available ‘starting materials’ aggregate and cement. Inconsistent analytical setup as well as of reporting results complicates direct comparison of data from different sources; 7. the crystalline structure of ASR gel can be represented as consisting of kanemite domains of ~10 Å, without long-range order between adjacent domains. The structure of the interstitial space between domains is presently unknown, but is assumed capable of accommodating additional water contributing to gel expansion; 8. alternative alkali-reactive species comprise natural and industrial glasses, as well as a range of common rock-forming silicate minerals and carbonates (also see point 3); 9. routine assessment of AAR concrete in a laboratory involves petrography on cores extracted from a damaged structure. Rigorous procedures are essential to minimize introduction of artifacts during extraction, handling, storage and assessment; 10. cracks and internal porosity can be visualized by impregnating plane and thin sections with fluorescent epoxy. The amount of AAR damage (and its progress over time) can be quantified with a number of alternative methods, the most accurate being timerobbing and hence costly, the quick-and-dirty alternatives being less expensive but less accurate; 11. assessment of bulk deleterious aggregate in set concrete is complicated by imperfect liberation, whether using chemical or mechanical procedures; 12. in situ analysis of ASR gel in thin section is complicated by a range of factors, most of which can be compensated for by rigorous specimen preparation procedures, finetuning instrument settings and operating conditions, and by critical post-processing of acquired data; 13. the crystallinity index for quartz QCI by Murata and Norman (1976) is a meaningless concept in its current form, and should be abandoned for the assessment of alkalisilica reactivity potential of concrete aggregate.

ACKNOWLEDGMENTS The author is indebted to Robin Charlwood/USA, Silvina Marfil/AR, Peter Laugesen/DK, Francisco Locati/AR, Karen Scrivener/CH, Ian Sims/UK, Catherine Wong Phui Chan/SG, and an anonymous other for their contribution to the AAR world map. Tetsuya Katayama/JP is thanked for sharing a number of references, and Isabel Fernandes/PT is thanked for a thorough review of an early version of this manuscript, and for suggestions improving this paper. Herbert PÖllmann/DE is thanked for his encouraging words and positive spirit, as well as for valuable technical comments and discussion during preparation. Last but certainly not least: great many thanks to my wife Roberta and my daughter Gaia, for your patience and inspiration.

REFERENCES Acker JG, Bricker OP (1992) The influence of pH on biotite dissolution and alteration kinetics at low temperature. Geochim Cosmochim Acta 56:3073-3092 Almond GG, Harris RK, Franklin KR (1997) A structural consideration of kanemite, octosilicate, magadiite and kenyaite. J Mater Chem 7(4):681-687 Andres U, Bialecki R (1986) Liberation of mineral constituents by high voltage pulses. Powder Technol 48:269277 Andres U (1989) Parameters of disintegration of rock by electrical pulses. Powder Technol 58:265-269

344

Broekmans

Andres U (1995) Electrical disintegration of rock. Miner Process Extr Metall Rev 14:87-110 Annehed H, Fälth L, Lincoln FJ (1982) Crystal structure of synthetic makatite Na2Si4O8(OH)2⋅4H2O. Z Kristallogr 159:203-210 Anthony JW, Bideaux RA, Bladh KW, Nichols MC (1990-2003) Handbook of Mineralogy. Volumes I-V. Mineral Data Publishing, Tucson, Arizona www.handbookofmineralogy.org Apperley C, Hudson MJ, Keene MTJ, Knowles JA (1995) Kanemite (NaHSi2O5·3H2O) and its hydrogenexchanged form. J Mater Chem 5(4):577-582 Applin KR, Hicks BD (1987) Fibers of dumortierite in quartz. Am Mineral 72:170-172 Aranda MAG, De la Torre AG, León-Reina L (2012) Rietveld quantitative phase analysis of OPC clinkers, cements and hydration products. Rev Mineral Geochem 74:169-209 Arya C, Xu Y (1995) Effect of cement type on chloride binding and corrosion of steel in concrete. Cem Concr Res 25:893-902 ASTM C227-10 (2010) Standard test method for potential alkali reactivity of cement-aggregate combinations (mortar bar method). American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 6 pp ASTM C294-98 (1998) Standard descriptive nomenclature for constituents of natural mineral aggregates. American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 10 pp ASTM C295-03 (2003) Standard Guide for petrographic examination of aggregates for concrete. American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 8pp ASTM C441-02a (2002) Standard test method for effectiveness of pozzolans or ground blast-furnace slag in preventing excessive expansion of concrete due to the alkali-silica reaction. American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 3 pp ASTM C1260-07 (2007) Standard test method for potential alkali reactivity of aggregates (mortar bar method). American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 4 pp ASTM C1293-08b (2008) Test method for determination of length change of concrete due to alkali-silica reaction. American Society for Testing and Materials, West Conshoshocken/PA, Annual Book of ASTM Standards (04.02): Concrete and aggregates: 7 pp Azaroual M, Fouillac C, Matray JM (1997a) Solubility of silica polymorphs in electrolyte solutions.I. Activity coefficient of aqueous silica 25 °C to 250 °C, Pitzer’s parameterisation. Chem Geol 140:155-165 Azaroual M, Fouillac C, Matray JM (1997b) Solubility of silica polymorphs in electrolyte solutions. II. Activity of aqueous silica and solid silica polymorphs in deep solutions from the sedimentary Paris Basin. Chem Geol 140:167-179 Bachiorrini A, Baronio G, Berra M, Delmastro A, Montanaro L, Negro A (1986) Infrared spectroscopy in the evaluation of aggregates in ASR deteriorated concretes from many parts of the world: comparison with other methods. In: Procs 7th ICAAR, Grattan-Bellew PE (ed), Ottawa, Canada, p 309-313 Bachiorrini A (1987) A method to test the alkali reactivity of siliceous aggregates: infrared spectroscopy. In: Katharine and Bryant Mather Intl Conf Concr Durability, Scanlon JM (ed), American Concrete Institute, Spec Pub 100:1821-1832 Bachiorrini A, Montanaro L (1988) Alkali-silica reaction (ASR) in carbonate rocks. Cem Concr Res 18(5):731738 Baker AF, Poole AB (1980) Cement hydrate development at opal-cement interfaces and alkali-silica reactivity. Q J Eng Geol 13(4):249-254 Bakker JD (1999) ASR in 20 bridges in and over motorway 59 in The Netherlands. In: Procs 8th Intl Conf Structural Faults and Repair. Forde MC (ed) London, United Kingdom cd-rom:1-8. Bambauer HU (1961) Spurenelementgehalte und γ-Farbzentren in Quarzen aus Zerrkluften der Schweizer Alpen. Schweiz Mineral Petrol Mitt 41:335-369 Bambauer HU, Brunner GO, Laves F (1961) Beobachtungen über Lamellenbau an Bergkristallen. Z Kristallog 116:173-181 Barry TI, Glasser FP (2000) Calculations of Portland cement clinkering reactions. Adv Cem Res 12(1):19-28 Baumann HN Jr (1957) Preparation of petrographic sections with bonded diamond wheels. Am Mineral 42:416421 Beaudoin JJ (1983) Properties of Portland cement paste reinforced with mica flakes. Cem Concr Res 13:153-160 Beilby GT (1921) Aggregation and Flow of Solids. Being the records of an experimental study of the microstructure and physical properties of solids in various states of aggregation, 1900-1921. MacMillan and Co, London Beneke K, Lagaly G (1977) Kanemite-inner crystalline reactivity and relations to other sodium silicates. Am Mineral 62:763-771

Deleterious Reactions of Aggregate With Alkalis in Concrete

345

Beneke K, Lagaly G (1983) Kenyaite - synthesis and properties. Am Mineral 68:818-826 Benmore CJ, Monteiro PJM (2010) The structure of alkali silicate gel by total scattering methods. Cem Concr Res 40:892-897 Bennett PC, Melcer ME, Siegel DI, Hassett JP (1988) The dissolution of quartz in dilute aqueous solutions of organic acids at 25°C. Geochim Cosmochim Acta 52:1521-1530 Bennett PC (1991) Quartz dissolution in organic-rich aqueous systems. Geochim Cosmochim Acta 55:17811797 Berger G, Cadore F, Schott J, Dove PM (1994) Dissolution rates of quartz in lead and sodium electrolyte solutions between 25°C and 300°C: effect of the nature of surface complexes and reaction affinity. Geochim Cosmochim Acta 52:541-551 Bertram MA, Mackenzie FT, Bishop FC, Bischoff WD (1991) Influence of temperature on the stability of magnesian calcite. Am Mineral 76:1889-1896 Bérubé MA, Fournier B (1986) Les produits de la réaction alcalis-silice dans le béton: étude de cas de la région de Québec. Can Mineral 24:271-288 Bérubé MA, Locat J, Choquette M, Fournier B (1988) Beneficial lime stabilization of clay soils and deleterious alkali-aggregate reactions in concrete: Two phenomena related to the chemical stability of minerals at high pH and in presence of Ca(OH2). In: Procs 41st Conf, Can Geotech Soc, Waterloo/Kitchener, Canada p 150-159 Bérubé MA, Choquette M, Locat J (1990) Effects of lime on common soil and rock forming minerals. Appl Clay Sci 5:145-163 Bérubé MA, Fournier B, Durand B (eds) (2000) Procs 11th ICAAR, Québec City, QC, Canada: pp 1406 Bérubé MA, Dorion JF (2000) Laboratory and field investigations of the influence of sodium chloride on alkalisilica reactivity, alkali-aggregate reaction in concrete. In: Procs 11th ICAAR, Bérubé MA, Fournier B, Durand B (eds), Québec City, QC, Canada:149-158 Bérubé MA, Duchesne J, Dorion JF, Rivest M (2002) Laboratory assessment of alkali contribution by aggregates to concrete and application to concrete surfaces affected by alkali-silica reactivity. Cem Concr Res 32:1215-1227 Bérubé MA, Fournier B (2004) Alkalis releasable by aggregates in concrete: significance and test methods. In: Procs 12th ICAAR, Tang M, Deng M (eds), Beijing, China p 17-30 Bhatty MSY (1985) Mechanism of pozzolanic reactions and control of alkali-aggregate expansion. Cem Concr Aggs CCAGDP 7:69-77 Bialecki R, Choi P, Andres U, Shaw CT (1992) Disintegration of rock by high voltage pulse discharge. In: Procs 8th IEEE Intl Pulsed Power Conf, Prestwich K, White R (eds), San Diego, California, p 838-841 Bickmore BR, Wheeler JC, Bates B, Nagy KL, Eggett DL (2008) Reaction pathways for quartz dissolution determined by statistical and graphical analysis of macroscopic experimental data. Geochim Cosmochim Acta 74:4521-4536 Bignozzi MC, Saccani, A (2008) Expansion behavior of ground glass cullet used as aggregate or cement replacement. In: Procs 13th ICAAR, Broekmans MATM, Wigum BJ (eds), Trondheim, Norway, p 11211130 Birnin-Yauri UA, Glasser FP (1998) Friedel’s salt, Ca2Al(OH)6(Cl,OH)⋅2H2O: its solid solutions and their role in chloride binding. Cem Concr Res 28(12):1713-1723 Bish DL, Reynolds RC Jr (1989) Sample preparation for X-ray diffraction. Rev Mineral 20:73-99 Bish DL, Plötze M (2010) X-ray powder diffraction with emphasis on qualitative and quantitative analysis in industrial mineralogy. Christidis GE (eds), EMU Notes Mineral 9:35-76 Bjørkum PA (1996) How important is pressure in causing dissolution of quartz in sandstones? J Sediment Res 66:147-154 Bluhm H, Frey W, Giese H, Hoppé P, Schultheiß C, Sträßner R (2000) Application of pulsed HV discharges to material fragmentation and recycling. IEEE Trans Dielectr Electr Insul 7:625-636 Blum AE, Yund RE, Lasaga AC (1990) The effect of dislocation density on the dissolution rate of quartz. Geochim Cosmochim Acta 54:283-297 Bogue RH (1929) Calculation of the compounds in Portland cement. Ind Eng Chem Anal 1(4):192-197 Boles JR, Johnson KS (1983) Influence of mica surfaces on pore-water pH. Chem Geol 43:303-317 Bonen D, Diamond S (1994) Interpretation of compositional patterns found by quantitative energy dispersive X-ray analysis for cement paste constituents. J Am Ceram Soc 77:1875-1882 Borbély G, Beyer HK, Karge HG, Schwieger W, Brandt AH, Bergk KH (1991) Chemical characterization, structural features, and thermal behavior of sodium and hydrogen octosilicate. Clays Clay Miner 39(5):490-497 Bowden FP, Hughes TP (1937) Physical properties of surfaces IV - Polishing, surface flow and the formation of the Beilby layer. Proc Roy Soc 160A:575-587 Brantley SL (1992) The effect of fluid chemistry on quartz microcrack lifetimes. Earth Planet Sci Lett 113:145156

346

Broekmans

Brindley GW (1969) Unit cell of magadiite in air, in vacuo, and under other conditions. Am Mineral 54:15831591 Brisard S, Chae RS, Bihannic I, Michot L, Guttmann P, Thieme J, Schneider G, Monteiro PJM, Levitz P (2012) Morphological quantification of hierarchical geomaterials by X-ray nano-CT bridges the gap from nano to micro length scales. Am Mineral 97:480-483 Broekmans MATM, Jansen JBH (1998) Silica dissolution in impure sandstone: application to concrete. J Geochem Explor 62:311-318 Broekmans MATM, Nijland TG, Jansen JBH (2001) Origin and mobility of alkalies in two Dutch ASRconcretes II: microscale element distribution around sandstone and chert. Implications for the mechanism of ASR. In: Procs 8th EMABM, Stamatakis M, Georgali B, Fragoulis D, Toumbakari EE (eds), Athens, Greece, p 85-92 Broekmans MATM (2002) The alkali-silica reaction: mineralogical and geochemical aspects of some Dutch concretes and Norwegian mylonites. PhD-thesis Utrecht University, The Netherlands. Geologica Ultraiectina 217:1-144 Broekmans MATM (2004a) Structural properties of quartz and their potential role for ASR. Mater Charact 53(2-4):129-140 Broekmans MATM (2004b) Qualities of quartz and the alkali-silica reaction in concrete. In: Procs 8th Intl Cong Appl Min, Pecchio M, Andrade FRD, D’Agostino LZ, Kahn H, Sant’Agostino LM, Tassinari MMML (eds), Aguas de Lindoia, Brazil, p 661-664 Broekmans MATM (2004c) The crystallinity index of quartz by XRD, its susceptibility for ASR, and a brief methodological review. In: Procs 12th ICAAR, Tang M, Deng M (eds), Beijing, China, p 60-68 Broekmans MATM (2006) Sample representativity: effects of size and preparation on geochemical analysis. In: Marc-André Bérubé Symposium on Alkali-Aggregate Reactivity in Concrete. Fournier B (ed), Montréal, Canada p 1-19 Broekmans MATM, Wigum BJ (eds) (2008) Procs 13th ICAAR, Trondheim, Norway: pp1336 + xvii. Broekmans MATM (2009) Petrography as an essential complementary method in forensic assessment of concrete deterioration: two case studies. Mater Charact 60(7):644-655 Broekmans MATM, Fernandes I, Nixon P (2009) A global petrographic atlas of alkali-silica reactive rock types: a brief review. In: Procs 12th EMABM. Middendorf B, Just A, Klein D, Glaubitt A, Simon J (eds) Dortmund, Germany, p 39-50 Brouxel M (1993) The alkali-aggregate reaction rim: Na2O, SiO2, K2O and CaO chemical distribution. Cem Concr Res 23:309-320 BS 7943-99 (1999) Guide to the interpretation of petrographical examinations for alkali-silica reactions. British Standard BSI 07. British Standard Institution, London, United Kingdom 20 p Buhrke VE, Jenkins R, Smith DK (eds) (2001) A Practical Guide for the Preparation of Specimens for X-ray Fluorescence and X-ray Diffraction Analysis. Wiley-VCH, New York Bulteel D, Garcia-Diaz E, Dégrugilliers P (2010) Influence of lithium hydroxide on alkali-silica reaction. Cem Concr Res 40(4):526-530 Burzo E (2009) 8.1.5.15 Kanemite and ekanite groups and related silicates. In: The Landolt-Börnstein Database, Wijn HPJ (ed), Springer Verlag, Berlin/Heidelberg: p 57 Cabri LJ, Rudashevsky NS, Rudashevsky VN, Oberthür T (2008) Electric-pulse disaggregation (EPD), hydroseparation (HS) and their use in combination for mineral processing and advanced characterization of ores. Canadian Mineral Processors 40th Annual Meeting, Proceedings 14:211-235 Carroll D, Starkey HC (1971) Reactivity of clay minerals with acids and alkalies. Clays Clay Miner 19:321-333 Casey WH, Bunker B (1990) Leaching of mineral and glass surfaces during dissolution. Rev Mineral 23:397426 Casey WH, Westrich HR, Banfield JF, Ferruzzi G, Arnold GW (1993) Leaching and reconstruction at the surfaces of dissolving chain-silicate minerals. Nature 366:253-256 Cembureau (2012) Construction in Europe Main activities 2010. Billion euro: 1312. www.cembureau.be/sites/ default/files/Construction2010.pdf, retrieved dd. 26 January 2012 Černý P, Burt DM (1984) Paragenesis, crystallochemical characteristics, and geochemical evolution of micas in granitic pegmatites. Rev Mineral 13:257-297 Chapman DL (1913) A contribution to the theory of electro capillarity. Philos Mag 25:475-481 Charlwood RG, Solymar ZV (1995) Long-term management of AAR affected structures: an international perspective. In: Procs 2nd Intl Conf on AAR in Hydroelectric Plants and Dams, Chattanooga, Tennessee. US Committee on Large Dams (USCOLD), Denver, Colorado, p 19-55 Chatterji S, Jensen AD, Thaulow N, Christensen P (1986) Studies of alkali-silica reaction. Part 3. Mechanisms by which NaCl and Ca(OH)2 affect the reaction. Cem Concr Res 16:246-254 Chatterji S, Thaulow N, Jensen AD (1987a) Studies of alkali-silica reaction. Part 4. Effect of different alkali salt solutions on expansion. Cem Concr Res 17:777-783

Deleterious Reactions of Aggregate With Alkalis in Concrete

347

Chatterji S, Thaulow N, Jensen AD (1988) Studies of alkali-silica reaction. Part 6. Practical implications of a proposed reaction mechanism. Cem Concr Res 18:363-366 Chatterji S, Thaulow N, Jensen AD (1989) Studies of alkali-silica reaction. Part 5. Verification of a newly proposed reaction mechanism. Cem Concr Res 19:177-183 Chatterji S, Thaulow N, Jensen AD, Christensen P (1987b) Mechanism of accelerating effects of NaCl and Ca(OH)2 on alkali–silica reaction. In: Procs 7th ICAAR, Grattan-Bellew PE (ed), Ottawa, Canada, p 115119 Chemtob SM, Rossman GR, Stebbins JF (2012) natural hydrous amorphous silica: quantitation of network speciation and hydroxyl content by 29Si MAS NMR and vibrational spectroscopy. Am Mineral 97:203-211 Chen CTA, Marshall WL (1982) Amorphous silica solubilities. IV. Behavior in pure water and aqueous sodium chloride, sodium sulphate, magnesium chloride, and magnesium sulphate solutions up to 350°C. Geochim Cosmochim Acta 46:279-287 Chernet T (2012) High voltage selective fragmentation for detailed mineralogical and analytical information, case study: Oiva’s gold-quartz-dyke in the lapland granulite belt, Laanilia, northern Finland. In: Procs 10th ICAM. Broekmans MATM (ed) Springer Verlag, Heidelberg, p 119-127 Choquette M, Bérubé MA, Locat J (1987) Mineralogical and microtextural changes associated with lime stabilization of marine clays from Eastern Canada. Appl Clay Sci 2:215-232 Choquette M, Bérubé MA, Locat J (1991) Behavior of common rock-forming minerals in a strongly basic NaOH solution. Can Mineral 29:163-173 Claeys PF, Mount JF (1991) Diagenetic origin of carbonate, sulfide and oxide inclusions in biotites of the Great Valley Group (Cretaceous), Sacramento Valley, California. J Sediment Res 61:719-731 Cole GAJ (1891) Devitrification of cracked and brecciated obsidian. Mineral Mag 9:272-274 Cole WF, Lancucki CJ, Sandy MJ (1981) Products formed in an aged concrete. Cem Concr Res 11:443-454 Cole WF, Lancucki CJ (1983) Products formed in an aged concrete. The occurrence of okenite. Cem Concr Res 13:611-618 Cong XD, Kirkpatrick RJ, Diamond S (1993) 29Si MAS NMR spectroscopy investigation of alkali-silica reaction product gels. Cem Concr Res 23:811-823 Constantiner D (1994) Factors affecting concrete pore solution: A) alkali fixation by drying and carbonation and (B) alkali released from feldspars. 2nd edition. PhD Thesis, Purdue University, Lafayette, Indiana Constantiner D, Diamond S (2003) Alkali release from feldspars into pore solutions. Cem Concr Res 33:549554 Çopuroğlu O, Andiç-Çakır Ö, Broekmans MATM, Kühnel RA (2009) Mineralogy, geochemistry and expansion testing of an alkali-reactive basalt from Western Anatolia, Turkey. Mater Charact 60/7:756-766 Coradin T, Lopez PJ (2003) Biogenic silica patterning: simple chemistry or subtle biology? Chem Bio Chem 3:1-9 Cross W, Iddings JP, Pirsson LV, Washington HS (1902) A quantitative chemico-mineralogical classification and nomenclature of igneous rocks. J Geol 10:555-690 Crowcroft PJ (1981) Demise of the Beilby-Bowden theory of polishing. J Gemology 17(7):459-465 Crowley MS, Roy R (1964) Crystalline solubility in the muscovite and phlogopite groups. Am Mineral 49:348362 Crumbie A, Walenta G, Füllmann T (2006) Where is the iron? Clinker microanalysis with XRD Rietveld, optical microscopy/point counting, Bogue and SEM-EDS techniques. Cem Concr Res 36:1542-1547 CSA A23.2-09 (2009a) Standard 26A – Determination of potential alkali-carbonate reactivity of quarried carbonate rocks by chemical composition. Test methods and standard practices for concrete. Canadian Standards Association, Mississauga, Ontario: 366-370 CSA A23.2-09 (2009b) Standard 15A – Petrographic examination of aggregates. Test methods and standard practices for concrete. Canadian Standards Association, Mississauga, Ontario: 297-328 Curie J (1889) Recherches sur la conductibilité des corps cristallisés. Ann Chim Phys 16(18): 321-325 CUR-Recommendation 89 (2008) Measures to Prevent Concrete Damage by the Alkali-Silica Reaction. Official English translation, 2nd revised edition. Centre for Civil Engineering Research and Codes, Gouda, The Netherlands p48 CUR-Recommendation 102 (2008) Inspection and assessment of concrete structures in which ASR is suspected or has been confirmed. Official English translation. Centre for Civil Engineering Research and Codes, Gouda, The Netherlands, p 31 Curtil L, Gielly J, Murat M (1992) Chemical behaviour of natural opal and silica in basic solutions. In: Procs 9th ICAAR, Poole AB (ed), London, United Kingdom, p 217-223 Damgaard Jensen A, Chatterji S, Christensen P, Thaulow N, Gudmundsson H (1982) Studies of alkali-silica reaction. Part 1. A comparison of two accelerated test methods. Cem Concr Res 12:641-647 Damgaard Jensen A, Chatterji S, Christensen P, Thaulow N (1984): Studies of alkali-silica reaction. Part 2. Effect of air-entrainment on expansion. Cem Concr Res (14):311-314

348

Broekmans

Danish Standard DS  423.39 (2002) Concrete testing - hardened concrete - production of fluorescenceimpregnated plane sections. (original Danish title: Betonprøvning - Hærdnet beton - Fremstilling af fluorescensimprægnerede planslib.) DS 423.39, p 8 Danish Standard DS  423.40 (2002) Concrete testing - hardened concrete - production of fluorescenceimpregnated thin sections. (original Danish title: Betonprøvning - Hærdnet beton - Fremstilling af fluorescensimprægnerede tyndslib.) DS 423.40, p2 Davies G, Oberholster RE (1986a) Chemical and swell properties of the alkali-silica reaction product. In: Procs 8th Intl Congr Chem Cem, Okada K, Nishibayashi S, Kawamura M (eds), Rio de Janeiro, Brazil, p 249-255 Davies G, Oberholster RE (1986b): The alkali-silica reaction product, a mineralogical and electron microscopic study. In: Procs 8th Intl Conf Cem Microscopy, Orlando, Florida, p 303-326 Davies G, Oberholster RE (1988): Alkali-silica reaction products and their development. Cem Concr Res 18:621-635 Davis MC, Wesolowski DJ, Rosenqvist J, Brantley SL, Mueller KT (2011) Solubility and near-equilibrium dissolution rates of quartz in dilute NaCl solutions at 398-473K under alkaline conditions. Geochim Cosmochim Acta 75:401-415 De Haas GJLM, Nijland TG, Valbracht PJ, Maijer C, Verschure R, Andersen T (2002) Magmatic versus metamorphic origin of olivine-plagioclase coronas. Contrib Mineral Petrol 143:537-550 Deer WA, Howie RA, Zussman J (1992) An Introduction to the Rock Forming Minerals. 2nd excerpted student edition. Longman, Burnt Mill, United Kingdom Demoulian E, Vernet C, Hawthorn F, Gourdin P (1980) Determination de la teneur en laitier dans les ciments par dissolutions sélectives. In: Procs 7th Intl Congr Chem Cements, Paris, France 2:111-151 Den Uijl JA, Kaptijn N, Walraven JC (2000) Shear resistance of flat slab bridges affected by ASR. In: Procs 11th ICAAR, Bérubé MA, Fournier B, Durand B (eds), Québec City, QC, Canada, p 1129-1138 Deng M, Han SF, Lu YN, Lan XH, Hu YL, Tang M (1993) Deterioration of concrete structures due to alkalidolomite reaction in China. Cem Concr Res 23:1040-1046 Detwiler RJ, Powers LJ, Jacobsen UH, Ahmed WU, Scrivener KL, Kjellsen KO (2001) Preparing specimens for microscopy. Concr Int 23(11):51-58 Deutsch Y, Sandler A, Nathan Y (1989) The high-low inversion of quartz. Thermochim Acta 148:467-472 Diamond S, White JL, Dolch WL (1964) Transformation of clay minerals by calcium hydroxide attack. In: Procs 12th Natl Conf Clays and Clay Mins, Ingerson E (ed), Atlanta, Georgia, p 359-379 Diamond S (1972) Identification of hydrated cement constituents using a scanning electron microscope energy dispersive X-ray spectrometer combination. Cem Concr Res 2:617-632 Diamond S, Thaulow N (1974) A study of expansion due to alkali-silica reaction as conditioned by the grain size of the reactive aggregate. Cem Concr Res 4:591-607 Diamond S, Young JF, Lawrence FV Jr (1974): Scanning electron microscopy - energy dispersive x-ray analysis of cement constituents - some cautions. Cem Concr Res 4:899-914 Diamond S (1976a) A review of the alkali-silica reaction and expansion mechanisms. 1. Alkalies in cements and in concrete pore solutions. Cem Concr Res 6:329- 346 Diamond S (1976b) A review of the alkali-silica reaction and expansion mechanisms. 2. Reactive aggregate. Cem Concr Res 6:549-560 Diamond S (1978) Beltane opal, its alkali reaction product, and some synthetic alkali-silica gels - a brief look at micromorphology. In: Procs 4th ICAAR, Diamond S (ed), West Lafayette, Indiana, p 181-197 Diamond S (1989) ASR - another look at mechanisms. In: Procs 8th ICAAR, Okada K, Nishibayashi S, Kawamura M (eds), Kyoto, Japan, p 83-94 Diamond S (1992) Alkali aggregate reactions in concrete: an annotated bibliography, 1939-1991. Strat Hwy Res Prog Rep SHRP-C/UWP-92-601, p 1-465 Diamond S (1997) Alkali-silica reactions – some paradoxes. Cem Concr Comp 19:391-401 Diamond S (2000) Chemistry and other characteristics of ASR gel. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec, Canada, p 31-40 Diamond S, Kotwica L, Olek J, Rangaraju PR, Lovell J (2006) Chemical aspects of severe ASR induced by potassium acetate airfield pavement de-icer solution. In: Marc-André Bérubé Symposium on AlkaliAggregate Reactivity in Concrete. Fournier B (ed) Montréal, Canada, p 261-277 Dietzel M (2000) Dissolution of silicates and the stability of polysilicic acid. Geochim Cosmochim Acta 64:3275-3281 Dove PM (1994) The dissolution kinetics of quartz in sodium chloride solutions at 25°C to 300°C. Am J Sci 294:665-712 Dove PM, Rimstidt JD (1994) Silica-water interactions. Rev Mineral 29:259-308 Dove PM (1995) Kinetic and thermodynamic controls on silica reactivity in weathering environments. Rev Mineral 31:235-290 Dove PM (1999) The dissolution kinetics of quartz in aqueous mixed solutions. Geochim Cosmochim Acta 63:3715-3727

Deleterious Reactions of Aggregate With Alkalis in Concrete

349

Dove PM, Han N, De Yoreo JJ (2005) Mechanisms of classical crystal growth theory explain quartz and silicate dissolution behavior. Procs Nat Acad Sci 102:15357-15362 Duchesne J, Bérubé MA (1996) Effect of de-icing salt and seawater on ASR: new considerations based on experimental data. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 830-837 Dunbar PA, Mukherjee PK, Bleszynski R, Thomas MDA (1996) A comparison of damage rating index with long-term expansion of concrete prisms due to alkali-silica reaction. In: Procs 10th ICAAR, Shayan A (ed) Melbourne, Australia, p 324-331 Duncan MAG, Swenson EG, Gillott JE (1973) Alkali-aggregate reaction in Nova Scotia III. Laboratory studies of volume change. Cem Concr Res 3:233-245 Eglinton MS, Higginbottom IE, Pettifer GS (1994) An unfamiliar alkali-aggregate reaction from Trinidad. Q J Eng Geol 27:363-373 Elsen J, Martens G, Snellings R (2010) Portland cement and other calcareous hydraulic binders: history, production and mineralogy. In: Advances in the Characterization of Industrial Minerals. Christidis GE (ed) EMU Notes Mineral 9:441-479 EN 932-3 (2010) Tests for general properties of aggregates - Part 3: Procedure and terminology for simplified petrographic description. CEN, Brussels: 14 pp Ettler V, Mihaljevič M, Touray JC, Piantone P (2002) Leaching of polished sections: an integrated approach for studying the liberation of heavy metals from lead-zinc metallurgical slags. Bull Soc Géol France 173(2):161-169 Eugster HP (1967) Hydrous sodium silicates from Lake Magadi, Kenya: precursors of bedded chert. Science 157:1177-1180 Eypert-Blaison C, Humbert B, Michot LJ (2001) Structural role of hydration water in Na- and H-magadiite: a spectroscopic study. Chem Mater 13:4439-4446 Eypert-Blaison C, Villiéras F, Michot LJ, Pelletier M, Humbert B, Ghanbaja J, Yvon J (2002) Surface heterogeneity of kanemite, magadiite and kenyaite: a high-resolution gas adsorption study. Clay Mineral 37:531-542 Feng X, Thomas MDA, Bremner TW, Folliard KJ, Fournier B (2010) Summary of research on the effect of LiNO3 on alkali–silica reaction in new concrete. Cem Concr Res 40(4): 636-642 Fernandes MIG (2005) Caracterização petrográfica, química e física de agregados graníticos em betões. Estudo de casos de obra. PhD-thesis Universidade do Porto, Portugal:334p + Appendices Fernandes I, Noronha F, Teles M (2007) Examination of the concrete from an old Portuguese dam: texture and composition of alkali-silica gel. Mater Charact 58:1160-1170 Fernandes I (2009) Composition of alkali-silica reaction products at different locations within concrete structures. Mater Charact 60:655-668 Ferrell RE Jr, Grim RE (1967) The influence of alkaline solutions on the alteration of the clay minerals. In: Procs 15th Natl Conf Clays and Clay Minerals. Bailey SW (ed) Pittsburgh, Pennsylvania, p 293-304 Figg JW, Bowden SR (1971) The Analysis of Concretes. Build Res Estb, Her Majesty’s Stationery Office: pp 68+folding plate Figg JW (1981) Reaction between cement and artificial glass in concrete. In: Procs 5th ICAAR, Oberholster RE (ed) Cape Town, South Africa, p S252/7 Fiore S, Huertas FJ, Tazaki K, Huertas F, Linares J (1999) A low temperature alteration of a rhyolitic obsidian. Eur J Mineral 11:455-469 Fischer R, Kuzel HJ, Schellhorn H (1980) Hydrocalumit: Mischkristalle von ‘Friedelschem Salz’ 3CaO⋅Al2O3⋅CaCl2⋅10H2O und Tetracalciumaluminat-hydrat 3CaO⋅Al2O3⋅CaOH2⋅12H2O? Neues Jahrb Mineral Monatsh 1980(7):322-334 Fleming BA (1986) Kinetics of reaction between silicic acid and amorphous silica surfaces in NaCl solutions. J Colloid Interface Sci 110:40-64 Fletcher RA, Bibby DM (1987) Synthesis of kenyaite and magadiite in the presence of various anions. Clays Clay Miner 35:318-320 Flörke OW, Flörke U, Giese U (1984) Moganite. A new microcrystalline silica mineral. Neues Jahrb Mineral Abh 149:325-336 Folk RL, Pittman JS (1971) Length-slow chalcedony: a new testament for vanished evaporites. J Sediment Petrol 41:1045-1058 Folliard KJ, Ideker JH, Drimalas T (2006) University of Texas at Austin (UTX), personal communication Fournier RO, Rowe JJ (1977) The solubility of amorphous silica in water at high temperatures and high pressures. Am Mineral 62:1052-1056 French WJ, Poole AB (1974) Deleterious reactions between dolomites from Bahrain and cement paste. Cem Concr Res 4:925-937 French WJ, Poole AB (1976) Alkali-aggregate reactions and the Middle East. Concrete 10(1):18-20 Friedlaender C (1951) Untersuchung über die Eignung alpiner Quarze für piezoelektrische Zwecke. Beiträge zur Geologie der Schweiz, Geotechnische Serie, Lieferung 29:1-98

350

Broekmans

Friedlaender C (1952) Structural imperfections in Alpine quartz crystals. Geol Mag 89:217-220 Friedman GM (1959) Identification of carbonate minerals by staining methods. J Sediment Petrrol 29:87-97 Frondel C (1945) Secondary Dauphiné twinning in quartz. Am Mineral 30:447-461 Frondel C (1978) Characters of quartz fibers. Am Mineral 63:17-27 Frondel C (1979) Crystalline silica hydrates from leached silicates. Am Mineral 64:799-804 Frondel C (1982) Structural hydroxyl in chalcedony (type B quartz). Am Mineral 67:1248-1257 Frondel C (1985) Systematic compositional zoning in the quartz fibers of agates. Am Mineral 70:975-979 Ftikos C, Parissakis G (1987) A study on the effect of some ions contained in seawater on hydrated cement compounds. In: Katharine and Bryant Mather Intl Conf Concr Durability, Scanlon JM (ed), American Concrete Institute, Spec Pub 100:1651-1665 Gaboriaud F, Nonat A, Chaumont D, Craievich A (1999) Aggregation and gel formation in basic silico-calcoalkaline solutions studied: a SAXS, SANS, and ELS study. J Phys Chem B 103/28:5775-5781 Galí S, Ayora C, Alfonso P, Tauler E, Labrador M (2001) Kinetics of dolomite-portlandite reaction: application to portland cement concrete. Cem Concr Res 31:933-939 Garcés JM, Rocke SC, Crowder CE, Hasha DL (1988) Hypothetical structures of magadiite and sodium octosilicate and structural relationships between the layered alkali metal silicates and the mordenite- and pentasil-group zeolites. Clays and Clay Mins 36/5:409-418 García E, Alfonso P, Labrador M, Galí S (2003a) Dedolomitization in different alkaline media: application to Portland cement paste. Cem Concr Res 33:1443-1448 García E, Alfonso P, Tauler E, Galí S (2003b) Surface alteration of dolomite in dedolomitization in alkaline media. Cem Concr Res 33:1449-1456 Garcia-Vallès M, Gimeno-Torrente D, Martínez-Manent S, Fernández-Turiel JL (2003) Medieval stained glass in a Mediterranean climate: typology, weathering and glass decay, and associated biomineralization processes and products. Am Mineral 88:1996-2006 Gartner E (2004) Industrially interesting approaches to “low-CO2 cements.” Cem Concr Res 34:1489-1498 Garvie LAJ, Devouard B, Groy TL, Cámara F, Buseck PR (1999) Crystal structure of kanemite, NaHSi2O5·3H2O, from the Aris phonolite, Namibia. Am Mineral 84:1170-1175 Gavrilenko E, García del Amo, D, Calvo Pérez, B, García García, E (2007): Comparison of ASR-gels in concretes against accelerated mortar bar test samples. Mag Concr Res (59):483-494 Giaramita MJ, Day HW (1990) Error propagation in calculations of structural formulas. Am Mineral 75:170-182 Giebson C, Seyfarth K, Stark J (2010) Influence of acetate and formate-based deicers on ASR in airfield concrete pavements. Cem Concr Res 40(4):537-545 Gifford PM, Gillott JE (1996) Alkali-silica reaction (ASR) and the alkali-carbonate reaction (ACR) in activated blast furnace slag cement (ABFSC) concrete. Cem Concr Res 26:21-26 Gillott JE, Swenson EG (1969) Mechanism of the alkali-carbonate reaction. Q J Eng Geol 2:7-23 Gillott JE, Rogers CA (1994) Alkali-aggregate reaction and internal release of alkalis. Magn Conc Res 46:99112 Gillott JE, Rogers CA (2003) The behavior of silicocarbonatite aggregates from the Montreal area. Cem Concr Res 33:471-480 Gimeno D (2003) Devitrification of natural rhyolitic obsidian glasses: petrographic and microstructural study (SEM+EDS) of recent (Lipari island) and ancient (Sarrabus, SE Sardinia) samples. J Non-Cryst Solids 323:84-90 Giovambattista A, Batic OR, Traversa LP (1986) Reactivity of alkalis and sandstones cemented with opal and chalcedony. In: Procs 7th ICAAR. Grattan-Bellew PE (ed) Ottawa, Canada, p 408-412 Gíslason SR, Heaney PJ, Veblen DR, Livi KJT (1993) The difference between the solubility of quartz and chalcedony: the cause? Chem Geol 107:363-366 Gíslason SR, Heaney PJ, Oelkers EH, Schott J (1997) Kinetic and thermodynamic properties of moganite, a novel silica polymorph. Geochim Cosmochim Acta 61:1193-1204 Goguel R (1995) Alkali release by volcanic aggregates in concrete. Cem Concr Res 25:841-852 Goguel, R (1996) Selective dissolution techniques in AAR investigation: application to an example of failed concrete. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 783-790 Goguel R, Milestone N (2000) Alkali release from basalts and alkalinity of pore solution in mortar. Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 179-188 Goldstein J, Newbury D, Joy D, Lyman C, Echlin P, Lifshin E, Sawyer L, Michael J (2003) Scanning Electron Microscopy and X-ray Microanalysis. A Text for Biologists, Materials Scientists, and Geologists. 3rd Edition. Springer Verlag, Berlin Götze J, Nasdala L, Kleeberg R, Wentzel M (1998) Occurrence and distribution of “moganite” in agate/ chalcedony: a combined micro-Raman, Rietveld and cathodoluminescence study. Contrib Mineral Petrol 133:96-105

Deleterious Reactions of Aggregate With Alkalis in Concrete

351

Götze J, Plötze M, Graupner T, Hallbauer DK, Bray CJ (2004) Trace element incorporation into quartz: a combined study by ICP-MS, electron spin resonance, cathodoluminescence, capillary ion analysis, and gas chromatography. Geochim Cosmochim Acta 68:3741-3759 Gouy G (1909) Sur la constitution de la charge électrique à la surface d’un électrolyte. Compt Rend Acad Sci 149:654-657 Graetsch H (1994) Structural characteristics of opaline and microcrystalline silica minerals. Rev Mineral 29:209-232 Graetsch HA, Grünberg JM (2012) Microstructure of flint and other chert raw materials. Archaeometry 54(1):1836 Grattan-Bellew PE, Beaudoin JJ (1980) Effect of phlogopite mica on alkali-aggregate expansion in concrete. Cem Concr Res 10:789-797 Grattan-Bellew PE (1981a) A review of test methods for alkali-expansivity of concrete aggregates. In: Procs 5th ICAAR. Oberholster RE (ed) Cape Town, South Africa, p S252/9 Grattan-Bellew PE (1981b) Evaluation of miniature rock prism test for determining the potential alkaliexpansivity of aggregates. Cem Concr Res 11:699-711 Grattan-Bellew PE, Danay A (1992) Comparison of laboratory and field evaluation of AAR in large dams. In: Proc Intl Conf Conc AAR in Hydroelectric Plants and Dams, Canadian Electrical Association (CEA), Montréal, in association with Can Natl ICOLD Committee, Fredericton, Canada, p 1-23 Grattan-Bellew PE (1995) Laboratory evaluation of alkali-silica reaction in concrete from Saunders Generating Station. ACI Mater J 92(2):126-134 Grattan-Bellew PE, Rogers CA (2000) Expansion due to alkali-carbonate reaction: laboratory prognosis versus field experience. In: Procs11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 41-50 Grattan-Bellew P, Mitchell LB (2006) Quantitative petrographic analysis of concrete - the damage rating index (DRI) method, a review. In: Marc-André Bérubé Symposium on Alkali-Aggregate Reactivity in Concrete. Fournier B (ed) Montréal, Canada, p 321-334 Grattan-Bellew PE, Margeson J, Mitchell LD, Deng M (2008) Is ACR just another variant of ASR? Comparison of acid insoluble residues of alkali-silica and alkali-carbonate reactive limestones and its significance for the ASR/ACR debate. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 685-695 Grattan-Bellew PE, Mitchell LD, Margeson J, Deng M (2010) Is alkali-carbonate reaction just a variant of alkali-silica reaction ACR = ASR? Cem Concr Res 40(4):556-562 Gregg JM, Goldstein ST, Walters Jr LJ (1977) Occurrence of strained quartz in the siliceous frustules of cultured freshwater diatoms. J Sediment Petrrol 47(4):1623-1629 Grice JD, Ferraris G (2000) New minerals approved in 1999 by the Commission on New Minerals and Mineral Names, International Mineralogical Association. [99-035: moganite.] Can Mineral 38:245-250 Grim RE (1953) Clay Mineralogy. McGraw-Hill Book Company Inc., New York Gudmundsson G, Asgeirsson H (1975) Some investigations of alkali-aggregate reaction. Cem Concr Res 5:211220 Gudmundsson G, Olafsson H (1996) Silica fume in concrete - 16 years of experience in Iceland. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 562-569 Guedon JS, Le Roux A, Martineau F (1992) Visualisation of alkali-reaction products by fluorescence. Development of the method for the survey of structures. In: Procs 9th ICAAR. Poole AB (ed) London, United Kingdom, p 395-399 Guggenheim S, Bain DC, Bergaya F, Brigatti MF, Drits VA, Eberl DD, Formoso MLL, Galán E, Merriman RJ, Peacor DR, Stanjek H, Watanabe T (2002) Report of l’Association Internationale Pour l’Etude des Argiles (AIPEA), Nomenclature Committee for 2001: Order, disorder, and crystallinity in phyllosilicates and the use of the crystallinity index. Clay Mineral 37:389-393 Guthrie Jr GD, Carey JW (1997) A simple environmentally friendly, and chemically specific method for the identification and evaluation of the alkali-silica reaction. Cem Concr Res 27:1407-1417 Gutteridge WA, Hobbs DW (1980) Some chemical and physical properties of Beltane opal rock and its gelatinous alkali silica reaction product. Cem Concr Res 10(2):183-193 Hadizadeh J, Law RD (1991) Water-weakening of sandstone and quartzite deformed at various stress and strain rates. Int J Rock Mech Mining Sci Geomech Abs 28:431-439 Hadley DW (1961) Alkali reactivity of carbonate rocks - expansion and dedolomitization. Hwy Res Bd Procs 40:462-474 Hamilton JP, Pantano CG, Brantley SL (1998) The dissolution behaviour of crystalline and amorphous albite in acidic and basic environments at 25 °C. Mineral Mag 62a:565-566 Hansen WC (1959) Release of alkalies by sands and admixtures in Portland cement mortars. ASTM Bull 236:353-359

352

Broekmans

Hart JR, Zhu Y, Pirard E (2010) Particle size and shape characterization: current technology and practice. In: Advances in the Characterization of Industrial Minerals. Christidis GE (ed) EMU Notes Mineral 9:77-127 Hasparyk NP, Lopes ANM, Cavalcanti AJCT, Silveira JFA (2004a) Expansions due to AAR verified in concrete cores and mortars from several power plants in Brazil. In: Procs 12th ICAAR. Tang M, Deng M (eds) Beijing, China, p 888-897 Hasparyk NP, Lopes ANM, Cavalcanti AJCT, Silveira JFA (2004b) Deterioration index and properties of concretes from Paolo Alfonso Power Plants - Brazil affected by the alkali-aggregate reaction. In: Procs 12th ICAAR. Tang M, Deng M (eds) Beijing, China, p 898-906 Haugen M, Skjølsvol, O, Lindgård J, Østnor T, Wigum BJ (2003) Experience from petrographic analyses on aggregate separated from concrete cores. In: Procs 9th EMABM. Broekmans MATM, Jensen V, Brattli B (eds) Trondheim, Norway, p 8 Hawthorne FC, Ferny P (1982) The mica group. MAC Short Course Handbook 8:63-98 Hayward SA, Salje EKH (2000) Twin memory and twin amnesia in anorthoclase. Minreal Mag 64:195-200 Hazen RM, Wones DR (1972) The effect of cation substitution on the physical properties of trioctahedral micas. Am Mineral 57:103-129 Hazen RM, Wones DR (1973) The effect of cation substitution on the physical properties of trioctahedral micas. Addenda. Am Mineral 58:137 He C, Osbæck B, Makovicky E (1995) Pozzolanic reactions of six principal clay minerals: activation, reactivity assessments and technological effects. Cem Concr Res 25:1691-1702 Heaney PJ, Veblen DR (1991) Observation and kinetic analysis of a memory effect at the α-β quartz transition. Am Mineral 76:1459-1466 Heaney PJ, Post JE (1992) The widespread distribution of a novel silica polymorph in microcrystalline quartz varieties. Science 255:441-443 Heaney PJ, Prewitt CT, GibbsGV (eds) (1994) Silica: physical behavior, geochemistry and materials applications. Rev Mineral 29:1-606 Heaney PJ (1994) Structure and chemistry of the low-pressure silica polymorphs. Rev Mineral 29:1-40 Heimann RB, Hooton, RD (1986) Mineralogical changes of various cement formulations during reaction with groundwater in the presence of Ca- and Na-bentonite at 150 C. Can Mineral 24:289-302 Helmuth R, Stark D (1992) Alkali-silica reactivity mechanisms. In: Alkali-silica reactivity mechanisms in materials science of concrete III. Skalny J (ed) American Ceramic Society, Westerville, Ohio, p 131-208 Hewlett PC (ed) (1998) Lea’s Chemistry of Cement and Concrete. 4th edition, Arnold Publishers, London Hime WG (2001) Chemical methods of analysis of concrete. In: Handbook of Analytical Techniques in Concrete Science and Technology. Principles, Techniques, and Applications. Ramachandran VS, Beaudoin JJ (eds) Noyes Publications, Park Ridge, New Jersey Hitzman MW (1999) Routine staining of drill core to determine carbonate mineralogy and distinguish carbonate alteration textures. Mineral Deposita 34:794-798 Hochella MF Jr, White AF (eds) (1990) Mineral-Water Interface Geochemistry. Reviews in Mineralogy Volume 23. Mineralogical Society of America Hochella MF Jr, Banfield JF (1995) Chemical weathering of silicates in nature: a microscopic perspective with theoretical considerations. Rev Mineral 31:353-406 Hong SY, Glasser FP (1999) Alkali binding in cement pastes, part I. The C-S-H phase. Cem Concr Res 29:18931903 Hong SY, Glasser FP (2002): Alkali sorption by C-S-H and C-A-S-H gels. Part II. Role of alumina. Cem Concr Res 32:1101-1111 Hou X, Struble LJ, Kirkpatrick RJ (2004) Formation of ASR gel and the roles of C-S-H and portlandite. Cem Concr Res 34:1683-1696 Hou X, Kirkpatrick RJ, Struble LJ, Monteiro PJM (2005) Structural investigations of alkali-silica gels. J Am Ceram Soc 88(4):943-949 Hudec PP, Larbi JA (1989) A study of alkali-aggregate reaction in concrete: measurement and prevention. Part I: measurement - development of a rapid AR test method. Cem Concr Res 19:193-198 Hudec PP, Larbi JA (1990) A study of alkali aggregate reaction in concrete: measurement and prevention. Part II: air saturated hot and cold NaCl solutions. Cem Concr Res 20:73-78 Hudec PP, Ghamari RC (2000) Ground waste glass as an alkali-silica reactivity inhibitor. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 663-671 Humphries DW (1992) The preparation of thin sections of rocks, minerals and ceramics. Roy Micr Soc, Oxford Science Publications, Microscopy Handbooks 24:1-83 Hünger KJ (2007) The contribution of quartz and the role of aluminum for understanding the AAR with greywacke. Cem Concr Res 37:1193-1205 Hünger KJ, Hübert C, Heidemann D (2008) Formation of kanemite from precursors as a model for ASR gel hydration. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 130-139 Icopini GA, Brantley SL, Heaney PJ (2005) Kinetics of silica oligomerization and nanocolloid formation as a function of pH and ionic strength at 25 °C. Geochim Cosmochim Acta 69:293-303

Deleterious Reactions of Aggregate With Alkalis in Concrete

353

Iler RK (1964) Ion-exchange properties of crystalline hydrated silica. J Colloid Sci 19:648-657 Iler RK (1979) The Chemistry of Silica. Solubility, polymerization, collid and surface properties, and biochemistry. 2nd edition. Wiley Interscience, Ithaca, New York Ingham J (2010) Geomaterials Under the Microscope. A Colour Guide. Manson Publishing, London, United Kingdom Jbara O, Cazaux J, Trebbia P (1995) Sodium diffusion in glasses during electron irradiation. J Appl Phys 78:868875 JCAS I-31 (1996) Quantitative measurement of quartz (phosphoric acid method). (in Japanese) Japan Cem Assoc, Tokyo, p 1-11 Jensen V (1993) Alkali aggregate reaction in southern Norway. PhD Thesis, Norwegian University of Science and Technology NTNU, Trondheim, Norway, pp 262 Jensen V, Haugen M (1996) Alkali reactions in Northern Norway. Report nr.2: Sedimentary deposits and petrographic analysis. (translated Norwegian title) SINTEF Report (STF22 A96806): pp 37 Jensen V (2004) Alkali-silica reaction damage to Elgeseter Bridge, Trondheim, Norway: a review of construction, research and repair up to 2003. Mater Charact 53(2-4):155-170 Johan Z, Maglione GF (1972) La kanemite, nouveau silicate de sodium hydrate de néoformation. Bull Soc Franç Mineral Crystallogr 95:371-382 Jones AEK, Clark LA (1994) The practicalities and theory of using crack width summation to estimate ASR expansion. Procs Inst Civ Eng Structs Builds 104:183-192 Juenger MCG, Ostertag CP (2004) Alkali-silica reactivity of large silica fume-derived particles. Cem Concr Res 34:1389-1402 Juilland P, Gallucci E, Flatt R, Scrivener K (2010) Dissolution theory applied to the induction period in alite hydration. Cem Concr Res 40:831-844 Juliani M, Cavalcanti AJCT, Carrazedo R, Gaspare JC (2008) Analysis of the effects caused by alkali aggregate reaction on the structures of Moxotó power plant. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 838-848 Justnes H (2012) Alternative low-CO2 “green” clinkering processes. Rev Mineral Geochem 74:83-99 Kalousek GL (1944) Studies of portions of the quaternary system soda-lime-silica-water at 25°C. J Natl Bureau Standards 32:285-304 Kalousek GL (1955) An interpretation of some published researches on the alkali-aggregate reaction. J Am Concr Inst Procs 51:812/1-812/20 Karlsson M, Craven C, Dove PM, Casey WH (2001) Surface charge concentrations on silica in different 1.0 m metal-chloride background electrolytes and implications for dissolution rates. Aquat Geochem 7:13-32 Katayama T, Kaneshige Y (1986) Diagenetic changes in potential alkali-aggregate reactivity of volcanic rocks in Japan - a geological interpretation. In: Procs 7th ICAAR. Grattan-Bellew PE (ed) Ottawa, Canada, p 489-495 Katayama T, Futagawa T (1989) Diagenetic changes in potential alkali-aggregate reactivity of siliceous sedimentary rocks in Japan. In: Procs 8th ICAAR. Okada K, Nishibayashi S, Kawamura M (eds) Kyoto, Japan, p 525-530 Katayama T, St. John DA, Futagawa T (1989) The petrographic comparison of some volcanic rocks from Japan and New Zealand - potential reactivity related to interstitial glass and silica minerals. In: Procs 8th ICAAR. Okada K, Nishibayashi S, Kawamura M (eds) Kyoto, Japan, p 537-542 Katayama T (1992) A critical review of carbonate rock interactions - Is their reactivity useful or harmful? In: Procs 9th ICAAR. Poole AB (ed) London, United Kingdom, p 508-518 Katayama T, Bragg DJ (1996) Alkali-aggregate reaction combined with freeze-thaw in Newfoundland petrography using EPMA. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 243-250 Katayama T, Helgason TS, Olafson H (1996) Petrography and alkali-reactivity of some volcanic aggregates from Iceland. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 377-384 Katayama T (1998) Petrographic diagnosis of alkali-aggregate reaction in concrete based on quantitative EPMA analysis. In: Procs 4th CANMET/ACI/JCI Intl Conf, Tokushima, Japan. Malhotra VM (ed) ACI Spec Pub 179:539-560 Katayama T (2000) Alkali-aggregate reaction in the vicinity of Izmir, Western Turkey. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 365-374 Katayama T (2004a) How to identify carbonate rock reactions in concrete. Mater Charact 53/2-4):85-104 Katayama T (2004b) Alkali-aggregate reaction under the influence of deicing salts in the Hokuriku district, Japan. Mater Charact 53/2-4:105-122 Katayama T, Sommer H (2008) Further investigations of the mechanisms of so-called alkali-carbonate reaction based on modern petrographic techniques. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 829-839 Katayama T (2010a) The so-called alkali-carbonate reaction (ACR) - its mineralogical and geochemical details, with special reference to ASR. Cem Concr Res 40(4):643-675

354

Broekmans

Katayama T (2010b) Diagnosis of alkali-silica reaction – polarizing microscopy and SEM-EDS analysis. In: Procs 6th CONSEC. Castro-Borges P, Moren, EI, Sakai K, Gjørv OE, Banthia N (eds) Mérida, Yucatan, Mexico, p 19-34 Katayama T (2011) So-called ACR - petrographic details of field concretes in Ontario. In: Procs 13th EMABM. Mauko A, Kosec T, Kopar T, Gartner N (eds) Ljubljana, Slovenia: pp15 Kats A (1962a) Hydrogen in alpha-quartz, part 1. Philips Res Reports 17:133-195 Kats A (1962b) Hydrogen in alpha-quartz, part 2. Philips Res Reports 17:201-279 Kawamura M, Takemoto K, Hasaba S (1983) Application of quantitative EDXA analysis and microhardness measurements to the study of alkali-silica reaction mechanisms. In: Procs 6th ICAAR. Idorn GM, Rostam S (eds) Copenhagen, Denmark, p 167-174 Kawamura M, Koike M, Nakano K (1989) Release of alkalis from reactive andesitic aggregates and fly ashes into pore solution in mortars. Procs 8th ICAAR, Kyoto, Japan, p 271-278 Kawamura M, Takeuchi K, Sugiyama A (1996) Mechanism of the influence of externally supplied NaCl on the expansion of mortars containing reactive aggregate. Mag Conc Res 48:237-248 Kawamura M, Arano N, Terashima T (1998) Composition of ASR gels and expansion of mortars. In: Materials Science of Concrete: The Sidney Diamond Symposium, Special Volume. Cohen M, Mindess S, Skalny J (eds) Am Ceram Soc, Westerville, Ohio, p 261-276 Keene MTJ, Knowles JA, Hudson MJ (1996) The reversible extraction of the hexaminecobalt (III) cation by kanemite (NaHSi2O5·3H2O): enhanced extraction in the presence of a cationic surfactant. J Mater Chem 6:1567-1573 Kekulawala KRSS, Paterson MS, Boland JN (1978) Hydrolytic weakening in quartz. Tectonophys 46(1-2):T1T6 Kennerley RA, St. John DA (1969) Reactivity of aggregates with cement alkalis. In: Procs Nat Conf Conc Aggs, Wellington, NZ. New Zealand Portl Cem Assoc, p 35-47 Kinrade SD, Hamilton RJ, Schach AS, Knight CTG (2001): Aqueous hypervalent silicon complexes with aliphatic sugar acids. J Chem Soc Dalton Trans 961-963 Kirkpatrick RJ, Kalinichev AG, Hou X, Struble L (2005) Experimental and molecular dynamics modeling studies of interlayer swelling: water incorporation in kanemite and ASR gel. Mater Struct 38(4):449-458 Kjellsen KO, Monsøy A (1996) Preparation of flat-polished specimens of high-performance concrete for SEM: backscattered electron imaging and microanalysis. In: Procs 18th Intl Conf Cem Micro, Houston, Texas, p 356-364 Kjellsen KO, Monsøy A, Isachsen K, Detwiler RJ (2003) Preparation of flat-polished specimens for SEMbackscattered electron imaging and X-ray micro-analysis - importance of epoxy impregnation. Cem Concr Res 33:611-616 Kloprogge JT (1992) Pillared clays. Preparation and characterization of clay minerals and aluminum-based pillaring agents. PhD-thesis Utrecht University, Geologica Ultraiectina 91:1-349 Klug HP, Alexander LE (1974) X-ray Diffraction Procedures. 2nd Edition. John Wiley and Sons, New York. Knauss KG, Wolery TJ (1986) Dependence of albite dissolution kinetics on pH and time at 25°C and 70°C. Geochim Cosmochim Acta 50:2481-2497 Knauss KG, Wolery TJ (1989) Muscovite dissolution kinetics as a function of pH and time at 70 °C. Geochim Cosmochim Acta 53:1493-1501 Kneller WA (1967) Aggregate solubility related to chert petrographic textures and electron surface morphologies. In: Procs 3rd Forum Geol Industr Mins. Angino EE, Hardy RG (eds) State Geol Surv Kansas, Special Distribution 34:110-119 Kneller WA, Kriege HF, Saxer EL, Wilband JT, Rohrbacher TJ (1968) The properties and recognition of deleterious cherts which occur in aggregate used by Ohio concrete producers. Aggregate Research Group, Geol and Civ Eng Depts, University of Toledo, Ohio, Final Report 1014:p201 Knudsen T, Thaulow N (1975) Quantitative microanalyses of alkali-silica gel in concrete. Cem Concr Res 5:443-454 Kojima T, Takagi N, Haruta K (2000) Expanding characteristics of mortar with glass powder produced from waste bottles. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 663-671 Komarneni S, Breval E, Miyake M, Roy R (1987) Cation-exchange properties of (Al+Na)-substituted synthetic tobermorites. Clays Clay Miner 35:385-390 Komarneni S, Guggenheim S (1988) Comparison of cation exchange in ganophyllite and [Na+Al]-substituted tobermorite: crystal-chemical implications. Minreal Mag 52:371-375 Korkanç M, Tuğrul A (2005) Evaluation of selected basalts from the point of alkali-silica reactivity. Cem Concr Res 35:505-512 Korovkin MV, Ananieva LG, Antsiferova AA (2012) Assessment of quartzite crystallinity index by FT-IR. In: Procs 10th ICAM, Broekmans MATM (edr), Springer Verlag, Heidelberg, p 403-409

Deleterious Reactions of Aggregate With Alkalis in Concrete

355

Kreshkov AP, Chivikova AN, Zagorovskava AA (1965) A rapid method for determining free amorphous silica in clay. J Anal Chem USSR 20:1295-1296 Kronenberg AK, Kirby SH (1987) Ionic conductivity of quartz: DC time dependence and transition in charge carriers. Am Mineral 72:739-747 Kühnel RA, Verkroost TW, Chaudry MA (1987) Thermal behavior of amorphous silica. In: Procs Conf on Geochim and Min Formation in the Earth Surface. Rodríguez-Clemente R, Tardy Y (eds) Madrid, Spain, p 863-875 Kühnel RA (1995) Problems of quantification of fabrics. In: Procs 5th EMABM. Elsen J (ed) Leuven, Belgium, p 58-67 Kühnel RA, Van der Gaast SJ (1996) Humidity sensitive minerals. In: Procs 14th Conf Clay Min and Petr, Banska Stiavnica, Slovakia. Acta Carpathica, Series Clays 4:100-101 Kurtis KE, Monteiro PJM, Brown JT, Meyer-Ilse W (1998) Imaging of ASR gel by soft X-ray microscopy. Cem Concr Res 28(3):411-421 Kurtis KE, Monteiro PJM, Brown JT, Meyer-Ilse W (1999) High resolution transmission soft X-ray microscopy of deterioration products developed in large concrete dams. J Microsc 196:288-298 Kurtis KE, Collins CL, Monteiro PJM (2002) The surface chemistry of the alkali-silica reaction: a critical evaluation and X-ray microscopy. Concr Sci Eng 4:2-11 Kurtis KE, Monteiro PJM (2003) Chemical additives to control expansion of alkali-silica reaction gel: proposed mechanisms of control. J Mater Sci 38:2027-2036 Lagaly G, Beneke K, Weiss A (1975a) Magadiite and H-magadiite: I. Sodium magadiite and some of its derivatives. Am Mineral 60:642-649 Lagaly G, Beneke K, Weiss A (1975b) Magadiite and H-magadiite: II. H-magadiite and its intercalation compounds. Am Mineral 60:650-658 Lagaly G, Beneke K (1991) Intercalation and ion exchange reactions of clay minerals and non-layer compounds. Colloid Polym Sci 269:1198-1211 Lagerblad B, Jacobsson B (1997) Smectite clays and concrete durability. In: Procs 19th Intl Conf Cem Microscopy, Cincinnati, Ohio, p 151-162 Laing SV, Scrivener KL, Pratt PL (1992) An investigation of alkali-silica reaction in seven-year old and model concretes using SEM and EDS. In: Procs 9th ICAAR, Poole AB (ed) London, United Kingdom, p 579-586 Lakshtanov LZ, Stipp SLS (2010) Interaction between dissolved silica and calcium carbonate. 1. Spontaneous precipitation of calcium carbonate in the presence of dissolved silica. Geochim Cosmochim Acta 74:26552664 Lane SJ, Dalton JA (1994) Electron microprobe analysis of geological carbonates. Am Mineral 79:745-749 Lastra R, Cabri LJ, Weiblen PW (2003) Comparative liberation study by image analysis of Merensky Reef samples comminuted by electric-pulse disaggregation and by conventional crusher. In: Procs 22nd Intl Min Processing Congr. Lorenzen L, Bradshaw DJ (eds) Cape Town, South Africa, p 251-260 LCPC (1993) Essai de granulats. Determination des alcalins solubles dans l’eau de chaux. Méthode d’essai LPC n 37. Laboratoire Central des Ponts et Chaussées LCPC, Paris, France, p 1-13 Lee CJ, Yu SC (1994) A crystallinity index measurement on quartz crystals in metapelite from the southern cross-island highway of Taiwan and its significance. J Geol Soc China 37:215-224 Lee MR, Hodson ME, Parsons I (1998) The role of intragranular microtextures and microstructures in chemical and mechanical weathering: direct comparisons of experimentally and naturally weathered feldspars. Geochim Cosmochim Acta 62:2771-2788 Lee H, Cody RD, Cody AM, Spry PG (2005): The formation and role of ettringite in Iowa highway concrete deterioration. Cem Concr Res (35):332-343. Lee MR, Brown DJ, Smith CL, Hodson ME, MacKenzie M, Hellmann R (2007) Characterization of mineral surfaces using FIB and TEM: A case study of naturally weathered alkali feldspars. Am Mineral 92:13831394 Leek DS (1991) The passivity of steel in concrete. Q J Eng Geol 24:55-66 Leemann A, Holzer L (2005) Alkali-aggregate reaction – identifying reactive silicates in complex aggregates by ESEM observation of dissolution features. Cem Concr Comp 27:796-801 Leeman A, Lothenbach B (2008) The influence of potassium-sodium ratio in cement on concrete expansion due to alkali-aggregate reaction. Cem Concr Res 38:1162-1168 Leeman A, Le Saout G, Winnefeld F, Rentsch D, Lothenbach B (2011) Alkali-silica reaction: the influence of calcium on silica dissolution and the formation of reaction products. J Am Ceram Soc 94(4):1243-1249 Li D, Bancroft GM, Kasrai M, Fleet ME, Secco RA, Feng XH, Tan KH, Yang BX (1994) X-ray absorption spectroscopy of silicon dioxide (SiO2) polymorphs: the structural characterization of opal. Am Mineral 79:622-632 Liang T, Meihua W, Sufen H, Tang M (1997) Method to evaluate the reactivity of microcrystalline quartz in carbonate rock. Adv Cem Res 9(34):55-63

356

Broekmans

Lindgård J, Wigum BJ (2003) Alkali aggregate reactions - Field experience. (translated Norwegian title) SINTEF Report No STF22 A02616: pp127 + appendices Lindgård J, Skjølsvold O, Haugen M, Hagelia P, Wigum BJ (2004) Experience from evaluation of degree of damage in fluorescence impregnated plan polished sections of half cores based on the “crack index method.” In: Procs 12th ICAAR. Tang M, Deng M (eds) Beijing, China, p 939-947 Lindgård J, Andiç-Çakır Ö, Fernandes I, Rønning TF, Thomas MDA (2012) Alkali-silica reactions (ASR): Literature review on parameters influencing laboratory performance testing. Cem Concr Res 42:223-243 Locati F, Marfil S, Baldo E, Maiza P (2010) Na2O, K2O, SiO2 and Al2O3 release from potassic and calcic-sodic feldspars into alkaline solutions. Cem Concr Res 40:1189-1196 Lofgren G (1970) Experimental devitrification of rhyolite glass. Geol Soc Am Bull 81:553-559 Lofgren G (1971) Experimentally produced devitrification textures in natural rhyolitic glass. Geol Soc Am Bull 82:111-123 Lohmann KC, Meyers WJ (1977) Microdolomite inclusions in cloudy prismatic calcites: a proposed criterion for former high-magnesium calcites. J Sediment Petrol 47:1078-1088 Lombardi J, Perruchot A, Massard P, Larive C (1996a) Étude comparée de gels silico-calciques produits des réactions alcalis-granulats dans les bétons et de gels synthétiques types. Cem Concr Res 26:623-631 Lombardi J, Perruchot A, Massard P (1996b) Les alcalins apartiennent - ils á la structure des gels, produits des alcalis-réaction dans les betons? Comptes Rendues de l’Academie des Sciènces Paris, Sciences de la Terre et des Planètes (323-IIa): 283-288 López-Buendía AM, Climent V, Verdú P (2006) Lithological influence of aggregate in the alkali-carbonate reaction. Cem Concr Res 36:1490-1500 Loste E, Wilson RM, Seshadri R, Meldrum FC (2003) The role of magnesium in stabilising amorphous calcium carbonate and controlling calcite morphologies. J Cryst Growth 254:206-218 Lu D, Fournier B, Grattan-Bellew PE (2004) Evaluation of the Chinese accelerated test for alkali-carbonate reaction. In: Procs 12th ICAAR. Tang M, Deng M (eds) Beijing, China, p 386-392 Lu D, Zhou X, Xu Z, Lan X, Tang M, Fournier B (2006) Evaluation of laboratory test method for determining the potential alkali contribution from aggregate and the ASR safety of the Three-Gorges dam concrete. Cem Concr Res 36:1157-1165 Ludwig U, Bauer W (1976) Untersuchungen zur Alkali-Zuschlag-Reaktion. Zement-Kalk-Gips 29(9):401-411 Lumley JS (1989) Synthetic cristobalite as a reference reactive aggregate. In: Procs 8th ICAAR. Okada K, Nishibayashi S, Kawamura M (eds) Kyoto, Japan, p 561-566 Lumley JS (1997) ASR suppression by lithium compounds. Cem Concr Res 27:235-244 Lüschen H (1979) Die Namen der Steine. Das Mineralreich im Spiegel der Sprache: mit einem Wörterbuch, enthaltend über 1300 Namen von Mineralien, Gesteinen, Edelsteinen, Fabel- und Zaubersteinen. 2. Ausgabe. Ott Verlag, Thun, Switzerland Lüttge A, Winkler U, Lasaga AC (2003) Interferometric study of the dolomite dissolution: a new conceptual model for mineral dissolution. Geochim Cosmochim Acta 67:1099-1116 Mackenzie FT, Bischoff WD, Bishop FC, Loijens M, Schoonmaker J, Wollast R (1983) Magnesian calcites: low-temperature occurrence, solubility and solid solution behavior. Rev Mineral 11:97-144 Malarkey J, Pearson DG, Kjarsgaard BA, Davidson JP, Nowell GM, Ottley CJ, Stammer J (2010) From source to crust: tracing magmatic evolution in a kimberlite and a melilitite using microsample geochemistry. Earth Planet Sci Lett 299:80-90 Malmström M, Banwart S (1997) Biotite dissolution at 25°C: the pH dependence of dissolution rate and stoichiometry. Geochim Cosmochim Acta 61:2779-2799 Manara D, Grandjean A, Neville DR (2009) Advances in understanding the structure of borosilicate glasses: A Raman spectroscopy study. Am Mineral 94:777-784 Marfil SA, Maiza PJ (1993) Zeolite crystallization in portland cement concrete due to alkali-aggregate reaction. Cem Concr Res 23:1283-1288 Marinoni N, Broekmans MATM (2012) Microstructure of quartz in twelve Italian sedimentary aggregates and a critical review of the crystallinity index for quartz. Am J Sci in review Marshall RR (1961) Devitrification of natural glass. Geol Soc Am Bull 72:1493-1520 Marshall WL, Warakomski JM (1980) Amorphous silica solubilities-II. Effect of aqueous salt solutions at 25 °C. Geochim Cosmochim Acta 44:915-924 Marshall WL, Chen CTA (1982) Amorphous silica solubilities. V. Predictions of solubility behavior in aqueous mixed electrolyte solutions to 300 °C. Geochim Cosmochim Acta 46:289-291 Marshall WL (1980) Amorphous silica solubilities. III. Activity coefficient relations and predictions of solubility behavior in salt solutions, 0-350 °C. Geochim Cosmochim Acta 44:925-931 Mather K, Buck AD, Luke WI (1964) Alkali-silica and alkali-carbonate reactivity of some aggregates from South Dakota, Kansas and Missouri. Natl Acad Sci - Natl Res Council. Hwy Res Rec 45, Pub 1167:72-109 Mather B (1976) A discussion of the paper “A review of alkali silica reaction and expansion mechanisms. 2. Reactive aggregates” by Sidney Diamond. Cem Concr Res 6:813-814

Deleterious Reactions of Aggregate With Alkalis in Concrete

357

McAtee Jr JL, House R, Eugster HP (1968) Magadiite from Trinity County, California. Am Mineral 53:20612069 McCoy WJ, Caldwell AG (1951) New approach in inhibiting alkali-aggregate expansion. ACI Mater J 22(9):693706 McCrea JM (1950) On the isotope chemistry of carbonates and a paleotemperature scale. J Chem Phys 18:849857 McDonald AM, Proenza JA, Zaccarini F, Rudashevsky NS, Cabri LJ, Stanley CJ, Rudashevsky VN, Melgarejo JC, Lewis JF, Longo F, Bakker RJ (2010) Garutiite, (Ni, Fe, Ir), a new hexagonal polymorph of native Ni from Loma Peguera, Dominican Republic. Eur J Mineral 22:293-304 McGee H (2004) On Food and Cooking. The Science and Lore of the Kitchen. 2nd Revised Edition. Scribner, New York McNally C, Richardson M, Carr A, Strogen P (2004) Domain size as a parameter for studying the potential alkali-silica reactivity of chert-bearing aggregates. Mag Concr Res 56:201-209 McNally C, Richardson MG (2005) Reactivity assessment of aggregates: role of chert crystallinity. ACI Mater J 102:163-169 McRae ND (1995) Secondary-ion mass spectrometry and geology. Can Mineral 33:219-236 Meral C, Benmore CJ, Monteiro PJM (2011) The study of disorder and nanocrystallinity in C-S-H, supplementary cementitious materials and geopolymers using pair distribution function analysis. Cem Concr Res 41:696710 Meier R, Anderson J, Verryn S (2012) Industrial x-ray diffraction analysis of building materials. Rev Mineral Geochem 74:147-167 Meyer EE, Greene GW, Alcantar NA, Israelachvili JN, Boles JR (2006) Experimental investigation of the dissolution of quartz by a muscovite mica surface: implications for pressure solution. J Geophys Res Solid Earth 111:B08202 Michel-Lévy A, Munier-Chalmas CPE (1892) Mémoire sur diverses formes affectées par le réseau élémentaire du quartz. Bull Soc Franç Min 15:159-190 Midgley HG (1976) The identification of opal and chalcedony in rocks and methods of estimating the quantities present. In: Procs 3rd ICAAR. Poole AB (ed) London, United Kingdom, p 193-201 Miehe G, Graetsch H (1992) Crystal structure of moganite: a new structure type for silica. Eur J Mineral 4:693706 Milanesi CA, Batic OR (1994) Alkali reactivity of dolomitic rocks from Argentina. Cem Concr Res 24:10731084 Milanesi CA, Marfil SA, Batic OR, Maiza PJ (1996) The alkali-carbonate reaction and its reaction products, an experience with Argentinean dolomite rocks. Cem Concr Res 26:1579-1591 Monteiro PJM, Mehta PK (1986) Interaction between carbonate rock and cement paste. Cem Concr Res 16:127134 Monteiro PJM, Wang KS, Dos Santos GMC, De Andrade WP (1997) Influence of mineral admixtures on the alkali-aggregate reaction. Cem Concr Res 27:1899-1909 Moore GSM (1986) The α-β inversion in quartz and the effects of structural damage. Phase Trans 7:25-40 Moranville-Regourd M (1989) Products of reaction and petrographic examination. In: Procs 8th ICAAR. Okada K, Nishibayashi S, Kawamura M (eds) Kyoto, Japan, p 445-456 Morino K (1989) Alkali-aggregate reactivity of cherty rock. In: Procs 8th ICAAR. Okada K, Nishibayashi S, Kawamura M (eds), Kyoto, Japan, p 501-506 Morino K, Iwatsuki E, Sarai Y (1997) ASR expansion and microstructure of concrete with cherty rock. In: Procs East Asia AAR Seminar. Nishibayashi S, Kawamura M (eds) Tottori, Japan, p 113-124 Motchany AI, Chvanski PP (2001) Crystal growth of an α-quartz like piezoelectric material, berlinite. Ann Chim Sci Mat 26:199-208 Moxon T, Rios S (2002) Agate: a study of ageing. Eur J Mineral 14:1109-1118 Moxon T, Rios S (2004) Moganite and water content as a function of age in agate: an XRD and thermogravimetric study. Eur J Mineral 16:269-278 Müller A, Wiedenbeck M, Van den Kerkhof AM, Kronz A, Simon K (2003) Trace elements in quartz; a combined electron microprobe, secondary ion mass spectrometry, laser-ablation ICP-MS, and cathodoluminescence study. Eur J Mineral 15:747-763 Multon S, Cyr M, Sellier A, Diederich P, Petit L (2010) Effects of aggregate size and alkali content on ASR expansion. Cem Concr Res 40(4):508-516 Muraishi H (1989) Crystallization of silica gel in alkaline solutions at 100 to 180 °C: Characterization of SiO2-Y by comparison with magadiite. Am Mineral 74:1147-1151 Muraishi H (1995) Experimental study of amorphous silica crystallization in NaCO3-NaHCO3 solutions as a source of chert formation in Lake Magadi. Bull Chem Soc Japan 68:3027-3033 Murata KJ, Norman-II MB (1976) An index of crystallinity for quartz. Am J Sci 276:1120-1130 Nagy KL (1995) Dissolution and precipitation kinetics of sheet silicates. Rev Mineral 31:173-233

358

Broekmans

Nassau K (1983) The Physics and Chemistry of Color: The Fifteen Causes of Color. 1st Edition. John Wiley and Sons, New York Natesaiyer K, Hover KC (1988) In situ identification of ASR products in concrete. Cem Concr Res 18:455-463 Natesaiyer K, Hover KC (1989) Further study of an in-situ identification method for alkali-silica reaction products in concrete. Cem Concr Res 19:770-778 Natesaiyer K, Stark D, Hover KC (1991) Gel fluorescence reveals reaction product traces. Concr Int 13(1):25-28 Nielsen CH, Sigurdsson H (1981) Quantitative methods for electron microprobe analysis of sodium in natural and synthetic glasses. Am Mineral 66:547-552 Nijland TG, Verschure RH, Maijer C (1994) Catalytic effect of biotite: formation of hydrogarnet lenses. Compt Rend Acad Sci Paris 318-II:501-506 Nijland TG (1997) personal communication Nixon PJ, Canham I, Page CL, Bollinghaus R (1987) Sodium chloride and alkali-aggregate reaction. In: Procs 7th ICAAR. Grattan-Bellew PE (ed) Ottawa, Canada, p 110-114 Normin2000 (1996) Alkali reactions in concrete. Pilot study. (translated Norwegian title) The Research Council of Norway, Trondheim: pp86. Normin2000 (1999) Alkali reactions in concrete. Project report. (translated Norwegian title) The Research Council of Norway, Trondheim: pp 166 + photo atlas: pp 14. Norwegian Concrete Association (2008) Durable concrete with alkali-reactive aggregate. Translated English edition. Norw Concr Assoc, Publication 21:p22+12 Oberti R, Camara F, Ottolini L, Caballero JM (2003) Lithium in amphiboles: detection, quantification, crystalchemical mechanisms and definition of a compositional space bridging sodic and BLi-amphibole. Eur J Mineral 15:309-319 Oelkers EH, Bjørkum PA, Murphy WM (1996) A petrographic and computational investigation of quartz cementation and porosity reduction in North Sea sandstones. Am J Sci 296:420-452 Oelkers EH, Schott J, Gauthier JM, Herrero-Roncal T (2008) An experimental study of the dissolution mechanism and rates of muscovite. Geochim Cosmochim Acta 72:4948-4961 Ong S, Diamond S (1993) Unexpected effects of alkali hydroxide added to cement mix water. In: Cement-based materials: present, future, and environmental aspects. Moukwa M, Sarkar SL, Luke K, Grutzeck MW (eds), Am Ceram Soc, Westerville, Ohio. Ceram Trans 37:37-48 Pagano MA, Cady PD (1982) A chemical approach to the problem of alkali-reactive carbonate aggregates. Cem Concr Res 12:1-12 Palmer DC (1994) Stuffed derivatives of the silica polymorphs. Rev Mineral 29:83-122 Papike JJ (1987) Chemistry of the rock-forming silicates: Ortho, ring, and single-chain structures. Revs Geoph 25(7):1483-1526 Papike JJ (1988) Chemistry of the rock-forming silicates: Multiple-chain, sheet, and framework structures. Revs Geoph 26(3):407-444 Pastore HO, Munsignatti M, Mascarenhas AJS (2000) One-step synthesis of alkyltrimethylammoniumintercalated magadiite. Clays Clay Mineral 48:224-229 Pearson JC, Laughlin GF (1923) An investigation case of dangerous aggregate. Proc Am Conc Inst 19:142-154 Pedersen BM (2004) Alkali-reactive and inert fillers in concrete. Rheology of fresh mixtures and expansive reaction. PhD Thesis, Norwegian University of Science and Technology (NTNU), Trondheim, Norway (2004/92): pp198 + pp48 (appendices). Peterson K, Gress D, Van Dam T, Sutter L (2006) Crystallized alkali-silica gel in concrete from the late 1890s. Cem Concr Res 36:1523-1532 Phair JW, Livingstone RA, Brown CM, Benesi AJ (2004) Investigation of the state of water in hydrating layered sodium disilicate in crystalline and amorphous forms by quasi-elastic neutron scattering. Chem Mater 16:5042-5050 Pinnavaia TJ, Johnson ID, Lipsicas M (1986) A 29Si MAS NMR study of tetrahedral site distributions in the layered silicic acid H+-magadiite (H2Si14O29⋅nH2O) and Na+-magadiite (Na2Si14O29⋅nH2O). J Solid State Chem 63:118-121 Plyusnina II (1978) Study of the disorder in the structure of chalcedony by infrared spectroscopy. (in Russian). Dokl Akad Nauk USSR 240:839-842 Poirier JP (1985) Creep of crystals: high-temperature deformation processes in metals, ceramics and minerals. Cambridge University Press, Cambridge, United Kingdom Poitevin P, Regourd M (1983) Durabilité des bétons. Cas des granulats réactifs. Annales de I’Institut Technique du Batiment et des Travaux Publics (ITBTP) 413:109-143 Pokrovsky OS, Golubev SV, Mielczarski JA (2006) Kinetic evidences of the existence of positively charged species at the quartz-aqueous solution interface. J Colloid Interface Sci 296:189-194 Pöllmann H (2012) Calcium aluminate cements – raw materials, differences, hydration and properties. Rev Mineral Geochem 74:1-82

Deleterious Reactions of Aggregate With Alkalis in Concrete

359

Poole AB (1976) Electron probe microanalyses of reaction zones at cement/opal interfaces. In: Procs 3rd ICAAR. Poole AB (ed), London, United Kingdom, p 163-175 Poole AB, Sotiropoulos P (1980) Reactions between dolomitic aggregate and alkali pore fluids in concrete. Q J Eng Geol 23:281-287 Poole AB (1981) Alkali-carbonate reactions in concrete. In: Procs 5th ICAAR. Oberholster RE (ed) Cape Town, South Africa: S252/34 Poole AB (ed) (1992) Procs 9th ICAAR. Westminster/London, UK, p 1-1130 Ponce JM, Batic OR (2006) Different manifestations of the alkali-silica reaction in concrete according to the reaction kinetics of the reactive aggregate. Cem Concr Res 36:1148-1156 Powers TC, Steinour HH (1955a) An interpretation of some published researches on the alkali-aggregate reaction. Part 1 - the chemical reactions and mechanism of expansion. J Am Concr Inst 26:497-516 Powers TC, Steinour HH (1955b) An interpretation of some published researches on the alkali-aggregate reaction. Part 2 - a hypothesis concerning safe and unsafe reactions with reactive silica in concrete. J Am Concr Inst 26:785-811 Previte RW (1971) Some insights on the mechanism of saccharide set retardation of Portland cement. Cem Concr Res 1:301-316 Prezzi M, Monteiro PJM, Sposito G (1997) The alkali-silica reaction, part I: use of the double-layer theory to explain the behavior of reaction-product gels. ACI Mater J 94:10-17 Prince W, Perami R (1992) Mechanism of the alkali-dolomite reaction. In: Procs 9th ICAAR. Poole AB (ed) London, United Kingdom, p 799-806 Prince W, Perami R, Espagne M (1994) Une nouvelle approache du mecanisme de la reaction alcali-carbonate. Cem Concr Res 24:62-72 Putnis A, McConnell JDC (1980) Principles of Mineral Behaviour. Geoscience Texts (1) Blackwell Scientific, London. Elsevier, New York. Qian G, Deng M, Tang M (2002): Expansion of siliceous and dolomitic aggregates in lithium hydroxide solution. Cem Concr Res 32:763-768 Radonjic M, Ragnarsdóttir KV, Allen GC, Livesey P (1997) Mineralogical changes occurring within aggregates affected by alkali-carbonate reactivity. In: Procs 6th EMABM. Sveinsdóttir EL (ed) Reykjavik, Iceland, p 177-186 Radonjic M, Allen GC (1999) An AFM/ESEM study of de-dolomitisation in the aggregate-carbonate reaction (ACR). In: Procs 7th EMABM. Pietersen HS, Larbi JA, Janssen HHA (eds) Delft, The Netherlands, p 547-555 Radonjic M, Allen G, Ragnarsdottír KV, Livesey P (2000) Novel techniques in a study of the mechanism of dedolomitisation in alkali-carbonate reaction. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 99-108 Rankin AH, Miller MF, Carter JS (1987) The release of trace elements and volatiles from crinoidal limestone during thermal decripitation. Mineral Mag 51:517-525 Ray S, Gault HR, Dodd CG (1957) The separation of clay minerals from carbonate rocks. Am Mineral 42:681686 Reed SJB (1995) Electronprobe microanalysis. In: Microprobe techniques in the earth sciences. Potts PJ, Bowles JFW, Reed SJB, Cave MR (eds) The Mineralogical Society Series 6:49-89 Reed SJB (2005) Electron Microprobe Analysis and Scanning Electron Microscopy in Geology. 2nd edition. Cambridge University Press, Cambridge Regourd M, Hornain H (1986) Microstructure of reaction products. In: Procs 7th ICAAR. Grattan-Bellew PE (ed), Ottawa, Canada, p 375-380 Reksten K (1990) Superstructures in calcite. Am Mineral 75:807-812 Rémy P, Brouxel M (1992) Wollastonite: a reactive mineral to alkali-aggregate reaction. In: Procs 9th ICAAR. Poole AB (ed), London, United Kingdom, p 851-858 Richardson IG, Groves GW (1993) The incorporation of minor and trace elements into calcium silicate hydrate (C-S-H) gel in hardened cement pastes. Cem Concr Res 23:131-138 RILEM (2012) AAR-1 - Detection of potential alkali-reactivity of aggregates - petrographic method. Special Issue RILEM TC219-ACS, Mater Struct 45: accepted for publication RILEM (2012) AAR-2 - Detection of potential alkali-reactivity of aggregates - the ultra-accelerated mortar bar test. Special Issue RILEM TC219-ACS, Mater Struct 45: accepted for publication RILEM (2012) AAR-3 - Detection of potential alkali-reactivity of aggregates - method for aggregate combinations using concrete prisms. Special Issue RILEM TC219-ACS, Mater Struct 45: accepted for publication RILEM (2012) AAR-4 - Detection of potential alkali-reactivity of 60°C accelerated method for testing aggregate combinations using concrete prisms. Special Issue RILEM TC219-ACS, Mater Struct 45: accepted for publication

360

Broekmans

RILEM (2012) AAR-5 - Rapid preliminary screening test for carbonate aggregates. Special Issue RILEM TC219-ACS, Mater Struct 45: accepted for publication Rimstidt JD (1997) Quartz solubility at low temperatures. Geochim Cosmochim Acta 61:2553-2558 Rivard P (1998) Quantification de l’endommagement du béton atteint de réaction alcali-silice par analyse d’images. Master Thesis, Université de Sherbrooke, Canada: pp 166. Rivard P, Fournier B, Ballivy G (2000a) Quantitative assessment of concrete damage due to alkali-silica reaction (ASR) by petrographic analysis. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds), Québec City, QC, Canada, p 889-898 Rivard P, Fournier B, Ballivy G (2000b) The damage rating index method for ASR affected concrete a critical review of petrographic features of deterioration and evaluation criteria. Cem Concr Aggs 24(2):1-11 Rivard P, Ballivy G (2005) Assessment of the expansion related to alkali-silica reaction by the damage rating index method. Constr Build Mater 19:83-90 Rivard P, Ballivy G, Gravel C, Saint-Pierre F (2008) Monitoring of a hydraulic structure affected by ASR: a case study. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 1103-1110 Rodrigues FA, Monteiro PJM, Sposito G (2001) The alkali-silica reaction, the effect of monovalent and bivalent cations on the surface charge of opal. Cem Concr Res 31:1549-1552 Rogers CA (1986) Alkali-aggregate reactions in Ontario. In: Procs 7th ICAAR. Grattan-Bellew PE (ed) Ottawa, Canada, p 5-9 Rooney TP, Jones BF, Neal JT (1969) Magadiite from Alkali Lake, Oregon. Am Mineral 54:1034-1043 Rosenberg PE, Kittrick JA, Alldredge JR (1984) Composition of the controlling phase in muscovite equilibrium solubility. Clays Clay Miner (32/6):480-482 Rosenberg PE, Kittrick JA (1990) Muscovite dissolution at 25°C: implications for illite/smectite-kaolinite stability relations. Clays Clay Miner 38(4):445-447 Rosenholtz JL, Smith DT (1941) Linear thermal expansion and inversions of quartz, var. rock crystal. Am Mineral 26:103-109 Rudashevsky NS, Burakov BE, Lupal SD, Thalhammer OAR, Saini-Eidukat B (1995) Liberation of accessory minerals from various rock types by electric-pulse disintegration – method and application. Trans Inst Min Metall Sect C 104:25-29 Rufe E, Hochella MF (1998) Characterization of mica dissolution in the presence of bulk and confined solutions: implications for natural weathering processes. Mineral Mag 62a:1299-1300 Rykart R (1995) Quarz-Monographie, 2nd edition. Ott Verlag, Thun, Switzerland Saccani A, Bignozzi MC (2010) ASR expansion behavior of recycled glass fine aggregates in concrete. Cem Concr Res 40(4):531-536 Saini-Eidukat B, Weiblen PW (1996) A new method of fossil preparation, using high-voltage electric pulses. Curator 39:139-144 Samson SD, Nagy KL, Cotton WB III (2005) Transient and quasi-steady-state dissolution of biotite at 22-25 °C in high pH, sodium, nitrate, and aluminate solutions. Geochim Cosmochim Acta 69(2):399-413 Samuel G, Mullick AK, Ghosh SP, Wason RC (1984) Alkali-silica reaction in concrete - SEM and EDAX analyses. In: Procs 6th Intl Conf Cem Microscopy, Albuquerque, New Mexico, p 276-291 Schneiderhöhn H (1958-1961) Die Erzlagerstätten der Erde. Band I-II. 1st edition. Gustav Fischer Verlag, Stuttgart Schoondheydt RA, Bergaya F (2010) Industrial clay minerals as nanomaterials. In: Advances in the characterization of industrial minerals. Christidis GE (ed), EMU Notes Mineral 9:415-440 Scrivener KL, Monteiro PJM (1994) The alkali-silica reaction in a monolithic opal. J Am Ceram Soc 77:28492856 Scrivener K, Kirkpatrick RJ (2008) Innovation in use and research on cementitious material. Cem Concr Res 38:128-136 Sergi G, Page CL, Thompson DM (1991) Electrochemical induction of alkali-silica reaction in concrete. Mater Struct 24: 359-361 Shao Y, Lefort T, Moras S, Rodriguez D (2000) Studies on concrete containing ground waste glass. Cem Concr Res 30:91-100 Shayan A (ed) (1996) Procs 10th ICAAR, Melbourne, Australia, p 1-1063 Shayan A (1998) Effects of NaOH and NaCl solutions and temperature on the behaviour of specimens subjected to accelerated AAR tests. Cem Concr Res 28(1):25-31 Shayan A, Xu A, Chirgwin G, Morris H (2010) Effects of seawater on AAR expansion of concrete. Cem Concr Res 40(4):563-568 Sheppard RA, Gude AJ, Hay RL (1970) Makatite, a new hydrous sodium silicate from Lake Magadi, Kenya. Am Mineral 55:358-366 Sherwood WC, Newlon HH Jr (1964) A survey for reactive carbonate aggregate grains in Virginia. Hwy Res Board, Record 45:1-222

Deleterious Reactions of Aggregate With Alkalis in Concrete

361

Shoval S, Ginott Y, Nathan Y (1991) A new method for measuring the crystallinity index of quartz by infrared spectroscopy. Mineral Mag 55:579-582 Shrimer F (2000) Application and use of damage rating index in assessment of AAR-affected concrete – selected case studies. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec City, QC, Canada, p 899-907 Shrimer F (2006) Development of the damage rating index method as a tool in the assessment of alkali-aggregate reaction in concrete: a critical review. In: Marc-André Bérubé Symposium On Alkali-Aggregate Reactivity in Concrete. Fournier B (ed) Montréal, Canada, p 391-411 Shrimer F, Briggs A, Hudson B (2008) Alkali-aggregate reaction in Western Canada: review of current trends. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 11-20 Sibbick RG, Page CL (1987) Effects of sodium chloride on the alkali-silica reaction in hardened concrete. In: Procs 10th ICAAR. Shayan A (ed) Melbourne, Australia, p 822-829 Sibson RH (1975) Generation of pseudotachylite by ancient seismic faulting. Geoph J Roy Astron Soc 43:775794 Siedlecka A (1979) Length-slow chalcedony and relicts of sulfates - evidences of evaporitic environments in the Upper Carboniferous and Permian beds from Bear Island, Svalbard. In: Silica in sediments: nodular and bedded chert. McBride EF (ed) Soc Econ Paleont Min, Reprint Series 8:73-77 Siegesmund S, Weiss T, Vollbrecht A (eds) (2003) Natural stone, weathering phenomena, conservation and case studies. Geological Society London, Spec Pub 205:1-496 Siivola J (1969) On the evaporation of some alkali metals during the electron microprobe analysis. Geol Soc Finland Bull 41:85-91 Sims I, Poole AB (1978) Potential alkali-reactive aggregates from the Middle East. In: Procs 4th ICAAR. Diamond S (ed) West Lafayette, Indiana, p 359-361 Sims I, Sotiropoulos P (1983) Standard alkali-reactivity testing of carbonate rocks from the Middle East and North Africa. In: Procs 6th ICAAR. Idorn GM, Rostam S (eds) Copenhagen, Denmark, p 337-350 Sims I, Hunt B, Maglio B (1992) Quantifying microscopical examinations of concrete for alkali aggregate reaction (AAR) and other durability aspects. Am Conc Inst Spec Pub 131-114:267-287 Smith RK, Tremallo RL, Lofgren GE (2001) Growth of megaspherulites in a rhyolitic vitrophyre. Am Mineral 86:589-600 Smykatz-Kloss W (1970) Die Hoch-Tiefquarz-Inversion als petrologisches Hilfsmittel. Contrib Mineral Petrol 26:20-41 Smykatz-Kloss W (1972) Das Hoch-Tiefquarz-Umwandlungsverhalten mikrokristalliner Quarze. Contrib Mineral Petrol 36:1-18 Snellings R, Mertens G, Elsen J (2012) Supplementary cementitious materials. Rev Mineral Geochem 74:211278 Sommer H, Grattan-Bellew PE, Katayama T, Tang M (2004) Development and inter-laboratory trial of the RILEM AAR-5 rapid preliminary screening test for carbonate aggregates. In: Procs 12th ICAAR. Deng M, Tang M (eds) Beijing, China, p 407-412 Spray JG (1998) Localized shock- and friction-induced melting in response to hypervelocity impact. In: Meteorites: flux with time and impact effects. Grady MM, Hutchison R, McCall GJH, Rothery DA (eds) Geol Soc London, Spec Pub 140:195-204 Spray JG, Rae DA (1995) Quantitative electron-microprobe analysis of alkali silicate glasses: A review and user guide. Can Mineral 33:323-332 Stanton TE (1940a) Influence of cement and aggregate on concrete expansion. Eng News Rec 123(5):59-61 + 124:171-173 Stanton TE (1940b) Expansion of concrete through reaction between cement and aggregate. Proc Am Soc Civ Eng 66:1781-1811 Stark J, Erfurt D, Freiburg E, Giebson C, Seyfarth K, Wicht B (2008) Alkali-Kieselsäure-Reaktion. FA FingerInstitut für Baustoffkunde, Bauhaus-Universität Weimar, Schriftenreihe 3:139p Št’astná A, Šachlová Š, Pertold Z, Přikryl R, Leichmann J (2012) Cathodoluminescence microscopy and petrographic image analysis of aggregates in concrete pavements affected by alkali-silica reaction. Mater Charact 65:115-125 Stevenson JS, Stevenson LS (1965) The petrology of dawsonite at the type locality, Montréal. Can Mineral 8:249-252 Stipp SLS, Gutmannsbauer W, Lehmann T (1996) The dynamic nature of calcite surfaces in air. Am Mineral 81:1-8 St. John DA, Goguel RL (1992) Pore solution/aggregate enhancement of alkalies in hardened concrete. In: Procs 9th ICAAR. Poole AB (ed) London, United Kingdom, p 894-901 St. John DA, Poole AB, Sims I (1998) Concrete Petrography, A Handbook of Investigative Techniques. 1st Edition. Arnold Publishers, London

362

Broekmans

Stoica G, Perez-Ramirez J (2010) Stability and inter-conversion of synthetic dawsonites in aqueous media. Geochim Cosmochim Acta 74:7048-7058 Strandh H, Petterson LGM, Sjöberg L, Wahlgren U (1997) Quantum chemical studies of the effects on silicate mineral dissolution rates by adsorption of alkali metals. Geochim Cosmochim Acta 61:2577-2587 Strogen P (1993) Study of cherts in fine and coarse aggregates in Ireland. Concrete 27(3):26-29 Strunge H, Chatterji S, Damgård Jensen A (1991) Studies of alkali-silica reaction. Part 8. Correlation between mortar-bar expansion and Δ values. Cem Concr Res 21:61-65 Suryavanshi AK, Scantlebury JD, Lyon SB (1996) Mechanism of Friedel’s salt formation in cements rich in tricalcium aluminate. Cem Concr Res 26:717-727 Swenson EG (1957a) Cement aggregate reaction in concrete of a Canadian bridge. ASTM Procs 57:1043-1056 Swenson EG (1957b) A reactive aggregate undetected by ASTM test. ASTM Bull 226:48-50 Swenson EG, Gillott JE (1960) Characteristics of Kingston carbonate rock reaction. Hwy Res Brd, Bull 275:1831 Swenson EG, Legget RF (1960) Kingston study of cement-aggregate reaction. Can Consultancy Eng 2:38-46 Swenson EG, Gillott JE (1964) Alkali–carbonate rock reaction. Hwy Res Record 45:21-40 Swenson EG, Gillott JE (1967) Alkali reactivity of dolomitic limestone aggregate. Mag Concr Res 19:95-104 Taijing L, Zhang X, Sunagawa I, Groves GW (1995) Nanometre scale textures in agate and Beltane opal. Mineral Mag 59:103-109 Talvitie NA (1951) Determination of quartz in the presence of silicates using phosphoric acid. Anal Chem 23:623-626 Tambelli CE, Schneider JF, Hasparyk NP, Monteiro PJM (2006) Study of the structure of alkali-silica reaction gel by high-resolution NMR spectroscopy. J Non-Cryst Solids 352:3429-3436 Tang M, Deng, M (eds) (2004) Procs 12th ICAAR, Beijing, China:1-1298+appendices Tang M, Xue WR (1962) Expansive reaction of cement and reactive silica in aggregate. (in Chinese) J Chin Silicate Soc 1/1:20-29 Taylor HFW (1989) Modification of the Bogue calculation. Adv Cem Res 2/6:73-77 Taylor HFW (1997) Cement Chemistry. 2nd edition, Thomas Telford, London. Teng FZ, McDonough WF, Rudnick RL, Dalpé C, Tomascak PB, Chappell BW, Gao S (2004) Lithium isotopic composition and concentration of the upper continental crust. Geochim Cosmochim Acta 68(20):41674178 Teza G, Calgaro A, Moro F (2009) Contactless recognition of concrete surface damage from laser scanning and curvature computation. Non-Dest Test Eng Int 42:240-249 Thaulow N, Knudsen T (1975) Quantitative microanalyses of the reaction zone between cement paste and opal. In: Procs 2nd ICAAR. Ásgeirsson H (ed), Reykjavik, Iceland, p 189-203 Thaulow N, Jakobsen UH, Clark B (1996) Composition of alkali silica gel and ettringite in concrete railroad ties: SEM-EDX and X-ray diffraction analyses. Cem Concr Res 26:309-318 Thomas MDA (1998) The role of calcium in alkali-silica reaction. In: Materials Science of Concrete: The Sidney Diamond Symposium, Special Volume. Cohen M, Mindess S, Skalny J (eds) Am Ceram Soc, Westerville, Ohio, p 325-337 Thomas M, Beaman N, Hayman S, Gilks P (2008) Evaluating ASR prevention strategies for the reconstruction of concrete structures at Mactaquac generating station. In: Procs 13th ICAAR. Broekmans MATM, Wigum BJ (eds) Trondheim, Norway, p 1271-1280 Tong L, Tang M (1995) Correlation between reaction and expansion of alkali-carbonate reaction. Cem Concr Res 25:470-476 Tong L, Deng M, La X, Tang M (1997) A case study of two airport runways affected by alkali-carbonate reaction. Part one: Evidence of deterioration and evaluation of aggregates. Cem Concr Res 27(3):321-328 Tong L, Tang M (1997) A case study of two airport runways affected by alkali-carbonate reaction. Part two: Microstructural investigations. Cem Concr Res 27(3):329-336 Tong L, Tang M (1999) Expansion mechanism of alkali-dolomite and alkali-magnesite reaction. Cem Concr Comp 21:361-373 Toriya S, Tamura Y, Takei T, Fuji M, Watanabe T, Chikazawa M (2002) Preparation and characterization of silica-pillared derivatives from kanemite. J Colloid Interface Sci 255:171-176 Tremblay C, Bérubé MA, Fournier B, Thomas MD, Folliard KJ (2010) Experimental investigation of the mechanisms by which LiNO3 is effective against ASR. Cem Concr Res 40(4):583-597 Trojer F, Blümel O (1969) Beitrag zur Alkali-Dolomit-Reaktion im Beton. Zement-Kalk-Gips 4:175-178 Tulloch AJ (1979) Secondary Ca-Al silicates as low-grade alteration products of granitoid biotite. Contrib Mineral Petrol 69:105-117 Van Aardt JHP, Visser S (1977a) Formation of hydrogarnets: calcium hydroxide attack on clays and feldspars. Cem Concr Res 7:39-44 Van Aardt JHP, Visser S (1977b) Calcium hydroxide attack on feldspars and clays: possible relevance to cementaggregate reactions. Cem Concr Res 7:643-648

Deleterious Reactions of Aggregate With Alkalis in Concrete

363

Van Aardt JHP, Visser S (1978) Reaction of Ca(OH)2 and of Ca(OH)2 + Ca(SO4)⋅2H2O at various temperatures with feldspars in aggregates used for concrete mixing. Cem Concr Res 8:677-682 Vandervoort GF (1999) Metallography, principles and practice. Am Soc Metals (ASM) Intl, Metals Park, Ohio Varma A (1991) Handbook of Inductively Coupled Plasma Atomic Emission Spectroscopy. CRC Press, Boca Raton, Florida Varshneya AK, Cooper AR, Cable M (1966) Changes in composition during electron microprobe analysis of K2O-SrO-SiO2 glass. J Appl Phys 37:2199-2202 Velbel MA (1984) Weathering processes of rock-forming minerals. MAC Short Course Handbook 10:67-111 Villeneuve V, Fournier B, Tremblay S, Fecteau PL, Desormeaux C, Sanchez L (2010) Description des indices pétrographiques pour analyse pétrographique semi-quantitative selon la méthode du Damage Rating Index. Report de la Faculté des Sciences et de Génie, Département de Géologie et de Génie Géologique, Université Laval, Québec City, QC, Canada: p49 Vivian HE (1950) The reaction product of alkalies and opal. Studies in cement-aggregate reaction. CSIRO Australia, Bull 256:60-81 Vogels RJMJ (1996) Non-hydrothermally synthesized trioctahedral smectites. Preparation, characterisation and catalytic properties of synthetic clay minerals. PhD-thesis University of Utrecht, Faculty of Chemistry: p191 Vortmann S, Rius J, Marler B, Gies, H (1999) Structure solution from powder data of the hydrous layer silicate kanemite, a precursor of the industrial ion exchanger SKS-6. Eur J Mineral 11(1):125-134 Wakizaka Y (2000) Alkali-silica reactivity of Japanese rocks. Eng Geol 56:211-221 Walker TM, Howitt DG (1989) Field-induced migration of sodium in soda silicate glasses during scanning electron microscopy. Scanning 11:5-11 Wang H, Gillott JE (1991) Mechanism of alkali-silica reaction and significance of calcium hydroxide. Cem Concr Res 21:647-654 Wang Y, Deng M, Tang M (2008) Alkali release from aggregate and the effect on AAR expansion. Mater Struct 41:159-171 Wark DA, Watson BE (2006) TitaniQ: a titanium-in-quartz geothermometer. Contrib Mineral Petrol 152(6):743754 Warne SStJ (1962) A quick field or laboratory staining scheme for the differentiation of major carbonate minerals. J Sediment Petrol 32(1):29-38 Way SJ, Cole WF (1982) Calcium hydroxide attack on rocks. Cem Concr Res 12:611-617 Way SJ, Shayan A (1993) Synthesis and characterization of crystalline analogues of alkali-aggregate reaction products. Cem Concr Res 23:471-479 Weaver CE (1958) The effects and geologic significance of potassium “fixation” by expandable clay minerals derived from muscovite, biotite, chlorite, and volcanic material. Am Mineral 48:859-861 Wedepohl KH (ed) (1969-78) Handbook of Geochemistry. Springer Verlag, Berlin Wedepohl KH (1995) The composition of the continental crust. Geochim Cosmochim Acta 59(7):1217-1232 Weissbart EJ, Rimstidt JD (2000) Wollastonite: incongruent dissolution and leached layer formation. Geochim Cosmochim Acta 64:4007-4016 White AF, Brantley SL (eds) (1995) Chemical weathering rates of silicate minerals. Reviews in Mineralogy, Volume 31. Mineralogical Society of America Wieker H, Hubert C, Heidemann D (1997) Recent results of solid-state NMR investigations and their possibilities of use in cement chemistry. In: Procs 10th Intl Congr Chem Cem. Justnes H (ed) Göteborg, Sweden, p 2-24 Wieker W, Hubert C, Heidemann D, Ebert R (1998) Alkali-silica reaction - a problem of the insufficient fundamental knowledge of its chemical base. In: Materials Science of Concrete: The Sidney Diamond Symposium, Special Volume. Cohen M, Mindess S, Skalny J (eds) Am Ceram Soc Westerville, Ohio:395408 Wieker H, Hubert C, Heidemann D, Ebert R (2000): Some experiences in chemical modelling of the alkali-silica reaction. In: Procs 11th ICAAR. Bérubé MA, Fournier B, Durand B (eds) Québec, Canada, p 119-128 Wigum BJ (1995) Alkali-aggregate reactions in concrete: properties, classification and testing of Norwegian cataclastic rocks. PhD thesis, Norwegian University of Science and Technology NTNU, Trondheim: p227 Wigum BJ, Haugen M, Skjølsvold O, Lindgård J (2004) Norwegian petrographic method - Development and experiences during a decade of service. In: Procs 12th ICAAR. Tang M, Deng M (eds) Beijing, China, p 444-452 Wijnen PWJG (1990) A spectroscopic study of silica gel formation from aqueous silicate solutions. PhD Thesis Tech Uni Eindhoven/NL: pp133 Wolf KH, Chilingar GV, Beales FW (1967) Elemental composition of carbonate skeletons, minerals, and sediments. In: Developments in sedimentology (9b), Carbonate rocks, physical and chemical aspects. Chilingar GV, Bissell HJ, Fairbridge RW (eds) Elsevier Publishing Company, Amsterdam, p 23-149

364

Broekmans

Wolff-Boenisch D, Gíslason SR, Oelkers EH, Putnis CV (2004) The dissolution rates of natural glasses as a function of their composition at pH 4 and 10.6 and temperatures from 25 to 74 °C. Geochim Cosmochim Acta 68:4843-4858 Wong HS, Buenfeld NR (2006) Monte Carlo simulation of electron-solid interactions in cement-based materials. Cem Concr Res 36:1076-1082 Wood JGM (2007) The application of forensic investigation to improving construction and maintenance of structures. In: Procs Intl Conf Forensic Eng: Failure, Diagnosis and Problem Solving, India Chapter of ACI (ICACI), Mumbai, India: 9p Xu H, Heaney PJ (1997) Memory effects of domain structures during displacive phase transitions: A hightemperature TEM study of quartz and anorthite. Am Mineral 82:99-108 Zinovyev NT, Siomkin BV, Tanbayev ZG (1988) On concrete demolition by electric discharges. In: Demolition and Reuse of Concrete and Masonry. Kasai Y (ed) Chapman and Hall, London, p 300-305 Zinovyev NT (1989) Electric pulse fracture of sub-standard concrete. (in Russian) Obogashcheniye Rud 34(4):43-45

RiMG Volume 74

Applied Mineralogy of Cement & Concrete CONTENTS 1-82

Calcium aluminate cements – raw materials, differences, hydration and properties. Pöllmann

83-99

Alternative low-CO2 “green” clinkering processes.

101-145

Justnes

Microscopy of clinker and hydraulic cements. Stutzman

147-167

Industrial x-ray diffraction analysis of building materials. Meier et al.

169-209 Rietveld quantitative phase analysis of OPC clinkers, cements and hydration products. Aranda et al. 211-278

Supplementary cementitious materials. Snellings et al.

279-364

Deleterious reactions of aggregate with alkalis in concrete. Broekmans

ISBN 978-0-939950-88-1

9 780939

950881