Liquid Chromatography: Fundamentals and Instrumentation 9780124158078, 1865843830, 0124158072

Table of contents :
Front Cover......Page 1
Liquid Chromatography: Fundamentals and Instrumentation......Page 4
Copyright......Page 5
Contents......Page 6
Contributors......Page 12
Chapter 1 - Milestones in the Development of Liquid Chromatography......Page 14
1.1.INTRODUCTION......Page 15
1.2.HPLC THEORY AND PRACTICE......Page 17
1.3.COLUMNS......Page 19
1.4.EQUIPMENT......Page 22
1.5.DETECTORS......Page 25
APOLOGIES AND ACKNOWLEDGEMENTS......Page 26
References......Page 27
Chapter 2 - Kinetic Theories of Liquid Chromatography......Page 32
2.2.MACROSCOPIC KINETIC THEORIES......Page 33
2.3.MICROSCOPIC KINETIC THEORIES......Page 43
2.4.COMPARISON OF THE MICROSCOPIC AND MACROSCOPIC MODELS......Page 51
References......Page 52
Chapter 3 - Column Technology in Liquid Chromatography......Page 54
3.1.INTRODUCTION......Page 55
3.2.COLUMN DESIGN AND HARDWARE......Page 57
3.3.COLUMN PACKING MATERIALS AND STATIONARY PHASES......Page 62
3.4.COLUMN SYSTEMS AND OPERATIONS......Page 76
3.5.CHROMATOGRAPHIC COLUMN TESTING AND EVALUATION......Page 80
3.6.COLUMN MAINTENANCE AND TROUBLESHOOTING......Page 83
3.7.TODAY’S COLUMN MARKET–AN EVALUATION, COMPARISON, AND CRITICAL APPRAISAL......Page 87
References......Page 96
Chapter 4 - Secondary Chemical Equilibria in Reversed-Phase Liquid Chromatography......Page 100
4.1.INTRODUCTION......Page 101
4.2.ACID–BASE EQUILIBRIA......Page 102
4.3.ION-INTERACTION CHROMATOGRAPHY......Page 103
4.4.MICELLAR LIQUID CHROMATOGRAPHY......Page 110
4.5.METAL COMPLEXATION......Page 113
Further Reading......Page 116
5.1.INTRODUCTION......Page 118
5.2.PRINCIPLES OF HILIC......Page 120
5.3.MOBILE AND STATIONARY PHASES COMMONLY EMPLOYED IN HILIC......Page 123
5.4.APPLICATION EXAMPLES......Page 127
References......Page 129
Chapter 6 - Hydrophobic Interaction Chromatography......Page 134
6.2.PRINCIPLES OF HYDROPHOBIC INTERACTION CHROMATOGRAPHY......Page 135
6.3.MAIN FACTORS THAT AFFECT HYDROPHOBIC INTERACTION CHROMATOGRAPHY......Page 138
6.4.PURIFICATION STRATEGIES......Page 143
6.5.PRACTICAL ASPECTS OF HYDROPHOBIC INTERACTION CHROMATOGRAPHY PURIFICATION......Page 144
6.6.SELECTED APPLICATIONS......Page 145
6.7.FUTURE TRENDS......Page 148
References......Page 149
7.1.INTRODUCTION......Page 156
7.2.RETENTION AND SEPARATION......Page 157
7.3.METHOD DEVELOPMENT......Page 162
7.4.PROBLEMS IN THE USE OF NORMAL-PHASE CHROMATOGRAPHY......Page 167
References......Page 168
8.1.INTRODUCTION......Page 170
8.2.BASIC PRINCIPLES AND SEPARATION MODES......Page 171
8.3.INSTRUMENTATION......Page 178
8.4.APPLICATIONS......Page 198
References......Page 201
Chapter 9 - Size-Exclusion Chromatography......Page 206
9.2.HISTORICAL BACKGROUND......Page 207
9.3.RETENTION IN SIZE-EXCLUSION CHROMATOGRAPHY......Page 210
9.4.BAND BROADENING IN SIZE-EXCLUSION CHROMATOGRAPHY......Page 215
9.5.RESOLUTION IN SIZE-EXCLUSION CHROMATOGRAPHY......Page 218
9.6.SIZE-EXCLUSION CHROMATOGRAPHY ENTERS THE MODERN ERA: THE DETERMINATION OF ABSOLUTE MOLAR MASS......Page 220
9.7.SIZE-EXCLUSION CHROMATOGRAPHY TODAY: MULTIDETECTOR MEASUREMENTS, PHYSICOCHEMICAL CHARACTERIZATION, TWO-DIMENSIONAL TECHNIQUES......Page 230
9.8.CONCLUSIONS......Page 232
References......Page 233
Chapter 10 - Solvent Selection in Liquid Chromatography......Page 238
10.1.ELUTION STRENGTH......Page 239
10.2.COLUMNS AND SOLVENTS IN RPLC, NPLC, AND HILIC......Page 241
10.3.ASSESSMENT OF THE ELUTION STRENGTH......Page 242
10.4.SCHOENMAKERS’S RULE......Page 244
10.5.ISOELUOTROPIC MIXTURES......Page 246
10.6.SOLVENT-SELECTIVITY TRIANGLES......Page 247
10.7.PRACTICAL GUIDELINES FOR OPTIMIZATION OF MOBILE PHASE COMPOSITION......Page 254
10.8.ADDITIONAL CONSIDERATIONS FOR SOLVENT SELECTION......Page 259
References......Page 261
Chapter 11 - Method Development in Liquid Chromatography......Page 264
11.1.INTRODUCTION......Page 265
11.3.A STRUCTURED APPROACH TO METHOD DEVELOPMENT......Page 266
11.4.METHOD DEVELOPMENT IN PRACTICE......Page 271
11.5.PREVALIDATION......Page 276
11.6.VALIDATION......Page 277
11.7.DOCUMENTATION......Page 278
11.8.SUMMARY......Page 279
References......Page 280
12.1.INTRODUCTION......Page 282
12.2.THE EFFECTS OF EXPERIMENTAL CONDITIONS ON SEPARATION......Page 285
12.3.METHOD DEVELOPMENT......Page 291
12.4.PROBLEMS ASSOCIATED WITH GRADIENT ELUTION......Page 294
References......Page 295
Chapter 13 - General Instrumentation......Page 296
13.1.INTRODUCTION......Page 297
13.3.PUMPING SYSTEMS......Page 299
13.4.GRADIENT-ELUTION MIXING SYSTEMS......Page 302
13.5.SAMPLE INJECTION......Page 304
13.6.COLUMN COMPARTMENT......Page 306
13.7.TUBINGS AND FITTINGS......Page 307
13.8.DETECTOR OVERVIEW......Page 309
13.9.ULTRAVIOLET–VISIBLE ABSORBANCE DETECTORS......Page 311
13.10.REFRACTIVE INDEX DETECTORS......Page 313
13.12.CHARGED AEROSOL DETECTORS......Page 314
13.14.FLUORESCENCE DETECTORS......Page 315
13.16.OTHER DETECTION METHODS......Page 316
References......Page 317
Chapter 14 - Advanced Spectroscopic Detectors for Identification and Quantification: Mass Spectrometry......Page 320
14.1.INTRODUCTION......Page 321
14.2.IONIZATION METHODS SUITABLE FOR LC COUPLING......Page 322
14.3.HOW TO INCREASE SPECIFICITY OF MS DATA......Page 330
14.4.MICRO- AND NANO-LC–MS......Page 334
14.5.CAPILLARY ELECTROCHROMATOGRAPHY......Page 338
References......Page 340
15.1.INTRODUCTION......Page 346
15.2.OFF-LINE HYPHENATION......Page 350
15.3.ON-LINE HYPHENATION......Page 352
15.4.CONCLUSIONS......Page 358
References......Page 359
Chapter 16 - Advanced Spectroscopic Detectors for Identification and Quantification: Nuclear Magnetic Resonance......Page 362
16.1.INTRODUCTION......Page 363
16.2.HYPHENATION OF NMR WITH HPLC......Page 364
16.3.ADVANCES IN NMR SENSITIVITY......Page 366
16.4.STRATEGIES FOR OBTAINING NMR INFORMATION FROM A GIVEN LC PEAK......Page 372
16.6.QUANTIFICATION CAPABILITIES......Page 381
16.7.FIELDS OF APPLICATION......Page 383
16.8.CONCLUSIONS......Page 392
References......Page 393
Chapter 17 - Quantitative Structure Property (Retention) Relationships in Liquid Chromatography......Page 398
17.2.METHODOLOGY AND GOALS OF QSRR STUDIES......Page 399
17.3.APPLICATIONS OF QSRR IN PROTEOMICS......Page 405
17.4.CHARACTERIZATION OF STATIONARY PHASES......Page 406
17.5.QSRR AND ASSESSMENT OF LIPOPHILICITY OF XENOBIOTICS......Page 408
17.6.QSRR ANALYSIS OF RETENTION DATA DETERMINED ON IMMOBILIZED-BIOMACROMOLECULE STATIONARY PHASES......Page 412
17.8.CHEMOMETRICALLY PROCESSED MULTIVARIATE CHROMATOGRAPHIC DATA IN RELATION TO PHARMACOLOGICAL PROPERTIES OF DRUGS AND “DRUG CA .........Page 413
17.9.CONCLUDING REMARKS......Page 414
References......Page 415
Chapter 18 - Modeling of Preparative Liquid Chromatography......Page 420
18.1.INTRODUCTION......Page 421
18.2.COLUMN MODEL......Page 422
18.3.ADSORPTION MODEL......Page 423
18.4.PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY......Page 428
References......Page 435
Chapter 19 - Process Concepts in Preparative Chromatography......Page 440
19.1.INTRODUCTION......Page 441
19.2.CLASSICAL ISOCRATIC DISCONTINUOUS ELUTION CHROMATOGRAPHY......Page 442
19.3.OTHER DISCONTINUOUS OPERATING CONCEPTS......Page 445
19.4.CONTINUOUS CONCEPTS OF PREPARATIVE CHROMATOGRAPHY......Page 450
19.5.OPTIMIZATION AND CONCEPT COMPARISON......Page 459
Acknowledgments......Page 461
References......Page 462
20.1.INTRODUCTION, DEFINITIONS, AND SCOPE......Page 466
20.2.ICROFLUIDIC SYSTEMS FOR SEPARATIONS......Page 468
20.3.COMMERCIAL INSTRUMENTATION......Page 472
References......Page 478
21.1.INTRODUCTION......Page 482
21.2.PRINCIPLES OF CAPILLARY ELECTROCHROMATOGRAPHY......Page 483
21.3.INSTRUMENTATION......Page 484
21.4.METHOD OPTIMIZATION IN CEC......Page 491
21.5.EXAMPLES OF SOME RECENT APPLICATIONS......Page 492
21.6.CONCLUSIONS AND FUTURE TRENDS......Page 499
References......Page 500
Index......Page 506

Citation preview

LIQUID CHROMATOGRAPHY: FUNDAMENTALS AND INSTRUMENTATION

This page intentionally left blank

LIQUID CHROMATOGRAPHY: FUNDAMENTALS AND INSTRUMENTATION

SALVATORE FANALI PAUL R. HADDAD COLIN F. POOLE PETER SCHOENMAKERS DAVID LLOYD

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO

Elsevier 225, Wyman Street, Waltham, MA 02451, USA The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands Copyright Ó 2013 Elsevier Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/ permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made Library of Congress Cataloging-in-Publication Data Application Submitted British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-415807-8 For information on all Elsevier publications visit our web site at store.elsevier.com Printed and bound in USA 13 14 15

10 9 8 7 6 5 4 3 2 1

Contents Contributors xi

1.

Milestones in the Development of Liquid Chromatography

1

L.R. SNYDER AND J.W. DOLAN

1.1. Introduction 2 1.2. HPLC Theory and Practice 4 1.3. Columns 6 1.4. Equipment 9 1.5. Detectors 12 Apologies and Acknowledgements References 14

13

2. Kinetic Theories of Liquid Chromatography

19

A. FELINGER AND A. CAVAZZINI

2.1. Introduction 20 2.2. Macroscopic Kinetic Theories 20 2.3. Microscopic Kinetic Theories 30 2.4. Comparison of the Microscopic and Macroscopic Models 38 References 39

3. Column Technology in Liquid Chromatography 41 K.K. UNGER, S. LAMOTTE AND E. MACHTEJEVAS

3.1. Introduction 42 3.2. Column Design and Hardware 44 3.3. Column Packing Materials and Stationary Phases 49 3.4. Column Systems and Operations 63 3.5. Chromatographic Column Testing and Evaluation 67 3.6. Column Maintenance and Troubleshooting 70 3.7. Today’s Column Marketean Evaluation, Comparison, and Critical Appraisal 74 References 83

v

vi

CONTENTS

4.

Secondary Chemical Equilibria in Reversed-Phase Liquid Chromatography 87 ´ LVAREZ-COQUE AND J.R. TORRES-LAPASIO ´ M.C. GARCI´A-A

4.1. Introduction 88 4.2. AcideBase Equilibria 89 4.3. Ion-Interaction Chromatography 90 4.4. Micellar Liquid Chromatography 97 4.5. Metal Complexation 100 Further Reading 103

5. Hydrophilic Interaction Liquid Chromatography 105 A. CAVAZZINI AND A. FELINGER

5.1. Introduction 105 5.2. Principles of HILIC 107 5.3. Mobile and Stationary Phases Commonly Employed in HILIC 5.4. Application Examples 114 References 116

110

6. Hydrophobic Interaction Chromatography 121 C.T. TOMAZ AND J.A. QUEIROZ

6.1. Introduction 122 6.2. Principles of Hydrophobic Interaction Chromatography 122 6.3. Main Factors that Affect Hydrophobic Interaction Chromatography 125 6.4. Purification Strategies 130 6.5. Practical Aspects of Hydrophobic Interaction Chromatography Purification 131 6.6. Selected Applications 132 6.7. Future Trends 135 References 136

7. LiquideSolid Chromatography 143 L.R. SNYDER AND J.W. DOLAN

7.1. Introduction 143 7.2. Retention and Separation 144 7.3. Method Development 149 7.4. Problems in the Use of Normal-Phase Chromatography References 155

8. Ion Chromatography

157

B. PAULL AND P.N. NESTERENKO

8.1. Introduction 157 8.2. Basic Principles and Separation Modes

158

154

vii

CONTENTS

8.3. Instrumentation 165 8.4. Applications 185 References 188

9. Size-Exclusion Chromatography 193 A.M. STRIEGEL

9.1. Introduction 194 9.2. Historical Background 194 9.3. Retention In Size-Exclusion Chromatography 197 9.4. Band Broadening in Size-Exclusion Chromatography 202 9.5. Resolution in Size-Exclusion Chromatography 205 9.6. Size-Exclusion Chromatography Enters the Modern Era: The Determination of Absolute Molar Mass 207 9.7. Size-Exclusion Chromatography Today: Multidetector Measurements, Physicochemical Characterization, Two-Dimensional Techniques 217 9.8. Conclusions 219 Acknowledgement and Disclaimer 220 References 220

10.

Solvent Selection in Liquid Chromatography 225 ´ LVAREZ-COQUE G. RAMIS-RAMOS AND M.C. GARCI´A-A

10.1. Elution Strength 226 10.2. Columns and Solvents in RPLC, NPLC, and HILIC 228 10.3. Assessment of the Elution Strength 229 10.4. Schoenmakers’s Rule 231 10.5. Isoeluotropic Mixtures 233 10.6. Solvent-Selectivity Triangles 234 10.7. Practical Guidelines for Optimization of Mobile Phase Composition 10.8. Additional Considerations for Solvent Selection 246 References 248

11.

Method Development in Liquid Chromatography J.W. DOLAN AND L.R. SNYDER

11.1. Introduction 252 11.2. Goals 253 11.3. A Structured Approach to Method Development 11.4. Method Development in Practice 258 11.5. Prevalidation 263 11.6. Validation 264 11.7. Documentation 265 11.8. Summary 266 References 267

253

241

251

viii

CONTENTS

12.

Theory and Practice of Gradient Elution Liquid Chromatography 269 J.W. DOLAN AND L.R. SNYDER

12.1. Introduction 269 12.2. The Effects of Experimental Conditions on Separation 272 12.3. Method Development 278 12.4. Problems Associated with Gradient Elution 281 References 282

13.

General Instrumentation 283 J.G. SHACKMAN

13.1. Introduction 284 13.2. Solvent Source 286 13.3. Pumping Systems 286 13.4. Gradient-Elution Mixing Systems 289 13.5. Sample Injection 291 13.6. Column Compartment 293 13.7. Tubings and Fittings 294 13.8. Detector Overview 296 13.9. UltravioleteVisible Absorbance Detectors 298 13.10. Refractive Index Detectors 300 13.11. Evaporative Light-Scattering Detectors 301 13.12. Charged Aerosol Detectors 301 13.13. Conductivity Detectors 302 13.14. Fluorescence Detectors 302 13.15. Electrochemical Detectors 303 13.16. Other Detection Methods 303 References 304

14.

Advanced Spectroscopic Detectors for Identification and Quantification: Mass Spectrometry 307 S. CROTTI, I. ISAK AND P. TRALDI

14.1. Introduction 308 14.2. Ionization Methods Suitable for LC Coupling 309 14.3. How to Increase Specificity of MS Data 317 14.4. Micro- and Nano-LCeMS 321 14.5. Capillary Electrochromatography 325 References 327

15.

Advanced Spectroscopic Detectors for Identification and Quantification: FTIR and Raman 333 ´S J. KULIGOWSKI, B. LENDL, AND G. QUINTA

15.1. Introduction 333 15.2. Off-Line Hyphenation 337

ix

CONTENTS

15.3. On-Line Hyphenation 339 15.4. Conclusions 345 References 346

16.

Advanced Spectroscopic Detectors for Identification and Quantification: Nuclear Magnetic Resonance 349 J.-L. WOLFENDER, N. BOHNI, K. NDJOKO-IOSET AND A.S. EDISON

16.1. Introduction 350 16.2. Hyphenation of NMR with HPLC 351 16.3. Advances in NMR Sensitivity 353 16.4. Strategies for Obtaining NMR Information from a Given LC Peak 359 16.5. Integration with a Multiple Detection System (LCeNMReMS) 368 16.6. Quantification Capabilities 368 16.7. Fields of Application 370 16.8. Conclusions 379 Acknowledgments 380 References 380

17.

Quantitative Structure Property (Retention) Relationships in Liquid Chromatography 385 R. KALISZAN

17.1. Introduction 386 17.2. Methodology and Goals of QSRR Studies 386 17.3. Applications of QSRR in Proteomics 392 17.4. Characterization of Stationary Phases 393 17.5. QSRR and Assessment of Lipophilicity of Xenobiotics 395 17.6. QSRR Analysis of Retention Data Determined on Immobilized-Biomacromolecule Stationary Phases 399 17.7. Quantitative Retentione(Biological) Activity Relationships 400 17.8. Chemometrically Processed Multivariate Chromatographic Data in Relation to Pharmacological Properties of Drugs and “Drug Candidates” 400 17.9. Concluding Remarks 401 Acknowledgment 402 References 402

18. Modeling of Preparative Liquid Chromatography

407

T. FORNSTEDT, P. FORSSE´N AND J. SAMUELSSON

18.1. Introduction 408 18.2. Column Model 409 18.3. Adsorption Model 410 18.4. Process Optimization of Preparative Chromatography 18.5. Case Example 422 References 422

415

x

CONTENTS

19.

Process Concepts in Preparative Chromatography

427

M. KASPEREIT AND A. SEIDEL-MORGENSTERN

19.1. Introduction 428 19.2. Classical Isocratic Discontinuous Elution Chromatography 429 19.3. Other Discontinuous Operating Concepts 432 19.4. Continuous Concepts of Preparative Chromatography 437 19.5. Optimization and Concept Comparison 446 19.6. Conclusions 448 Acknowledgments 448 References 449

20.

Miniaturization and Microfluidics 453 F. FORET, P. SMEJKAL AND M. MACKA

20.1. Introduction, Definitions, and Scope 453 20.2. Microfluidic Systems for Separations 455 20.3. Commercial Instrumentation 459 20.4. Conclusion 465 Acknowledgment 465 References 465

21.

Capillary Electrochromatography

469

A. ROCCO, G. D’ORAZIO, Z. ATURKI AND S. FANALI

21.1. Introduction 469 21.2. Principles of Capillary Electrochromatography 470 21.3. Instrumentation 471 21.4. Method Optimization in CEC 478 21.5. Examples of Some Recent Applications 479 21.6. Conclusions and Future Trends 486 References 487

Index 493

Contributors Z. Aturki Institute of Chemical Methodologies, Italian National Council of Research (IMC-CNR), Area della Ricerca di Roma I, Rome, Italy N. Bohni School of Pharmaceutical Sciences, EPGL, University of Geneva, University of Lausanne, Geneva, Switzerland A. Cavazzini Department of Chemistry and Pharmaceutical Sciences, University of Ferrara, Ferrara, Italy S. Crotti DAIS, Ca` Foscari University, Venezia, Italy G. D’Orazio Institute of Chemical Methodologies, Italian National Council of Research (IMC-CNR), Area della Ricerca di Roma I, Rome, Italy J.W. Dolan LC Resources, Inc., Walnut Creek, CA, USA A.S. Edison Department of Biochemistry and Molecular Biology and National High Magnetic Field Laboratory, University of Florida, Gainesville, Florida S. Fanali Institute of Chemical Methodologies, Italian National Council of Research (IMC-CNR), Area della Ricerca di Roma I, Rome, Italy A. Felinger Department of Analytical and Environmental Chemistry, University of Pe´cs, Pe´cs, Hungary F. Foret Institute of Analytical Chemistry ASCR, Brno, Czech Republic T. Fornstedt Department of Engineering and Chemical Sciences, Karlstad University, Karlstad, Sweden; Analytical Chemistry, Department of Chemistry - BMC, Uppsala University, Uppsala, Sweden P. Forsse´n Department of Engineering and Chemical Sciences, Karlstad University, Karlstad, Sweden ´ lvarez-Coque Department of Analytical Chemistry, University of M.C. Garcı´a-A Vale`ncia, Spain I. Isak CNR-ISTM, Corso Stati Uniti 4, Padova, Italy R. Kaliszan Department of Biopharmaceutics and Pharmacodynamics, Medical University of Gda nsk and Department of Biopharmacy, Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University in Toru n, Gda nsk, Poland M. Kaspereit Friedrich-Alexander-Universita¨t Erlangen-Nu¨rnberg, Lehrstuhl fu¨r Thermische Verfahrenstechnik, Erlangen, Germany J. Kuligowski Department of Analytical Chemistry, University of Valencia, Edificio Jero´nimo Mun˜oz, Burjassot, Spain S. Lamotte Competence Center Analytics, Ludwigshafen, Germany B. Lendl Institute of Chemical Technologies and Analytics, Vienna University of Technology, Vienna, Austria

xi

xii

CONTRIBUTORS

E. Machtejevas Product Managment Analytical Chemistry, Merck Millipore, Merck KGaA, Darmstadt, Germany M. Macka University of Tasmania, Hobart, Australia K. Ndjoko-Ioset School of Pharmaceutical Sciences, EPGL, University of Geneva, Geneva, Switzerland P.N. Nesterenko ACROSS, University of Tasmania, Hobart, Tasmania, Australia B. Paull ACROSS, University of Tasmania, Hobart, Tasmania, Australia J.A. Queiroz Department of Chemistry, University of Beira Interior, Covilha˜, Portugal G. Quinta´s Leitat Technological Center, Bio In Vitro Division, Terrassa, Spain G. Ramis-Ramos Department of Analytical Chemistry, University of Valencia, Spain A. Rocco Institute of Chemical Methodologies, Italian National Council of Research (IMC-CNR), Area della Ricerca di Roma I, Rome, Italy J. Samuelsson Department of Engineering and Chemical Sciences, Karlstad University, Karlstad, Sweden A. Seidel-Morgenstern Otto-von-Guericke-Universita¨t Magdeburg, Institut fu¨r Verfahrenstechnik: Max-Planck-Institut fu¨r Dynamik Komplexer Technischer Systeme, Magdeburg, Germany J.G. Shackman Bristol-Myers Squibb, New Brunswick, NJ P. Smejkal Institute of Analytical Chemistry ASCR, Brno, Czech Republic; University of Tasmania, Hobart, Australia L.R. Snyder LC Resources, Inc., Walnut Creek, CA, USA A.M. Striegel Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, MD C.T. Tomaz CICS-UBIdHealth Sciences Research Centre, University of Beira Interior, Covilha˜, Portugal J.R. Torres-Lapasio´ Department of Analytical Chemistry, University of Vale`ncia, Spain P. Traldi CNR-ISTM, Corso Stati Uniti 4, Padova, Italy K.K. Unger Institute of Inorganic Chemistry and Analytical Chemistry, Johannes Gutenberg-University, Duesbergweg, Germany J.-L. Wolfender School of Pharmaceutical Sciences, EPGL, University of Geneva, University of Lausanne, Geneva, Switzerland

C H A P T E R

1

Milestones in the Development of Liquid Chromatography L.R. Snyder, J.W. Dolan LC Resources, Inc., Walnut Creek, CA, USA O U T L I N E 1.1. Introduction 1.1.1. Developments before 1960 1.1.2. HPLC at the Beginning

2 2 3

1.2. HPLC Theory and Practice 1.2.1. New HPLC Modes and Techniques 1.2.2. Selection of Conditions for the Control of Selectivity

4 5 5

1.3. Columns 1.3.1. Particles and Column Packing 1.3.2. Stationary Phases and Selectivity

6 6 9

1.4. Equipment

9

1.5. Detectors

12

Apologies and Acknowledgements

13

References

14

The importance of liquid chromatography (LC), and especially highperformance LC (HPLC), in today’s world hardly needs stating. It is the most widely used technique for the analysis of chemical mixtures and has contributed in a major way to science (especially the biological sciences) and everyday laboratory practice. Liquid chromatography is primarily a practical technique, so our story is limited to those innovations that Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00001-8

1

Copyright Ó 2013 Elsevier Inc. All rights reserved.

2

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

contributed significantly to its present use in “working” laboratories. In reflecting on the history of LC, it appears to us that only a few “essential” actors are in this drama: single individuals whose absence might have delayed the technique by more than a year or two. Thus, the development of present-day LC has largely been an evolutionary, rather than a revolutionary, process. Furthermore, many important innovations within the past 50 years have occurred within industrial research and development (R&D) groups, where it is often not possible to assign credit for a final product to a single person. Finally, every attempt at history suffers from incomplete and conflicting accounts of who did whatdand when. In the present “history,” we try to emphasize “what” and “when” rather than “who.”

1.1. INTRODUCTION Several important innovations in the history of liquid chromatography are reviewed by Ettre [1]: • Invention of chromatography in the early 1900s [2]. • Invention of partition and paper chromatography in the early 1940s [3]. • Development of ion-exchange chromatography [4] and the amino-acid analyzer during the 1950s [5]. • Invention of gel-filtration chromatography in the late 1950s [6,7]. • Development of the gel-permeation chromatograph in the early 1960s [8,9]. • Development of high-performance LC in the mid-1960s [8,10e17]. The present chapter emphasizes work on HPLC, while noting major, prior contributions that made this technique possible. Most advances in HPLC can be organized as follows: • Development or application of basic theory, combined with empirical observations of the separation process. • Invention of new chromatographic modes (e.g., ion-pair chromatography, hydrophilic-interaction chromatography) and the development of HPLC columns for new applications (chiral separation, large biomolecules). • Development and improvement of equipment and columns.

1.1.1. Developments before 1960 A good account of the discovery of chromatography by Tswett is given in [2] and [14, pp. 4e6]. Despite a few subsequent applications of chromatography in other laboratories [14, pp. 7e9], this technique became generally accepted only after its reintroduction in 1931 by Kuhn,

1.1. INTRODUCTION

3

Winterstein, and Lederer [18]. The invention of liquid-partition chromatography was reported by Martin and Synge in 1941, followed soon after by its extension to paper chromatography in 1944 and the first application of two-dimensional chromatography [3]. Significant work on ion-exchange separation began in the 1930s, with the subsequent development and application of ion-exchange chromatography (IEC) for separation of the rare earths and transuranium elements [4]. The extension of IEC to organic compounds was next, implemented by Cohn and Samuelson [14, pp. 17e21]. By 1958, Moore, Stein, and Spackman reported the automatic analysis of amino acid samples by means of IEC [5]. Their system was a precursor of HPLC that incorporated automatic pumping, efficient IEC columns, and continuous colorimetric detection. This system was later improved and commercialized as the Beckman-Spinco model 120B amino acid analyzer. Still another major development, in the later 1950s, was the invention of gel filtration [6,7] for the separation of large biomolecules by molecular size; this was followed a few years later by the development of gelpermeation chromatography (GPC) for the similar separation of synthetic polymers [9]. The latter then led to the development of a commercial GPC system by Waters Associates (the GPC-100 [8]), which would morph into an early commercial HPLC system (the ALC-100).

1.1.2. HPLC at the Beginning Prior to the development of the first HPLC systems, gas chromatography (GC) provided an example of what HPLC might be capable of: automation, speed, and detection sensitivity, as well as the separation of higher-boiling and thermally unstable compounds. By the early 1960s, the automation of LC had been demonstrated for amino-acid analysis and gel-permeation chromatography (Section 1.1.2). By then, it was appreciated that smaller particles in well-packed beds could increase both separation speed and efficiency. Fast separations with small-particle columns also require higher pressures to pump the mobile phase through the column. Detectors that could be used with LC for most samples presented a major challenge at this timedand for several years thereafter (Section 1.5). Before 1965, the possibility of using HPLC for separating samples other than amino acids or polymers had undoubtedly occurred to many people. However, the exploitation of this idea required its reduction to practice, followed by the production of commercially available equipment, as in the case of the amino-acid analyzer and the gel-permeation chromatograph (Section 1.1.2). Viewed in these terms, high-performance LC was first reduced to practice in ~1964, in the United States under the direction of Csaba Horva´th [16] and in Holland by Josef Huber (see [10, pp. 159e166

4

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

and 209e217]). The system developed by Horva´th was subsequently the basis for the LCS 1000 Nucleic Acid Analyzer sold by Picker Nuclear (later acquired by Varian) and contributed to the first general-purpose HPLC (Waters ALC-100) [8]. Jack Kirkland had visited Huber’s lab in 1964 and subsequently began an HPLC program at DuPont [19e22], which culminated in the Model 820 at about the same time as the ALC-100. By 1970, sales of HPLC systems were dominated by Waters Associates and Du Pont. Superficially porous Zipax [22] was initially the most popular column packing. In our opinion, Horva´th, Huber, and Kirkland can be considered the “fathers” of HPLC. Some closely related work at this time by others [23e29] is also relevant to the origin of HPLC. For a description of the columns, equipment, and practice at that time, see [30].

1.2. HPLC THEORY AND PRACTICE Separation as a function of experimental conditions was understood in general terms for GC, and similar principles were expected to apply to HPLC. Resolution, Rs, can be described by the Purnell equation [31] in terms of the plate number, N, separation factor, a, and retention factor, k: Rs ¼ 0:25ðN 0:5 Þða  1Þ ½k=ðk þ 1Þ

(1.1)

A semi-quantitative understanding of column efficiency (plate number, N) existed prior to 1965, based on the further development of the van Deemter equation for GC [32] and its extension to LC by Giddings [23]. Later work resulted in the highly useful and widely applied Knox equation [33]: h ¼ Av0:33 þ B=v þ Cv

(1.2)

where the reduced plate height, h, is related to the reduced velocity, v, of the mobile phase. The later development of “Poppe” (or “kinetic”) plots further advanced our understanding and use of column efficiency [34]. For further details on Eq. (1.2) and values of N, see Chapter 2 and [35]. When developing an HPLC procedure, the main challenge has been the selection of separation conditions for the control of peak spacing, that is, “optimum” values of a (Section 1.2.2). Basic theory played an important role in the development of HPLC, but its implementation was primarily the result of (a) the introduction of new separation modes or techniques (Section 1.2.1), (b) a better understanding of how best to vary conditions for a satisfactory separation (Section 1.2.2), and (c) improved columns (Section 1.3) and hardware (Sections 1.4 and 1.5).

1.2. HPLC THEORY AND PRACTICE

5

1.2.1. New HPLC Modes and Techniques Many of the separation “modes” used today in HPLC were described prior to 1960; for example, reversed-phase chromatography (RPC) was first used by Martin in 1950 [36]. While a few names are often associated with the rapid development of RPC over the past four decades (e.g., Horva´th, Kirkland, Knox), the present dominance of this technique can be attributed to the efforts of numerous practitioners in both industry and universitiesdas well as its inherent advantages. A history of the development of gradient elution is provided in [37e39]; the group of Tiselius is generally given priority for its first implementation and theoretical treatment in the 1950s. A practical understanding of gradient elution has since been facilitated by the linear-solvent-strength model [39]. The technique of ion-pair chromatography became a useful supplement to RPC for the separation of ionizable compounds that were often poorly retained by RPC. Schill and Knox are usually associated with the introduction of this mode [40,41]. Another technique for the separation of more polar compounds, such as sugars (which have poor retention in RPC), was used in the 1970s and later improved by Alpert [42] to become hydrophilic-interaction chromatography (HILIC). Separations of large biomolecules by HPLC required the development of suitable columns, featuring rigid, large-pore particles and lesshydrophobic stationary phases. While RPC has been used for separating proteins, these large, hydrophilic compounds are more often separated by gel filtration or ion exchange. Beginning about 1975, Chang, Gooding, and Regnier [43]; Kato et al. [44]; and others pioneered the development of columns for bio-HPLC. The first such column (SynChropak GPC100) was sold in 1978. Chiral separation was another area that awaited the development of suitable enantiospecific columns (Pirkle, Davankov, Okamoto, Armstrong, and others; see [35] for details). No more will be said about the latter columns for chiral separation, and very little about columns for biochromatography.

1.2.2. Selection of Conditions for the Control of Selectivity By the 1960s, it was appreciated that values of a for different pairs of solutes can be affected by the column, the mobile phase, and temperature. There are many different columns, and the mobile phase may vary according to the concentrations of its constituent solvents, various buffers, and additives, as well as pH. There are thus a very large number of possible combinations of these separation conditions; only gradually was it learned which conditions are the most useful for controlling a and how best to minimize the number of necessary experiments for a successful

6

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

separation. As the result of much practical experience and a few systematic studies (e.g., [45e48]), successful strategies for optimizing selectivity are now available for different samples (Chapters 11, 12, and [35]). Some noteworthy developments between 1970 and 2010 include • The introduction of resolution maps (for a or Rs as a function of temperature for IEC [49]). • Development of computer-assisted mapping of a in RPC as a function of mobile-phase mixtures of acetonitrile, methanol, and tetrahydrofuran [50]. • Development of computer-assisted mapping of Rs in either isocratic or gradient RPC as a function of (a) simultaneous change in two or more conditions that affect a and (b) all conditions that affect N (e.g., DryLab [35, Chapter 11]). • Development of a reliable procedure for characterizing column selectivity [51] and its use in various practical applications [52].

1.3. COLUMNS The development of HPLC depended on new columns, which in turn required new particles, new stationary phases (particle coatings), and improved procedures for packing the column. For details on column developments before 1994, see the review of Majors [53]. Table 1.1 summarizes several of these column innovations, with the dates of their introduction to the marketplace.

1.3.1. Particles and Column Packing In 1939, Martin noted that (a) small particles in well-packed beds would be needed for increased separation efficiency, and (b) such columns would require higher pressures to operate. Prior to 1960, particles for chromatography usually had diameters 100 mm and consisted of either polymeric spheres or irregular silica (prepared by sieving crushed silica). Polymeric materials typically gave lower plate numbers, so that inorganic oxides (mainly silica) became preferred for HPLC columns. HPLC columns at first used coated glass beads (e.g., Pellosil) or beads coated with a porous layer of silica (e.g., Zipax). The thickness of the stationary phase was only a fraction of the particle diameter, thereby reducing the diffusion distance within the stationary phase and yielding higher values of Ndbut with decreased loadability. Smaller, fully porous particles with a narrow size range were expected to provide more efficient columns (as well as superior loadability), but

7

1.3. COLUMNS

TABLE 1.1 Some Highlights in Commercial HPLC Column Development Datea

Column

Description

Company

1967

Pellosil

Pellicular ion exchange (40 mm)

Northgate

1969

Zipax

Porous-layer silica (40 mm)

DuPont

1971

MicroPak

Irregular porous silica (5e10 mm)

Varian

1972

Zorbax

Spherical porous silica (7 mm)

DuPont

1972

Permaphase

Silane phase (7 mm)

DuPont

1973

mBondapak C18

Silane phase (10 mm)

Waters

1978

SynChropak GPC100

Gel filtration column

SynChrom

1988

Rx-silica

Type-B silica

DuPont

1989

StableBond

Stable bonded phases

DuPont

1994

Hypercarb

Porous graphitic carbon

Hypersil

1996

ZirChrom PBD

Zirconia particles

Keystoneb

2000

SilicaRod

Monolith

Merck

1999

XTerra

Hybrid particles (3e5 mm)

Waters

2003

Rapid Resolution

Porous silica (1.8 mm)

Agilent

2004

Acquity

Porous silica (1.7mm)

Waters

2007

Halo

Shell particles (2.7 mm)

AMTc

a

Date of commercial introduction. Produced by Zirchrom, distributed by Keystone Scientific. Advanced Materials Technology. Source: Adapted from [53]. b c

such particles could not be produced by sieving. The introduction of air classification for particle sizing overcame this difficulty, and about 1970, Merck offered irregular silica in diameters of 5 and 10 mm. While larger particles were easily packed by tapping the column until the bed settled, packing particles of 10-mm diameter (or smaller) required a different approach. The first published procedure for reproducibly packing HPLC columns with particle diameters 10 mm used a balanced-density approach [54] similar to that used earlier for GPC columns; the resulting MicroPak columns were offered for sale by Varian in 1971. Over the years, many procedures have been described for packing particles with diameters 10 mm [35]; in practice, each new particle requires customized conditions for best results. About the same time, spherical particles with 7-mm diameters were produced by DuPont and sold under the name Zorbax. The latter particles were prepared by the aggregation of colloidal particles, much as popcorn balls are assembled from individual kernels of popcorn. Other procedures

8

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

have been developed for the manufacture of porous, spherical particles [55]. Today, most analytical columns use spherical particles. The silica used for HPLC columns before 1988 was usually derived from natural sources (water-glass solution) contaminated by various metals (e.g., iron, aluminum) that can increase its acidity. This often results in tailing peaks for basic compounds, poor column reproducibility, and other problems. These difficulties were largely overcome by preparing silica particles from the hydrolysis and polycondensation of pure tetraethoxysilane (TEOS). The latter, less-acidic silica is now referred to as type-B silica, in contrast to the older, less-pure, and more-acidic typeA silica. The first, widely used columns based on type-B silica were introduced by DuPont in 1988, and today most analytical columns use type-B silica. One disadvantage of bonded silica is that it is unstable at both low and high pH. Particles of porous, spherical, graphitic carbon [56] provided one answer to this problem; the resulting Hypercarb columns were offered in 1994 by Hypersil. Another approach to pH-stable particles is the use of porous, spherical zirconia in place of silica [57]. The first widely distributed column of this type (ZirChrom PBD) was introduced in 1996. Similar, but less dramatic improvements in column stability can be achieved by changes in bonding chemistry (Section 1.3.2). A quite different approach to HPLC columns was the development of so-called monoliths. Monolithic columns are cast as a porous, rigid cylinder by in situ polymerization and can be made from either silica or polymer. Several groups have contributed to their development, as reviewed in [58]. One advantage of monolithic columns is their greater permeability, allowing separations at lower pressures. While the first monolithic column (SilicaRod, Merck) was introduced for sale in 2000, the use of these columns (as of 2012) has remained somewhat limited. Hybrid particles are formed of an organic/inorganic structure, based on the reaction of TEOS and an alkyl triethoxysilane [59]. The resulting particles are more stable than their silica counterparts, with reduced acidity. Hybrid columns are therefore well suited for the separation of basic samples. The first hybrid column (XTerra, Waters) was offered for sale in 1999. From 1970 to 1990, the preferred particle size gradually evolved from 10 mm to about 3 mm with a resulting improvement in separation speed. The use of very small particles requires an increase in the maximum operating pressure of the equipment (>6000 psi). Equipment for separations at much higher pressures was first described by Rogers et al. [60], later perfected by MacNair, Lewis, and Jorgenson [61], and eventually commercialized by Waters (ultra-HPLC, or UPLC) in 2004. This enabled the use of 1.8-mm (Rapid Resolution, Agilent, 2003) and 1.7-mm fully porous particles (Acquity, Waters, 2004). A little later, so-called core-shell particles (Halo, AMT, 2007) were introduced for use at pressures

1.4. EQUIPMENT

9

6000 psi, but with similar performance as for smaller-particle columns and UHPLC systems. The latter shell particles consist of a nonporous core coated with a 0.5-mm layer of porous silica (total diameter 2.7 mm). Surface-coated particles of various kinds have played an important role in HPLC from the beginning; see the review of [62].

1.3.2. Stationary Phases and Selectivity Three stationary phases were used with surface-coated particles at the beginning of HPLC: an attached polymeric layer for ion exchange, a mechanically held liquid (liquid-liquid partition), and bare silica (adsorption). Liquid-liquid partition was initially most popular, but the mechanically held liquid proved unstable and operationally inconvenient. Various attempts were made to permanently bond an organic layer to a silica particle, with eventual success using a silicone polymer as stationary phase (marketed in 1972 as Permaphase; DuPont). Subsequently, organosilanes were used as reactants; for example, Cl-SiðCH3 Þ2 C18 þ hSiOH5hSiO-SiðCH3 Þ2 C18

(1.3)

(e.g., mBondapak, Waters, 1973). The latter approach has since been preferred for the preparation of RPC and other columns. One shortcoming of the original silane phases is their instability at both low and high pH, which can limit their application. The development of sterically protected stationary phases provided stability at low pH [63]; the first column of this kind (StableBond, DuPont) was introduced in 1989. These columns use hindered silanes, where the methyl groups in the silane of Eq. (1.3) are replaced by bulky groups such as i-butyl. Many other silanes have found use for HPLC stationary phases [35], for various purposes. The C18 group used initially has been followed by other ligands to yield phenyl, cyano, embedded-polar-group, and other columns. These column types can result in large differences in selectivity (Section 1.2.2), as well as provide other advantages (e.g., avoidance of stationary phase dewetting with highly aqueous mobile phases). For a review of column packings, including both the particle and stationary phase, see [17,55].

1.4. EQUIPMENT Many of the innovations in HPLC equipment remain as proprietary company secrets or are buried in the patent literature. Most of the dates listed here are estimates based on personal knowledge, interviews, and (limited) patent information. Additional information on instrumentation before 1980 can be found in [64].

10

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

Early HPLC systems contained most of the same components used today, but with less-sophisticated execution. A reservoir, pump, injector, column, detector, and data collection system were required. Reservoirs most commonly consisted of laboratory glassware. Microfiltration of the mobile phase (solvents) was used. Solvent outgassing was not a big problem until on-line mixing and gradient elution became more important. At first, solvents were degassed prior to use by means of vacuum or heating, followed in the late 1970s by the more reliable helium sparging (Spectra-Physics [65]), and later in-line vacuum degassing. Although complete HPLC systems were available from some vendors, many early workers built their own HPLC systems or purchased components and assembled HPLC systems from the “best” available modules. In the late 1960s and early 1970s, the Milton-Roy Mini-Pump formed the core of most home-built systems, as well as some commercial units. This was a single-piston pump with the stroke length adjusted to control the flow rate. It used a mechanical pressure gauge, whose Bourdon tube also acted as a pulse damper. Additional pulse dampening might be added with gas ballasts or additional Bourdon tubes. The Waters M-6000 (dual-reciprocating-piston ) pump was introduced in 1972 [17,66] and rapidly became the gold standard. With two variable-speed, reciprocating pistons operating out of phase, pulsations and system volume were reduced, with piston volumes in the 100-mL range. Nester-Faust (later acquired by Perkin Elmer) and Varian took a different approach, with large (250e1000 mL) pistons, one for each solvent, driven by screw drive. Depending on the solvent volume used per run, the pistons might need to be refilled after each run. Syringe pumps dropped from the market soon after a publication pointed out serious limitations in their use [67]: mobile-phase compressibility could result in poor reproducibility of retention times, especially in gradient elution. About 1976, Altex (later purchased by Beckman) introduced the Model 110A pump with a fast-refill feature [68]; its single piston could spend more than 50% of its duty cycle delivering solvent, thus reducing pump pulsations. A year or two earlier, Altex introduced the Model 100 pump for operation at 10,000 psi, far exceeding the 6000 psi upper limits of other systems and foreshadowing today’s UHPLC systems. Another innovation in the late 1970s was Perkin Elmer’s tandem-piston or accumulator-piston pump, where the first piston feeds solvent to the second piston. This design reduced both pulsations and the number of required check valves from four to three (or, in some cases, two) and improved reliabilitydthis feature is still common in many of today’s pumps. The earliest gradients were generated with a single pump; the weak solvent A was placed in a beaker on a stir plate, and the strong solvent B was added to the beaker (by siphoning) as its contents were delivered to the pump. More reliable, on-line mixing was subsequently accomplished

1.4. EQUIPMENT

11

by controlling the relative flow rate of two solvents, each pumped separately by a dedicated pump to a high-pressure mixer. In the late 1970s, Spectra-Physics introduced a three-solvent, low-pressure mixing system [69] coupled with its patented helium degasser [65]. This was the first practical low-pressure mixing system and allowed the simultaneous use of up to three solvents. At about the same time, an alternative multisolvent design was featured in Varian’s 5000 pumping system [70]; this design introduced each of three solvents directly into the pump head, minimizing bubble problems. These pumps also feature active inlet check valves, an innovation that eliminated most problems associated with these valves. As patents expired and licensing agreements were reached, the best of these featuresdlow-pressure mixing, accumulator-piston pumps, and active check valvesdbecame standard features on pumps from many manufacturers. Sample introduction was a challenge for early HPLC users. For lowerpressure operation, a septum-type injector could be used, but these tended to leak and were difficult to use; stop-flow injection represented an alternative. The Waters U6K [71] was introduced in 1973 and allowed convenient and reliable injection into the flowing stream. The U6K also included an innovative flow-bypass channel that reduced the pressure shock when the valve was cycled, a common source of column collapse with the less-robust columns in use at that time. At about the same time, Valco introduced six-port rotary injectors, a design that is now the industry standard. Rheodyne was formed soon after, and its injection valve [72] eventually became the industry leader. An automated valve inevitably led to an autosampler. One of the first widely used autosamplers, the Micromeritics model [73], used sealed vial caps that acted as a syringe plunger to deliver sample. A needle was inserted through the cap, and the cap was depressed so as to displace sample into the needle and sample loop. Several companies later developed autosamplers that used a motorized syringe to draw samples from a vial into the loop of a fixed-loop injector. This and a needle-in-loop design [74] are the most common autosampler configurations today. Although some of the early instruments (e.g., DuPont) included column ovens, many users did not consider column ovens necessary; others used hot water baths, heat tapes, or retired GC ovens to control the column temperature. Today’s ovens are designed specifically for HPLC, with either a resistance or Peltier heater to both maintain column temperature and preheat the mobile phase. Data were first collected on strip-chart recorders; every user had a favorite technique to keep the paper from jamming or pens from clogging, either of which could mean loss of data. Manual quantification was the rule, using either peak height or areadthe latter by triangulation, planimetry, or cutting out the peak from the paper and weighing it. Disc

12

1. MILESTONES IN THE DEVELOPMENT OF LIQUID CHROMATOGRAPHY

integrators were an innovation of the 1960s for GC and later provided some automation to the LC data-collection process. Spectra-Physics and Hewlett-Packard pioneered digital integrators in the mid-1970s, which revolutionized HPLC data collection. In the late 1970s Nelson Analytical introduced software that could integrate peaks automatically, even with drifting baselines; this software became part of data systems sold by many manufacturers. With the introduction of the IBM Personal Computer in 1981, the death knell rang for stand-alone integrators. Peak integration and system control gradually migrated to this now universally accepted computing platform. A recent advance in instrumentation is the development of ultrahighpressure LC (UHPLC) systems that exceed the traditional 6000 psi/400 bar pressure barrier; some systems (e.g., Shimadzu’s Nexera) offer pressures up to 19,000 psi/1300 bar. Waters introduced the UHPLC system in 2004, and many other manufacturers now supply competitive instrumentation. For optimal performance, these systems use sub-2-mm particles in short, narrow-diameter columns, which generate peaks of very small volume. Chromatographic band broadening due to noncolumn sources (especially injection, plumbing, detector configuration, and data processing) then became critical. The small-volume design of UHPLC systems (and many newer HPLC systems) reduce, but do not eliminate, the band-broadening problems that restricted the acceptance of 1.0

>1.0

>1.0

>1.0

>1.0

0.25

0.08

0.48 >1.0

>1.0

>1.0

>1.0

>1.0

>1.0

0.10

Ethyl ether

0.43 >1.0

>1.0

0.46

0.27

0.18

0.10

0.05

Hexane, heptane

0.00 0.35

0.20

0.07

0.03

0.02

0.01

0.00

0.70 >1.0

0.53

0.23

0.10

0.04

0.02

0.01

0.30 >1.0

>1.0

>1.0

>1.0

0.09

0.00

0.00

0.48 >1.0

0.70

0.54

0.45

0.28

0.10

0.05

0.60 >1.0

0.65

0.35

0.15

0.07

0.03

0.01

0.60 >1.0

0.44

0.20

0.11

0.05

0.03

0.02

0.53 >1.0

>1.0

0.60

0.40

0.21

0.18

0.09

Chloroform

b

Ethyl acetate

d,e

a,f

Methanol

c

Methylene chloride

d

Methyl-t-butyl ether f

n-Propanol

f

i- Propanol

d,e

Tetrahydrofuran a

Immiscible with hexane. Nonbasic localizing. c Nonlocalizing d Basic localizing. e Easily oxidized and therefore less useful in practice. f Very strong (localizing), proton-donor solvent; classification as “basic” or “nonbasic” may not be relevant. g Values for the pure solvent [2], derived as described in [1]. b

ester-anhydride]) [14]. Because the drug is more polar than the polymer, an assay by RPC would have required either prior separation of the polymer from the drug (because of very late elution of the polymer in RPC) or gradient elution. For this reason, NPC separation was explored, so that the polymer would leave the column before the drug, thereby precluding a need for either sample pretreatment or gradient elution. Initial studies were carried out by means of TLC with silica plates. The use of 100% methylene chloride yielded RF ¼ 0.00 for the drug, so the stronger B solvent methanol (MeOH, ε0 ¼ 0.70) was investigated, in mixture with methylene chloride. Mobile phases composed of 2e5% MeOH-CH2Cl2 appeared promising, and 1.5% MeOH-CH2Cl2 provided the acceptable HPLC separation of Figure 7.6(a), with UV detection at 240 nm. The method of Figure 7.6(a) was also intended for use with thermally stressed samples, so the experiments of Figures 7.6(b) and (c) were carried out. It appears that thermal degradation of the polymer, Figure 7.6(b),

154

7. LIQUIDeSOLID CHROMATOGRAPHY

Paclitaxel

Fresh sample

(a)

Polymer

(b)

Degraded polymer

(c)

0

Degraded sample

2

4

6

8

10

Time (min) FIGURE 7.6 NPC assay of paclitaxel in the presence of a polymeric additive. Conditions: 250
4.0-mm silica column (5-mm particles); 1.5% methanol-methylene chloride; 25 C; 1 mL/min. (a) fresh sample; (b) degraded sample of polymer (stored at pH 7.4 and 37 C for 60 days); (c) degraded sample of paclitaxel plus polymer. Source: Adapted from [14].

does not result in the formation of peaks that overlap the paclitaxel peak and thereby compromise its quantification. For other details of this NPC method development, see [14].

7.4. PROBLEMS IN THE USE OF NORMAL-PHASE CHROMATOGRAPHY Various problems can occur when carrying out NPC separation with silica as column packing: • Poor separation reproducibility (including extreme sensitivity to mobile-phase water content). • Solvent demixing. • Slow column equilibration when changing the mobile phase. Water is a very strong solvent for NPC, and traces of water in the mobile phase can markedly affect sample retention, especially for weaker

REFERENCES

155

mobile phases (e.g., ε < 0.3). Therefore, as ambient humidity increases, the concentration of water in a solvent can increase, leading to decrease in sample retention. This problem can be minimized by maintaining a constant water concentration in the mobile phase. A mobile phase with a fixed water content can be prepared by blending specified volumes of the mobile phase that is either water free or water saturated. For example, mixing equal volumes of the two mobile phases results in 50% water saturation. For further details, see [1,3]. Solvent demixing refers to the preferential retention of the more-polar B solvent and its removal from the mobile phase. This process is more important for TLC, where its effects can be minimized by a pre-equilibration of the plate in a mobile-phase-saturated atmosphere, prior to adding the sample. While demixing can also occur in HPLC operation, passage of a suitable volume of mobile phase through the column eliminates this problem. Column equilibration after a change of the B solvent may require a longer time than in RPC, so it is advisable to confirm that retention does not change with time by repeated injections of the sample. These problems are usually of minor importance and should not discourage the use of NPC for the right application. For further details on NPC, see [4].

References [1] Snyder LR. Principles of adsorption chromatography: the separation of nonionic organic compounds. New York: Marcel Dekker; 1968. [2] Snyder LR. In: Horva´th C, editor. Higheperformance liquid chromatography: advances and perspectives, vol. 3. New York: Academic Press; 1983. p. 157. [3] Snyder LR. Solvent Selectivity in Normal-phase TLC. J Planar Chromatogr 2008; 21:315. [4] Snyder LR, Kirkland JJ, Dolan JW. Introduction to modern liquid chromatography. Chapter 8. 3rd ed. New York: Wiley-Interscience; 2010. [5] Truedsson L-A, Smith BEF. Study of retention behaviour of primary, secondary and tertiary anilines in normal- and reversed-phase liquid chromatography, J Chromatogr 1981:214:291. [6] Karger BL, Snyder LR, Horva´th C. An introduction to separation Science. Chapter 19. New York: Wiley-Interscience; 1973. [7] Snyder LR, Poppe H. The Mechanism of Solute Retention in Liquid-solid Chromatography and the Role of the Mobile Phase in Affecting Separation. Competition vs ‘Sorption’, J Chromatog 1980;184:363. [8] Soczewinski E. Solvent composition effects in thin-layer chromatography systems of the type silica gel-electron donor solvent, Anal Chem 1969;41:179. [9] Meyer VR, Palamareva MD. New graph of binary mixture solvent strength in adsorption liquid chromatography, J Chromatogr 1993;641:391. [10] Soczewinski E, Dzido T, Go1kiewicz W. Comparison of high-performance liquid chromatographic and thin-layer chromatographic data obtained with various types of silica, J Chromatogr 1977;131:408.

156

7. LIQUIDeSOLID CHROMATOGRAPHY

[11] Glajch JL, Kirkland JJ, Snyder LR. Practical Optimization of Solvent Selectivity in Liquid-solid Chromatography Using a Mixture-design Statistical Technique, J Chromatogr 1982;238:269. [12] Fried B, Sherma J. Thin layer chromatography. 4th ed. New York: Marcel Dekker; 1999. [13] Solvent guide. Muskegon, MI: Burdick & Jackson Labs. 1980. [14] Vaisman B, Shikanov A, Domb AJ. Normal phase high performance liquid chromatography for determination of paclitaxel incorporated in a lipophilic polymer matrix, J Chromatogr A 2005;1064:85.

C H A P T E R

8

Ion Chromatography B. Paull, P.N. Nesterenko ACROSS, University of Tasmania, Hobart, Tasmania, Australia O U T L I N E 8.1. Introduction 8.1.1. Definitions 8.1.2. History

157 157 158

8.2. Basic Principles and Separation Modes 8.2.1. Ion-exchange Chromatography 8.2.2. Ion-exclusion Chromatography 8.2.3. Chelation Ion Chromatography 8.2.4. Zwitterionic Ion Chromatography 8.2.5. Eluents for Ion Chromatography

158 158 160 161 162 164

8.3. Instrumentation 8.3.1. Ion Chromatography Columns 8.3.2. Eluent Generators and Eluent Converters 8.3.3. Detection in Ion Chromatography

165 165 175 179

8.4. Applications 8.4.1. Industrial Applications 8.4.2. Environmental Applications

185 185 186

References

188

8.1. INTRODUCTION 8.1.1. Definitions As a rule the terminology used for the classification of various chromatographic techniques is based on either key separationeretention Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00008-0

157

Copyright Ó 2013 Elsevier Inc. All rights reserved.

158

8. ION CHROMATOGRAPHY

mechanism or the instrumental process utilised. Ion chromatography is an exception to this general rule, as the name was derived from the class of separated or target solutes and not the type of chromatographic interaction. This produces some significant difficulties when attempting to correctly define what chromatographic separations can be collectively termed ion chromatography, as the potential to utilise many varied chromatographic methods for the “high-performance” separation of ions exists. However, for the majority of cases, it is appropriate to define modern ion chromatography as the “high-performance liquid chromatography of small charged and polar solutes, predominantly based on ion-exchange and electrostatic interactions with an oppositely charged stationary phase.” In this chapter, those chromatographic methods that closely fit the definition are presented. Methods for ion separation based on mainly mobile phase interactions, such as ion-pairing and ion-interaction liquid chromatography are not included here.

8.1.2. History Staying true to the preceding definition, the liquid chromatographic separation of 13 lanthanides on a cation-exchange column, as a classified process within the famous Manhattan Project during the Second World War [1] and the ion-exchange based chromatography of 22 natural amino acids as a part of amino-acid analysis project established in 1958 by S. Moore and W. H. Stein, together with D.H. Spackman [2], are excellent very early examples of what can be considered the origins of today’s modern ion chromatography, both of which are examples of remarkably efficient ion-exchange separations for the period. However, the term ion chromatography, as we know it today, appeared only in 1975, as a result of the ground-breaking discovery of Hamish Small, T. S. Stevens, and W. C. Bauman [3], who provided a solution to the detection of small inorganic ions separated on ion-exchange resins. Small and co-workers were able to improve the sensitivity of conductimetric detection through post-column suppression of eluent conductivity (see Section 8.3.3), at once establishing a whole new chromatographic technique for the rapid and quantitative determination of inorganic anions and cations, the fundamentals principles of which have to this day changed only marginally.

8.2. BASIC PRINCIPLES AND SEPARATION MODES 8.2.1. Ion-exchange Chromatography An ion-exchange separation mechanism is mainly based on electrostatic interactions between hydrated ions from a sample (A1, A2, ., An),

8.2. BASIC PRINCIPLES AND SEPARATION MODES

159

and oppositely charged functional groups of an ion-exchanger or stationary phase housed within a chromatographic column. Solute ions are delivered to the stationary phase via a flow of an eluent or mobile phase, possessing an appropriate elution strength arising from the presence of competing eluent ions, E. In the process of the separation, the ion-exchanger is equilibrated with eluent ions, and in the case of anion exchange chromatography, the following basic interaction to describe the exchange process can be written: þ   Anion exchangerþ E þ A j ¼ Anion exchanger Aj þ E

The relative strength of interaction between any particularly anion, A-j, and the anion exchanger defines the ion-exchange selectivity of the system. The affinity of an anion-exchanger for a specific anion, A-j, can be expressed by either a distribution coefficient: D ¼

½Aj s ½Aj m

or by a selectivity coefficient: A

KE j ¼

½Aj s ½Em ½Aj m ½Es

where ½Aj s and ½Aj m are equilibrium concentrations of the solute anion in the stationary and mobile phases, respectively. The selectivity coefficient takes into account the affinity of the ion-exchanger toward eluent competing ions, E, (EL for anion exchange systems). Similar, but precisely opposite, expressions are of course true for cation-exchange based processes. The retention factor, k, for an ion retained within an ion-exchange column is proportional to the distribution coefficient, D, and to the phase ratio, 4, which is constant for each chromatographic column, containing ms grams of stationary phase and Vm milliliters of mobile phase: tR  to ms k ¼ ¼ D4 ¼ D to Vm Correspondingly, the separation selectivity a ¼ kj/ki is equal to the ratio of distribution coefficients Di/Dj for the pair of separated ions. In practice, for each of the methods presented within this Chapter, selectivity depends on the following four parameters: 1. 2. 3. 4.

Properties of the stationary phase (see Section 8.3.1). Properties of the mobile phase (see Section 8.3.3). Type and nature of the solute ion. Secondary phase interactions and equilibria, such as solvate effects, hydrophobic interactions for organic ions, hydrogen bonding, Donnan ion exclusion, and phase complexation.

160

8. ION CHROMATOGRAPHY

In the absence of any such secondary interactions, the retention of an ion in ion-exchange chromatography is simply proportional to its charge, polarizability, and size. Theoretically, among ions of a similar charge, stronger interaction, and therefore retention, should be expected for smaller ions. However, due to their interaction with water molecules, the actual size of hydrated ions changes drastically and in the opposite order to the size of the original ion, as demonstrated by Figure 8.1. This results in a reversal of selectivity to the theoretical, following the approximate orders shown graphically in Figure 8.1, where affinity toward standard ion-exchange phases increases from left to right.

8.2.2. Ion-exclusion Chromatography Ion-exclusion chromatography (IEC), like most modes of high-performance chromatography, takes its name from the dominant retention or separation mechanism being exploited, in this case, “ion exclusion.” The technique is applied mainly to the separation of weak acids, particularly carboxylic acids, but has also been applied to the separation of carbohydrates, phenols, and amino acids and can also be used for the separation of weak bases. The commonly accepted retention mechanism at play within IEC involves the formation of a pseudo semi-permeable membrane around the resin stationary phase. This is often referred to as a Donnan membrane equilibrium, through the accumulation of water

FIGURE 8.1 Variation in hydrated radius of selected ions. Source: Reproduced with permission from Ref. [4].

8.2. BASIC PRINCIPLES AND SEPARATION MODES

161

molecules within a dense hydrated layer on the surface and within the pore structure of a high-capacity strong ion-exchange resin, forming what is often termed the occluded phase. Ionic solutes of similar charge to the stationary phase (generally sulfonated cation exchangers are used for weak acid analytes), experience repulsion from the resin surface, whereas neutral species can penetrate the resins pores and stationary occluded phase, thus experiencing retention. Separation selectivity in such a system is primarily solute dependent, as IEC generally uses very simple eluent systems, such as dilute sulfuric acid (for weak acid solutes), and the exchange groups on the resin surface do not interact with the similarly charged solutes. Therefore, the potential of solutes to partition from the mobile phase into the stationary phase is governed by solute charge (excluding secondary hydrophobic interactions, which in practice can be significant). For weak acids, this of course is dependent on acid dissociation constants (pKa values). Therefore, retention, in the form of elution volume, within IEC can be directly correlated to pKa values for a range of weak acid and basic solutes. Excluding secondary interactions, the solute retention factor in IEC can be described by the following basic equation [5]: k ¼

Cs;HR Vs CHR Vm þ CR Vm

where Vm and Vs are the volumes of the mobile phase (eluent) and occluded liquid stationary phase, respectively, CHR is the concentration of acid within the mobile phase, Cs, HR is the concentration of acid in the stationary phase, and C-R is the concentration of anion R- in the mobile phase. In practice, IEC is accomplished over a limited working range, scaling between totally excluded fully ionized species and the retention volume of a completely neutral marker. As mentioned previously, hydrophobic adsorption of larger organic acids is a significant contributory factor to selectivity in IEC, and often 10e30% of organic solvent is required within the mobile phase to reduce these interactions. Stationary phase capacity and degree of cross-linking (for polystyrene-based resins) has also been shown to play a role in the resultant selectivity [6].

8.2.3. Chelation Ion Chromatography Chelation ion chromatography (CIC) is the name given to a mode of IC where the stationary phase used is one that not only interacts with the solute ions through a simple ion-exchange mechanism but has the potential to form coordinate bonds based on multipoint interactions (chelate formation) with the solute ions, typically inorganic cations. This

162

8. ION CHROMATOGRAPHY

coordination capacity comes from the stationary-phase-bound chelating functional groups used in CIC, typical of which are iminodiacetic acid (IDA) or aminophosphonic acid (APA) [7]. The strength of these interactions is closely described by the known stability constants for each metal cation and the specific immobilised polydentate ligand, and as the ligands applied in CIC are typically weak acid based, retention is heavily dependent on eluent pH. In CIC, retention can result from simultaneous ion-exchange and coordination type interactions. The contribution of the former can be minimised through the use of a relatively high ionic strength eluent, typically 0.5 to 1 M of an inorganic salt, such as KNO3. The following expression describes the distribution ratio, DM, of a metal cation, Myþ, between the negatively charged chelating ion-exchanger and the eluent [7]: DM ¼ ðynÞþ

½MR

ðynÞþ

e þ ½MR ½Myþ 

ðynÞþ

c

ðynÞþ

where ½MR e and ½MR c are equilibrium concentrations of the metal cation retained through electrostatic-based ion-exchange and stationary phase complexation (chelation), respectively. The free metal ion concentration in the eluent is given by ½Myþ . Inclusion of a phase ratio constant to describe column specific variables, 4 ¼ Vs/Vm, relates DM to retention factor, k: k ¼ DM $4 Assuming electrostatic-based ion-exchange interactions can be reduced to negligible levels through the use of high ionic strength eluents and that surface complexation occurs with a 1:1 stoichiometry only, the retention of cations in CIC can be directly related to the formation constant of the metal-ligand complex formed on the stationary phase, b1: n

k ¼

b1 ½Myþ ½R ½Myþ 

 $4

n

where ½R  is the concentration of the chelating functional groups. In practice, although eluent pH and ionic strength each exert a major influence on solute retention, the addition of a secondary complexing ligand to the eluent is often utilised to provide additional control over system selectivity [8].

8.2.4. Zwitterionic Ion Chromatography Zwitterionic and amphoteric ion-exchangers, which can be defined as those within which both positive and negative charges (ion-exchange

8.2. BASIC PRINCIPLES AND SEPARATION MODES

163

sites) are located in close proximity, often exhibit alternative ion selectivity to the standard anion or cation exchangers used in IC. With a large variety of possible chemistries, including both strong and weak anion and cation exchange groups in many structural arrangements and combinations, these versatile phases can allow greater control of selectivity through manipulation of the simultaneous electrostatic attraction and repulsion forces acting between the solute ions and the stationary phase. In so-called zwitterionic ion chromatography (ZIC), researchers have exploited these dual functionality phases for the simultaneous separation of both anionic and cationic solutes, often using relatively dilute eluents compared to traditional IC, which additionally acts to increase detector sensitivity, particularly when using conductivity detection. Several models have been developed to describe retention of ionic solutes on zwitterionic and amphoteric stationary phases, although due to the complexity of the stationary phase chemistries, no single model is adequate to describe all varieties of ZIC. Both covalently bound and dynamically coated phases have been applied to ZIC [9] and utilised with a large variety of eluent systems for the separation of inorganic and organic anions and cations. These range from pure water, in what was initially termed electrostatic ion chromatography (EIC) [10], to the use of high concentrations of organic solvent in the eluent, under so-called hydrophilic interaction liquid chromatography (HILIC) conditions [11]. Cook, Dicinoski, and Haddad [12] suggested a retention mechanism for ZIC, where the stationary phase was one in which the zwitterionic groups were present as a coating of a zwitterionic surfactant, with an inner strong anion exchange site and an outer strong cation exchange group. The retention model, based on an eluent consisting of a simple inorganic salt, described two effects, one of ion-exclusion and one of chaotropic interaction. The ion-exclusion effect arose from the cationexchange group on the outer part of the zwitterionic stationary phase having a negative charge that repels solute anions, essentially acting as a Donnan membrane. However, the magnitude of this repulsion depends on the strength of interaction between eluent cations and the cationexchange functional groups, and how strongly eluent anions interact with inner anion exchange groups. Strong interaction of eluent cations acts to effectively decrease the surface negative charge, while strong interactions with eluent anions acts in an opposite fashion. The same model was later expanded to include so-called carboxybetaine-type coatings, whereby the outer group was a weak cation exchanger, which allowed eluent pH be utilised as an additional variable to control retention. Figure 8.2 shows a schematic representation of the proposed model for the surfactant coated phase [13], while Figure 8.3 shows the simultaneous separation of inorganic anions and cations obtained on a bonded zwitterionic phase operated under HILIC conditions [14].

164

8. ION CHROMATOGRAPHY

FIGURE 8.2 Proposed schematic representation of retention mechanism for ZIC using

carboxybetaine-type surfactant modified stationary phases: (a) establishment of Donnan membrane, (b) use of NaClO4 eluent, (c) use of CeCl3 eluent, (d) use of low pH eluent. Source: Reproduced with permission from Ref. [13].

FIGURE 8.3 Simultaneous analysis of anions and cations using hydrophilic interaction liquid chromatography (HILIC) on a 5-mm ZIC-pHILIC stationary phase with charged aerosol detection (CAD). Source: Reproduced with permission from Ref. [14].

8.2.5. Eluents for Ion Chromatography Typical Eluents for Anion Exchange The vast majority of eluents used with IC are water based. In some specific applications, a small percentage of organic solvent may be required to limit unwanted hydrophobic interactions. The eluents applied to anion exchange chromatography range from dilute electrolyte solutions to complex multicomponent buffer solutions and can be inorganic or organic in nature. The elution strength of an eluent for anion exchange chromatography, and therefore solute retention, depends on the concentration of the eluent competing anion and its selectivity for the particular

8.3. INSTRUMENTATION

165

anion exchange stationary phase. Choice of eluent electrolytes or buffer solutions often depends primarily on the mode of detection being used with the particular variety of IC. For the majority of IC applications, conductivity detection is used either directly, indirectly, or following post-column eluent conductivity suppression (see Section 8.3.3);, and in each instance, an eluent is chosen to match the requirements for solute detection, namely, either a low-conductivity eluent, a high-conductivity eluent, or an eluent with which conductivity can be suppressed. Table 8.1 shows typical eluents used in anion exchange chromatography, together with the typical mode of detection associated with each eluent. Figure 8.4 shows a typical anion exchange separation of common inorganic and organic anions using a sodium hydroxide eluent, delivered as a gradient, with suppressed conductivity detection. Typical Eluents for Cation Exchange Two main types of eluent systems are commonly applied to the IC separation of inorganic and organic cations, these being either dilute inorganic and organic acids or, in certain applications, protonated organic bases. The most popular eluents for the standard IC separation of inorganic monovalent and divalent alkali and alkaline earth metal cations, together with small organic amines, are dilute nitric or methane sulfonic acids (MSA) (see Section 8.3.2). In both cases, hydronium (H3Oþ) is the active eluent competing ion; therefore, solute retention is directly related to eluent pH. As hydronium ions have a very high equivalent ionic conductance, detection is generally in the form of suppressed conductivity (Section 8.3.3), although sensitive indirect conductivity is also possible. For separations of transition and heavy metal cations, selectivity can often be manipulated through the inclusion of a complexing ligand to the eluent, often an aliphatic or aromatic carboxylic acid, with detection achieved using the post-column reaction with a suitable colour-forming ligand prior to visible absorbance detection (see Section 8.3.3).

8.3. INSTRUMENTATION 8.3.1. Ion Chromatography Columns In the last decade, remarkable progress in the development and production of new IC columns has been achieved [16]. A few general parameters for ion-exchangers should be considered when attempting column selection. Ion-exchange selectivity depends on the type of ionexchange groups, ion-exchange capacity (or more correctly, charged group distribution on the surface), accessibility (with regard to the interaction with solute ions), and to some degree, on the hydrophobicity

166

8. ION CHROMATOGRAPHY

TABLE 8.1 Common Eluents for Anion Exchange Chromatography Common detection mode

Eluent system

Comments

Hydroxide (sodium or potassium)

The hydroxide ion is the weakest eluent competing anion but has a very high equivalent ionic conductance. It is suitable for weakly retained anions and is used with weak or hydroxide selective anion exchange columns

Suppressed conductivity or indirect conductivity.

Carbonate/bicarbonate (sodium or potassium)

A stronger eluent system than simpler hydroxide-only eluents, allows greater control of selectivity through control of carbonate/bicarbonate ratio

Suppressed conductivity.

Aliphatic carboxylic acids (e.g., citric, tartaric, acetic, and formic acids)

These are weak eluents (bar citrate), and relatively highly conducting, with weak to moderate UV absorption. They are useful the elution of weakly retained anions.

Direct and indirect conductivity.

Aromatic and aliphatic sulfonic acids

Aromatic sulfonic acids are strong eluents, with relatively low to moderate conductivity and highly UV absorbing. Aliphatic sulfonic acids have moderate ionexchange selectivities and are weakly UV absorbing.

Aromatic sulfonic acids: direct and indirect conductivity, indirect UV absorbance. Aliphatic SAs - direct UV absorbance.

Aromatic carboxylic acids

Examples include benzoate/ benzoic acid phthalate/phthalic acid buffer systems, which are commonly employed in nonsuppressed IC. They have low ionic conductances and are strongly UV absorbing. Eluents can provide buffering action over a relatively wide range of pH.

Direct conductivity or indirect UV absorbance.

Inorganic Salts

Strong inorganic eluents can be prepared from chloride, sulfate and phosphate salts. These are incompatible with conductimetric detection.

Direct UV absorbance, electrochemical detection.

8.3. INSTRUMENTATION

167

FIGURE 8.4 Typical modern IC separations of inorganic anions using anion exchange chromatography and a NaHCO3/Na2CO3-based eluent (a), inorganic cations using cation exchange chromatography with an HCl-based eluent (b), both with suppressed conductivity detection. Source: Reproduced with permission from Ref. [15].

of the matrix of the stationary phase. Particle size and structure of the bonded layer define separation efficiency. 1. Ion-exchange functionality. Functional groups differ by charge, bulkiness, polarizability, hydrophobicity, and ability for multipoint solute interactions. The type of functional group defines not only the retention mechanism but also separation selectivity and type of eluent suitable for the separation of a specific group of ions. 2. Ion-exchange capacity, expressed as meq/g or meq/ml or meq/ column, is responsible for the total amount of interaction between solute ions and the ion-exchanger. Thus, ion-exchange capacity predefines the concentration of the eluent required to elute ions in the shortest possible time without loss of resolution. In addition, the suitability of the detection system may also depend on the concentration of the eluent. The first ion-exchangers applied in IC had very low ion-exchange capacities, approximately 0.01e0.05 meq/ column, and were of size 250  4.0 mm i.d., due to the limited capability of packed suppressors to suppress background conductivity of eluents (see Section 9.3.3) and, hence, provide sensitive conductimetric detection of inorganic anions and cations. Modern

168

8. ION CHROMATOGRAPHY

suppressors can effectively minimize conductivity of eluents at concentrations up to 200 mN ml/min [17], so the current trend in IC sees the introduction of new ion-exchangers with significantly higher ion-exchange capacity (Figure 8.5). It should be noted that the high ionexchange capacity of IC columns (up to 0.5e0.7 meq per standard size anion-exchange column and up to 2.8e8.4 meq per cation-exchange column), extends the limits of IC to the analysis of more complex sample matrices and the handling of larger sample volumes. 3. Ion-exchanger matrix. Relatively few inorganic (silica, alumina, titania, zirconia, porous graphitic carbon) materials and organic polymers are suitable for the preparation of ion-exchangers. Chemical inertness and hydrolytic stability of the matrix are crucial requirements for ion-exchangers to be used in IC. Inorganic materials have limited hydrolytic stability in alkaline eluents (silica) or increased reactivity toward phosphates and carboxylates (titania, zirconia), so organic polymer-based ion-exchangers, including highly cross-linked poly(styrene-divinylbenzene), polymethacrylate, and poly(vinyl alcohol) functionalised materials are more commonly used for the separation of anions with alkaline eluents. More-efficient columns packed with silica-based ion-exchangers are often preferable for the separation of cations under acid conditions (see Section 8.2.5). Other important characteristics of the matrix are porous structure, specific

FIGURE 8.5 Trends in ion-exchange capacity (closed circles) and particle size (open circles) of anion-exchangers produced by Dionex Corporation (part of Thermo Scientific) in the last three decades. Source: Adapted from [17].

8.3. INSTRUMENTATION

169

surface area, and hydrophobicity in the case of organic polymers. Obviously, the last property should be taken into consideration when optimising the eluent system for the separation of hydrophobic ions. 4. Particle size. Both column efficiency and peak resolution are dependent on the particle size of the adsorbent, so there is a clear current trend to progressively decrease particle sizes for anionexchange columns, again shown in Figure 8.5. Modern IC columns are generally packed with 3e5 micron ion-exchangers. The other path to improve column efficiency is to further develop monolithic porous ionexchange columns for IC; however, only silica monolithic columns have to date demonstrated impressive efficiencies for small ions, but these are not stable hydrolytically [18]. Two very important characteristics of ion-exchangers are their porous structure and the method of immobilisation of the functional groups. These characteristics are responsible for the kinetics of mass transfer of ions within the chromatographic column. According to the structure and method of preparation, the following types of ion-exchangers can be collectively classified: 1. Ion-exchangers with grafted functional groups. These are the most common type of ion-exchangers, in which the surface of a suitable material (inorganic oxides, organic polymers) is modified via one or more surface reactions, providing a monolayer of uniformly distributed charged groups on the outer surface. Typical examples include cation-exchange resins prepared by sulfonation of porous poly(styrene-divinylbenzene) matrix and silica gels with immobilised N,N,N-trimethyl-propylammonium groups prepared by the treatment of the silica surface with 3-chloropropylsilane followed by reaction with trimethylamine. 2. Ion exchangers with grafted layers of ionogenic polymers and polymer-coated ion-exchangers. This type of modification of the surface is used when extra ion-exchange capacity is required or when the central core or matrix displays unsatisfactory properties, such as a lack of hydrolytic stability or the presence of strong binding sites, which must be shielded. An example of this type of ion-exchanger is one commonly used for the simultaneous separation of alkali and alkaline-earth metal cations, which is based on a silica particle coated with a layer of poly(butadiene-maleic) acid (PBDMA) [19]. 3. Agglomerated ion-exchangers. These are prepared using composite bead technology, which generally involves the coating, either electrostatically or covalently, of a surface layer of charged nanoparticles on the surface of a larger supporting substrate particle. The presence of the layer of nanoparticles provides favorable masstransfer kinetics and simultaneously delivers an ion-exchanger with

170

8. ION CHROMATOGRAPHY

increased surface area and higher ion-exchange capacity. For many years, this type of adsorbents was produced by Dionex Corp. (now part of ThermoFisher Scientific) and recently by Phenomenex [20]. 4. Dynamically modified ion-exchangers. These are usually prepared by the saturation of a column packed with a hydrophobic adsorbent with hydrophobic ionic molecules. These ionic molecules are strongly retained by hydrophobic interactions on the column and form a stable coating with aqueous eluents. This method gained popularity because of the ability to prepare various types of ion-exchanger, based on the same commercially available reversed-phase column. With this approach, ionexchangers with complex functional groups, such as zwitterionic groups, can be easily obtained, (see Section 8.2.4) [13]. Anion-Exchange Columns The effective separation of anions formed through dissociation of weak acids is possible only in alkaline conditions, so hydrolytic stability of modern anion exchangers is a crucial condition. Poly(methacrylate)-, poly(vinyl alcohol)-, and PS-DVB-based anion exchangers constitute the most common types of commercially available columns. Most organic polymers exhibit reasonable stability in alkaline eluents (see Table 8.2), but the low mechanical stability of PVA and PMA limits column length and particle size, due to limited column back-pressure tolerance. In the case of polystyrene-based ion-exchangers, hydrophobic interactions between hydrophobic anions and the substrate contribute significantly to retention and can cause peak distortion. To avoid such issues and combine certain advantages of various polymers, a new generation of anion exchangers was recently introduced by Dionex Corp. For example, the IonPac AS24 anion exchanger is composed of a core macroporous polystyrene microparticle, coated with a hydrophilic multilayer, obtained by consecutive treatment of the surface with epoxy and amino substrates. The resulting material has both excellent mechanical and hydrolytic stability. TABLE 8.2 Properties of Common Substrate Matrices Used for the Preparation of Ion Exchangers

Matrix Silica

Hydrolytic stability 1e7

Mechanical stability

Residual ion-exchange activity

Hydrophobicity

Excellent

Significant

Low

PS-DVB

0e14

Good

No

Substantial

PMA

2e12

Moderate

Some

Lowemoderate

PVA

3e12

Moderate

Low

Low

8.3. INSTRUMENTATION

171

Anion exchangers bearing quaternary ammonium groups cover over 95% of all anion exchangers produced for IC. Among these, two types of anion-exchange functionalities can be outlined. Anion-exchange “Type-I” groups consist of a trialkylammonium head connected through an alkyl spacer to the surface. Anion exchangers of this type are also known as providing carbonate-type selectivity. The effect of the structure of quaternary alkylammonium sites (type of substituent, length and bulkiness, length of spacer, etc.) on separation selectivity is well studied and described in a monograph by Fritz and Gjerde [21]. Anion-exchange “Type-II” groups consist of two alkyl substituents and one hydroxyalkyl substituent at the nitrogen and provide so-called hydroxyl-type selectivity. Weak-acid anion exchangers bearing primary, secondary, and tertiary amino groups may be also used in IC [22]. The properties of some anion exchangers and anion-exchange columns for IC are presented in Table 8.3. Cation-Exchange Columns Commercially available cation exchangers for ion chromatography can be divided into three groups: 1. Strong-acid cation exchangers bearing sulfonic acid functional groups. 2. Weak-acid cation exchangers, which includes carboxylic, phosphonic, and phosphoric acid functional groups or thereof combination. 3. Complexation-type ion-exchangers, for example, ion-exchangers with immobilized crown ether molecules, capable of forming inclusion complexes with specific cations and thus selectively retain them [16,23]. Sulfonic-acid-type ion-exchangers have a high affinity toward alkaline earth metal cations, which increases analysis time and restricts their practical application for the simultaneous determination of alkali and alkaline-earth metals. However, they can be used for the efficient and selective separation of transition metals by using ethylenediamineeorganic acid (tartaric, citric) based eluents [24] or for the separation of lanthanides with hydroxyisobutyric-acid-containing eluents [25]. Optimum selectivity for alkali- and alkaline-earth metal cations and their simultaneous separation in acceptably short run times can be achieved with carboxylic-type cation exchangers. This type of cation exchanger usually contains an immobilised layer of poly(butadieneemaleic acid) copolymer [19] or, for example, poly(itaconic acid) [26]. In some ion-exchangers, such as the IonPac CS-12A, the addition of a bonded-phosphonic-acid functionality is used to improve the separation selectivity for manganese in relation to the group of alkaline-earth metals [27]. Crown-ethers or macrocycles containing cation exchangers provides the unique selectivity for the separation of cations with similar charge and ionic radius, for example, potassium and

172

TABLE 8.3 Properties and Common Applications of Commercially Available Anion Exchangers for IC Column properties Stationary phase

Bonded groups þ

dp, mm

Column size, mm

Capacity, meq/col

Matrix

Applications

Ref

eN R2R’OH

9

250  4.0

190

Macroporous 200 nm pores; EVB-DVB, 55%; 90 nm latex beads with 15% cross-linking

Determination of trace anions in concentrated hydrofluoric and glycolic acids with ion-exclusion pretreatment, organic solvents

[29]

IonPac AS11-HC

eNþR2R’OH

9

250  4.0

290

Macroporous 200 nm pores; EVB-DVB, 55%; 70 nm latex beads with 6% cross-linking

Determination of trace anions in methanesulfonic and phosphoric acids with ion-exclusion pretreatment

[29,30]

IonPac AS14A

eNþR2R’OH

9

250  4.0

65

EVB-DVB, 55%; 10 nm pores

Determination of hydrolysis products in hexafluorophosphates salts

[31]

IonPac AS15

eNþR2R’OH

8.5

250  2.0

56

EVB-DVB, 55%; 10 nm pores

Determination of free cyanide in drinking water

[32]

IonPac AS16

eNþR2R’OH

9

250  4.0

170

Macroporous EVB-DVB, 55%; 80 nm latex beads with 1% cross-linking

Determination of oxyhalides and haloacetic acids in drinking water

[33,34]

IonPac AS17

eNþR2R’OH

10.5

250  4.0

30

Microporous EVB-DVB, 55%; 75 nm latex beads with 6% cross-linking

The determination of trace level phosphorus in purified quartz, trace anions in boric acid

[35]

8. ION CHROMATOGRAPHY

IonPac AS9-HC

eNþR2R’OH

7.5

250  4.0

160

Metrosep A Supp 1

eNþR3

7.0

250  4.6

Metrosep A Supp 4

eNþR3

9

Metrosep A Supp 5,

eNþR3

Metrosep Anion Dual 1

The determination of trace bromate in drinking water

[36]

64

PS-DVB

Speciation analysis of selenium

[37]

250  4.0

46

Polyvinylalcohol

Suitable for all routine tasks in water analysis

[38]

5

100  4.0

39

Polyvinylalcohol

Determination of bromide in canine plasma

[39]

eNþR3

10

150  3.1

9

Macroporous polyhydroxymethacrylate, 20e60 m2/g

Determination of inorganic anions in commercial seed oils and in virgin olive oils

[40]

Metrosep Anion Dual 2

eNþR3

6

75  4.6

34

Polymethacrylate

Determination of chloride in magnesium metal

[41]

IonPac Cryptand A1

2,2,1 cryptand

5

150  3.0

73*

Macroporous, 100 nm pores, EVB-DVB 55%

Determination of polyvalent anions, including polyphosphates and polysulfonates; alkanesulfonic acids in a chromic acid plating bath

[42,43]

IonPac AS18

eNþR2R’OH

7.5

250  4.0

285

Macroporous, 200 nm pores, EVB-DVB 55%, 65 nm latex beads with 8% cross-linking

Determination of anions in toothpaste

[44]

IonPac AS20

eNþR2R’OH

7.5

250  4.0

310

Macroporous, 200 nm pores, EVB-DVB 55%

Determination of perchlorate in complex samples

[45]

(Continued)

173

Macroporous EVB-DVB, 55%

8.3. INSTRUMENTATION

IonPac AS19

174

TABLE 8.3

Properties and Common Applications of Commercially Available Anion Exchangers for ICdcont’d Column properties Bonded groups

IonPac AS21

eNþR2R’OH

7

250  2.0

IonPac AS22

eNþR2R’OH

6.5

IonPac AS23

eNþR2R’OH

6

Hamilton PRPx100

eNþ(CH3)3

dp, mm

Capacity, meq/col

Matrix

Applications

Ref

45

Macroporous, 200 nm pores, EVB-DVB 55%

Determination of perchlorate in fertilizer

[46]

250  4.0

210

Macroporous, 200 nm pores, EVB-DVB 55%

Determination of anions in ionic liquids

[47]

250  4.0

320

Macroporous, 200 nm pores, EVB-DVB 55%

Determination of oxalate, phosphate and citrate in human urine

[48]

250  4.0

190

Mesoporous PS-DVB, 10 nm, 415 m2/g

Analysis of inorganic arsenic species

[49]

8. ION CHROMATOGRAPHY

Column size, mm

Stationary phase

8.3. INSTRUMENTATION

175

FIGURE 8.6 Ion-exchange selectivity for cation exchangers with sulfonic acid groups and anion exchangers with quaternary ammonium groups.

ammonium. The cation exchanger IonPac CS15 (Dionex) combines three types of functional groups, namely, carboxylic, phosphonic, and 18-crown6-ether, displaying perfectly attuned selectivity for the separation of alkali, alkaline earth, and ammonium cations [28]. The use of acidic eluents for the separation of metal cations is favorable, as this prevents unwanted secondary equilibria in chromatographic systems, such as hydrolysis, formation of aqua complexes, and other possible complexation with the active constituents of the eluent. Being hydrolytically stable in acidic eluents, silica provides a good base substrate for the preparation of efficient cation-exchange columns. Therefore, unlike anion exchangers, silica-based cation exchangers for IC are produced by several companies. Table 8.4 shows a summary of the common cation exchangers available for use in IC, while Table 8.5 shows the range of cation-exchange columns applied in IEC.

8.3.2. Eluent Generators and Eluent Converters One of the more significant advances in IC over the past decade has been the development of eluent generators. These devices require a concentrated source of electrolyte solution, and act to generate the IC eluent (isocratic or gradient) using a flow of deionised water from the IC pump module prior to sample injection. Commercial devices are based on either electrolysis together with an ion-exchange membrane (e.g., Dionex EG50 Eluent Generator) or a controlled dosing arrangement (e.g., Metrohm 845 Eluent Synthesizer). Additional eluent conditioning devices can be included within each system to remove ionic contaminants or adjust eluent pH prior to the eluent entering the sample injection module. In the case of electrolytically generated eluents, an additional degassing unit is

176

TABLE 8.4 Properties and Common Applications of Cation Exchangers for IC Column properties Bonded groups

IonPac CS5A

eSO3H

9

250  4.0

IonPac CS12A

eCOOH ePO3H2

8.5

250  4.0

IonPac CS15

eCOOH ePO3H2 e18-crown e6 ether

8.5

IonPac CS16

eCOOH

5.5

dp, mm

Capacity, meq/col

Matrix

Applications to analysis of complex matrices

Ref

EVB-DVB, 55%;

Speciation of aluminium complexes

[50]

2800

EVB-DVB, 55%; 15 nm pores; 450 m2/g

Determination of Mg and Ca in 30% NaCl brine

[51,52]

250  4.0

2800

EVB-DVB, 55%; 15 nm pores; 450 m2/g

Determination of trace-level Naþ in cooling waters, trace-level NHþ 4 in environmental waste water containing a high Naþ concentration, trace level NHþ 4 in a KCl soil extract

[28]

250  5.0

8400

EVB-DVB, 55%; 15 nm pores; 450 m2/g

Determination of trace level NHþ 4 in high concentrations of Naþ; trace level Naþ in high concentrations of NHþ 4 or amines, alkali- and alkaline-earth metal ions in acid samples (up to 0.1 M) without pH adjustment

[53]

8. ION CHROMATOGRAPHY

Column size, mm

Stationary phase

eCOOH

7

250  4.0

1450

EVB-DVB, 55%; 15 nm pores; 450 m2/g

Simultaneous separation of alkali-, alkaline-earth cations and boiler water amine additives

[54]

IonPac CS18

eCOOH

6

250  4.0

1160

EVB-DVB, 55%; 15 nm pores; 450 m2/g

Determination of biogenic amines in alcoholic beverages

[55]

Universal cation

eCOOH

7

100  4.6

d

Silica based

Determination of cations at trace levels in ice core samples

[56]

Metrosep Cation 1-2

eCOOH

7

125  4.0

122

Silica-based, pores 10 nm, 350 m2/g

Simultaneous determination of alkali-, alkaline-earth, and transition metal elements in uranium and thorium-based nuclear fuel materials

[57]

Metrosep C2

eCOOH

5

250  4.0

194

Silica based

Determination of methylamines and trimethylamine-N-oxide in particulate matter air samples

[58]

IonPac SCS 1

eCOOH

4.5

250  4.0

318

Silica based, 12 nm pores; 300 m2/g

Simultaneous determination of alkali-, alkaline-earth, and transition metal cations

[59,60]

Hamilton PRPx200

eSO3H

10

250  4.1

35 meq/g

PS-DVB, 10 nm pores

Determination of mercury and methylmercury in seafood

[61]

Hamilton PRPx800

eCOOH

5

150  4.0

3700 meq/g

PS-DVB, macroporous

Trace alkaline-earth and transition metals in brines

[26]

8.3. INSTRUMENTATION

IonPac CS17

177

TABLE 8.5

Properties and Common Applications of Ion Exchangers for Ion-Exclusion IC

178

Column properties Bonded groups

dp, mm

Column size, mm

Capacity, meq/col

IonPac ICE-AS1

eSO3H

7.5

250  9

IonPac ICE-AS6

Mixed eSO-3 and eCOOH

8

IonPac ICE-Borate

eSO3H

Hamilton PRPx300

Matrix

Applications

Ref

27,000

Microporous, PSDVB, 8% crosslinking, hydrophilic surface

Trace analysis of anions in hydrofluoric acid

[62]

250  9

27,000

Microporous PSacrylate -DVB, 8% cross-linking

Application to environmental analysis; determination of trace anions in concentrated weak acids

[29,63]

7.5

250  9

27,000

Microporous, hydrophilic surface, 8% cross-linking

Trace borate analysis in deionised water

[64]

eSO3H

7

250  4.1

170 meq/g

PS-DVB

Determination of arsenate and arsenite

[65]

Aminex HPX-87H

eSO3H

9

300  7.8

1,700 meq/g

PS-DVB, 8% crosslinking

Analysis of carbohydrates and organic acids

[66,67]

TSKgel OApak A

eCOOH

5

150  6.0

100 meq/ml

Polymethacrylate

Simultaneous determination of inorganic anions and cations in acid rainwaters; organic acids

[68]

TSKgel Super IC-A/C

eCOOH

3

150  6.0

200 meq/ml

Polymethacrylate

Simultaneous determination of inorganic anions and cations

[68,69]

TSKgel SCX

eSO3H

5

150  6.0

>1,500 meq/ml

Polymethacrylate, 6 nm pores

Organic acids

[70]

8. ION CHROMATOGRAPHY

Stationary phase

8.3. INSTRUMENTATION

179

necessary. The advantages of these systems are the production of purer eluents (lower detection background and noise), precise computer control of eluent ion concentration (the ability to produce complex linear and step gradients with an isocratic pump module), and the overall savings in analysis times such automation can provide. Figure 8.7 shows the Dionex EG50 eluent generation device. In the example shown, the device is configured to generate a potassium hydroxide (KOH) eluent, although similar devices are available to  produce carbonateebicarbonate eluents (CO2 3 /HCO3 ), also for anionexchange, or indeed MSA eluents for cation-exchange-based separations.

8.3.3. Detection in Ion Chromatography Electrochemical NON-SUPPRESSED CONDUCTIVITY

By far the most commonly employed mode of detection in IC is conductivity detection. Conductivity detectors are universal (nonselective) bulk property detectors for ionic solutes, which measure the overall conductivity of the eluent and the transient bands of eluting analyte ions. As the majority of common anions and cations have no, or only weak, UV chromophores, conductivity is the detector of choice for the majority of IC applications. Conductivity detection is based on measurement of the resistance (or strictly the impedance) between two electrodes within the post-column analytical flow cell. The response of a standard conductivity cell (here for anion-exchange chromatography) can be described by the following equation: DG ¼

ðlS  lE ÞCS 103 K

FIGURE 8.7 An eluent generator module for production of potassium hydroxide eluents. Source: Reproduced with permission of Ref. [71].

180

8. ION CHROMATOGRAPHY

where DG is the conductance signal, lS- and lE- are the limiting equivalent ionic conductances of the solute (analyte) and eluent anions, respectively; CS is the concentration of the solute (analyte) anion; K is a constant (called the cell constant) taking into account the physical dimensions of the cell. Conductivity can be measured in two ways. Firsty, in a “direct” mode, an eluent with an overall low ionic conductance is used, commonly a dilute solution of a weak organic acid for anion-exchange, under which conditions eluting analyte ions are detected as positive peaks, due to their greater limiting equivalent ionic conductance, compared to the eluent ion. In this configuration, a relatively low-capacity ion-exchange analytical column is typically required. Alternatively, a highly conducting eluent, such as a dilute solution of a strong acid solution or alkali hydroxide can be used, whereby analyte ions exhibit a lower equivalent ionic conductance in relation to the eluent ion and thus appear as negative peaks within the resultant chromatogram. This mode of conductivity detection is termed indirect conductivity. SUPPRESSED CONDUCTIVITY

The breakthrough in the development of what today we regard as modern IC was achieved in the mid-1970s by Small, Stevens and Bauman [3], who introduced what was initially called the stripper column and which we now know as the eluent suppressor. The first commercial suppressors were developed by Dionex Corp., in 1975, and the company has remained at the forefront of eluent suppressor technology to this day. An eluent suppressor acts to reduce or “suppress” the background conductivity of the IC eluent through the exchange of ions across an ionexchange membrane (membrane suppressor) or through the use of a high-capacity ion-exchange resin (packed bed suppressor). In IC methods where the eluent is composed of a simple dilute solution of either a strong acid (e.g., HNO3) or base (e.g., KOH) (see Section 8.2.5), the very high background conductance generated by the eluent can essentially be eliminated through the exchange of eluent counter ions, namely, NO3- and Kþ in the preceding examples, with either hydroxide (OHe) or hydronium ions (H3Oþ), forming simply H2O in both instances. The provision of these suppressing ions can occur, as mentioned, from either a packed bed suppressor containing a high-capacity anion- or cationexchange resin in either basic or acid form, respectively; delivered across an anion- or cation-exchange membrane, from an ion source, such as a stream of strong acid or base; or generated electrolytically. In a similar fashion. Other common eluent systems for anion-exchange chromatography, such as sodium or potassium carbonate/bicarbonate (CO32-/ HCO3-)ebased eluents, can be converted to a less conducting solution of carbonic acid (H2CO3) through exchange of the alkali cation counter ions

181

8.3. INSTRUMENTATION W aste

Waste Analyte in Na + and OH -

Na + HSO 4 -

Na + HSO 4 Na

+

Na

H+ HSO 4 + H+

+

H+

Analyte in H 2 O

H + HSO 4 -

HSO 4 + H+

H + HSO 4 To Detector

Cation Exchange Membranes

FIGURE 8.8 A membrane-based suppressor showing chemical suppression of a sodium hydroxide eluent using sulfuric acid.

with hydronium ions. Figure 8.8 shows the exchange of ions within a membrane-based suppressor module (Dionex Corp.) for a NaOH eluent used for anion-exchange-based IC. As can be seen, the analyte counterion within the sample (Naþ in the diagram) is simultaneously exchanged for H3Oþ (Hþ) during passage through the suppressor module or column, acting to additionally increase the relative conductance of the analyte band as it passes through the subsequent detector cell. Recent developments in IC, such as the commercial production of capillary IC columns, has necessitated the development of new suppressor technology capable of eluent suppression at the capillary scale without introduction of significant post-column band broadening. Capillary suppressor modules, such as the Dionex ACES300 system, use an ion-exchange membrane capillary, coiled within an ion-exchange resin-filled chamber, through which a continuous flow of regenerant solution flows. Suppression of eluent ions occurs across the capillary membrane in a similar fashion to the much larger standard membrane suppressor; however, in the capillary format, the internal capillary void volume is as little as 1.5 ml. Two electrode chambers separated from the resin-filled chamber by ion-exchange membranes are used to electrolytically supply the regenerant ions for continuous eluent suppression.

182

8. ION CHROMATOGRAPHY

AMPEROMETRY

Amperometric detection can be used for the detection of electroactive solutes, specifically those readily amenable to reduction or oxidation. A certain potential is applied between a working and a reference electrode, and where solutes passing over the working electrode are reduced or oxidized, current will flow, this being the principle of the detection method. Amperometric detection requires close control over eluent temperature, pH, and eluent flow (eluent flow through the reactor should be continuous and pulse free) to obtain stable, reproducible results. Pulsed amperometric detection (PAD) is the most common operating mode for IC applications, which utilises a measuring potential and two cleaning potentials to provide constant electrochemical regeneration of the electrode surface. In PAD, the analysis is performed by a series of cyclic potentials applied to the working electrode, with a measuring potential applied first, and the current is measured after a suitable equilibration time. Following this, a large positive potential is applied to the electrode, causing the oxidative removal of any reaction products, and a subsequent negative potential to reduce the working electrode to its original state. The whole process is constantly repeated, typically lasting Mw > Mn ; Mz ¼ Mw ¼ Mn. The discussion here focuses exclusively on the former scenario, in which case Mz is characteristic of the higher (larger molar mass) end of the MMD, Mn of the lower end, and Mw of an intermediate region near the mode. An example of this is seen in Figure 9.1 for the monomodal MMD of a disperse, linear polystyrene (PS) homopolymer. Because the z-average molar mass is usually located in a region of the MMD occupied by long chains, this average can inform knowledge of processing characteristics, such as flex life and stiffness. Conversely, being located in the small-molecule region of the MMD, the Mn of a polymer can provide some idea as to the brittleness and flow properties of the material. Given the statistical nature of the various molar mass (M) averages, however, it is possible for polymers with very different MMDs to have identical Mn, Mw, Mz, and so on [2]. Additionally, certain processing and end-use properties, such as elongation, hardness, and yield strength, may increase with increasing M but decrease with a narrowing of the MMD. As such, examination of the MMD should almost always be performed in conjunction with an examination of the M averages of a polymer.

9.2. HISTORICAL BACKGROUND Many of the aforementioned relationships between macromolecular properties and the various M averages and the MMD were well-recognized by the early 1960s. However, no convenient method existed through which to determine the MMD and accompanying M averages of a polymer in a single experiment. To address this shortcoming, John Moore at the Dow Chemical Company, developed a technique he termed gel permeation chromatography, or GPC. His paper Gel Permeation Chromatography, I. A New Method for Molecular Weight Distribution of High Polymers was published in 1964 [3]. 1

For an excellent discussion of the concept of statistical moments as it applies to the MMD of polymers, the reader is referred to Section 2.4 of reference [1].

195

9.2. HISTORICAL BACKGROUND

Mw

Differential weight fraction

1.25

Mn

1.00

Mz

0.75

0.50

0.25

0.00 4

10

5

10

6

10

Molar mass M (g mol-1) FIGURE 9.1 Differential MMD and M averages of broad dispersity, linear PS. Mn [ 2.77 3 105 g mole1, Mw [ 5.38 3 105 g mole1, Mz [ 8.29 3 105 g mole1, Mw/Mn [ 1.94. Determined by SECeMALSeDRI employing a set of three PSS GRALlinear 10-mm particle size columns and one PSS GRAL10000 10-mm particle size column, preceded by a guard column. Solvent: N,N-dimethyl acetamide þ 0.5% LiCl; temperature: 35 C; flow rate: 1 ml mine1. Detectors: DAWN E MALS, and Optilab DSP DRI. (A. M. Striegel, unpublished results).

The work of Moore built upon earlier research by Wheaton and Bauman and by Porath and Flodin [4e6]. In 1953, the former researchers noted the fractionation of nonionic substances during passage through an ion exchange column, which indicated that separation of molecules based on size should be possible in an aqueous solution. This type of separation was demonstrated in 1959 by Porath and Flodin, who employed columns packed with cross-linked polydextran gel, swollen in aqueous media, to effect the size-based separation of various water-soluble macromolecules. This aqueous-based technique became known as gel filtration chromatography, or GFC. While other hydrophobic gels were also developed for the separation of compounds of biological interest, the fact that the gels swelled only in aqueous media limited their application to water-soluble substances. In his pioneering work, Moore employed styreneedivinylbenzene gels cross-linked to a degree that balanced rigidity and permeability. Columns packed with these gels where connected to a differential refractometer, specially designed by James Waters with an optical cell smaller than what was commercially available at the time, with continuous flow in both the sample and reference sides of the cell, and capable of operating at

196

9. SIZE-EXCLUSION CHROMATOGRAPHY

temperatures up to 130 C [7]. Moore recognized that, with proper calibration, GPC could provide both the MMD and M averages of synthetic polymers, a capability quickly capitalized on by many scientists in the polymer industry, who had been longing for just such a technique. In the words of A. C. Ouano, “With the introduction of gel-permeation chromatography (GPC) by Moore, molecular weight distribution data for polymers took a sudden turn from near nonexistence to ready availability” [8]. The early history of the instrumental development and commercialization of GPC by the Waters Corporation is elegantly recounted in the articles by McDonald [9,10]. We reconcile at this point in the chapter the terms gel permeation chromatography and gel filtration chromatography under the common term size-exclusion chromatography, or SEC. There are a number of good reasons for doing this: First, elution in both GPC and GFC proceeds by a common size-exclusion mechanism. Second, while many SEC columns are still packed with cross-linked gels, just as many are packed with “nongel” materials, such as porous silica and alumina and, more recently, monoliths. Last, because GPC was the term used when operating in organic solvents, while GFC denoted experiments in aqueous media, it is difficult to avoid pointing out that a particular researcher might perform GPC experiments on a Monday and GFC experiments on a Friday of the same week, using the exact same hardware (and, perhaps, even the same columns), separating analytes via the same chromatographic mechanism, only employing a different solvent. For these reasons, the all-inclusive and more aptly descriptive term size-exclusion chromatography is preferred and employed from here onward. The column packings employed both by Moore and by Porath and Flodin were lightly cross-linked, semi-rigid networks of large (z75 mm to 150 mm) particles that could be used only at low flow rates and operating pressures (\250 psi), resulting in long, relatively inefficient analyses. The introduction, in the 1970s, of m-Styragel, which consisted of semi-rigid 10 mm cross-linked PS particles capable of withstanding pressures of several thousand psi, simultaneously allowed for both faster analysis and superior performance compared to what had been previously possible. Since then, a variety of packing materials have been introduced, ranging in size from around 3 mm to 20 mm and capable of separating anywhere from monomers and oligomers to ultrahigh-M polymers and, even, particles. In the intervening decades, advances in SEC have been guided mostly by the need for absolute (i.e., calibrant-independent) M determination, by a multidetector approach aimed at determining physicochemical distributions and heterogeneities, and by the incorporation of SEC into twodimensional liquid chromatographic (2D-LC) scenarios to further understanding of complex polymers and blends. These topics are treated

9.3. RETENTION IN SIZE-EXCLUSION CHROMATOGRAPHY

197

in Sections 9.6 and 9.7. First, however, we review some of the fundamental chromatographic principles of SEC.

9.3. RETENTION IN SIZE-EXCLUSION CHROMATOGRAPHY “Retention in SEC is an equilibrium, entropy-controlled, size-exclusion process” [2]. This statement merits some attention, and each part of it is examined individually here, in reverse order.

9.3.1. A Size-Exclusion Process Let us suppose we inject a disperse analyte (e.g., the PS in Figure 9.1), or a mix of narrow dispersity analytes, onto an SEC column packed an inert, porous substrate. Due to their size, the larger components of the analyte or mix sample either a smaller number of pores or, within a given pore, smaller pore volume than the smaller components of the sample. Because of this behavior, the larger components elute from the column first and the smaller components elute later. Elution in SEC is in reverse order of analyte size, and the method may be thought of as an “inversesieving” technique. Do samples actually behave in the manner just described? Figure 9.2 shows that the size (as exemplified here by the radius of gyration;2 see Section 9.6) of the broad dispersion PS in Figure 9.1 decreases steadily as retention volume increases, that is, the sample elutes by a size-exclusion process. An abundance of similar examples can be found in the literature, including almost any SEC calibration curve, in which narrow dispersity standards are observed to elute in order of decreasing M (for a homologous series at a given set of solventetemperature conditions, M f size).

9.3.2. An Entropy-Controlled Process For dilute solutions at equilibrium (the equilibrium nature of SEC is examined next), solute distribution is related to the standard free-energy difference (DGo) between the phases at constant temperature and pressure: DG ¼ RT ln K

(9.1)

2 It should be noted that a quantitative relationship has not yet been established between the radius of the hydrodynamic sphere occupied by the analyte in solution and any main macromolecular radius, such as the radius of gyration RG. For a more detailed discussion, the reader is referred to Section 2.6.2 of reference [2].

198

9. SIZE-EXCLUSION CHROMATOGRAPHY

1.0

DRI response (V)

0.8

0.6

0.4

10

0.2

Radius of gyration RG (nm)

100

0.0 15

20

25

30

35

Retention volume VR (mL) FIGURE 9.2 SEC chromatogram as monitored by differential refractometer, DRI

(solid line), and radius of gyration versus retention volume relationship (open squares), of PS sample in Figure 9.1. Experimental conditions same as in legend for Figure 9.1. Scatter in RG when VR > 27 ml corresponds to molecules with insufficient angular dissymmetry, at experimental conditions, to allow for accurate measurement of RG; see Section 9.6 for details.

with DG ¼ DH   TDS

(9.2)

where K is the solute distribution coefficient, R is the gas constant, T is the absolute temperature, and DHo and DSo are, respectively, the standard enthalpy and entropy differences between the phases. In “traditional” LC (e.g., normal- and reversed-phase LC), retention is generally governed by solute-stationary phase interactions, sorptive or otherwise, and solute transfer between phases is associated with large enthalpy changes. In SEC, noninteracting (ideally) column packing materials are employed, corresponding to DHo z 0, and it is the change in entropy between phases that governs solute retention, as per KSEC z eDS

o

=R

(9.3)

where KSEC is the solute distribution coefficient in SEC, corresponding to the ratio of the average solute concentration inside the pores of the column packing material to the concentration outside the pores. Because solute mobility is more limited inside than outside the pores, solute permeation in SEC is associated with a decrease in solution conformational entropy, corresponding to negative values of DSo.

199

9.3. RETENTION IN SIZE-EXCLUSION CHROMATOGRAPHY

Equation (9.3) predicts that solute retention in SEC should be temperature independent. While it is realized that the size of polymers in solution, and hence, their SEC retention volume, has a modest temperature dependence (and, sometimes, a larger dependence, as when transitioning through the theta point of the solution), this does not affect the mechanism by which these analytes elute. The entropic nature of SEC retention has been confirmed by noting the virtual lack of change in KSEC with changes in temperature for a large number of mono-, di-, and oligosaccharides in both aqueous and nonaqueous systems [11e15]. Some of these data are given in Table 9.1. It is a truism that there are enthalpic contributions to all real SEC separations; it is only the magnitude of these contributions that differ. In certain specialized cases, such as when using SEC to calculate the DS of monodisperse analytes in solution, it is essential that DH z 0. As seen in Table 9.1, this can be verified by varying the temperature of the experiment and noting that, if the enthalpic contribution is small then, with a relatively large (z10 C to 20 C or more) change in temperature, KSEC changes by only a few percent. Even in nonspecialized cases, a large DH component of the separation should be avoided, as this may cause coelution due to some of the smaller components in a sample eluting by an entropically dominated process while the elution of larger components reflects a substantial enthalpic contribution to their separation.

9.3.3. An Equilibrium Process The thermodynamic equilibrium of the SEC retention process has been confirmed through two independent sets of experiments. The first set TABLE 9.1 Temperature Independence of KSEC KSEC Oligosaccharide



25 C

37 C

Maltose

0.683

0.675

Maltoheptaose

0.430

0.415

Cellobiose

0.657

0.648

Cellopentaose

0.431

0.419

Isomaltose

0.626

0.613

Isomaltoheptaose

0.356

0.341

Laminaribiose

0.657

0.643

0.369

0.354

Laminariheptaose

e1

Note: Solvent, H2O; pH, 7.4; flow rate, 1 ml min . Columns: Four ultrahydrogel ˚ pore size; detector: Optilab rEX. See references [13,15] for details. 6 mm-particle size, 120 A

200

9. SIZE-EXCLUSION CHROMATOGRAPHY

demonstrated that the solute distribution coefficient is independent of flow rate, over a flow rate range of approximately one order of magnitude and a molar mass range of approximately three orders of magnitude, for two types of analytes, narrow dispersity PS and poly(methyl methacrylate) (PMMA) standards [2,16]. Results for the latter are shown in Figure 9.3, demonstrating that retention in SEC is governed by the extent of analyte permeation into the pores of the column packing material, not by the rate of permeation. It should be noted that, 40 years earlier, James Waters and colleagues found the retention volume of both solvent tracers and large polymeric analytes to be essentially flow-rate independent for flow rates spanning two orders of magnitude employing, individually, columns of three particle sizes [17]. The second set of experiments confirming the thermodynamic equilibrium of SEC retention compared results from flow and static mixing experiments for a set of narrow dispersity PS standards covering approximately three orders of magnitude in M [2,18,19]. In the flow experiments, solute distribution coefficients were calculated from the

1.0 0.9 0.8 0.7

KSEC

0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

-1

Flow rate (mL min ) FIGURE 9.3 Flow rate independence of KSEC. Narrow dispersity linear PMMA stan-

dards in tetrahydrofuran (THF) at 30 C, with peak-average molar mass Mp in g mole1 of (-) 1.27  103, (C) 4.91  103, (:) 2.70  104, (;) 1.07  105, (=) 2.65  105, ( 0; and at poor conditions A2 \0. This “theta” should not be confused with the scattering angle q, to which it is unrelated. A2 can usually be determined by using the MALS photometer off-line, for so-called batch mode experiments; see Section 9.3.3 of reference [2] for details. 6

At a fundamental level, the calculations rely on the relationship between index of refraction, dielectric constant, and polarizability, as given by the Clausius-Mosotti equation and Maxwell’s theory of radiation.

9.6. SIZE-EXCLUSION CHROMATOGRAPHY ENTERS THE MODERN ERA

213

monodisperse, so that Mw,i z Mn,i z Mz,i z. z Mi. Incorporating the Mi from SLS and the ci from DRI into the Meyerhoff equation, P ci Mxi i ; when x ¼ 0; b ¼ n; when x ¼ 1; b ¼ w; Mb ¼ P ci Mx1 i i

when x ¼ 2; b ¼ z

(9.22)

it can be seen how a SECeSLSeDRI experiment allows the determination of the various M averages of a polymer, as well as of the polymer MMD, on an absolute basis and without the need to construct a calibration curve. These determinations of molar mass are performed most accurately by coupling SEC with LALS, the type of detector employed by Ouano and Kaye in their experiments, because in LALS, the data need not be corrected for angular effects. On the other hand, LALS provides no information about the size of the molecule (as, at q ¼ 0 , the RG,z term vanishes from Eq. (9.18c)). Also, LALS experiments are notoriously sensitive to dust and other particulate matter (e.g., from the shedding of fines from column packing material), which scatters preferentially in the forward direction, that is, at low angles, rendering SECeLALS data inherently noisy and plagued by spikes. It was these drawbacks of LALS (sensitivity to dust, no size information) that prompted the introduction of multiangle static light-scattering (MALS) detection. In MALS, the scattered light is measured at a multiplicity of angles, simultaneously, using a number of individual photodiodes or chargecoupled devices placed at discrete angular intervals around the MALS cell (an example of the placement of photodiodes around a commercial MALS cell is shown in Figure 9.8(a), while the cell itself is shown in Figure 9.8(b)). The combined results are then extrapolated to q ¼ 0 to obtain Mw, while the angular dependence of the scattered light (angular dissymmetry, due to intramolecular interference effects) is used to provide a measure of the size of the molecule, by way of the z-average radius of gyration RG,z.7 Figure 9.9 shows a plot of K c/R(q) versus sin2(q/ 2) for the slice eluting at the SEC peak apex, as measured by the 90 photodiode of the MALS detector, for the broad PS sample from Figure 9.1. Each data marker represents the measurement from each photodiode of the MALS (the angular placement of the photodiodes is 7

Molecules that are small compared to l, the wavelength of radiation in the medium, are considered near-isotropic scatterers, e.g., they scatter almost equally in all directions and, consequently, their RG cannot be determined by MALS. A rule of thumb is that the cutoff for accurate determination of RG by MALS is when RG \l/40, where l h l0/n0.

214

9. SIZE-EXCLUSION CHROMATOGRAPHY

FIGURE 9.8 (a) Read head of a commercial 18-angle MALS unit. (b) Flow cell assembly of 3-angle and 18-angle commercial MALS units. Source: Figures courtesy of Wyatt Technology Corp.

given by the numbers next to the individual data markers), with concentration c provided by the DRI subsequent to correction for interdetector delay. Because, as described previously, each slice eluting from the SEC column(s) is considered to be virtually monodisperse, coupling MALS and a concentration-sensitive detector to SEC allows calculation of the statistical averages and distribution of not just M but also RG.

215

9.6. SIZE-EXCLUSION CHROMATOGRAPHY ENTERS THE MODERN ERA

-6

1.7x10

145.2

152.3

-6

1.6x10

137.2 128.4

-6

1.6x10

118.7

K*c/R(θ )

109.0 -6

1.5x10

99.5 90.0

-6

1.5x10

80.5 71.0

-6

1.5x10

62.4 47.2 40.6 33.7

26.4

-6

1.4x10

21.2

54.9

-6

1.4x10

0.0

0.2

0.4

0.6

0.8

1.0

2

sin (θ/2 ) FIGURE 9.9 Angular variation of scattered light intensity. Sample and experimental conditions as given in legend for Figure 9.1. Data are for SEC slice eluting at peak apex, as monitored by 90 MALS photodiode, for which Mw ¼ 7.33  105 g mole1 and RG,z ¼ 32 nm. Error bars, representing instrumental standard deviation, are substantially smaller than data markers and, therefore, not shown. Solid line represents nonweighted, first-order linear fit to the data, with r2 > 0.999. Numbers next to individual markers denote angle q of measurement, in degrees ( ), subsequent to correction for reflectance at the solventeglass interface of the MALS cell.

Figure 9.10(a) shows the RG distribution of the PS sample with the MMD shown in Figure 9.1. Given in Figure 9.10(b) is the dependence of RG on M. This so-called conformation plot has the ability to inform our knowledge of polymer architecture and conformation. For example, in the present case, the slope of 0.53 is in the range expected for linear random coils at good solventetemperature conditions, 0.5 to 0.6. MALS detection for SEC was introduced in the late-1980s [45]. While it took some time for this detection method to gain acceptance, it is now considered the benchmark to which other determinations are compared. Current commercial offerings include systems that measure scattered light at 2, 3, 7, 8, 18, and 21 angles, simultaneously (the two last systems actually provide measurements of, at most, 17 and 20 angles when the MALS is connected to a SEC system or other type of separation device). As was the case with intrinsic viscosity, a (long-chain-)branched polymer has a smaller RG than its linear counterpart of the same M and composed of the same monomeric repeat unit. As such, a comparison of the RG of the linear and branched molecule, at the same M, can provide an indication of branching. With the MALS detector connected to a SEC system (or other

216

9. SIZE-EXCLUSION CHROMATOGRAPHY

(a)

2.5

2.0 0.8 1.5 0.6 1.0

0.4

0.5

0.2

0.0 100

0.0 1

Cumulative weight fraction

Differential weight fraction

1.0

10

Radius of gyration RG (nm)

Radius of gyration RG (nm)

(b)

100

0.53

10

5

10

6

10

7

10

Molar mass M (g mol-1) FIGURE 9.10 (a) Differential (solid line) and cumulative (open circles) distributions of RG. Sample and experimental conditions as given in legend for Figure 9.1. Both distributions based on first-order polynomial fits to the experimental data. (b) Conformation plot for same sample as in (a), at same experimental conditions (for explanation of scatter at lower M values, see legend to Figure 9.2). Slope is for a nonweighted, first-order linear fit of M data between 2  105 g mole1 and 3  106 g mole1, for which r2 ¼ 0.999.

9.7. SIZE-EXCLUSION CHROMATOGRAPHY TODAY

217

suitable separation method) and assuming a suitable linear standard can be found,8 branching can be determined across the chromatogram and, hence, across the MMD of the polymer. This determination of branching can be made fully quantitative by applying the theory developed by Bruno Zimm and Walter Stockmayer in their classic 1949 paper, “The Dimensions of Chain Molecules Containing Branches and Rings” [46].

9.7. SIZE-EXCLUSION CHROMATOGRAPHY TODAY: MULTIDETECTOR MEASUREMENTS, PHYSICOCHEMICAL CHARACTERIZATION, TWO-DIMENSIONAL TECHNIQUES While multidetector SEC combining DRI, VISC, and MALS has been in use for at least the last two decades, the combination of detectors being employed and the information obtained from them has certainly proliferated in this century [47,48]. The differential viscometer, for one, has grown beyond its role as a universal calibration tool; and like MALS, both these detectors are now also appreciated for the valuable architectural and conformational information they can provide when used together. The addition of quasi-elastic light-scattering (QELS, also known as dynamic light scattering) detection, with the detector usually housed in the same unit as the MALS, added another tool to the arsenal of detectors employed for characterizing physical aspects of macromolecules (consequently, these detectors are often referred to as physical detectors). In addition to long-chain branching information, these detectors can determine the persistence length, characteristic ratio, fractal dimension, and so on of macromolecules across the MMD. Details of how this is done are given in Chapter 11 of reference [2]. Macromolecular behavior is often dictated not only by the topology but also by the chemistry of polymers, with observed conformational differences among physically similar species [49]. This is especially so in the case of copolymers, where the relative amounts of the different monomers, and the arrangement of these monomers both within the chain and across the MMD, influences processing, end-use, and dilute solution properties [50]. To better understand the underlying basis of this behavior, results from the preceding physical detectors have been augmented by the addition of mass spectrometry; ultraviolet, infrared, fluorescence, and nuclear magnetic resonance spectroscopy; conductivity; and similar 8

The requirements for accurately performing long-chain branching calculations, including the suitability qualifications for a linear standard, are described in Section 11.2 of reference [2] and in reference [38].

218

9. SIZE-EXCLUSION CHROMATOGRAPHY

detection. When used in conjunction with MALS and a suitable concentration-sensitive detector, these detectors (often referred to as chemical detectors) can measure how the ratio of the various monomers in a copolymer changes as a function of molar mass, a datum known as chemical heterogeneity, or how tacticity or polyelectrolytic charge changes as a function of M. Ludlow et al. combined various chemical detectors, namely, UV, 1H NMR, ESIeMS, and off-line continuous FT-IR in the study of polymer additives [51], while Striegel and colleagues applied quadruple- and quintuple-detector SEC with a combination of physical and chemical detectors to the study of copolymers and blends [52,53]. For example, using SECeMALSeQELSeVISCeUVeDRI, with all detectors on line, Rowland and Striegel recently determined the chemical heterogeneity in a poly(acrylamide-co-N,N-dimethyl acrylamide) copolymer, the chemical-heterogeneity-corrected molar-mass averages and distribution of the copolymer, its dilute solution conformation and how this conformation changed across the MMD, as well as the physicochemical basis for the observed changes, all in a single analysis [53]. While many characteristics of homo- and copolymers can be characterized by multidetector SEC, a number of other separation techniques provide complementary information [54]. Examples are the chemical composition distribution (CCD) of copolymers, obtained via so-called interactive macromolecular separation methods such as gradient polymer elution chromatography (GPEC) or liquid adsorption chromatography at the critical condition (LACCC) [55e57]. Alternatively, techniques such as hydrodynamic chromatography or field-flow fractionation may be employed to study molecules or particles that are either too large or too fragile to analyze successfully by SEC [20,21,26,27,58]. A corollary of these needs is that, for a more-complete understanding of complex polymers (understood as polymers with distributions in more than one property) and blends, it is necessary to couple SEC to other types of separation methods. In these two-dimensional liquid chromatographic (2D LC) separations, a polymer with distributions in both molar mass and chemical composition may be analyzed by, for example, GPEC  SEC, to determine the combined CCD  MMD of the macromolecule. Figure 9.11 shows the results of an LACCC  SEC analysis of a block copolymer of PMMA and poly(isobornyl acrylate) or PiBoA. The contour plot shows the presence not only of the PMMA-b-PiBoA block copolymer but also of the remaining PiBoA homopolymer that was employed in the copolymer synthesis. Relative abundances are given by the scale bar and, more informatively, by the three-dimensional rendering of each peak, to the right of the two-dimensional chromatogram. Recent reviews of 2D LC of polymers include references [60,61], while the theory and applications 2D LC with SEC as one of the dimensions are treated in Chapter 14 of reference [2]. It can safely be said that

9.8. CONCLUSIONS

219

FIGURE 9.11 Two-dimensional LACCC 3 SEC chromatogram of PMMA-b-PiBoA, showing remaining PiBoA homopolymer after synthesis of the block copolymer. Relative abundances are given by the scale bar and the three-dimensional rendering of each peak, to the right of the two-dimensional chromatogram. See reference [59] for experimental details. Source: Reprinted with permission from reference [59].

multidimensional, multidetector macromolecular separations will be the growth area in polymer chromatography in upcoming years, due to the power of this group of techniques with respect to both peak capacity and to the wealth of information they can provide about the physicochemical phase space occupied by complex polymers and blends.

9.8. CONCLUSIONS Size-exclusion chromatography can be said to have “come of age” in the 1980s with the ability to determine absolute M, as imparted by on-line static light-scattering and viscometric detection (the latter because it permitted construction of universal calibration curves), using robust, commercially available detectors. The next decade was chiefly governed by tripledetector methods involving MALS, VISC, and DRI to determine a number of physical properties, with many studies involving the measurement (with varying levels of accuracy) of long-chain branching across the MMD of both natural and synthetic polymers. During this period, the coupling of SEC to chemical detectors also grew. Today, multidetector SEC experiments may involve three, four, and even five detectors, all on-line, to characterize a wide range of physicochemical properties such as M averages and distributions, chemical and sequence-length heterogeneity, longand short-chain branching, fractal dimension and persistence length, and more. With a growing knowledge of the power of SEC has come the realization of many of its limitations, especially when a more complete

220

9. SIZE-EXCLUSION CHROMATOGRAPHY

characterization of complex polymers and blends is desired [2,62]. To deconvolute from each other the multiple physical and chemical distributions that may be present in these types of materials, two-dimensional separations are usually necessary, and SEC has found a central role in these methods, as well. Indeed, it is in the acceptance and popularization of multidetector, multidimensional techniques that macromolecular separation science can be expected to grow and demonstrate its full power in the upcoming years. This is likely to be accompanied by advances in the synthesis of stationary phases specifically tailored for interactive macromolecular separations and by continued computer modeling and simulations of the various distributions and heterogeneities present in copolymers and related materials [42,50,63].

ACKNOWLEDGEMENT AND DISCLAIMER I am most grateful to Dr. Philip J. Wyatt for insightful discussions on the early days of interfacing MALS to separation systems in general and SEC in particular. The identification of certain commercial equipment, instruments, or materials does not imply recommendation or endorsement by the National Institute of Standards and Technology. These identifications are made only to specify the experimental procedures in adequate detail.

References [1] Rudin A. The elements of polymer science and engineering. New York: Academic Press; 1982. [2] Striegel AM, Yau WW, Kirkland JJ, Bly DD. Modern size-exclusion liquid chromatography. 2nd ed. New York: Wiley; 2009. [3] Moore JC. Gel permeation chromatography. I. A new method for molecular weight distribution of high polymers. J Polym Sci A 1964;2:835. [4] Wheaton RM, Bauman WC. Non-ionic separations with ion exchange resins. Ann NY Acad Sci 1953;57:388. [5] Porath J, Flodin P. Gel filtration: A method for desalting and group separation. Nature 1959;183:1657. [6] Striegel AM. Chapter 1. In: Striegel AM, editor. Multiple detection in size-exclusion chromatography. ACS Symp Ser 893. Washington, DC: American Chemical Society; 2005. p. 2. [7] Moore JC. Gel permeation chromatography: Its inception. J Polym Sci C 1968;21:1. [8] Ouano AC. J Polym Sci A-1 1972;10:2169. [9] McDonald PD. Waters Corporation: Fifty years of innovation in analysis and purification. Chem Heritage Summer 2008:33. [10] McDonald PD. James Waters and his liquid chromatography people: a personal perspective. www.waters.com. [11] Striegel AM. Anomeric configuration, glycosidic linkage, and the solution conformational entropy of O-linked disaccharides. J Am Chem Soc 2003;125:4146 (see Erratum in J. Am. Chem. Soc. 2004:126:4740).

REFERENCES

221

[12] Boone MA, Striegel AM. Influence of anomeric configuration, degree of polymerization, hydrogen bonding, and linearity versus cyclicity on the solution conformational entropy of oligosaccharides. Macromol 2006;39:4128. [13] Boone MA, Nymeyer H, Striegel AM. Determining the solution conformational entropy of O-linked oligosaccharides as quasi-physiological conditions: size-exclusion chromatography and molecular dynamics. Carb Res 2008;343:132. [14] Buley TD, Striegel AM. Relation between the delta 2 effect and the solution conformational entropy of aldohexoses and select methyl glycosides. Carbo Polym 2010;79:241. [15] Striegel AM, Boone MA. Influence of glycosidic linkage on solution conformational entropy of oligosaccharides: Malto- vs. isomalto- and cello- vs. laminarioligosaccharides. Biopolym 2011;95:228. [16] Richard DJ, Striegel AM. The obstruction factor in size-exclusion chromatography. 1. The intraparticle obstruction factor. J Chromatogr A 2010;1241:7131. [17] Little JN, Waters JL, Bombaugh KJ, Pauplis WJ. Fast gel-permeation chromatography. I. A study of operational parameters. J Polym Sci A-2 1969;7:1775. [18] Yau WW, Malone CP, Fleming SW. The equilibrium distribution coefficient in gel permeation chromatography. J Polym Sci B: Polym Lett 1968;6:803. [19] Striegel AM. Thermodynamic equilibrium of the solute distribution in size-exclusion chromatography. J Chromatogr A 2004;1033:241. [20] Striegel AM. Hydrodynamic chromatography: packed columns, multiple detectors, and microcapillaries. Anal Bioanal Chem 2012;402:77. [21] Striegel AM, Brewer AK. Hydrodynamic chromatography. Ann Rev Anal Chem 2012;5:15. [22] Uliyanchenko E, van der Wal S, Schoenmakers PJ. Deformation and degradation of polymers in ultra-high-pressure liquid chromatography. J Chromatogr A 2011;1218:6930. [23] Striegel AM. Longitudinal diffusion in size-exclusion chromatography: a stop-flow size-exclusion chromatography study. J Chromatogr A 2001;932:21. [24] Striegel AM. Observations regarding on-column, flow-induced degradation during SEC analysis. J Liq Chromatogr Rel Techn 2008;31:3105. [25] Striegel AM, Isenberg SL, Coˆte´ GL. An SEC/MALS study of alternan degradation during size-exclusion chromatographic analysis. Anal Bioanal Chem 2009;394:1887. [26] Isenberg SL, Brewer AK, Coˆte´ GL, Striegel AM. AM. Hydrodynamic versus size exclusion chromatography characterization of alternan and comparison to off-line MALS. Biomacromol 2010;11:2505. [27] Brewer AK, Striegel AM. Characterizing string-of-pearls colloidal silica by multidetector hydrodynamic chromatography and comparison to multidetector size-exclusion chromatography, off-line multiangle static light scattering, and transmission electron microscopy. Anal Chem 2011;83:3068. [28] Giddings JC. Dynamics of chromatography. New York: Marcel Dekker; 1965. [29] Giddings JC. Unified separation science. New York: Wiley; 1991. [30] Snyder LR, Kirkland JJ, Dolan JW. Introduction to modern liquid chromatography. 3rd ed. New York: Wiley; 2010. [31] Karger BI, Snyder LR, Horvath C. An introduction to separation science. New York: Wiley; 1973. [32] Yau WW, Kirkland JJ, Bly DD, Stoklosa HJ. Effect of column performance on the accuracy of molecular weights obtained from size exclusion chromatography (gel permeation chromatography). J Chromatogr 1976;125:219. [33] Grubisic Z, Rempp P, Benoit H. A universal calibration for gel permeation chromatography. J Polym Sci Polym Lett 1967;5:753. [34] Benoit H, Grubisic Z, Rempp P, Decker D, Zilliox J- G. E´tude par chromatographie en phase liquid de polystyre`nes line´aires et ramifie´s de structures connues. J Chim Phys 1966;63:1507.

222

9. SIZE-EXCLUSION CHROMATOGRAPHY

[35] Benoit HC. Reflections on “A universal calibration method for gel permeation chromatography.” J Polym Sci 1967;5:753. [36] Haney MA. The differential viscometer. I. A new approach to the measurement of specific viscosities of polymer solutions. J App Polym Sci 1985;30:3023. [37] Haney MA. The differential viscometer. II. On-line viscosity detector for size-exclusion chromatography. J App Polym Sci 1985;30:3037. [38] Striegel AM. In: Cazes J, editor. Encyclopedia of chromatography. 3rd ed. New York: Taylor & Francis; 2010. p. 1417. [39] Dubin PL, Principi JM. Failure of universal calibration for size-exclusion chromatography of rodlike macromolecules versus random coils and globular proteins. Macromol 1989;22:1891. [40] Pannell J. Gel permeation chromatography: the behavior of polystyrenes with longchain branching. Polym 1972;13:277. [41] Temyanko E, Russo PS, Ricks H. H. Study of rodlike homopolypeptides by gel permeation chromatography with light scattering detection: Validity of universal calibration and stiffness assessment. Macromol 2001;34:582. [42] Striegel AM, Plattner RD, Willett JJ. Dilute solution behavior of dendrimers and polysaccharides: SEC, ESI-MS, and computer modeling. Anal Chem 1999;71:978. [43] Wyatt PJ. Light scattering and the absolute characterization of macromolecules. Anal Chim Acta 1993;272:1. [44] A.C. Ouano AC, W.Kaye W. Gel-permeation chromatography, X. Molecular weight detection by low-angle laser light scattering. J Polym Sci Polym Chem Ed, 1974:12:1151. [45] (a) Wyatt PJ, Jackson C, Wyatt GK. Absolute GPC determinations of molecular weights and sizes from light scattering. Am Lab 1988:20:86. (b) Wyatt PJ, Hicks DL, Jackson C, Wyatt GK. Absolute GPC determinations of molecular weights and sizes. Am Lab 1988:20:108. [46] Zimm BH, Stockmayer WH. The dimensions of chain molecules containing branches and rings. J Chem Phys 1949;17:1301. [47] Striegel AM, editor. Multiple detection in size-exclusion chromatography. ACS Symp Ser, 893. Washington, DC: American Chemical Society; 2005. [48] Striegel AM. Multiple detection in size-exclusion chromatography of macromolecules. Anal Chem 2005;77:104A. [49] Haidar Ahmad IA, Striegel AM. Influence of second virial coefficient and persistence length on dilute solution polymer conformation. Anal Bioanal Chem 2011;399:1515. [50] Haidar Ahmad IA, Striegel DA, Striegel AM. How does sequence length heterogeneity affect the dilute solution conformation of copolymers? Polym 2011;52:1268. [51] Ludlow M, Louden D, Handley A, Taylor S, Wright B, Wilson ID. Size-exclusion chromatography with on-line ultraviolet, proton nuclear magnetic resonance and mass spectrometric detection and on-line collection for off-line Fourier transform infrared spectroscopy. J Chromatogr A 1999;857:89. [52] Haidar Ahmad IA, Striegel AM. Determining the absolute, chemical-heterogeneitycorrected molar mass averages, distribution, and solution conformation of random copolymers. Anal Bioanal Chem 2010;396:1589. [53] Rowland SM, Striegel AM. Characterization of copolymers and blends by quintupledetector size-exclusion chromatography. Anal Chem 2012;84:4812. [54] Striegel AM. Separation science of macromolecules. Anal Bioanal Chem 2011;399:1399. [55] Striegel AM. Determining the vinyl alcohol distribution in poly(vinyl butyral) using normal-phase gradient polymer elution chromatography. J Chromatogr A 2002;971:151. [56] Striegel AM. Determining and correcting “moment bias” in gradient polymer elution chromatography. J Chromatogr A 2003;996:45.

REFERENCES

223

[57] Philipsen HJA. Determination of chemical composition distribution in synthetic polymers. J Chromatogr A 2004;1037:329. [58] Podzimek S. Light scattering, size exclusion chromatography and asymmetric flow field flow fractionation. New York: Wiley; 2011. [59] Inglis AJ, Barner-Kowollik C. Visualizing the efficiency of rapid modular block copolymer construction. Polym Chem 2011;2:126. [60] Berek D. Two-dimensional liquid chromatography of synthetic polymers. Anal Bioanal Chem 2010;396:421. [61] Baumgaertel A, Altunas E, Schubert US. Recent developments in the detailed characterization of polymers by multidimensional chromatography. J Chromatogr A 2012;1240:1. [62] Berek D. Size exclusion chromatography - A blessing and a curse of science and technology of synthetic polymers. J Sep Sci 2010;33:315. [63] Dong S, Striegel AM. Monte Carlo simulation of the sequence length and junction point distributions in random copolymers obeying Bernoullian statistics. Int J Polym Anal Charac 2012;17:247.

This page intentionally left blank

C H A P T E R

10

Solvent Selection in Liquid Chromatography G. Ramis-Ramos, M.C. Garcı´a-A´lvarez-Coque Department of Analytical Chemistry, University of Vale`ncia, Spain

O U T L I N E 10.1. Elution Strength

226

10.2. Columns and Solvents in RPLC, NPLC, and HILIC

228

10.3. Assessment of the Elution Strength 10.3.1. The Hildebrand Solubility Parameter and Other Global Polarity Estimators 10.3.2. Global Polarity for Solvent Mixtures

229

10.4. Schoenmakers’s Rule

231

10.5. Isoeluotropic Mixtures

233

10.6. Solvent-Selectivity Triangles 10.6.1. The Snyder’s Solvent-Selectivity Triangle 10.6.2. Prediction of the Character of Solvent Mixtures 10.6.3. A Solvatochromically Based Solvent Selectivity Triangle

234 234 239 240

10.7. Practical Guidelines for Optimization of Mobile Phase Composition 10.7.1. Selection of the Chromatographic Mode 10.7.2. Systematic Trial-and-Error Mobile-Phase Optimization for Isocratic Elution

Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00010-9

225

229 231

241 241 242

Copyright Ó 2013 Elsevier Inc. All rights reserved.

226

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

10.7.3. Systematic Trial-and-Error Mobile-Phase Optimization for Gradient Elution 10.7.4. Computer-Assisted Interpretive Optimization

244 245

10.8. Additional Considerations for Solvent Selection

246

References

248

10.1. ELUTION STRENGTH In liquid chromatography (LC), the elution strength is the ability of the mobile phase to sweep away the solutes retained on the stationary phase. This strength depends on the nature of the stationary phase and solutes, as well as on the mobile phase composition (i.e., nature and concentration of the solvents and additives), pH, and column temperature. Therefore, the elution strength is not a property exclusively related to the solvent, since solutes undergo different elution strengths depending on their particular molecular structures. In spite of this, the elution strength is a very practical concept in LC, commonly used to adjust the overall retention for a group of solutes inside the target retention region (optimally, in the 1 < k < 5 range, or at least 0.2 < k < 20, k being the retention factor). For a given stationary phase and set of solutes, if the elution strength is too high, retention times will be too short, and consequently, the resolution will be poor. Conversely, if the elution strength is too low, retention times will be excessive, and consequently, the analysis time will be too long and the signal-to-noise ratio at the peak maxima will decrease significantly. Once the elution strength has been adjusted, the selectivity (i.e., elution order and peak distribution) can be optimized without modifying significantly the overall retention. In addition to water, many organic solvents can be used to prepare the mobile phase (Table 10.1). Also, it is possible to use mixtures of solvents in different ratios, to modify the solvent properties (e.g., the elution strength and selectivity). This can make solvent selection for a given purpose a difficult task, unless suitable guidelines are followed. The purpose of this chapter is to summarize the most common strategies used by skilled chromatographers. Although mostly developed and used for reversedphase liquid chromatography (RPLC), the guidelines should be useful for normal-phase liquid chromatography (NPLC) as well, including the aqueous-compatible normal mode known as hydrophilic interaction liquid chromatography (HILIC). The elution strength can be either maintained constant (isocratic elution) or gradually increased (gradient elution).

227

10.1. ELUTION STRENGTH

TABLE 10.1 Solvent Properties

Solvent

Normal boiling point (oC)a

Cutoff wavelength (nm)a

Viscosity at 20
C (mPa$sec)a

Solubility parameter, db

Snyder global polarity, P’c

Isooctane

99.2

200e210

0.50

7.0

e0.4

Diisopropyl ether

68.0

380

0.33

7.1

1.8

n-Heptane

98.4

200

0.42

~7.5

0.0

n-Hexane

68.7

200

0.31

~7.5

0.0

Triethylamine

89.5

235

0.38

7.5

1.8

Cyclohexane

80.7

200

0.98

8.2

0.0

Carbon tetrachloride

76.8

263

0.97

8.6

1.7

Ethyl acetate

77.1

256

0.46

8.9

4.3

110.6

284

0.59

8.9

2.3

Tetrahydrofuran

66.0

212

0.55

9.1

4.2

Chloroform

61.2

245

0.58

9.2

4.4

Dichloromethane

40.0

232

0.44

9.6

4.3

Methyl ethyl ketone

79.6

329

0.42 (15
C)

9.5

4.5

Acetone

56.3

330

0.30 (25
C)

9.6

5.4

Carbon disulfide

46.0

220

0.36

10.0

1.1

Toluene




1,4-Dioxane

101.3

215

1.44 (15 C)

10.1

4.8

Pyridine

115.3

330

0.95

10.6

5.3




82.3

205

2.86 (15 C)

11.4

4.3

1-Butanol

117.7

215

2.95

11.6

3.9

2-Methoxyethanol

124.6

210

1.72

11.7

5.7

Dimethylformamide

153.0

268

0.92

11.8

6.4

78.3

205e210

1.2

12.0

5.2

189.0

286

2.20

12.0

6.5

Isopropanol

Ethanol Dimethyl sulfoxide

(Continued)

228

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

TABLE 10.1 Solvent Propertiesdcont’d Normal boiling point (oC)a

Solvent

Cutoff wavelength (nm)a

Viscosity at 20
C (mPa$sec)a

Solubility parameter, db 12.1

6.2

12.2

3.9

Acetonitrile

81.6

190

0.34

1-Propanol

97.2

210

2.26


Snyder global polarity, P’c

117.9

210

1.31 (15 C)

13.0

6.2

64.7

205

0.55

14.5

6.6

Formamide

210.5

210

3.5

19.2

7.3

Water

100.0

1.00

23.5

9.0

Acetic acid Methanol

dX > 10 are properly eluted using a 0 to 100% ACN gradient. Less polar solutes, going down to dX z 8.5, are eluted by substituting ACN with THF. Similarly, the polarity range of solutes properly eluted from a silica column with heptaneeisopropanol mixtures in NPLC is depicted in Figure 10.1(b). As observed, the solute polarity range is narrower in NPLC with regard to RPLC, approximately, 11.5 < dX < 13.5. Finally, in HILIC, where solutes are retained on a water layer (dS z 23.5, Figure 10.1(c)), highly polar solutes in the 18 < dX < 21 range (mainly, ions, polyions, or zwitterions) are eluted with watereACN mixtures by increasing water from 5% to 50%.

10.5. ISOELUOTROPIC MIXTURES Fine tuning of the polarity through discrete or continuous changes of the mobile phase composition in the isocratic and gradient elution modes, respectively, is mainly achieved by adjusting the modifier concentration. On the other hand, the selectivity is controlled by changing the composition of the solvent mixture, and for some solutes, by also modifying the mobile phase pH or column temperature. The selectivity depends mainly on the specific interactions of solutes with the stationary and mobile phases, that is, on the profile of the contributions to the global polarity of solutes and phases. A basic question in selectivity optimization is how to modify the nature of a solvent mixture without altering the selected elution strength. Mixtures with the same elution strength but prepared with different modifiers are called isoeluotropic mixtures. For binary mixtures of MeOH, ACN, or THF with water, from Eq. (10.2), dMeOH 4MeOH þ dH2O ð1  4MeOH Þ ¼ dACN 4ACN þ dH2O ð1  4ACN Þ ¼ dTHF 4THF þ dH2O ð1  4THF Þ (10.6) By substituting the polarity values given in Table 10.1, 4MeOH ¼ 1:27 4ACN ¼ 1:60 4THF

(10.7)

Hence, the elution strength of an aqueous mobile phase with 20% MeOH is approximately the same as for 15.7% ACN or 12.5% THF. Since THF is the most hydrophobic solvent, the same elution strength is achieved with a smaller percentage of organic solvent. As indicated

234

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

previously, the predictions of elution strength depart from linearity at large modifier concentrations. To address this problem, nonlinear relationships and nomograms, such as that shown in Figure 10.2, can be used. On this nomogram, all possible isoeluotropic binary mixtures constituted by water and either ACN, MeOH, or THF can be estimated. ACN is generally stronger than MeOH, and THF appreciably stronger than ACN. Note that the scale for ACN is linear, making it necessary to draw nonlinear scales for MeOH and THF. However, due to the limitations inherent in the global polarity parameters, predictions are rough and depend largely on the solute type.

10.6. SOLVENT-SELECTIVITY TRIANGLES 10.6.1. The Snyder’s Solvent-Selectivity Triangle Mobile phase selectivity is understood as a consequence of the particular profile of the contributions of solventesolvent intermolecular interactions to the global polarity. Six types of interactions contribute to the Hildebrand solubility parameter: interactions between permanent dipoles, between induced dipoles, between permanent and induced dipoles, hydrogen ion donation (acidity), hydrogen ion acceptance (basicity), and electrostatic interactions. Owing to the different contributions, if solutes with exactly the same global polarity but structural differences are separated by chromatography, retention times will be close but still different. We could add “fortunately different,” because otherwise selectivity optimization would not be possible. To deal with up to six parameters, multivariate statistics is required; however, in the strategy proposed by Snyder in 1974 [2], electrostatic interactions are neglected and some of the most akin interactions (among permanent and induced dipoles) are summarized in a single

ACN-water 0

20

10

30

40

50

60

70

80

90

100

MeOH-water 0

20

0

10

20

30

40

100

80

60

40

50

60

70

THF-water 80

90

100

FIGURE 10.2 Nomogram showing isoeluotropic binary mixtures in RPLC. The compositions are obtained by connecting the solvent scales with a vertical line; the example indicates that aqueous binary mixtures having 60% ACN, 70% MeOH, or 46% THF are isoeluotropic. Source: Adapted from Sigma-Aldrich.com/Supelco 2009e2010 chromatography products catalog, p. 38.

235

10.6. SOLVENT-SELECTIVITY TRIANGLES

property called dipolarity (i.e., polarity and polarizability). Accordingly, mobile phase selectivity is characterized by only three parameters: acidity, basicity, and dipolarity. This made possible plotting solvent properties on a triangular diagram, called the solvent-selectivity triangle (SST), where each corner represents one of the properties (Figure 10.3). The solvent properties were estimated using three probes: ethanol (e), 1,14-dioxane (d), and nitromethane (n), which were formerly proposed by Rohrschneider to represent each one of the intended properties: “hydrogen ion donor” (ethanol), “hydrogen ion acceptor” (1,14-dioxane), and “polar or polarizable” (nitromethane). In fact, none of the three probes represents these characteristics uniquely: Ethanol is predominantly a hydrogen ion donor but also a weak acceptor and is moderately dipolar; 1,14-dioxane is a good hydrogen ion acceptor, weakly dipolar and a non-hydrogen-ion donor; and nitromethane is strongly dipolar but also both weakly acidic and weakly basic. Although far from ideal, the selected probes led to a useful classification of solvents.

Basic 0.1

0.7 Triethylamine

0.2

0.6

isoPrOH

xe

MeOH

Diisopropyl ether

II

0.3

0.5 I

H2O HAcO

0.4

THF

DMF

xd

IV

CHCl3

VIII

0.5

Acidic 0.6 0.2

III

CS2 0.3

0.4

0.4

Ethyl acetate

VI ACN

0.3 V

CH2Cl2 VII 0.5

xn

0.6

0.2

Dipolar

FIGURE 10.3 Snyder’s solvent-selectivity triangle, indicating the eight solvent families (large circles). The location of several solvents, including those most commonly used in RPLC and NPLC, is indicated (DMF, dimethylformamide; HAcO, acetic acid; isoPrOH, isopropanol). The arrows starting from chloroform illustrate how to read the scales.

236

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

Solvents were characterized according to their capacity to interact with the three probes, which was estimated from gaseliquid partition equilibria. Snyder’s global polarity, P0 (Tables 10.1), was defined as the sum of the three contributions: 0

0

0

P 0 ¼ log ke þ log kd þ log kn 0

0

(10.8)

0

where ke ; kd ; and kn are the gaseliquid partition coefficients for the probes, which were determined from their equilibrium concentrations in a sealed vial, containing a fixed volume of solvent to be characterized. The partition coefficients were defined as the ratio of the solute concentration in the solvent and in the vial void volume, after making two corrections to eliminate the effect of the solvent volume and the nonspecific contributions (CeH weak permanent or induced dipole interactions, obtained with n-octane). Finally, to eliminate the differences among the global polarities of the solvents, normalization was performed: 0

0

0

log ke log km log kn þ þ ¼ xe þ x d þ x n 1 ¼ P0 P0 P0

(10.9)

where xe represents the basic character, xd is the acidic character, and xn is the dipolar character of the solvent (Table 10.2). It is therefore assumed that a solvent that strongly retains ethanol or 1,14-dioxane should have a predominantly basic and acidic character, respectively; and a solvent that strongly retains nitromethane has a strong polar character or is readily polarizable. The xe, xd, and xn data for a large number of solvents are plotted on the SST (Figure 10.3). Solvents are grouped according to their properties in eight families: (I) aliphatic ethers and amines; (II) aliphatic alcohols; (III) pyridine and THF; (IV) glycols and acetic acid; (V) dichloromethane and dichloroethane; (VI) aliphatic ketones, esters, 1,4-dioxane, and nitriles; (VII) aromatic hydrocarbons and nitrocompounds; and (VIII) phenols and water. The scales should be read counterclockwise: xe is represented on the right side (the higher on the scale, the stronger is the basic character of the solvent), xd is on the left side (the lower on the scale, the stronger is the acidic character), and xn is on its base (with the solvent dipolarity increasing to the right). The diagram shows that the most common solvents in RPLC provide different selectivity, since they have rather different profiles of the three properties defined in the SST: (a) water is a strong hydrogen ion donor and acceptor (it is situated at half-height in the SST) but a weak dipole (it is on the left); (b) ACN is less acidic than water but appreciably more dipolar; (c) MeOH is appreciably more basic (higher in the diagram), more dipolar than water, and less dipolar than ACN; (d) THF has both acidic and basic character, but it is more dipolar than water.

TABLE 10.2 Normalized Selectivity Factors Derived from gaseliquid partition data of Rohrschneider’s probesb Solventa

xd

Diisopropyl ether

0.10

xe

d

xn

0.51 e

d

0.39 e

d

Derived from Kamlet-Taft solvatochromic parametersc a

b

p*

0.00

0.64

0.36

0.00

0.00

0.00

d

Carbon disulfide

0.39

0.22

0.39

0.00

0.10

0.90

Triethylamine

0.07

0.61

0.32

0.00

0.84

0.16

d

0.59

Carbon tetrachloride

0.38

0.30

0.32

0.00

d

Ethyl acetate

0.23

0.34

0.43

0.00

0.45

0.55

Toluene

0.28

0.25

0.47

0.00

0.17

0.83

Tetrahydrofuran

0.20

0.38

0.42

0.00

0.49

0.51

Chloroform

0.35

0.31

0.34

0.43

0.00

0.57

Dichloromethane

0.33

0.27

0.40

0.30

0.09

0.78

d

d

d

d

d

Methyl ethyl ketone

0.22

0.35

0.43

d

Acetone

0.23

0.35

0.42

0.06

0.38

0.56

Carbon disulfide

0.39

0.22

0.39

0.00

0.10

0.90

1,4-Dioxane

0.24

0.36

0.40

0.00

0.40

0.60

Pyridine

0.22

0.41

0.36

0.00

0.42

0.58

237

(Continued)

10.6. SOLVENT-SELECTIVITY TRIANGLES

Hexane

Derived from gaseliquid partition data of Rohrschneider’s probesb

Derived from Kamlet-Taft solvatochromic parametersc

xd

xe

xn

a

b

p*

Isopropanol

0.19

0.55

0.27

0.35

0.43

0.22

1-Butanol

0.19

0.59

0.25

0.37

0.41

0.22

d

d

dd

d

2-Methoxyethanol

0.24

0.38

0.38

d

Dimethylformamide

0.21

0.39

0.40

0.00

0.44

0.56

Ethanol

0.19

0.52

0.29

0.39

0.36

0.25

Dimethylsulfoxide

0.27

0.35

0.38

0.00

0.43

0.57

Acetonitrile

0.27

0.31

0.42

0.15

0.25

0.60

1-Propanol

0.19

0.54

0.27

0.36

0.40

0.24

Acetic acid

0.31

0.39

0.30

0.54

0.15

0.31

Methanol

0.22

0.48

0.31

0.43

0.29

0.28

Formamide

0.33

0.38

0.30

0.33

0.21

0.46

Water

0.37

0.37

0.25

0.43

0.18

0.45

a

Solvents ordered according to Table 10.1. b Large values of xd, xe, and xn denote good hydrogen ion donor, good hydrogen ion acceptor, and large permanent or induced dipole moments, respectively [2,11, and 29]. c a, b, and p* represent solvent ability to interact as hydrogen ion donor, hydrogen ion acceptor, and by polar and polarization effects, respectively [14 and 15]. d Not available.

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

Solventa

238

TABLE 10.2 Normalized Selectivity Factorsdcont’d

10.6. SOLVENT-SELECTIVITY TRIANGLES

239

The SST scales should not be interpreted as “percentages” of the intended properties, since solvent properties were obtained from solutes with a mixed character, and therefore the vertices do not represent “pure” properties. For example, a strongly basic solvent such as triethylamine is not located close to the upper vertex due to its basicity but because it strongly retains ethanol and weakly retains 1,4-dioxane and nitromethane. Ideally, if the SST scales would correspond to pure properties (each vertex representing 100% acidity, 100% basicity, and 100% dipolarity), mixtures of three hypothetical solvents, each one located at each vertex, would provide a whole universe of possibilities. However, such solvents do not exist. Furthermore, real solvents located close to the SST vertices are not mutually miscible or are not compatible with common stationary phases. ACN, MeOH, and THF are at intermediate locations in the SST, being excellent choices to achieve a wide range of properties in RPLC. Not surprisingly, these solvents were already popular by the time the SST was developed.

10.6.2. Prediction of the Character of Solvent Mixtures The SST allows predicting whether the elution strength will increase or decrease for certain solutes when one modifier is replaced by another. For example, substituting a MeOHewater mixture with an isoeluotropic ACNewater mixture will reduce the ability of the mobile phase to accept hydrogen ions, so the elution strength will be reduced for acidic solutes. Simultaneously, the dipolar character of the mobile phase will increase so that dipolar and polarizable compounds will elute earlier. This reasoning can be of help in solute identification. Thus, if a solute elutes earlier when a MeOHewater mixture is substituted with an isoeluotropic ACNewater mixture, then, the solute should have a basic or a dipolar character or both. As shown in the SST of Figure 10.4, the character of all possible mixtures of water, ACN, MeOH, and THF is delimited by straight lines connecting the four solvents. This figure illustrates how wide the selectivity range in RPLC is. The character of isoeluotropic mixtures of the four solvents, at increasing elution strength, is indicated by the three small a, b, and c triangles. The location of these isoeluotropic mixtures on the SST was established according to their compositions obtained from the nomogram of Figure 10.2. A linear variation of the properties with modifier concentration is also assumed. The small triangles illustrate how the character of a mixture of solvents is modified by varying its composition, while maintaining a constant elution strength.

240

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

0.2

0.6 Basic 100% MeOH

0.3

xd 0.4

xe

39% H2O

21% 46% 72%

a

30%

b

100% THF

c

0.4

60%

Acidic

0.5

0.5

70%

100% ACN 0.3

0.4

xn

0.5

Dipolar

0.3

FIGURE 10.4 Snyder’s solvent-selectivity triangle indicating the character of mixtures of water, ACN, MeOH, and THF. The small triangles a, b, and c describe isoeluotropic mixtures at increasing elution strength. In a, the lowest vertex corresponds to 30:70 ACNewater, the upper vertex to 39:61 MeOHewater, and the left vertex to 21:79 THFewater. Other points on the sides of the small triangle a correspond to ternary mixtures, and points inscribed in a correspond to quaternary mixtures. Similarly, the small triangles b and c correspond to isoeluotropic mixtures with respect to 60:40 ACNewater and 100% ACN, respectively.

10.6.3. A Solvatochromically Based Solvent Selectivity Triangle According to the “mixed” character of the probes used to construct the SST, xe reflects, in fact, a composite of hydrogen bond basicity, hydrogen bond acidity, and dipolarity; xd reflects a composite of solvent acidity and dipolarity; and xn reflects predominantly solvent dipolarity with small contributions from hydrogen bond basicity and acidity. In 1989, Rutan et al. [11] substituted the gaseliquid partition coefficients obtained with Rohrschneider’s probes by the Kamlet-Taft “solvatochromic parameters” (Table 10.2). These parameters, mainly derived from spectroscopic measurements, separately estimate the hydrogen bond donor (a), hydrogen bond acceptor (b), and dipolarity/polarizability (p*) properties of solvents as contributors to the global solvent polarity. Solvatochromic parameters are averages over results obtained with several probes; then, it is normally assumed that they provide more “pure” measurements of the addressed properties than gaseliquid partition coefficients derived from only three probes. Reconstruction of the SST using normalized solvatochromic parameters led to diagrams that are of interest from a theoretical point of view; however, they do not provide practical advantages

10.7. PRACTICAL GUIDELINES FOR OPTIMIZATION

241

with regard to the classical Snyder’s SST. Also, it should be noted that the essential conclusion of SST diagrams, independent of the approach used to construct them, is that, to explore the full range of possibilities during mobile-phase selectivity optimization, solvents with maximal differences in their properties should be selected.

10.7. PRACTICAL GUIDELINES FOR OPTIMIZATION OF MOBILE PHASE COMPOSITION 10.7.1. Selection of the Chromatographic Mode The optimization of the modifier type and volume fraction in the mobile phase is frequently performed on a trial-and-error basis. Next, some guidelines to rationalize and speed up this process are given. After selecting the chromatographic mode (e.g., RPLC, NPLC, or HILIC) and deciding between isocratic or gradient elution, the elution strength should be adjusted, and finally, the selectivity optimized until all peak pairs of interest are resolved. To select the chromatographic mode, two criteria are important: 1. Solute nature. If the solute molecules contain extensive hydrophobic regions in “external” structural parts, they are retained on the hydrophobic RPLC stationary phases. In contrast, if the influence of ionic or polar groups (e.g., COOH, OH, or NH2) predominates, the solute experiences poor retention and requires polar stationary phases typical in NPLC. Permanent ions and other highly polar solutes are not compatible with NPLC mobile phases, HILIC could be the correct choice. 2. Sample compatibility with the mobile phase. Direct injection of samples soluble in water or in hydro-organic mixtures (e.g., serum, urine, and other aqueous samples or aqueous extracts) require RPLC or HILIC. If HILIC is selected, the elution strength should be decreased by evaporation of water in the sample, followed by redissolution in a rich ACN mixture, or by dilution with ACN at the cost of a higher limit of detection. For hydrophobic samples (oils, greases, hydrocarbons, or extracts in heptane, dichloromethane, or other hydrophobic solvents), NPLC is needed. Extracts in solvents that provide high elution strength, such as ethyl acetate in NPLC or isopropanol in both RPLC and NPLC, should be avoided. However, it is often possible to change the solvent initially used to extract the sample. For instance, an aqueous sample can be extracted with heptane or chloroform, a vegetable oil can be extracted with an aqueous buffer or MeOH, and compounds of interest in an environmental aqueous sample can be concentrated on a solid phase, followed by elution with an appropriate solvent. Therefore, within the limits of the analyte’s solubility or

242

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

stability, it is possible to change the solvent nature by evaporation and dilution to make the medium compatible with a given chromatographic mode.

10.7.2. Systematic Trial-and-Error Mobile-Phase Optimization for Isocratic Elution Isocratic elution can be selected if the polarities of the compounds in the sample are similar. In contrast, if the polarities span a wide range then, gradient elution is needed. For an unknown problem, it is preferable to start the optimization in the gradient elution mode. However, we focus first on the development of an isocratic method. Usually, in RPLC, a C18 stationary phase is tried first. If no previous information about solute polarities is available, starting with a mobile phase of high elution strength, such as 95% ACN, is advisable. This ensures elution of most compounds in the sample, although many may elute close to the dead time. If the retention of one or more solutes is still too high (k > 20), NPLC is probably preferable, although other options are changing the C18 column for C8 or C4 columns or using a higher column temperature. Next, the retention of solutes eluting close to the dead time should be increased by using progressively smaller modifier concentrations (e.g., 60, 40, and 20%). At this stage, gradient elution is probably necessary if all solutes of interest cannot be moved to the target retention factor range. An analogous strategy can be followed by using NPLC: Initially, a polar column (e.g., silica) and a mobile phase with high elution strength are selected. However, the chromatographer should be aware that, in NPLC, a few parts percent of a polar modifier added to the alkane in the mobile phase can cause dramatic effects on retention. For instance, a smaller increase in retention can be produced by decreasing the ethyl acetate concentration from 40% to 2% than from 2% to 0%. This is because, contrary to RPLC, where the “strong” solvent is water and not the modifier, in NPLC, the “strong” solvent, which mainly determines the solvating properties of the mixture, is the modifier. Therefore, in NPLC with moderate modifier concentrations, most solutes probably elute close to the dead time. In the absence of excessively retained solutes, the elution strength should be progressively reduced by decreasing the amount of modifier until appropriate retention factors are obtained. Similarly, for HILIC, aqueous mixtures containing up to 50% water can be initially tried, followed by the stepwise reduction of the water concentration. In the three chromatographic modes (RPLC, NPLC, and HILIC), the selectivity can be further optimized to improve the resolution between all peak pairs. For this purpose, solvent mixtures of similar elution strength,

10.7. PRACTICAL GUIDELINES FOR OPTIMIZATION

243

another pH or column temperature, or if necessary, a different stationary phase, can be tried. Here, we will discuss the selection of the isoeluotropic mixture. This may be based on solute properties guided by the SST. For example, in the RPLC elution of two solutes with the same retention in RPLC but one more than the other, the former elutes earlier if ACN is replaced by MeOH. However, often solute properties are not known or the interpretation of the possible solute-solvent interactions in multifunctional solutes is not straightforward. Therefore, the selectivity is most frequently optimized in an empirical fashion. In RPLC, the first modifier to be trialed is ACN, due to its low viscosity and short UV cutoff wavelength (190 nm) (Table 10.1), which allow a low backpressure and a wide UV detection window. If the separation is not satisfactory, the second option is MeOH. The viscosity of MeOHewater mixtures is much higher than for ACNewater mixtures, with a maximum at 40% MeOH, which makes them unsuitable for working at high flow rates, with long packed columns, or small particle sizes. Also, the cutoff wavelength of MeOH is higher (205 nm). The third option, THF, has a higher viscosity, a cutoff wavelength of 212 nm, and requires long equilibration times. Therefore, not surprisingly, these solvents are always tried in the same order: ACN, MeOH, and THF. This is indicated by the AeBeC vertices of the method development triangle (Figure 10.5). If one of the three isoeluotropic mixtures is successful, the problem is over. If some peaks remain unresolved, ternary or even quaternary isoeluotropic mixtures may be trialed. For this purpose, the order of the DeG mixtures in Figure 10.5 is usually followed. After selecting the optimal isoeluotropic mixture, its composition can be slightly changed until all the peaks of interest are satisfactorily resolved. Let us consider

A

D

F G

B

E

C

FIGURE 10.5 Method development triangle. A, B, and C represent isoeluotropic binary mixtures of water with ACN, MeOH, and THF, respectively; DeF are isoeluotropic ternary mixtures (e.g., point D is an ACNeMeOHewater mixture, where half of one modifier has been substituted by an isoeluotropic amount of the other modifier). The central point G is the ACNeMeOHeTHFewater isoeluotropic quaternary mixture, where two thirds of one modifier have been substituted by isoeluotropic amounts of the two other modifiers.

244

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

a 70:30 ACNewater mixture, for which all peaks for a given sample are in the target range of k values. If the resolution between some peak pairs is unsatisfactory, following the scheme in Figure 10.5 and the nomogram in Figure 10.2, the mobile phase to try next is 78:22 MeOHewater (point B in Figure 10.5). If required, we continue with 52:48 THFewater (point C), 35:39:26 ACNeMeOHewater (point D), 39:26:35 MeOHeTHFewater (point E), and so on. This trial-and-error method is more common in practice than the SST approach, owing to its simplicity and because it requires no knowledge of solute properties. However, when the problem remains unresolved, either the SST or a computer-assisted interpretive optimization (see Section 10.7.4) is of help. Similarly, selectivity optimization in NPLC and HILIC can be conveniently carried out by systematically substituting the modifier by other miscible solvents exhibiting a different location on the SST.

10.7.3. Systematic Trial-and-Error Mobile-Phase Optimization for Gradient Elution When analyzing samples with solutes covering a wide range of polarities, gradient of elution strength is needed to get both adequate retention of the first peaks in the chromatogram and progressively expedite the elution of the most retained solutes. For this purpose, at least two solvent mixtures with different elution strength (mixtures A and B, with B stronger) should be combined. The gradient starts at the time of sample injection (actually, when the gradient reaches the column). During the gradient time, tG (the time the gradient is run), the flow of B and A are increased and decreased, respectively, keeping the sum of the two flows constant, until only B is pumped. To reduce the baseline noise due to fluctuations in the mixture composition, which can be particularly large with quaternary pumps, A and B mixtures containing at least 5% of the minor solvent, should be used. Initially, for an unknown sample, a broad gradient with a small slope is used to ensure the elution of all solutes (e.g., in RPLC, from 5 to 100% ACN). The ratio of the difference between the retention times of the first and last peaks in the chromatogram and the gradient time (Dt/tG) provides a criterion for deciding whether the sample can be separated isocratically or gradient elution is required. If Dt/tG < 0.25, the sample is isocratically eluted within the k target region, using a mobile phase composition close to that of the midpoint in Dt. In contrast, Dt/tG > 0.25 means that the solutes elute in a wide k range and isocratic elution is not practical. The new gradient should now be focused between the mobile phase composition at the time of the first eluting peak (start of Dt; new mixture A) and the time for the last peak (end of Dt, new phase B).

10.7. PRACTICAL GUIDELINES FOR OPTIMIZATION

245

If some peak pairs remain unresolved, the composition of mixtures A and B should be modified without altering significantly their respective elution strengths. In RPLC, this can be achieved by substituting ACN with MeOH or THF, or by using isoeluotropic ternary or quaternary mixtures, as discussed for isocratic elution. When all solutes are satisfactorily resolved, the gradient time can be further reduced without losing resolution. The easiest way is to increase the gradient slope as much as tolerated by the resolution of the least resolved peak pair. Another option is a segmented gradient, that is, a gradient whose slope changes according to the peak distribution: The slope is smaller in time regions of poorly resolved peaks and steeper in regions without peaks. In addition to elution strength gradients, it is possible to establish selectivity gradients by increasing the mobile phase acidity, basicity, or dipolarity, at either constant or increasing elution strength. Therefore, in principle, there are four possibilities: 1. Isocratic isoselective elution where the mobile phase composition is constant. 2. Isocratic elution with a selectivity gradient, obtained with isoeluotropic mixtures where the acidity, basicity, or dipolarity is changed. This entails the continuous modification of the coordinates on the SST, for example, by following any line along the sides or inscribed in the a, b, or c small triangles in Figure 10.4 or in equivalent triangles. 3. Isoselective gradient elution where the elution strength is increased but the selectivity is not modified. Isoselective gradients are implemented by using A and B mixtures corresponding to the same point in the SST, where at least one of the solvents is different. 4. Double gradient elution where both elution strength and selectivity are modified. These are the most common gradients: When the ACN or MeOH content is increased in a mixture with water, not only the elution strength increases, but also the coordinates in the SST change. Double gradients can be programmed by progressively decreasing the water flow while simultaneously increasing the flow for one or even two modifiers at different rates. In this way, the elution strength is increased and, simultaneously, the selectivity is continuously modified in the desired direction (higher acidity, basicity, or dipolarity).

10.7.4. Computer-Assisted Interpretive Optimization Finding the best mobile phase composition or gradient to obtain good peak resolution within a short analysis time is not easy. In spite of being particularly slow and inefficient, the trial-and-error strategies explained previously (or other less systematic ones) are still frequent.

246

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

Many solute mixtures, however, are so complex that the protocol can be too long and, often, the best (or at least acceptable) conditions are not found. Fortunately, method development can be expedited with more reliable results by applying computer-assisted interpretive strategies. The optimization process includes two steps: system modeling using data from experimental chromatograms and resolution prediction through computer-simulated chromatograms. In the first step, to fit equations or train algorithms that allow the prediction of retention, a number of experiments, as reduced and informative as possible, are carried out. Incidentally, other properties that summarize a chromatogram, such as peak width and asymmetry, are also inferred from the experiments. The aim is to develop models capable of predicting the separation at any new arbitrary condition. Next, based on the models, the separation quality is scanned for a large number of separation conditions, to find the one giving the maximal (or at least an appropriate) resolution. In practice, this is done by simulating the sample separation inside a fixed factorial space and calculating a numerical value that qualifies the chromatograms, ideally according to the analyst’s appraisal of resolution. In addition to resolution, properties, such as short analysis time, minimal solvent consumption, or desirable peak profiles (i.e., high efficiencies and low asymmetries), can be optimized. To assist interpretive optimization, several software packages, such as DryLab, ChromSword, Osiris, PREOPT-W, and MICHROM, have been commercialized. The user can also develop his or her own software with the aid of a spreadsheet or a high-efficiency calculation tool, such as MATLAB. More information on computer-assisted method development can be found in Chapters 11 and 12.

10.8. ADDITIONAL CONSIDERATIONS FOR SOLVENT SELECTION In addition to the limits established by solvent viscosity and cutoff wavelength (Table 10.1), there may be other reasons to change solvents. Solvents producing high backgrounds or baseline drift with the selected detector cannot be used. In this connection, the continuous modification of the concentration of a minor component in the mobile phase might be far more significant in gradient methods than in isocratic approaches. This occurs, for instance, when an absorbing solvent is used with UV detection or when one of the components of the mixture contains a conducting buffer with conductimetric detection, and in all instances with refractometric detection. Also, lot-to-lot variability of solvents can affect UV detection, particularly when working near the cutoff wavelength.

10.8. ADDITIONAL CONSIDERATIONS FOR SOLVENT SELECTION

247

A wider range of solvents is compatible with evaporative light scattering, corona charged aerosol, mass spectrometric and ion-mobility spectrometric detectors; however, nonvolatile buffers cannot be used with these detectors. The instability or degradation of solvents (such as ethers and chlorinated alkanes) need special attention in their use and handling. Thus, THF has the drawback of its relative instability. However, using other ethers instead of THF can be problematic, due to their limited miscibility with water. Analytes can also be affected by reactivity with certain solvents. For example, higher alcohols (e.g., isopropanol) tend to be less denaturing to biomolecules than MeOH. In fact, one of the reasons that made ACN a popular choice for LC is its ability to dissolve a wide range of compounds with minimal chemical change. Care should be also taken with bacterial growth, which is a source of unexpected and unexplained chromatographic peaks, promoted by certain reagents added to aqueous mobile phases. There is a concern that many volatile organic solvents are toxic or hazardous to human health or the environment (e.g., chlorinated solvents deplete the ozone layer). Therefore, legislation restricting the use of certain solvents can affect their choice or impel finding alternatives for established methods in analytical laboratories. To reduce solvent consumption columns with a narrower internal diameter and/or smaller particle size can be used. Also, solvent recycling technologies can be a solution. All these reduced consumption patterns are supported by commitments to “greener” strategies in an effort to minimize pollution and wastes and increase sustainability. The organic solvent required in RPLC for a given separation can be reduced by using high column temperatures. Commercial equipment for control and programming of column temperature up to 200
C, with mobile phase preheating and postcolumn cooling, as well as bondedsilica columns capable of routinely supporting high temperatures, are now available. Water becomes less polar at high temperature, which increases its elution strength. From room temperature to 200
C, a 5
C increase is equivalent to approximately a 1% and 1.3% increase in ACN and MeOH, respectively. This allows the development of water-based greener, environmentally friendly RPLC methods, although at the cost of the additional energy needed to maintain the oven temperatures and preheating and cooling systems. Selectivity changes achieved by increasing the temperature are complementary with respect to those produced by modifying the mobile phase composition. These changes are mainly due to a different polarity of the solvent mixture, also depending largely on the solute molecules (derived from entropic, steric, conformational, and ionization effects). Unfortunately, the elution strength of water is still relatively low below 200
C, which in most cases hinders total

248

10. SOLVENT SELECTION IN LIQUID CHROMATOGRAPHY

replacement of organic solvents by water. Further reduction of water polarity can be achieved at temperatures over 200
C, but commercial equipment is not available and the choice of suitable stationary phases is rapidly reduced.

References [1] Hildebrand JH, Scott RL. The solubility of non-electrolytes. 3rd ed. New York: Dover Publications; 1964. [2] Snyder LR. Classification of the solvent properties of common liquids. J Chromatogr 1974;92:223e30. [3] Karger BL, Snyder LR, Eon C. An expanded solubility parameter treatment for classification and use of chromatographic solvents and adsorbents: parameters for dispersion, dipole and hydrogen bonding interactions. J Chromatogr 1976;125: 71e88. [4] Snyder LR, Glajch JL, Kirkland JJ. Theoretical basis for systematic optimisation of mobile phase selectivity in liquid-solid chromatography. J Chromatogr 1981;218: 299e326. [5] Schoenmakers PJ, Billiet HAH, de Galan L. The solubility parameter as a tool in understanding liquid chromatography. Chromatographia 1982;15:205e14. [6] Kamlet MJ, Abboud JLM, Abraham MH, Taft RW. Linear solvation energy relationships. 23. A comprehensive collection of the solvatochromic parameters, p*, a, and b, and some methods for simplifying the generalised solvatochromic equation. J Org Chem 1983;48:2877e87. [7] Nyiredy S, Meier B, Erdelmeier CAJ, Sticher O. Prisma: a geometrical design for solvent optimization in HPLC. J High Res Chromatogr. Chromatogr Comm 1985;8: 186e8. [8] Chastrette M, Rajzmann M, Chanon M, Purcell KF. Approach to a general classification of solvents using a multivariate statistical treatment of quantitative solvent parameters. J Am Chem Soc 1985;107:1e11. [9] Schoenmakers PJ. Optimisation of chromatographic selectivity: a guide to method development. Amsterdam: Elsevier; 1986. [10] Snyder LR, Quarry MA, Glajch JL. Solvent-strength selectivity in reversed-phase HPLC. Chromatographia 1987;24:33e44. [11] Rutan SC, Carr PW, Cheong WJ, Park JH, Snyder LR. Re-evaluation of the solvent triangle and comparison to solvatochromic based scales of solvent strength and selectivity. J Chromatogr A 1989;463:21e37. [12] West SD, Mowrey DH. Characterization of reversed-phase HPLC solvent selectivity for the prediction of adjusted retention indices and resolution. J Chromatogr Sci 1991;29: 497e502. [13] Abbott TP, Kleiman R. Solvent selection guide for counter-current chromatography. J Chromatogr 1991;538:109e18. [14] Li J, Zhang Y, Carr PW. Novel triangle scheme for classification of gas chromatographic phases based on solvatochromic linear solvation energy relationships. Anal Chem 1992;61:210e8. [15] Snyder LR, Carr PW, Rutan SC. Solvatochromically based solvent-selectivity triangle. J Chromatogr 1993;656:537e47. [16] Galushko SV, Kamenchuk AA, Pit GL. Calculation of retention in reversed-phase liquid chromatography. IV. ChromDream software for the selection of initial conditions and for simulating chromatographic behaviour. J Chromatogr A 1994;660:47e59.

REFERENCES

249

[17] Cela R, Leira E, Cabaleiro O, Lores M, PREOPT-W. off-line optimization of binary gradient separations in HPLC by simulationdIV. Phase 3. Comp Chem 1996;20: 315e30. [18] De Juan A, Fonrodona G, Casassas E. Solvent classification based on solvatochromic parameters: a comparison with the Snyder approach. Trends Anal Chem 1997;16: 52e62. [19] Snyder LR. Changing reversed-phase high performance liquid chromatography selectivity. Which variables should be tried first? J Chromatogr B 1997;689:105e15. ´ lvarez-Coque MC, Baeza-Baeza JJ. Global treatment of [20] Torres-Lapasio´ JR, Garcı´a-A chromatographic data with MICHROM. Anal Chim Acta 1997;348:187e96. [21] Barwick VJ. Strategies for solvent selection: a literature review. Trends Anal Chem 1997;16:293e309. [22] Heinisch S, Lesellier E, Podevin C, Rocca JL, Tchapla A. Computerized optimization of RP-HPLC separation with nonaqueous or partially aqueous mobile phases. Chromatographia 1997;44:529e37. [23] Snyder LR, Kirkland JJ, Glajch JL. Practical HPLC method development. 2nd ed. New York: Wiley; 1997. [24] Hanai T. HPLC: a practical guide. In: Smith RM, editor. RSC Chromatographic monographs. London: Royal Society of Chemistry; 1999. [25] Palamarev CE, Meyer VR, Palamareva MD. New approach to the computer-assisted selection of mobile phases for high-performance liquid chromatography on the basis of the Snyder theory. J Chromatogr A 1999;848:1e8. [26] Pawlowski TM, Poole CF. Solvation characteristics of pressurized hot water and its use in chromatography. Anal Comm 1999;36:71e5. ´ lvarez-Coque MC, Hinze WL, Quina FH, Berthod A. Effect of [27] Lo´pez-Grı´o S, Garcı´a-A a variety of organic additives on retention and efficiency in micellar liquid chromatography. Anal Chem 2000;72:4826e35. [28] Sadek PC. The HPLC solvent guide. 2nd ed. New York: Wiley; 2002. [29] Poole CK. The essence of chromatography. Amsterdam: Elsevier; 2003. p. 369. [30] Bruno TJ, Svoronos PDN. Handbook of basic tables for chemical analysis. 2nd ed. Boca Raton, FL: CRC Press; 2003. ´ lvarez-Coque MC, Torres-Lapasio´ JR, Baeza-Baeza JJ. Models and objective [31] Garcı´a-A functions for the optimisation of selectivity in reversed-phase liquid chromatography: a review. Anal Chim Acta 2006;579:125e45. ´ lvarez-Coque MC. Levels in the interpretive optimisation [32] Torres-Lapasio´ JR, Garcı´a-A of selectivity in high-performance liquid chromatography: a magical mystery tour: a review. J Chromatogr A 2006;1120:308e21. [33] Smith RM. Superheated water chromatographyda green technology for the future. J Chromatogr A 2007;1184:441e55. [34] Izutsu K. Electrochemistry in nonaqueous solutions. 2nd ed. Weinheim, Germany: Wiley-VCH; 2009. [35] Snyder LR, Kirkland JJ, Dolan JW. Introduction to modern liquid chromatography. 3rd ed. New York: Wiley; 2010. [36] Dolan JW. Selectivity in reversed-phase LC separations, part I: solvent-type selectivity. LCeGC North America 2010;28:1022e7. [37] Johnson AR, Vitha MF. Chromatographic selectivity triangles. J Chromatogr A 2011;1218:556e86.

This page intentionally left blank

C H A P T E R

11

Method Development in Liquid Chromatography J.W. Dolan, L.R. Snyder LC Resources, Inc., Walnut Creek, CA, USA O U T L I N E 11.1. Introduction

252

11.2. Goals

253

11.3. A Structured Approach to Method Development 11.3.1. Column Plate Number, N: Term i of Eq. (11.1) 11.3.2. Retention Factor, k: Term ii of Eq. (11.1) 11.3.3. Selectivity, a: Term iii of Eq. (11.1) 11.3.4. Gradient Elution

253 254 254 255 258

11.4. Method Development in Practice 11.4.1. Resolution-Modeling Software 11.4.2. Priority of Column Screening 11.4.3. HPLC vs. UHPLC 11.4.4. A Systematic Plan

258 258 259 260 261

11.5. Prevalidation

263

11.6. Validation

264

11.7. Documentation

265

11.8. Summary

266

References

267

Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00011-0

251

Copyright Ó 2013 Elsevier Inc. All rights reserved.

252

11. METHOD DEVELOPMENT IN LIQUID CHROMATOGRAPHY

There are many different ways to develop a new HPLC method, and one is not necessarily better than another as long as it attains the goals of the developer in a timely manner. Here, we share one approach based on our experience in a method development environment, years of interacting with clients in the pharmaceutical industry, and working with widely accepted scientific principles. It should be noted that method development is an involved process, so this discussion should be considered as an overview rather than a stepby-step instruction manual. For in-depth information, consult references [1e3], as well as the scientific literature and column manufacturer’s technical notes.

11.1. INTRODUCTION In recent years, the pharmaceutical industry has been applying quality by design (QbD) to various tasks in the laboratory and manufacturing environment. QbD is based on an ICH (International Committee on Harmonization) document [4], which states that, to have a high-quality product (e.g., method), quality must be designed into the product, not tested into it. Another concept of QbD is the design space, which is the multidimensional space of operational variables within which a method is valid and does not require revalidation. In practical terms, changes in conditions are allowed within the design space to meet system suitability requirements (without revalidating the method). For HPLC, the design space encompasses the range of allowed values of various conditions (%B, pH, oC, etc.); changes in any combination of these variables are allowed. This approach provides the flexibility to adjust a method to restore performance, if necessary. A requirement of QbD is that the effects of various conditions on the separation must be defined, so that the design space limits can be identified. QbD, although new in name, is not a new concept to experienced chromatographers. It does, however, provide a practical organizational structure for method development, which we apply in this chapter. The method development process comprises six consecutive steps: 1. 2. 3. 4. 5. 6.

Define the goals of the method (Section 11.2). Determine the method development approach (Section 11.3). Develop the method (Section 11.4). Perform prevalidation experiments (Section 11.5). Validate the method (Section 11.6). Document the process (Section 11.7).

We emphasize step 2 but also examine each step in more or less detail. Reversed-phase separation is assumed unless otherwise noted.

11.3. A STRUCTURED APPROACH TO METHOD DEVELOPMENT

253

A survey of the scientific literature, existing in-house methods, or other resources may provide leads on how to proceed with a particular sample. Such information can be helpful at the outset, but be careful, because information about method robustness rarely is available; also, starting method development from a poorly developed existing method seldom is a good approach. It is better to use available information for choosing starting conditions, such as the initial column, organic solvent, and (maybe) mobile phase pH.

11.2. GOALS Before method development can start, the goals (or practical application) of the method must be delineated. Related questions may include • How will the method be used (research, production, quality control, random generic samples, high throughput, etc.)? • Who will use the method (location, training, special communication problems, etc.)? • What are the chromatographic goals (resolution, run time, number of samples to analyze per batch, detection and quantification limits, linearity, range, etc.)? • Are any restrictions or limitations placed on the method (laboratory environment, isocratic only, UV detection only, etc.)? • What level of validation is required (R&D method, regulatory approval, etc.)? • Are sufficient resources available for adequate method development (time, personnel, budget, equipment, etc.)? For example, a method for the content assay of a pharmaceutical product for regulatory purposes has different requirements than a method used to support a synthetic chemist or one used for in-process monitoring. Once the goals are established, the method development process can proceed.

11.3. A STRUCTURED APPROACH TO METHOD DEVELOPMENT Adequate resolution, Rs, between adjacent peaks of interest is a primary goal of most HPLC methods. For method development, a fundamental resolution equation for isocratic separation can serve as a useful guide: Rs ¼ 0:25N 0:5 ½k1 =ð1 þ k1 Þða  1Þ i ii iii

(11.1)

254

11. METHOD DEVELOPMENT IN LIQUID CHROMATOGRAPHY

where N is the column plate number, k1 is the retention factor, k, for the first peak, and a is the separation factor (selectivity): a ¼ k2 =k1

(11.2)

where k2 is the retention factor of the second peak. We recommend using Eq. (11.1) as a guide for method development. First, start with a column that has an adequate value of N. The value N z 10,000 is recommended unless other factors suggest larger or smaller N values. Usually a C8 or C18 column is chosen at the start, because these columns often can provide a successful separation (see Section 11.4.2 for column-type screening). Next, use a gradient scouting run to determine if isocratic or gradient conditions should be used (Section 12.3 in the next chapter and the discussion of Figure 12.7). The adjustment of either the isocratic percent of organic solvent, %B, or gradient time, tG, may be sufficient to obtain the desired separation. If greater resolution is needed, explore each of the various factors that influence a. The simplest approach is to use a combination of tG (or %B) and temperature,  C (e.g., Figure 12.8), then change solvent or column type if necessary. It usually is prudent to select a pH (e.g., pH 2.5) for initial experiments and reserve changes in pH for later. After changes in a have been explored, the value of N can be revisited. If there is excess resolution, the run time (and N) can be reduced by using a shorter column and increased flow rate. Conversely, a limited increase (generally, no more than 25e40%) in Rs can be gained by using a longer or smaller-particle column. Additional choices are discussed in Section 11.4 and Chapter 2.

11.3.1. Column Plate Number, N: Term i of Eq. (11.1) For most separations, values of N fall within a range of 5,000  N  20,000, corresponding to a maximum twofold change in resolution. Larger values of N require longer run times, so changes in a often are preferable. N increases for longer columns, smaller particles, and lower flow rates d but flow rate usually has a relatively small effect on plate number and resolution. For conventional HPLC operation with a maximum pressure of 400 bar (6000 psi), a 100  4.6-mm column packed with 3-mm particles represents a good starting point (N z 10,000). For UHPLC operation and a maximum pressure of 1000 bar, shorter columns with smaller particles (< 2 mm) are used. For more details on the dependence of N and column pressure-drop on separation conditions, see Chapter 2.

11.3.2. Retention Factor, k: Term ii of Eq. (11.1) Examination of a plot of term ii of Eq. (11.1) vs. Rs leads to the conclusion that 2 < k < 10 is a favorable retention range. Retention

11.3. A STRUCTURED APPROACH TO METHOD DEVELOPMENT

255

times are not inconveniently long and Rs is not strongly affected by small changes in k. For practical purposes, however, 2 < k < 10 may not be possible, so a range of 1 < k < 20 usually is acceptable. For k < 2, Rs can be affected strongly by changes in k, and interference with nonretained materials may create problems; for k > 10, excessive run times and undesirable peak broadening can occur. When the range of k values exceeds 0.5 < k < 20, gradient elution usually is recommended (Chapter 12). The retention factor is controlled most easily by adjustment of the mobile-phase strength (% B solvent). For isocratic conditions, this can be achieved by progressively reducing %B in a sequence of 90% B, 80% B, 70% B, ., until the desired k-range is reached. An alternative approach is to use gradient scouting runs (Sections 11.3.4, 12.3). Fine-tuning k often provides additional benefits (Section 11.3.3).

11.3.3. Selectivity, a: Term iii of Eq. (11.1) Selectivity, which defines the spacing of two peaks, is influenced by different chromatographic variables. Unfortunately, without prior knowledge (experimental data, sample-structure information, etc.), it is not possible to predict the influence of a particular variable on a for a given pair of peaks. It is possible, however, to make general statements about the influence of different variables on a. One such study examined 67 chemically diverse solutes in this regard, with the results summarized in Table 11.1. The study determined the average change in a (jd log aj) for the sample set for a defined change in a variable; we refer to this as the orthogonal power, OP, for that variable. If OP  0.1, it is likely that a significant change in selectivity will occur. This, of course, does not guarantee the separation of any particular peak pair, but it is a good starting point. We can approximately rank values of OP: ½Buffer ðleast effectiveÞ < %B z tG z  C < solvent type z column type ¼ pH ðmost effectiveÞ where tG is gradient time (Chapter 12). The OP values of different variables will be examined next (in the order presented in Table 11.1). • %B, tG. According to the linear-solvent-strength model [3], %B and tG (or gradient steepness) are equivalent variables for controlling a separation. A change of 10% B (e.g., from 50% ACN to 60% ACN) changes k values by about 2.5-fold (p. 58 of [2]). Similarly, a 2.5-fold change in tG (e.g., from a 10-min to a 25-min gradient) changes retention about 2.5-fold. Either such change has OP z 0.07e0.08 (Table 11.1), slightly less than the target minimum of OP  0.1. However, these

256

11. METHOD DEVELOPMENT IN LIQUID CHROMATOGRAPHY

TABLE 11.1 Comparison of Orthogonal Power of Chromatographic Variables Variablea

Change

Example

Orthogonal powerb (OP)b

%B

10%

50% ACN to 60% ACN

0.08

tG

3x

10 min to 30 min

0.07



20  C

35 to 45  C

0.07

ACN (MeOH)

To MeOH (ACN)

Replace ACN by MeOH (or vice versa)

0.20

Column

Fs > 65c Fs > 100d

pH

5 units

pH 2.5 to pH 7.5

[0.7e

[Buffer]

2x

25 mM to 50 mM

0.02

C

0.19

%B ¼ %-organic solvent; tG ¼ gradient time;  C ¼ column temperature; [buffer] ¼ concentration of buffer. b Average jd log aj; OP  0.1 needed for “orthogonal” conditions. c Fs ¼ F value in [7]; for ionic or ionizable compounds. d For nonionized compounds. e Ionic samples only. a

variables are easy to change while maintaining k values in an acceptable range (Section 11.3.2). Furthermore, changes in %B or tG may provide sufficient changes in a to obtain adequate Rs. For these reasons, we recommend that %B or tG should be investigated early in the method development process, despite their relatively lower OP values. •  C. The value of OP ¼ 0.07 for a 20 C change in column temperature (Table 11.1) suggests that temperature is somewhat limited in its ability to increase resolution. However, for partially ionized solutes, a change in column temperature can have a dramatic effect on selectivity [5]. Furthermore, the nature of the selectivity change for  C may be different than that of %B or tG, so that a combined change of  C and %B or tG may be especially effective. The convenience of temperature changes leads us to recommend simultaneous changes in  C and either %B or tG at an early stage in method development (in this connection, see also Figure 12.8 in Chapter 12). The column temperature should be controlled in all cases (usually slightly above room temperature, e.g., 30e35  C). • Solvent type. A change in the B solvent (e.g., methanol, MeOH, vs. acetonitrile, ACN) can be effective for changing a during method development. According to Table 11.1, replacing ACN with MeOH (or vice versa) has OP ¼ 0.2, double the minimum desired OP  0.1. Any of the three popular organic solvents (ACN; MeOH; tetrahydrofuran, THF) can be blended for improved control of selectivity [5]. One approach to method development is to screen two or more solvents

11.3. A STRUCTURED APPROACH TO METHOD DEVELOPMENT

257

early in the development process to see which one separates more peaks. Then, the chosen solvent can be fine-tuned (as previously) by adjusting %B or tG; the use of mixtures of two or more B solvents can also be considered. • Column type. For years it has been known that changing from one column type to another (e.g., C18 to cyano) can result in a significant change in a; however, changing from one C18 column to another can sometimes also provide an adequate change in selectivity. Recent developments (the hydrophobic-subtraction model [6]) have led to a better understanding of column selectivity, as well as its implementation by means of free column-comparison software (USPPQRI database [7]). Using the latter software, we can identify similar (equivalent) or different (“orthogonal”) columns by means of a derived comparison function, Fs. Two columns with Fs  3 can be assumed to be equivalent. As Fs increases, the columns become more different. For maximum change in column selectivity (OP  0.19 in Table 11.1), a value of Fs > 65 is sufficient for ionizable solutes (acids or bases), while Fs > 100 is adequate for neutral or nonionized compounds. The USP-PQRI database [7] can be used to select columns during method development. If one or more additional columns is used to vary selectivity, columns with significantly different Fs values should be chosen. This approach is generally more effective than other criteria for column orthogonality. Also, to ensure additional method robustness, it is best to begin method development with columns for which equivalent columns (Fs  3) are available. When starting method development, columns should be picked that are based on the newer, high-purity (type-B) silica; such columns are more reproducible and less likely to generate tailing peaks than older, lower-purity (type-A) packings. Method development should be started with a new (virgin) column, and column-to-column reproducibility for different column manufacturing lots should be verified after a column is selected for the final method. • pH. A change in mobile-phase pH can be one of the most powerful ways to change a if the analytes are ionizable. For such samples, a 5-unit change in pH (e.g., pH 2.5 to pH 7.5) can have OP > 0.7 (Table 11.1); an operating range of 2 < pH < 8 generally is advised for silicabased columns,. At pH < 2, the bonded phase hydrolyzes and is lost; at pH > 8, the silica dissolves. For most columns, a low pH buffer of 2.5  pH  3 is a good starting place. Low pH suppresses the ionization of column silanols and acidic analytes, providing better peak shape. Many basic analytes have sufficiently high pKa values that they remain ionized at pH < 8. For work at pH > 8, several manufacturers offer silica-based columns that are stable at pH > 8.

258

11. METHOD DEVELOPMENT IN LIQUID CHROMATOGRAPHY

• Buffer concentration. For most reversed-phase separations, a change in mobile-phase buffer concentration has little effect on selectivity (OP ¼ 0.02, Table 11.1). Exceptions exist for mixed-mode or HILIC separations (Chapter 5), where ionic or electrostatic interactions play a significant role in the separation. A buffer concentration of 5e10 mM (measured in the total mobile phase) is recommended. Higher buffer concentrations (e.g., >50 mM) can result in buffer solubility problems.

11.3.4. Gradient Elution Many samples have a sufficiently wide polarity range that 1 < k < 20 is not possible for any isocratic condition. Furthermore, even when isocratic separation is possible, identifying those conditions by stepwise changes in %B can be time consuming. An initial gradient separation is instead recommended prior to method development, to determine whether isocratic separation is possibledand if so, what %B provides 1 < k < 20 for the sample. A free calculator [8] can use the results of this initial gradient to determine approximate isocratic separation conditions. If only gradient elution is feasible, the calculator also can be used to trim “wasted” time off the beginning or end of the gradient. Resolution-modeling software (Section 11.4.1) can further increase the information content of a limited number of experimental runs. We recommend starting method development with gradient runs that can be used with resolution-modeling software.

11.4. METHOD DEVELOPMENT IN PRACTICE Implementation of the method development approach of Section 11.3 involves several additional choices, as presented in this section. The method development process should represent a best compromise among the factors that affect method development for a given sample.

11.4.1. Resolution-Modeling Software A linear relationship exists between retention (log k) and mobile-phase %B: log k ¼ a þ b%B

(11.3)

a and b are constants for a given solute and separation conditions. Similar relationships exist between values of k and  C; other curve fits can be used to describe the relationship between k and other variables (pH, ionpair-reagent concentration, etc.). These relationships allow accurate prediction of retention as a function of separation conditions, based on two

11.4. METHOD DEVELOPMENT IN PRACTICE

259

or more experimental measurements for changes in each condition. This in turn allows predictions of Rs for simultaneous changes in eone to three variables, such as temperature and gradient time. It is convenient to display the results of such calculations as resolution maps, where Rs is plotted vs. one to three conditions, using resolution-modeling (“computersimulation”) software (e.g., DryLab, Molnar Institute, Berlin). Using data from the initial “calibration runs,” resolution maps allow optimum conditions for a separation to be determined quickly. So, for example, 12 experimental runs (2 tG values  2  C values  3 pH values) can give a three-dimensional model (cube) allowing prediction of Rs under any combination of these three variables, as well as any isocratic %Be CepH combination. Thus, just a few runs can answer the following questions: • • • •

Can an adequate separation be obtained using the tested variables? If so, what conditions should be used? How sensitive is the separation to each (or a combination of) variables? What conditions should be tested to demonstrate robustness in QbD (Sections 11.5e11.6)?

A further benefit of resolution-modeling software is that it requires high-quality input data for accurate predictions. This adds discipline to the method development process, so that, even if the software is not used, the quality of the experimental datadand the results of method developmentdtend to be better. We strongly recommend using resolutionmodeling software during method development for both improved productivity and higher-quality methods.

11.4.2. Priority of Column Screening All the variables listed in Table 11.1 can be varied in a continuous mannerdexcept column selectivity. Optimization of these “continuous” variables can be achieved by incrementally changing the variable or by using resolution-modeling software (Section 11.4.1). Column selection, on the other hand, requires a choice between one column and anotherdcolumns cannot be blended conveniently for intermediate results. Historically, predictability of differences in column selectivity was poor, so successfully changing a separation by changing columns was often more luck than skill. Today, column selectivity differences can be predicted (Section 11.3.3), improving the chance of changing a separation by using a different column. This leads to two general approaches: • Screen continuous variables first. This is the traditional approach, where a single column is chosen, then a systematic investigation of other variables (e.g., tG,  C, pH) is carried out. This approach is easily automated with modern HPLC equipment, and the number of

260

11. METHOD DEVELOPMENT IN LIQUID CHROMATOGRAPHY

experiments can be reduced if resolution-modeling software is used. This can be an efficient way to conduct method development. • Column screening first. An alternative approach is to screen two or more columns of different selectivity at the beginning of the method development process, to pick a column for further method development. The problem with this approach historically is that candidate columns were chosen for reasons that may not have reflected the orthogonal nature of the column; each lab had a favorite column set but often could not offer a solid rationale for the selection. With recent advances in the understanding of column selectivity [6], and the availability of a free database for selecting orthogonal columns [7], column screening now makes more sense. A simple switching-valve system can facilitate screening several columns in an unattended manner. Visual inspection or peak counting can facilitate choosing the most promising column for further method development.

11.4.3. HPLC vs. UHPLC A thorough investigation of several variables can be time consuming. Consider first a conventional HPLC system ( 100 [7]. Third column (if desired; e.g., Hipurity Advance), which is orthogonal to both the first and second columns (e.g., ACE C18 vs. Zorbax Bonus RP). Other variables as for continuous variable example. Procedure: 1. Two tG values  two  C values þ blank (five runs) for primary column and ACN. 2. Repeat for second column (and third, if using one). 3. Repeat steps 1 and 2 for MeOH. 4. Check resolution map for promising conditions and refine. 5. If no success, change to pH 7.0 and start over. Alternate procedure: 1. 15-min gradient (HPLC) and one blank at 35 C (two runs) for primary column and ACN. 2. Repeat for second column (and third, if using one). 3. Choose most promising column (number of peaks, peak shape, etc.). 4. Follow continuous-variable screening procedure with best column. If the separation is not successful after investigating all the variables just discussed, a different chromatography mode may be required.

11.5. PREVALIDATION

263

For example, normal phase (Chapter 7), HILIC (Chapter 5), mixed-mode, or another technique may be needed. More-detailed instructions can be found in general references, such as [1e3], the scientific literature, or column manufacturers’ technical notes. Before moving on to prevalidation, check the goals of the method to ensure there is no more development work to do.

11.5. PREVALIDATION At this stage, the method should be ready to validate, but it is prudent to perform some prevalidation experiments first, especially if a formal validation with submission to a regulatory agency is planned. A formal validation must be performed under an approved protocol; any changes, failures, or deviations from that plan require additional documentation and proof statements that are not required prior to formal validation. Prevalidation can minimize such problems by making a “dry run” of some or all of the validation tests; if additional adjustment is needed, it can be done under method development rules. In other words, you want to be reasonably certain that formal validation will be successful before you start that process. Prevalidation is an ideal time to establish limits for each selectivity variable that determines the boundaries of the QbD design space. Robustnessdthe ability of a method to withstand small, intentional changes in the values of different variablesdis a method characteristic that must be confirmed during validation. The design space and robustness can be estimated by using resolution-modeling software to simulate runs based on data gathered during method development. For example, experiments based on a full factorial design can be used to test the limits (high and low) for each of five variables (e.g., %B,  C, pH, flow rate, and equivalent column). This would require 25 ¼ 32 experiments. At 30 min per experiment, 16 hr would be required for this work, often with the result that one or more of the variables is not as robust as anticipated. This would require repeating the matrix of experiments with different values. With computer-simulated runs using resolution-modeling software, these 32 “runs” could be made and evaluated in vt vt vx vx2 > > > < 0  x  L; t  0; i ¼ 1; .; n; (18.1) > > > > Ci ðx; 0Þ ¼ C0;i ; > > > : Ci ð0; tÞ ¼ 4i ðtÞ: In the mass balance equation, Ci(x, t) is the mobile-phase concentration of component i at a distance x from the column inlet and at a time t after sample injection; qi is the stationary-phase concentration as described by the adsorption isotherm, F is the volumetric phase ratio, u is the linear flow rate, Da is the apparent dispersion constant, and L is the column length. F can be expressed as F ¼

Vs 1  εt ¼ V0 εt

(18.2)

where Vs and V0 are the stationary and mobile phase partial volumes and εt is the total porosity of the column. The apparent dispersion constant, Da, can be calculated from Da ¼

Lu ; 2N

(18.3)

where N is the number of theoretical plates. The last two lines in Eq. (18.1) are the initial and boundary conditions. The initial condition describes the mobile phase concentration, C0,i, at all positions x prior to the injection. Normally, C0,i ¼ 0, but as we see later, this is not always so. The boundary condition is the injection profile, 4i, that is, the shape of the injected sample zone. LC separations can be simulated by numerically estimating the solution of Eq. (18.1) at the outlet Ci(L, t), that is, estimating the elution profiles. This is usually done by the Rouchon finite-difference method or by orthogonal collocation.

18.3. ADSORPTION MODEL Functions describing the relationship between the component concentrations in the mobile and stationary phases, at a specific and constant temperature (isothermal conditions), are called adsorption isotherms. Several adsorption isotherm models are available for describing single component as well as multicomponent systems at constant temperature.

18.3. ADSORPTION MODEL

411

18.3.1. Band Shape Dependence on Adsorption In analytical LC, sample concentrations are normally very low and the corresponding adsorption isotherms are practically linear in this concentration range. This means that the adsorbed concentration is proportional to the concentration in the mobile phase. All molecules then migrate through the column, adsorb and desorb, independent of the other molecules, so each solute elutes as a Gaussian peak. The retention time of each peak depends on the initial slope of the corresponding adsorption isotherms. The peak shape deviates only slightly from the Gaussian ideal, but the chromatograms may become complex due to the multitude of solutes in the sample. In preparative LC, where concentrations are generally much higher, the adsorption-isotherm curvature and saturation capacity have an enormous impact on the peak shapes. Molecules in high-concentration zones spend relatively more time in the mobile phase due to the difficulty of finding free adsorption sites. Because of this overload, the sample zone becomes asymmetrical, elongated, and strongly dependent on the shape of the adsorption isotherm. An adsorption isotherm can be classified according to the shape of the isotherm curves, see Figure 18.1. Most reported adsorption isotherms have a convex curvature, approaching a maximum adsorbed concentration, the saturation capacity. Such models are classified as type I according to the IUPAC standard, see Figure 18.1(a), left side.

FIGURE 18.1 The most typical adsorption isotherms and the corresponding shapes of the elution profiles: (a) Type I, (b) Type II, (c) Type III adsorption behavior.

412

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

Type III adsorption isotherms are concave with an increasing slope at high concentrations, see Figure 18.1(c), left side, whereas type II isotherms are initially convex but, after an inflection point, turn concave, see Figure 18.1(b) left,side. From Figure 18.1, it is further concluded that, if a Type I adsorption isotherm describes the adsorption process best, that is, a convex upward shape, then the overloaded eluted band has a sharp front and a diffusive rear, see Figure 18.1(a), right side. The reason is that the higher eluted concentration strives for smaller retention times, because the higher the concentration, the smaller is the degree of adsorption. But, if a Type III adsorption isotherm describes the adsorption process best, that is, a concave upwards shape, then the overloaded eluted band has the opposite shape; that is, a diffusive front and a sharp rear, see Figure 18.1(c), right side. The reason is that the higher eluted concentration strives for longer retention times, because the higher the concentration, the larger is the degree of adsorption. Type II adsorption isotherms are composed of both type I and III and have vertical asymptotes that are unrealistic in LC, because this means the saturation capacity is unlimited, resulting in a most complex band shape, see Figure 18.1(b), right side. Most reported liquidesolid adsorption processes are described with Type I adsorption isotherms. In a competitive situation, the adsorption of a component between the mobile and stationary phases depends not only on the local concentration of the component itself but also on all the other components. This ultimately results in complex chromatograms. Figurer 18.2 shows the resulting preparative chromatogram after the injection of an equal mixture of two components, assuming type I adsorption behavior. The first eluted component is displaced by the second eluted one with a mixed zone in between. We can also see that the second eluted component has a hump on its rear. This situation is much more advantageous for

FIGURE 18.2 A typical binary elution profile, assuming Type I adsorption behavior.

18.3. ADSORPTION MODEL

413

fractionation of component 1 that is enriched than if injected alone; however, it is of crucial importance to understand, predict, and control the competitive forces creating this situation. By using numerical simulation, we can predict the optimal experimental conditions for collecting pure amounts of component 1 or component 2.

18.3.2. Adsorption Isotherms Adsorption isotherms describe the equilibrium distribution of solutes between the mobile and stationary phases, q(C), in a chromatography column. The nature of the interactions varies from system to system, so there are many adsorption isotherm models. Each model consists of a number of model parameters, which define the specific adsorption isotherm for the different components. If the adsorption isotherms can be measured and fitted to the appropriate model, a lot of information is obtained about system characteristics. Furthermore, it is then possible to perform computer simulations, such as by solving Eq. (18.1). The Langmuir Adsorption Isotherm To understand the fundamentals of the adsorption mechanism at overloaded preparative conditions, it can be necessary to first study the adsorption of a single component in a chromatographic system. The Langmuir adsorption isotherm is a very simple Type I model. It assumes ideal solutions, homogeneous and independent monolayer adsorption [33]:   qs;i bi Ci ai Ci P P ¼ ; i; j ¼ 1; .; n (18.4) qi C1 ; C2 ; .; Cn ¼ 1 þ j bj C j 1 þ j bj Cj In this expression, ai is the initial slope, bi is the equilibrium constant, and qs,i ¼ ai/bi is the saturation capacity of component i. It holds that ki ¼ Fai, so the retention times of the Gaussian peaks in analytical separations are given by the initial slope of the corresponding Langmuir adsorption isotherm. Figure 18.1(a) (left side) shows an example of a single-component Langmuir adsorption isotherm. Most separation problems of practical interest involve more than a single component and the Langmuir adsorption isotherm in Eq. (18.4) can be used to model the competition between the components. Here, the adsorbed concentration of any component i depends on the concentration of all components present. Because of this, all n mass-balance equations in Eq. (18.1) are coupled and cannot be treated independently. Another common adsorption isotherm model that accounts for this experimental condition is the bi-Langmuir model [34], which is an empirical extension of the Langmuir model, with two Langmuir terms

414

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

added to each other describing two different types of adsorption sites. The bi-Langmuir model applies to heterogeneous systems containing two separate types of adsorption sites. Examples are alkyl and silanol groups in C18 reversed-phase systems [35] and chiral selective and nonselective sites in chiral stationary phases [36]. More complex models of type II and III exist, which take lateral surface interactions, multilayer adsorption, adsorption energy distribution heterogeneity, are described elsewhere [2].

18.3.3. Determination of Adsorption Data There are several methods for determining the adsorption isotherm [2,37]. The most accurate technique today is frontal analysis [2,15], whereas the most recently developed method, the inverse method, is a better choice for process chromatography because of its rapidity [27e29]. Frontal Analysis Frontal analysis is usually carried out in a series of increasing concentration pulses [2,37]. The adsorption data for a single-component case is calculated by integrating the mass balance for those pulses: q ¼ C

VR  V0 Vs

(18.5)

where C and q are the solute concentrations in the mobile and stationary phase, and VR is the frontal breakthrough volume. FA can be used for any type of adsorption isotherm, is not effected by slow or concentrationdependent kinetics, and is therefore considered to be the most-accurate method for adsorption-isotherm determination. Due to these advantages, FA is more or less known as a reference method. For multicomponent cases, an intermediate plateau must be determined, which means that a fractionation and reinjection procedure must be followed for systems with more than two compounds. Unfortunately, it has been found that, for ternary mixtures, FA works only for high-efficiency separation systems [16]; otherwise, the erosion of the intermediate plateau is too pronounced. The major disadvantage with FA is that it is tedious and consumes a large amount of solvent and pure solute. The Inverse Method With the inverse method, adsorption-isotherm parameters are determined from overloaded elution profiles (peak shapes at sample overload are treated in Section 18.3.1). The solute consumption and time requirements are modest compared to other methods. The adsorption isotherm

18.4. PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY

415

cannot be obtained directly from this data (as opposed to FA data). Instead, the parameters are estimated by solving the inverse partialdifferential-equation problem: Elution profiles are simulated iteratively, by solving Eq. (18.1) numerically, and the parameters are tuned by numerical optimization until the simulated and experimental profiles coincide in the least square sense. The IM is not as accurate as the FA method. However, in process chromatography, we are mainly interested in the column model’s ability, in combination with the determined adsorption isotherm, to predict elution profiles that later are going to be used for process optimization. If the determined adsorption-isotherm parameters or model is physicochemically correct or not is not a major concern in process chromatography, this makes IM a perfect candidate for adsorption-isotherm estimation.

18.4. PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY The goal of chromatographic processes optimization, in most cases, is to produce and manufacture a high-quality product as fast and cheaply as possible, and the optimization can be performed both empirically and numerically. The empirical-process optimization approach requires extensive laboratory work to find the optimal conditions, which can be time consuming.

18.4.1. Empirical Optimization In the end, what matters to preparative chromatographers is not which model applies or the values of the model parameters. The most important thing is how much sample can be separated in a single injection and how quickly a pure component can be produced. To characterize a system, one could calculate the maximal amount of substance that is possible to inject under the condition that the component bands, or profiles, are separated. This empirical quantity is called the loading capacity. Ideally, one should use saturated solutions of the components when doing these calculations. Here, often the flow rate is fixed at as high a value as allowed and only the injection volume is varied. For example, the injection volume is increased until the maximum cross-contamination of a component exceeds 1%, that is, the component bands are no longer separated. A better optimum can be reached by allowing greater cross-contamination, performing a series of experiments with different injection volumes, then plotting the measured yield and production rate. However, this requires that the

416

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

individual component bands be measured or calculated from the total response. Notice that the component bands are allowed to overlap if the minimum yield constraint is set lower than 100%, that is, if the compound is readily available and cheap compared to the chromatographic process, highly overloaded overlapping bands [38,39] are preferable.

18.4.2. Numerical Optimization The numerical approach requires considerably less laboratory work and the ability to generate fast simulations and analyze the results qualitatively and quantitatively [2]. The objective function in numerical process optimization, usually the productivity PR , is a function of the experimental conditions and depends on the adsorption mechanisms of the system, such as thermodynamics, mass-transfer rate, and dispersion. To simplify the problem, it is common to keep some process optimization parameters fixed in the objective function, that is, perform a “suboptimization.” Then, it has to be decided which parameters to include in the optimization procedure and which ones should be kept fixed. Constraints are also usually present in the optimization problem, such as on the yield (Y) and purity (PU). Increasing the yield demand often lowers the productivity, and decisions about an acceptable limit must be made. The productivity (PR,i) is how much of component i is retrieved during one injection cycle per kg CSP, mCSP, and can be written as FV PR;i ¼

tstop;i R tstart;i

Ci ðtÞdt

tcycle mCSP

(18.6)

where FV is the volumetric flow rate, tcycle is the cycle time, tstop,i and tstart,i are the end and start of the fraction collection for the ith component. The yield, Yi, is defined as how much of the injected amount of component i is collected during one injection cycle. The purity, PU,i, is the amount of component i in the collected fraction as a percent of the total amount of all collected components. General Procedures Numerical optimization of batch-process chromatography can be represented by a workflow according to Figure 18.3. First, it is important to find a suitable experimental-separation system by thorough screening of the available stationary and mobile phases. Thereafter, the adsorption isotherms for the major components must be determined. The inverse method is one of the most convenient and rapid methods for determination of competitive adsorption isotherms aimed at process

18.4. PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY

417

Experimental system Hold up volume Calibration curve (Injection profiles)

Elution profiles

Inverse method

Column model Adsorption model Simulation algorithm

Adsorption parameters

Van Deemter curve

Process optimization

Objective function Constraints Optimization algorithm

Optimal conditions

FIGURE 18.3 Flowchart describing the most important steps in numerical optimization of process chromatography.

chromatography. The heart of the method comprises defining and solving a column mass-balance model, and the procedure involves setting an adsorption-isotherm model and parameter guesses, solving the equation and seeking to maximize the overlap of the simulated profiles to a number of elution profiles for defined overloaded injections. Therefore, a set of proper elution profiles at varying loads must be provided. The inverse solver then produces a set of adsorption isotherm parameters that best describes the system. Now it is possible to use these parameters, together with measured van Deemter functions and process conditions of the large-scale separation system, to perform process optimization with a given objective function and constraints (see Figure 18.3) [2,40]. Decision variables in the objective function are typically injection volume, injection concentration, and flow rate. Constraints are typically the maximum allowed pressure and minimum purity and yield for the target component. Numerical Injection Volume Optimization The empirical injection volume optimization just described can also be done by using computer simulations, if the adsorption-isotherm parameters and the number of theoretical plates, N in Eq. (18.3) are measured for

418

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

both components. This numerical approach requires no advanced optimization routines, for example, “gridding” can be used that is, dividing the injection volume range using a finite number of equidistant points and calculating the objective function and the constraints in each, the one with the maximum value of the objective function that also fulfills the constraints is the estimated optimal injection volume. In a pharmaceutical setting, the yield constraint is often set to 75% and the Purity constraint to 99%. Numerical Full Optimization Here, all relevant parameters are allowed to vary: injection volume, sample concentration, and flow rate. Full optimization is difficult to perform without computer simulations and requires more-advanced optimization algorithms. We often use a response-surface global-optimization algorithm for “costly” problems, such as TOMLAB [41] combined with a modified NeldereMead simplex algorithm [42] that supports inequality constraints. In the response-surface algorithm, a global optimum is sought; this is crucial, as an ordinary local optimization algorithm might get stuck in a local minimum far from the optimal solution.

18.4.3. Important Operational Conditions Several parameters are important to model correctly if a highly accurate model prediction is needed. In this section, we discuss the holdup volume, injection profiles, and the correct accounting for additives in the modeling procedure. Holdup Volume Many articles have demonstrated the need to use the correct holdup volume (porosity) in the adsorption-isotherm determination [43e46]. All these studies were performed for single-component cases. However, from a process chromatographic point of view, it is more interesting to know how such an error affects the prediction of productivity. This was recently investigated by the determination of optimum experimental conditions using erroneous adsorption isotherms combined with a wrong holdup volume and applied in the true system to evaluate the objective functions and constraints: productivity, yield, purity [47]. It was shown that, for underestimated holdup volumes, the purity requirements are fulfilled for only the second eluted component, whereas for overestimated holdup volumes, the process requirements are fulfilled for only the first eluted component. The decreased productivity is larger for overestimated holdup volumes than underestimated volumes.

18.4. PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY

419

Injection Profiles Injection profiles are used as boundary conditions for solving the column model, and often numerical process optimization is conducted using rectangular injection profiles, see Eq. (18.1). Normally, in numerical optimization of preparative chromatography, a rectangular injection profile is assumed instead of the “true” injection profile. The reason is that it is very time consuming and tedious to determine the injection profiles for all the different operational conditions used in the numerical optimization. However, by assuming rectangular injection profiles, large errors are introduced. In Figure 18.4, an experimental injection profile is plotted together with the corresponding rectangular injection profile, the difference in shape is striking. The injection profile depends on the flow rate, injection volume, viscosity of the solvent, and the solute size [48]. The eroded injection profiles are mainly due to radial diffusion in the injection loop that transports solutes back and forth from faster-moving regions to the slower-moving regions in the in the parabolic flow. Since it is so tedious to determine all possible injection profiles in an optimization procedure, another approach is to determine, from a few measured injection profiles, a function that describes the flow-rate and injection-volume dependence of the injection. In [47,49], it was shown that the injection profile 4i(t) in Eq. (18.1) can be described as a convolution of a Gaussian peak and an exponentially decaying pulse that has an initial constant part, the length

1

Norm. resp.

0.8 0.6 0.4 0.2 0 0

0.5

1

1.5

Volume [mL]

FIGURE 18.4 An experimental injection profile (black line) of solute omeprazole injected in 600 ml into an eluent of pure methanol at flow rate 1 ml/min overlaid with the corresponding rectangular injection profile (gray line).

420

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

of which is given by q. Expressed in eluted volume, V, the convolution can be written 

    A 2V0  2V þ q 2V0 þ 2V þ q p ffiffi ffi p ffiffi ffi þ erf þ erf C V ¼ 2 2s 2s   2  A 2V þ 2V0 þ q s2 s  2Vs þ 2V0 s þ sq pffiffiffi exp þ 2 þ ln erfc 2 s 2s 2ss (18.7) where A is the area of the injection profile (related to the amount injected); s, s, V0, q are parameters and erf and erfc are the error function and the complementary error function, respectively. Here, we have that V0, Vinj/s, s, and q have a linear relationship with the injection volume, Vinj, for a constant flow rate. It is also possible to include the volumetric flow rate dependency by letting these linear relationship parameters depend on the flow rate. Modeling Additives Most often, additives are used in the mobile phases of modern separation systems, this is especially the case for chiral separation systems. However, in almost all cases of numerical modeling of such preparative systems, the additives are neglected in the modeling. The reason is that the additive is invisible to the detector, especially in such ranges, so that reliable measurements of their adsorption isotherms cannot be performed. However, this problem can be bypassed by using the inverse method. One then assumes that an “invisible” additive is present and the adsorption of it can be described by an adsorption isotherm function, such as the Langmuir function. Elution profiles are then measured for different additive levels, and due to the effect on the visible eluted peaks, the adsorption isotherm parameters of the additive also can be estimated. For example, we used the inverse solver to characterize the adsorption behavior of the FMOCeallylglycine enantiomers on the quinidinecarbamate anion exchanger by estimating both the enantiomer and the additive (acetic acid) adsorption-isotherm parameters [10]. In was shown that a simulation based on adsorption-isotherm parameters estimated by the inverse method, neglecting the additive (acetic acid) in the mobile phase, fitted an experimental overloaded profile well. However, these adsorption-isotherm parameters failed to predict the experimental elution profile when using a mobile phase with a somewhat higher level of additive concentration (see Figure 3 in [10]). Then, we used the inverse solver and accounted for the additive concentration by using elution profiles from two additive levels and including the additive level in the model. With these parameters for the

421

18.4. PROCESS OPTIMIZATION OF PREPARATIVE CHROMATOGRAPHY

two enantiomers, as well as for the invisible additive, it was possible to successfully predict the elution profiles for any additive levels in between. Figure 18.5 shows the experimental and predicted binary profiles for two intermediate mobile phases and two injection volumes. It remains, however, to investigate how well the concept can be used for process optimization of a real experimental system accounting for the additive. In this context, it should be mentioned that, if the additive adsorption strength is larger than any of the injected components, very strange band shapes occur. Such profiles have been described in the literature since the 1990s [2] and the phenomenon has also been verified by computer simulations. The systems used at that time were often not of practical interested but were designed to provoke the strange effects. More recently, however, we found that the effects also take place in modern systems aimed at preparative chiral separations [50], and we can use the inverse solver approach to accurately simulate and predict cases where strong additives results in strange band shapes [51,52].

12

aII

aI

10 8 6 4

C [mg/ml]

2 0 12

bI

bII

10 8 6 4 2 0

4

6

8

10

12 4 t [min.]

6

8

10

12

FIGURE 18.5 Experimental (symbols) and predicted elution profiles using the adsorption parameters in Table 3 of [10] accounting for the additive: (a) For mobile phase 2 and (b) for mobile phase 3; here, I is a 350 ml injection and II is a 500 ml injection. The solid line corresponds to the Langmuir parameters, and the dotted line corresponds to the bi-Langmuir parameters. See Section 3 and Table 1 of [10] for experimental conditions. Source: From [10]. Copyright 2008, with permission from Elsevier.

422

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

(a)

S

C [g/L]

6

Sum S R Cut

4 2 0

R

(b) C [g/L]

4 3 2 1 0

3

4

t [min.]

5

6

FIGURE 18.6 Overlay of experimental and simulated elution profiles for the optimal conditions on 10 mm Kromasil AmyCoat 25 3 0.46 cm column: (a) S- and (b) R-omeprazole. The solid lines are experimental UV signals representing the sum of the enantiomers, symbols are analyzed fractions, and the dashed horizontal lines represent the cut points.

18.5. CASE EXAMPLE Enantiomeric separation of omeprazole has been extensively studied regarding both product analysis and preparation using several different chiral stationary phases. We recently made a full optimization of the preparative purification of R- and S-omeprazole using columns packed with amylose tris(3,5-dimethyl phenyl carbamate) on 10 mm particles. The resulting optimal chromatogram can be seen in Figure 18.6. The thick gray lines are the experimental and the thick black line is the simulated chromatogram, thin lines are simulated for R- and S-omeprazole and symbols are fractions taken during the elution. As one can see, the model predicts the process rather well with an error in the cut point of approximately 5 sec.

References [1] Helfferich F, Klein G. Multicomponent chromatography. In: Chromatographic science series, vol. 4. New York: Marcel Dekker; 1970. [2] Guiochon G, Felinger A, Shirazi DG, Katti AM. Fundamentals of preparative and nonlinear chromatography. 2nd ed. Amsterdam: Elsevier Academic Press; 2006.

REFERENCES

423

[3] FDA’S policy statement for the development of new stereoisomeric drugs. Chirality 1992;4:338e40. [4] Francotte ER. Enantioselective chromatography as a powerful alternative for the preparation of drug enantiomers. J Chromatogr A 2001;906:379e97. [5] Samuelsson J, Arnell R, Fornstedt T. Potential of adsorption isotherm measurements for closer elucidating of binding in chiral liquid chromatographic phase systems. J Sep Sci 2009;32:1491e506. [6] Marchetti N, Cavazzini A, Pasti L, Dondi F. Determination of adsorption isotherms by means of HPLC: Adsorption mechanism elucidation and separation optimization. J Sep Sci 2009;32:727e41. [7] Asnin LD, Guiochon G. The adsorption of Naproxen enantiomers on the chiral stationary phase (R,R)-Whelk-O1 under reversed-phase conditions: The effect of mobile phase composition. J Chromatogr A 2010;1217:2871e8. [8] Zabkova M, Zabka M, Rodrigues AE. Separation of Racemic Chiral Drugs Using Immobilized CHIRALPAK IA: Methodology for Preparative Scale Development. Sep Sci Technol 2009;44:275e303. [9] Mihlbachler K, De Jesus MA, Kaczmarski K, Sepaniak MJ, Seidel-Morgenstern A, Guiochon G. Adsorption behavior of the ()-Tro¨ger’s base enantiomers in the phase system of a silica-based packing coated with amylose tri(3,5-dimethyl carbamate) and 2-propanol and molecular modeling interpretation. J Chromatogr A 2006;1113:148e61. [10] Arnell R, Forsse´n P, Fornstedt T, Sardella R, Lammerhofer M, Lindner W. Adsorption behaviour of a quinidine carbamate-based chiral stationary phase: Role of the additive. J Chromatogr A 2009;1216:3480e7. [11] Seidel-Morgenstern A, Keßler LC, Kaspereit M. New developments in simulated moving bed chromatography. Chem Eng Tech 2008;31:826e37. [12] Piatkowski W, Antos D, Kaczmarski K. Modeling of preparative chromatography processes with slow intraparticle mass transport kinetics. J Chromatogr A 2003;988:219e31. [13] Tiselius A, Claeson S. Arkiv Kemi Mineral Geol. Adsorption Analysis by Interferometric Observation 1942;15B(18):1e6. [14] James DH, Phillips CSG. The chromatography of gases and vapours. Part III. The determination of adsorption isotherms. J Chem Soc 1954:1066e70. [15] Lindholm J, Forsse´n P, Fornstedt T. Validation of the accuracy of the perturbation peak method for determination of single and binary adsorption isotherm parameters in LC. Anal Chem 2004;76:4856e65. [16] Lisec O, Hugo P, Seidel-Morgenstern A. Frontal analysis method to determine competitive adsorption isotherms. J Chromatogr A 2001;908:19e34. [17] Blu¨mel C, Hugo P, Seidel-Morgenstern A. Quantification of single solute and competitive adsorption isotherms using a closed-loop perturbation method. J Chromatogr A 1999;865:51e71. [18] Forsse´n P, Lindholm J, Fornstedt T. Theoretical and experimental study of binary perturbation peaks with focus on peculiar retention behaviour and vanishing peaks in chiral liquid chromatography. J Chromatogr A 2003;991:31e5. [19] Lindholm J, Forsse´n P, Fornstedt T. Validation of the Accuracy of the Perturbation Peak Method for Determination of Multicomponent Adsorption Isotherm Parameters in LC. Anal Chem 2004;76:5472e8. [20] Glueckauf E. Theory of chromatography. Part 10.eFormulæ for diffusion into spheres and their application to chromatography. Trans Faraday Soc 1955;51:1540e51. [21] Ravald L, Fornstedt T. Theoretical study of the accuracy of the elution by characteristic points method for bi-Langmuir isotherms. J Chromatogr A 2001;908:111e30. [22] Samuelsson J, Fornstedt T. Injection Technique for Generating Accurate Adsorption Isotherm Data Using the Elution by Characteristic Points Method. Anal Chem 2008;80: 7887e93.

424

18. MODELING OF PREPARATIVE LIQUID CHROMATOGRAPHY

[23] Samuelsson J, Undin T, To¨rncrona A, Fornstedt T. Improvement in the generation of adsorption isotherm data in the elution by characteristic points method - The ECPslope approach. J Chromatogr A 2010;1217:7215e21. [24] Dose EV, Jacobson S, Guiochon G. Determination of isotherms from chromatographic peak shapes. Anal Chem 1991;63:833e9. [25] James F, Sepulveda M, Charton F, Quin˜ones I, Guiochon G. Determination of binary competitive equilibrium isotherms from the individual chromatographic band profiles. Chem Eng Sci 1999;54:1677e96. [26] Zhang L, Selker J, Qu A, Velayudhan A. Numerical estimation of multicomponent adsorption isotherms in preparative chromatography: implications of experimental error. J Chromatogr A 2001;934:13e29. [27] Felinger A, Cavazzini A, Guiochon G. Numerical determination of competitive isotherm of enantiomers. J Chromatogr A 2003;986:207e25. [28] Arnell R, Forsse´n P, Fornstedt T. Accurate and rapid estimation of adsorption isotherms in liquid chromatography using the inverse method on plateaus. J Chromatogr A 2005;1099:167e74. [29] Forsse´n P, Arnell R, Fornstedt T. An improved algorithm for solving inverse problems in liquid chromatography. Comp Chem Eng 2006;30:1381e91. [30] De Jong AWJ, Kraak JC, Poppe H, Nooitgedacht F. Isotherm linearity and sample capacity in liquid chromatography. J Chromatogr 1980;193:181e95. [31] Petkovska M, Seidel-Morgenstern A. Nonlinear frequency response of a chromatographic column. part I: Application to estimation of adsorption isotherms with inflection points. Chem Eng Comm 2005;192:1300e33. [32] Zhong G, Fornstedt T, Guiochon G. Profiles of large-size system peaks and vacancy bands in liquid chromatography I. Analytical solution of the ideal model. J Chromatogr A 1996;734:63e74. [33] Langmuir I. The constitution and fundamental properties of solids and liquids. part i. solids. J Am Chem Soc 1916;38:2221e95. [34] Graham D. The Characterization of Physical Adsorption Systems. I. The Equilibrium Function and Standard Free Energy of Adsorption. J Phys Chem 1953;57:665e9. [35] Samuelsson J, Franz A, Stanley BJ, Fornstedt T. Thermodynamic characterization of separations on alkaline-stable silica-based C18 columns: Why basic solutes may have better capacity and peak performance at higher pH. J Chromatogr A 2007;1163: 177e89. [36] Enmark M, Samuelsson J, Undin T, Fornstedt T. Characterization of an unusual adsorption behavior of racemic methyl-mandelate on a tris-(3,5-dimethylphenyl) carbamoyl cellulose chiral stationary phase. J Chromatogr A 2011;1218:6688e96. [37] Seidel-Morgenstern A. Experimental determination of single solute and competitive adsorption isotherms. J Chromatogr A 2004;1037:255e72. [38] Jacobson SC, Guiochon G. Experimental study of the production rate of pure enantiomers from racemic mixtures. J Chromatogr A 1992;590:119e26. [39] Ziomek G, Antos D, Tobiska L, Seidel-Morgenstern A. Comparison of possible arrangements of five identical columns in preparative chromatography. J Chromatogr A 2006;1116:179e85. [40] Degerman M, Westerberg K, Nilsson B. A Model-Based Approach to Determine the Design Space of Preparative Chromatography. Chem Eng Tech 2009;32:1195e202. [41] Bjo¨rkman M, Holmstro¨m K. Global Optimization of Costly Nonconvex Functions Using Radial Basis Functions. Optim Eng 2000;1:373e97. [42] Lagarias JC, Reeds JA, Wright MH, Wright PE. Convergence properties of the NelderMead simplex method in low dimensions. Siam J Optim 1998;9:112e47. [43] Sajonz P. Influence of the column hold-up time measurement accuracy on the prediction of chromatographic band profiles. J Chromatogr A 2004;1050:129e35.

REFERENCES

425

[44] Gritti F, Guiochon G. Systematic errors in the measurement of adsorption isotherms by frontal analysis: Impact of the choice of column hold-up volume, range and density of the data points. J Chromatogr A 2005;1097:98e125. [45] Samuelsson J, Sajonz P, Fornstedt T. Impact of an error in the column hold-up time for correct adsorption isotherm determination in chromatography: I. Even a small error can lead to a misunderstanding of the retention mechanism. J Chromatogr A 2008;1189:19e31. [46] Samuelsson J, Zang J, Murunga A, Fornstedt T, Sajonz P. Impact of an error in the column hold-up time for correct adsorption isotherm determination in chromatography: II. Can a wrong column porosity lead to a correct prediction of overloaded elution profiles? J Chromatogr A 2008;1194:205e12. [47] Samuelsson J, Enmark M, Forssen P. Highlighting Important Parameters Often Neglected in Numerical Optimization of Preparative Chromatography. Chem Eng Technol 2012;35:149e56. [48] Samuelsson J, Edstro¨m L, Forsse´n P, Fornstedt T. Injection profiles in liquid chromatography. I. A fundamental investigation. J Chromatogr A 2010;1217:4306e12. [49] Forsse´n P, Edstro¨m L, Samuelsson J, Fornstedt T. Injection profiles in liquid chromatography II: Predicting accurate injection-profiles for computer-assisted preparative optimizations. J Chromatogr A 2011;1218:5794e800. [50] Arnell R, Forsse´n P, Fornstedt T. Tuneable peak deformations in chiral liquid chromatography. Anal Chem 2007;79:5838e47. [51] Forsse´n P, Arnell R, Kaspereit M, Seidel-Morgenstern A, Fornstedt T. Effects of a strongly adsorbed additive on process performance in chiral preparative chromatography. J Chromatogr A 2008;1212:89e97. [52] Forsse´n P, Arnell R, Fornstedt T. A quest for the optimal additive in chiral preparative chromatography. J Chromatogr A 2009;1216:4719e27.

This page intentionally left blank

C H A P T E R

19

Process Concepts in Preparative Chromatography M. Kaspereit *, A. Seidel-Morgenstern y *

Friedrich-Alexander-Universita¨t Erlangen-Nu¨rnberg, Lehrstuhl fu¨r Thermische Verfahrenstechnik, Erlangen, Germany y Otto-von-Guericke-Universita¨t Magdeburg, Institut fu¨r Verfahrenstechnik; Max-Planck-Institut fu¨r Dynamik Komplexer Technischer Systeme, Magdeburg, Germany O U T L I N E 19.1. Introduction

428

19.2. Classical Isocratic Discontinuous Elution Chromatography 19.2.1. Mathematical Modeling and Typical Effects

429 429

19.3. Other Discontinuous Operating Concepts 19.3.1. Gradient Elution Chromatography 19.3.2. Recycling Techniques

432 433 435

19.4. Continuous Concepts of Preparative Chromatography 19.4.1. Multicolumn Countercurrent Concepts: SMB Chromatography 19.4.2. Annular Chromatography

437

19.5. Optimization and Concept Comparison

446

19.6. Conclusions

448

Acknowledgments

448

References

449

Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00019-5

427

437 445

Copyright Ó 2013 Elsevier Inc. All rights reserved.

428

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

19.1. INTRODUCTION Chromatographic separation processes using solid stationary and fluid mobile phases are frequently applied industrially with a preparative purpose to isolate and purify specific target compounds [1]. Required purity criteria can often be met only if chromatographic techniques are applied, as in the field of enantioseparations [2]. As in analytical chromatography, the basis for every successful preparative chromatographic separation is the proper choice of the chromatographic system, that is, the combination of stationary and mobile phases. However, in addition to this selection it plays an important role how the separation process is carried out, that is the operating mode. It decides the achievable productivity and yield, the specific desorbent consumption, and thus, the overall economy. Traditionally, a specific separation process is first developed and realized, exploiting the principle of isocratic batch elution. However, nowadays, a large arsenal of more flexible process options is available that is increasingly applied to improve performance criteria. More straightforward modifications are here based on keeping the most essential features of the classical process, namely, the exploitation of a single column and its discontinuous character. Gradient elution processes exploit the fact that, during the chromatographic process, a modulation of certain operating parameters (temperature, pressure, flow rate, and composition of the mobile phase) can improve the performance of conventional isocratic operation. Promising other additional degrees of freedom are exploited in concepts based on recycling insufficiently separated fractions back into the column. From a chemical engineering point of view, continuously operated separation processes are most attractive. Inspired by the success of applying other separation principles in a continuous mode (distillation, extraction, .), nowadays various concepts for continuous chromatographic separation are also available. Two of the concepts are explained later, namely, different variants of multicolumn simulated moving-bed (SMB) chromatography, which exploits several columns connected in series and mimics a continuous countercurrent movement between the mobile and stationary phases, and annular chromatography. It should be mentioned that the increased application of moresophisticated operating regimes in preparative chromatography is also driven by our improved understanding of the dynamics of concentration profiles moving in chromatographic columns under nonlinear (overloaded) conditions [3e6].

19.2. CLASSICAL ISOCRATIC DISCONTINUOUS ELUTION CHROMATOGRAPHY

429

19.2. CLASSICAL ISOCRATIC DISCONTINUOUS ELUTION CHROMATOGRAPHY The classical single-column principle of batch chromatography is wellknown from analytical chromatography, and Figure 19.1 is sufficient to explain the general principle. Traditionally, a specific separation processes is first developed and realized for preparative application by injecting, in a repetitive periodic manner, samples of the feed mixture and collecting the target products batchwise. As indicated in the bottom of Figure 19.1, subsequent feed samples are typically dosed before complete elution of the previous sample. Nevertheless, the stationary phase is often not efficiently exploited. Therefore, in preparative chromatography, there is a strong interest in more-productive and possibly continuous separation modes. There is large experience and good theoretical understanding of this classical and straightforward batch-operating mode. Therefore, it is typically considered a benchmark process, serving as a reference for evaluating the potential of alternatives. To perform such a comparative analysis, quantitative models are most suitable. A frequently applied standard model of a chromatographic column follows below, together with simulation results that illustrate typical features of nonlinear chromatography. The model presented can be seen as a building block for deriving modified models capable of quantifying all other operating modes of preparative chromatography discussed in this chapter.

19.2.1. Mathematical Modeling and Typical Effects A large number of models are available to describe the migration processes of concentration fronts in chromatographic columns [6]. Injection

Column

Detector

Fractionation

Feed

Product 1 ...

Product 2 Product n Δtcycle

Time

Oulet conc.

Inlet conc.

Eluent Δtcycle 1 2

Δtcycle 1 2

3

3 Time

FIGURE 19.1 Top: Principle of classical discontinuously operated isocratic batch

elution. Bottom:Injection policy and resulting chromatograms for three components and two consecutive injections.

430

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

A general classification is to distinguish between discrete equilibrium-stage models and continuous models leading to systems of partial differential equations (PDEs). The latter type of models allows for a more straightforward implementation of the various mass-transfer processes occurring in a chromatographic column, as demonstrated in the general-rate model (GRM) [6]. Here, we introduce briefly the frequently applied simple equilibriumdispersion model (EDM). It is based on the following assumptions: • The two phases are considered as quasi-homogeneous; that is, adsorption equilibria are permanently established throughout the column. • For relatively small particles, intraparticle concentration gradients are negligible. • In well-packed columns, radial concentration gradients can be neglected. • Isothermal conditions can be assumed, due to relatively low thermal effects. In the EDM, all effects causing band broadening are lumped into an apparent dispersion coefficient, Dapp. The corresponding mass balance for a component i in a mixture of K components is vci 1  ε vqi ðc1 ; c2 ; .; cK Þ vci v2 c i þ ¼ Diapp 2 þu vt vz ε vt vz

i ¼ 1; .; K (19.1)

Most essential are the equilibrium functions relating the loading of a component i, qi, to the fluid phase concentrations of all components, c1, c2, ., cK. In Eq. (19.1), ε stands for the total column porosity and u for the linear velocity of the mobile phase. To use this model for simulations, the typically nonlinear functions qi ¼ i 1 2 q (c , c , ., cK) must be provided. Due to its simplicity and capability of fitting many experimental results, the following multi-Langmuir equation is often applied to describe these functions based on component-specific parameters ai and bi [5,6]: qi ¼

1þ

ai ci PK

j¼1 b

jc j

i ¼ 1; .; K

(19.2)

For efficient columns, the apparent dispersion coefficient is closely related to the number of theoretical plates, N, by the following relation: Diapp ¼ Here, L is the column length.

uL 2N i

i ¼ 1; .; K

(19.3)

431

19.2. CLASSICAL ISOCRATIC DISCONTINUOUS ELUTION CHROMATOGRAPHY

To describe batch elution, typically, completely regenerated columns are considered to specify the initial conditions at t ¼ 0: ci ðz;t ¼ 0Þ ¼ 0 i ¼ ; 1; .; K

(19.4)

A rectangular injection profile is assumed as the boundary condition at the column inlet (z ¼ 0): ci ðz ¼ 0; tÞ ¼ ciinj

0  t < tinj:

ci ðz ¼ 0; tÞ ¼ 0

t  tinj:

ðInjectionÞ ðElutionÞ

i ¼ 1; .; K

i ¼ 1; .; K

(19.5) (19.6)

Other operating regimes, described later, essentially differ with respect to these conditions. Since no general analytical solutions are available for the set of PDEs as given by Eq. (19.1), one of the numerous available numerical methods must be used [6]. Summarizing the main results of a simulation study carried out solving Eq. (19.1), two figures illustrate typical features of nonlinear chromatography. In contrast to the small sample sizes processed for analysis, larger injection concentrations or volumes are relevant in a preparative scale. Under these conditions, the underlying distribution equilibria are no longer linear, and concentration-dependent migration speeds and competition effects arise. Figure 19.2(a) illustrates the significant impact of the plate number N (or the apparent dispersion coefficient Dapp) on the variance of chromatographic bands under linear conditions. Figure 19.2(b) reveals that, under nonlinear conditions, that is, for higher injection amounts, the plate numbers (or efficiencies) lose importance, and thermodynamics start to influence more strongly band shapes and retention

(a)

(b)

Linear conditions (cinj = 1)

N

0.002

Concentration

Concentration

Nonlinear conditions (cinj = 200)

0.20

0.003

0.001

0.15

N

0.10 0.05 0.00

0.000 5

5,5

6

Time

6,5

7

0

2

4

6

8

Time

FIGURE 19.2 Elution profiles as a function of the plate number N according to Eqs.

(19.1e19.6) under linear (cinj [ 1) and non-linear (cinj [ 200) conditions, respectively. Other parameters (in compatible units): ε ¼ 0.5, u ¼ 20, L ¼ 20, a ¼ 5, Injections: c(t, z ¼ 0) ¼ cinj for 0 < t < 0.001. Plate numbers: N ¼ 500, 1000, 2000.

432

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

(a)

Displacement effect (c1inj = 20,c2inj = 180)

Concentration

0.20 0.15 0.10 0.05 0.00

0

1

2

3

4

5

6

7

5

6

7

Time

(b)

Tag-along effect (c1inj = 180,c2inj = 20)

Concentration

0.20 0.15 0.10 0.05 0.00

0

1

2

3

4

Time

FIGURE 19.3 The effects of competitive adsorption. Elution profiles according to Eqs. (19.1e19.6) for injections of the same amounts as a single component (dashed lines) and in a binary mixture (solid lines): (a) The displacement effect for c1ing [ 20, c2ing [ 180; (b) the tag-along effect for c1ing [ 180, c2ing [ 20. Other parameters (in compatible units): ε ¼ 0.5, u ¼ 20, L ¼ 20, a1 ¼ 4, b1 ¼ 4, a2 ¼ 5, b2 ¼ 5, N ¼ 1000. Injections for 0 < t n s and n i i i i For these amounts, we hold n_ f ¼ V_ f;III cIII and n_ s ¼ V_ s qIII ; where V_ f;III and V_ s are the volumetric flow rates of the liquid and the solid phases, respectively. Combining Eqs. (19.8) and (19.9) provides B _ A A _ _ B B V_ f;III cA III > V s a cIII and V f;III cIII < V s a cIII

Dividing Eq. (19.10) by the concentrations leads to

ciIII

(19.10) and the solid flow rate V_ s

V_ f;III V_ f;III > aA and < aB : V_ s V_ s

(19.11)

440

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

These inequalities define the adjustable flow-rate region that leads to the separation of A and B in zone III as a function of the isotherm parameters aA and aB only. For designing the process, it is more convenient to apply dimensionless flow rates mj for each zone j: j V_ f ; mj ¼ V_ s

j ¼ I . IV

(19.12)

These mj values serve as the main design parameters. Defining corresponding tasks and inequalities to all four zones of the TMB process and requiring a nonnegative feed flow rate, mIII  mII, leads to the following set of design constraints: m I  aB

(19.13a)

aA  mII  mIII  aB

(19.13b)

mIV  aA

(19.13c)

Here, in particular, the separation zones II and III are of interest. The inequality (19.13b) defines a triangular region for (mII, mIII) in which any chosen flow rate combination leads to a complete separation, provided the other inequalities are also fulfilled. An example is shown in 3.5

m III

aB 3.0 cFi ↑

2.5

aA 2.0

1.5 1.5

2.0 aA

2.5 m II

3.0 aB

3.5

FIGURE 19.9 Design of SMB chromatography based on triangular regions of complete separation for the dimensionless flow rates in zones II and III of an ideal TMB process. Solid lines and filled symbol are the separation region and optimal operating point for linear isotherms, Eq. (19.8). Dashed lines and open symbols are the separation regions and optimal operating points for Langmuir isotherms, Eq. (19.2), for increasing feed concentrations ciF as determined according to [30]. Parameters: aA ¼ 2, bA ¼ 0.02 g/l, aB ¼ 3, bB ¼ 0.03 g/l, feed (inlet) concentrations c AF ¼ c BF ¼ 0, 1.0, 2.5, 5.0, 10.0, 20.0 g/l.

19.4. CONTINUOUS CONCEPTS OF PREPARATIVE CHROMATOGRAPHY

441

Figure 19.9. Operating inside of the triangle guarantees pure products, while outside the purity of one or both outlets is less than 100%. The indicated optimal operating point lies on the vertex of the region. Furthermore, the figure contains examples of separation regions for the case of nonlinear Langmuir adsorption isotherms, Eq. (19.2) [30]. While the performance of the process in terms of productivity and eluent consumption improves with increasing the feed concentration, the region becomes smaller, which corresponds also to an increased sensitivity to disturbances. An optimal feed concentration can be found making a compromise between robustness and performance. Finally, the volumetric flow rates in the four zones and the switching time of the corresponding SMB process are found from the m-values using the following expression [30] mj ¼

SMB V_ j DtS  Vc ε

Vc ð1  εÞ

; j ¼ ðI.IVÞ

(19.14)

where Vc denotes the volume of a single column. Extensions of this shortcut method have been established by Mazzotti, Morbidelli and coworkers for other isotherm types as well [31e35]. Furthermore, application is possible to design processes where the purity requirements are lower than 100%, as shown for linear [36] and Langmuir isotherms [37,38]. The operating parameters found by this shortcut procedure represent valuable initial estimates that can be applied as starting values for further model-based or experimental investigation and optimization of the process. Improved Operating Concepts Although the conventional SMB process just described is already a powerful technology, a number of possibilities exist to further improve its performance or broaden its scope of application. The periodic nature and its peculiar setup offer a multitude of options to intervene and improve performance. Details can be found in several reviews [39e41]. Here, only a few relevant examples can be discussed. A rough classification of the major options suggested to enhance SMB processes include: • Implementation of gradients of, for example, solvent composition, salt concentration, temperature, or pH value. • Periodic manipulation of parameters like flow rates, concentration, and column configuration. • Modification of the setup by, for example, enrichment steps, using fewer or more zones, internal recycles, or a combination with further separation or reaction steps.

442

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

Zone II

Extract (B)

Raffinate (A)

Zone III

Zone I

Feed (A + B) (weak solvent)

Zone IV

Eluent (strong solvent)

Concentration Concentration of A and B of modifier

Process variants that combine one or more of these modifications have also been suggested. The gradient concept introduced in Section 19.3.1 for single-column chromatography can be applied in SMB chromatography, too. This approach, originally suggested for separations using supercritical fluids [42], is capable of increasing productivity and product concentrations, as well as of reducing eluent consumption. In liquid chromatography, a gradient SMB process is easily accomplished by using mobile phases of different elution strength for the feed and the eluent (desorbent), respectively [43]. Theoretical and experimental studies reveal that it is particularly attractive to apply, as a desorbent, a solvent that has stronger elution strength than the solvent containing the feed [44]. As illustrated in Figure 19.10, the resulting internal solvent strength profile leads to an enhanced desorption in zones I and II, while adsorption is favored in zones III and IV. An increasingly important application of gradient SMB chromatography is the separation of proteins by ion-exchange or hydrophobicinteraction chromatography using salt gradients [45e47]. Other examples of successful application of gradient SMB are given in [48,49]. When designing a gradient SMB process, special care has to be taken in describing the dependency of the adsorption isotherm parameters on the modifier concentration, see Eq. (19.7). Furthermore, it has to be verified whether components of the solvent adsorb or not, since this is crucial for determining the internal solvent strength profiles. These information are required by the different suggested design methods, which range from extensions of the “triangle theory” introduced previously [50] to applying more detailed simulation routines [51]. Another attractive option developed to enhance SMB processes is to exploit their periodic nature directly by manipulating, in an optimal

„Weak solvent“

„Strong solvent“

A

B

I

II

III

IV

FIGURE 19.10 Gradient SMB chromatography using a two-step gradient exploiting

a (nonadsorbing) modifier. Left: Process setup, right: illustration of the solvent strength profile developing in the apparatus (top) and the resulting effect on the internal concentrations of two components, A and B (bottom; thin lines indicate profiles attained under conventional isocratic conditions).

19.4. CONTINUOUS CONCEPTS OF PREPARATIVE CHROMATOGRAPHY

443

manner, the time course of certain operating parameters during each switching interval. Most of these concepts aim at optimizing the dynamic propagation behavior of the concentration bands within the unit, which is controlled by the nonlinearity of the adsorption isotherms (see Section 19.2.1). Various parameters can be easily manipulated. Figure 19.11 shows a few selected examples. In the “Power feed” concept (Figure 19.11, left) [52,53], some or all internal flow rates are changed between subswitching intervals. This leads to time-dependent volumetric flow rates for the outlet streams. In the “Modicon” concept (Figure 19.11, middle) [54,55] the feed concentration is varied periodically. This directly affects the internal concentration-dependent migration velocities of the components. For systems with Langmuir isotherms, Eq. (19.2), it was found most beneficial, during each switching interval, to first apply a low then a high feed concentration. Figure 19.11 (right) shows the idea of an asynchronous column shifting regime, denoted as “Varicol” [56]. In contrast to conventional SMB processes, here, the columns are switched not simultaneously but independently in an optimized fashion. Apart from the “superperiodic” processes just discussed, also rather simple changes of the process setup can create significant benefits. For example, Figure 19.12 (left) shows a simple three-zone, open-loop setup. Here, zone IV of the process is abandoned. This is of interest when

FIGURE 19.11 Selected improved concepts for SMB chromatography based on a “superperiodic” variation of operating parameters. Top: Schematic representations; F, R, X, and E denote feed, raffinate, extract, and eluent, respectively; main manipulated parameters are marked by bold lines. Bottom: Temporal profiles of the manipulated parameter(s) over two consecutive switching intervals. Left: “Power feed” operation with variation of flow rates (only flow profile for zones II and III are shown). Middle: “ModiCon” process with modulation of feed concentration. Right: “Varicol” process, based on asynchronous column shifting.

444

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY Buffer tank F Zone II

R

F

R

R

Zone III

Zone I

F

X

E

X E

Partial solvent removal

E

X

E

FIGURE 19.12 Selected improved concepts for SMB chromatography based on modification of the process setup. Left: Simple three-zone, open-loop setup for applications where a regeneration of the fluid phase is abandoned due to specific requirements (see text). Middle: Extract-enrichment (EE) SMB with an increased recirculated extract concentration to enhance separation efficiency. Right: Fraction- feedback (FF) SMB process that improves a partially successful separation by an internal recycle.

a regeneration of the fluid phase is less important than preventing a possible breakthrough of weakly retained components; for example, in bioseparations or when using gradients. Also, additional regeneration zones or Cleaning in Place (CIP) procedures can easily be added. Figure 19.12 (middle) illustrates a partial enrichment of the extract (“EESMB”) before its partial reintroduction into zone II [57,58]. The higher concentration established in this way causes a beneficial internal displacement effect, see Figure 19.3(a). Figure 19.12 (right) shows the “Fraction-Feedback” concept (“FFeSMB”) [59]. Here, the insufficiently separated proportions of the raffinate stream are collected when they elute and then recycled. This process has, thus, also a “superperiodic” character. Finally, a number of promising approaches extend the application of SMB units beyond binary separations with two product streams; that is, they perform multicomponent separations by SMB chromatography, see for example [60e65]. Figure 19.13 shows three examples for the VIII IV III

IV A B C

I

II C

VII

VIII

A

VI B C

B Purge

IV III II I

VII

Purge A B

V

II

III

I

A

A B C

B

C

VI V

A

B

IV A B C

III II I

C

FIGURE 19.13 Selected options for performing ternary separation by SMB chromatography (according to [40]).

19.4. CONTINUOUS CONCEPTS OF PREPARATIVE CHROMATOGRAPHY

445

separation of ternary mixtures. The left of the figure shows a simple serial connection of two SMB processes. In the middle and right, additional zones are included that allow for an internal recycling of mixture streams, which corresponds to implementing a “two-in-one” SMB process. Regarding the latter concepts, one has to account for dilution effects, which increase the volume of such recycle streams. This can necessitate purge streams, as shown in the figure, or an additional solvent removal [62]. In addition to the improved SMB concepts discussed previously, a number of further concepts was proposed, like “partial feed” [66] and “selective withdrawal” [67]. Also, the combination of the technology with further alternative separation processes, like crystallization [68,69], can be beneficial. All mentioned advanced SMB process concepts were demonstrated to be capable of achieving a significantly improved performance in comparison to conventional SMB chromatography. A comparison of selected concepts is given, for example, in [70]. However, despite their superior performance, the optimal design of such processes remains a challenging task. This can be solved using proper mathematical column models, such as the one introduced in Section 19.2, and by formulating an optimal control problem [52] or applying specialized optimization techniques [70,71].

19.4.2. Annular Chromatography The process of annular chromatography (Figure 19.14) was anticipated by Martin [72]. It uses a stationary phase fixed in the annular space Eluent Feed

Rotation

Products

FIGURE 19.14

Principle of annular chromatography.

446

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

between two concentric cylinders. Through the formed fixed bed, the mobile phase is transported from top to bottom. In a certain angular region, the mixture to be separated is continuously dosed on top. The fixed bed rotates slowly. Thus, a cross-flow between mobile and stationary phases is established. The components of the feed mixture separate from each other in helical bands on their way downward (see for example [73,74]). Extending Eq. (19.1), it is possible to derive the following mass balance describing AC [73,75]: vci 1  ε vqi ð! cÞ vci vci 1  ε vqi ð! cÞ þu þ þu þu ε vt ε v4 vt vz v4 (19.15) 2 i v2 c i i i v c ¼ Dapp 2 þ Dtan 2 i ¼ 1; .; K vz v4 Here, the angle 4 appears as an additional spatial coordinate. Equation (19.15) contains the rotation speed u and the tangential dispersion coefficient Dtan as further parameters. With the assumptions that the AC process operates at steady state (v/vt ¼ 0) and the tangential dispersion is negligible (Dtan ¼ 0), a transformation, t* ¼ 4/u, simplifies Eq. (19.15): vci 1  ε vqi ð! cÞ vci v2 ci i ¼ D þ u  þ app ε vt v vz vz2

i ¼ 1; .; K

(19.16)

Obviously Eq. (19.16) corresponds to Eq. (19.1). Thus, the temporal concentration profile at a certain position of a conventional column corresponds to a profile over the angular coordinate in steady-state annular chromatography. Appreciating the advantage of the continuous process character, this prohibits AC from providing improvements compared to conventional batch elution with respect to typically evaluated performance parameters, such as productivity and solvent consumption [73,75].

19.5. OPTIMIZATION AND CONCEPT COMPARISON As already mentioned above, model equations like Eq. (19.1) can be used to evaluate the performance of various configurations of arranging and operating chromatographic columns. In addition to classical elution, also recycling chromatography, various variants of the SMB process and AC can be described quantitatively. For example, to describe discontinuous closed-loop recycling just the following condition at the column inlet has to be respected, while the recycle is active: ci ðz ¼ 0; tÞ ¼ ci ðz ¼ L; tÞ

i ¼ 1; .; K

(19.17)

19.5. OPTIMIZATION AND CONCEPT COMPARISON

447

Since in the continuous SMB process several columns j are connected in series, for all M columns (with the exception of the positions where feed or solvent are introduced) must hold the following boundary conditions: cij ðz ¼ 0; tÞ ¼ cij1 ðz ¼ L; tÞ

i ¼ 1; .; K; j ¼ 2; M

(19.18)

Based on the transformation introduced above, the corresponding boundary conditions for annular chromatography are similar to the ones for the conventional batch column elution process, see Eqs. (19.5) and (19.6): 0  t < tinj ¼ 4F =u : ci ðz ¼ 0; t Þ ¼ ciF

ðInjectionÞ i ¼ 1; .; K (19.19)

t  tinj ¼ 4F =u :

ci ðz ¼ 0; t Þ ¼ 0

ðElutionÞ i ¼ 1; .; K (19.20)

Hereby 4F stands for the size of the injection segment (angular coordinate) and cF is the feed concentration. More details on formulating chromatographic process models and the corresponding boundary conditions can be found, for example, in [6,76]. To evaluate the performance of chromatographic separation processes, frequently the specific production rate PR and the recovery yield REC are applied as objective function: PR ¼

REC ¼

Amount of target component collected Time  Amount of solid phase

Amount of target component collected Amount of the target component in the feed

Further, it might be important to consider a specific desorbent consumption DC: DC ¼

Amount of desorbent introduced Amount of product collected

The specific definitions of these performance criteria depend on the actual process concept. For details, the reader is referred to, for example, [77e79]. It can be also useful to consider combined objective functions, for example, the product PR  REC, or a goal function that considers different objectives in combination with weighting factors. Another, computationally more extensive approach available is to apply multiobjective optimization [80]. Various reliable optimization techniques are available to find the process-specific free operating parameters. Most frequently, nonlinear

448

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

programming methods are applied using, for example, sequentialquadratic-programming (SQP) methods e.g. [37]. Increasingly used are also evolutionary algorithms [81]. An important and often neglected aspect is that when evaluating different process options, only optimized scenarios should be compared with each other. Finally, a remark on the applicability of advanced process concepts should be given. An important rule is that a more-sophisticated concept that exploits additional degrees of freedom should be applied only if it provides the chance for a substantial improvement. Otherwise, the simpler process concepts should be preferred. In addition to identifying optimal operating parameters, it is important to check the processes with respect to their robustness. Again, simple reliable processes should be favored. However, the knowledge and experience acquired in the last decades has clearly shown the potential of preparative chromatographic processes that go beyond the possibilities of standard elution. This holds true in particular for the SMB technology.

19.6. CONCLUSIONS In addition to standard isocratic elution, various alternative process concepts in preparative chromatography are available. The principles of the most important options are explained, including gradient elution, recycling chromatography, simulated moving-bed chromatography, and annular chromatography. In particular, the various variants of SMB chromatography offer the potential to develop efficient separation processes. It is pointed out that all the concepts mentioned can be quantitatively described, exploiting basic models developed for the conventional standard process by adjusting the corresponding initial and boundary conditions and installing correct connections in multicolumn processes. This allows selecting and optimizing the most suitable process concept for a specific separation problem. However, a necessary assumption for the application of this approach is the availability of validated parameters characterizing the chromatographic system under consideration.

Acknowledgments The essential contributions of numerous PhD students and our lab crew in Magdeburg are gratefully acknowledged. We are further grateful for the financial support of the European Union (Collaborative research project IntEnant, FP7-NMP2-SL2008-214129), the Deutsche Forschungsgemeinschaft (SFB 578), and Wissenschaftliche Gera¨tebau Herbert Knauer GmbH, Berlin.

REFERENCES

449

References [1] Ganetsos G, Barker PE. Preparative and production scale chromatography. New York: Marcel Dekker; 1993. [2] Cox GB, editor. Preparative enantioselective chromatography. Oxford: Blackwell Publishing; 2005. [3] Rhee H-K, Aris R, Amundson NR. First-order partial differential equations, vols. 1 and 2. Englewood Cliffs, NJ: Prentice-Hall; 1986 and 1989. [4] Guiochon G. Preparative liquid chromatography. J Chromatogr A 2002;965:129e61. [5] Schmidt-Traub H, Schulte M, Seidel-Morgenstern A, editors. Preparative chromatography of fine chemicals and pharmaceutical agents. 2nd ed. Weinheim, Germany: Wiley-VCH; 2012. [6] Guiochon G, Felinger A, Shirazi DG, Katti AM. Fundamentals of preparative and nonlinear chromatography. 2nd ed. Amsterdam: Elsevier; 2006. [7] Colin H, Hilaireau P, Martin M. Flip-flop elution concept in preparative liquid chromatography. J Chromatogr A 1991;557:137e53. [8] Frenz J, Horvath C. Displacement chromatography. In: Horvath C, editor. Highperformance liquid chromatography-advances and perspectives, vol. 5. New York: Academic Press; 1988. [9] Jandera P, Churacek J. Gradient elution in column liquid chromatography. Amsterdam: Elsevier; 1985. [10] Antia FG, Horvath C. Gradient elution in non-linear preparative liquid chromatography. J Chromatogr 1989;484:1e27. [11] El Fallah Z, Guiochon G. Prediction of protein band profile in preparative reverse phase gradient elution. Biotech Prog 1992;39:877e855. [12] Gallant SR, Vunnum S, Cramer SM. Optimization of preparative ion-exchange chromatography of proteins: linear gradient separations. J Chromatogr A 1996;725:295e314. [13] Felinger A, Guiochon G. Optimizing experimental conditions in overloaded gradient elution chromatography. Biotech Prog 1996;12:638e44. [14] Wang A, Carr PW. Comparative study of the linear solvation energy relationship, linear solvent strength theory, and typical-conditions model for retention prediction in reversed-phase liquid chromatography. J Chromatogr A 2002;965:3e23. [15] Snyder LR, Dolan JW. High performance gradient elution. New York: John Wiley & Sons; 2007. [16] Seidel-Morgenstern A. Preparative gradient chromatography. Chem Eng Tech 2005;28: 1265e73. [17] Antos D, Seidel-Morgenstern A. Continuous step gradient elution for preparative separations. Sep Sci Tech 2002;37:1469e87. [18] Bombaugh KJ, Dark WA, Levangie RF. High resolution steric chromatography. J Chromatogr Sci 1969;7:42e7. [19] Seidel-Morgenstern A, Guiochon G. Theoretical study of recycling in preparative chromatography. AIChE J 1993;39:809e18. [20] Heuer C, Seidel-Morgenstern A, Hugo P. Experimental investigation and modelling of closed-loop recycling in preparative chromatography. Chem Eng Sci 1995;50:1115e27. [21] Bailly M, Tondeur D. Recycle optimization in non-linear productive chromatographydI. Mixing recycle with fresh feed. Chem Eng Sci 1982;37:1199e212. [22] Grill CM. Closed-loop recycling with periodic intra-profile injection: a new binary preparative chromatographic technique. J Chromatogr A 1988;796:101e13. [23] Schlinge D, Scherpian P, Schembecker G. Comparison of process concepts for preparative chromatography. Chem Eng Sci 2010;65:5373e81. [24] Charton F, Bailly M, Guiochon G. Recycling in preparative liquid chromatography. J Chromatogr A 1994;687:13e31.

450

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

[25] Sainio T, Kaspereit M. Analysis of steady state recycling chromatography using equilibrium theory. Sep Purif Tech 2009;66:9e18. [26] Kaspereit M, Sainio T. Simplified design of steady-state recycling chromatography under ideal and nonideal conditions. Chem Eng Sci 2011;66:5428e38. [27] Broughton CB, Gerhold CG. U.S. Patent 2 985 589 1961. [28] Subramanian G. Chiral separation techniquesda practical approach. Weinheim, Germany: Wiley-VCH; 2000. [29] Mihlbachler K. Simulated moving bed chromatographyda promising alternative for the purification of biopharmaceuticals. In: Langer ES, editor. Advances in large-scale biopharmaceutical manufacturing and scale-up production. Washington, DC: ASM Press; 2007. [30] Mazzotti M, Storti G, Morbidelli M. Optimal operation of simulated moving bed units for nonlinear chromatographic separations. J Chromatogr A 1997;769:3e24. [31] Mazzotti M, Storti G, Morbidelli M. Robust design of countercurrent adsorption separation processes: 2. Multicomponent systems, AIChE J 40 1994:1825e42. [32] Mazzotti M, Storti G, Morbidelli M. Robust design of countercurrent adsorption separation: 3. Nonstoichiometric systems. AIChE J 1996;42:2784e96. [33] Mazzotti M, Storti G, Morbidelli M. Robust design of countercurrent adsorption separation processes: 4. Desorbent in the feed. AIChE J 1997;43:64e72. [34] Migliorini C, Mazzotti M, Morbidelli M. Robust design of countercurrent adsorption separation processes: 5. Nonconstant selectivity. AIChE J 2000;46:1384e99. [35] Gentilini A, Migliorini C, Mazzotti M, Morbidelli M. Optimal operation of simulated moving-bed units for non-linear chromatographic separations: II. Bi-Langmuir isotherm. J Chromatogr A 1998;805:37e44. [36] Rajendran A. Equilibrium theory-based design of simulated moving bed processes under reduced purity requirements: Linear isotherms. J Chromatogr A 2008;1185: 216e22. [37] Kaspereit M, Seidel-Morgenstern A, Kienle A. Design of simulated moving bed chromatography under reduced purity requirements. J Chromatogr A 2007;1162:2e13. [38] Fu¨tterer M. Design of simulated moving bed plants for reduced purities. Chem Eng Tech 2010;33:21e34. [39] Gomes PS, Minceva M, Rodrigues AE. Simulated moving bed technology: old and new. Adsorption 2006;12:375e92. [40] Seidel-Morgenstern A, Keßler LC, Kaspereit M. New developments in simulated moving bed chromatography. Chem Eng Tech 2008;31:826e37. [41] Kaspereit M. Advanced operating concepts for simulated moving bed processes. In: Grushka E, Grinberg N, editors. Advances in chromatography. Boca Raton, FL: CRC Press; 2009. p. 165e92. Taylor & Francis. [42] Clavier J-Y, Nicoud R-M, Perrut M. A new efficient fractionation process: the simulated moving bed with supercritical eluent. In: Rudolf van Rohr P, Trepp C, editors. High pressure chemical engineering. London: Elsevier; 1996. [43] Jensen TB, Reijns TGP, Billiet HAH, van der Wielen LAM. Novel simulated movingbed method for reduced solvent consumption. J Chromatogr A 2000;873:149e62. [44] Antos D, Seidel-Morgenstern A. Application of gradients in simulated moving bed processes. Chem Eng Sci 2001;56:6667e82. [45] Keßler LC, Gueorguieva L, Rinas U, Seidel-Morgenstern A. Step gradients in three-zone simulated moving bed chromatography: application to the purification of antibodies and bone morphogenetic protein-2. J Chromatogr A 2007;1176: 69e78. [46] Houwing J, Van Hateren SH, Billiet HAH, van der Wielen LAM. Effect of salt gradients on the separation of dilute mixtures of proteins by ion-exchange in simulated moving beds. J Chromatogr A 2002;952:85e98.

REFERENCES

451

[47] Wekenborg K, Susanto A, Schmidt-Traub H. Modelling and validated simulation of solvent-gradient simulated moving bed (SG-SMB) processes for protein separation. Comp Aided Chem Eng 2005;20:313e8. [48] Abel S, Mazzotti M, Morbidelli M. Solvent gradient operation of simulated moving beds. I. Linear isotherms. J Chromatogr A 2002;944:23e39. [49] Houwing J, Billiet HAH, van der Wielen LAM. Optimization of azeotropic protein separations in gradient and isocratic ion-exchange simulated moving bed chromatography. J Chromatogr A 2002;944:189e201. [50] Abel S, Mazzotti M, Morbidelli M. Solvent gradient operation of simulated moving beds. II. Langmuir isotherms. J Chromatogr A 2004;1026:47e55. [51] Beltscheva D, Hugo P, Seidel-Morgenstern A. Linear two-step gradient counter-current chromatography: analysis based on a recursive solution of an equilibrium stage model. J Chromatogr A 2003;989:31e45. [52] Kloppenburg E, Gilles ED. A new concept for operating simulated moving-bed processes. Chem Eng Tech 1999;22:813e7. [53] Zhang Z, Mazzotti M, Morbidelli M. PowerFeed operation of simulated moving bed units: changing flow-rates during the switching interval. J Chromatogr. A 2003; 1006:87e99. [54] Schramm H, Kaspereit M, Kienle A, Seidel-Morgenstern A. Simulated moving bed process with cyclic modulation of the feed concentration. J Chromatogr A 2003; 1006:77e86. [55] Schramm H, Kienle A, Kaspereit M, et al. Improved operation of simulated moving bed processes through cyclic modulation of feed flow and feed concentration. Chem Eng Sci 2003;58:5217e27. [56] Ludemann-Hombourger O, Nicoud R-M, Bailly M. The “VariCol” process: a new multicolumn continuous chromatographic process. Sep Sci Tech 2000;35: 1829e62. [57] Bailly M, Nicoud RM, Adam P. Ludemann-Hombourger O. US Patent 2006124549, 2004. [58] Paredes G, Rhee H-K, Mazzotti M. Design of simulated-moving-bed chromatography with enriched extract operation (EE-SMB): Langmuir isotherms. Ind Eng Chem Res 2006;45:6289e301. [59] Keßler LC, Seidel-Morgenstern A. Improving performance of simulated moving bed chromatography by fractionation and feed-back of outlet streams. J Chromatogr A 2008;1207:55e71. [60] Wooley R, Ma Z, Wang N- HL. A nine-zone simulating moving bed for the recovery of glucose and xylose from biomass hydrolyzate. Ind Eng Chem Res 1998;37:3699e709. [61] Hashimoto K, Shirai Y, Adachi S. A simulated moving-bed adsorber for the separation of tricomponents. J Chem Eng Japan 1993;26:52e6. [62] Keßler LC, Seidel-Morgenstern A. Theoretical study of multicomponent continuous countercurrent chromatography based on connected four-zone units. J Chromatogr A 2006;1126:323e37. [63] Chin CY, Wang N- HL. Simulated moving bed equipment designs. Sep Purif Rev 2004;33:77e155. [64] Hur JS, Wankat PC. New design of simulated moving bed (SMB) for ternary separations. Ind Eng Chem Res 2005;44:1906e13. [65] Kurup AS, Hidajat K, Ray AK. Comparative study of modified simulated moving bed systems at optimal conditions for the separation of ternary mixtures of xylene isomers. Ind Eng Chem Res 2006;45:6251e65. [66] Zang Y, Wankat PC. SMB Operation strategydpartial feed. Ind Eng Chem Res 2002;41:2504e11. [67] Zang Y, Wankat PC. Three-zone simulated moving bed with partial feed and selective withdrawal. Ind Eng Chem Res 2002;41:5283e9.

452

19. PROCESS CONCEPTS IN PREPARATIVE CHROMATOGRAPHY

[68] Kaspereit M, Gedicke K, Zahn V, Mahoney AW, Seidel-Morgenstern A. Shortcut method for evaluation and design of a hybrid process for enantioseparations. J Chromatogr A 2005;1092:43e54. [69] Amanullah M, Mazzotti M. Optimization of a hybrid chromatography-crystallization process for the separation of Tro¨ger’s base enantiomers. J Chromatogr A 2006; 1107:36e45. [70] Zhang Z, Mazzotti M, Morbidelli M. Continuous chromatographic processes with a small number of columns: comparison of simulated moving bed with Varicol, PowerFeed, and ModiCon. Korean J Chem Eng 2001;21:454e64. [71] Kawajiri Y, Biegler LT. Large scale nonlinear optimization for asymmetric operation and design of simulated moving beds. J Chromatogr A 2006;1133:226e40. [72] Martin AJP. Summarizing paper. Discuss Faraday Soc 1949;7:332e6. [73] Wankat PC. The relationship between one-dimensional and two-dimensional separation processes. AIChE J 1977;23:859e67. [74] Begovich JM, Byers CH, Sisson WG. A high-capacity pressurized continuous chromatograph. Sep Sci Tech 1983;18:1167e91. [75] Thiele A, Falk T, Tobiska L, Seidel-Morgenstern A. Prediction of elution profiles in annular chromatography. Comp Chem Eng 2001;25:1089e101. [76] Guiochon G, Lin B. Modeling for preparative chromatography. San Diego, CA: Academic Press; 2003. [77] Felinger A, Guiochon G. Comparing the optimum performance of the different modes of preparative liquid chromatography. J Chromatogr A 1998;796:59e74. [78] Heuer C, Kniep H, Falk T, Seidel-Morgenstern A. Comparison of various process engineering concepts of preparative chromatography. Chem Eng Tech 1998;21:469e77. [79] Seidel-Morgenstern A. Optimization and comparison of different modes of preparative chromatography. Analusis Mag 1998;26:M46e55. [80] Zhang Z, Hidajat K, Ray AK, Morbidelli M. Multiobjective optimization of SMB and VariCol process for chiral separation. AIChE J 2002;48:2800e16. [81] Zhang Z, Mazzotti M, Morbidelli M. Multiobjective optimization of simulated moving bed and VariCol processes using a genetic algorithm. J Chromatogr A 2003;989:95e108.

C H A P T E R

20

Miniaturization and Microfluidics F. Foret *, P. Smejkal *, y, M. Macka y *

Institute of Analytical Chemistry ASCR, Brno, Czech Republic y University of Tasmania, Hobart, Australia O U T L I N E

20.1. Introduction, Definitions, and Scope

453

20.2. Microfluidic Systems for Separations 20.2.1. Microfabrication Technologies 20.2.2. Miniaturization of HPLC Systems

455 455 458

20.3. Commercial Instrumentation 20.3.1. Electrophoretic Systems 20.3.2. HPLC Systems

459 461 462

20.4. Conclusion

465

Acknowledgment

465

References

465

20.1. INTRODUCTION, DEFINITIONS, AND SCOPE Until recently, the users of HPLC systems relied on instrumentation, which has gradually evolved since the 1960s. This trend has been, to some extent, disrupted by the introduction of microfabrication techniques, leading to the most dramatic changes in the practice of HPLC in the past decades. Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00020-1

453

Copyright Ó 2013 Elsevier Inc. All rights reserved.

454

20. MINIATURIZATION AND MICROFLUIDICS

Since the breakthrough in microelectronics in the 1960s, miniaturization has begun establishing itself as a fresh exciting trend with an impact on all fields of science and technology and opening countless new possibilities. In chemical analysis, it enables the quest for faster, higherthroughput analysis with portable or field-deployable instrumentation, enabling on-site analysis (in biomedical analysis often called point of care) and remote analysis. The inception and development of microfluidics from the early 1990s provided another significant boost to miniaturization in all fields of science. Specifically, in chemical analysis, it led to the concepts of lab-on-a-chip and micro-total analytical systems (m-TAS)[1]. Miniaturization in analytical chemistry progressed significantly when several large-scale projects gave impetus to the development of microfluidic systems. In addition to the project funded by the Defense Advanced Research Project Agency (DARPA), aiming at equipping the U.S. military with small, portable, and user-friendly analyzers for monitoring the environment and personnel health, it was especially Important for the Human Genome Project (HGP) funded by the U.S. Department of Energy and National Institutes of Health. While the DNA sequencers used to finish this project were still based on capillary electrophoresis, the microfluidics-based systems have also been developed, and new generations of instruments generally termed the NextGen sequencers are now in routine use. The financial support for these projects initiated astounding research activity in the field of microfluidic instrumentation. Electrophoresis, electrochromatography, and liquid chromatography are the most common separation methods applying miniaturization. While electrophoresis is a method naturally predetermined for miniaturization, in case of liquid chromatography, it has been a much more challenging task. When looking into its history, electrophoresis was originally realized in the format of a planar layer of gel (slab gel). Later electrophoresis methods were introduced in fused silica capillaries as capillary electrophoresis (CE). Nowadays, research and commercial applications indicate that the future of electrophoresis may be in the microfluidic-chip format (chip-CE). Microfluidic chips enable fluid multiplexing and therefore allow easier combination of several methods on one chip, which can be designed for multiple analyses as well. The definitions of microfluidic or nanofluidic are usually based on requirements that at least one dimension of the microfluidic system in question has a longitudinal dimension of 100 mm or less and in a nanofluidic system it has a longitudinal dimension of 100 nm or less. In this context, it is important to stress that commercially available LC instruments (either capillary or recently also increasingly chip based) are often called nanosystems. Here the prefix nano does not mean that microfluidics systems would be sized in nanometers but rather that the volume of samples used for analysis is on the order of nanoliters. The microfluidic device that

20.2. MICROFLUIDIC SYSTEMS FOR SEPARATIONS

455

contains the microfabricated microfluidic channels or similar microfluidic features, usually on a flat material, such as a silicon wafer, of a small size compatible with portable instrumentation, is called microfluidic chip or often just chip, following the original inspiration by microelectronics chips. Fundamentals and applications of microfluidics have been compiled in a recent monograph [2]. Selected references from 2008 to 2010 are summarized in Table 20.1. The following sections contain selected developments, techniques, applications, and trends in microfluidics with a focus on separation methods.

20.2. MICROFLUIDIC SYSTEMS FOR SEPARATIONS The term m-TAS appeared for the first time more than 20 years ago in 1990 [24]. The first experiments using microchips implementing CE on a chip (chip CE) were relatively simple with geometrically regular channels [25]. The results were convincingly successful and channels with semi-circular or rectangular cross sections showed similar results as obtained in classical cylindrical silica capillaries. Later chips progressed rapidly to a higher degree of complexity, so that they could combine several steps of analysis. The main aims and benefits of miniaturization in analytical chemistry and separation methods are • Smaller amounts of samples can be analyzed, with a lower mass detection limit. • Shorter separation time and overall faster analysis. • Lower consumption of chemicals, resulting in a lower cost per sample and a more environmentally friendly analysis. • Minimizing the contact between the sample and the environment, with ideally no delay between sampling and sample analysis. • On-site (point-of-care, POC) analysis with portableefield-deployable analyzers. • Minimization of dead volumes and leakages in the channel junctions. • Better compatibility with in-line detection techniques, such as MS.

20.2.1. Microfabrication Technologies Methods for the microfabrication of microfluidic chips were originally adopted from electronics. For this reason, silicon wafers were also among the first used material for miniaturized analyzers, such as the gas chromatograph on a chip developed in the early 1980s [26]. Later glass was the most commonly used material, especially thanks to its optical properties. Channels in glass are usually fabricated by photolithography and wet etching. Typically, a protective metal film (e.g., 50-nm Cr) is first deposited

456

Field

Main focus

Year

Pages

Figs./ Tables

Citations

Reference

Fabrication of microfluidic chips

Materials and techniques used for fabrication of microfluidic chips

2008

23

20/4

306

[3]

Multichannel chips and their application

2008

10

4/1

62

[4]

Fabrication of chips from thermoplastic materials

2009

16

9/3

111

[5]

Free-flow electrophoresis on a chip

2009

12

11/0

57

[6]

Microfluidic systems applying Native Fluorescence for analytes’ detection

2009

11

9/1

84

[7]

Electrochemical sensors in microfluidics

2009

9

5/0

104

[8]

Optical detectors for microfluidic analysers

2010

27

19/1

95

[9]

Microfluidic systems nomenclature/terminology

2010

21

1/9

163

[10]

Techniques use for solvents mixing in microfluidic format

2010

14

6/4

120

[11]

PDMS surface modification

2010

15

7/1

108

[12]

20. MINIATURIZATION AND MICROFLUIDICS

TABLE 20.1 Selected publications on microfluidics organized according to the main focus of the review from the years 2008-2010

Genomics

2009

12

5/1

65

[13]

Diagnoses of DNA mutations

2010

20

6/6

193

[14]

Genomics, proteomics

Analysis of Nucleic acids and proteins

2009

10

9/1

129

[15]

Proteomics

Protein analysis by CE on a chip

2008

22

13/0

151

[16]

2D protein separation on a chip

2009

8

7/0

62

[17]

Protein analysis by CE on a chip (continuing the article 69)

2010

27

20/0

177

[18]

m-TAS for cell membrane analyses

2008

8

3/0

90

[19]

One drop cell analyses

2009

10

13/0

47

[20]

Separation of cells based on their properties

2010

19

6/1

122

[21]

Culturing the cells on a chip

2010

13

7/0

56

[22]

Culturing and studying tissue cultures

2010

13

6/1

133

[23]

Cell analysis

20.2. MICROFLUIDIC SYSTEMS FOR SEPARATIONS

Microfluidic systems applying Electrochemical detection for analysis of Nucleic acids

457

458

20. MINIATURIZATION AND MICROFLUIDICS

on the glass surface. Next, a thin layer (~1 mm) of a photoresist is spun on the metal layer. After exposure of the desired pattern through a mask or using a direct laser writer, the photoresist is developed and the exposed metal layer etched away. In the next step, a suitable etchant (e.g., HF based for glass) is used to etch channels in the exposed wafer surface. In glass, which etches isotropically, the resulting channels have a semi-circular profile. Finally, the microstructures are enclosed by glass bonding, where two glass wafers are bonded together under temperature close to the glass softening point (500e600 C). Many alternative materials are commonly used for microfabrication. Probably the most popular of all the materials is polydimethylsiloxane resin (PDMS), an attractive material especially for rapid prototyping suitable for low-pressure devices made by casting. Other popular materials include epoxy photoresists such as SU-8 [27], polycarbonates (PC), poly(methylmethacrylate) (PMMA), polyimide resin (PI), polyetheretherketone (PEEK) [28], cyclic olefin copolymers (COC) [29], or ceramics [30,31].

20.2.2. Miniaturization of HPLC Systems The first miniaturized separations in microfluidic chips relate primarily to electro-driven methods, while miniaturized LC has been pursued mainly in capillary-based systems. The following captures some of the most important aspects. • Pumps. The challenges have been increasingly a combination of columns with smaller particles (down to about 1 mm) that result in extreme back pressures and the requirement for pumps to work with extremely low flows in the range of nanoliters per minute [32]. • Dead volumes. The volumes related to the standard “nuts-and-bolts” plumbing found in the LC systems of the past is unacceptable in microcolumn arrangements. The potential for extremely low or zero dead volumeezero leakage connections of the microfabricated systems represents one of their major advantages. • Injection valve. A smaller column diameter means that a small sample volume injection is required. The required volume of the sample can be injected into the system either by a nanoliter injection loop or the amount of injected sample has to be adjusted by flow splitting. • Gradient elution. Mobile phase properties change with an altered composition. The most apparent change relates to mobile phase viscosity changes, resulting in flow rate variations. This problem can be solved either on the software side (viscosity of different mobile mixtures are known and tabulated) or by employing more expensive systems with precise flow meters. The flow meter gives a feedback to

20.3. COMMERCIAL INSTRUMENTATION

459

the pump, adjusting the pressure as required. Gradient elution in some nano-LC systems use mobile phase splitting, resulting in a waste of solvents that unnecessarily increases the price of analysis and has a negative impact on the environment. Multiloop injection valves, where each loop is filled with a different solvent mixture, is an example of an alternative solution to this problem [33]. • Separation column. Nano-LC or chip techniques use columns with equivalent inner diameters, ranging from 10 to 100 mm. In addition, the common 3e5-mm particles, packing with the diameter of 1.5e1.8 mm is gaining popularity for ultrahigh-performance liquid chromatography (UPLC)[32,34]. While the microfabricated systems typically do not have channels with circular cross sections, it has been shown that excellent performance can be obtained in columns with a noncylindrical shape[35]. As the process of packing is a tricky task, some scientists prefer to use alternatives, such as microfabricated or polymerized monolith columns [36]. The chemistry for preparing organic polymeric monoliths was introduced in the 1980s [37], and is gaining popularity, especially in the past decade. Similarly becoming popular are the silica-based monoliths with extremely high separation efficiencies demonstrated for peptides [38]. Monoliths do not need frits, and their preparation may be easier when compared to packing the capillaries with particle sorbents. The chemistry of monoliths also makes it possible to modify their surface structures. An interesting application of monolithic materials is for porous-layer, open-tubular (PLOT) columns showing promising performance for proteomic separations[39]. Chip-based systems are still in the minority; however, the first commercial instruments are already on the market, proving performance advantages related to the overall system miniaturization. • Detector. As the separation volume decreases, so does the detection volume. The sensitivity of the most widespread UVeVis detection systems can be improved when a U- or Z-shape detection cell is used [40]. Fluorescence detection, especially laser induced, does not suffer from the sensitivity issues; however, it is limited to naturally fluorescent or labeled analytes. While there are additional detection modes, including electrochemistry and conductivity, at present, the most attractive, miniaturization-compatible detection system for microfluidic LC is electrosprayemass spectrometry [41,46].

20.3. COMMERCIAL INSTRUMENTATION Many companies are designing, developing, manufacturing, or just using microfluidic systems of some form. A short sampler is given in Table 20.2.

460

20. MINIATURIZATION AND MICROFLUIDICS

TABLE 20.2 A sampler of microfluidics related companies Advanced Liquid Logic

http://www.liquid-logic. com/

digital microfluidics

Agilent Technologies

http://www.agilent.com

bio-analytical instruments

Biacore

http://www.biacore.com

technology for measuring protein-protein interaction based on SPR (Surface Plasmon Resonance).

Bio-Rad

http://www.bio-rad.com

chip base capillary electrophoresis

Caliper Life Sciences

http://www.caliperls. com/

number of microfluidic based products

Cellix

http://www.cellixltd.com

chips for studying blood cells

Cepheid

http://www.cepheid.com

real-time PCR for DNA analysis

Dolomite

http://www.dolomitemicrofluidics.com

designs and manufactures microfluidic chips

Eksigent Technologies

http://www.eksigent.com

liquid chromatography systems

Fluidigm

http://www.fluidigm.com

BioMark System - dPCR (digital PCR); TOPAZ protein crystallography on a chip

Gyros

http://www.gyros.com

Gyros immunoassay platform - ‘compact disks’ for microfluidic imunoanalyses

Helicos

http://www.helicosbio. com

HeliScope Single Molecule DNA Sequencer

Ibidi Integrated BioDiagnostics

http://www.ibidi.de/

Microfluidic chips for bacteria cultivation

IntegenX

http://integenx.com/ contact/

Microfluidic DNA sequencer Apolo 100

Micralyne Inc.

http://www.micralyne. com/

designs and manufactures i.a. microfluidic chips

Micronics

http://www.micronics.net

microfluidic instruments for invitro diagnosis

Micronit Microfluidics BV

http://www.micronit. com/

designs and fabricates microfluidic chips from glass

20.3. COMMERCIAL INSTRUMENTATION

461

TABLE 20.2 A sampler of microfluidics related companiesdcont’d Nanoterra

http://www.nanoterra. com/

focuses on Soft Lithography

Network Biosystems

http://www. networkbiosystems.com

Microfluidic systems for molecular biology

PRECISIONmicro

http://www. precisionmicro.com/23/ microfluidic-components/

designs and fabricates microfluidic chips

Shimadzu

http://www.shimadzu. com/

MultiNA chip electrophoresis for RNA, DNA separations

ThinXXS

http://www.thinxxs.com

designs and manufactures plastic microfluidic chips

Translume

http://www.translume. com/

designs and fabricates microfluidic chips from glass

Veredus laboratories

http://www.vereduslabs. com/

screening of DNA/RNA on a chip (VereFlu)

Waters Corporation

http://www.waters.com/

TrizaicÔ LC chips

20.3.1. Electrophoretic Systems Some of the very first microfluidic separation systems were based on electrophoresis. Presently, electrophoresis chip-based instruments are available mainly for separation of RNA, DNA, and SDS proteins, typically with fluorescence detection. Caliper Life Science (www.caliperls.com) developed one of the first commercial systems and has continued to produce microfluidic instruments, for a period of more than 10 years. A typical example is the LabChip GX/GXII (Figure 20.1) for analysis of nucleic acids and proteins loaded from 96- or 384-well microplates [42]. Caliper’s development is also behind the two very similar microfluidic bioanalyzers marketed by Agilent (2100 Bioanalyzer) since the late1990s and more recently by Bio-Rad (Experion Automated Electrophoresis Station). The Bioanalyzer can be used either for analysis of nucleic acids (DNA, RNA) and proteins or for cell cytometry. For each sample, the instrument uses different chips made from soda-lime glass wafers. The analyses of DNA, RNA, and proteins utilize an electrophoretic platform with 16 electrodes compatible with three chips. Chips for DNA, RNA, and proteins are designed for analyses of up to 12 samples, the protein chip for up to 10 samples. The flow cytometry uses a pressure-driven adapter. The detection of samples is completed by fluorescence detection, excited

462

20. MINIATURIZATION AND MICROFLUIDICS

FIGURE 20.1 (a) - Caliper’s LabChip GX/GXII is a bench microfluidic bioanalyzer for DNA, RNA and protein analysis (http://www.caliperls.com/). (b) - The 2100 Bioanalyzer sold by Agilent (https://www.genomics.agilent.com/). (c) -The Experion from Bio-Rad (http:// www.bio-rad.com/).

either by solid-state laser (lex1 ¼ 635 nm) or light-emitting diode (LED, lex2 ¼ 470 nm). In both instruments, the chips are disposable. A conceptually similar instrument is also produced by Shimadzu in Japan; however, at this time, the instrument is available only in selected markets. The main difference from the previous instruments is the use of a single-channel chip made of quartz and automated loading of the samples stored in a standard 96-well microtiter plate. LED detection is similar to the previous models as well as the separation length of 23 mm and a high-voltage power supply delivering up to 1.5 kV at maximum current of 250 mA. The instrument MCE-202 MultiNA, introduced in 2007, is designed for DNA and RNA analysis on a chip by agarose-gel electrophoresis. MultiNA has space for four chips and the process of analysis can be enacted simultaneously. The chips are automatically cleaned after analysis. The company claims that the lifetime of the chips is 3600 analyses if used correctly.

20.3.2. HPLC Systems The 1260 Infinity HPLC-Chip/MS instrument introduced by Agilent Technologies was the first system integrating the HPLC system with

463

20.3. COMMERCIAL INSTRUMENTATION

electrospray ionization. A credit cardesized chip contains enrichment and analytical columns, injection channel, electrospray tip, and all the necessary connections. The chip is positioned in a stainless steel holder (HPLCChip Cube), forming an interface for connecting HPLC pumps and positioning the chip in front of the mass spectrometer (series Agilent 6000). Recently, chips for different applications have become available, including proteomics, metabolomics, small-molecule analysis, biopharmaceutical analysis, and nucleotide analysis. The existing chips can be filled with a variety of sorbents, increasing the number of applications. Recently, new multilayered chip designs for multidimensional separations have also been developed. The HPLC-Chip Cube does not include an integrated solvent delivery device. Instead it is connected to an external system, the 1260 Infinity HPLC system (Figure 20.2). Since the polyimide chip limits the maximum applicable pressure due to mechanical stability, new chips, based on stainless steel, for operation at pressures over 1000 atmospheres have been under development at the writing of this text. A dedicated nano-HPLC system with nanoliter flow rates without splitting the mobile phase is produced by Exsigent. The pumps use a system called Microfluidic Flow Control,’ where the flow rate of the mobile phase is monitored and adjusted by changing the pressure supplied by compressed gas (airenitrogen). The compressor is able to increase the pressure difference by a factor of 36. For example, if the input pressure is 10 bars, the pressure inside the system can be set up to 360 bars. Two systems using MFC, the nanoLC and nanoLC-Ultra, were introduced. Both systems work with either capillary columns or a system upgrade called cHiPLCnanoflex, designed to work with microfluidic chips (Figure 20.3). Eksigent sells two cHiPLC-compatible chips, with either an analytical (70 mm  15 cm) or a trap (200 mm  0.5 mm) column. Both analytical and trap columns are packed with C18 particles (3 or 5 mm). The declared dead volume of the connection between capillaries and chips in cHiPLC is under 1 nl. External connection with nano-ESI MS is also possible. Solvent delivery system HPLC-chip cube

Rotor

Stator

Spray tip

FIGURE 20.2 HPLC-Chip/MS System 1260 Infinity. The polyimide chip is fixed between the rotor and stator of the injection valve. The chip’s integrated spraying tip is placed in front of the mass spectrometer in the HPLC-Chip Cube - www.agilent.com.

464

20. MINIATURIZATION AND MICROFLUIDICS

(a) (c)

(b)

Ch1 waste

Connerctor

Connerctor

Connerctor

The cHiPLCÔ is compatible with all nano-LC systems from Eksigent. (a) The chip, (b) cHiPLC-nanoflex. (c) The scheme of a possible connection of chips in cHiPLC http://www.eksigent.com

FIGURE 20.3

(a)

(b)

(c)

FIGURE 20.4 The TRIZAICÔ UPLC-chip/MS design by Waters e http://www.waters.com.

REFERENCES

465

The second product dedicated for nano-LC/MS applications introduced by one of the major instrument manufacturers is called TRIZAICÔ UPLC nanoTile, produced by Waters. While the dimensions of the integrated system with a trap column, analytical column, and ESI emitter are similar to the one produced by Agilent ( the analytical column has an inner diameter of 80 mm filled with 1.7-mm particles), the concept is different, with the main stress on the compatibility with the popular very high-pressure (>800 atm) nanoACQUITYÔ UPLC system produced by Waters. To withstand the pressure, a ceramic material, known in the electronic industry as Green TapeÔ [31] (www.dupont.com), was selected as the chip material. The fabrication process involves laser ablation and thermal sintering. The ceramic chip is placed in a plastic frame with the spray tip for online connection to the mass spectrometer and should include also some of the related electronics (Figure 20.4).

20.4. CONCLUSION While miniaturization is a prerequisite of achieving portability in LC systems, microfluidics represents a qualitative step toward chip-based LC systems. This short chapter provides an excursion into the microfluidics for separations with a brief overview of some of the commercial systems. It is worth stressing that, while the development and introduction of new microfluidic instrumentation is just at its beginning, the miniaturized technology is being used quite often in commercial systems without stressing it in the marketing of the final products. Typical examples include the fluidic manifolds, mixers, and flow sensors included in some of the high-performance chromatography systems. For more in-depth reading, one can consult the literature referenced here or some of the recent review articles [43e48].

Acknowledgment This research was supported by the Grant Agency of the Czech Republic (P301-11-2055).

References [1] Reyes DR, Iossifidis D, Auroux PA, Manz A. Micro total analysis systems. 1. Introduction, theory, and technology. Anal Chem 2002;74(12):2623e36. [2] Nguyen NT, Wereley ST. Fundamentals and applications of microfluidics. 2nd ed. Norwood, MA: Artech House; 2010. [3] Becker H, Ga¨rtner C. Polymer microfabrication technologies for microfluidic systems. Anal Bioanal Chem 2008;390(1):89e111. [4] Dishinger FJ, Kennedy RT. Multiplexed detection and applications for separations on parallel microchips. Electrophoresis 2008;29(16):3296e305.

466

20. MINIATURIZATION AND MICROFLUIDICS

[5] Tsao SW, DeVoe DL. Bonding of thermoplastic polymer microfluidics. Microfluid Nanofluid 2009;6(1):1e16. [6] Turgeon RT, Bowser MT. Micro free-flow electrophoresis: theory and applications. Anal Bioanal Chem 2009;394(1):187e98. [7] Schulze P, Belder D. Label-free fluorescence detection in capillary and microchip electrophoresis. Anal Bioanal Chem 2009;393(2):515e25. [8] Wei D, Bailey MJA, Andrew P, Ryha¨nen T. Electrochemical biosensors at the nanoscale. Lab Chip 2009;9(15):2123e31. [9] Borecki M, Korwin-Pawlowski ML, Beblowska M, Szmidt J, Jakubowski A. Optoelectronic capillary sensors in microfluidic and point-of-care instrumentation. Sensors 2010;10(4):3771e97. [10] Lim YC, Kouzani AZ, Duan W. Lab-on-a-chip: a component view. Microsyst Tech 2010;16(12):1995e2015. [11] Jeong GS, Chung S, Kim C, Lee S. Applications of micromixing technology. Analyst 2010;135(3):460e73. [12] Zhou JW, Ellis AV, Voelcker NH. Recent developments in PDMS surface modification for nmicrofluidic devices. Electrophoresis 2010;31(1):2e16. [13] Mir M, Homs A, Samitier J. Integrated electrochemical DNA biosensors for lab-on-achip devices. Electrophoresis 2009;30(19):3386e97. [14] Lien KY, Lee GB. Miniaturization of molecular biological techniques for gene assay. Analyst 2010;135(7):1499e518. [15] Ali I, Aboul-Enein HY, Gupta VK. Microchip-based nano chromatography and nano capillary electrophoresis in genomics and proteomics. Chromatographia 2009;69: S13e22. [16] Peng YY, Pallandre A, Tran NT, Taverna M. Recent innovations in protein separation on microchips by electrophoretic methods. Electrophoresis 2008;29(1):157e78. [17] Chen H, Fan ZH. Two-dimensional protein separation in microfluidic devices. Electrophoresis 2009;30(5):758e65. [18] Tran NT, Ayed I, Pallandre A, Taverna M. Recent innovations in protein separation on microchips by electrophoretic methods: an update. Electrophoresis 2010;31(1):147e73. [19] Suzuki H, Takeuchi S. Microtechnologies for membrane protein studies. Anal Bioanal Chem 2008;391(8):2695e702. [20] Chiu DT, Lorenz RM. Chemistry and biology in femtoliter and picoliter volume droplets. Acc Chem Res 2009;42(5):649e58. [21] Gossett DR, Weaver WM, Mach AJ, Hur SC, Tse HTK, Lee Amini H, et al. Label-free cell separation and sorting in microfluidic systems. Anal Bioanal Chem 2010;397(8): 3249e67. [22] Young EWK, Beebe DJ. Fundamentals of microfluidic cell culture in controlled microenvironments. Chem Soc Rev 2010;39(3):1036e48. [23] Gupta K, Kim DH, Ellison D, Smith C, Kundu A, Tuan J, et al. Lab-on-a-chip devices as an emerging platform for stem cell biology. Lab Chip 2010;10(16):2019e31. [24] Manz A, Graber N, Widmer HM. Miniaturized total chemical-analysis systemsda novel concept for chemical sensing. Sens Actuators B 1990;1(1e6):244e8. [25] Woolley AT, Mathies RA. Ultra-high-speed DNA fragment separations using microfabricated capillary array electrophoresis chips. Proc Natl Acad Sci USA 1994;91(24): 11348e52. [26] Terry SC, Jerman JH, Angell JB. A gas chromatographic air analyzer fabricated on a silicon wafer. IEEE Trans Electron Devices 1979;26(12):1880e6. ED-. [27] Abgrall P, Conedera V, Camon H, A.M. Gue AM, Nguyen NT. SU-8 as a structural material for labs-on-chips and microelectrochemical systems. Electrophoresis 2007; 28(24):4539e51. [28] Becker H, Locascio LE. Polymer microfluidic devices. Talanta 2002;56(2):267e87.

REFERENCES

467

[29] Jena RK, Chester SA, Srivastava V, Yue CY, Anand L, Lam YC. Large-strain thermomechanical behavior of cyclic olefin copolymers. Application to hot embossing and thermal bonding for the fabrication of microfluidic devices. Sens Actuators B 2011;155(1):93e105. [30] Gongora-Rubio MR, Espinoza-Vallejos P, Sola-Laguna L, Santiago-Avile´s JJ. Overview of low temperature co-fired ceramics tape technology for meso-system technology (MsST). Sens Actuators A 2001;89(3):222e41. [31] DuPont microcircuit materials. www2.dupont.com/MCM/en_US/tech_info/products/ltcc.html#951, 2011; accessed 07.08.2011. [32] Jorgenson JW. Capillary liquid chromatography at ultrahigh pressures. In: Yeung ES, Zare RN, editors. Annual review of analytical chemistry, vol. 3; 2010; p. 129e50, doi: 10.1146/annurev.anchem.1.031207.113014; 2010. [33] Cappiello A, Famiglini G, Florucci C, Mangani M, Palma P, Siviero A. Variable-gradient generator for micro- and nano-HPLC. Anal Chem 2003;75(5):1173e9. [34] Hernandez-Borges J, Aturki Z, Rocco A, Fanali S. Recent applications in nanoliquid chromatography. J Sep Sci 2007;30(11):1589e610. [35] Khirevich S, Holtzel A, Hlushkou D, Tallarek U. Impact of conduit geometry and bed porosity on flow and dispersion in noncylindrical sphere packings. Anal Chem 2007;79(24):9340e9. [36] Regnier FE. Microfabricated monolithg columns for liquid chromatographydsculpting supports for liquid chromatography. High Res Chromatogr 2000;23(1):19e26. [37] Tennikova TB, Blagodatskikh IV, Svec F, Tennikov MB. Phase-transition chromatography of polyesters on macroporous glycidyl methacrylate ethylene dimethacrylate copolymers. J Chromatogr 1990;509(1):233e8. [38] Minakuchi H, Nakanishi K, Soga N, Ishizuka N, Tanaka N. Octadecylsilylated porous silica rods as separation media for reversed-phase liquid chromatography. Anal Chem 1996;68(19):3498e501. [39] Luo Q, Yue G, Valaskovic GA, Gu Y, Wu SL, Karger BL. On-line 1D and 2D porous layer open tubular/LC-ESI-MS using 10-mu m-i.d. poly(styrene-divinylbenzene) columns for ultrasensitive proteomic analysis. Anal Chem 2007;79(16):6174e81. [40] Chervet JP, Ursem M, Salzmann JB. Instrumental requirements for nanoscale liquid chromatography. Anal Chem 1996;68(9):1507e12. [41] Y. Ishihama Y. Proteomic LC-MS systems using nanoscale liquid chromatography with tandem mass spectrometry. J Chromatogr A 2005;1067(1e2):73e83. [42] www.caliperls.com/products/labchip-systems/labchip-gx.htm, accessed August 2011. [43] Lazar IM, Grym J, Foret F. Microfabricated devices: a new sample introduction approach to mass spectrometry. Mass Spectrom Rev 2006;25:573e94. [44] Foret F, Kusy´ P. Microdevices in mass spectrometry. Eur J Mass Spec 2007;13:41e4. [45] Lee J, Soper SA, Murray KK. Microfluidic chips for mass spectrometry-based proteomics. J Mass Spec 2009;44(5):579e93. [46] Sikanen T, Franssila S, Kauppila TJ, et al. Microchip technology in mass spectrometry. Mass Spec Rev 2010;29(3):351e91. [47] Lavrik NV, Taylor LT, Sepaniak MJ. Nanotechnology and chip level systems for pressure driven liquid chromatography and emerging analytical separation techniques: a review. Anal Chim Acta 2011;694(1e2):6e20. [48] Kitagawa E, Otsuka K. Recent progress in microchip electrophoresis-mass spectrometry. J Pharm Biomed Anal 2011;55(4):668e78.

This page intentionally left blank

C H A P T E R

21

Capillary Electrochromatography A Look at Its Features and Potential in Separation Science A. Rocco, G. D’Orazio, Z. Aturki, S. Fanali Institute of Chemical Methodologies, Italian National Council of Research (IMC-CNR), Area della Ricerca di Roma I, Rome, Italy O U T L I N E 21.1. Introduction

469

21.2. Principles of Capillary Electrochromatography

470

21.3. Instrumentation 21.3.1. Stationary Phases and Capillary Columns in CEC 21.3.2. Detectors and Hyphenation

471 473 476

21.4. Method Optimization in CEC

478

21.5. Examples of Some Recent Applications

479

21.6. Conclusions and Future Trends

486

References

487

21.1. INTRODUCTION Capillary electrochromatography (CEC) is a recently developed separation technique offering high efficiency and high selectivity toward a large number of compounds. These two features are the results of the Liquid Chromatography: Fundamentals and Instrumentation http://dx.doi.org/10.1016/B978-0-12-415807-8.00021-3

469

Copyright Ó 2013 Elsevier Inc. All rights reserved.

470

21. CAPILLARY ELECTROCHROMATOGRAPHY

combination of the best properties of high-performance liquid chromatography (HPLC) and capillary electrophoresis (CE). In practice, CEC is an HPLC technique where a strong electro-osmotic flow (EOF), instead of pressure-driven flow, is used to move both the mobile phase and analytes into the capillary column [1]. Because the use of a stationary phase, CEC can offer a higher selectivity than CE, such as employing a reversed-phase separation for neutral analytes. In addition, separation mechanisms alternative to HPLC for charged compounds are also possible. In fact, these analytes can be differentiated not only by interactions with the stationary phase but also by their electrophoretic mobility, which occurs when the electric field is applied. Finally, the flat profile of EOF is responsible for very high efficiency. CEC is usually performed in fused silica capillaries containing stationary phases either bonded to the wall (open-tubular CEC) or to silica or polymeric particles (packed CEC); recently, monolithic stationary phases have also been used for CEC separations.

21.2. PRINCIPLES OF CAPILLARY ELECTROCHROMATOGRAPHY In CEC, the driving force responsible for the transport of analytes and mobile phase to the detector is the EOF generated by the application of an electric field to the capillary containing an electrolyte in contact with a charged surface (capillary wall or stationary phase). Specifically, this surface is covered by a diffuse layer of electrolyte ions with a charge opposite to that of the surface. The presence of this double layer, influenced by the adsorption of ions, is responsible for EOF generation. A theoretical explanation of electro-osmosis has been reported since 1974 by Pretorius, Hopkins, and Schieke, comparing experiments carried out in an electrochromatographic or chromatographic mode [2]. In addition to the good results and the novelty concerning EOF, in this publication an editor’s note remarks that electroosmosis, as a mechanism, was already used by other authors [3,4] in “Electrokinetic ultrafiltration.” Further detailed theoretical studies were carried out in the years 1980e1990 by Knox and Grant [5,6]. Later on, Rathore and Horvath reported an interesting comparison among HPLC, CE, and CEC [7]. Based on the Knox and Grant’s theory, utilizing packed capillaries, the EOF velocity can be calculated by the following equation [1]: ueof ¼ ε0 εr zE=h

(21.1)

21.3. INSTRUMENTATION

471

where ε0 is the vacuum dielectric constant, εr is the dielectric constant of the mobile phase, z is the “zeta” potential generated by the double layer, E is the applied field strength, and h is the viscosity of the mobile phase. Each experimental parameter, such as the buffer concentration, pH, or temperature, that can affect these variables and has to be properly selected. As can be observed in this equation, the EOF velocity is not influenced by the particle diameter for packed columns; on the contrary, in LC, the flow velocity decreases with the square root of the particle diameter. Therefore, in CEC, the particle diameter can be decreased, affording advantages, such as higher efficiency, without affecting EOF velocity, and consequently short analysis time. It is well known that in LC the plate heights (H), a useful indication of the chromatographic efficiency, can be calculated by using the following van Deemter equation: H ¼ A þ B=v þ Cv

(21.2)

where A, B, and C are parameters influenced by eddy diffusion, axial molecular diffusion, and resistance to mass transfer in the mobile and stationary phases, respectively. In CEC, the terms A and C are lower than those calculated in LC. For a correct evaluation of the column efficiency, some additional effects must be considered. Among them, thermal band broadening can play a negative role and should be minimized [6], by reducing the capillary i.d., the applied voltage, and the background electrolyte ionic strength. Substantially, in this way, a decrease of the current, with a consequent reduction of Joule heating, is advantageously generated.

21.3. INSTRUMENTATION Figure 21.1 depicts a schematic diagram of a commercial instrument widely used for CEC separations. The instrument can be divided into several key parts that ensure the correct operation of the tool. They include (a) the capillary with the electrode compartments and thermostating system, (b) the detector, (c) the high-voltage power supply, and (d) the data handling system and control. The capillary column is the most important part of the CEC system, because together with the selected mobile phase, it affects the selectivity of the chromatographic separation. A specific section is dedicated to the description of the main characteristics of the capillary column. The capillary is positioned in a cartridge and slotted into a thermostating system, where a detector (UV or fluorescence) reveals the separated zones. Alternatively, a special cartridge can be used for coupling the CEC

472

21. CAPILLARY ELECTROCHROMATOGRAPHY

CEC instrumentation

Thermostated system Detector

Frit Handling data system control

Capillary column

+

Electrode

Outlet vial

High-voltage power supply

N2

Inlet vial

Frit

FIGURE 21.1

Stationary phase

Instrumentation used for capillary electrochromatography.

capillary with a mass spectrometer (MS). Injection is usually carried out by application of external pressure (8e12 bar) or high voltage (5e10 kV) for a certain time (0.5e2.0 min). For CEC, mobile phases containing aqueouse polar organic mixtures, are usually employed, mainly in an isocratic mode. Acetonitrile and methanol are the most commonly used organic solvents for achieving excellent separations of several chargedechargeable and uncharged compounds. In addition, the instrumentation illustrated in Figure 21.1 can be used for capillary-zone electrophoresis (CZE), OT-CEC, capillary-liquid chromatography (CLC), and CEC. In the case of CEC, the two electrode compartments need to be pressurized during runs (8e12 bar) to avoid bubble formation. Usually, this commercial instrumentation (Agilent Technologies) allows us to run only isocratic elution, which enormously limits its use. The possibility of applying a gradient, employing different approaches, has been demonstrated in CEC [8e12]. Although some different approaches have been proposed, here, as an example, we briefly report on interesting work done by Yan, Dadoo, and Zare [9], where two electro-osmotic flows were merged and controlled by computer. Two capillaries were immersed in two reservoirs containing different mobile phases: the outlet ends were connected using a T-connector, in turn joined to the packed capillary. To obtain the gradient, the two capillaries, used for EOF generation, were operated with different power supplies, where the voltage could be controlled [9]. In a previous work, instead, Behnke and Bayer generated a gradient connecting the inlet capillary used for CEC with an HPLC pump [8]. This second system seems to work more easily

21.3. INSTRUMENTATION

473

and the idea was, most probably, taken from one of the commercially available instruments.

21.3.1. Stationary Phases and Capillary Columns in CEC The role of a stationary phase in CEC is essential not only to guarantee selective interactions toward analytes but also to assure the presence of a quite strong and constant electro-osmotic flow. For this reason, an intrinsic characteristic of stationary phases for CEC is to possess chargedechargeable groups [13,14]. This behavior reduces the availability of particulate stationary phases used in chromatography that can be utilized in CEC. Further, due to the still reduced use of this electro-driven technique on a large scale, few types of capillary columns are commercially available. As a consequence, most columns are prepared in analytical laboratories that make use of CEC [15]. However, the studies involving CEC are numerous, despite these limitations [16e23]. Depending on the nature of the stationary phases, capillary columns can be distinguished as packed with particulate material, monoliths with a polymeric network forming a continuous bed inside the capillary and open-tubular (OT), where the stationary phase is bonded to the inner wall of the capillary [24,25]. Packed capillary columns are most commonly employed. They are prepared with stationary phases arising from HPLC, usually utilizing fused-silica capillaries with an i.d. in the range 50e100 mm. The use of a lower i.d. can be advantageous, to increase efficiency and better dissipate heat during the run. The column can be prepared either completely packed throughout its length or with particles occupying only part of the capillary. The second approach has been used by several groups, because it allows the preparation of the detector window directly on the capillary (at a short distance from the outlet frit). In such a way, dead volumes can be avoided. Although detection through the packed stationary phase has been used [26,27], this approach cannot be recommended because of the drawbacks related to sensitivity [28], especially when the particles absorb UV light (silica derivatized with phenyl, vancomycin, etc.). Because most stationary phases (SPs) are represented by bonded silica materials (porous or non-porous), the residual silanol groups (SieOH) are enough to promote a high EOF [29]. However, to improve the chromatographic performance, there is a tendency to derivatize residual silanol groups (end-capping procedures) and the resulting SPs provide lower EOF and are less useful for CEC applications [15]. The stationary-phase particle size can range from 1 and 10 mm, but usually 3e5 mm particles are predominant. Regarding this subject, it is

474

21. CAPILLARY ELECTROCHROMATOGRAPHY

worth considering that the use of an electro-osmotic flow as driving force allows the use of very small particles, without high back pressure [30]. Packed capillary columns offer high efficiency, selectivity, and good loading capability, but special attention should be paid to the packing procedure, which can compromise all these parameters if it is not correctly carried out. Although different approaches have been used for packing capillary columns [1,31], the slurry packing method is the one most utilized. Here, particles are suspended in a selected solvent, such as acetone or hexane for an RP18 phase, sonicated, and introduced into a precolumn connected to the capillary that is closed at one end with a mechanical frit. Afterward, the slurry is pumped into the capillary. When the desired column length has been packed, the slurry is removed and water is pumped through the column to eliminate the organic solvent. To retain the stationary phase in the capillary column, inlet and outlet retaining frits are required. The two frits are prepared by heating the capillary flushed with water at high pressure (20e30 MPa). The capillary is cut close to the inlet frit and flushed in the opposite direction to eliminate excess stationary phase. The detector window is prepared using a razor at 0.5e1 cm from the outlet frit. Typical temperatures and times used for frit preparation are about 700 C for 6e8 sec. It is worth noting that, for frit preparation, the presence in the stationary phase of some silanol groups is essential. In some cases (absence of silanol groups), NaCl solutions (1e8 mM) can be helpful to obtain frits of good quality. Functional frits have to resist the high pressure applied and permit the mobile phase to flow through the column [32]. As a result, frit fabrication represents a major challenge for the preparation of columns of high efficiency. The procedure is quite simple to describe but complicated to realize, mainly because it needs some practical experience. In fact, to achieve useful results, it is necessary to control several experimental parameters, such as the solvent flow rate and type through the column, the wire temperature, and the time used to make the frit. Unfortunately, this modus operandi causes the removal of the polyimide coating at the frit zone and makes the column more fragile. Furthermore, a defective sintering can compromise the permeability of the column and change the properties of the packing material where the heat has been applied, generating nonhomogeneous packing. This is particularly disadvantageous for CEC, because it produces irregularity in the EOF, which can lead to bubble formation [33,34]. To bypass all these inconveniences, different solutions have been suggested to retain the stationary phase. Externally and internally tapered columns have been proposed as well as columns with inlet, or outlet, or both monolithic frits [35e38]. Even the entrapment of particulate material

21.3. INSTRUMENTATION

475

by sintering and solegel technology have been reported, where a column, hybrid between packed and monolith, is obtained [32,39]. A complete alternative to frit fabrication is the use of monolithic columns. They present several advantages, such as the synthesis in situ of a continuous bed, realizable with the capillary format of the columns; high permeability; high efficiency; and availability of monomers, which allow a separation medium with unique chromatographic properties to be obtained [15,40,41]. The continuous porous structure of monoliths, due to the presence of macropores and mesopores of different sizes, permits the use of high flow rates, due to the lower back pressure generated, compared with that of packed columns [42]. Monolithic columns can be obtained by the polymerization of different types of inorganic (e.g., silica, zirconia, carbon, titania) and organic monomers (e.g., polymethacrylate, poly(styreneedivinylbenzene), and polyacrylamide) or by the solegel process in the case of silica-based monoliths [43,44]. The organic monolithic columns can be prepared after the appropriate derivatization of the inner capillary wall for the anchoring of the polymer. Radical polymerization can be accomplished thermally or photochemically. The porosity of the bed can be controlled, varying the amount of porogenic solvent or salt, the concentration and ratio of monomers. Depending on the type of monomers, it is possible to distinguish water soluble comonomers (acrylamideebase monoliths), where a salt is used as porogen, and comonomers soluble in organic solvent (polystyrene- and methacrylate-based monoliths), where polymerization takes place in the presence of a porogen cosolvent (usually isopropanol). When monolithic beds are employed in CEC, a charged monomer is included in the polymeric mixture to assure the EOF [15,41,45e47]. Even if monoliths can be easily prepared, such a continuous bed may present low efficiency and poor reproducibility [48]. Silica monolithic stationary phases possess a well-defined structure, with a narrow pore size distribution of both macropores (a few micrometers in size) and mesopores (a few nanometers in size), which allow, respectively, high flow rates, high efficiency, and high specific area. Their preparation by the solegel process is based on several steps that ensure independent control of the size of the silica skeleton and through pores [42,49,50]. As a last step, chemical modification gives the surface feature desired for chromatographic separations. Alternatively, functional groups can be directly introduced by incorporation of functional monomers in the polymeric mixture [51]. Open-tubular columns, where the stationary phase is directly bonded or adsorbed to the inner wall of the capillary, have usually an i.d. between 10 and 60 mm [52]. The adsorption of the stationary phase, or modifier, can

476

21. CAPILLARY ELECTROCHROMATOGRAPHY

occur dynamically (weak interactions) or physically (modifier strongly adsorbed). Modifiers utilized for the preparation of adsorbed stationary phases can be grouped into (a) cationic surfactants, (b) polymeric surfactants, and (c) charged polymers [53]. The layer of the stationary phase can be fixed by covalent bonding or cross-linking. This approach offers a long capillary lifetime, but it usually requires a more complicated coating procedure [54]. This type of column is characterized by low convective dispersion, good efficiency, and high permeability [55]. The columns possess a very low phase ratio and low sample capacities. As a result, column overloading can easily occur, causing peak asymmetry and decreasing efficiency [53]. The conditioning of this kind of columns is fast, and the length can be increased without the occurrence of long analysis times and high back pressure. For the highest performance, an i.d.  20 mm is recommended. However, the small internal diameter aggravates the sensitivity when optical detection methods are employed [56]. To enlarge the wall surface available for “coating” with the stationary phase and to increase the phase ratio, chemical etching of the inner wall is used [53,57].

21.3.2. Detectors and Hyphenation Commercial instrumentation dedicated to CE is usually employed for CEC experiments and UV-Vis detectors (fixedemultiwave length or diode array) are the commonly used tools. Among them, diode array systems of the last generation offer valuable information concerning sensitivity, dynamic range, and the like. Despite the good sensitivity achieved, their applicability can be troublesome because of the short path length of the detector (usually the diameter of the capillary) and the low sample volume injected (nl). To improve the detector sensitivity, enlarged path lengths have been proposed, such as the use of bubble cells or a z-shaped cell [58e60]. A laser-induced fluorescence (LIF) detector, offering higher sensitivity, has also been applied in CEC experiments; however, it can be utilized for compounds exhibiting native fluorescence, such as the analysis of trace levels of flavins, [61] or for those that can be easily derivatized [62,63]. Another detector used in CEC is the conductivity detector that measures the signal in either a direct or indirect mode [64e67]. In addition, some authors employed amperometric detection [66,67]. Consequently, considering the limitations of the aforementioned systems, other detectors types were studied with the aim to (a) increase the sensitivity and (b) utilize modern tools offering the possibility to characterize the separated analytes by CEC. Among these, mass spectrometry has been widely investigated and successfully applied. Until now, the best results have been achieved

21.3. INSTRUMENTATION

477

coupling CEC with electrospray ionization (ESI) MS (single quadrupole, ion-trap, or time of flight). Although the coupling of these modern miniaturized techniques seems easy, considering their main characteristics, such as the low flow rate, at the moment, the hyphenation of CEC with MS cannot be considered satisfactory, mainly because appropriate interfaces are not commercially available. The coupling of CEC with MS was proposed by several authors, employing mainly a coaxial interface with an assisted sheath flow, as usually employed for CE experiments. Since the sheath liquid, necessary for the CEC run (application of the high electric field) and the ionization, is operated at a relatively high flow rate (3e5 ml/min), compared with the one through the column, it can be troublesome. This is recognized by electric instability (poor repeatability), reduced sensitivity, and turbulence (reducing the efficiency and resolution) [68e70]. Schmeer, Behnke, and Bayer [71] coupled a commercial CE instrument with the MS. The capillary (100 mm i.d.), packed with 1.5 mm reversed-phase particles was connected at one end with a commercial HPLC pump equipped with a T split, while the second end was positioned directly into the MS orifice. CEC separation of peptides was achieved using assisted pressure at the inlet side of the capillary. Pressure-assisted operation is available with commercial instruments such as the instrument produced by Agilent Technologies; however, in some cases it is not satisfactory because of its limitation (max 12 bar). It is worth mentioning that, the use of pressure assistance during the CEC run can cause some problems, especially related to the efficiency. In fact, lower efficiency must be expected because of working in a mixed separation mode, namely, LC (parabolic flow) and CEC (plug flow). Recently, CEC was coupled with MS, utilizing a laboratory-made liquidejunction interface. The system was used for the separation of some selected pesticides and drug enantiomers [72]. Capillaries were packed with silica RP18 or silica modified with vancomycin, respectively. The interface uses a small electrode compartment connected to the high voltage of the MS where the column met a tip, for the transfer of mobile phase and analytes to the MS. The opposite end of the capillary was connected to the inlet compartment for the application of high-voltage. In the liquidejunction compartment, a selected solvent mixture (e.g., methanoleformic acid) was added and an assisted hydrostatic pressure (10e30 mbar) was applied to improve analyte ionization. It is worth mentioning that, at the inlet end of the capillary, no pressure was applied. Employing such a tool, excellent analyte separations were achieved. The advantages in using the described interface include (a) higher sensitivity, easily achieved because all of the separated compounds reach the MS, and (b) no pressure (usually 8e12 atm) is necessary during the run. Finally, ordinary CEC instrumentation can be employed. Different types of liquidejunction interfaces were previously used by other groups for

478

21. CAPILLARY ELECTROCHROMATOGRAPHY

FIGURE 21.2 (a) Scheme of a liquidejunction interface. (b) Extraced ion chromato-

gram of CECeMS separation of some chiral drugs. Experimental conditions: capillary column, 100 mm i.d., 26.0 cm packed with (5 mm) silica modified with vancomycin. Mobile phase composition: 500 mM ammonium acetate pH 6/H2O/MeOH/ACN, 1:9:20:70 (v/v/ v/v); applied voltage, þ20 kV. Electrokinetic injection: 12 kV for 5 sec. The sheath liquid was 0.5% (v/v) AcH in MeOH/H2O 80:20 (v/v). 2.5 kPa, applied hydrostatic pressure to outlet compartment only. Sample: 1, alprenolol; 2, propranolol; 3, acebutolol; 4, nadolol. Source: Modified from [72].

coupling CEC with MS, very often applying inlet pressure during the run [73] or pressure at both electrode compartments [74]. A schematic diagram for a liquid-junction interface is shown in Figure 21.2 and was used for the enantiomeric separation of some selected racemic drugs.

21.4. METHOD OPTIMIZATION IN CEC Although individual researchers generally have their own approach for optimizing a separation by CEC some general guidelines can be given. First of all, a suitable solvent must be selected for the analytes (sample

21.5. EXAMPLES OF SOME RECENT APPLICATIONS

479

solution). Usually, the same mobile phase is selected as the solvent, but the buffer is diluted to obtain an ionic strength 100 times lower than the run buffer. In general, solvents with lower ionic strength than the mobile phase positively affect the injection of samples due to a stacking effect [31]. Since compounds of a different nature (positively or negatively charged, uncharged, and chargeable) can be analyzed by CEC, special attention must be paid to the choice of the mobile phase (especially its pH) and the stationary phase. Furthermore, it has to be taken into account that both the mobile and stationary phases have to be selected with the goal of obtaining a high EOF. As previously mentioned, most of the separations achieved with CEC are carried out by isocratic elution in the reversed-phase mode. The most commonly used mobile phases contain polar organic solvents, such as acetonitrile (ACN), methanol (MeOH), tetrahydrofuran (THF) in a mixture with water and buffer. Generally, CEC experiments are carried out at relatively high concentrations of ACN (70e80%, v/v) mixed with 30e20%, v/v of 5e25 mM TRIS buffer at pH 8. Such mixtures provide a relatively high EOF due to the high pH, which influences the dissociation of silanols of the stationary phase surface (or other chargeable groups), and to the presence of an organic solvent with a high dielectric constant and low viscosity. To improve the selectivity, MeOH can be added to the mobile phase either alone or as a mixture with ACN. Ethanol and isopropanol can also be used as organic modifiers to improve the selectivity of the method. Using solvents with different properties, a reversion of elution order can be obtain, for example, substituting THF (80%, v/v) with 20% TRIS changes the elution order of fluorene, anthracene, and fluoranthene. This effect is due to the properties of the solvent and not to the CEC mechanism; the same results are observed in HPLC [75]. Of course, a change in pH can strongly alter the selectivity, since it can affect both the EOF strength and the ionization of analytes. Two other parameters to consider for method optimization are the applied voltage and the temperature. By increasing both of them, it is possible to significantly reduce the analysis time, but very high values of applied voltage and temperature cause problems due to bubble formation.

21.5. EXAMPLES OF SOME RECENT APPLICATIONS Only recently has CEC research addressed application studies, including the analysis of pharmaceutical, biochemical, food, chiral, and environmental samples. Only a few reviews have been entirely devoted to applications in CEC [18,19,76].

480

21. CAPILLARY ELECTROCHROMATOGRAPHY

A significant number of these applications involve the use of relatively new column technologies, such as monolithic columns used especially in biochemical research. Mixed-mode stationary phases, suitable for the separation of hydrophilic compounds, HILIC, and novel chiral stationary phases (CSPs) have been evaluated as new packing materials. In this section, some applications concerning the analysis of chiral compounds, drugs, proteins and peptides, and natural products in food have been selected and summarized in Table 21.1. Anti-inflammatory drugs, barbiturates, antidepressants, benzodiazepines [77e81] were separated with high efficiency in a complex matrix with stable retention times and peak areas as well as good detector sensitivity. In recent times, CEC has been successfully applied to forensic sciences for the determination of illicit drugs. The use of a polar stationary phase instead of a conventional RP C18 [82,98] and coupling with massspectrometry [99] allowed the complete separation of a mixture of 10 illicit drugs and analyzed in urine samples with high sensitivity. Recently, particular attention has been paid to the application of CEC in food analysis, with the determination of endogenous components, such as vitamins, proteins, natural products, and phytochemical bioactive compounds (polyphenols, carotenoids, sterols) [20,100,101]. Many food components that exhibit strong antioxidant activities have been studied as cardiovascular and cancer chemopreventive agents. The beneficial effects on the human health resulting from the intake of these substances have been demonstrated. In this regard, several studies were performed by means of CEC to evaluate the nutritional value, bioactivity, and functional properties of these compounds [83e89]. The CEC methods were applied to real matrices, indicating the high performance of the CEC system. One of the most promising application fields for CEC is the separation of relatively complex polar biomolecules, such as peptide mixtures and protein digests obtained in proteomic research. These separations are characterized by high resolving power, fast separations, and the ability to handle complex matrices. In addition to conventional reversed phases, mixed mode (e.g., C18/SCX or C18/SAX) packing materials, in which hydrophobic interactions and ion-exchange properties are combined, were proposed for the separation of synthetic peptides mixtures [90e92]. Monoliths, created in situ from a wide variety of monomers, crosslinkers, and porogenic solvents have been evaluated for the separation of peptide mixtures. However, in most cases, peptide or protein separations have been used as models to illustrate the performance of CEC methods [102,103]. A few examples of peptides mixture separations are reported in Table 21.1. In the last decade, one of the most widely studied applications of CEC are chiral separations. Enantiomer separation represents a field of great

TABLE 21.1

Selected Applications by Capillary Electrochromatography Column

Detector

Mobile phase

Observations

Reference

Ketorolac, (K), 1-hydroxy analog of ketorolac (HK), 1-keto analog of ketorolac (KK), ketorolac decarboxylated (DK)]

Coated tablets

C18, packed column, 100 mm i.d.  23 cm, 5 mm

UV (195 nm)

50 mM ammonium formate/water/ ACN (10:20:70, v/v/v)

Separation in less than 9 min. Recovery of ketorolac in tablets: 98.5 0.8%

[77]

Amobarbital, phenobarbital, barbital, sulphanilamide, theophilline, 2,4-dimethyl quinoline, propranolol

Spiked serum samples

PolySULFOETHYL A, UV packed column, (214 nm) 50 mm i.d.  20 cm, 5 mm. Hydrophilic interaction capillary electrochromatography (HI-CEC)

100 mM TEAP buffer pH 2.8 in 80% v/v ACN

Column efficiencies for all analytes > 200,000 plates/m. Recoveries of basic drugs in serum: 97.86e100.21%

[78]

Barbital, phenobarbital, secobarbital

Spiked serum samples

NAIP, packed column, 75 mm i.d.  9 cm, 5 mm

UV (210 nm)

1.0 mM citrate buffer pH 5/MeOH (60:40, v/v)

Separation in less than 4.5 min

[79]

Caffeine, paracetamol, acetylsalicylic acid

Analgesic tablets

C8, packed column, 75 mm i.d.  21.5 cm, 5 mm

UV (210 nm)

25 mM ammonium formate pH 3/ACN (30:70, v/v)

Separation in less than 8 min. RSD values for peak area ratios 1.9e2.9%

[80]

Hexyl acrylatebased porous monolithic column, 100 mm i.d.  50 cm

MS (TOF)

5 mM ammonium acetate pH 7/ACN (30:70, v/v)

Limit of detection [81] (LOD) 0.6e1.8 ng/ml

Spiked urine Alprazolam,triazolam, samples chlordiazepoxide, lorazepam, nitrazepam, clonazepam flunitrazepam clorazepate, diazepam, prazepam

481

Matrix

21.5. EXAMPLES OF SOME RECENT APPLICATIONS

Analyte

(Continued)

482

TABLE 21.1 Selected Applications by Capillary Electrochromatographydcont’d Matrix

Column

Detector

Mobile phase

Observations

Reference

AM, MA, MDA, MDMA, MDEA, cocaine, codeine, heroin, morphine, 6-MAM

Spiked urine samples

CN, packed column, 75 mm i.d.  23 cm, 3 mm

UV (200 nm)

20 mM phosphate pH 2.5/ACN (80:20, v/v)

Column efficiencies for all analytes 205,800e301,700 plates/m. LODs: 5e15 ng/mL

[82]

a-tocopherol, g-tocopherol, d-tocopherol, a-tocopherol acetate

Vegetable oils

C18, packed column, 75 mm i.d.  7 cm, 3 mm

UV (205 nm)

0.01% ammonium acetate in ACN/ MeOH (50:50, v/v)

Separation within 2.5 min RSD values for peak areas 1.64e2.51 %

[83]

Hydroxytyrosol, protocatechuic acid, tyrosol, caffeic acid, o-coumaric acid, vanillic acid, oleuropein, p-coumaric acid, ferulic acid and syringic acid

Extra virgin olive oils

Cogent Bidentate C18, packed column, 75 mm i.d.  23 cm, 4.2 mm

UV (200 nm)

100 mM ammonium formate pH 3/water/ACN (5:65:30, v/v/v)

LODs: 0.015e2.5 mg/ [84] ml; recovery: 87e99%

Eriocitrin, narirutin, naringin, hesperidin, neohesperidin

Citrus juices

C18, packed column, 75 mm i.d.  8 cm, 5 mm

UV (195 nm)

2.5 mM ammonium formate pH 2.5/ACN (80:20, v/v)

Column efficiencies [85] for all analytes in the range 87,875 e 126,975 plates/m. Separation of all analytes within 10 min

21. CAPILLARY ELECTROCHROMATOGRAPHY

Analyte

SCX/C18, packed column, 50 mm i.d.  25 cm, 5 mm

UV (337 nm)

50 mM phosphate buffer pH 2.8 containing 50% ACN

Complete separation [86] of 11 phenolic compounds in less than 7.5 min. A partial resolution was obtained with a m-HPLC system

Eugenol, piperine, terpinen-4-ol, caryophyllene oxide, a-phellandrene, limonene, b-pinene, 3-carene, a-pinene, a-humulene, b-caryophyllene

Pepper extracts

C18, packed column, 100 mm i.d.  23 cm, 5 mm

UV (210, 265, 338 nm)

50 mM ammonium acetate pH 6/ACN (10:90, v/v)

RSD values for peak areas < 5.5%.

[87]

Ergosterol, campesterol, stigmasterol, b-sitosterol, avenasterol, cholesterol

Vegetable oils

LMAeEDMA monolithic column, 100 mm i.d.  8.5 cm

UV (210 nm)

ACN/2-propanol/ water containing 5 mM Tris, pH 8 (85:10:5, v/v/v)

Resolution of the analytes in less than 7 min. Column efficiencies 28800e60900 plates/m

[88]

Catechin, epicatechin, epigallocatechin, theophylline, caffeine

Green and black teas

Cogent Bidentate C18, packed column, 75 mm i.d.  23 cm, 4.2 mm

UV (200 nm)

Water/ACN (80:20, v/v) containing 5 mM ammonium acetate pH 4

LODs 1 mg/ml Recoveries ranged 90e112% for all analytes

[89]

Gly-Val, Tyr-Glu, Phe-Gly, Trp-Gly, Ala-Trp

d

Monolithic column with in situ copolymerization of SEMA and EDMA 100 mm i.d.  9.5 cm

UV (214 nm)

32 mM phosphate pH 3 with 50% ACN

RSD values for [90] retention times in the range 0.32e0.71%

(Continued)

483

Chamomile extracts

21.5. EXAMPLES OF SOME RECENT APPLICATIONS

Herniarin, caffeic acid, chlorogenic acid, umbelliferone, apigenin, luteolin, quercetin, rutin, naringenin, apigenin7-O-glucoside, luteolin7-O-glucoside

484

TABLE 21.1 Selected Applications by Capillary Electrochromatographydcont’d Matrix

Column

Detector

Mobile phase

Observations

Reference

Bradykinin, substance P, vasopressin, LHRH, bombesin, oxytocin, bradykinin (fragment1-5) methionine enkephalin, leucine enkephalin

d

Butylmetacrylate/SCX monolithic column 100 mm i.d.  25 cm

UV (206 nm)

(a) 50 mM ammonium formate pH 2.8/ACN/ water (20:70:10, v/v/v) (b) 50 mM sodium borate pH 9.5/ACN/ water (20:70:10, v/v/v)

Complete resolution [91] and higher separation efficiency with respect to CE system

Thrombin receptor antagonistic peptide TRAP-1, alanine-scan analogues TRAP 2- 6

d

C18/SCX, packed column, 100 mm i.d.  26.5 cm

UV (214 nm)

100 mM phosphate pH 6.5/ACN/water (10:32:58, v/v/v)

Mixed mode mechanism

[92]

Venlafaxine, O-desmethyl venlafaxine

Plasma samples

Vancomycin bonded to 5 mm diol silica, packed column, 75 mm i.d.  23 cm

UV (206 nm)

100 mM ammonium acetate pH 6/ water/ACN (5:5: 90, v/v/v)

LODs for each enantiomer 0.01 mg/ml. Column efficiencies for thr first enantiomers 110,634 and 110,958 N/m, respectively.

[93]

Racemic norgestrel, levonorgestrel

Enantiomeric purity in pharmaceutical formulations

25% CDCPC, packed column 100 mm i.d.  24 cm

UV (254 nm)

5 mM ammonium acetate in MeOH

Separation efficiencies [94] for the two enatiomers of norgestrel 106,725, 113,413 N/m with CEC and 58.313, 54.692 N/m in CLC

21. CAPILLARY ELECTROCHROMATOGRAPHY

Analyte

D, L-ephedrine

Enantiomeric impurity profiling

SCX-type CPS, packed column 100 mm i.d.  25 cm, 3.5 mm

UV (200 nm)

50 mM formic acid and 25 mM 2-amino1-butanol in ACN-MeOH (80:20, v/v)

Column efficiency 320.000 N/m

D, L-loxiglumide

Pharmaceutical formulations

Hepta-Tyr bonded to 5 mm diol silica, packed column, 75 mm i.d.  7 cm, 5 mm

UV (206 nm)

5mM phosphate pH 6/ACN (50:50, v/v)

Complete resolution [96] of the two enantiomers in less than 5 min. LODs: 0.5 and 0.3 mg/ml, for the two enantiomers, respectively

Mirtazapine, 8-hydroxymirtazapine, N-desmethylmirtazapine

Urine samples

Vancomycin bonded to 5 mm diol silica, packed column, 75 mm i.d.  23 cm

UV (200 nm)

100 mM ammonium acetate pH 6/ water/ MeOH/ ACN (5:15:30:50, v/v/v/v)

Simultaneous resolution of 6 enantiomers

[95]

NAIP: 3-(1,8-naphtalimido)propyl-modified silyl silica gel; amphetamine (AM), methamphetamine (MA), 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine,(MDMA), 3,4-methylenedioxyethylamphetamine (MDEA), 6-monoacethylmorphine (6-MAM). LMA: Lauryl metacrylate; LHRH: hormone-realizing hormone; SEMA, 2-(sulfooxy)ethyl methacrylate; EDMA, ethylenedimethacrylate; CDCPC; cellulose tris(3, 5-dichlorophenylcarbamate); SCX: N-(ACB)-N-tert-butylamido penicillamine sulfonic acid bon. RSD is relative standard deviation.

21.5. EXAMPLES OF SOME RECENT APPLICATIONS

[97]

485

486

21. CAPILLARY ELECTROCHROMATOGRAPHY

importance in pharmaceutical, biomedical, agrochemical, and environmental research: to monitor appropriate chiral discrimination content for drug quality control, to check for enantiomeric impurities in drugs, to investigate the toxicological effects of pesticides or pharmaceutical enantiomers, to evaluate the nutritional effects in functional food, and so forth. CEC research in this area has grown significantly, in both column technology development and applications in different fields. Chiral CEC can be performed in different modes, employing open-tubular, packed, or monolithic columns. The chiral selector can be present in the mobile or stationary phase. Chiral selectors such as cyclodextrins (CDs), polysaccharides, proteins, macrocyclic antibiotics, and ion exchangers are generally used to prepare CSPs. Until now, several reviews on enantioselective separations by CEC concerning general concepts and applications, especially in pharmaceutical and biomedical fields, have been published [104e107]. In addition to packed columns, monolithic type CSPs have been developed for enantioselective separations in CEC and gained considerable attention. The chiral resolving capability of monolithic columns is comparable to the corresponding packed columns with their inherent advantages, such as fritless design, ease of preparation with variable porosity and pore diameter, and fast diffusional mass transfer [108]. Despite many applications of chiral CEC, those in the field of pharmaceutical and biomedical analysis are represented by greater numbers [93e97]. The use of chiral drugs as models to evaluate novel chiral stationary phases in real matrices is often the basis of these studies.

21.6. CONCLUSIONS AND FUTURE TRENDS As reported in this paper, CEC is a modern separation technique with great potentiality, especially on an analytical scale. It can be applied to the separation of compounds of agrochemical, food, pharmaceutical, environmental, and biomedical interest (see Table 21.1). A large selection of stationary phases, mainly arising from HPLC, are available as column packings. On the other hand, monolithic columns can offer good performance in CEC as well because of the possibility of avoiding the use of frits. Commercial instrumentation is available with the possibility to use different detectors. Among these, MS has great potential for (a) improving the sensitivity and (b) determining the analytes’ structure. Because of these advantages, the coupling of CEC with MS is an issue that needs further attention by researchers. Although electrochromatography has been carried out mainly in capillary format, other approaches have been proposed, namely, chip and planar electrochromatography. These two formats are quite interesting,

REFERENCES

487

and show potential for future development, for example, miniaturization and speed for chip-based instruments and the possibility (a) to run two dimensional separations, (b) to increase sample loading in the planar format [109,110]. Recently, there has been a strong trend to develop miniature versions of laboratory instruments for field analysis (i.e., micro-total analysis system, m-TAS). Enhancement of extraction procedures, microfluidics, and chip devices allow researchers to obtain interesting results, especially in those areas where minute amounts of a sample are available (i.e., “omics” and forensic science). If, on the one hand, microdevices were quite easily developed for electrophoretic techniques, on the other hand, the miniaturization of the chromatographic system has met with several difficulties, such as the introduction of the stationary phase into narrow channels, the implementation of injectors, mechanical valves, and micro-pumps compatible with the chip format. CEC techniques can take advantage of the use of a voltage for both injection and developing a driving force through the separation media. With nearly established monolithic material, it is possible to synthesize in situ the stationary phases, simply filling chip channels with a liquid solution. Furthermore, due to the high permeability of this material, pressurization to eliminate bubble formation can be completely avoided.

References [1] Robson MM, Cikalo MG, Myers P, Euerby MR, Bartle KD. Capillary electrochromatography. J Microcol Sep 1997;9:357e72. [2] Pretorius V, Hopkins BJ, Schieke JD. Electro-osmosis: a new concept for high-speed liquid chromatography. J Chromatogr 1974;99:23e30. [3] Mould DL, Synge RLM. Electrokinetic ultrafiltration analysis of polysaccharides. A new approach to the chromatography of large molecules. Analyst 1952;77:964e70. [4] Mould DL, Synge RLM. Separations of polysaccharides related to starch by electrokinetic ultrafiltration in collodion membranes. Biochem J 1954;58:571e80. [5] Knox JH, Grant IH. Electrochromatography in packed tubes using 1.5 to 50 mm silica gels and ODS bonded silica gels. Chromatographia 1991;32:317e28. [6] Knox JH. Thermal effects and band spreading in capillary electro-separation. Chromatographia 1988;26:329e37. [7] Rathore AS, Horvath C. Separation parameters via virtual migration distances in high-performance liquid chromatography, capillary zone electrophoresis and electrokinetic chromatography. J Chromatogr A 1996;743:231e46. [8] Behnke B, Bayer E. Pressurized gradient electroehigh performance liquid chromatography. J Chromatogr A 1994;680:93e8. [9] Yan C, Dadoo R, Zare RN, Rakestraw DJ, Anex DS. Gradient elution in capillary electrochromatography. Anal Chem 1996;68:2726e30. [10] M. R, Taylor MR, P, Teale P. Gradient capillary electrochromatography of drug mixtures with UV and electrospray ionisation mass spectrometric detection. J Chromatogr A 1997;768:89e95. [11] Huber CG, Choudhary G, Horvath C. Capillary electrochromatography with gradient elution. Anal Chem 1997;69:4429e36.

488

21. CAPILLARY ELECTROCHROMATOGRAPHY

[12] Wang S, Jia L, Chen D. Pressurized CEC with gradient elution for separation of flavonoids from corn. J Sep Sci 2009;32:388e93. [13] Bartle KD, Myers P. Theory of capillary electrochromatography. J Chromatogr A 2001;916:3e23. [14] Smith NW, Carter-Finch AS. Electrochromatography. J Chromatogr A 2000;892: 219e55. [15] Mistry K, Krull I, Grinberg N. Capillary electrochromatography: an alternative to HPLC and CE. J Sep Sci 2002;25:935e58. [16] Dermaux A, Sandra P. Applications of capillary electrochromatography. Electrophoresis 1999;20:3027e65. [17] Colon LA, Burgos G, Maloney TD, Cintron JM, Rodriguez RI. Recent progress in capillary electrochromatography. Electrophoresis 2000;21:3965e93. [18] Eeltink S, Kok WT. Recent applications in capillary electrochromatography. Electrophoresis 2006;27:84e96. [19] Huo Y, Kok WT. Recent applications in CEC. Electrophoresis 2008;29:80e93. [20] Yang FQ, Zhao J. CEC of phytochemical bioactive compounds. Electrophoresis 2010;31:260e77. [21] Asensio-Ramos M, Hernandez-Borges J, Rocco A, Fanali S. Food analysis: a continuous challenge for miniaturized separation techniques. J Sep Sci 2009;32:3764e800. [22] Wu M, Wu R, Zhang Z, Zou H. Preparation and application of organic-silica hybrid monolithic capillary columns. Electrophoresis 2011;32:105e15. [23] El Rassi Z. Electrophoretic and electrochromatographic separation of proteins in capillaries: an update covering 2007e2009. Electrophoresis 2010;31:174e91. [24] Guihen E, Glennon JD. Recent highlights in stationary phase design for open-tubular capillary electrochromatography. J Chromatogr A 2004;1044:67e81. [25] Dong X, Wu R, Wu M, Zhu Y, Zou H. Recent progress of polar stationary phases in CEC and capillary liquid chromatography. Electrophoresis 2009;30:141e54. [26] Li S, Lloyd DK. Direct chiral separations by capillary electrophoresis using capillaries packed with an alpha(1)-acid glycoprotein chiral stationary phase. Anal Chem 1993;65:3684e90. [27] Desiderio C, Aturki Z, Fanali S. Use of vancomycin silica stationary phase in packed capillary electrochromatography I. Enantiomer separation of basic compounds. Electrophoresis 2001;22:535e43. [28] Behnke B, Grom E, Bayer E. Evaluation of the parameters determining the performance of electrochromatography in packed capillary. J Chromatogr A 1995;716:207e13. [29] Pursch M, Sander LC. Stationary phases for capillary electrochromatography. J Chromatogr A 2000;887:313e26. [30] Svec F. Capillary electrochromatography: a rapidly emerging separation method. In: Shepers T, Freitag R, editors. Advances in biochemical engineering/biotechnology: modern advances in chromatograph. Berlin, Heildelberg: Springer-Verlag; 2002. p. 1e47. [31] Poole CF. The essence of chromatography. Amsterdam: Elsevier Science B.V; 2003. [32] Colo´n LA, Maloney TD, Fermier AM. Packing columns for capillary electrochromatography. J Chromatogr A 2000;887:43e53. [33] Legido-Quigley C, Marlin ND, Melin V, Manz A, Smith NW. Advances in capillary electrochromatography and micro-high performance liquid chromatography monolithic columns for separation science. Electrophoresis 2003;24:917e44. [34] Rapp E, Bayer E. Improved column preparation and performance in capillary electrochromatography. J Chromatogr A 2000;887:367e78. [35] Lord GA, Gordon DB, Myers P, King BW. Tapers and restrictors for capillary electrochromatography and capillary electrochromatographyemass spectrometry. J Chromatogr A 1997;768:9e16.

REFERENCES

489

[36] Choudhary G, Horvath C, Banks JF. Capillary electrochromatography of biomolecules with on-line electrospray ionization and time-of-flight mass spectrometry. J Chromatogr A 1998;828:469e80. [37] Chen JR, Dulay MT, Zare RN, Svec F, Peters E. Macroporous photopolymer frits for capillary electrochromatography. Anal Chem 2000;72:1224e7. [38] Rocco A, Fanali S. Capillary electrochromatography without external pressure assistance. Use of packed columns with a monolithic inlet frit. J Chromatogr A 2008;1191: 263e7. [39] Chirica GS, Remcho VT. A simple procedure for the preparation of fritless columns by entrapping conventional high performance liquid chromatography sorbents. Electrophoresis 2000;21:3093e101. [40] Hilder EF, Svec F, Frechet JMJ. Polymeric monolithic stationary phases for capillary electrochromatography. Electrophoresis 2002;23:3934e53. [41] Maruska A, Kornysova O. Continuous beds (monoliths): stationary phases for liquid chromatography formed using the hydrophobic interaction-based phase separation mechanism. J Biochem Biophys Meth 2004;59:1e48. [42] Tanaka N, Kobayashi H, Ishizuka N, Minakuchi H, Nakanishi K, Hosoya K, et al. Monolithic silica columns for high-efficiency chromatographic separations. J Chromatogr A 2002;965:35e49. [43] Ou J, Dong J, Dong X, Yu Z, Ye M, Zou H. Recent progress in polar stationary phases for CEC. Electrophoresis 2007;28:148e63. [44] Guillarme D, Ruta J, Rudaz S, Veuthey J-L. New trends in fast and high-resolution liquid chromatography: a critical comparison of existing approaches. Anal Bioanal Chem 2010;397:1069e82. [45] Svec F, Huber CG. Monolithic materials: promises, challenges, achievements. Anal Chem 2006;78:2100e7. [46] Svec F, F. Porous polymer monoliths: amazingly wide variety of techniques enabling their preparation. J Chromatogr A 2010;1217:902e24. [47] Eeltink S, Svec F. Recent advances in the control of morphology and surface chemistry of porous polymer-based monolithic stationary phases and their application in CEC. Electrophoresis 2007;28:137e47. [48] Myers P, Bartle K. Miniaturization in LCeMS. LCeGC Eure 2002;2:2e5. [49] Rieux L, Niederlander H, Varpoorte E, Bischoff R. Silica monolithic columns: synthesis, characterisation and applications to the analysis of biological molecules. J Sep Sci 2005;28:1628e41. [50] Allen D, El Rassi Z. Silica-based monoliths for capillary electrochromatography: methods of fabrication and their applications in analytical separations. Electrophoresis 2003;24:3962e76. [51] Wu R, Hu L, Wang F, Ye M, Zou H. Recent development of monolithic stationary phases with emphasis on microscale chromatographic separation. J Chromatogr A 2008;1184:369e92. [52] Ishii D, Takeuchi T, Wada A. Introduction to microscale high-performance liquid chromatography. In: Ishii D, editor. Microscale columns. New York: VCH Publisher; 1988. p. 3369. [53] Kapnissi-Christodoulou CP, Zhu X, Warner IM. Analytical separations in opentubular capillary electrochromatography. Electrophoresis 2003;24:3917e34. [54] Nyholm LM, Markides KE. Column preparation for reversed-phase high-temperature open tubular column liquid chromatography. J Chromatogr A 1998;813:11e20. [55] Swart R, Kraak JC, Poppe H. Recent progress in open tubular liquid chromatography. TRAC Trends Anal Chem 1997;16:332e42. [56] Bruin GJM, Tock PPH, Kraak JC, Poppe H. Electrically driven open-tubular liquid chromatography. J Chromatogr 1990;517:557e72.

490

21. CAPILLARY ELECTROCHROMATOGRAPHY

[57] Pesek JJ, Matyska MT. Open tubular capillary electrokinetic chromatography in etched fused-silica tubes. J Chromatogr A 2000;887:31e41. [58] Banholczer A, Pyell U. In-column versus on-column photometric detection in capillary electrochromatography with capillaries packed with octadecylsilica gel. J Microcol Sep 1998;10:321e8. [59] Steiner F, Scherer B. Instrumentation for capillary electrochromatography. J Chromatogr A 2000;887:55e83. [60] Desiderio C, Rudaz S, Veuthey JL, Raggi MA, Fanali S. Use of vancomycin silica stationary phase in packed capillary electrochromatography, IV. Enantiomer separation of fluoxetine and norfluoxetine employing UV high sensitivity detection cell. J Sep Sci 2002;25:1291e6. [61] Wu Y, Lin J, Wu X, Lin X, Xie Z. Rapid analysis of trace levels of flavins by pressurized capillary electrochromatographyelaser induced fluorescence detection with sulfonated N-octadecyl methacrylate monolith. J Pharm Biomed Anal 2010;53:1324e31. [62] Wu Y, Wang X, Wu Q, Lin X, Xie Z. Separation and determination of structurally related free bile acids by pressurized capillary electrochromatography coupled to laser induced fluorescence detection. Anal Meth 2010;2:1927e33. [63] Oguri S, Okuya Y, Yanase Y, Suzuki S. Post-column derivatization capillary electrochromatography for detection of biogenic amines in tuna-meat. J Chromatogr A 2008;1202:96e101. [64] Hilder EF, Zemann AJ, Macka M, Haddad PR. Anion-exchange capillary electrochromatography with indirect UV and direct contactless conductivity detection. Electrophoresis 2001;22:1273e81. [65] Kuban P, Kuban P, Kuban V, Hauser PC, Bocek P. Capillary electrochromatography of inorganic cations in open tubular columns with a controllable capacity multilayered stationary phase architecture. J Chromatogr A 2008;1190:377e82. [66] Hilmi A, Luong JHT. In-line coupling capillary electrochromatography with amperometric detection for analysis of explosive compounds. Electrophoresis 2000;21:1395e404. [67] Wu W, Yuan X, Wu X, Lin X, Xie Z. Analysis of phenolic xenoestrogens by pressurized CEC with amperometric detection. Electrophoresis 2010;31:1011e8. [68] Lord GA, Gordon DB, Tetler LW, Carr CM. Electrochromatography electrospray mass spectrometry of textile dyes. J Chromatogr A 1995;700:27e33. [69] Strickmann DB, Blaschke G. Capillary electrochromatographyeelectrospray ionization mass spectrometry for the qualitative investigation of the drug etodolac and its metabolites in biological samples. J Chromatogr B 2000;748:213e9. [70] Strickmann DB, Chankvetadze B, Blaschke G, Desiderio C, Fanali S. Separation and identification of etodolac and its urinary phase I metabolites using capillary electrochromatography and on-line capillary electrochromatographyeelectrospray ionisation mass spectrometry coupling. J Chromatogr A 2000;887:393e407. [71] Schmeer K, Behnke B, Bayer E. Capillary electrochromatographyeelectrospray mass spectrometry: a microanalysis technique. Anal Chem 1995;67:3656e8. [72] D’Orazio G, Fanali S. Coupling capillary electrocheomatography with mass spectrometry by using a liquidejunction nano-spray interface. J Chromatogr A 2010;1217: 4079e86. [73] Koerner T, Turck K, Brown L, Oleschuk RD. Porous polymer monolith assisted electrospray. Anal Chem 2004;76:6456e60. [74] Dekkers SEG, Tjaden UR, Van der Greef J. Development of an instrumental configuration for pseudo-electrochromatography electrospray mass spectrometry. J Chromatogr A 1995;712:201e9. [75] Dittmann MM, Rozing GP. Capillary electrochromatographydinvestigation of the influence of mobile phase and stationary phase properties on electroosmotic velocity, retention, and selectivity. J Microcol Sep 1997;9:399e408.

REFERENCES

491

[76] Van Hoenacker G, Van den Bosch T, Rozing G, Sandra P. Recent applications of capillary electrochromatography. Electrophoresis 2001;22:4064e103. [77] Orlandini S, Furlanetto S, Pinzauti S, D’Orazio G, Fanali S. Analysis of ketorolac and its related impurities by capillary electrochromatography. J Chromatogr A 2004;1044: 295e303. [78] Fu H, Jin W, Xiao H, Xie C, Guo B, Zou H. Determination of basic pharmaceuticals in human serum by hydrophilic interaction capillary electrochromatography. Electrophoresis 2004;25:600e6. [79] Ohiama K, Wada M, Lord GA, Ohba Y, Fujishita O, Nakashima K, et al. Capillary electrochromatographic analysis of barbiturates in serum. Electrophoresis 2004;25: 594e9. [80] Pucci V, Mandrioli R, Raggi MA, Fanali S. Reversed-phase capillary electrochromatography for the simultaneous determination of acetylsalicylic acid, paracetamol and caffeine in analgesic tablets. Electrophoresis 2004;25:615e21. [81] Blas M, McCord BR. Determination of trace levels of benzodiazepine in urine using capillary electrochromatography-time of flight mass spectrometry. Electrophoresis 2008;29:2182e92. [82] Aturki Z, D’Orazio G, Fanali S, Rocco A, Bortolotti F, Gottardo R, et al. Capillary electrochromatographic separation of illicit drugs employing a cyano stationary phase. J Chromatogr A 2009;1216:3652e9. [83] Aturki Z, D’Orazio G, Fanali S. Rapid assay of vitamin E in vegetable oils by reversedphase capillary electrochromatography. Electrophoresis 2005;26:798e803. [84] Aturki Z, Fanali S, D’Orazio G, Rocco A, Rosati C. Analysis of phenolic compounds in extra virgin olive oil by using reversed-phase capillary electrochromatography. Electrophoresis 2008;29:1643e50. [85] Desiderio C, De Rossi A, Sinibaldi M. Analysis of flavanone-7-O-glycosides in citrus juices by short-end capillary electrochromatography. J Chromatogr A 2005;1081: 99e104. [86] Fonseca FN, Tavares MFM, Horvath C. Capillary electrochromatography of selected phenolic compounds of Chamomilla recutita. J Chromatogr A 2007;1154:390e9. [87] Musenga A, Mandrioli R, Ferranti A, D’Orazio G, Fanali S, Raggi MA. Analysis of aromatic and terpenic constituents of pepper extracts by capillary electrochromatography. J Sep Sci 2007;30:612e9. [88] Lerma-Garcia MJ, Simo-Alfonso EF, Ramis-Ramos G, Herrero-Martinez JM. Rapid determination of sterols in vegetable oils by CEC using methacrylate ester-based monolithic columns. Electrophoresis 2008;29:4603e11. [89] Uysal UD, Aturki Z, Raggi MA, Fanali S. Separation of catechins and methylxanthines in tea samples by capillary electrochromatography. J Sep Sci 2009;32:1002e10. [90] Wu R, Zou H, Fu H, Jin W, Ye M. Separation of peptides on mixed mode of reversedphase and ion-exchange capillary electrochromatography with a monolithic column. Electrophoresis 2002;23:1239e45. [91] Adu JK, Lau SS, Watson DG, Euerby MR, Skellern GG, Tettey JNA. Capillary electrochromatography of therapeutic peptides on mixed-mode butylmethacrylate. Electrophoresis 2005;26:3445e51. [92] Yang Y, Y, R. I, Boysen RI, Hearn TW. Use of mixed-mode sorbents for the electrochromatographic separation of thrombin receptor antagonistic peptides. J Chromatogr A 2005;1079:328e34. [93] Fanali S, Rudaz S, Veuthey JL, Desiderio C. Use of vancomycin silica stationary phase in packed capillary electrochromatography, II. Enantiomer separation of venlafaxine and O-desmethylvenlafaxine in human plasma. J Chromatogr A 2001;919:195e203. [94] Chankvetadze B, Kartozia I, Yamamoto C, Okamoto Y, Balschke G. Comparative study on the application of capillary liquid chromatography and capillary

492

[95] [96]

[97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110]

21. CAPILLARY ELECTROCHROMATOGRAPHY

electrochromatography for investigation of enantiomeric purity of the contraceptive drug levonorgestrel. J Pharm Biomed Anal 2003;30:1897e906. Bicker W, Hebenstreit D, Lammerhofer M, Lindner W. Enantiomeric impurity profiling in ephedrine samples by enantioselective capillary electrochromatography. Electrophoresis 2003;24:2532e42. Fanali S, D’Orazio G, Quaglia MG, Rocco A. Use of a Hepta-Tyr antibiotic modified silica stationary phase for the enantiomeric resolution of d, l-loxiglumide by electrochromatography and nano-liquid chromatography. J Chromatogr A 2004;1051: 247e52. Aturki Z, Scotti V, D’Orazio G, Rocco A, Raggi MA, Fanali S. Enantioselective separation of the novel antidepressant mirtazapine and its main metabolites by CEC. Electrophoresis 2007;28:2717e25. Lurie LS, Conver TS, Ford VI. Simultaneous separation of acidic, basic, and neutral organic compounds, including strong and moderate acids and bases, by capillary electrochromatography. Anal Chem 1998;70:4563e9. Aturki Z, D’Orazio G, Rocco A, Bortolotti F, Gottardo R, Tagliaro F, et al. CEC-ESI ion trap MS of multiple drugs of abuse. Electrophoresis 2010;31:1256e63. Scherz H, Huck CW, Bonn GK. CEC and EKC of natural compounds. Electrophoresis 2007;28:1645e57. Herrero M, Garcia-Canas V, Simo C, Cifuentes A. Recent advances in the application of capillary electromigration methods for food analysis and foodomics. Electrophoresis 2010;31:205e28. Bandilla D, Skinner CD. Capillary electrochromatography of peptides and proteins. J Chromatogr A 2004;1044:113e29. Miksik I, Sedlakova P. Capillary electrochromatography of proteins and peptides. J Sep Sci 2007;30:1686e703. Fanali S, Catarcini P, Blaschke G, Chankvetadze B. Enantioseparations by capillary electrochromatography. Electrophoresis 2001;22:3131e51. Gubitz G, Schmid MG. Recent advances in chiral separation principles in capillary electrophoresis and capillary electrochromatography. Electrophoresis 2004;23:3981e96. Mangenlings D, Maftouth M, Vander Heyden Y. Monoliths with chiral surface functionalization for enantioselective capillary electrochromatography. J Sep Sci 2005;28:691e709. Preinerstorfer B, La¨mmerhofer M. Recent accomplishments in the field of enantiomer separation by CEC. Electrophoresis 2007;28:2527e65. La¨mmerhofer M, Gargano A. Monoliths with chiral surface functionalization for enantioselective capillary electrochromatography. J Pharm Biomed. Anal 2010;53: 1091e123. Masa´r M, Sydes D, Luc M, Kaniansky D, Kuss H- M. Determination of ammonium, calcium, magnesium, potassium and sodium in drinking waters by capillary zone electrophoresis on a column coupling chip. J Chromatogr A 2009;1216:6252e5. Dzido TH, Plocharz PW, Chomicki A, Halka-Grysinska A, Polak B. Pressurized planar electrochromatography. J Chromatogr A 2011;1218:2636e47.

Index Page numbers with “f” denote figures; “t” tables.

A

Absolute molecular mass, determination of, 207e217 Accuracy, defined, 319 Accurate mass measurements, 318e319 Acetonitrile (ACN), 46, 68, 73, 99e100, 105e107, 109e110, 148, 228, 269e271, 279, 299e300, 391e395, 471e473, 479 Acidebase equilibria, 89e90 pH, buffers and measurement of, 89e90 pH, changes in retention with, 89 Adsorbents, 52te53t properties of, 126e128 alkyl chain length, 127 base matrix, 126 binding capacity, 127e128 coupling chemistry, 126e127 ligand density, 127e128 ligand type, 127 Adsorption band shape dependence on, 411e413 chromatography, See Normal-phase chromatography data, determination of, 414e415 frontal analysis, 414 inverse method, 414e415 isotherms, 413e414 Langmuir, 413e414 model, 410e415 principles, in hydrophobic interaction chromatography, 123e125 Agglomerated ion exchangers, 169e170 Alkyl chain length, 127 Aminophosphonic acid (APA), 161e162 Amperometric detection, 182 Analytical columns, column packing procedures for, 61e63 Anion exchange columns, ion chromatography, 170e171, 172e174 eluents for, 164e165

Annular chromatography, 445e446 Artificial neural networks (ANN), 390 At-line LCeNMR, quantitative capabilities of, 369e370 At-line micro-NMR analysis, 364e367 Atmospheric-pressure chemical ionization (APCI), 312e313 Atmospheric-pressure photoionization (APPI), 313e315 Average particle size, choice of, 63e65

B

Back pressure, of LC columns, 66 Band broadening, in size-exclusion chromatography, 202e205 extracolumn effects, 204e205 Baseline problems, associated with gradient elution, 281 Base matrix, 126 Bi-Langmuir model, 413e414 Biological activities, of QSRR relationships, 400 Biomolecules properties of, 130 purification of practical aspects of, 131e132 strategies, 130e131 Bio-partitioning micellar chromatography (BMC), 398e399 Bio-partitioning micellar liquid chromatography, 99 Body fluids, metabolite identification in, 376e377 Bonded silica, physicochemical characterization of, 61 Bonding, 58e59 Butanol, 99e100, 230

C

Caliper Life Science, 460te461t, 461, 461e462 Capacity, of columns, 66e67

493

494

INDEX

Capacity, of columns (Continued) in capillary electrochromatography, 473e476 Capillary electrochromatography (CEC), 325e327, 469e492 applications of, 479e481 future trends of, 486e487 instrumentation of, 471e478 capillary columns, 473e476 detectors, 476e478 hyphenation, 476e478 stationary phases, 473e476 interfacing with mass spectrometry, 326e327 method optimization in, 478e479 principles of, 470e471 stationary phases, characterization of, 395 Capillary electrophoresis (CE), 454 Carbonate-type selectivity, 161 Cation exchange columns, ion chromatography, 171e175 eluents for, 165 Cell constant, 179e180 Cetylpyridinium chloride, 100 Cetyltrimethylammonium bromide (CTAB), 99e100 Cetyltrimethylammonium chloride, 93f Chaotropic effect, 96 Chaotropic ions, as additives, 96e97 Charged-aerosol detectors (CAD), 12e13, 301e302 Charged residue mechanism (CRM), 311 Chelation ion chromatography (CIC), principles of, 161e162 Chemical detectors, 217e218 Chemically bonded stationary phases, 113 Chemometric background correction, 336e337 CHI index, 396 Chiral stationary phases (CSPs), 408, 480, 486 Chromatographic testing, of columns, 67e70 hydrophobicity, 67 metal content, 69e70 polar selectivity, 68 shape selectivity, 68e69 silanophilic activity, 67e68 Cleaning in Place (CIP) procedures, 129, 132, 443e444 Closed-loop recycling (CLR) chromatography, 436

Coherent anti-Stokes Raman scattering (CARS), 335 Column(s) compartment, 293e294 conditions, effect on separation, 275e277 equilibration, associated with gradient elution, 281 internal diameter, 63e65 model, 409e410 packing, 6e9 plate number, 254 resistance factor, 61e62 screening, priority of, 259e260, 262e263, See also Column technology, in liquid chromatography Column technology, in liquid chromatography, 41e86 average particle size, choice of, 63e65 back pressure, 66 capacity, 66e67 chromatographic testing, 67e70 hydrophobicity, 67 metal content, 69e70 polar selectivity, 68 shape selectivity, 68e69 silanophilic activity, 67e68 classification of, 50e51 column internal diameter, 63e65 comparison of, 77e82, 77t development during 2000e2011, 74e77 equilibration time, 65 hardware, 46e48, 47f loadability, 66e67 maintenance and troubleshooting of, 70e74 polymer-based columns, 70e72 silica-based columns, 70e73 miniaturization, 48e49 mobile phase, 49 optimum-flow conditions, choice of, 66 packing materials, 51e63 bonded silica, physicochemical characterization of, 61 particle shape, 54e56 particle size, 54e56 particle size distribution, 54e56 pore structure parameters, 57e58 silica, surface functionalization of, 58e61 stationary phase, 49 temperature, choice of, 66, See also Column(s)

INDEX

Complexation chromatography, See Dynamic chelating chromatography Computer-assisted interpretive optimization, of mobile-phase composition, 245e246 Condensation-nucleation light-scattering detector (CNLSD), 301 Conductivity detectors, 302 Continuous-variable screening, 262 Copolymers, chemical composition distribution, 218 Core performance tests, 264 Core-shell particles, general rate model for, 27e28 Coupling chemistry, 126e127 Critical micelle concentration (CMC), 97e99 Cryogenic probes, 357

D

Detector overview, 296e298, 297t Differential refractometer (DRI), 207e208 Dipolarity, 234e235 Direct LCeNMR hyphenation, 359e363 Disordered silica vs. ordered mesoporous silica, 76e77 Displacement chromatography, 432e433 effect, 431e432 process, 146 Donnan membrane equilibrium, 160e161 DryLab, 246 Dwell-volume changes, associated with gradient elution, 281 Dynamically modified ion-exchangers, 170 Dynamic axial compression (DAC), 63 Dynamic chelating chromatography, 102 Dynamic light scattering, See Quasi-elastic light-scattering Dynamic nuclear polarization (DNP), 354

E

Electrochemical detection, in ion chromatography, 179e182 amperometry, 182 nonsuppressed conductivity, 179e180 suppressed conductivity, 180e181 Electrochemical detectors (ECDs), 303 coulometric, 303 polarographic, 303 Electrokinetic ultrafiltration, 470 Electronic records, 265

495

Electron ionization, in liquid chromatographyemass spectrometry, 316e317 Electrophoresis (CE), 454 capillary, 454 Electrospray ionization (ESI), 183e184, 309e312 Electrospray ionization mass spectrometry (ESI-MS), 476e477 Electrostatic ion chromatography (EIC), 94e95, 162e163 Eluent(s) converters, 175e179 generators, 175e179 for ion chromatography, 164e165, 166t for anion exchange, 164e165 for cation exchange, 165 suppressor, 180 Elution additives, 129 Elution strength, 226e228 assessment of, 229e231 global polarity estimators, 229e230 Hildebrand solubility parameter, 229e230 solvent mixtures, global polarity for, 231 Empirical optimization, of preparative liquid chromatography, 415e416 Entropy-controlled process, of SEC retention, 197e199 Equilibration time, 65 Equilibrium-dispersive model (EDM), 409e410, 430 Equilibrium process, of SEC retention, 199e201 Evaporative light-scattering detectors (ELSD), 12e13, 301 Excess Rayleigh ratio, 211e212 Extracolumn band-broadening (ECBB) effect, 284e285, 294e295 Extracolumn effects, 204e205 Extract-enrichment SMB (EE-SMB) chromatography, 443e444 Extrathermodynamic approach, 386

F

Field-flow fractionation, 218 First-passage time, 34 Fittings, 294e296 Fluorescence detectors, 302e303 Fourier-transform mass spectrometry (FTMS), 320

496

INDEX

Fraction-feedback SMB (FF-SMB) chromatography, 443e444 Frontal analysis, 414 Fundamental resolution equation, 205e206

G

Gas chromatography (GC), 3 Gel filtration chromatography (GFC), 195e196 Gel permeation chromatography (GPC), 3, 194, 196 General rate model (GRM), 24e28, 430 for core-shell particles, 27e28 moment analysis, 28 of monolith columns, 26e27 Genetic algorithms and multiple linear regression (GAeMLR), 389 Giddings plate-height equation, 34e36, See also Plate-height equation Global polarity estimators, 229e230 for solvent mixtures, 231 Gradient conditions, effect on separation, 274 Gradient elution, 5, 258 linear-solvent-strength model of, 274 mixing systems, 289e291 systematic trial-and-error mobile-phase optimization for, 244e245 use of, 271 Gradient elution liquid chromatography, 433e435 method development, 278e281 problems associated with, 281e282 separation, experimental conditions effect on, 272e278 column conditions effect on, 275e277 gradient conditions effect on, 274 gradient separation versus isocratic, 272e274 irregular sample, 277e278 theory and practice of, 269e282 Gradient polymer elution chromatography (GPEC), 218 Gradient separation versus isocratic separation, 272e274 Grafting, See Bonding

H

Hardware, column, 46e48, 47f Height equivalent of a theoretical plate (HETP), 43

Hetaeric chromatography, 91e92 1-Hexyl-3-tetrafluoroborate, 93f High-performance liquid chromatography (HPLC), 284f at the beginning, 3e4 column development, 7t columns, 6e9 column packing, 6e9 particles, 6e9 stationary phases and selectivity, 9 detectors, 12e13 equipment, 9e12 history of, 44e46, 45t method development for, 260e261, 271 with NMR, hyphenation of, 351e353 silica-based columns, proper storage of, 72 stationary phases, characterization of, 395 systems instrumentation for, 462e465 miniaturization of, 458e459 theory and practice, 4e6 new modes and techniques, 5 selectivity control, condition selection for, 5e6 High-temperature superconducting coils, 357e358 Hildebrand solubility parameter, 229e230 Hindered diffusion, defined, 49 Hofmeister series, 96, 124t Holdup volume, 418 Hybrid micellar liquid chromatography, 99e100 Hydrodynamic chromatography, 218 Hydrophilic interaction liquid chromatography (HILIC), 5, 105e120, 162e163, 226e228 applications of, 114e116 columns in, 229 mobile phases in, 110 principles of, 105e107 solvent selection in, 229 stationary phases in, 110e114, 111f chemically bonded phases, 113 ion exchange stationary phase, 113e114 silica gel, 110e113 Zwitterionic stationary phase, 113e114 Hydrophobic-adsorption chromatography, 123 Hydrophobic-affinity chromatography, 123

INDEX

Hydrophobic chromatography with dynamically coated stationary phases, 91e92 Hydrophobic interaction chromatography (HIC), 121e141 applications of, 132e135, 133t, 135t factors affecting, 125e130 adsorbents, properties of, 126e128 biomolecules, properties of, 130 mobile phase, 128e129 temperature, 129 future trends of, 135e136 principles of, 122e125 adsorption principles, 123e125 retention mechanism, 123e125 purification practical aspects of, 131e132 strategies, 130e131 Hydrophobic interactions, 122e123 Hydrophobicity, 67 8eHydroxyquinoline, 101 Hyphenation capillary electrochromatography, 476e478 direct LCeNMR, 359e363 indirect LCeNMR, 363e364 of NMR with HPLC, 351e353 off-line, 337e339 on-line, 339e345

I

Iminodiacetic acid (IDA), 161e162 Immobilized artificial membrane (IAM), 398 Immobilized-biomacromolecule stationary phases, QSRR retention data analysis on, 399e400 Indirect conductivity, 180 Indirect LCeNMR hyphenation, 363e364 Injection profiles, 419e420 Inorganic anions with surfactant-coated stationary phases, separation of, 94e95 Intermediate precision, 264 Intraassay precision, See Repeatability Inverse method, 414e415 Ion chromatography, 157e192 applications of, 185e186, 187t environmental, 186 industrial, 185e186 chelation, 161e162 columns, 165e175 anion-exchange columns, 170e171

497

cation-exchange columns, 171e175 defined, 157e158 electrochemical detection in, 179e182 amperometry, 182 nonsuppressed conductivity, 179e180 suppressed conductivity, 180e181 electrostatic, 162e163 eluent converters, 175e179 eluent generators, 175e179 eluents for, 164e165, 166t anion exchange, 164e165 cation exchange, 165 history of, 158 principles of, 158e165 spectroscopic detection in, 182e183 mass spectrometry, 183e184 photometric detection, 182e183 postcolumn-reaction detection, 183 Zwitterionic, 162e163 Ion evaporation mechanism (IEM), 311e312 Ion-exchange chromatography (IEC), 3 capacity of, 167e168 principles of, 158e160 Ion exchange packings, regeneration of, 73 Ion-exchanger(s) agglomerated, 169e170 dynamically modified, 170 with grafted functional groups, 169 with ionogenic polymer grafted layers, 169 matrix, 168e169 polymer-coated, 169 polymer-based, 73 Ion exchange stationary phase, 113e114 Ion-exclusion chromatography (IEC) ion exchangers, applications of, 178t principles of, 160e161 Ionic liquids, as additives, 96e97 Ion-interaction chromatography (IIC), 90e97 additives, 96e97 inorganic anions with surfactant-coated stationary phases, separation of, 94e95 operational modes, 92e94 permanent coating, 91e92 reagents, 92e94 retention mechanism, 90e92 silanol, effect of, 95e96 silanol, suppression with amine compounds, 95e96 Ion-modified chromatography, 91e92 Ion-pair chromatography (IPC), 5, 91e92

498

INDEX

Ion-suppression chromatography, 89 Isocratic discontinuous elution chromatography, 429e432 mathematical modeling of, 429e432 typical effects of, 429e432 Isocratic elution, systematic trial-and-error mobile-phase optimization for, 242e244 Isocratic separation versus gradient separation, 272e274 Isoeluotropic mixtures, 233e234

K

Kinetic theories, of liquid chromatography, 19e40 macroscopic, 20e30 equivalence of, 29 general rate model, 24e28 lumped kinetic model, 22 lumped-pore diffusion model, 28e29 of nonlinear chromatography, 29e30 macroscopic versus microscopic, 38e39 microscopic, 30e37 comparison with macroscopic kinetic theories, 38e39 Giddings plate-height equation, 34e36 nonlinear chromatography, Monte Carlo simulations of, 36e37 stochastic model, 31e34 Kozeny-Carman equation, 55

L

Lab-on-a-chip (LOC), 325 Laboratory notebook, 265 Lambert-Beer law, 298, 335 Langmuir adsorption isotherm, 413e414 Laser-induced fluorescence (LIF) detector, 476 LCeNMReMS, 368 Ligand(s), 59, 60t accessibility, 60t density, 60t, 127e128 distribution, 60t homogeneity, 60t stability, 60t topography, 60t type, 127 uniformity, 60t utility, 60t Linear free-energy relationships (LFER), 386 Linear salvation energy relationships (LSER), 388, 390

Linear-solvent-strength (LSS) model, 5, 274, 390 Liquid adsorption chromatography at the critical condition (LACCC), 218 Liquid chromatography (LC) column packing, 6e9 column technology in, 41e86 coupling, ionization methods for, 309e317 development of, 1e18 before 1960, 2e3 functioning of, 42e43, 43f ideal detectors for, characteristics of, 308t kinetic theories of, 19e40 particles, 6e9 Liquid chromatographyeinfrared (LCeIR), 333e348 off-line hyphenation, 337e339 on-line hyphenation, 339e345 Liquid chromatographyemass spectrometry (LC-MS) electron ionization in, 316e317 micro-, 321e325 nano-, 321e325 Liquid chromatographyeRaman (LCeRaman), 333e348 off-line hyphenation, 337e339 on-line hyphenation, 339e345 Liquidesolid chromatography, See Normalphase chromatography Loadability, column, 66e67 Loading capacity, 415e416 Localization, 147e148 solute, 147e148 solvent, 147e148 Low-angle static light-scattering (LALS) detector, 212e213 Lumped kinetic model, 22 Lumped-pore diffusion model, 28e29

M

Macroscopic kinetic theories, 20e30 comparison with microscopic kinetic theories, 38e39 equivalence of, 29 general rate model, 24e28 for core-shell particles, 27e28 moment analysis, 28 of monolith columns, 26e27 lumped kinetic model, 22 lumped-pore diffusion model, 28e29 of nonlinear chromatography, 29e30

INDEX

Magnetic field, NMR spectroscopy, 353e354 Mass-balance equation, 20e21 Mass spectrometry, 183e184, 307e332 data specificity, 317e320 accurate mass measurements, 318e319 MS/MS, 319e320 interfacing with capillary electrochromatography, 326e327 Mechanical stability, in silica-based columns, 71 Metabolomics, metabolite identification in, 376 Metal complexation, 100e103 metal ions, determination of, 100e102 organic compounds, determination of, 102e103 Metal content, 69e70 Metal ions, determination of, 100e102 Methanol (MeOH), 46, 68, 92e94, 110, 153, 228, 277, 391e395, 471e473, 479 Method development, in liquid chromatography, 251e268 documentation for, 265e266 goals of, 253 gradient elution liquid chromatography, 278e281 in practice, 258e263 column screening, priority of, 259e260 HPLC vs. UHPLC, 260e261 resolution-modeling software, 258e259 systematic plan, 261e263 prevalidation of, 263e264 structured approach to, 253e258 column plate number, 254 gradient elution, 258 retention factor, 254e255 selectivity, 255e258 validation of, 264e265 Method document, 264e265 Method optimization, in capillary electrochromatography, 478e479 Micellar electrokinetic chromatography (MEKC), xenobiotics lipophilicity assessment using, 398 Micellar liquid chromatography (MLC), 97e100 mobile phase, secondary equilibrium in, 97e99 xenobiotics lipophilicity assessment using, 398e399 Micelles, 97

499

Micro-emulsion electrokinetic chromatography (MEEKC), xenobiotics lipophilicity assessment using, 398 Microfabrication technologies, 455e458 Microfluidic chip, 454e455 Microfluidic devices, 323e325 Microfluidic flow control (MFC), 463 Microfluidics, 453e468 applications of, 456te457t defined, 454e455 instrumentation for, 459e465 electrophoretic systems, 461e462 HPLC systems, 462e465 systems for separations, 455e459 HPLC systems, miniaturization of, 458e459 microfabrication technologies, 455e458 Microfractionation, NMR spectroscopy, 364e367 Micro-LCeMS, 321e325 classical approach, 322e323 microfluidic devices, 323e325 Microscopic kinetic theories, 30e37 comparison with macroscopic kinetic theories, 38e39 Giddings plate-height equation, 34e36 nonlinear chromatography, Monte Carlo simulations of, 36e37 stochastic model, 31e34 Micro-total analysis system (m-TAS), 454e455, 487 Miniaturization, 453e468 column, 48e49, 74 of HPLC systems, 458e459 Mobile phase(s) affecting hydrophobic interaction chromatography, 128e129 elution additives, 129 pH, 129 salt, type and concentration of, 128 composition optimization, practical guidelines for, 241e246 chromatographic mode, selection of, 241e242 computer-assisted interpretive optimization, 245e246 gradient elution, systematic trial-anderror mobile-phase optimization for, 244e245

500

INDEX

Mobile phase(s) (Continued) isocratic elution, systematic trial-anderror mobile-phase optimization for, 242e244 defined, 49 in hydrophilic interaction liquid chromatography, 110 secondary equilibrium in, 97e99 selection of, 152 in silica-based columns, 71e72 Modeling additives, 420e421 Molar mass distribution (MMD), 194, 217e218 Moment analysis, 28 Monolith columns, general rate model of, 26e27 Monte Carlo simulations of nonlinear chromatography, 36e37 MS/MS, 319e320 Multiangle static light-scattering (MALS) detector, 213e218, 214f Multiple regression analysis (MRA), 390 Multivariate curve resolution with alternating least squares (MCReALS), 340 Multivariate data analysis, using QSRR relationships, 400e401

N

Nanofluidics, defined, 454e455 Nano-LCeMS, 321e325 classical approach, 322e323 microfluidic devices, 323e325 Nanosystems, 454e455 Natural product identification, using LCeNMR spectroscopy, 371 NextGen sequencers, 454 Nonlinear chromatography kinetic theory of, 29e30 Monte Carlo simulations of, 36e37 Nonsuppressed conductivity, 179e180 Normal-phase liquid chromatography, 143e156, 226e228 columns in, 228e229 method development, 149e154 example of, 152e154 mobile phase, selection of, 152 problems in use of, 154e155 retention in, 144e149 selectivity, 148e149 separation of mixtures by, 144e149 solute localization, 147e148

solvent localization, 147e148 solvent selection in, 228e229 Normal phase packings, regeneration of, 72 Nuclear magnetic resonance (NMR) spectroscopy, 349e384 fields of application, 370e372 body fluids, metabolite identification in, 376e377 metabolomics, metabolite identification in, 376 natural product identification, 371 pharmaceutical impurities, identification of, 377 unstable compounds, analysis of, 376 with HPLC, hyphenation of, 351e353 information from LC peak, obtaining strategies for, 359e368 at-line micro-NMR analysis, 364e367 direct LCeNMR hyphenation, 359e363 indirect LCeNMR hyphenation, 363e364 microfractionation, 364e367 microgram sample amounts detection, practical considerations for, 367e368 integration with multiple detection system, 368 quantification capabilities of, 368e370 at-line LCeNMR, 369e370 on-line LCeNMR, 369e370 qNMR, general considerations on, 369 sensitivity, advances in, 353e359 cryogenic probes, 357 high-temperature superconducting coils, 357e358 magnetic field, 353e354 probe design, 354 samples, amount of, 358e359 size, 354e357 Numerical optimization, of preparative liquid chromatography, 416e418 full, 418 general procedures, 416e417 injection volume, 417e418

O

Occluded phase, 160e161 Off-line hyphenation, 337e339 On-line hyphenation, 339e345 On-line LCeNMR, quantitative capabilities of, 369e370 On-line viscometry, 208e210

INDEX

Optimization of preparative liquid chromatography, 415e421 empirical, 415e416 numerical, 416e418 Optimum-flow conditions, choice of, 66 Orbitrap, 320 Ordered mesoporous silica vs. disordered silica, 76e77 Organic compounds, determination of, 102e103 Orthogonal power, 255e258

P

Packing defined, 49 materials, 51e63 procedures, for analytical columns, 61e63 Pairedeion chromatography, 91e92 Parallel factor analysis (PARAFAC), 340 Partial-differential equations (PDEs), 430e431 Partial feed, 445 Partial least square (PLS) method, 390 Particle(s), 6e9 shape, 54e56 size, 54e56 Particle size distribution, 54e56 Perfluorinated carboxylate anions, as additives, 96e97 Perfusion, 58 Perfusive particles, 58 Permanent coating ion-interaction chromatography, 91e92 pH affecting protein retention in HIC, 129 buffers and measurement of, 89e90 changes in retention with, 89 stability, in silica-based columns, 70e71 Pharmaceutical impurities, identification of, 377 1,10-Phenanthroline, 101 Photometric detection, 182e183 Physical detectors, 217 Plate-height equation, 24e26, 28 Giddings, 34e36 van Deemter, 24, 36 Point of care, 454 Polar selectivity, 68 Polycyclic aromatic hydrocarbons (PAH), 389e390 Poly ether ether ketone (PEEK), 46, 48, 50e51

501

Polymer-based columns, maintenance and troubleshooting of, 73e74 general considerations, 73e74 Polymer-coated ion exchangers, 169 Polymer materials, regeneration of, 73e74 Polyoxyethylene-(23)-dodecyl ether (Brij-35), 99 Polystyrene-divinylbenzene (PS-DVB), hydrophobic unmodified, 73 Pore connectivity, 57 Pore structure parameters, 57e58 Postcolumn-reaction (PCR) detection, 183 Power feed concept, 442e443 Precision defined, 319 intermediate, 264 intraassay, See Repeatability Precursor ion, 319 Preparative liquid chromatography, 427e452 continuous concepts of, 437e446 annular chromatography, 445e446 improved operating concepts, 441e445 simulated moving-bed chromatography, 437e445 gradient elution chromatography, 433e435 isocratic discontinuous elution chromatography, 429e432 mathematical modeling of, 429e432 typical effects of, 429e432 modeling of, 407e426 adsorption, See Adsorption model case example, 422 column model, 409e410 empirical, 415e416 numerical, 416e418 operational conditions of, 418e421 optimization of, 415e421 optimization of, 446e448 recycling chromatography, 435e437 closed-loop, 436 steady-state, 436e437 Presaturation, 361e362 Principal-component analysis (PCA), 341e343, 400e401 Probe(s) cryogenic, 357 design, NMR spectroscopy, 354 Product ions, 319 Proteomics, QSRR applications in, 392e393 Pseudophase liquid chromatography, 97

502

INDEX

Pulsed amperometric detection (PAD), 182 Pumping systems, 286e289

Q

Quality by design (QbD), 252, 263, 265 Quantitative NMR (qNMR) spectroscopy, general considerations on, 369 Quantitative structure property (retention) relationships (QSP(R)R), 385e406 goals of, 386e392, 387f methodology of, 386e392, 387f retention prediction, 389e392 structural descriptors, 387e389 Quantitative structure-retention relationships (QSRR) biological activities of, 400 immobilized-biomacromolecule stationary phases, retention data analysis on, 399e400 multivariate data analysis using, 400e401 in proteomics, applications of, 392e393 stationary phases, characterization of, 393e395 xenobiotics lipophilicity assessment using, 395e399 Quantum-cascade lasers (QCLs), 335 Quasi-elastic light-scattering (QELS), 217

R

Rayleigh-Gans-Debye approximation, 211e212 Reactionedispersive model, 22 Recycling chromatography, 435e437 closed-loop, 436 steady-state, 436e437 Refractive-index (RI) detectors, 12e13, 300 Regeneration of columns, 72e73 ion exchange packings, 73 normal phase packings, 72 RP packings, 72 Repeatability, 264 Reproducibility, 264 Resolution maps, 258e259 Resolution-modeling software, 258e259 Retention factor, 205e206, 254e255 Retention mechanism, 90e92 in hydrophobic interaction chromatography, 123e125 on immobilized-biomacromolecule stationary phases, data analysis of, 399e400

in normal-phase chromatography, 144e149 in quantitative structure property (retention) relationships, 389e392 of size-exclusion chromatography, 197e201 entropy-controlled process, 197e199 equilibrium process, 199e201 size-exclusion process, 197 Reversed-phase liquid chromatography (RPLC), 5, 134e135, 226e228 acidebase equilibria, 89e90 pH, buffers and measurement of, 89e90 pH, changes in retention with, 89 columns in, 228 ion-interaction chromatography, 90e97 additives, 96e97 inorganic anions with surfactantcoated stationary phases, separation of, 94e95 operational modes, 92e94 reagents, 92e94 retention mechanism, 90e92 silanol, effect of, 95e96 silanol, suppression with amine compounds, 95e96 metal complexation, 100e103 metal ions, determination of, 100e102 organic compounds, determination of, 102e103 micellar liquid chromatography, 97e100 hybrid micellar liquid chromatography, 99e100 mobile phase, secondary equilibrium in, 97e99 secondary chemical equilibria in, 87e104 solvent selection in, 228 Robustness, 265

S

Salting-out chromatography, 123 Salt-promoted adsorption, 123 Sample injection, 291e293 Schoenmakers’s rule, 231e233 Science-based calibration (SBC), 341e343 Selective withdrawal, 445 Selectivity, 148e149, 255e258 experimental conditions affecting, 277e278 shape, 68e69 Separation corresponding, 274

INDEX

experimental conditions effect on, 272e278 column conditions effect on, 275e277 gradient conditions effect on, 274 gradient separation versus isocratic, 272e274 irregular sample, 277e278 factor, 205e206 microfluidic systems for, 455e459 HPLC systems, miniaturization of, 458e459 microfabrication technologies, 455e458 Shape selectivity, 68e69 Silanol effect of, 95e96 suppression with amine compounds, 95e96 Silanophilic activity, 67e68 Silica -based columns, maintenance and troubleshooting of, 70e73 general guidelines, 70e72 HPLC columns, proper storage of, 72 regeneration of, 72e73 bonded silica, physicochemical characterization of, 61 gel, 110e113 major synthesis routes, 59e61 ordered mesoporous vs. disordered, 76e77 surface functionalization of, 58e61 Simple-to-use self-modeling algorithm (SIMPLISMA), 341e343 Simulated moving-bed (SMB) chromatography, 437e445 extract-enrichment, 443e444 fraction-feedback, 443e444 multicomponent separations by, 444e445 Size-exclusion chromatography (SEC), 12e13, 193e224 absolute molecular mass, determination of, 207e217 static light-scattering detection, 211e217 universal calibration, 208, 211 band broadening in, 202e205 extracolumn effects, 204e205 historical background of, 194e197 multidetector separations, 217e219 physicochemical characterization, 217e219

503

resolution in, 205e206 retention in, 197e201 entropy-controlled process, 197e199 equilibrium process, 199e201 size-exclusion process, 197 two-dimensional techniques, 217e219 Slurry packing method, 474 Snyder’s global polarity, 236 Snyder’s solvent-selectivity triangle, 234e239 Sodium dodecyl sulfate (SDS), 93f, 99 Soft-pulse multiple irradiation, 361e362 Solegel process, 474e475 Solidefilm linear driving-force model, 22e23 Solute localization, 147e148 Solvatochromically based solvent selectivity triangle, 240e241 Solvent(s) demixing, 155 properties of, 227te228t localization, 147e148 mixtures, global polarity for, 231 source, 286 Solvent selection, in liquid chromatography, 225e250 additional considerations for, 246e248 elution strength, 226e228 assessment of, 229e231 isoeluotropic mixtures, 233e234 mobile-phase composition optimization, practical guidelines for, 241e246 chromatographic mode, selection of, 241e242 computer-assisted interpretive optimization, 245e246 gradient elution, systematic trial-anderror mobile-phase optimization for, 244e245 isocratic elution, systematic trial-anderror mobile-phase optimization for, 242e244 Schoenmakers’s rule, 231e233 solvent-selectivity triangle, 234e241 Snyder’s, 234e239 solvatochromically based, 240e241 solvent mixtures character, prediction of, 239 Solvent-selectivity triangle (SST), 234e241 Snyder’s, 234e239 solvatochromically based, 240e241

504

INDEX

Solvent-selectivity triangle (SST) (Continued) solvent mixtures character, prediction of, 239 Specific resolution, 206 Spectroscopic detection, in ion chromatography, 182e183 mass spectrometry, 183e184 photometric detection, 182e183 postcolumn-reaction detection, 183 Static light-scattering (SLS) detection, 211e217 low-angle, 212e213 multiangle, 213e218, 214f Stationary phase(s) in capillary electrochromatography, 473e476 chiral, 408, 479 defined, 49 in hydrophilic interaction liquid chromatography, 110e114, 111f chemically bonded phases, 113 ion exchange stationary phase, 113e114 silica gel, 110e113 Zwitterionic stationary phase, 113e114 mass transfer, 202 in quantitative structure-retention relationships, 393e395 immobilized-biomacromolecule stationary phases, retention data analysis on, 399e400 sterically protected, 9 Steady-state recycling (SSR) chromatography, 436e437 Stochastic model, 31e34 stochasticedispersive model, 33e34 Stripper column, 180 Suppressed conductivity, 180e181 Surfactant chromatography, 91e92 Systematic plan, for method development, 261e263 System suitability tests, 265

T

Tag-along effect, 431e432 Temperature affecting hydrophobic interaction chromatography, 129 column, 66 Tetrahydrofuran (THF), 228, 479

Thin-layer chromatography (TLC), 151e152 Time-of-flight (TOF), 320 Transportedispersive model, 22 Triethylamine, 93f Trifluoroacetic acid (TFA), 96 True moving bed (TMB) chromatography, 437e438 Tubings, 294e296 Two-dimensional liquid chromatography, 217e219

U

Ultrahigh-performance liquid chromatography (UHPLC), 12e13, 76e80 method development for, 260e261 Ultravioletevisible (UV-Vis) absorbance detectors, 298e300 Uninformative variable elimination partial least squares (UVEePLS), 389 Universal calibration, for size-exclusion chromatography, 208, 211 Unstable compounds, analysis of, using LCeNMR spectroscopy, 376

V

Validation documentation for, 265 of method development, 264e265 protocol, 264 van Deemter plate-height equation, 24, 36, See also Plate-height equation Varicol, 442e443 Vibrational spectroscopy (VS), 333e337

W

Water-suppression enhanced through T1 effects (WET), 361e362 Wheatstone bridgeetype differential viscometer, 210f

X

Xenobiotics lipophilicity assessment, using quantitative structure-retention relationships, 395e399

Z

Zwitterionic ion chromatography (ZIC) principles of, 162e163 Zwitterionic stationary phase, 113e114