Investment Casting 9781906540579, 9781003419228, 1906540578

Table of contents :
Cover
Half Title
Title
Copyright
Contents
Editors and Authors
Foreword
Acknowledgements
1. Introduction
2. Tooling
3. Pattern Technology
4. Investment Materials and Ceramic Shell Manufacture
5. Melting and Casting
6. Gating and Feeding Investment Castings
7. Finishing Investment Castings
8. Health, Safety and Environmental Legislation
9. Defects and Non-Destructive Testing
10. Metallurgical Aspects: Structure Control
11. Design for Investment Casting
12. Review of Applications
12.1 Application to Aerospace
12.2 General Applications of Investment Castings
12.3 Jewellery Investment Casting
12.4 Investment Casting in Surgery and Dentistry
Index

Citation preview

INVESTMENT CASTING

INVESTMENT CASTING EDITED BY

Peter R. Beeley and

Robert F. Smart

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

Book 511 First published in 199 5 by The Institute of Materials This paperback edition first published in 2008 by Maney Publishing Published 2021 by CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 1995 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works ISBN 13: 978-1-906540-57-9 (pbk) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please accesswww.copyright.com(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com Statements in this book reflect those of the authors and not those of the Institute or publisher Typeset in the UK by Dorwyn Ltd, Rowlands Castle. Hampshire

DOI: 10.1201/9781003419228

Contents

Editors and Authors...........................................................................................

vii

Foreword................................................................................................................

ix

Acknowledgements...............................................................................................

xi

1. Introduction...............................................................................................

1

2. T o o lin g ........................................................................................................

30

3. Pattern Technology...................................................................................

43

4. Investment Materials and Ceramic Shell Manufacture...................

65

5. Melting and C asting................................................................................. 123 6. Gating and Feeding Investment Castings............................................. 150 7. Finishing Investment C astin g s............................................................. 183 8. Health, Safety and Environmental Legislation.....................................212 9. Defects and Non-Destructive T estin g ................................................... 240 10. Metallurgical Aspects: Structure C o n tro l............................................. 293 11. Design for Investment C asting ................................................................ 334 12. Review of Applications...............................................................................373 12.1 Application to Aerospace..........................................................354 12.2 General Applications of Investment Castings......................392 12.3 Jewellery Investment Casting................................................... 408 12.4 Investment Casting in Surgery and Dentistry......................441 In d ex ....................................................................................................................... 474

Editors and Authors

Peter R. Be e l e y DMet, PhD, CEng, FIM, FIBF. Life Fellow and formerly Senior Lecturer in Metallurgy, University of Leeds, UK. Robe rt F. Smart BSc , PhD, CEng, FIM. Director, British Investment Casting Trade Association and Secretary, European Investment Casters' Federation. Geo f f re y Be l l MIBF. Managing Director of A W Bell Australia PTY Ltd; Member of Investment Casting Institute and Past President of Investment Casters' Association of Australia. H e n r y T. Bid w e ll MIM, CEng. Executive Director, Investment Casting Institute, USA and President, Investment Casting Resource International. CEng, MIM. EA Technology, UK.

M ike Bo n d

Da v id B. C ritc hl ey NFCDipl, DMS. Technical Officer, British Investment Casting Trade Association. Did a r Sin g h Du l a y BSc , FInstNDT. Managing Director, NDT Consultants Ltd and Member of Technical Committee, British Institute of Nondestructive Testing. P ete r Ga ins bu ry CEng, FIM. Former Director, Design and Technology, The Worshipful Company of Goldsmiths, UK. PhD, CChem, MRSC. E r ic F. H art mann OEH Scientific Ltd, Aston Science Park, Birmingham, UK.

viii

Investment Casting

Philip Johnson CChem, MRSC. OEH Scientific Ltd, Aston Science Park, Birmingham, UK. Mau ri ce F. L ec ler c BSc , PhD, CEng, MIM, MIQA. Director of Regulatory Affairs, Quality Assurance and Engineering, Vida Med International Ltd, UK. Da v id Mil ls BSc , CEng, MIM. Manager, Manufacturing Technology (Foundry Ceramics) Rolls Royce PLC, UK. Thomas S. P iwo nk a Sc D. Director, Metal Casting Technology Center, The University of Alabama, USA. Ro n a l d W ill ia ms LRCS, MIMgt. Managing Director, Blayson Olefines Ltd, UK.

Foreword

The concept of a modern book on investment casting originated in the work of the Books Committee of the Institute of Materials, which had identified a substantial gap in the literature of metal founding. The investment casting sector of the foundry industry has seen rapid growth, exemplified in the United Kingdom, where the financial turnover has reached a level well over £250 m per annum and is surpassed only in the USA. Despite this, the literature devoted specifically to the process and its products has remained relatively sparse, even though the industry itself has ready access to the proceedings of conferences organised through its own collaborative bodies, and the subject is treated to a limited extent in more general works on metal casting. A brief survey as undertaken in the Introduction portrays a process which is clearly of the distinction and importance to merit a separate and comprehensive treatment. In the production of the book, the aim has been to draw upon the knowledge of authorities within or closely associated with the industry, facilitated by co-operation with the British Investment Casting Trade Association, and to examine the process and its products in a way which will be useful both to the industry itself and to engineers involved with the selection, design and use of investment castings. To this end the earlier chapters are devoted to each of the main production stages from tooling to finishing, with a separate treatment of health, safety and environmen tal issues, commensurate with the importance now given to this topic. Subsequent chapters are concerned mainly with the quality and characteristics of investment castings, including considerations of defects and of methods of inspection and testing. Metallurgical characteristics are reviewed against a background of the basic phenomena of solidification and subsequent treatment and their effects on structure and properties, including the techniques used to develop these to the best advantage in major groups of alloys. Design aspects of investment castings are also examined, with guidance to alloy selection and to capabilities and

x Investment Casting limitations in respect of shape and dimensions. In these and other cases, recommendations are given to sources of further information where this is felt to be useful. The last chapter of the book, arranged in four parts, brings together many examples of applications in a variety of fields. The aerospace section describes the progressive evolution of gas turbine rotor blade castings, based on the combination of sophisticated alloy developments with enhanced capability of the casting process. The major expansion into the broader engineering arena is then demonstrated, in a section containing a wide range of illustrated examples, whilst the long-established and im portant presences of the process in the jewellery - art and medical-dental fields provide the substance for two further accounts, which include details not only of the applications themselves but of the specialised production techniques and equipment associated with them. In any multi-author work there are inevitable differences of style, structure and scope of treatment, and the present book is no exception. The Editors have nevertheless endeavoured to achieve full and effective coordination of the contributions and are grateful to the individual authors for their collaboration in this respect. Apart from acknowledgements made elsewhere, they and the Institute are also grateful to all who have given encouragement and practical help in achieving production of the book. P.R.B. R.F.S.

Acknowledgements

The editors and authors are grateful for the provision of advice, data and illustrations from many sources; the illustrations are individually attributed where they appear. In respect of Chapter 12 thanks are due to Mr Donald Pratt of AE Turbine Components Ltd for helpful comments on the original draft text of Part 1, and to Dr David Driver for providing accompanying photographs. The author of Part 4 acknowledges valuable help from casting producers and specialist practitioners in the surgical and dental fields as follows: Tim Band (Precision Cast Parts Ltd - Sheffield) Ken Brummitt (De Puy International Ltd - Leeds) Andy Crosbie (Department of Health - Supplies Technology Division London) Don McKenna (McKenna Precision Castings Ltd - Rotherham) Fred Norris (Howmet Turbine Components Comp - Whitehall Mich. USA) Rex Palmer (Truecast Ltd - Ryde, Isle of Wight) Brian Penn (Howmedica International Inc - Limerick, Ireland) Prof. John Scales OBE (Mount Vernon Hospital - Middlesex) Phil Whateley (Deritend Precision Castings Ltd - Droitwich Spa) Marion Broomes (British Standards Institution - London) Keith Day (Biomet Ltd - Bridgend, South Glamorgan) Derek Johnson (Yeovil Precision Castings Ltd - Yeovil) Peter D. Gordon LDS, RCS, Dental Surgeon (Upper Wimpole Street London) George Ashton (Ashton Dental Laboratories, Boston Place - London) Ian Waterhouse (De Puy International - Leeds)

1 Introduction P.R. BEELEY and R.F. SMART

The process of investment casting has come to occupy a key position in the range of modern metal casting techniques. Over the half-century dating from 1940, what had been a small and highly specialised sector of casting activity developed into a worldwide and distinctive industry, reflecting the importance of the product in the intensifying search for close accuracy of shape and dimensions in materials forming. The nearnet-shape objective is seen, not only as a means of providing the engineer with a direct, efficient and economical route to the manufacture of a finished component, but also as a contribution to the conservation of costly materials and energy. The term investment casting derives from the characteristic use of mobile ceramic slurries, or 'investments', to form moulds with extremely smooth surfaces. These are replicated from precise patterns and transmitted in turn to the castings. Although certain variants employ permanent patterns and multi-part moulds analogous to those used in sand casting, investment casting has become closely identified with the expendable pattern principle typified in the long-established lost wax process. In brief, disposable replicas of the required casting are formed by injecting molten wax into a die with the appropriately shaped cavity. The wax patterns are connected, singly or in groups, to a wax sprue and gating system and the whole is clothed in investment slurry. The wax is melted out and the investment consolidated by heating, leaving a hard ceramic mould to receive the molten metal. The mould is finally broken up to extract the solidified product. A special feature conferred by the use of expendable patterns is the one-piece mould; the absence of the partings normally required for pattern extraction eliminates a major source of errors arising from misalignment of separate mould parts on assembly. Smooth, hard, precise jointless moulds are the key to the product characteristics that have given DOI: 10.1201/9781003419228-1

2 Investment Casting investment casting its increasing importance in the wider world of metal manufacture.

CASTING PROCESSES AND THE CONCEPT OF PRECISION Casting has, through most of its long history, been primarily associated with sand moulding. Apart from early production of copper alloys, the development of a distinctive foundry industry also remained closely identified with cast iron as a metal, until the mid-nineteenth century brought the onset of diversification into the comprehensive modern range of cast alloys. Production of these too remained the almost exclusive preserve of sand casting, until the limitations of that versatile process testified to a need for more precise moulding techniques. Whilst the ad vent of die casting met some of the criteria for enhanced precision, this group of techniques embodied limitations of a different kind, most notably the restriction in the range of alloys compatible with metal moulds, and on the shapes capable of being produced and extracted from them at reasonable cost. The concept of precision can be portrayed as in Fig. 1, and is seen to embrace, not only the aspect of dimensional accuracy and tolerances, but also surface quality and capability to reproduce intricate cast detail; either of the latter can be the critical factor in the choice of a forming process for a particular application. All three attributes are significant at the interface with machining operations, affecting datum points and location in fixtures or, indeed, determining whether such operations are needed at all. The capacity of a casting process to meet these criteria is determined by the opportunity for departure from predicted behaviour during successive stages of production. A brief review of these stages will identify

Figl

Aspects of precision in casting.

Introduction

3

problems that need to be addressed in the search for greater precision, and will enable investment casting to be perceived against the broader background of other casting processes. Tooling and Equipment These are fundamental to the production of satisfactory castings, irrespective of process. Original shape or mounting errors in patterns, dies and coreboxes are automatically transferred to the product. Excessive clearance in coreprints, mould registers, box pins or dowels, whether original or from wear or distortion, produce mismatch of mould parts, with similar consequences. Any advance must address at least some of these poten tial sources of inaccuracy. A further aspect of the tooling is the incorporation of contraction allowances with their inherent uncertainties. This and similar aspects require clear understanding between the casting user and producer. The Mould Mould dimensions, besides being dependent on the foregoing tooling stage, can change when the mould leaves the pattern, whether through cavity enlargement, sagging, or volume change on drying or setting. Manual finishing of moulds, often traceable back to worn equipment or machinery, is clearly incompatible with precision. Mould surfaces are themselves imperfect to the extent of their particulate and pore structures, including larger voids from incomplete compaction. Modern approaches to the improvement of mould precision aim at the production of strong, dense moulds with fine surface textures, and mechanical support of the compacted material is wherever possible main tained by the pattern surface until the mould is strong enough to resist distortion. Casting Further shape imperfections can occur at the casting stage, including swelling under the influence of metallostatic pressure and roughness from metal penetration into surface pores and imperfections. Answers to these problems lie along the lines mentioned above. Other shape defects arise in the reverse direction, resulting from failure of the melt to conform to the full shape of the mould cavity before freezing. The result is seen as imperfect definition of corners and surface detail, one of the three param eters stressed in Fig. 1. Avoidance of these faults requires smooth and unimpeded metal flow with minimum friction. Smoother mould surfaces make an important contribution, but hot moulds and pressure or vacuum assisted mould filling are introduced in some processes, including investment casting, to further this aim.

4

Investment Casting

Solidification and Cooling Casting shape and dimensions undergo further changes during cooling in the mould. Normal contraction begins as soon as the cast component acquires enough cohesion to behave as a solid body and should, at least in theory, influence all dimensions by an amount predictable from the co efficient of expansion of the alloy. This is the basis for the contraction allowances incorporated in foundry patterns, which can exceed 2% in linear terms; precision of the cast product clearly depends on the reliability of this estimate. A freely contracting body will conform closely to expected behaviour, but under real casting conditions there can be significant resistance to contraction from two sources. The mould has sufficient compressive strength to induce high temperature plastic deformation in the cast metal, particularly where casting features enclose bodies of mould material, as in cored cavities and between flanges; contraction is then lower than might otherwise be expected. Thermal stresses too are generated within the metal itself; these result from differential contraction associated with local variations in cooling rate, as between thick and thin members or between surface and interior. Such hindrances to free contraction can in severe cases cause tears in the casting, but their effect is otherwise to reduce the theoretical contraction in affected members, mak ing prediction more uncertain. Clearly all casting processes, and indeed all other metal shaping pro cesses involving high temperature, are subject to this fundamental limitation to dimensional accuracy. Variations will depend on the design of the individual component, the arrangement of the particular dimension, and the production conditions. In these circumstances the advantage will lie with processes which can offer two characteristics, namely maximum consistency of manufacturing conditions, and the readiness with which tooling can be modified to take account of experience with prototypes or early production runs. Finishing The precision of a cast component is obviously influenced by the nature of the cleaning and finishing operations, including cutting and surface dressing. A precise process must minimize the need for interference with the original cast skin, giving the maximum advantage to processes generating smooth and clean surfaces in the first place, as the casting solidifies in the mould. The concept of precision must, naturally, be viewed in conjunction with other quality attributes of castings, for example low incidence of nonmetallic inclusions, oxide films, porosity and cracks. This represents a further factor in the choice of casting process, in which all the technical

Introduction

5

considerations must in the end be balanced against the acceptable cost of the finished component.

THE RANGE OF CASTING PROCESSES Precision is a comparative term. The modern era of casting manufacture has seen the emergence of new process developments which have con tributed to a broad advance in the quality and precision of cast products, including those from long established processes. Some have involved the mould and its manufacture, others have been concerned with the method of introducing the metal into the mould cavity, and yet others are of a general character, suitable for application across the entire field: molten metal filtration is one example. It is not practicable to examine other casting processes in detail in a work concerned primarily with investment casting, but it will be useful to summarize the characteristics of the established groups of techniques for the production of shaped castings, and to refer to some of the notable advances of recent years within them. The major process groups can be largely identified with the three distinctive routes from tooling to casting sketched in Fig. 2. Figure 2(a) shows the simple and direct die casting system. Sand casting and its many variants correspond to route (b), which is also, however, the basis for the production of investment castings from permanent patterns; the expendable pattern route as used in most investment castings production is represented in (c). Sand Casting The traditional process of sand casting, employing clay-bonded sand compacted around permanent patterns in moulding boxes and using oil bonded cores, has undergone dramatic changes. This situation has arisen with the advent of new systems of moulding material bonding and new types of machinery, both for sand preparation and for compacting and handling moulds and cores. Sand casting can now be seen as a family of processes, in many of which loose patterns and hand moulding have given way to techniques and equipment based on the modern toolroom. Greensand remains highly competitive in its own field, and the key to improved surface quality and accuracy of the products has been the greater stress on the achievement of rigid, high density moulds by using combinations of jolting, squeezing and blowing actions, in conjunction with well engineered boxes and patterns in integrated layouts. At the lighter end of the product range the outstanding modern development is the boxless high pressure automatic machine, in which dense block

6 Investment Casting

Cast (a)

Die

Cast (b)

Cast

o pattern Fig 2

Mould

Alternative routes from tool to casting.

mould parts are produced by blow -squeeze sequences at rates of several hundred per hour. The blocks, with pattern impressions on both vertical faces, are successively transferred to form a continuous strand for pouring (see Fig. 3). The use of both faces of each block provides a mould cavity, complete with its own gating and feeding system, per block. Chemical bonding systems for moulds and cores have replaced traditional binders in many sand casting applications. Various hardening pro cedures are used, including thermal curing in hot boxes and cold setting. In the latter case the hardening reaction is induced either by a liquid

Introduction

7

Blow fill

To pouring Fig 3 Principle of high pressure block moulding system. catalyst, applied by blending into the sand mix just before or during delivery into the moulding container, or by diffusing a reactive vapour or gas through the compacted unit. The original application of the latter principle is seen in the C 0 2 process, with sodium silicate used as the gashardened binder. A subsequent development was the phenol formaldehyde/isocyanate cold set system, in which final curing is in duced by an amine vapour catalyst. The use of inorganic and organic chemical reactions in these two developments has been paralleled in numerous other cold-set and cold box binder systems, using both liquid and gaseous catalysts. Taken as a whole, these materials and processes have grown in relative importance, given their clear advantage of operations at or near ambient temperature. They offer high standards of

8 Investment Casting mould part precision over a wide range of sizes, given rigid and accurate pattern equipment. Although moulding boxes remain as a mainstay of the sand casting process, the chemical bonding systems facilitate the production of high strength boxless block mould parts. These are increasingly used in the core-assembly mode, requiring moulded location features to replace the normal moulding box alignment system. The ultimate extension of the same principle is the shell mould, a sufficiently radical concept to be regarded as a separate process. Before further reference to shell mould casting, mention should be made of the unbonded sand system employed in the vacuum sealed moulding or V -process (see Fig. 4). This relies for compaction upon vaccum extraction of air from the sand body held in a box between two plastic films. The first of these is softened by radiant heat and suction formed on to the pattern plate, using a vacuum applied from below. The flask is filled with sand, vibrated, sealed with a backing film and evacu ated. The lower vacuum is then released, and the mould part lifted off the pattern; the shaped cavity in the compressed sand is retained by the upper vacuum until the mould has been closed and poured. The V -process requires a special layout with vacuum line connections, but its products have acquired a high reputation for surface finish and

Radiant heater

Plastic film 2

l QQQQQQQQQQOQOflQ^QQ^QQQQfl&gd

I

I

i

I 1 J

(b) Plastic film 2

Fig 4

Principle of V-process moulding.

Introduction

9

dimensional accuracy. They can also be produced with negligible pattern draft allowances. Shell moulding The Croning resin shell process introduced a radical new principle in mould making. It was the first to depart from the concept of a mould as a cavity within a solid block of material. The basic feature is the use of a moulding medium in which fine sand grains are coated with a solid synthetic resin. The action of heat on the resin produces initial softening, followed by thermal curing to form a strong, solid bond within a few minutes. The process sequence is illustrated in Fig. 5. The metal pattern plate, fitted with an ejector pin system, is heated at about 200 °C and a quantity of moulding material deposited on to its surface. As the heat penetrates the sand-resin mixture, a layer some 5-10 mm thick softens and adheres to the pattern plate when the container is turned over to dump the excess

(d)

Fig 5

Principle of resin shell moulding.

(e)

10 Investment Casting material. The assembly is further heated to accelerate the cure, forming a strong, smooth shell which can be pushed off the pattern by actuating the ejector pins. Core production follows an analogous principle. The shells are mutually aligned by moulded registers and glued or clamped together for casting. The products have a reputation for smooth surfaces and ready production of thin sections and intricate detail. The shell mould principle clearly offers major savings in materials consumption and handling and has since been adopted in other pro cesses. Shells can, for example, be formed in cold -setting sand mixtures by using contoured backs to follow the general shape of the pattern plate, and by employing similar hardening reagents to those used in block mould and core production. One such development employs a flexible diaphragm to form the back of the shell, followed by gas hardening through vents. The shell principle has also been adopted in the ceramic shell processes, both in normal investment casting and in the Replicast CS system. In these cases shells are formed by applications of slurries and solid grains to form coatings on expendable patterns. The patterns are eliminated by burnout and the shells hardened by firing. The ceramic shell system will be fully examined in later chapters. Shell moulds are often poured without further complication after closing, but in some cases require support in a backing medium of sand or shot to resist dilation under metallostatic pressure. Vacuum extraction of air from the granular surround can be used to increase this support and minimize fume on casting. A further important modern sand casting development is represented in the Cosworth process (Fig. 6), in which molten metal is gently raised by electro-magnetic pumping into a precise resin-bonded zircon sand mould placed above the metal reservoir. The counter-gravity metal transfer prin ciple offers many advantages and is also featured in some investment casting systems, besides having been long established in the low pressure die casting process described in the following section. Die Casting This group of processes is characterized by the most direct of all casting systems, in which molten metal is introduced directly into the permanent tooling (Fig. 2a). The processes generally involve high tooling costs and embody more shape restrictions than those based on refractory moulds, but offer high production rates and low costs for intricate parts in com patible alloys, especially of zinc, aluminium and magnesium. Within these limitations the capacity for precision is high. The main process variations concern the method of introduction of metal into the die cavity, as summarised in Fig. 7. In gravity die or

Introduction

11

Zircon sand mould

Large melter/holder furnace to allow sink or float of impurities Fig 6

Principle ofCosxuorth process. (Courtesy of Professor John Campbell).

permanent mould casting there are certain similarities with sand casting. Gating and feeding systems are embodied in vertically jointed dies analo gous to the sand blocks produced in the high pressure automatic mould ing machine. The dies, commonly of cast iron, are dressed periodically with a protective refractory coating. Metal cores are used where these can be retracted, but collapsible refractory cores are required for more com plex features. Mechanisms are often provided for die opening and closing and for casting ejection, to facilitate high rates of production. Low -pressure die casting has features in common with the gravity process but in this case the molten metal is displaced upwards into the die cavity by a modest air pressure applied to the space above the bath. The metal held in the tube or 'stalk' provides feeding during solidification of the casting, after which it is allowed to drain back into the reservoir. Smooth upward filling minimises turbulence and high casting yields are achieved. The process is mainly used for aluminium alloy production, although the principle has long been applied to the production of cast steel railway wheels in graphite moulds; its more recent introduction in the sand casting field has already been referred to. High pressure die casting is radically different from these processes, in that the metal is injected into the die cavity at high velocity, providing exceptional capacity for the rapid quantity production of thin walled,

Investment Casting

12

Gravity

Pressure

Low pressure

Fig 7

Die casting systems.

intricate components at low cost, although initial die costs are extremely high and metallurgical quality is reduced by the turbulent flow associated with die-fill times in the range 0.05-1 seconds, causing air bubbles to be retained within the solidified casting. Surface quality can nevertheless be very high and new approaches to the methods of casting are reducing levels of internal porosity.

Introduction

13

Die casting machines embody systems for metal injection and for die motions and locking. In the hot chamber machine, a reservoir of molten metal is maintained at the operating temperature, whilst successive strokes of a plunger in a submerged chamber force metal up an inclined tube into the die, the chamber being refilled from the reservoir with each return stroke. In the cold chamber machine, separate shots of metal are transferred manually or automatically from an external holding vessel into a shot sleeve, whence the piston forces the metal into the die. Carefully controlled injection pressure sequences are employed to optimise the pattern of metal flow and solidification. Pressure die casting dies, usually machined from steel, are of complex construction and need to be engineered to high standards. Apart from the main casting cavities and gates, whether for single or multiple casting, metal cores are incorporated to form holes and recesses. These and the ejector system are mechanically actuated as part of the die opening sequence. Water cooling passages are a further feature in some dies to regulate the temperature distribution during production. Die lives of 100,000 castings and more are feasible, and quantity production is required for the process to be economic; for alloys of higher melting point, die life is greatly reduced. High standards of precision can be achieved, with tolerances on some small dimensions closer than those obtainable from any other casting process. A further process loosely related to pressure die casting is squeeze casting, in which molten metal is poured into the hollow lower half of a twopart die, after which the upper half, in the form of a positive punch, is brought down to close the die and displace the liquid to fill the cavity; the process is especially suitable for cup-shaped components and gives products of high integrity. Not only cast alloys but metal matrix composites containing strengthening fibres can be squeeze-cast into simple components. Investment Casting Although investment casting is the subject of the main body of the book, a brief outline of its important features will be introduced here to complete the broad picture of the range of casting processes. The central and predominant process, that based on the expendable pattern principle as characterised in Fig. 1, is pictured in more detail in Fig. 8 and the production stages will now be briefly reviewed. Full treatments are featured in Chapters 2 -7. Tooling

The permanent tooling for a cast component takes the form of a die rather than a pattern as employed in sand casting. This is constructed in two or

14 Investment Casting

Casting Fig 8

Production sequence in investment casting using expendable patterns.

more pieces, with as many partings and inserts as are needed to permit extraction of the expendable patterns. Locators are embodied in the die parts and, given the low temperatures entailed in wax injection, close alignment in die assembly ensures precise and reproducible pattern dimensions. Choice of die material and method of manufacture depends on the quantity requirements and the nature of the product. Steel, brass,

Introduction

15

aluminium alloy, fusible alloys, polymers, plasters and rubbers are used. For the harder metals, dies are produced by conventional toolroom machining techniques, but direct casting on to metal master patterns is widely used for other materials: an example of this principle is included in the illustration in Fig. 8. Pattern Production and Assembly Wax is the most commonly used pattern material. Natural and synthetic waxes and various additives are blended to achieve minimum shrinkage and close reproducibility of pattern dimensions, together with strength for stability in handling and storage. Melting points are in the range 55 - 90 °C and the molten wax is usually introduced into the die from an injection machine, under either manual or automatic control. The expendable patterns readily incorporate most holes and cavities forming part of the casting design, the pattern emerging from the die as a full replica. In some cases, however, these pattern features can be more readily formed by using a soluble wax core insert of higher melting point. This is placed in the die before injection of the standard wax and subsequently dissolved out to leave the required cavity. Pre-fired ceramic cores are similarly embodied in some patterns, being left in position in the ceramic mould when the wax is melted out. Patterns for small castings are normally assembled in clusters around a common sprue and feeder system, similarly formed in wax, for mould making and casting. Mould Production The original block mould process used in investment casting is still retained in some applications, as, for example, for the small moulds used in dental casting, but the ceramic shell system has become stand ard practice through most of the industry. Pattern assemblies are dip coated in investment slurry, beginning with a primary coat. This is followed by alternate applications of further slurry dips and granular stucco material to build up a thick layer on the pattern surfaces. Although prolonged, this process does lend itself to automatic handling and control in special plants. Investment slurries contain graded suspensions of refractory particles, with binders which are most commonly based on soluble silicates. Setting and hardening are induced by controlled reactions and the shells are then ready for dewaxing and further consolidation by heating. Special heating conditions are required for the dewaxing stage to avoid shell cracking, after which high temperature firing eliminates residual volatiles to pro duce a strong, inert mould.

16 Investment Casting Casting and Finishing The metal melting equipment and techniques employed are not unique to investment casting. There is heavy emphasis on the production of high quality melts, whether in air, under controlled atmosphere, or in vacuum as used for much superalloy casting. Special techniques such as those used for controlled directional solidification will be detailed in other chapters. Centrifugal casting and vacuum - or pressure-assisted upward fill systems find increasing application. After knockout, dry, wet and chemical cleaning processes are used, the castings are cut from the feeding system and dressed, and the inspection and testing techniques appropriate to a precision cast component are applied. Related processes The Replicast CS process sequence is essentially similar to that just described, incorporating the same principle of an expendable pattern, coated and fired to produce a ceramic shell. In this case, however, the pattern is made from expanded polystyrene, using a specialized system involving injection of solid beads into an aluminium die, and the process is normally used for heavier castings than those typically produced by the investment casting industry. The process is a development of the lost foam or evaporation casting process, in which the same type of pattern is embedded in dry unbonded sand, being left in the mould to be displaced and evaporated by the incoming molten metal. Investment casting using jointed moulds and orthodox patterns is represented in various processes, including plaster moulding for non-ferrous castings, and the Shaw or ceramic mould process, which uses similar silicate bonded investments to those employed with the expendable pattern system. These again are mainly used for heavier products than the typical lost wax casting, although the latter is now being adopted over an expanding weight range. Some Characteristics of Investment Castings The outstanding feature of the process is the design freedom afforded by the capacity for intricate shaping, especially the production of thin sections and sharp detail. Investment casting offers all the general advantages of the casting route in respect of complex curves and contours, with the additional ability to dispense with the draft taper required in most other casting processes. Internal features present no problem, given the versatile alternative coring options. Exploiting these qualities, it is often possible to design complex single investment castings to replace assem blies of several separate components, so eliminating joining operations.

Introduction

17

Surface finish and dimensional accuracy are of a high standard; these and other attributes will be detailed and quantified in Chapter 11. There are few restrictions on investment casting in terms of available alloys and the process is particularly suitable for the production of intricate components in materials such as wear resisting and tooling alloys. In such cases much of the finished detail including holes, slots and fins can be formed in the original casting. The range of investment casting alloys will be reviewed in Chapters 10 and 11. The variety of applications determined by these exceptional quality characteristics will be demonstrated with practical examples in Chapter 12.

HISTORY OF THE PROCESS The basic technique of investment casting, under its traditional name of lost wax (or cire perdue) casting, has been known for well over six millen nia. The precise origin of the process is a matter of some doubt and various claims have been made. Table 1 shows the estimated ages of lost wax objects, plotted according to the area where they were manufactured or recovered.1 Archaeological investigations have indicated Mesopotamia, around 3000 -4000 bc , as the location of a civilised society of city states possessing skills in engineering and metallurgy, including the knowledge and the

Table 1. Estimated ages of lost wax objects (from P.R. Taylor1) 5000 BC 4000 Thailand Mesopotamia Israel India/S.E. Asia Anatolia China Aegean/Greece Etruscans Celtic N. Europe Roman South/Central America West Africa West Europe (Medieval to Victorian times) Renaissance Italy

3000

2000

1000

0

1000

t

: • ••« *•

• A *

* | •

•

•H

•

2000

18 Investment Casting means to produce a range of gold, silver and copper artefacts made by lost wax casting. Another candidate location for the original use of the technique is Thailand/South East Asia, where it is believed that metallurgical activities were carried out by local tribes rather than by urban populations. There is evidence that elaborate bronze artefacts were made by the lost wax method as early as 4500 bc in South East Asia. The Chinese were using the technique from 2000 bc onwards and the Egyptians from around 1400 bc . An archaeological excavation in 1972 of a first-century bc Iron Age factory, at Gussage All Saints in the UK, was particularly interesting, since it provided one of the few examples where clay-based investment moulds were recovered. Over 7000 fragments were found, for leaded bronze harnesses and chariot fittings. It is thus clear that knowledge of the investment casting process was widely dispersed in the ancient world, and by the time of Christ appears to have been known and practised in China, South East Asia, Mesopotamia, Egypt, Greece, Italy and Northern Europe, and possibly elsewhere as well. During the next 1000 years there are isolated references to the process. One remarkable example, dating from the 11th century a d or earlier, is Shiva, the Lord of the Dance, a 96cm high bronze figure surrounded by a circle representing the cycle of creation, destruction and birth. This in vestment casting was produced by the Chola dynasty in India, and emphasies their cult of the god-king; the statue is unsurpassed in technical skill and delicacy of design. Well before Columbus set sail, the Aztecs in Mexico and the aboriginal Quimbaya goldsmiths in the Cauca Valley, Colombia, were familiar with the process, producing remarkable hollow gold castings. The details of the actual processing were recorded by Friar Bernandino de Sahagua, who extensively studied these peoples.2 At about the same time (the 13th century ad ) investment casting was the chosen production method for a number of bronze tomb effigies for kings and queens; examples of these are the effigies of King Henry III and Queen Eleanor in Westminster Abbey. It may be noted that the 14th and 15th centuries represented the flowering of lost wax bronze casting in mid and western Nigeria, particularly in Benin, the capital of the Bini region of the country. Probably introduced into the area from the nearby Ife region some centuries before, the techniques became very sophisticated, but were restricted in use to artefacts for the royal household. The Benin bronze of Iyoba, the Queen mother, was chosen in 1972 as the crest for the British Investment Casting Trade Association, the head (Figure 9) being framed in a geometrical configuration favoured by the tribe. Figure 10 shows a West African bronze dating from approximately 1800 ad .

Introduction

19

Fig 9 Lost wax bronze head oflyoba, Queen Mother of the Benin tribe in Nigeria, chosen as the logo for BICTA.

20

Investment Casting

Fig 10

Taylor1).

West African warrior, bronze, approximately 1800 a d (Courtesy of P.R.

Introduction

Fig 11 Renaissance Italy: equestrian monument, bronze, approximately 1620 (Courtesy ofP.R. Taylor1),

21

ao

Early castings were produced from a wide range of patterns. Taylor1 claims that archaeological finds indicate that production of identical wax patterns was achieved by the use of dies of carvable stone, cast bronze and carved wood. It is possible that the bronze dies were also used to cast lead by the gravity die process. The Quimbaya workers probably mass produced patterns for ornamental castings by pressing sheet wax on to the carved surface of stone matrices. In regions as far apart as Africa, India, and South and Central America, it was traditional to build elaborate patterns from wax thread and wire; for small or medium single hollow castings, the wires or thread were wound round a clay or clay/ charcoal core, either to cover it wholly or as an open network. Lost wax casting reached possibly its highest artistic expression in Renaissance Italy, an example being shown in Figure IT Benvenuto Cellini produced many masterpieces by the process, one of the most outstanding being a bronze statue of Perseus holding the head of Medusa.

22

Investment Casting

Investment casting turnover by geographical area (after R.B. Williams5).

Cellini has left a detailed description of the process, both in a treatise of 1568 and in his autobiography, claiming to have learnt about the casting method from a description by the monk Theophilus Presbyster in his Schedula Diversarum Strium, dating from about 1100 a d . Other written evidence of investment casting has come down from about the same time, when Varrinee Krickes of Prague described the use of the lost wax method to produce bronze gun barrels. In 1538, Vannoccio Biringuccio, head of the Papal foundry and a contemporary of Leonardo da Vinci, wrote in his Pirotechnia: There are likewise moulds for large statues which, if one desires to make them of bronze, are first made of wax according to the ordinary procedure.

This procedure involved creating an original model or sculpture in wax, which was subsequently polished and embellished by its creator. Each

Introduction

23

item was a unique work of art; the image was then coated with a milky slurry of plaster, building up successive layers until a strong shell com pletely enveloped the wax. After melting out the residual wax, molten metal was poured into the void, which after removal of the plaster shell left a perfect duplicate of the original pattern form, complete even to undercuts and folds. The lost wax casting method continued to be the preserve of artistic applications and the statue of Eros in London's Piccadilly, which dates from 1893, is a notable application of investment casting, and indeed an early use of aluminium castings; with an overall height of 2.5 metres, this comprises an assembly of aluminium investment castings supported on a leg of solid metal. Around the 1900s the use of the process was extended to the manufacture of gold fillings and dental inlays for false teeth, and in 1932 the lost wax ceramic block mould process was developed with cobalt-chromium alloys known as "Vitallium ", for dental applications and orthopaedic components. These developments are more fully examined in Chapter 12. THE MODERN INVESTMENT CASTING INDUSTRY By the 1930s investment casting ranked as a useful specialised casting method, but with little relevance to mainstream engineering. It was the requirements of the Second World War that changed this situation and laid the foundations of the modern investment casting industry. An urgent demand for finished components could not be met by the capacity of the machine tool industry and attention turned to investment casting to produce precision components for armament and aircraft parts. The pace of development accelerated with the introduction of the aircraft gas turbine, where designers, seeking increased efficiency by the use of higher operating temperatures, were attracted to investment casting to form the refractory alloys specified (or developed) for turbine blading.3'4 To meet this challenge, the traditional process had to address four new requirements: (a) (b) (c) (d)

reproducibility of castings within close dimensional limits production of castings in high melting point alloys high standards of metallurgical quality cost savings over parts produced by alternative manufacturing techniques

It was the solution to these problems that laid the foundation of the modern investment casting industry; established firstly in the US and the UK, the industry was mainly allied to aircraft and military applications.

24

Investment Casting

Europe

Note: The largest individual national producers are the USA, which accounts for some 95% of the North American sector of the chart, the UK with 39% of the European sector, and Japan with 60% of the Asian sector. Fig 13

World investment casting market, share of turnover 1993 (from R.B. Williams6).

The introduction of the jet engine for civil aviation after the war proved a real opportunity for investment casting and strengthened the links between it and high quality, critical component manufacture. Expansion continued through the 1950s, with a growing list of applications and the beginning of a general commercial market. The range of metals and alloys cast became more diverse, with steel, superalloy and non-ferrous (copper and especially aluminium alloy) markets being established. The wider scope of the process was facilitated by the introduction, from the mid-fifties, of the ceramic shell process of mould pro duction in place of the original block mould technique; this gave more versatility to the process and allowed much larger parts than hitherto to be cast. Growth of the industry was initially slow. By 1958 UK output was only £5 m a year and by 1972 £29m a year; by 1982, however, it had reached over £100 m per annum and by 1986 it had doubled again to exceed the sales output of the steel casting industry. By 1990 output was in excess of £300 m , supplied by about 50 investment casting foundries employing 6500 people; this, representing 18% of the entire value of UK castings, made investment casting the largest sector of the foundry industry save for iron castings. The US investment casting market grew, with a series of peaks and troughs, from $35m a year at the end of the Second World War to $70 m a year by 1958 and to $1000m a year by 1979. There was very strong growth in the second half of the 1980s, so that the most recent sales figures are placed at about £1400 m (~$2200m ). There are some 385 investment casting foundries in the US and a further 30 in Canada.

Introduction

25

‘Western Europe Fig 14

Ratio of commercial to non-commercial output of investment castings (after R.B. Williams5).

In addition to the USA and the UK, Western (Continental) Europe and japan have also been recognised as principal producers of investment castings: Figure 12 compares their outputs for the years 1988 -92.5 The world market for investment castings has declined in the last few years, in line with general economic trends, and recent estimates6 (based on 1993 data) place the total at some £2200M ($3600M) per annum (see Fig 13). Flowever, it should be stressed that, in the absence of reliable production data, output for certain Eastern European countries, for the former USSR, China and some other states have been omitted from the overall total, which may thus underestimate the size of the investment casting market. Two types of application served by investment castings should be distinguished. One is for the aircraft, aerospace and military applications and is known as released castings in the UK and as documented castings in the US. The other type of application is for the general commercial market. Figure 14 shows the ratio of released/documented (i.e. noncommercial) to commercial castings for the four geographical areas mentioned above. These ratios differ significantly and this affects other performance parameters for the industry, e.g. output per employee. In both the US and

26

Investment Casting

the UK, non-commercial applications predominate. In the US, documented castings account for about 60% of all output by value, but only some 15% of investment casting foundries are involved in such work; the remaining 85% of foundries share the remaining 40% of the output. In the UK, however, while the proportion of released business is even higher than in the US and reaches about 70%, nearly all investment casting foundries are involved in this business as well as in routine commercial work. The Japanese investment casting market has a low non-commercial/ commercial ratio, indicating relatively little use of investment castings in the aircraft and defence sector; at the same time, investment casting in Japan has achieved a much higher penetration than elsewhere into the automotive industry. The performance of the investment casting industry over recent years has tended to be in sharp contrast with other foundry sectors, which have experienced difficult trading conditions, with foundry closures and minimal growth or reduction of output. This performance arises in part from the growing awareness and interest of designers in the design capabilities of precision castings. However, it is believed that the main engine for growth in the latter half of the 1980s was the buoyancy of the aircraft industry. This is exemplified by the US where, between 1985 and 1990, new orders for US-built aircraft increased by 59% and provided a unique opportunity for investment castings, which find use as turbine blades, nozzles and vanes, instrument housing support structures, valves, pumps and other items.7 In addition, more sophisticated systems on both military and civilian aircraft have led to a demand for new and more intricate castings with greater added value. The increasing use of investment castings in land-based gas turbines has also assisted the growth. Thus, whilst the commercial sector of the industry has grown significantly, it is the aircraft market that has, both technically and commercially, spearheaded the advance. Investment castings can account for 50% of the total cost of a modem jet engine.

Trade Associations A number of national or international associations exist to promote the practice of investment casting. In the UK, the British Investment Casting Trade Association, widely known by its initials BICTA, exists to assist the investment casting industry. It is located at Bordesley Hall, The Holloway, Alvechurch, Birmingham. B48 7QA, UK (Telephone: 0527 584770; fax 0527 584771). The Association was formed in 1958 by a group of UK investment casters who wished to improve the technical base of the industry by

Introduction

27

Table 2. Typical trade association activities

1. 2. 3. 4. 5. 6. 7.

Voice for investment casting Technical/commercial assistance Statistics and standards Research & Development Training Conferences and publications7 Promotional a c t i v i t y ___________________

collaborative efforts. It now operates with members throughout the world and includes investment casting foundries, suppliers of materials and equipment to the industry, and companies or individuals who are interested in precision casting. Membership has grown throughout the years and today UK foundry members account for about 80% by value of the output in the UK; the Association can claim, therefore, to speak as the voice of the British in vestment casting industry. Table 2, which summarises the principal activities of BICTA, may be taken as representative of the scope of a leading industry trade association. A similar body to BICTA in the USA is the Investment Casting Institute (or ICI) located at 8350 N Central Expressway, Suite 1110, Dallas Texas 75206 - 1601 (Telephone: 214 368 8896; fax: 214 368 8852). This has a membership comprising investment casters, suppliers and educational establishments and its activities cover a range similar to those described above for BICTA. In Germany, the Verein Deutscher Giessereifachleute (VDG) has a number of separate technical groups, one of which deals with investment casting. This association has in membership leading German investment casting foundries and is concerned generally with technical developments in that country. The VDG is located at Terstergenstrasse 28, Dusseldorf, Germany (Telephone: 49 211 687 10; fax: 49 211 687 1333). The French SGFF (Société Générale de Fonderies de France) has, since the middle 1980s, operated an investment casting professional group based on principal French investment casting companies. This group seeks to represent the investment casting industry in France and to promote the greater use of investment castings by publications and lectures. The SGFF is located at 2 Rue de Bassano, 75783 Paris Cedex 16, France (Telephone: (1) 47 23 5550; fax: (1) 47 20 40 15). In the Czech and Slovak Republics, an investment casting society has been active since 1960 and organises meetings and seminars. In Australia, a small investment casting group is in existence. In other countries, precision casting matters are generally dealt with by the national foundry associations. Slightly different in concept is the European Investment Casters' Federation (EICF) which operates with foundries and suppliers

28

Investment Casting

throughout Western Europe. This is a truly pan-European body, whose formation predates any of the other European national investment casting groupings. Governed by an international Board, the Secretariat is provided by BICTA from Alvechurch in the UK. The activities of the EICF have centred on the organisation of biennial European conferences and, in conjunction with BICTA and the Investment Casting Institute, of World Conferences on investment casting. These various meetings have become established as major venues for the announcement and discussion of significant technical developments.

Publications Investment casting receives periodic coverage in the technical foundry press in the different producer countries but there are only two regular periodicals that are entirely devoted to the process. One is the monthly magazine Incast, published by the Investment Casting Institute, and the other is the quarterly BICTA Bulletin; these circulate widely in the industry and carry articles, news and views of current interest. The main technical developments are to be found in the technical pub lications of ICI and BICTA and in the papers and proceedings of the major conferences. RESEARCH AND DEVELOPMENT The investment casting industry has a long history of technical develop ment and innovation. While much of the underlying R & D has been undertaken by individual companies, the UK industry has pioneered collaborative pre-competitive R & D on subjects of general interest to the industry; organised and managed by the Trade Association, valuable progress has been made in the development of techniques to ensure high integrity investment castings, with greater consistency of properties. This has involved the better understanding and optimisation of the process variables, and the majority of the UK members of BICTA have played an active part in such researches. Collaborative research is now being undertaken in other countries, and is helping to secure the technical future of the investment casting industry.

REFERENCES 1. P.R. Taylor:

M eta ls a n d M a te r ia ls ,

2 (11), 1986, 705-710.

Introduction

29

2. Designers Handbook for Investment Casting. British Investment Casting Trade Association, 1990. 3. R.F. Smart: Industrial Minerals, (2), 1992 51-58. 4. 'Investment Casting', Industry and Development Global Report 1992/93, United Nations Industrial Development Organisation, Vienna, 296-307. 5. R.B. Williams: 22nd EICF Conference on Investment Casting, Paris, April 1992, Paper 1. 6. R.B. Williams: 23rd EICF Conference on Investment Casting, Prague, June 1994, Paper 2. 7. T. Gibson: 'The Future for Foundry Casting', Industrial Minerals, 1991, 9-17.

2

Tooling G.A. BELL

Tooling is a crucial consideration to investment casters and their customers. The quality of the tool used has a major influence on the price and quality of the casting. As a consequence it is essential that: 1. Both the caster and the customer spend time defining the casting requirements prior to tooling manufacture. This must include Estimated volume. Batch volumes. Castings end use. Final product construction. Casting quality requirements. 2. The caster takes time to evaluate various tool designs prior to com mitting the design to manufacturing. 3. The customer is made aware of the various options of tooling types and costs and the effect these options have on the casting price and quality. If the above three suggestions are carried out it is more likely that the customer will receive: 1. The casting he actually requires and not the one the caster may have thought he required. 2. A tool whose cost and life will reflect the quality and produce the estimated volume. A saying attributed to Gucci sums it up perfectly — The sweet taste of a cheap price is long forgotte?i after the sour taste of poor quality begins.

There are several tooling methods and types, with various combinations available to be chosen. Many things affect this choice, but the main factor is generally the quantity of units required and the period over which they will need to be produced. DOI: 10.1201/9781003419228-2

Tooling

31

Tooling types range from plaster cast to rubber, resin, metal spray and metal dies, generally in order of increasing casting volume requirement: each of these types will be briefly reviewed and characterized in the present chapter. PLASTER CAST DIES These are produced by casting a plaster mix over a pattern. They are rarely used in industry for production parts because they increase the time the wax is in the mould, due to their poor heat conductivity. Unless they are encapsulated in a flask the wax can only be gravity poured and because they are plaster they are easily damaged. Plaster dies are best produced by mixing the slurry and casting the die under vacuum. This avoids inclusions of trapped air on the mould surface and hence a wax casting problem. Their main application is in the art world, but they can be used extremely effectively for specialized short run gating dies.

Fig. 1

Gravity poured rubber die (held together by rubber band).

32

Investment Casting

Fig. 2 Rubber die (from previous figure), split open to show the cavity and rough cutting of the joint.

Fig. 3

A rubber die and its retaining flask. On the left is the clamp injection plate. When

the rubber die is assembled and inserted in the aluminium flask (right) the clamp injection plate is inserted on top. The press then compresses this plate until it is flush with the top of the flask and the wax is then injected. This pressing of the complete unit together pressurises the rubber and enables the die to be injected under pressure without distorting the rubber.

RUBBER DIES Like the plaster dies these are not com m only em ployed throu ghou t the m anu factu ring ind ustry, as they are slow to operate, inaccurate and have lim ited life. They find their use m ainly in the art and jew ellery foundries

Tooling

33

as discussed more fully in Chapter 12. There are several methods of producing rubber dies, but two basic systems are as follows: 1. A pattern is produced and a layer of wax, of the thickness of the rubber required, is laid over it. This assembly is then placed on a rough joint and resin or some like material is cast over one half. The joint line part is then removed and the other half is similarly cast. This is called the flask. The flask is then opened and the pattern and wax film are removed. The pattern alone is placed back in the flask and rubber is cast in where the wax film was previously located. When the rubber has cured, the rubber encased pattern is removed and a rough joint line is cut in the rubber using a sharp knife (Figure 2). Using this method, wax patterns can be injected under pressure since the rubber is supported by the rigid flask (see Figure 3). 2. A pattern is placed in a chamber and rubber is poured over it. When the rubber cures, the mould is cut with a sharp knife to create the opening joint and the pattern is removed. These moulds can then be used to create wax patterns. The main difference from the above method is that the wax pattern can only be hand poured, as the rubber mould has no support and any pressure would distort the cavity. RESIN DIES The choice of a resin die is generally made when low volume, low tolerance products are to be manufactured and there would be no opportunity for the amortisation of more expensive tooling, such as that made in aluminium or steel, over the expected total product run. The main advantages of resin tooling are:1. Short tooling lead time. 2. Complex joint lines are relatively easy to produce. 3. Alterations, adjustments and design modifications are easy to achieve. 4. Multiple dies are inexpensive. 5. Overall tooling cost is low. The main disadvantages are:1. 2. 3. 4. 5.

Insulation properties increase wax injection cycle times. Life of tooling is substantially less than those of aluminium or steel. More release agent is generally required. Patterns usually require trimming. Dimensional tolerances of the final product are hard to maintain.

34

Investment Casting

The manufacture of resin tooling is relatively simple in comparison to that of metal tooling, because the resin is cast over a wooden, plastic or other easily carved material to produce the die. Naturally, accurate resin dies require accurate models. However, although highly skilled tradesmen make the model and check its dimensions using height gauges, verniers and similar equipment, they are unable to achieve the accuracy obtainable on the milling machines, spark eroders and similar modem automatic plant used for most metal die production. As a consequence of this, drawing-to-part tolerance cannot be held as tightly, since a resin die generally uses up more of the total allowable tolerance than a fully machined die. Added to this, the die-to-part tolerance is more variable than that achieved with metal tooling, because the cooling rate in a resin die varies substantially from injection to injection and hence the wax patterns vary correspondingly in size. This problem can be overcome, to some degree, by having multiple dies. The ability to rotate several dies through the wax press assists in maintaining a fixed cycle time for each injection. Additional resin dies are relatively inexpensive, as once the master model is produced and the jointline generated it is simple to cast additional resins. Likewise, a damaged resin die is easily replaced for the same reason. Alterations to resin dies are inexpensive, as the model can be readily modified to the new design. The area of the die requiring modification is then carved clear and a hole is drilled through the die. Once this is done the modified model is placed back into the die and resin is cast down the hole, filling the remaining cavity. The resin bonds well to itself, so a modified or repaired die is generally as strong as the original. If a part cannot be readily machined using conventional methods, e.g. because of compound curves, complex jointlines or similar features, then a resin die will be anywhere from three to twenty times cheaper than a machined die. Wax patterns from resin dies generally require more trim ming than those from metal dies, especially in the jointline area. This is mainly due to the surface tension of the resin when it is cast over the model, which detracts from the production of a sharp joint between the upper and lower halves of the die. Resin dies also require more release agent to allow the wax pattern to strip easily. The additional release agent reduces the surface finish of the wax pattern and increases the need for wax pattern cleaning. It also increases the chance of release agent build-up in virgin wax through the recycling of wax sprues that have been coated with the material, and may thus increase the incidence of wax pattern imperfections. It is advisable to cast an aluminium insert into a resin die in the injection area. If joint line injection presses are used, the nozzle pressure will damage a resin die very quickly and the addition of a small aluminium block in the immediate nozzle area will eliminate this problem. Vertical

Tooling

35

injection presses are subject to a similar problem, in that the nozzle pin can damage the resin; consequently an aluminium insert in this area is again advisable. It is important when using such inserts to ensure that they are properly keyed into the die. Parts from resin tools are generally more expensive to produce because of the poor conductivity of resins and the consequent increase in wax pattern cycle time. Wherever possible, therefore, foundries should encourage their clients to purchase the more expensive metal dies; these have a longer life, the wax pattern cycle time is much shorter and the castings should thus be cheaper. However, where product volume cannot sustain the initial metal tooling cost, then resin dies are the answer. Foundries should maintain their awareness of the dimensional instability of resin dies when quoting, and should pay particular attention to the drawing tolerances required. Should the customer be willing to accept resin tooling, the foundry should make the customer aware of the problems of holding tight tolerances, and of the relatively short life expectancy of the dies. It is recommended that these points be made in writing to ensure that there is no misunderstanding. METAL SPRAY TOOLING Metal spray tooling is an excellent option for manufacturing products where complex jointlines are required but where solid metal tooling is too expensive. Generally spray metal tooling will produce a wax pattern nearly as quickly and consistently as full metal tooling and can be produced at little more cost than that of resin tooling. The basic advantages of metal spray tooling are:1. 2. 3. 4. 5. 6.

Relatively low die costs. Reasonable wax pattern quality. Short tooling lead time. Good tooling life. Complex joint lines are readily achievable. Multiple dies are relatively inexpensive.

The basic disadvantages of metal spray tooling are:1. 2. 3. 4.

Parts generally require trimming. Modification of parts is difficult. Ejectors are difficult to fit. Repairs must generally be made in resin.

Spray metal dies are produced using basically the same technique as that for resin dies. A model can be made in wood, resin, plaster or wax and a

36

Investment Casting

Fig. 4

Master pattem set in a joint line ready io spray.

jointline developed as shown in Figure 4. Metal is then sprayed on to the pattern and joint surface (see Figures 6 and 7). In principle any alloy can be used for the purpose but the metal most commonly employed is a zinc alloy, chosen as it is easier and faster to spray. Basically a metal spraying machine is similar to a MIG welder, but instead of one wire two wires are fed. As they touch, the wire vaporises and an air blast blows the vaporised metal on to the pattern. Very little heat is generated in the vaporised metal and it is quite normal to spray on to wax patterns without any deterioration of the wax. A thin layer is applied and air is then blown over the layer to cool it. Subsequent layers are applied until there is a build-up of about 3mm, a process very similar to spray painting. Care must be taken not to build up too much heat between the layers, as this can tend to pull the previous layers away from the pattern and joint. Aluminium, brass and steel can also be sprayed, but they are far more difficult and hence increase the cost of the tool. It is important when making a spray metal die to ensure that all sprues and ingates are incorporated in the pattern and joint before spraying, since machining of the metal spray deposit tends to leave chipped edges. If ejectors or vents are required it is advisable to fit an aluminium pin on to the pattern so that it protrudes out of the back of the die. This can then be drilled to accommodate the ejection

Tooling 37

Fig. 5 Metal spray die completed using master patterns from Figure 4. In this picture, the end has been cut away to highlight the 3 mm of sprayed metal and the coarse granular aluminium and resin backing. On the right, the aluminium frame can be seen. pin or vent at a later stage. After spraying, an aluminium frame is attached with resin to the back of the spray and the inside is filled with a mix of resin, aluminium powder and aluminium chips (see Figure 5). These reduce the heat retention effect of the resin and act as heat conductors. The cost of metal spray dies is about 20% more than that of resin dies. Because of their superior heat conductive properties wax pattern cycle times are generally only 25% of those of resin dies, so the additional cost can quickly be made up in wax pattern production. A spray metal die takes from 2 - 4 hours longer than a full resin die to produce. The life of a spray metal die is extremely good: depending on the complexity of the tool it can be anywhere between 10,000 to 100,000 injections. The main areas that can cause trouble are knife edges and the die joint, otherwise the dies have a similar life to aluminium dies. As with resin dies, complex joint lines are easy to achieve and once a joint and pattern are made replacement or multiple dies are relatively inexpensive and quick to produce. Deep blind pockets are a problem in the manufacture of spray dies as it is extremely difficult to spray into

38

Investment Casting

Fig. 6

Metal is being sprayed on to form a die. In the top right corner is the metal spray

gun. The dull pattern has been sprayed with metal while the brighter one is yet to be sprayed.

these areas. Modification or repair of a cavity is also extremely difficult, since quite a large area needs to be cut away, right through the die. The actual spray deposit is relatively open grained as compared to solid metal

Tooling 39

Fig. 7

Completed metal spray die, manufactured from the part shown in Figure 6.

and consequently the edges tend to chip when machined. Because of this difficulty, repairs are generally done with resin. As a consequence the repaired dies have cooling problems in the resin area and wax pattern cycle times increase. Also the pattern finish in the resin repaired area is not as good as that in the unmodified area, because more release agent is needed to free the wax pattern. Given the difficulties associated with major modifications or repairs, it is generally more economical to replace a spray metal die totally than to modify the existing one. TIN - BISM UTH DIES The main advantage of cast tin-bismuth over spray metal dies is the wax pattern production rate. They are, however, generally more difficult to make and hence rather more expensive in terms of initial cost. The main advantages of tin-bismuth dies are:1. 2. 3. 4. 5.

Good production rates. Relatively low die costs. Relatively short lead times. Relatively easy to repair or modify. Complex joint lines are easily achievable.

The main disadvantages are:1. Waxes generally need trimming due to joint lines not being sharp.

40 2. 3. 4. 5.

Investment Casting The tin-bismuth material is expensive. Dies are easily damaged. Cavities require a fair amount of hand clean-up. Suitable for small parts only.

Tin-bismuth dies are produced using a wooden pattern and joint line. The joint and pattern surfaces are covered with a release agent, normally a carbon film. An aluminium block is then machined out to relieve the area around the pattern, leaving usually about 5mm of clearance. This clearance is kept to a minimum as the tin -bismuth alloy is quite expensive. It is important to ensure that the aluminium blocks are totally dry, as any moisture will turn to steam when the tin-bismuth is cast, causing blow holes in the die. The aluminium block is placed over the pattern and the tin -bismuth is poured through a hole in the block, filling the cavity be tween pattern and block. To improve definition a small positive air pressure can be applied to the molten metal; this also assists in reducing the radius created at the joint lines by the surface tension of the alloy. Modifications are achievable by adjusting the pattern and machining out the die in the area to be modified. Tin-bismuth alloy is then poured back into the modified area. Repairs can be carried out on the same principle. Because the die is made totally of metal and there is no insulating resin material in its construction, this makes a tin-bismuth die marginally better than a spray metal die, which does have resin acting as an insulator between the spray metal layer and the aluminium frame, thus giving longer cooling times for the wax patterns. The production of this type of die is usually confined to small parts and it is uncommon to see large tools produced. This is not just because of the alloy cost, but primarily because of inherent problems in the ability of the tinbismuth alloy to flow over large areas. As in the case of metal sprayed tooling, this type of die is probably not used to its full capacity, mainly because of lack of knowledge by toolmakers about the manufacturing process.

FULL METAL DIES Full solid metal tooling is used where high production rates or long runs are expected, and for the manufacture of components requiring supreme accuracy. Adequate design time is essential before any manufacturing begins if the ultimate ease and standard of wax production are to be achieved. It is a matter for individual preference whether an aluminium or steel die is to be manufactured. The foundry must decide which attributes of aluminium or steel dies will best suit the individual application.

Tooling 41 The benefits of aluminium over steel are: 1. Wax production is marginally faster, as the aluminium has greater heat transfer ability. 2. Aluminium can be machined more easily and rapidly, so that an aluminium die will normally be less expensive. 3. Dies are easier to load and unload from the wax press due to their lighter weight. The benefits of steel over aluminium are: 1. Where fine knife edges cannot be designed out of the die, steel is less likely to distort than, or damage as readily as, aluminium. 2. Joint edges are not damaged as easily. 3. Repair is generally simple, since steel is relatively easy to weld as compared with aluminium. 4. Ejection pin holes do not wear as rapidly and hence do not allow wax to flow down them. Metal dies are produced from solid stock or cast preforms using all the normal types of machining methods such as milling, turning and spark erosion, and modern CAD/CAM developments can be exploited to achieve optimal design and cost. The construction time of a die for basic parts that can be readily milled or turned is not much greater than for the previously discussed forms of tooling. However, once a part requires more than a flat joint line, or involves more complex machining operations, then both cost and lead times escalate. If the die maker has CAD/CAM available, complex components and joint lines are less of a problem. Also, once the programme for one cavity is generated, CAD/CAM has the distinct advantage of facilitating the production of multiple cavities relatively cheaply. Cavities produced by CAD/CAM are far more accurate, and complex joints require a minimum of bedding. The life of these dies far exceeds that associated with the other methods and materials. Casting dimensional tolerances are also easier to achieve, since the die maker can achieve much closer tolerances on the cavity, thus leaving more of the drawing tolerance available to the foundryman. Figure 8 (a) and (b) shows a metal die embodying some of the special features required for the production of patterns for a complex component. Although the manufacture of solid metal dies is still based largely on long-established toolroom techniques, future advances in machining plant and practice are likely to ensure the continued use of this form of tooling for the production of investment castings of the highest quality for the forseeable future.

42

Investment Casting

Fig. 8 (a) and (b) A full metal die, with loose inserts required to ease the stripping operation. The square and round boss features on the inserts have been made out of brass. This is done to reduce the chance of them being damaged when the die is stripped.

3

Pattern Technology R.B. WILLIAMS

INTRODUCTION The objective in this chapter is to examine how investment casting wax has developed, with a review of structure, categories of casting wax available, properties and wax pattern production. The writer then moves on to look at the possible direction wax may follow in the future, considering quality control, choice of wax, future materials, reclamation and reconstitution, and cost. BRIEF HISTORY AND STRUCTURE OF INVESTMENT CASTING WAX Wax is the oldest thermoplastic material known to man and, because it can be cast or formed while in a liquid, semi-liquid or plastic state, its history has been closely linked with the arts and crafts and the growth of the investment casting industry. In early times craftsmen of China and Egypt used the lost wax process but the name referred only to beeswax. However, today in the investment casting industry, the name applies to any substance having a wax-like property. Modern blends of investment casting wax are complex compounds containing numerous components, such as natural hydrocarbon wax, natural ester wax, synthetic wax, natural and synthetic resins, organic filler materials and water. Many variations of such compounds have been formulated to suit various requirements; properties such as melting point, hardness, viscosity, expansion/contraction and setting rate are, of course, all influenced by the structure and composition of any wax compound. When we are dealing with hydrocarbon wax, natural ester wax, many types of synthetic wax and some of the resins used, we are usually DOI: 10.1201/9781003419228-3

44

Investment Casting

dealing with compounds of straight-chained carbon atoms. However, some of the resins and filler materials used could also be compounds of ring structured carbon atoms. Generally, the shorter these chains are, the lower the melting point of the wax and the less its hardness. With increasing chain length both hardness and melting or congealing point rise. The chain length will also influence the viscosity and solubility of the wax. The fact that casting wax is a mixture of a large number of components of different chain lengths results in wax manifesting a physical behaviour different from other substances. Wax does not melt immediately on heating like other homogeneous chemical compounds but passes through an intermediate state. This is illustrated in Figure 1. As seen from the shape of the curve, with gradual heating, wax of an initially solid consistency becomes softer, then plastic and with further heating semi-plastic. At higher temperature it acquires the consistency of a thick liquid (semi-liquid), finally passing on complete melting to a New tonian liquid. It is worth mentioning here that filled wax is not a true Newtonian liquid, but would usually still show a behaviour similar to the one depicted by the curve. This gradual change in the overall state occurs because short chain fractions melt first while longer chains remain solid. With further increase in temperature the latter melt progressively until the liquid state is reached. The actual shape of the curve and the temperature range of each condition is naturally a reflection of the specific makeup of the blend.

Fig 1 Hardness of a typical wax against temperature.

Pattern Technology 45 Of course, on cooling the reverse takes place and will again occur according to the make up of the blend, over a longer or shorter temperature range, giving rise to a range of setting rates. The structure or components of a casting wax will also affect its expansion/contraction. Wax expands like other materials under the influence of heat and on cooling it contracts. In comparison with a metal, the expansion of a wax is relatively high. In waxes the expansion and contraction rates over a range roughly between 20°C and the melting point are not uniform but change over the temperature range as a function of their structure. It may be useful to demonstrate this by showing some typical expan sion curves for the following three types of material: a homogeneous crystalline organic substance, a wax and a non-crystalline resin (see Figure 2). The crystalline substance behaves like any solid and undergoes relatively little expansion. At its melting point, the crystalline structure suddenly breaks down and a sharp transition into the liquid state occurs, which is characterised by a sudden increase in expansion. In the liquid state the expansion is again small. In wax the short chain fractions become soft even at low temperature, giving a gradual rise in the expansion curve. In the case of the higher molecular weight, crystalline fractions the curve assumes a steeper increase and then rises slowly again on the transition to the liquid state.

Expansion

Fig 2

Comparison of expansion behaviour.

46

Investment Casting

The non-crystalline resin behaves differently. It has a uniform pattern of expansion from the start of heating to the liquid state. No sharp increase in expansion occurs since no crystalline elements are present. Hence the addition of certain resins to wax can reduce the crystalline structure and help to reduce this expansion/contraction capacity. In this brief review of structure we have a simplified view of why numerous components are added to a wax blend and of the properties that result. We can now consider the types of investment casting wax available and how these are categorised.

CATEGORISATION OF INVESTMENT CASTING WAX For ease of reference casting wax can be divided into the categories shown in Table 1.

Table 1. Types of casting wax

Pattern wax Runner wax Reclaim or reconstituted wax Water soluble wax Other special wax - including dipping, patching and adhesive (sticky wax). Pattern wax can be further divided into the following three main types: Straight or unfilled pattern wax Emulsified pattern wax Filled pattern wax Straight or unfilled pattern waxes These are in effect complex compounds of many wax and resin components. The surface finish of straight wax would normally be shiny and the compounds can usually be reclaimed and reconstituted for use for both runner systems and patterns. Emulsified pattern waxes These have similar base materials to the straight wax compounds but are emulsified with water, normally between 7 -12%. The surface finish is extremely smooth and because the water acts partially as a filler very little cavitation takes place. Handling of emulsified waxes is quite simple providing the foundry keeps to the guidelines laid down by the supplier. They have become extensively used due to their versatility and again these compounds can usually be reclaimed and reconstituted for use for both runner systems and patterns. Filled pattern waxes Here again the base materials are similar to those of the other two categories, but into the compound is blended a powdered, inert filler material, insoluble in the base wax, to give the compound greater stability and less cavitation. It is essential that the filler used is organic to ensure complete burnout, leaving no ash, and a number of

Pattern Technology

47

different filler materials are used. It is also critical to use a filler of fine particle size so that surface finish is not impaired, and to have the density of the filler as near as possible to that of the base wax to ensure that minimum separation takes place when the wax is liquid. Here again they are widely used and with advanced reclaim technology can usually be reclaimed and reconstituted for use in runner systems or patterns. Runner waxes These have similar base materials to straight waxes and are blended to give the strength runner bar systems demand. Reclaim or reconstituted waxes This is a service carried out by the wax manufacturer, whereby a foundry's used wax can be thoroughly cleaned and blended or reconstituted to an agreed specification. The material is then returned for use for runner systems or patterns again. Straight or unfilled wax, emulsified wax and filled wax can all be reclaimed and reconstituted in this way. Water soluble waxes These are designed to produce internal shapes which are difficult to produce by other means, as explained previously on page 15. The waxes are soluble in water or mildly acidic solutions. Other special waxes These are unfilled wax compounds used in dipping, patching, or repair and adhesive applications. PROPERTIES OF INVESTMENT CASTING WAX AND THEIR INFLUENCE ON QUALITY As explained before, the majority of investment casting wax materials are complex compounds of numerous components. Each component has been included to influence the final properties of the compound in some way. These properties of the wax are obviously of critical importance to the foundry for the production of good castings. Once a specification for a casting wax has been agreed between wax manufacturer and foundry, it is essential that the material is manufactured, tested and supplied within these limits. In looking at general properties of a casting wax it will be useful for both foundry and wax manufacturer to consider a series of points that affect the quality of a casting wax and hence pattern production. As the industry moves forward, so even more emphasis will be placed on understanding the control of these points or properties. The points are listed in Table 2. Once a foundry has decided on a particular wax compound it is extremely important that a consistent contraction/cavitation rate be main tained. We have already discussed how structure and composition can affect contraction, emphasising that certain components will influence this property and highlighting the importance of the quality control tests carried out by both manufacturer and foundry.

48

Investment Casting Table 2.

1 2 3 4 5 6 7 8 9 10

Important characteristics of casting waxes

Contraction and cavitation Congealing point or melting point Ash content Hardness and elasticity Viscosity Surface finish Setting rate Oxidation stability Reclaimability Other characteristics

Again we have considered how structure influences the congealing point and melting point of a casting wax. These in turn have a major influence on the required injection temperature. As was explained in the section on the structure of wax, casting wax passes through a number of phases on heating and/or cooling. Congealing point and melting point will represent temperatures at the beginning and end of the semi-liquid state respectively. With the knowledge of either temperature the correct wax conditioning and injection machine temperatures can be set. Most foundries would be aware of the importance of using and main taining a wax with a low ash content and of the detrimental effect of ash. The limit recommended by BICTA is 0.05% maximum. However it is also important not to place all the emphasis on the percentage weight of ash without considering the nature of the residue left and whether this could cause problems in the mould and affect the casting. We have discussed how structure can affect the hardness of a wax. The wax must have sufficient hardness and elasticity to reduce the possibility of rejects due to breakages, bending or other undesirable phenomena during the subsequent processing of the wax pattern. Different compo nents will affect the wax compound in different ways. Again we consider how structure influences the viscosity of a wax. The viscosity of a casting wax compound is critical to successful pattern pro duction. Where large fine sections need to be produced a low viscosity wax is required to enable the wax to penetrate into the finest spaces in the die. For heavier sections a less fluid wax may be preferred. If a wax with the incorrect viscosity is used for a particular application then the flowability of the wax into the die will be wrong. This highlights the need for quality control in respect of this critical property. Again good surface finish is an important property for successful pattern production. It goes without saying that a poor quality wax pattern surface will give the same poor quality to the resultant casting. The three major categories of pattern wax — straight, emulsified or filled — will give, as mentioned earlier, different surface finishes. In general, straight

Pattern Technology

49

waxes are more shiny on the surface and emulsified waxes more smooth, whereas filled waxes are slightly rougher. In their own ways all three are satisfactory and foundries have their own preferences. Examples of the types of surface that could prove detrimental are the soft, easily damaged surface, or the 'pitted' surface usually associated with coarse particle sized filler being used. The foundry must have a knowledge of the setting rate of a wax for the successful production of wax patterns, and mention has already been made of how different structures or components give different setting rates. At one extreme, foundries are producing parts where they require a very fast set and release from the die, whereas at the other extreme a slower setting wax is an advantage. Stability of the wax compound is a property worth consideration. Here one is thinking in terms of the ability of the compound to resist oxidation or breakdown of certain of its components, due to the action of heat or simply to ageing. Some components have a greater tendency to oxidise than others and it is necessary for the manufacturer to use antioxidant materials where this could occur. If oxidation of the wax does occur then the overall properties will markedly change and the compound may be unsuitable for use. The reclaimability of a wax is important from both ecological and econ omical standpoints. Whilst it is possible to reclaim and reconstitute all three categories of wax to an agreed specification, strict quality control over the process is recommended. The topic of reclaim wax is discussed further later in this chapter. No doubt there are other properties that could be considered. For example, the fact that such compounds should be non-toxic is obvious. However, the items considered should cover the majority of general properties required of an investment casting wax, and how these can affect quality of the wax and wax pattern production. It is the emphasis that foundries and wax manufacturers place on these properties, linked with quality control and commercial considerations, that will determine how compounds develop in the future. WAX INJECTION EQUIPMENT* A wax injector is a machine that takes a preconditioned wax and injects it into a die, creating a wax pattern. Injectors are classified by the state of the * Section based on a text provided by courtesy of M.T. Pinczes, Mueller Phipps International, with additional illustrations by courtesy of Schott GMBH and Tempcraft.

50

Investment Casting

wax that the machine is capable of injecting. There are three basic types of injector, namely liquid, paste and solid (billet). Liquid injectors inject wax at a consistency ranging from that of a cooking oil to that of honey. A liquid injector consists of a heated wax reservoir which is agitated to prevent any filled waxes from separating while held in the machine awaiting injection. Wax is transferred from the reservoir into an injection unit by a vacuum created by a retracting hydraulic or pneumatic cylinder, and by the weight of the wax itself (gravity feed). When the injection cylinder is full, valving in the injection unit closes, separating the injection unit and reservoir, and the machine is now ready to inject. A die is loaded into the machine and clamped. When sufficient pressure is reached to hold the die together, the injection cylin der compresses the wax in the injection unit to a preset pressure, which pressurizes a manifold or heated wax hose feeding a nozzle assembly. The nozzle assembly moves forward and contacts the die, which opens an internal valve, which in turn fills the die. The machine then holds pressure in the die for a predetermined dwell time, to allow the wax to cool and harden to a state in which it may be removed from the die and hold its shape within tolerances, so that a good part may be cast from the pattern. At the end of this time the injection unit depressurizes and the machine opens, allowing for part removal. Liquid wax injectors should be refilled on a constant basis with a conditioned wax for best results. Wax that is put into the reservoir too cold can cause air entrapment in the wax, which ends up as air bubbles in the wax pattern. Wax that is put in the reservoir too hot can overheat the injection system, which can cause dimensionally unacceptable patterns because the part has had insufficient cooling time under pressure, which creates a smaller pattern (all wax patterns shrink to some degree during cooling). All zones of heating should be set to the same temperature for the best results. A paste injector injects wax that has a consistency ranging from that of a toothpaste to that of vegetable shortening. There are two types of paste injector on the market today. The first is a 'canister7 type injector. A canister, or cylinder, is filled with a liquid wax and is then placed in a tempering oven until the wax conditions to a preset temperature. After a canister is conditioned, it is placed in the machine (which is at the same temperature as the tempering unit) to be injected. The canister is the injection cylinder of the machine. It contains the injection piston, which mates with a hydraulic cylinder that is permanently mounted to the machine. When the canister is loaded, a switch is thrown which causes the hydraulic cylinder to compress the wax, pressurizing the canister and an injection manifold which contains a nozzle. The nozzle may have an internal valve like the liquid injector, or may be of a manual ball valve type. Some canister type paste injectors may also allow multiple nozzle

Pattern Technology 51 assemblies on the injection manifold. It is important, when filling the canisters, that the wax is entered in a condition liquid enough to allow any air to escape before the wax becomes too thick. It is also important to top up the cylinders with liquid wax after the cylinder has tempered, so that only a minimal amount of air needs to be bled out of the injection manifold after a new cylinder is loaded. The second type of paste injector available today is a hybrid liquid machine that has a two-stage wax reservoir. The top section of the reservoir is kept hot to keep the wax liquid, which allows it to let any air out of the wax. The bottom section consists of a scraped wall heat exchanger, which can cool the wax on the reservoir walls and then scrape it back to the centre of the reservoir, where it is blended with the hotter wax to create a smooth paste. This type of injector must be attached to a liquid wax supply that can keep the upper section of the reservoir at an acceptable level for the conditioning reservoir to work properly. This is the only type of wax injector in which conditioning is done on the machine. The rest of the machine works like a standard liquid injector. A solid or billet injector uses a pre -made tempered wax billet that is loaded into an injection chamber. The billet is heated wax that can main tain its shape outside the machine. The wax reduces in viscosity when squeezed under pressure through a nozzle assembly, so that the solid form becomes soft enough to flow into the die and fill the cavity. The liquid injector is the most popular wax injector in the investment casting industry today. The liquid injector benefits from the liquid wax in that this is easily pumped through piping from a central wax transfer system, making it the easiest to maintain wax levels in the machine with a minimum amount of human intervention. Hotter wax temperatures that are associated with a liquid injector, however, increase cycle times and increase the amount of shrinkage that will occur. Wax manufacturers have helped in reducing these problems by adding various fillers to reduce wax shrinkage. Canister type paste and billet injectors, although they inject at lower temperatures, have their own problems. Loading of these types of injector is labour intensive and requires tempering ovens which take up valuable floor space. These types of machine are also not very efficient, due to air bleeding when a new canister or billet is loaded. Approximately 20% of a billet is wasted before it can make an acceptable pattern. The hybrid liquid/paste injector offers the best of both worlds but must be hooked to a liquid wax supply that can keep up with the output of the machine. There are four different types of clamping unit available in different size ranges on today's wax injector. There is a four-post horizontal type, which covers most of the automatic wax injectors. Dies must be permanently mounted (bolted on to the platen face) on this type of press.

52

Investment Casting

Fig 3

A fully automatic wax injection machine (Courtesy of Mueller-Phipps International Inc).

This configuration is used because it allows the pattern to be removed from the die automatically using knockout cylinders. Then there are vertical four post and C-frame designs, which are the most popular. Tools may be either permanently mounted or moved in and out of the press for disassembly. All of these types of press are available with bottom or parting line injection. The last is hand clamping, using bolts or manual clamps which must be removed for disassembly. The clamp force needed to run a particular tool is determined by the product of the injection pressure and the cavity area at the parting line of the tool. There are three parameters that are the most crucial in making a good wax pattern, regardless of which type of injector is used: wax temperature, injection pressure and injection flow rate. The wax temperature should be kept constant throughout the injector. Reservoir temperature should be the same as the nozzle temperature; repeatable temperature yields repeatable wax patterns. The injection pressure should be high

Pattern Technology

53

Fig 4 A semi-automatic wax injection machine (Courtesy of Meuller-Phipps International Inc).

enough to give a good surface finish (if the tool is splitting, the clamp pressure should be increased or it should be run on a larger press). The injection flow rate is the most important in injecting thin-wall or irregularly shaped patterns. The flow should be such that the wax can enter the die quickly enough to eliminate flow and knit lines, but slowly enough to prevent turbulence and trapped air. There are commercially available electronic flow controls which can help alleviate these problems. Machines are made in a wide range of capacities, characterized by the available clamping force, and offering varying degrees of automation. Two typical modem machines embodying full control of wax temperature, pressure and flow rate are illustrated in Figures 3 and 4. In the fully

54 Investment Casting automatic version shown in Figure 3 the ejected patterns are collected in a stainless steel water tank and steered by nozzles to a variable speed conveyor for removal to the left of the machine, whilst the automatic cycle includes facilities for programmed die core pull sequences and for lubrication. The action of the semi-automatic machine shown in Figure 4 is initiated by push buttons for die clamping, followed by an automatic injection cycle with programmable electronic control of a valve governing wax flow rate. Figure 5 shows interesting internal details of a similar type of machine by a different manufacturer, indicating how a double-walled container system with circulating heating fluid can be used for precise regulation of the wax temperature right up to the injector nozzle. Yet another type of machine is shown in Figure 6. This highly flexible unit incorporates dual work stations, dual injection and dual die shuttle S tirre r tan k

Fie 5

A wax heating system within an injection machine (Courtesy of P. Leveringhaus Schott, Schott GmbH).

Pattern Technology

55

Fig 6 A dual system wax injection machine operating on the shuttle principle (Courtesy of Temper aft).

tables. It can employ either one or two operators and up to four dies for larger or small patterns, with two different waxes if so required. These examples indicate the scope and range of plant now available for quality pattern production. WAX PATTERN PRODUCTION AND THE MONITORING OF FAULTS If problems with wax pattern production are being encountered, it is very important to consider with the wax supplier a number of fault guidelines. The most common faults encountered during wax injection are as shown in the following list:1 2 3

Flow lines Trapped air Lubricant marks

Investment Casting

56 4 5 6 7 8

Chill breakthrough Incomplete coverage of chill Surface finish (orange peel effect) Misrun Cavitation

Flow lines are usually associated with:a Cold die b Cold wax c Incorrect injection pressure d Incorrect flow rate setting e Injecting a thick section through a thin section 2 Trapped air is usually associated with:a Wax too hot — causing turbulence during injection, b Flow rate too high — the wax flowing into the die faster than the air escaping through the joints. c Air entrapped in the wax in the machine, causing air bubbles to be injected with the wax. d Air entrapped in the patching wax when filling in slots in ceramic cores. 3 Lubricant marks occur with over-lubrication of the die, allowing wax to push lubricant into the folds/creases, giving the appearance of flow lines. 4 Chill breakthrough is associated with:a Chill too large b Distorted chill c Chill too small (floating to one side) d Pips missing from the chill e Sinks on the chill in pip location area f Chill movement due to force of wax, especially if located near the sprue 5 Incomplete coverage of chill is associated with:a Too much lubricant on the chill b Trapped air around the chill (injection rate too fast) c Insufficient injection pressure 6 Orange peel effect is associated with:a Die too cold b Wax too cold c Insufficient injection pressure 7 Misrun is usually associated with:a Cold wax b Cold die c Injection rate too low 1

Pattern Technology 57

8

d Wax flow restriction in the die, predominantly with thin wall sections Cavitation is usually associated with:a Die temperature too high b Wax temperature too high c Insufficient injection pressure d Sprue too small e Sprue in wrong position f Chill left out of die g Chill required h Injecting a thick section through a thin section

This long list only highlights the many variables that exist in wax injection technology and the object is to illustrate how important it is for the foundry, with or without the supplier's assistance, to check each area thoroughly before assuming that the problem is simply due to the characteristics of the wax.

POSSIBLE FUTURE TRENDS In the future the industry is likely to become more sophisticated and therefore wax and its quality control will increase in sophistication also. The balancing factor will be cost, as there is obviously a limit to what foundries will pay for a wax, depending on its application. It would be inevitable that if we asked a question of what a foundry will want from a wax in the future, the majority would specify a low price, high quality material that can be reclaimed. In other words, there would be no real change from past or present needs. In a competitive world it would be good to think that wax manufacturers would aim to achieve this ideal. However, the reality is that with increased emphasis on understanding properties, quality and quality control, a compromise must be made on cost, depending largely on the nature of the casting to be produced, the process used and the market the foundry is operating in. Let us consider in more detail some of the trends that could develop in the future. 1. Quality control of investment casting wax As the industry has developed, so the importance of quality control of all materials has grown. With further, future development it would seem essential that even greater emphasis should be placed on quality control of wax. In a previous section the properties of wax and their importance to pattern production were discussed. Now it can be added that it is

58

Investment Casting Table 3. Parameters in pattern production

Wax temperature in injection machine Nozzle temperature Die and/or platen temperature Injection pressure Flow control Injection and hold times etc. equally important to monitor these properties, by both manufacturer and foundry using a strict quality control procedure. For the foundry this ensures that the material purchased is within the specification issued, or as agreed with the manufacturer, and will therefore produce patterns as good as those produced from the previous batch of material supplied. For the manufacturer, it will ensure that the material is within specification and that the correct compounds have been blended. In the UK there is growing emphasis on the quality system BS 5750. Certainly such systems put quality control on a much higher level and aim to ensure that all products meet the necessary specification. There are similar systems in many other countries and we are likely to see a much greater emphasis on their application in the control, not just of wax, but of all products and processes within the industry. It is appreciated that some companies are operating sound quality control procedures now, but there will be much wider emphasis on this in the future. When a foundry produces wax patterns it will usually do so against set machine and die parameters for specific patterns. Examples are shown in Table 3. It goes without saying that if there is a variation in material specification, such as congealing point, penetration or viscosity, and the foundry has not been informed, then considerable time can be wasted producing reject patterns before the machine variables are adjusted satisfactorily. A concise quality control system should overcome this and help to reduce any wasted time and cost. Most associations/institutes issue their own recommended test methods. They are sometimes varied by individual manufacturing com panies, but as long as the particular manufacturer and foundry are looking at the same test methods, this is not critical. The wax tests recommended by BICTA include those shown in Table 4.

Table 4. Tests for wax quality

Melting point (drop point) Congealing point Ash content Penetration Viscosity

Pattern Technology

59

Melting point (drop point) and congealing point The melting point is defined as the temperature at which a drop of the sample detaches itself from the main bulk. As the melting point is closely allied to the congealing point test we can deal with them together, but first the definition of the latter, as that temperature at which molten wax, when allowed to cool under prescribed conditions, ceases to flow. The results give different temperatures but for practical purposes they give a picture of what is happening to the compound. Most important is that for the customer they give a guide to temperatures required in the injection machine tank and the injection temperature itself, whereas for the manufacturer they provide a further check on materials used. Ash content No definition is required for represents the percentage of compound and provided that accepted by the customer and

ash content as this is self-explanatory. It non-combustible solids contained in the the figure is below the required limit, it is manufacturer.

Penetration The penetration of a wax compound is defined as the distance in tenths of a millimetre that a standard needle penetrates vertically into a sample of the material under fixed conditions of loading, time and temperature. Penetration thus gives the customer a guide to the hardness of the wax. If the penetration figure has increased but is still within the limit, then the compound is slightly softer, and it may be necessary to increase the hold time in the die to maintain pattern dimensions. If the penetration has decreased then the converse applies. For the manufacturer the test is again a further check on the materials used. Viscosity (kinematic and dynamic) The kinematic and dynamic viscosities are defined as follows:Kinematic viscosity is a measure of the time for a fixed volume of liquid to flow through a capillary. In the SI system the property is expressed in m2/s, but a widely accepted unit is the stokes (St), which has the dimension cm2/s. In the petroleum industry kinematic viscosity is usually expressed in centistokes, so that 1 cSt = l_St = 1 mm2/s.

Too

Dynamic viscosity is numerically the product of the kinematic viscosity and density of the liquid, both at the same temperature. The SI unit is Ns/m2 but the poise (P) is often employed, where IP = 0.1Ns/m2 = 0.1 kg/ms = lg/cm s.

60

Investment Casting

For Newtonian fluids, the absolute (dynamic) viscosity is defined as a quantitative measure of the tendency of a fluid to resist shear. The results of these tests give the customer a guide to the flowability of the wax, the pressure required to transfer wax from machine to die and the size of the injection channel required to maintain the pressure applied. Again for the manufacturer they are a further check on materials used and the general properties of the wax. Finally, there are a number of other tests sometimes applied to a wax. These include dimensional, volumetric contraction/expansion, linear contraction/expansion, strength and specific gravity measurements. 2. The choice of wax and changing wax In the past it was invariably the case that once a foundry had chosen to use a particular grade of pattern wax they would tend always to use that wax. This was to avoid the risk of dimensional variation of patterns, coupled with a basic fear of change. Pattern wax compounds have been chosen by foundries for numerous different reasons, for example historical, the wax having been the only suitable compound at the time, recom mendation, or copying of another foundry. It is not advocated that a foundry should change its wax for the sake of changing. There are many foundries content with their existing wax. There must be fundamentally sound reasons for wanting to change, such as superior quality and quality control, increased production from a quicker setting wax, less cavitation, low price, better service from the supplier, or new injection machines with different injection criteria, and the list must be much longer. As mentioned before, it has always been a difficult decision to change wax, but now, with a better understanding of materials and a close liaison between foundry and supplier, the process is easier. It is possible for the supplier to develop wax compounds with a foundry's specific requirements in mind, and in the majority of cases to submit a wax that meets these requirements. In looking to the future it is conceivable that more foundries will consider this option in their efforts to satisfy requirements of quality and cost. 3. Materials for the future There are from time to time discussions about alternative pattern materials to wax. Polystyrene, expanded polystyrene, and urea are used. As mentioned earlier in the chapter, casting wax blends are complex com pounds of many different components. Wax is a loose definition of a large class of chemical compounds and it would be difficult to see other mater-

Pattern Technology

61

ials totally replacing these. It would seem more likely that as the industry moves forward, so wax manufacturers will continue to work with foundries so as to expand their existing knowledge and produce further waxtype materials to suit specific requirements. We have discussed the three main categories of pattern wax available today — straight or unfilled, emulsified and filled. While there will be modifications to certain of the blends to meet the specific requirements discussed earlier, no doubt waxes in these categories will continue to be used in the applications to which they are most suited. However there may well be a greater tendency, for certain applications, to change from one category to another. For example, foundries manufacturing large, thin -walled aluminium castings may prefer a low viscosity filled wax compound for greater dimensional stability of the thin walled patterns, whereas foundries producing numerous commercial castings may tend towards the use of straight or emulsified wax. Again it becomes a com promise between quality and price, depending on the market the foundry is serving. In recent years more emphasis has been placed on filled wax com pounds and their possible applications. We could see a greater trend in this direction in the future and it will therefore be useful to make a few further points concerning these materials. In the earlier categorisation of wax, stress was laid on the importance of the filler properties of particle size, its insolubility in the base wax, low ash, specific gravity and the fact that it should be inert to the process. Most fillers used today are relatively well known compounds and have been used extensively for a number of years. New materials may come along, but provided that the filler powder achieves the properties mentioned, it is the base wax that will influence the wax compound in other areas such as congealing point, viscosity, ash, penetration, setting rate, oxidation stability and also cavitation, in conjunction with the filler. Filler just blended into any wax compound will not necessarily contribute to lower cavitation or stability unless it is suspended in the correct base. If, therefore future consideration is given to greater use of filled wax materials for certain applications, it is important to put emphasis on the com pound as a whole and not simply on the filler. Finally, for some foundries, reclaiming of the filled wax may be critical and discussion with the supplier on the various options open would need to be analysed from both cost and quality viewpoints. 4. Reclaiming and reconstitution Earlier in the chapter it was highlighted how important reclaiming is to the industry. Traditionally investment casters have tended to use reclaim

62

Investment Casting

wax mainly for runner systems, or for certain patterns when using an unfilled or emulsified wax. Now, with the advance,in reclaim technology, coupled with strict quality control measures, it is possible for a foundry to consider the use of reclaim and reconstituted wax irrespective of whether they were unfilled, emulsified, or filled wax. Such technology offers a foundry the opportunity of considering its autoclaved or used wax being reclaimed and reconstituted within a specification for virgin wax. When following this route a number of critical points need to be considered. 1) It is necessary to ensure that there is only one base wax material in the system. 2) It is unsatisfactory to mix different pattern wax materials. 3) A separate runner wax should not be used in the system. 4) All wax for reclamation should be processed at one reclaimer's plant to avoid contamination. 5) It is important to have a general appreciation of wax reclamation and quality control. A foundry must develop controls on the quality of the wax it generates for reclamation and reconstitution. For example a) Waste products must not be mixed with the wax. b) The amount of silicone used should be reduced as far as possible. c) Water mixed with the wax should be minimised. d) A filter cloth placed over the autoclave tray can prevent ceramic sand from entering the wax during dewaxing. e) The size of autoclaved blocks should be considered for easy packing and optimum use of transport. f) The wax blocks should be strapped and wrapped to reduce the chance of contamination in storage. If such guidelines are adopted, and by working closely with the wax reclaimer, a foundry can have large volumes of autoclaved or used wax reclaimed and reconstituted to a specification for virgin wax. With economic and environmental considerations likely to be im portant, future emphasis on reclaim and reconstitution of wax is also likely to increase. 5. Cost considerations Obviously it is difficult to make predictions on cost trends. Cost has already been stressed as a limiting factor but only a foundry can say what the limit is. Perhaps the matter could be viewed in the following way. With existing wax compounds on the market it is unlikely that major reductions in cost will occur. The formulae would not normally be

Pattern Technology

63

changed otherwise they would become different compounds. If any major change is made then under a strict quality control system such a change should be notified to the foundry. The raw materials used in the compounds are unlikely to exhibit large decreases in cost unless extraordinary circumstances prevail; they are, unfortunately, more likely to show increases over a period of time. However, cheaper wax compounds can be formulated and supplied in certain cases. In the section concerned with changing a wax it was pointed out that wax could be designed with specific requirements in mind. Lower cost raw materials can be looked at with the aim of main taining the major characteristics of the wax, but the cost saving needs to balance the cost of testing and changing. Two examples can be considered in which changes in formulae were made to counteract high cost. Firstly, in the 1970s, carnauba wax was widely used in wax formulations. There was an acute shortage and consequently the price rose astronomically, having a great effect on the cost of certain casting wax compounds. Substitute compounds without carnauba wax were manufactured, approved and used by numerous foundries, thus overcoming some very large cost increases. When carnauba wax returned to a lower price, so the original compounds could be reduced in price. Secondly and more recently, polystyrene filler, used in numerous filled waxes for many years, rose steeply in price due to high increases in price of the feedstock of raw styrene. The various options were to pay a much higher price for filled wax using the material, to use a compound with a substitute filler material at a much lower cost, or to consider reclaiming and reconstituting the used wax. Some foundries opted to change to the lower cost compound or the reconstituted wax. What is now being stressed is that if an increase in cost reaches a limit the foundry cannot tolerate, then by working with the supplier it is often possible for cheaper alternatives to be offered. However the overriding object should be not to detract from the quality of compound needed by the foundry to produce patterns successfully, especially with the increasing emphasis on quality and quality control as the industry moves forward. CONCLUSION As demonstrated in the above review, both materials and equipment for the production of investment casting patterns have undergone major developments since the early days of the process, and patterns are produced with the aid of highly advanced plant and process controls to ensure reproducibility of properties and dimensions. Some of the economic

64

Investm ent C asting

factors influencing the choice of materials and practice have also been discussed: both these and further technical developments will no doubt continue to affect the nature of pattern production in the future.

4

Investment Materials and Ceramic Shell Manufacture D. MILLS

INTRODUCTION The ceramic nature of the mould in investment casting is crucial to the process and lends itself to a wide variety of casting applications and an even wider selection of alloys. There are, however, certain characteristics of the process which require special attention. Waxes have quite large coefficients of thermal expansion, whereas ceramics have low expansion coefficients; alloys are intermediate between the two. These differences create imbalances in the manufacture of castings which need to be addressed and controlled. One consequence of the differential expansion of wax compared with ceramic, is that if, at any time during the 'mould build' stage, the ambient temperature should rise then the encased wax pattern will crack the brittle ceramic surrounding it. Many large coating rooms are temperature controlled to avoid this potential problem. The act of melting the wax from the mould may even have prevented the development of the shell mould process because, if the mould were placed in an oven to melt out the wax, it would always crack. Special procedures have been developed to avoid this and will be discussed later. As the metal solidifies and cools, a contraction stress may be imposed on certain geometries of casting because the mould will contract at a much lower rate. This differential can damage the casting, giving rise to the well known hot tearing or hot cracking defects. These mould strength related defects are encountered in many types of mould system and are not peculiar to the ceramic shell mould. Nevertheless, control of mould strength is an important feature of the investment casting process. Indeed the requirement that the mould should be strong, to avoid breakage in general handling and resist the expansion of the wax as it melts, conflicts DOI: 10.1201/9781003419228-4

66

Investment Casting

with the desirability of a weaker mould to avoid tears or cracks in the casting and to achieve easy removal of the shell after casting. Another characteristic of the process is that of dimensional contraction. The metal shrinks slightly as it cools, the ceramic shell mould shrinks slightly as it sinters during its firing stage, and the wax pattern also shrinks after injection. The final size of the casting is therefore not exactly the same as that of the wax pattern. This also implies that the larger the component, the greater the size reduction in real terms and the more difficult it is to control it to close tolerances.

BASIC SHELL BUILD It has already been said that the ceramic shell is built up around the wax pattern assembly. This procedure is summarized in Fig. 1 and involves the application of a number of separate layers or 'coats', usually between

(a)

Fig 1 Manufacturing the ceramic shell mould, (a) Wax assembly dipped into ceramic slurry and drained to give an even wet coating (b) Stucco applied with coarse ceramic grit and the layer hardened; 5-15 layers produce a suitable shell

Investment Materials and Ceramic Shell Manufacture

67

five and fifteen according to the required strength of the mould. A large, heavy casting obviously needs more layers to contain the molten alloy than a smaller one. Each layer is produced by immersing the pattern assembly in a ceramic slurry of paint-like consistency. The coated assembly is drained for a short time and manipulated to give as even a covering as possible, free from drips and runs. Before the coating has had time to dry, the surface is sprinkled or 'stuccoed' with a coarse ceramic grit, usually by using a simple machine to supply a continuous 'rain' of grit through which the coated assembly is passed, while applying both rotary and transverse motions to ensure that all surfaces are covered as evenly as possible. The grit adheres to the wet ceramic coating, which is then hardened before repeating the sequence and thus building up the shell thickness layer by layer. Although additives are often used to give improved rheology to the slurry, it is basically composed of a hardenable liquid binder with a filler of ceramic powder. It should be noted that the first or 'primary' coat applied to the wax will ultimately be in contact with the molten alloy, and its ingredients therefore often differ from those of the secondary or 'back up' coats. These differences apply both to the type of ceramic used as a slurry filler and to the nature of the binder liquid which, after hardening, will cement the particles together. Binders are usually made with silica, itself a ceramic material, but these silica binders can either be water based or alcohol based. The water based system is usually air dried after coating and is almost universally used for the primary coat. The slower drying of the water based, as compared with the alcohol based, silica binders is useful in that it allows sufficient time for the manipulating operation to ensure a smooth and even coating, and total surface coverage at the stuccoing stage. The evaporation rate can influence the quality of the primary coat and ultimately the surface finish of the casting — emphasising the importance of controlling the working environment with respect to both temperature and relative humidity. Alcohol based binders not only dry at a faster rate but can be hardened by exposure to an ammonia atmosphere. These binders are used extensively in secondary coat formulations, particularly when coatings are being applied in robot units. With alcohol based slurries and ammonia hardening, coatings can be applied within minutes of each other. Second ary coats are required to build up the total shell thickness, and the choice of materials in combination influences the bulk properties of the shell system. Water based binders are also widely used in back-up slurries, by foundries which prefer their more environmentally and user friendly

68

Investment Casting

nature to the rapid-drying advantage of the alcohol based system. The need for flameproof electrical wiring and motors in the coating shop is avoided, and hazards to operators reduced, by the exclusive use of water based binders for both primary and secondary coats. Materials used for the slurry filler and the coarse stucco grits may or may not be the same. Ceramic materials are used in a very wide range of combinations and include silica sand, alumino-silicates, alumina and zirconium silicate. The choice is governed by availability, cost, and foundry performance. Alumino-silicates range from fired, crushed and ground clays, to mullite or sillimanite. Zirconium silicate, or zircon, is used widely, particularly as a primary coat filler, because of its excellent high temperature inertness and stability. Zircon exists naturally in the form of a fine sand, mined in various parts of the world. It is sometimes used for the fine primary coat stucco as well as, in a pulverised form, as a filler material. The particle size of natural zircon sand is too fine for its convenient use as a secondary stucco, for which other and coarser synthetic materials are preferred. In addition to refractory sands, synthetic materials are sintered or even fused in the manufacture of stuccos, followed by crushing and sieving. Further grind ing of remaining material will produce the powders used in the slurries. The range of materials used for stuccos is thus also available as a choice for slurry fillers. Some indication of the complexity of processing involved in the production of these types of refractory materials is given in Fig. 2. A high degree of mechanisation is possible for the shell building operation and fully integrated coating 'cells' are used in all the large enterprises throughout the world. A cell contains a robot which will handle the dipping and stuccoing operations, continuous mixing tanks for the slurries, conveyors for transporting the mould from one operation to the next and a suitable means of rapid drying for water based binders, or an ammonia cabinet for hardening alcohol based slurries. An example of the integrated cell concept is illustrated in Fig. 3.

RAW MATERIALS, THEIR PROPERTIES AND CONTROL All ceramic shell moulds are built up from three components, the binder, the filler and the stucco materials. Binders are divided into two groups, water based and alcohol based, and various ceramics are used for the filler and stucco. Choice of binder and ceramic materials determines the properties of the whole of the shell. Certain characteristics are of particu lar importance and warrant full discussion, because without a full understanding of properties and of the interactions between any combinations

Fig 2 Complex processing stages are required to produce high quality and consistency of product in the manufacture of stucco materials. Molochite processing procedures (Courtesy ofECC International Limited, St. Austell.)

70

Investment Casting

Autom ated investment coating cell AMMONIA TUNNEL PRIMARY STUCCO SANDER

UPPER CONVEYOR - 3 0 STA TIO N S

MEZZANINE FLOOR

LOWER CONVEYOR - 62 STATIONS CELL LOAD/UNLOAD AREA SECONDARY STUCCO SANDER STRUCTURAL MOULD STATION WAX WASH TANK

SECONDARY SLURRY TANK

C EIL CONTROL ROOM

PRIMARY SLURRY TANK

Fig 3

A fully automated coating cell for the manufacture of investment moulds in the Company Research and Development foundry of Rolls Royce pic at Patchway, Bristol. Notice the slurry tank covers which only open as the mould is dipped. This avoids solvent evaporation and slurry contamination. (Courtesy of Rolls Royce pic).

of materials, control tests might be irrelevant to shop floor performance, and foundry problems and casting non-conformance would be difficult to understand and correct. In the author's opinion this section covers by far the most important aspect of the manufacture of investment castings, because casting dimensions and surface finish can clearly only be as good as the mould into which metal is poured. Some other aspects of non-conformance directly related to mould behaviour are not concerned with finish or dimensions. Thermal characteristics of the shell will influence alloy solidification, and mould strength relates to other casting defects. Lack of understanding of the fundamentals of the shell and its raw materials will lead to an uncontrolled process. This in turn will lead to a system in which the shell properties will vary from day to day. Casting

Investment Materials and Ceramic Shell Manufacture

71

non-conformance will also vary, and in the extreme case it will become difficult to plan the production of castings, because it will never be certain how many defective components will be produced. In fact a good measure of total foundry performance can be obtained from the day-to-day or week-to-week casting yield and its fluctuations. Large fluctuations re quire a closer look at the shell mould manufacturing procedures.

Water Based Silica Binders and the Colloidal State These binders are correctly referred to as 'aqueous colloidal silica sols'. The significant colloidal state lies between a solution and a slurry. A solution is a combination of a solvent and another substance dispersed within it at a molecular level. A typical foundry slurry is composed of a liquid and finite particles of ceramic powder, although these are very small. Slurries are visibly opaque and the particles of powder will settle or 'sediment' within the liquid. Solutions, on the other hand, may be coloured but are transparent and no sedimentation occurs. With an intermediate particle size, too big to give a true solution and too small to sediment, the appearance would be between that of solution and slurry, i.e. it would be semi-transparent or opalescent. An aqueous silica sol is an example of this. The colloidal state also has its own unique properties, not found in either slurries or solutions, and to differentiate the special behaviour of colloidal solutions they are referred to as 'sols'. In the mould binder application of sols the particles are spherical and just big enough to be seen under an electron microscope. A typical microstructure is shown in Fig. 4. There is no sharp dividing line between solutions, sols and slurries. The larger sizes of molecule and the smallest particle sizes of colloid are about the same, as, for example, with plastic polymers or resins and alcohol based silica binders. A major distinction between alcohol based and water based binders is in the relative sizes of the colloidal particles. Water based sols can be manufactured with a range of particle sizes, the foundry industry generally using very small sizes between seven and thirty nanometres in diameter. Aqueous colloidal sols with larger particle sizes (around two hundred nanometres in diameter) can be obtained for other applications. With these larges sizes we can start to detect sedimentation, which represents the border line between sols and slurries. The main unique characteristic of all colloids is that they can convert from a water-like consistency to a jelly -like substance, which is, in fact, the means by which the slurry layers are hardened. Silica sols are chosen because this jelly will eventually dry and convert to silica, which is a refractory in its own right. Additives similar to emulsion paint binders

72

Investment Casting

Fig 4 Highly magnified view of water based colloidal silica sol showing its spherical shape. Each particle carries a small repulsive charge in the sol condition. (Courtesy of Bayer UK Limited, Newbury , Berkshire.)

are sometimes included to enhance the strength of the unfired mould and to produce a slurry that is easier to apply and less liable to flake or spall later. The term for this conversion from a liquid to a jelly is sol-gel and in the field of engineering ceramics this has become very important as a means of producing and shaping high purity material. The investment casting industry had been using sol-gel technology to produce moulds long before the potential of these colloids was appreciated in other ceramic fields. The Hardening Process In the liquid or 'sol' condition the spherical particles are freely circulating through the liquid and are kept from colliding by each carrying a small electrical repulsive charge. In silica sols all particles are charged nega tively. Although this state is highly stable (binders of this type are stable for at least a year and some have been kept for twenty years), the finely balanced charges can be easily upset, so that some of the repulsive charges which keep the particles apart are reversed, causing them to attract each other. Particles will then start linking together into a loose, open, three-dimensional network throughout the liquid, and when this

Investment Materials and Ceramic Shell Manufacture

73

structure is well established but not complete we can see the physical change to a jelly -like consistency. The electric charges of the particles are determined by the pH of the liquid medium, in this case water. Aqueous silica sols are at a pH of around 9.5 (the most stable condition) and any change, by introducing material contamination, will alter the pH and accelerate the linking process. Sol-gel conversion can be very slow or very fast — less than a second but never instantaneous. Slight changes in pH invariably happen when mixing the ceramic filler with the binder. Slurries have a shorter life than that of binder in an unopened drum because of the introduction of im purities from the filler and the surrounding air. The Ageing Process This slow interparticle linking is the actual mechanism of what is often called 'ageing' and gives rise to important consequences in the behaviour of slurries on the shop floor, and in the properties of the moulds subsequently manufactured from them. Apart from a straightforward change in pH, more subtle effects can be demonstrated. Salts which dissolve to form true solutions dissociate on a molecular level into charged particles or ions. The fact that other electrical charges have been introduced into a finely balanced silica/water environ ment also contributes to binder degeneration — an apt term for the 'ageing' process because the slurry properties degenerate and the strength of the mould is reduced. Changes of pH can arise from the use of fillers that have an acidic nature. The rate of gelation is at its maximum around pH 7, neutral. Carbon dioxide from the surrounding air can dissolve to form carbonic acid. Slurries should, therefore, always be covered when not in use to avoid this form of contamination. Incidentally, carbon dioxide gas in high concentration will rapidly gel water based binders, but this is not used in the manufacture of ceramic moulds (in the way that ammonia is used to gel alcoholic slurries) because the larger size of colloidal particle as com pared with that of alcohol systems would not give sufficient green strength to the mould. Degeneration from contamination by soluble salts is a real possibility. Zircon slurry problems have been known to have occurred because the zircon was transported to its destination over stormy seas which caused salt spray contamination. Knowledge of the basic chemistry of the colloidal state can explain the behaviour of water based binders and may enable the manufacturer to use a slurry for up to a year without prob lems, particularly if a quantity of slurry is discarded at regular intervals and fresh additions made. By operating in this way the degeneration of

74

Investment Casting

a slurry can be contained and consistent properties of the mould assured. It has been mentioned that the condition of the binder significantly affects the strength of the shell. Any ageing effect may go unnoticed until the binder has almost completely gelled. Long before this obvious condition, however, shell mould properties, particularly strength, degenerate throughout the operational temperature range. This effect is obscured and confused by the fact that slurry is being used and replenished all the time. High volume production can offset any real problem by regular replenishment. But in waiting or slack periods (particularly if coupled with a slight increase in contaminants) shell strength can deteriorate and there could be a sudden unexplained increase in de-wax cracking defects, or a mould bursting when the metal is poured. Even more subtle effects can be observed with a slurry that is out of condition. Viscosity of the binder can increase, and if the slurry is being controlled by viscosity, more binder or less filler will be added than previously because this unnoticed change in the viscosity of the binder has altered the rheology of the slurry. Water based shells are invariably air dried and in this case the evaporation of the water from the binder effectively concentrates the silica. Most foundries usually employ a content of between 15 and 30wt% of silica in the binder. Manufacturers are even able to produce sols with a 60% silica content by dispersing, in a watery medium, a high packing concentration of spherical particles, out of contact and suitably charged for mutual repulsion, thus avoiding collisions. However, these high concentrations easily become unstable as various impurities are introduced, and the solto-gel process, or ageing, is slowly initiated. In the air drying stage, depending on the condition of the binder, the maximum concentration achieved as evaporation proceeds will be different, with aged binders gelling prematurely at lower concentrations. Fresh binder will thus achieve a higher silica concentration than an aged binder and will pro duce a stronger bond. Mould strength reduction over a period of time is generally related to this slurry condition. A method of assessing this is to evaporate a slurry of known volume at room temperature by reducing the air pressure and measuring the volume at gelation. Water based slurries that are out of condition will also appear to dry faster than fresh ones. They are, however, actually hardening with a greater concentration of retained water, so that the linked up structure of the silica particles will be looser and hence weaker. (If we were to attempt to gel a slurry by gassing with carbon dioxide the gel would similarly contain a high proportion of water and produce a very weak structure). On the shop floor this aged condition can be detected at the stucco stage, because the premature 'drying' effect can prevent full adhesion and

Investment Materials and Ceramic Shell Manufacture

75

coverage of the stucco, even with consistent draining times. At thin edge features such as those found on the trailing edges of turbine blades this lack of stucco adhesion, if unnoticed, will also lead to locally reduced shell thickness and weakness that can cause crack propagation later. If such a condition is observed, the only possible corrective action is to discard the slurry and start again.

Drying the Shell Mould Concentrating the binder to induce gelation by air drying is only the first step in the total sequence. After the initial gelation further linkage be tween particles takes place, i.e. the observed gel point (which can be induced by adding a salt or altering the pH) and the structure contracts for some time. Practical tests on specimens indicate that this could be as long as a day after the first hardening stage. This process is also accompanied by the displacement of water, which exudes from the gel even if a specimen of shell is immersed in alcohol, for example. Water is forced out of the material as the linkage proceeds further and the structure starts to shrink. This shrinkage is another characteristic of the gel stage and invariably causes the gel to microcrack. Clearly the ultimate binder strength is related to the degree of disruption by cracking, and the strength of the shell is affected not only in the green state but later at elevated temperature. This also explains why a sol which will gel at high silica concentration by air drying will also give a stronger mould — because less shrinkage occurs after gelation. It must be remembered that the gel is physically binding all the shell ceramic materials together, both in the green state and subsequently at elevated temperature. With normal drying procedures it should be emphasised that most of the water is removed from the binder, which changes to a brittle, porous condition. Pores within the loose silica structure still tenaciously hold some water. Indeed this water will not be completely removed until a temperature around 1000°C is reached — a highly relevant fact for later consideration. The silica gel in a dried green mould is the same material as that used as a desiccant. It is capable of absorbing moisture which will, as with a desiccant, be driven off at about 300°C or above. At this stage we have a green mould with properties relating directly to the binder condition and also to the impurity contents of the filler and stucco. Other features of the ceramics also play some part in determining the mould condition at this stage. The process of binder shrinkage and gel cracking will also be influenced by the ceramic particle size distribution. Usually the gel cracking is restricted to microcracks because the presence of stucco prevents larger cracks forming.

76

Investment Casting

Another important contribution to the drying process is the wettability of the ceramic with the water in the binder. This will also affect the ultimate microstructure of the ceramic shell mould. It is very enlightening to observe the drying mechanisms under the microscope. A simple experiment will demonstrate the features described above. Take a glass microscope slide, cover one side with adhesive tape, immerse the slide into a slurry, drain and then strip off the tape so that only one side of the slide is coated with wet slurry. The slide can be observed at fairly low magnifications around 30-50X. Shrinkage of the binder and gelation can be seen, during which particles of filler are rearranged and small 'holes' in the film are formed — one reason why moulds are porous. This observable process of water loss, gelation and shrinkage is complex and any change in external conditions can affect the ultimate properties of the green and fired shell mould. Ethyl Silicates Whilst it is usual to employ a water based slurry for the primary coat, either water or alcohol based binders can be chosen for the back-up coats for the reasons already mentioned. Pure ethyl silicate can be used to produce a foundry binder but it is more usual to employ a condensed or concentrated form containing about 40% silica by weight. Many of the available commercial materials have '40' in their descriptive code or title. Ethyl silicate 'as received' has no binding properties, but must be chem ically decomposed by reacting with water. The reaction produces alcohol and silica in an active state. This colloid has a much smaller particle size than that in water based sols and the particles are not spherical, their branched chain structure being more akin to that of partly polymerised resins. Technically, therefore, this condition can be called 'polymeric', although because alcoholic binders exhibit the property of gelation they can also be considered as colloids. The chemical process of hydrolysis, or release of silica particles, is always accelerated by the addition of a small amount of acid, which acts as a catalyst. The acid is added to bring the pH of the liquor to about 2, at which there is a further condition of sol stability. Changing the pH of the liquor to the unstable condition of about 7 will cause gelation. Ammonia gas achieves just this. Unlike the water sol-gel mechanism of spherical particles linking up to form a loose structure, in this case the polymeric silica, in short chains of molecules, starts to link up to form a three-dimensional structure such as occurs in the curing or polymerisation of a resin. The pH 2 value is significant because it is the 'isoelectric' point of this system, at which particles contain no charge, (cf. aqueous silica

Investment Materials and Ceramic Shell Manufacture

77

sols, which are negatively charged.) This feature gives rise to some of the differences in the properties of an alcohol based sol, the main one being that since there are no repulsive charges present to prevent particles colliding and linking up, this will occur, although very slowly, at pH 2. As soon as we start to hydrolyse ethyl silicate we start the ageing process, because silica molecules are linking together to form twins and triplets and so on. Chemically, the resulting particles are called dimers or trimers etc., representing two, three or more linked silica molecules. Sophisticated methods of analysis have been developed to identify the species or chains at any point in the life of one of these sols and, as might be expected, external influences can alter the rate of polymerisation. High concentrations of silica and high temperature are particularly effective in creating rapid ageing in a slurry. For these reasons alcohol should always be added to replace any evaporation from the tank, to control concentration and also counteract the warming of the slurry due to heat from the tank drive unit. Lids are recommended on the slurry tanks to suppress evaporation. An example to demonstrate the effect of concentration would be to use hydrolysed ethyl silicate with a silica concentration of say 25% to produce moulds on the first day after preparing the slurry, and again after one week. Reduction in shell strength can be as much as 40% with shells manufactured after one week, due to slurry ageing. This is why the practical operating limit for a 'straight7 alcoholic binder is about 20% silica, and 15% is even better if the resulting reduction in general strength is acceptable. Various chemical modifications have been made to improve the stability of these binders. A strict and regular routine of diluting ageing slurry with fresh additions in periods when slurry is not being used is beneficial to the consistency of mould strength. A further consequence of the absence of electric charge at pH 2 is that, unlike aqueous sols, alcoholic sols are far less sensitive to the addition of salts which dissolve in alcohol to form ions. The author once made up a 20% sol and added some copper chloride which coloured it blue green and this remained stable in a bottle for two years, whereas adding the same salt to a water based sol would have shortened its life to a few minutes. This implies that salts can be dissolved in the binder from the filler impurities (particularly as there is hydrochloric or sulphuric acid present to dissolve some of the oxide impurities) with some impunity to the balance of the system, provided that any spent acid is replaced to maintain the binder at the isoelectric point. Alkaline impurities will alter the pH and accelerate the ageing process. Particular attention must be given to avoid contamination by ammonia, either from the air or from a

78

Investment Casting

previously gelled coat which has had insufficient time for the ammonia to clear. Although dissolved salts are less important at this stage of the process, their retention in the gel will modify the binder's high temperature behaviour, particularly its refractoriness.

Prehydrolysed Alcohol Binders It is a peculiarity of water based colloids that they can be changed from pH 9.5 to pH 2 so that the particles are stripped of their charge. This can only be accomplished by moving very rapidly through the highly un stable pH 7 region, taking advantage of the finite time it takes to gel. This is done by adding acid rapidly with intense stirring. The sol then behaves in the same way as an alcoholic slurry. It is not, however, possible to make the alcohol based silica alkaline without immediate gelation. This is not to say that this might not become possible in the future. It took a long period of research in the early days of colloidal silica to develop a process that could create the currently used materials without gelation. It is possible to produce a binder that is acid and contains both the polymeric species of the alcohol based system and the acidified particles of the water based system. By using a water based silica sol instead of pure water to hydrolyse the ethyl silicate, a 'hybrid' binder is obtained. This binder is alcoholic and free of water, since all the available water has been used up in the hydrolysis reaction. The stability of each system is enhanced in some respects, giving a much more stable sol. The availability of prehydrolysed ethyl silicate shows that manufacturers can now hydrolyse binders with a sufficiently long life for distribution to and storage by customers, whereas the 'straight' hydrolysed materials must be produced by the foundry as and when required. Other additions can be made to obtain even higher stability and less sensitivity to external influences, making it easier to achieve consistent shell strength. In the hydrolysis of the various silicate binders the full amount of water is never added to complete the reaction for releasing active silica; always less than the stoichometric amount is added. This partial hydrolysis gen erally improves shelf life, but makes the system sensitive to moisture contamination. Any extra water will allow further reaction to proceed, with resulting shortening of the binder life. This is an important factor to be taken into account, both in the preparation of the binder and in its subsequent use. In preparation, all water, even as impurity in the alcohols used to adjust the final silica content, must be calculated. In use, damp powders and moist air are obvious sources of potential slurry problems.

Investment Materials and Ceramic Shell Manufacture

79

Some prehydrolysed binders have additives which suppress this sensitivy to water but their formulations are usually protected by patents, or remain proprietary information. In general, the modern prehydrolysed binder produces adequate mould strength and day-to-day consistency, provided that care is taken to avoid contaminants entering the slurry.

RAW MATERIAL TESTING Detailed information on test procedures is available in the literature. The various trade associations (e.g. BICTA) publish information for guidance to users. The amount of material testing needs to be clearly thought out in relation to the manufacturing system in use. There should be a clear understanding between user and supplier as to exactly what criteria are applied in designing a reasonable and reliable series of tests. For this, a supply specification is needed which will not only describe test methods but will define test parameters and rejection procedures agreeable to both parties. The sort of areas that can cause problems are the definition of a 'batch', or of what constitutes a 'major change' in raw material production processes; clearly these aspects of quality control may not mean the same to both parties. Audit procedures also need to be identified, and the system for full interpretation of data defined. Laboratories sometimes carry out regular control tests rigorously but do not have the necessary expertise for full interpretation of the results. It is also important to define authority concerning acceptance, because where materials are found to be outside the agreed specification many pressures (for example for the delivery of castings) may otherwise overrule a decision to reject the material, with dire consequences. Regarding the actual tests, it is often the case that a foundry will develop specific tests which, in their particular situation, relate more closely to the behaviour of raw materials in the mould manufacturing stages. These tests need to be defined and agreed with the supplier. The foundry will use its experience to decide on testing frequency. It will also be of great importance to develop a data base on the raw materials used, particularly the ceramic powders and grits, because these will vary from batch to batch. These data will prove invaluable if any problem arises and powders or stucco materials are suspect. It will be of little value to obtain particle size and trace element data in a crisis situation if no previous data exist for comparison. Where foundries have full laboratory facilities, agreement with the supplier on the method of test and equipment used will ensure that both parties are measuring the same properties. Where a foundry has no

80

Investment Casting

laboratory facilities for control of incoming raw material, it should obtain as much information as possible from the supplier. Data base values may be obtained from independent testing concerns to ensure a good understanding of the materials used in the foundry, and as a double check on manufacturers' data. The most useful data on the relevant ceramic materials are the chemical and ceramic phase analyses and the particle size distribution. In the manufacture of superalloys, for example, it is essential to control iron spots in the primary coat surface, to avoid local metal/mould reactions at the casting surface, particularly with vacuum casting. Certain trace elements need to be absent or at levels of only a few parts per million; silver, lead and bismuth are examples of metal contaminants that can occur in some ceramics and which, if they find their way into a superalloy, can dramatically degrade its properties. Other elements, such as sodium, usually associated with alumina, can reduce mould refractoriness and influence the pH of the slurry. In short, some trace elements that are of no concern to other industries may cause great problems to foundries — which is the main reason why specification of raw materials is so vital to the supplier as well as to the user. It is usual to report the analysis of ceramic material as equivalent levels of oxides, although the constituents may not necessarily be present in this form and may be combined as salts. Chlorides, phosphates, and sulphates may be present. In considering the types of test that are most useful, materials can be classified in three groups: pure chemicals, formulated products and ceramic powders and grits. In the case of pure chemicals such as octyl alcohol or isopropinol a simple specific gravity value will characterise the batch, simply to ensure that the chemical supplied is correctly identified. A method of quarantine for incoming raw materials can be used to ensure that tested and approved batches are released for use. Isopropinol, which is frequently used to produce alcohol based slurries, is available at various purity levels of differing water contents; the level must be known because excess water can cause gelation in most alcohol systems. Formulated products are the most difficult to characterise because of the problem of devising simple tests to indicate potential problems in use. Silica contents of binders are widely used to confirm product consistency. Difficulties arise with formulated additives such as anti-foams and wetting agents. Here the possibility of pH change and undesirable levels of salts may, in the absence of compositional information about these additives, result in accelerated ageing of the slurry. Ceramic fillers and stuccos will vary both in particle size distribution and trace element levels. Many large foundries take special care to

Investment Materials and Ceramic Shell Manufacture

81

control and test these materials, knowing that sporadic problems with moulds will otherwise occur. Knowledge of the relationship between the powders and grits used and mould performance has improved in the past few years, owing to the introduction of equipment that can determine particle size or complete chemical analysis in a few minutes. This has meant that considerably more data have become available to foundries as an aid to the interpretation of problems. The automatic sedimentometer (for measuring particle size within a few minutes) and X -ray diffraction equipment, have proved invaluable for rapid data collection on incoming raw materials. In the early days of the foundry industry particle size determination of powders required manual sedimentation techniques taking a full working day to produce one result. It is no wonder that understanding of raw materials and their effect on mould performance has been slow to develop and is still incomplete. SLURRY PREPARATION, MIXING AND HOLDING TANKS In order to achieve consistency of slurry characteristics correct preparation and maintenance must be ensured. Because the operations involve working with dusty powders and corrosive liquids, health and safety on the shop floor also need to be considered. Having established a suitable formulation for a primary or secondary slurry, it is advisable to define procedures for its preparation and maintenance. Most water based primary coats are composed of the colloidal silica binder and a suitable ceramic filler. Zircon powder, silica, alumina and clay based alumino-silicates are all widely used in this application. With water based slurries benefits can be gained by using a non-ionic wetting agent to ensure good coverage of the wax assembly. This addition requires in turn an anti-foam agent to suppress bubbles which are liable to find their way on to the surface of the wax. Proprietary materials are used for this and it has been common practice to add octyl alcohol to the slurry, to float on its surface and suppress foam formation. The disadvantage of octyl alcohol, however, is its poor wetting capability in contact with water, so that excessive amounts can be deposited on the wax surface, resulting in poor slurry coverage. Wax washing may also be necessary to remove silicone residue from the wax injection stage, even though wetting agents are added to the mix. Other additives can be applied to improve the slurry properties in use. Careful addition of clays can help in reducing sedimentation and some latex additives produce a stronger, more flexible green bond to assist later at the de-wax stage.

82

Investment Casting

A peculiarity of slurry is that its rheology is to some extent influenced by the type of mixer used (due to the different shear characteristics of different tanks). This can become painfully obvious when a change of mixing tank is undertaken. Previous filler loadings and control viscosities often need to be adjusted to produce an acceptable primary coat. Water based secondary slurries also contain additives, but it is normal to use these slurries at a lower viscosity, so that coating and draining are less critical. With alcohol based slurries, particularly prehydrolysed materials, a simple binder/filler system will suffice because good wetting is characteristic of alcohol based systems. Another additive often used in primary coats is a grain refiner, usually based on cobalt aluminate and used to control the grain structure of the metal in critical casting applications (this is dealt with more fully in later chapters). it is sometimes desirable to make these various additions in strict sequence; for example, it is always beneficial to add filler to the liquid rather than the other way round, to ensure easy mixing. In this part of the process it is also important to realise that when slurry ingredients are mixed there is an initial period of changing rheology in the system. In particular, an early mud-like consistency very quickly changes to a smooth cream, and later the viscosity will drop to a steady value over a period of hours. As it is customary to control the slurry by monitoring its viscosity, difficulties can arise in any attempts to adjust the consistency in this unstable period. Tests have shown that the viscosity may continue to fall for at least 24 hours when using the low-shear conditions of a typical rotary mixing/hold tank. Much faster equilibrium can be effected by premixing using a high speed mixer, to increase mixing shear energy. High energy is needed to break up the agglomerated powder and to release trapped air. Incorrect premix conditions with low-shear mixers will not fully disperse the system, and cannot be corrected in the holding tank, with its own low-shear mixing action. High energy input at the preparation stage can also lead to high friction which, if not controlled, can lead to undesirably high temperature rises and more rapid ageing of silica binders. The importance of premixing, and of full and complete dispersion, cannot be overemphasised. The vital link between variations in raw materials, the slurry and the mould is still being developed and progress in this area will only continue with com mitment of the industry to regular raw material testing and the full use of the resulting data in the analysis of casting results. In other areas of ceramic technology there has been a similar recognition of the importance of mixing ingredients homogeneously, particularly in the field of engin eering ceramics.

Investment Materials and Ceramic Shell Manufacture

83

SLURRY BEHAVIOUR Although most slurry compositions are simply mixtures of binder and ceramic powder, with small quantities of additives to enhance slurry behaviour in specific ways, their rheology is highly complex and is influenced not only by the ingredients but by a number of external factors. Primary and secondary coat slurries have slightly different requirements for the manufacture of the mould. Perhaps the most important slurry is that for the primary coat, because much of the non-conformance that can be attributed to the mould is derived from deficiencies in this coat. Only in those processes where moulds are expected to hold large volumes of metal, or to withstand very high temperatures for long periods (as in directional solidification) do we see casting nonconformance, particularly dimensional control, being influenced significantly by the back-up coats. From the aspect of maintaining a stable, homogeneous mixture in the slurry tank, we have already noted some major differences in binder chemistry as between alcohol and water based slurries, that can affect changes in the slurry due to external influences. Because most primary coat formulations use a water based colloidal silica sol, only this type of primary coat will be considered, but comments will be extended to both binder systems for secondary coats. The primary coat slurries eventually provide the ceramic face of the mould in contact with the molten alloy, and as surface finish will be an important characteristic of the casting, great attention must be given to the nature of the ceramic filler. The density of the filler and its particle size distribution will affect slurry behaviour. The higher-density fillers such as zircon or alumina will tend to sediment faster than alumino-silicate or silica fillers. As might be expected, the design and agitation characteristics of the mixing tank will also influence the homogeneity of the slurry. It is obviously highly desirable in all slurries to ensure full dispersion and to avoid sedimentation and entrapped air. Because there are endless permutations and combinations of slurry composition, mixing method and external influences, it is possible to create a 'knife edge7situation, whereby a number of factors unite to create conditions in which the slurry is difficult to maintain in a homogeneous state. These critical combinations of variables can lead to sudden sporadic increases in cast scrap although 'nothing has been altered'. It is in the nature of the process that, at the mould build stage, slurry will be gradually used up and fresh material added (it has already been noted that there is a period of instability as the powders and binders disperse, which has a marked effect on the rheology of the slurry).

84

Investment Casting

Fig 5

The Malvern Mastersizer allows the characterisation of ceramic powders in a few minutes (Courtesy of Malvern Instruments, Worcester, England).

W ith the use of high shear prem ixing, the stabilisation tim e can be reduced bu t not com p letely elim inated. This im plies that unless sufficient tim e is given for rheology ad justm ent after every fresh addition to the slurry, it w ill never behave quite consistently from one hour to the next. Som e operators allow very little tim e to establish slurry equilibrium . Batch variation in raw m aterials is another source of variability and can only be elim inated by m ore stringent control of particle size distribution than has hitherto been considered acceptable. A dvanced equipm en t is available for pow d er characterization and an exam ple is illustrated in Fig. 5. In consid ering the influence of particle size distribution on slurry b e haviou r, although the coarser fractions play som e part in any sed im en ta tion effect it is the subm icron content of a filler that can change the slurry rheology to a rem arkable degree. Because the specific surface of the filler, rather than the p article size distribution, has the m ost significant influ ence on slurry rheology, specific surface values are quite suitable as a m onitor. Filler loading is an im portant consideration in the slurry form ulation but, unfortu nately, coating characteristics can be so altered by sm all vari ations in subm icron particles, that they cannot be controlled if filler sur face areas vary from batch to batch.

Investment Materials and Ceramic Shell Manufacture 85 External influences have to a large extent been discussed in relation to the chemistry of the binders. Impurities from the filler, carbonic acid from the air, and the use of tap water, can all affect water based binders — causing them to thicken and thus upsetting the balance between filler loading and draining and coating characteristics — and can also cause further problems at a later stage in the process. Good control of water based primary coats necessitates the use of distilled or de-ionised water and close attention to raw material impurities. This applies particularly if the general usage of the slurry is relatively low, so that the natural replenishment is not being maintained. In such cases it is advisable to remove and discard some of the slurry being held, and to replenish with fresh material on a regular basis. The essence of consistent shell manufacture is to devise suitable procedures and adhere strictly to them. What might be described as a 'circumstantial' system of control must be instituted, because the evidence of problems is always hidden, only to be revealed after the event. The first examination of a casting and discovery of a defect usually takes place only after shell removal and casting cleaning, when any clue to the nature of the problem has been lost. Mixing tank design and efficiency is not the only plant related variable that influences a process. Sometimes the coating workplace is without such desirable features as humidity control and/or temperature control. The former is relevant to slurry draining behaviour and any change in humidity will alter coating characteristics. Temperature variations will not only affect the long term stability of the slurry but also its rheology. Even if air controls are in place, the plant and shop layout makes a contribution to local differences in humidity and slurry temperature which are often overlooked. Heat, not only from the tank drive unit but also from the friction of mixing, can raise the temperature of the slurry by a surprising amount. If, of course, the local process instruction includes a regular slurry temperature monitor, then these problems will be eliminated. Humidity measurements in the coating work area may not reveal a local build-up of moisture vapour in some restricted spaces, and the thickness and quality of coating may well be determined by precisely where the operator stood in the critical period of draining the slurry prior to the application of stucco. Here again consistency of procedure is an essential part of a good process. The above comments apply equally to secondary coats, but here slurries are usually 'thinner' and drain faster than with primary coats; the secondary coat and stucco in combination provide the means of building up the layers of the shell. Apart from its formulation, the combination of slurry rheology and dip/drain technique of the coating will either

86

Investment Casting

produce a shell of excellent consistency or one which has uneven thickness and variable strength. It is generally considered that the particle size and packing of the secondary stucco determine the thickness for a given number of coats, but this is not necessarily true. Wetting characteristics between the stucco and slurry play a significant part, as does the condition of the wet slurry when the stucco is applied. Many parameters could be listed which control mould thickness and consistency, but finding those relevant to a particu lar problem may not be easy. One significant feature regarding wetting between slurry and stucco should be noted. If two ceramic shells are produced with different stucco materials being the only difference, and both shells have as close as possible to identical particle size distributions, but one is highly wetted by the slurry and the other is not wetted to a great extent, the highly wetted shell will be significantly thicker than the other. More than one layer of stucco will adhere to the wetting combination, as compared with only a single layer in the non-wetting case. This is because, with highly wetting materials, as each particle falls and adheres, the liquid immediately coats the whole of the particle, thus allowing further particles to adhere to it and developing a thicker mould compared with the monolayer. This will affect the high temperature properties of the shell and may even modify the way in which the casting solidifies. It should also be noted that the thicker (wetting) stucco layer can be much more easily eroded on sharp edges at the next stucco operation, particularly with fluidised bed application. This again is an example of the dramatic effects that physical features of the raw materials used in com bination (and probably never really considered previously) can have on mould performance. It can be appreciated that the complexity of slurry behaviour is still not fully understood, and again operations must be conducted on a strict procedural basis, even if some of the controls are apparently unimportant. The main criteria must be the quality of casting and the cost-effective value of the many available control tests.

SLURRY CONTROL AND TEST PROCEDURES Given sufficiently elaborate test procedures it might be possible to analyse the precise condition of the slurry and be able, if necessary, to take almost ideal remedial action. In the real world, however, many compromises have to be made and only limited testing is usual. By far the most common slurry control is the so-called 'viscosity' test, preferably referred to as the 'flow' test. This is carried out by filling a specially designed cup with slurry and taking the time (in seconds) it

Investment Materials and Ceramic Shell Manufacture

87

takes to empty the cup through an orifice in its base. Unfortunately there are two different types of cup design in general use, the Ford cup and the Zahn cup, and each will give a different flow time on a given slurry. Conversion graphs should be viewed with suspicion because the practical test procedure is controlled by the complexities of the rheology of different slurry compositions, and flow times do not necessarily relate to all situations. Having once established the control level for a particular plant, any changes will indicate some change in the slurry and corrective action can be taken. However, by controlling the viscosity value without regard to other factors, the slurry behaviour may be unwittingly changed in the interests of a consistent test result - which would not have been quite the objective in mind in designing the test procedures. The first consideration is the temperature of the slurry, because any increase in temperature will reduce the viscosity or 'flow time'. The second consideration is slurry sedimentation. As soon as slurry is removed for testing, the mixing action is stopped and sedimentation starts to occur in the sample. Any delay in releasing the flow of slurry can create a partial blockage near the orifice and alter the reading. This effect will depend on the nature of the slurry and may influence the test results more in some cases than in others. Sedimentation will, however, also occur with some methods of measuring viscosity used for general scien tific purposes. The Brookfield viscometer, for example, requires significant time to measure viscosity because its rotating spindle needs to equilibrate before a reading is taken. Great difficulty may thus be found in obtaining reliable readings, due to sedimentation. The Ostwald U-tube viscometer relies upon a liquid being set at differing heights in each arm of the tube, the time to reach equal level then being determined. Here again sedimentation can result in errors in the measured value. The third consideration in testing, which has already been discussed in detail, is the ageing effect of the binder. Any increase in viscosity of the liquid portion of the slurry will modify the overall viscosity. In this event the increase in flow time would generally be interpreted as indicating too much filler loading and probably, in error, more liquid would be added to the mix to reduce the 'flow time'. Ideally, more elaborate testing should be carried out in conjunction with the flow test, to ensure that the proper corrective action is applied to the slurry. Another effect encountered in practical monitoring and controlling of slurries by means of the flow test is that of changes in the thixotropy and plastic behaviour of slurries with different types of filler. There is a great amount of literature on this aspect of slurry technology, particularly from the pottery industry where slurries or 'slips' have been used for centuries.

88

Investment Casting

In investment casting application, the viscosity of a slurry can be shear sensitive, that is to say the readings will alter depending on the amount of mixing shear being applied by both the mixer and the physical act of carrying out the flow test. If we suddenly stop the mixing action and remove shear from the slurry, the viscosity will slowly change over a period of time ranging from a few seconds to perhaps even some minutes. An alternative method of controlling slurries is to monitor and main tain their densities to as close as possible to constant. This should, in practice, avoid some of the possible problems described above, but takes no account of the variables that will lead to changes in coating/draining characteristics. It would also be found, if both density and flow time were monitored, that the relationship between these would not be consistent, primarily due to variation in the particle sizes of filler batches. Density relates to the relative amounts of solid and liquid present, provided that the highest accuracy is observed in carrying out the test. For primary coats it may be desirable to have a control test that relates much more to their dipping and draining characteristics, which are the crucial requirements. Various types of 'plate weight' test have been ap plied. This test employs a standard piece of thin metal plate, which has been weighed, dipped and allowed to drain without any movement of the plate, as would be the case with a wax pattern assembly. The gain in weight represents the amount of slurry retained and hence the 'coatability' of the slurry. This simple test, although useful, is not without pitfalls in interpretation. The amount of slurry left on the plate is determined not only by the characteristics of the slurry but by external influences, particularly those that affect its rate of drying. Some years ago the author was attempting to measure the relationship between slurry thixotropy and the plate weight retention value without the added influence of drying. The simple test used for this involved filling a test tube with slurry and emptying it by suspending the tube from its base — a kind of dip weigh test but draining from the inside where the locally high humidity prevented any drying. Surprisingly, nearly all of the slurry drained out. It was, therefore, concluded that the rheology of the coating was more dependent on how fast the slurry lost solvent than on the rheology of the slurry. Temperature and humidity are critical factors in providing a consistent and even coating on the wax patterns. Because of the complex geometries involved in some wax assemblies, humidity is quite likely to be higher in the central regions of the assembly than in the outer regions. This means that coating thickness is affected by mould geometry as well as the other factors already mentioned.

Investment Materials and Ceramic Shell Manufacture

89

Experience over some years with this test, using a metal plate, shows that a different reading can be obtained with a new plate compared with the reading from a well-used plate. Evidently again, wetting characteristics between surface and slurry can change the coated layer. The same change must also be true of the wax patterns, although by how much would be hard to establish. It is surprising how much a piece of metal equipment can be eroded by regular use (something which also applies to the flow cup test). A 'master' therefore needs to be kept to ensure that the state of the equipment is not influencing the readings. THE MOULD BUILD PROCESS In this section the relative influences of all the plant and ceramic materials on shell manufacture are considered. Many of the factors which influence the slurry and stucco have already been discussed but what must now be considered is how they combine together, and how the plant design and materials may affect the final mould quality. It will be assumed that the surface of the pattern assembly will have been cleaned by pre-washing in a solvent to remove wax-release agent residues, so that wetting between slurry and wax surface is effective. To ensure full and even coverage of the pattern, a suitable wetting agent should be included in the slurry formulation to overcome any deficiencies of the wax surface due to the last traces of wax-release agent. It will also be assumed that the mixer/slurry combination does not result in air bubbles being drawn into the slurry by poor design, excessive mixing turbulence or an insufficient stabilising period, and that it is within the usual control test limits. When considering what else might contribute to a defective primary coating, and therefore to scrap or reworked castings, the first consideration is how the pattern assembly is dipped into the slurry. This operation could introduce air bubbles where none previously existed, and later, when the rough surface of the casting comes to be examined and the roughness is apparent in tiny spherical protrusions, it will be impossible to know which of the two original possibilities caused the defect. To avoid such frustrations, it must be ensured right from the start that these deficiencies do not occur. It is very difficult to see this type of problem at the time of coating. Pattern assemblies should be immersed slowly, to ensure the minimum chance of air entrapment. If this is done any bubbles are usually large enough to be seen and removed after draining. With water based slurries there can be instances where it may be obviously difficult, because of poor wax wetting, to obtain an even slurry

90

Investment Casting

coverage. In this event, immediate action will be necessary to add more wetting agent to the slurry and to wash off the faulty coating and re -clean the surface. One good thing here is that some water based primary coats can be easily washed off with running water, if a defect is seen, and a new coatings applied. Primary coats may even be washed off and new coatings applied after they have been dried for some hours, provided that they have not been left for too long. It is a good practice to issue operators with small artists' brushes to break bubbles and touch up surface blemishes. This will encourage the careful examination of the wet layer before the stucco covers any defects. The next stage is to withdraw the pattern and drain off the excess slurry. Many variables can now occur during the drain period which, if the conditions are not controlled, can prevent the attainment of a very even layer of wet slurry before applying the stucco. Manipulation should be so designed as to clear any local surface drips and produce an even coating. A correctly formulated and controlled slurry will greatly assist in this operation. Too thick a slurry will require more manual skill to produce an even layer than one with the 'correct' rheology, but a thicker slurry will be less susceptible to another defect, which can be called 'stucco penetration'. This is more likely to occur with the coarser stuccos around 12 - 20 mesh (British Standard sieve) — which is why it is usual to select a finer grade of stucco for the primary coat. Stucco penetration at this stage refers to a condition in which the stucco, as it strikes the wet slurry, penetrates through to the wax surface, drawing down air pockets and producing small voids in the primary coat, which metal may then penetrate at the time of casting. Too thin slurries will allow very easy drainage but will cause a high risk of stucco penetration. Clearly, if a finer grit size is used for the stucco, a lower viscosity slurry can be used without fear of this penetration de fect. In some cases natural zircon sand is used, with a relatively thin slurry, to provide a primary coat free from voids. Zircon sand has a typical particle size of 100 mesh B.S. and the particles are less angular than in many synthetic materials that have been crushed to obtain the desired size. No hard and fast rules exist as to the 'correct' or optimum slurry consistency. The penetration defect leading to rough casting surfaces can be influenced by a range of other factors, but provided that these are controlled, smooth casting surfaces will be achieved. The kinetic energies of the particles of stucco, as they fall, are not only controlled by the height of fall (i.e. by the plant design) but by the density of the material. The drain time of the slurry will also affect the degree of penetration of the slurry grit. As manipulation proceeds water will evaporate from the slurry, changing its rheology, and the partially dried coatings will be

Investment Materials and Ceramic Shell Manufacture

91

more resistant to stucco penetration. This is why it is so important to optimise and fix time sequences to avoid mishap. We must not allow a condition where the stucco has not adhered to the wet slurry properly. A balance must therefore be struck, by correct slurry rheology and control of the draining operation, to avoid the two extremes of stucco penetration or insufficient adhesion. The latter gives rise to the possibility of later delamination of the affected layer, which will ultimately cause pieces of primary coat to flake off into the molten alloy. Poor stucco adhesion, even if the drainage is consistent, can be influ enced by the rate of drying at the draining stage and by the condition of the silica binder. Indeed the change in draining rate or early onset of gelling of the slurry is the most probable cause of poor stucco adhesion. Primary coat inclusions and/or rough casting surfaces will inevitably occur if the process is out of control. It should be noted that some geometries of wax assembly set up local conditions of high humidity, particularly in the centre of the cluster, causing local over-draining, so that rough surfaces can be seen only in certain central areas of the cast assembly. Inspection of each assembly before cutting off the components can therefore be extremely useful in diagnosing the true cause of any rough surfaces. Inspection of the condition of the shell mould and the surface of the casting at the time of shell removal is also necessary, to establish the cause of any mould -generated casting defect. Even with this adverse condition present at the interior surface of the mould, it does not necessarily follow that a rough casting surface will be produced. Here again we must consider that important property of wetting, in this case between the molten metal and the porous primary coat. Penetration is much more likely to occur with the conditions of high metal/mould wetting. A sessile drop test will provide a measure of the ability of an alloy to wet the surface of a mould. The contact angle of a drop of molten alloy resting on the surface can be measured, and if the angle is above 90° the system is considered to be non-wetting. Conversely, angles below 90° indicate wetting. Values above 90° could be represented in everyday experience by drops of water on a greasy glass plate. But if the plate is cleaned and a drop of alcohol is applied the contact angle will be very low, allowing the alcohol to spread over the surface. Wetting is also encouraged by high pouring temperature, which reduces the contact angle. Sporadic outbreaks of rough casting finish may therefore indicate that the potential adverse condition is present all the time on the mould surfaces but only shows when the pouring temperature is slightly higher than normal. Without taking this into consideration it might have been concluded that the quality of the primary coat was

92

Investment Casting

varying. Penetration is also influenced by the pressure head of molten metal. With knowledge of all these contributing factors, it may be possible to obtain circumstantial evidence as to the source of the problem by examining the precise position of the defect on a complete mould assem bly, immediately after knockout. Yet another factor is that thick casting sections generally provide hotter metal/mould interfaces and may therefore determine the areas where the defect occurs. Alternatively, if found on thin metal sections, the defect is more likely to be due to the coating process and to the formation of an uneven coating layer at the time of draining. This is yet another example of a defect that can suddenly become evident, and just as quickly disappear despite an apparent absence of change. As might be expected, those foundries that employ robot primary coating application generally see a considerable reduction both in coating inclusions and rough castings. A simple and revealing test to examine the typical defect described again uses a glass microscope slide, with one side sealed with adhesive tape. Dip and drain the slurry and stucco with an appropriate material. Repeat this process over a series of increasing drain times from zero to, say, five minutes. After drying the slide, the tape can be stripped away and the surface examined at X 20 magnification through the glass, for air voids and stucco penetration. It will be noticed that there is a 'window' of drain times between which the coating has good adhesion without penetration because of the thickening effect of the partial evaporation of water from the slurry. Conditions may be experienced, however, particularly with poor slurry-to-stucco particle size mismatch or with an aged slurry in which this ideal situation is not achieved. With aged slurry, the premature gelling as the solvent evaporates will require the stucco to be applied earlier in the drain cycle in order to ensure good adhesion and complete coverage. This will allow deep stucco penetration into the slurry and a situation may occur, in an extreme case, of an old or contaminated slurry with which it is not possible to obtain optimum coating at all. If this situation arises the only course is to discard the slurry and make up fresh material. While it is not essential to carry out this test on a routine basis it is useful in order to confirm poor slurry condition if this is suspected. The next operation in the mould build is to dry the coating, and here the post-gel shrinkage of the binder has a great influence on the structure of the dried coating. As drying proceeds, the binder shrinkage will give rise to stresses in the coating, leading to cracking. One of the main uses of the stucco is to disperse the cracks on a microscopic scale, because the cracking will be restricted to areas between the coarse particles. If, for example, there is a greater local thickness of slurry in a sharp corner of the pattern, or a drip or run due to poor draining, then the cracking will be on

Investment Materials and Ceramic Shell Manufacture 93 a macro not on a micro scale, because there will not be stucco particles in the locality to disperse the cracks. This may create a condition at the firing stage in which local pieces of thicker slurry can fall off the mould surface, again creating the possibility of ceramic entrapment at the casting stage. Although the possibility of such inclusions is relatively unlikely, they will lead to rejection of castings requiring high metallurgical integrity. Shrinkage of prime coat slurry is also affected by other variables. An aged slurry will not only create stuccoing problems but will also shrink more than a fresh slurry because of the greater amount of retained water at the point of gelation. The author has used an injected wax test piece to examine this shrinkage cracking in relation to batch variation of raw materials. The test piece was a plate with a number of circular depressions in its surface, the shallowest being 0 1 mm deep and the others of different depths up to 1 0 mm. By filling each depression with slurry and wiping across the surface with a straight-edge, different thicknesses of slurry were obtained. Gross cracking or mud cracking could be seen on the thicker sections but none on the thinner ones. It was found that layers of different thickness would crack and could be related to different batches of filler, no doubt owing to particle size differences. Thus, at any one time, the possibility of pieces of primary coat flaking would change depending on the filler. A range of experimental slurries was produced to examine this phenomenon and less cracking was found to occur with a lower concentration of silica, presumably because there was less of the material to shrink. Drying should ideally be carried out under constant conditions of humidity and temperature, to ensure that the whole process is as near constant as possible from one mould to the next. Sometimes two primary coats are applied as standard procedure, but such a process difference, which may give extra insurance against poor casting surfaces in one foundry, may also give a less permeable shell in another, with subsequent difficulties with gas entrapment or misrunning of the casting. The drying time for the primary coat is usually between one and two hours and it is not advisable to extend this period. The primary coat offers only a thin insulation barrier against ambient temperature fluctuations which will expand or contract the wax. As successive secondary coats are applied the increasing insulation reduces the possibility of mould crack ing with changes in external temperature. This should be borne in mind with those processes involving an overnight hold in the shell-build cycle. The primary coat should not be applied just before the end of the shift, because minor cracks can develop later into more serious problems. The secondary coats have a different function from that of the primary coat and, therefore, the types of material used may be different from

94

Investment Casting

those of the primary coat, which is in direct contact with molten alloy. Special formulations are, however, necessary for the arduous conditions found in directional solidification casting, in order to ensure dimensional stability of the mould at much higher temperatures and for longer periods. In general, slurries for secondary coats can be thinner in consistency to facilitate draining (as the penetration problem does not exist with second ary coats). Stuccos can also be coarser, to allow rapid increase in shell thickness. With very coarse stuccos an intermediate or 'coupling' stucco can be used for the first of the secondary coats, to avoid delamination. Because of the relaxed requirement at the draining stage, secondary coating is an ideal process for the advantages of robot controlled coating. Many of the inadequacies possible with manual shell build can be overcome by the introduction of automation. Robots must, however, be initially programmed with great care to avoid the pitfalls described above. Different mould geometries may need special programming to ensure good coating. This complexity is unfortunate, requiring measures to ensure that the correct programme for each type of mould is used. Alcohol based slurries offer rapid hardening by exposure to ammonia, whilst water based slurries require longer drying periods between coats. For moulds manufactured totally from water based binders, a special tunnel with reduced humidity is often employed to accelerate the drying process. Many types of drying equipment are used, all based on reducing humidity and forced air circulation by fans. PRINCIPLES AND PROBLEMS OF THE DEWAX OPERATION Dewaxing refers to the removal of the wax pattern assembly from the completed ceramic shell mould. At this stage it is possible to detect some of the mould cracks that may have been introduced, because wax itself tends to act as a crack detection medium, particularly if a dark coloured wax is used, which stains the outside of the mould where it seeps through. Defects such as delamination, described earlier, can also sometimes be seen in the exposed mould throat. The dewaxing itself can also cause delamination, owing to liquid residues in the shell boiling violently as heat is applied (in the same way that popcorn blows up by evaporation of its own moisture). Mould cracking is partly due to the mould and partly to the dewax operation. This, like the preceding operations, can be subject to a 'knifeedge' condition in which adverse combinations of variables give rise to occasional spasms of cracking.

Investment Materials and Ceramic Shell Manufacture

95

Many types of dewaxing equipment and process are used, but discussion here will be limited mainly to the steam autoclave, of which an example is shown in Fig. 6. 'Flash firing' will also be considered, as a less frequently used alternative. Wax is used in most foundries as the pattern material, although plastic patterns may be preferred for special applications — for example, for very thin and delicate patterns where wax would be subject to excessive breakage or distortion. Plastics such as polystyrene offer much greater handling strength than wax but are unfortunately much more difficult to remove, requiring flash firing rather than steam to melt them out of the mould. The key point to be appreciated at this stage is the significant difference between the low thermal expansion of the ceramic materials and the high thermal expansion of the waxes. If a ceramic mould were simply to be placed in an oven to melt out the wax (melting point 60 -90°C) it would certainly crack because of this differential.

Fig 6 The 'Boilerclave', a compact dewaxing autoclave with integral boiler unit. Note the facility for rapid introduction of moulds into the vessel to ensure the highest rate of heating. (Courtesy of Leeds and Bradford Boiler Company, West Yorkshire.)

96

Investment Casting

Successful dewaxing depends on applying sufficient heat to the exte rior of the ceramic shell so that a thin skin of wax adjacent to the primary coat melts before the bulk of the wax. The molten wax will then, if the conditions are suitable, soak into the ceramic shell progres sively, allowing the wax to expand freely without distorting the mould. The pattern material, the ceramic shell and the nature of the heat source interact to determine the success or failure of the dewaxing operation. Ideal criteria can be listed for these three components to achieve trouble free dewaxing. The conductivity of the pattern material needs to be low because a high value would allow the heat to diffuse rapidly into the bulk of the wax, without forming the essential thin layer of molten wax to soak into the mould and leave space for the bulk of the wax to expand. For tunately, one of the properties of waxes is their very poor heat transfer. Any filler or additive for other purposes would need to preserve this property. Viscosity is another consideration here, because high viscosity would imply that penetration of the molten layer of wax into the mould would be more diffcult. Indeed it has been found in practice that filled waxes with higher viscosities produce cracked shells more readily than unfilled waxes. There are also considerable differences between waxes in their change in viscosity in the temperature range from 50 to 80°C, which is the melting region for most waxes used in foundries. Some waxes, like alloys, have abrupt melting points, while others have a wide 'mushy' range in which their viscosity is high. With regard to heat supply, steam autoclaves usually operate at 150 180°C and flash firing ovens at 1000°C. Most waxes melt below 100°C, while plastic pattern materials such as polystyrene have viscosity tem perature ranges and melting points approaching 200°C and carbonisation and decomposition may begin before a low viscosity is achieved. A steam autoclave would not be suitable for plastic pattern removal. Flash firing is carried out at a temperature above 1000°C and is the main method em ployed for plastic patterns, which give more removal problems than wax. An ideal material combining the high strength of plastic with the low viscosity of wax has yet to be developed. Other methods of pattern removal such as solvent extraction may be used for special purposes. Water soluble waxes are extensively used but not generally for the complete mould assembly. Solvent extraction of paraffin waxes has largely been dropped in favour of the steam autoclave method. In melt extraction simultaneous and uniform heating of all surfaces is necessary. With steam heating, losses through over-long pipe work will cause a reduction of pressure and temperature in the steam. The

Investment Materials and Ceramic Shell Manufacture 97 Boilerclave unit in Fig. 6 is an example of good design for a dewax autoclave. Both boiler and pressure vessel are integrated into one unit to minimise losses. With flash dewaxing there will always be 'shadow' effects shielding certain parts of the mould, because the heat transfer is by radiation. The ideal of instantaneous exposure of all surfaces to a high heat flux can never be fully achieved because of the variable geometry of moulds, but sound design of equipment is a prime requirement for successful dewaxing. The classic difficult shape is a pattern taking the form of a hollow cylinder, because heat cannot be transferred to its interior surfaces as efficiently as to the exterior. Local insulation by packing paper or refractory wool inside the cylinder can help by preventing heat from reaching the insulated interior, so that the wax melts progressively in wards from the exposed surfaces. Metal inserts in a wax pattern assembly can cause similar problems, because heat may be conducted prematurely into the interior of the wax; local insulation of the exposed metal can again be effective in this in stance. A more radical measure for a difficult assembly is to place the whole mould into a refrigerator for a period to cool the wax surface and thus develop a greater temperature differential between wax and heat source. This can be surprisingly effective against shell cracking when all else fails. Because of the importance of maintaining constant room temperature at the shell build stage it is usual to position the autoclave or flash firing oven well away from the controlled temperature area. Moulds which still contain wax are sometimes taken out of this protective environment and stacked alongside the hot autoclave or close to the firing oven — just the right places to ensure premature wax expansion and shell cracking. Moulds should always be heated rapidly from cold, and in flash firing must be plunged straight into the furnace. In autoclave operation it is, similarly, poor practice to place a number of moulds in the pressure chamber one after another, because the first is likely to crack before the door is shut and the pressure applied. Apart from careless work procedures and unsuitable equipment or plant, the mould itself may contribute towards cracking. To ensure that the skin of molten wax will soak into the shell, adequate mould permeability is required. This is also needed at the casting stage to ensure that air and generated gases can escape. Such is the force of the wax expansion that an impermeable mould will burst no matter how strong it is. To ease pressure during dewaxing of particularly difficult mould configurations venting can also be considered. Primary coat porosity can vary, and in many cases no information is available to a foundry on the overall permeability of the mould. It remains a mystery to the author that in the sand casting industry mould

98

Investment Casting

permeability is one of the properties that is regularly, sometimes hourly, monitored, whereas in the investment casting industry permeability has not been researched in any detail. We know from published work that there is a general permeability level for a given type of filler and that this can vary significantly between materials. Fused silica, for example, generally gives a highly permeable mould, because of the redistribution of filler and binder at the gelation stage (as demonstrated in the slide test used to observe the drying effect). Slurries with silica fillers give rise to what has been referred to as the 'lace curtain' effect, by which quite an open structure is formed on drying. The slide test demonstrates the formation of this structure far better than any description. The slurry tends to dry initially at the top edge, if dipped vertically, tending to be slightly thinner at this point due to draining. Increased wetting agent will reduce but not com pletely eliminate this lace-like appearance. Here again, the wetting characteristics as between filler material and slurry have an important effect. Those fillers with different wetting characteristics from those of silica will produce a much denser structure and therefore have lower permeability. 'Lace curtain' effect with water based silica primary coats resembles the open structure of a silica gel as it transforms from the sol, and the submicron particles of silica in a slurry could possibly become negatively charged in much the same way. Because ideal dewaxing conditions are never achieved there will always be some stress on the ceramic shell. Satisfactory green strength of the shell will therefore contribute to successful dewaxing. A careful balance between the binder liquid and the filler contents in a slurry will avoid too much or too little binder in relation to the particles of filler it has to bind together. Because the binder shrinks after gelation, a surplus will cause excessive micro -cracking and a weak green shell. For certain mould systems and particularly with alcohol based moulds there is an optimum interval between the drying of the last coat and dewaxing. This is because the initial binder contraction may increase green strength, which later begins to decrease because of the progress of microcracking. There are fundamental differences in drying water based and alcohol based moulds. A gelled alcohol based binder is far more resistant to disruption in steam than a freshly gelled water based material. A water based shell, however, becomes progressively insoluble to steam over many months, as changes occur within its internal structure. It should be noted that the visible stages of gelation and bulk hardening do not indicate that the process of linking between particles is completed. The initial gel stage only indicates sufficient linkage to form the jelly -like solid, still containing all the liquid, which can then migrate slowly through the mould and take any unlinked silica with it. With long waiting periods

Investment Materials and Ceramic Shell Manufacture 99 (perhaps a few days) before dewaxing, migration to the wax/mould in terface will subsequently produce a powdery white deposit on the fully fired primary coat surface. This material is of very high specific surface and can aggravate metal/mould reactions. It is claimed that additions of film forming latex to the slurry can eliminate this problem by preventing migration. Unlike the straightforward air-dry hardening of water based shells, alcohol based systems are frequently hardened 'on command' by exposure to ammonia. This method produces a weaker shell than full air drying, because the ammonia gassed shell is gelled at a lower silica concentration and has a more porous structure. Premature gassing with ammonia may therefore cause dewaxing problems. A pre-drying period prior to gassing will give an intermediate strength of shell. It is important to be aware of the above problems, and to ensure consistent conditions and time cycles for the shell building operation. By adjusting the predrying period, some control over ultimate shell strength is possible. A longer period may be beneficial for large castings, in order to reduce the possibility of mould failure on casting. Errors in mould design can cause uneven thickness in local areas of the mould. Sharp edges may produce a much thinner layer than over the bulk of the shell, causing lines of weakness where cracks may appear. Characteristic long cracks occurring always in the same area may be due to this fault. Both slurry rheology and mould erosion at the stucco stage can contribute to thick and thin areas on the ceramic mould. Fluidised bed stuccoing can be particularly aggressive and erode some of the previous layer. In summary, it can be seen that there are many possible factors that may contribute to dewaxing problems and without very careful examination of the defective shell it can be difficult to pinpoint the cause in any particular case. Having said this, it must also be added that many foundries operate without ever experiencing problems at the dewaxing stage. A number of adverse conditions may come together to cause a sudden increase in mould cracking, which may then disappear just as rapidly. Most of the problems that do occur are sporadic and it is therefore often difficult to determine their real causes. MOULD FIRING Unlike most other types of mould, the ceramic shell has to be fired before casting. The resulting moulds can withstand very high temperatures and with careful selection of slurry compositions and stucco materials, can be used for a very wide range of alloys and casting techniques. The

100

Investment Casting

refractory nature of the ceramic mould is the main factor which makes the process so versatile, allowing casting into moulds preheated to temperatures as high as 1550°C. Fig. 7 shows a typical mould for the directional solidification of superalloy blades, involving extended exposure to pro cess temperatures of around 1500°C. There are three reasons for firing the green mould before casting. These are: 1. to remove residual pattern material and solvents remaining in the ceramic after dewaxing, 2. to sinter the structure of the ceramic, and 3. to present the mould for casting at a predetermined and consistent temperature. Although this part of the process is relatively trouble free, it is perhaps the least understood, and yet it can significantly influence the metallurgical and dimensional integrity of the castings in subtle ways. Very little has been written about the basic mechanisms occurring on heating the mould, and without this knowledge it will be difficult to understand the influ ence of mould firing on casting quality. With the advent of sophisticated high temperature testing equipment many data have now been accumulated. Further pertinent information is

Fig 7 Typical DS mould for use with a large water cooled chill casting furnace. The base diameter is about 500 mm and the mould has to remain dimensionally stable and chemically inert in contact with molten superalloy for a number of hours at temperatures of around 1500°C as the castings are progressively solidified. (Courtesy of Howmet (UK), Ltd. Exeter Casting Division.)

Investment Materials and Ceramic Shell Manufacture

101

now being made available from research work carried out in other fields, such as sol- gel technology in the manufacture of pure ceramic materials for engineering applications. For example, very pure silica based glasses can be made by mixing colloidal silica or ethyl silicate (as used in ceramic moulds) with other soluble oxides. These mixtures are gelled and fired and, because the various oxides are completely mixed on an atomic level, the resulting glasses will fuse at a much lower temperature than if powders were mixed and heated. As might be expected, one of the most intense areas of research concerns the ability to control the gel shrinkage to avoid gross cracking when discrete ceramic shapes are required. This is a similar problem to that encountered with investment casting. Investigations into shrinkage and firing of mixtures containing the silica colloid are thus of direct interest to this field. The first requirement of the heating process is to burn off residual wax from the dewaxing operation and to remove free volatile liquids. Alcohol is removed below 100°C as it has a low boiling point, but water contained in the gelled structure of the silica binder will not be completely removed at 100°C — in fact, some of the combined water will require a temperature above 1000°C to be completely expelled. Before discussing reactions that occur at these higher temperatures, it is first necessary to examine changes in the mould structure that occur as the mould is heated to the required temperature. Wax residues will only be completely removed from the ceramic if they are volatile or can be burnt away. Waxes should therefore be of high quality and contain not more than 01% ash residue. Burning requires oxygen, and it is important to ensure that there is sufficient in the mould firing oven to eliminate any carbonised wax within the established firing time. Both gas and electric ovens can lack sufficient oxygen if insufficient care is taken in the operational procedures. Around 8 -10% free oxygen should ideally be present for residue removal. Electric ovens should have provision for free through-flow of air by efficient venting. It is more difficult to achieve the necessary oxygen levels with gas fired ovens because gas burners tend to consume any free oxygen. Special burners will, however, maintain a good flame with more air than is necessary to burn the gas entering the oven. Gas and air settings are usually adjustable in order to achieve and maintain this condition. Regular monitoring of the firing atmosphere is required, particularly for steel and superalloy casting, to avoid metal reaction with residual carbon from incomplete combustion of wax on the surface of the mould. A minimum temperature of 500°C should be maintained, but it is preferable to increase this to around 800°C to ensure rapid removal of residue. With gas fired ovens it becomes progressively more difficult to maintain

102

Investment Casting

an acceptable oxygen atmosphere at higher firing temperatures. This is one of the reasons for operating a two-stage firing process, i.e. we can either fire the mould at a high temperature then immediately cast without cooling, or we can fire initially to remove residue at the lower temperature, cool down, and reheat to the higher temperature required for casting. This second firing will not require an oxidising atmosphere and will allow a simpler oven design. It should be remembered that large ovens always contain suspended debris which could enter the mould and cause metal reaction later. Ceramic particles and rust can be circulated through the oven by atmospheric turbulence, particularly with gas burners. Double firing with an intermediate cool-down has the further benefit of allowing limited visual inspection, and the pre-fired shell can be shaken to see if ceramic particles are present. Inspection for cracks using a simple dye penetration method is also possible at this stage. These cracks may have been undetected after dewaxing but may have opened up after prefiring and could cause the mould to burst on casting. Firing the mould to 500 -800°C will not by itself be sufficient to sinter the mould and render it inert to molten metal. Many foundries fire within the range 950 - 1100°C to achieve reasonable inertness and high mould stability. There are exceptions however — much higher temperatures, in the region of 1500°C, are used for the directional solidification process. In this process molten alloy must remain liquid for some time as the solid ification proceeds and the mould needs to be above the melting point, or liquidus, of the alloy. Superalloys, for example, melt in the range 1200° 1400°C. Special alloys may require even higher mould temperatures, but the use of silica-bearing ceramic is restricted to a maximum of around 1550°C — which is approaching the melting point of silica. Some vacuum casting furnaces are designed to heat the mould to these high temperatures within the casting chamber, in which case a pre-fire is essential to remove wax residues before the second heating which, because it is carried out in vacuum, does not provide the necess ary oxygen. In other applications, particularly with alloys of lower melting point, the maximum firing temperature can be restricted to around 850°C, primarily to limit the increase in strength of the ceramic structure with firing. Excessive strength in the shell can result in hot tears and cracks in the casting. These occur when the alloy is relatively weak and brittle at temperatures close to the solidus and are characterized by oxidized surfaces. Control of the mould strength is one of the main requirements for all investment casting applications. Some alloys are particularly crack sensitive owing to their very low hot strength, while some single crystal alloys

Investment Materials and Ceramic Shell Manufacture

103

can be stressed on cooling by amounts insufficient to cause cracking but enough to recrystallise the alloy later when it is heat treated, making it suitable only for scrap. Other results of unsuitable high temperature properties of the ceramic are sometimes inherent to the material but always influenced by the firing process. The mould can become deformable at the time of casting. Gases can be evolved, giving rise to gas entrapment in the casting or, more subtly, creating a back pressure in the mould cavity and restricting the ability of the molten alloy to fill the mould before it solidifies — producing the well known 'misrunning' defect. Various other metal/ mould reactions can occur and cause casting blemishes, particularly with alloys containing reactive alloy constituents. Many of these defects can result from inadequate mould firing, but the possibility of any problem occurring is heavily dependent on the geometry of the casting as well as the process conditions and materials. Recent advances in ceramic testing at high temperatures have brought a clearer understanding of changes that occur within the ceramic shell. Properties such as strength, deformability, thermal expansion and permeability have been studied on a variety of mould formulations, but temperature cycles and dwell times must also be taken into account be cause they too fundamentally influence mould behaviour at the time of pouring. Prior to the development of modern testing equipment capable of measuring mould properties at temperatures and heating cycles which occur in practice, operators had to rely on simple tests such as heating a shell test piece to a given temperature for, say, one hour, cooling back to room temperature, then breaking it to establish mould strength. Many references will be found to such procedures, but the information gained from them gave no indication of the strength of the mould at the high temperatures experienced in actual casting. One test that could be carried out was the determination of thermal expansion from room temperature to around 1200°C, relevant for moulds for conventional 'equiaxed' casting, although now insufficiently high for moulds for the directional solidification process. In this special process testing up to 1550°C is necessary to cover the operating range of the mould. A typical thermal expansion curve is shown in Fig. 8. One of the salient features that can be noticed in any mould expansion curve is that it is not straight over the whole range, but at some temperature above 900°C it starts to flatten out, followed by a dip on further heating. Indeed the material will then continue to shrink for some time, even if only held at the higher temperature. The temperature at which the characteristic dip in the expansion curve occurs will not only be different for different mould formulations but will

104

Investment Casting

Temperature, °C

The linear thermal expansion is largely governed by the nature of the stucco, but the strength and refractoriness of the mould is governed by the nature and contamination of the silicon bond.

Typical thermal expansion graph of a ceramic mould, indicating changes occurring in the silica bond on heating.

Fig 8

also vary marginally with different batches of filler or stucco used in the test specimens. This inflection point is sometimes referred to as the 'sinter start' temperature and will be different, even in the same test piece, if we determine the thermal expansion curve for a second time, after cooling down to room temperature. It is also significant that many mould material systems show quite a different curve on cooling compared with the initial heating curve, particularly after being heated above the point of inflection, indicating changes within the ceramic structure. A further word of caution in interpreting thermal expansion data is that even the heating rate will affect the curve. All these observations can be explained, and must be taken into account when using such data to interpret the foundry performance of a mould. It would be difficult to explain many of the observed characteristics of mould systems without understanding the fundamental chemical and physical changes occurring within the ceramic structure at the mould firing stage. The following section deals extensively with this aspect of the mould, because little information has previously been published on the subject.

Investment Materials and Ceramic Shell Manufacture

105

MECHANISMS OPERATING WITHIN THE CERAMIC MOULD STRUCTURE As some 70% of the mould is stucco it is not surprising that the initial slope of the thermal expansion curve always reflects the expansion characteristics of the stucco material rather than the filler. A fused silica stucco, for example, has a much lower expansion slope than an alumina stucco, reflecting the relative expansion coefficients of these two minerals. The slurry material does not greatly influence the mould expansion curve on initial heating. On cool-down, however, the observed differences with some mould formulations may reflect the expansion or contraction characteristics of the filler material rather than the stucco, particularly if the filler has much the lower expansion of the two. This is because the stucco particles contract at a significantly higher rate than the filler and therefore do not contribute to the bulk contraction (the stucco particles can contract within the slurry network without affecting the bulk). Another well known phenomenon is the phase changes in many oxide ceramics on heating or cooling, and these changes also can often be seen as a contribution to the shell curve. Silica can exist either as a glassy material (fused silica) or in a number of crystalline forms, each with a different density. All these forms of silica can be converted to others, accompanied by changes in density, reflected in changes in the thermal expansion slope; in the crystalline forms the volume change can occur suddenly at a specific temperature. Amorphous or non -crystalline silica, such as the widely employed fused silica stucco or the gel form of the binder, will crystallise at high temperatures, with measurable quantities of cristobalite being formed above 1000°C, but the rate of crystallisation will depend on the time and temperature of the firing cycle and the presence or absence of certain impurities. This combination of amorphous and crystalline silica, holding the shell together, is the major contribution to the foundry behaviour of all silica bonded ceramic moulds, the stucco and filler materials only modifying the basic behaviour of the silica bond. The nature of the silica binder and the effect of heat on it requires further consideration to explain mould behaviour. The hardening or gelling mechanism of silica binders has been seen to be based upon linking small particles of silica together to form a loose three-dimensional structure consisting of minute spherical particles in water based sols, and chain -like particles in ethyl silicates. Immediately after the initial gelation, this structure starts to lose water from the interior of the gel; this is accompanied by a volume contraction. After the initial drying stage the shrinkage stops, because the silica network has gained sufficient rigidity to be self supporting; some water

106

Investment Casting

remains, held tenaciously inside the pores. In fact we have now produced the well known desiccant 'silica gel' which absorbs moisture and can be reactivated by heating to 300°C to drive off the adsorbed water. The binder in the ceramic mould will act in exactly the same way; silica gel can be heated to even higher temperatures and still retain the ability to absorb moisture, albeit at lower efficiency. This ability is reduced to a very low level after heating to around 800 -850°C — which is therefore an ideal pre-fire temperature to avoid renewed moisture pickup in storage. Apart from drying, the structure will be further consolidated, with an increase in strength from that in the green condition. The ability to adsorb water on to the internal surfaces of the silica is a chemical rather than just a physical effect. The silica bond will also be microcracked because of the primary shrinkage of the gel and, depending on the nature and particle size of the filler and stucco, its morphology will influence the strength of the mould throughout the later heating cycles in the foundry process. As the shell mould is heated from room temperature the silica bond softens sufficiently to start to coalesce or sinter. This process begins at a temperature well below the 'sinter start' temperature as indicated by the thermal expansion dip. Even firing to 600°C will produce some coalescence of the silica particles, indicated by an increase in mould strength. At the same time some of the internal porous channels will start to close up and this will affect moisture removal, because there will be less internal surface for the water to be adsorbed. A surprising observation of research workers examining colloidal silica sintering is that although pore closure proceeds with increasing temperature, it will only continue for a limited time on holding at a fixed temperature. The process will be resumed if the temperature is raised by, say, 50°C, but will then stop again. This phenomenon remains unexplained. Water based shells behave similarly, with mould strength depending on the degree of bond sintering. Firing a mould at a fixed temperature be tween 950 and 1100°C, the strength will increase to a maximum and then remain stable. On increasing the temperature by 50°C, pore closure in the binder starts again and the strength increases. This attainment of a given strength at a particular mould firing temperature is of great importance with water based shells. Part of the mould firing process is to remove as much of the adsorbed water as possible from the silica binder, because if water is still present on casting it will give rise to gas evolution. To avoid this, the ideal remedy would be to fire the mould to a higher temperature than experienced at the time of casting. For low melting point alloys, such as aluminium based materials, a firing temperature somewhat below 1000°C is acceptable, but for steel a higher temperature improves mould stability and reduces gas evolution on casting. Generally, however, the traditional fir-

Investment Materials and Ceramic Shell Manufacture

107

ing temperature for steel, cobalt and nickel based castings has been no higher than 1100°C, because of the limited refractoriness of commercial mould systems and the maximum working temperature of conventional ovens. The presence of excessive moisture also gives rise to reactions with aggressive alloy constituents, which combine with the oxygen in the water molecules to create hydrogen and form metal oxide slag inclusions. To eliminate misrun castings it is often more effective to increase the firing temperature rather than the pouring temperature. Failure to understand the mechanisms involved can lead to excessively high pouring temperatures which will have no effect on the misrun defect. Further consideration needs to be given to the glassy and crystalline forms of silica because the behaviour of the binder in the heating cycle determines that of the ceramic mould, regardless of the filler or stucco employed. In the glassy form, all the silica molecules are randomly positioned, whilst in a crystal structure they are fixed in a regular lattice. The basic reason for the changes that take place on heating is increasing atomic mobility. In a glass the molecules can move in relation to each other and the material becomes soft and deformable on heating. Because glasses are supercooled liquids viscosity is used as a measure of softening, a process which is adverse to dimensional stability on casting. Crystals, however, are more rigid than glasses because of their lattice structure. Another characteristic of the crystalline state is the possibility of a phase transformation, involving a change to another lattice structure more stable at the particular temperature. Crystalline zirconia, for example, undergoes a sudden rearrangement of its lattice structure at approximately 1050°C, producing a sharp expansion curve on cooling. This sudden rearrangement is very disruptive and in some phase changes the materials may literally disintegrate into powder. With zirconia the phase change can be suppressed by adding a small amount of lime to form the widely used 'lime stabilised zirconia' which can be heated through the 1050°C region without disruption. Silica cristobalite has a phase change at around 220°C and there is no known additive that will prevent this. Cristobalite is much stiffer than silica gel, the glassy phase of silica which progressively softens on heating, but also strengthens with sintering. If cristobalite could be used there would be no sintering because, as a crystalline form of silica, it does not soften to any great extent. Silica glass will crystallise to cristobalite at high temperature with very little change in volume, because the glass and crystal densities are similar. The silica bond therefore starts to stiffen, which is just what is required of a mould binder. If pure silica glass were left for a sufficiently long time at a high enough temperature it would all

108

Investment Casting

turn to cristobalite, with a significant increase in refractoriness and rigidity, a process analogous to the crystallisation of jam with ageing. It would be difficult to use pure cristobalite as a filler because of the disruptive phase change at 220°C on heating, but if it forms at high temperature below 220°C it will disrupt on cooling, which is ideal for assisting easy removal of moulds after casting (as is well known to users of silica fillers and stuccos). These peculiarities of silica can also be used to great advantage in preformed ceramic core technology, where it also meets the other essential requirement of being soluble in caustic soda solutions and therefore easily removed from the inside of a casting. To tie the whole technology together it is necessary to consider tem peratures and rates of softening of the silica glass within the mould structure, and the rate of conversion to cristobalite. Silica glass becomes sufficiently soft at around 800°C to enable a fused silica rod to be bent in a gas flame. The rod will start to deform under its own weight in about an hour. This gives an indication of the effect on viscosity. A similar heating process applied to a shell mould enables sintering to increase the strength of the mould, but the softening effect may cause bulging in particularly heavy casting sections as the mould overheats. The stiffening effect of crystallisation must also be considered. A pure silica glass rod would have to be heated to about 1400°C and held at this temperature for some time before any cristobalite could be detected. A final point is that many oxides can dissolve in silica glass and reduce its viscosity or allow the softening process to occur at a lower temperature. This increased mobility also allows ordering of the molecules (crystallisation) to proceed at a faster rate. Examples of oxides that have this effect on silica bonds are water and sodium oxide, both conveniently available in conventional water based silica binders. Ethyl silicates used for producing hydrolysed binders do not have the sodium content, but alcohol based binders produced in-house are slightly different in this respect from prehydrolysed binders, which do contain a trace of sodium — derived from the use of mixed silica sol and ethyl silicate in their formulation. Water based binders soften at fairly low temperatures and start to crystallise at around 800°C, although at a very slow rate. Increasing the temperature to 1100°C brings increasing conversion of the glassy bond to rigid crystalline bond in about two hours of firing and gives a measurable improvement in the stiffness of the mould at this temperature, which is an important factor in silica core behaviour. The rate of heating also has a significant influence on the behaviour of the mould. Rapid heating gives insufficient time for crystallisation to stiffen the mould so that it will progressively soften with increasing tern-

Investment Materials and Ceramic Shell Manufacture

109

perature, and to a greater extent than if it were to be held isothermally in order to form cristobalite. Rapid heating is in fact what occurs on casting, whilst slow heating occurs at the mould firing stage. These conditions need to be right to achieve the most stable moulds and the most accurate castings. If we consider two identical moulds using a fused silica stucco and filler, but one made with a water based binder and the other with a hydrolysed ethyl silicate binder, the difference is that two accelerators, sodium oxide and water, are naturally present only in the former. If both moulds were rapidly heated to 1200°C would we see any difference? If the previous explanation of the behaviour of silica binders is correct, the water based shell would be expected both to soften and stiffen more rapidly, because of the presence of sodium. Crystallisation would not have time to proceed in the pure ethyl silicate bonded mould, which would sag under its own weight as the silica bond progressively softened. Water based shells are found, in practice, to gain sufficient stiffness from accelerated crystallisation to maintain a dimensionally stable condition. In the same way, a thermal expansion curve can be modified by altering the rate of heating or introducing 'hold periods' in the cycle. Other oxides also will increase the softening effect in the silica bond if allowed time on heating to diffuse or dissolve into the glassy silica. There are plenty of oxide impurities, in commonly used fillers and stuccos, which modify both the softening of the glass and the formation of cristobalite. Although the levels of impurities are in themselves quite low, the concentrations built up in the small amount of silica bond can be sufficient to depress its viscosity curve by 100-200°C. Returning to consider the thermal expansion curve, the glass can be envisaged as starting to soften and the pores to contract at an increasing rate until the bond shrinkage exceeds the natural expansion of the stucco material, resulting in a net contraction of the shell. Different batches of fillers will contain different levels of impurities and thus modify this dual process. These considerations complete the picture of the internal structural changes in the binder, including the difference between alcohol and water based bonds and the role of the filler in modifying the fundamental behaviour of the different forms of silica by the introduction of impurities. These impurities further complicate the chemical situation, because some of them will also react together to form new compounds which will themselves affect mould behaviour. Consideration of the possible silicates that could be formed by combination of impurity oxides having low melting points can guide the choice of materials for the ceramic mould. Oxides such as those of calcium or magnesium are individually highly

110

Investment Casting

refractory but in combination with silica will produce silicates that melt below 1400°C, which is unsatisfactory if the mould is to be used for directionally solidifying a casting with a mould temperature of 1500°C. Even in conventional casting, in which the mould never reaches these temperatures, traces of impurities form low melting point glasses which combine with the silica, reducing its viscosity and modifying the bulk mould dimensional stability. In addition, glassy phases are formed which harden after casting and cause great difficulty in removing the mould without damaging the casting. There are some impurities that increase the rate of crystallisation and others that retard it. It is also well known that two or more components in a ceramic system may depress melting points by the formation of eutectics. While it has not been possible to interpret fully the highly complex interactions between the mould materials and the silica bond, it should be evident that these will only modify the basic behaviour of the mould by modifying the silica bond. An example of the fundamental combination of glass and crystal properties can be seen in a domestic ceramic hob. The heating elements of the unit are under what appears to be a sheet of glass. This did indeed start its life as glass and advantage would have been taken of its soften ing on heating to form the sheet easily. However the softening could also cause the sheet to deform under the weight of a saucepan at cook ing temperatures. However, by careful and stepwise heating cycles and the addition of small quantities of certain oxides as /mineralisers,/crys tallisation was induced in the glass sheet to stiffen it by partial crystallisation — hence the term 'glass ceramic'. This technology could well be applied to the firing of ceramic moulds to achieve the maximum dimen sional stability. Because sodium salts are virtually insoluble in alcohol, very little sodium is retained in prehydrolysed binders. These always contain a stabilising acid such as hydrochloric or sulphuric. Many other compounds are, however, soluble in alcohol, particularly chlorides. This means that any acid soluble oxide in the filler will dissolve into the binder in the slurry and introduce impurities that will modify the properties of the silica bond even before the application of the primary coat. This could be turned to advantage by deliberately adding suitable chlorides to act as 'mineralisers' at the mould firing stage. Again it must be remembered that alcohol based binders are not so sensitive to added salts, so such 'doping' is quite possible. A further point which should be made on post-casting mould removal is that certain combinations of silica with other oxides will form excessive amounts of glass in the mould structure. Where silica binder is used with other oxide materials, such as alumino-silicate fillers and stuccos, these

Investment Materials and Ceramic Shell Manufacture

111

glasses will not only soften the mould and slow down crystallisation but after casting there will be a significant increase in mould strength as the glass hardens on cooling. This can cause great difficulty in shell removal. As with other ceramic materials, excessive formation of glassy phases is highly undesirable. However, with a basic knowledge of the high temperature chemistry of, and the interaction between, various mould materials, it may be possible to modify offending mould formulations. Future development could well proceed through further understanding of the roles of impurities, particle size control, binder doping and unconventional firing cycles. PRESENTING THE MOULD FOR CASTING After a suitable mould firing period at a temperature sufficiently high to stabilise the mould, it is necessary to select the mould temperature required for pouring. One of the important and unique features of this casting process is the availability of mould temperatures up to 1550°C. This flexibility aids in filling thin sections when casting alloys of low fluidity, and contributes to the wide range of complex and thin walled castings that can be produced. Mould temperature also influences the solidification of the alloy, and correct temperature selection can improve metallurgical integrity by reducing casting defects; conversely incorrect or variable mould temperature can create problems. In most foundries using ceramic shell moulds there is a distinct time lapse between removing the mould from the oven and pouring the alloy. Any variation here can give rise to misrun castings. This can be aggravated by incorrect mould firing conditions leading to incomplete removal of moisture. Gases then released on casting create a back pressure which hinders mould fill. While it has been generally considered that misrun defects are due mainly to 'low7 firing temperatures, attention needs to be given to the full time-and-temperature profile of mould firing and to the temperature at which metal enters the mould cavity. The effect of gas release can also be eliminated by mould venting or by ensuring high permeability. The optimum mould temperatures for firing and casting may not coin cide. For example, in the casting of highly reactive alloys such as titanium, it is desirable to fire at as high a temperature as possible to make the mould inert, but to have the mould temperature for casting as low as would be consistent with complete mould fill, in order to reduce the reaction level between mould and metal. These conflicting requirements will influence the mould temperatures chosen in the casting of highly reactive alloys. A lower firing temperature to limit shell strength would

112

Investment Casting

be beneficial with light alloys. Casting quality can be improved not only by careful choice of casting temperature, but also by adjustment of the mould temperature. Many operators insulate the mould to minimise the temperature drop when there is a time lapse while transferring the mould to a casting chamber. This is particularly important in vacuum casting, where the hot mould has to be transferred into a vacuum chamber via a mould lock, which takes additional time. One method of insulation is to place the mould in a temperature resistant can and surround it with a refractory grog prior to mould firing. This maximises the mould temperature at the time of casting and reduces the subsequent casting solidification rate, although it will also slow down the heating of the mould in the firing oven. The reduced rate of heating will affect the sintering process, and consideration should be given to the previous discussion on the mechanisms occurring at this stage. An example will show the effects of mould insulation on sintering due to slow ing down the rate of heating. Tests can be applied to produce a curve of strength against time for a water based system. At 950°C, for example, the shell strength might increase to a maximum over some hours, but at 1050°C rapidly stabilise to a constant value in about 10 minutes, indicating the real rates of sintering of the silica bond. With a 10-hour firing cycle at 950°C, and using backed insulated moulds, it may take the full ten hours to reach the oven temperature and some hours more to attain a stable strength, so full stability will not be attained at this temperature. Furthermore, if some moulds are held at 950°C for 10 hours and others are left over a weekend, the strengths will not be the same. At 1050°C, however, because only 15 minutes at this temperature will be sufficient to stabilise the shells, they will remain much more consistent in strength even if oven dwell times vary from the stipulated 10 hours. Where a lower mould temperature is desirable it is better to fire at the higher temperature and then lower it by moving the moulds into a cooler zone of the oven, although this might require more elaborate equipment. Many foundries use insulating wrappings; employed selectively, for example by applying them locally to feeders, improved soundness and increased metal yield can sometimes be obtained. It must be remembered that as the mould heats up at the firing stage it can become deformable under external forces. Canisters filled with refractory grog can exert an inward force on the mould owing to natural packdown with movement through the oven, resulting in distortion. An advanced method of ensuring the correct mould temperature is to use a casting furnace with integral heating elements; directional solid -

Investment Materials and Ceramic Shell Manufacture

113

ification furnaces operate in this manner. This elegant system has also been applied in conventional casting, where selective heating of the feeders can also be provided to improve metal integrity. While insulation will usually improve the metallurgical quality of the casting, it can allow excessive peak mould temperatures at the time of metal pour, with the risk of mould deformation or poor surface finish. Thus, and as illustrated by this further example, although the firing stage is often considered to be relatively trouble free, casting defects can occur due to incorrect procedure, and the resulting nonconformance may not even be associated with this vital operation.

PREFORMED CERAMIC CORES Although many complex solid castings are produced, it is often necessary to 'core out' the casting to form hollow internal features. The simplest of these can be formed directly by the mould investment being allowed to flow into re-entrant cavities incorporated in the wax pattern. The production of such pattern features can in turn be assisted by the use of separately formed soluble wax cores, which can be dissolved out to leave the hollow features in the wax proper. Using these techniques the entire mould is formed from the single investment material. Limitations are, however, evident in using the mould investment to form its own 'core'. Deep pockets and small holes are difficult to invest and dry. Also, the normal mould materials need to be removed by mechanical means. This restricts the geometries to those mechanically accessible after casting. It thus becomes necessary to employ separate cores to form the more intricate and inaccessible internal features, in much the same way as in sand casting. But for investment casting a unique approach is necessary. The ceramic core must have properties to match those of the mould, being capable of remaining inert and dimensionally stable at molten metal tem peratures, producing smooth internal cast surfaces. The core will however, be placed in the wax tool before wax injection, so that it be comes embodied in the shell mould when the wax pattern is invested, and it must be capable of being removed from the casting afterwards by dissolution rather than mechanically. These pre-formed ceramic cores incur high additional cost compared with the direct formation of hollow features by the mould investment, but are necessary for all but the simpler features. Unlike cores used in sand casting, they are produced by specialist manufacturers and not by the foundry — mainly because of the complexities of the manufacturing process.

114

Investment Casting

Silica is the material used for virtually all pre-formed ceramic cores, because after casting the core can be removed by dissolution in aqueous caustic alkali, which does not corrode iron, nickel or cobalt alloys. In the same way acid -soluble cores have been produced for non-ferrous alloys that would dissolve in alkaline solutions. The leachable nature of the silica core has enabled the complexity of cast internal hollow features to be developed, limited only by the skill of the core manufacturer and by the core being sufficiently robust to survive the casting process. Because one of the main applications for pre-formed ceramic cores has been to make internal cooling passages in turbine blades, the advancement of blade cooling design has been the driving force which has developed the technology to produce cores of immense complexity. Some typical cores are shown in Fig. 9. The preferred method of forming ceramic cores is by injection mould ing. A ceramic dough is forced under pressure into a die cavity and hardened prior to removal. This is followed by high temperature firing to sinter the core material. Other methods of forming have been used to a limited extent where the shape of the article and the economics of manufacture offer advantages. These include pressing semi-dry powders to

Fig 9 Selection of preformed ceramic cores used to form cast internal features. (Courtesy of Fairey Industrial Ceramics Ltd., Stone, Staffs.)

Investment Materials and Ceramic Shell Manufacture

115

compact the material and slurry casting. Any forming method must rely on the use of a suitable silica powder formulation including a hardenable binder. In slurry casting, additions of high temperature hydraulic cements can provide the means of hardening, but the very high production rate by injection moulding is difficult to match using alternative systems. Injection moulding doughs are made with closely controlled silica powders, sometimes with additions of other ceramic materials. The binders used fall into two categories: thermoplastic materials, which will harden simply by solidifying from the melt in a cold injection die, and thermoset binders, which will polymerise with applied heat on injection moulding into a hot die. Waxes are traditionally employed for thermoplastic doughs, and silicone resins are used in formulations for hardening on heating (thermosetting). Other additives are used, but the formulations have always been closely guarded proprietary information. There are significant processing differences between a moulded 'green' core produced with a wax binder, and a thermoset silicone bonded core. In its green condition, a cured silicone bonded core is very strong, being similar to many 'filled' thermoset resins used without further processing for other applications, whereas a wax bonded core is more fragile and needs to be treated as such. Both systems, prior to the sintering operation, will need a special heat treatment for 'debonding', which entails removing the temporary binder materials. Debonding requires a very slow heating cycle that may range from one to five days, to ensure that the core integrity is preserved. The thermal decomposition of the organic binder materials could otherwise blow the core to pieces, because any gaseous decomposition products formed within the core need time to diffuse to the surface and escape; this is a long process because, unlike a ceramic shell mould, a green core is impermeable. To extend the temperature range over which the wax binder decom poses, a blend of waxes with different decomposition characteristics can be used to speed up the time cycle for the removal of the resulting volatile products. The use of binder materials with a wide spread of molecular weights facilitates easier debonding. Waxes with low viscosity are generally used and a valuable technique can be employed at the debonding stage to shorten the process of removing volatile substances. The green cores are packed into a box filled with a fine inert powder such as alumina. As the temperature increases above the melting point of the wax binder, this powder not only supports the core against sagging but draws out much of the liquid content by capillary action, providing a degree of permeability to the core to facilitate rapid removal of volatiles. A modern debonding process can be as short as twenty-four hours. After that time the temperature can be increased to

116

Investment Casting

sinter the silica at a much faster rate. Temperatures around 1100°C and dwell times of a few hours are usual for sintering the core and developing its structure. The thermoset silicone resin binders are methyl or phenyl silanes; in effect the silica molecule is bonded to certain organic groups. If these materials are heated to around 500°C the organic groups are decomposed and burned away while the silica remains as a residue. Methyl silanes may contain up to 80% silica in the structure of the resin molecule; phenyl silanes have about 40% silica content. By using one of these resins, or a combination of the two, in the formulation of dough with silica powder, a silica bonded silica core is obtained with the bond originating in the resin. In such a thermoset system there is no remelting of the polymerised binder on reheating to debond the green core, so that cores of thermoset binders can be fired without the additional support needed for wax bonded materials. Another aspect of using silicone resin binders is that the high level of residual silica reduces the permeability of the core when it is debonded, whereas wax is completely removed. This prevents rapid debonding, which therefore means that three to five days are required to debond without disruption of the silica ceramic. The added source of silica in silicone formulations results in cores of lower permeability than wax bonded cores, and of high physical strength. By using a combination of purely organic resins and silicones the silica concentration can be adjusted, and the strength of the core controlled, by varying the proportions of the two materials. Neither of the binder systems has a clear advantage over the other. While silicones are more costly than waxes, the resultant cores are stronger and more able to resist breakage in processing. Other differences include the much higher viscosity of molten silicone resin than of liquid wax. This gives a stiffer hot moulding dough that needs considerable pressure to mould by injection, and therefore requires the use of hardened steel dies. The special characteristic of the green investment shell which makes it relatively easy to fire without disruption and distortion is totally lacking in the green preformed ceramic core. The preformed core cannot be plunged straight into a firing oven at 1100°C, as can a ceramic shell mould. The mould is composed of a blend of fine powder from the slurry with the coarse grit from the stucco. The aggregate grading is therefore about 70% coarse material and 30% fine, and bulk firing shrinkage is reduced to a very low level for reasons examined in detail earlier in the chapter. The relative absence of shrinkage and the slight softening of the silica binder allow the mould to sustain the thermal shock in firing, while the

Investment Materials and Ceramic Shell Manufacture

117

natural permeability of the shell structure also allows free removal of any gases arising from decomposition. This is not, however, possible with a preformed core, which contains fine powder to allow the moulding of thin, delicate parts. The addition of a coarser fraction would reduce firing shrinkage but would impair the 'injectability', or plasticity of the mould able dough, especially for the manufacture of thin, intricate cores. Core manufacturers have learned to overcome these difficulties by long practical experience, and many compromises were required, particularly in the development of a manufacturing process for aeroengine cores. Most of the major suppliers undertook the grinding and preparation of the silica powders for their own cores in order to maintain quality — because the commercially available powders used to produce the shell moulds were not controlled to the close grading required for core manufacture. Even with this attention to detail, the core manufacturer is still faced with a significant 'die-to -fired' contraction which has to be allowed for in the injection moulding tools. Once the tool is cut, it is a principal objective of the ceramic core technologist to maintain a highly consistent firing shrinkage to ensure a dimensionally acceptable product. By careful formulation of the mould compound and controlling of the injection moulding cycle it is possible to limit the contraction of the core to below 2%. Any allowance designed into the die has ideally to be controlled to ± 0.1%. This will produce, for example, a 250 mm long turbine blade core to a tolerance of ± 0.25 mm, which is just sufficient to be accommodated in a precision wax die with the necessary accuracy. To accommodate this tolerance every batch of moulding dough must be kept to a die-to-fired shrinkage of 1.9-2.1% on a nominal 2% contraction allowance, which is a stringent demand upon even the most skilled operation. Cores are located in the wax die by 'coreprints' at their ends. These enable the core to be positioned and held in the wax tool and subsequently in the mould cavity. Here one of the difficulties in producing cores may become apparent. Particle size and binder-to-powder ratio control the firing shrinkage, and any minor separation or segregation as the moulding dough is injected under pressure will lead to uneven shrinkage as the core is sintered. Segregation is virtually impossible to eliminate entirely with complex core geometries, and may lead to twisting and bowing at the firing stage. This distortion, however small, is a real problem in core technology, because the core will be placed into a metal die for wax injection, and there may be occasions when the brittle ceramic will crack even as the tool is closed prior to injection. Every core, therefore, needs to be thoroughly inspected before use, especially for bow and distortion. Coreprints are made slightly smaller than the metal tool in

118

Investment Casting

order to accommodate minor variations in dimensions. Failure to do this will cause the core to be crushed in the tool. The clearance between die and coreprint must always be a compromise. If it is too large, core crush ing is avoided but the positioning tolerance of the core in relation to the wax walls is reduced. In general, clearances of about 012 mm are recom mended. Given these difficulties, some incidence of core scrap would seem to be inevitable. Three types of core dimensional error can be identified, which are often very difficult to resolve in practice. The core die is itself subject to error, the core can have variable shrinkage, and firing distortions must be added to these variations. Of course the wax die too can have errors at the coreprint locations. All these variables necessitate considerable 'tuning' of core and die when they are first introduced into production, to ensure a good fit. The tuning can continue after the dimensional results of the first batches of castings are made available, and this sometimes results in long and unplanned lead times. Efforts to introduce process modelling to get this aspect of the casting process 'right first time' should eventually minimise these difficulties. Once positioned inside the mould cavity the core needs to maintain a precise location relative to the mould walls throughout the remainder of the process. This is not easy because we are dealing with silica, with its unusual physical properties. Fused silica has a very low expansion. In the early days of core manu facture this was considered to be one of its good points, but most moulds have higher expansions unless a silica stucco is used, so there is a differential between mould and core when heat is applied. Because of the nature of the mould firing process the core too is subjected to thermal shock by the mould being plunged into an oven at about 1000°C. The silica core, being composed of a glassy material, starts to soften and will be susceptible to bow or twist if the differential expansion forces are imposed on the system. To minimise this problem it is usual to paint one end print of a core with a lacquer which bums off when the mould is fired to form a slip joint. If this is not completely free then core bowing is possible even before metal is poured, although this will not be evident until the casting is inspected. Research into slip joints, core positioning and wall section tolerances is difficult because of this inaccessibility, but further consideration of the technology associated with silica can assist in understanding such aspects of core behaviour. Fused silica, the basic ingredient in the core, will progressively soften on heating, and in its pure state may even start to sag under its own weight at the core firing stage. This is why stable ceramic setters are often used to support the core during firing. Because the core undergoes a prefire, conversion to cristobalite helps to stiffen the structure. It is not,

Investment Materials and Ceramic Shell Manufacture

119

however, possible to employ cristobalite from the outset because its disruptive phase change at 220°C would destroy the core as it cooled after pre-firing. A balance is therefore required between the silica glass and the crystalline cristobalite to optimise dimensional stability; 10% cristobalite will have a significant beneficial effect on the higher temperature stiffness of the core, but 20% will be much better. At this level, however, the phase change on cooling the fired core will considerably weaken the structure, and the strength may be reduced to half that of a 100% silica glass core. The maximum coarseness of silica particle size has already been incorporated to achieve the lowest firing shrinkage compatible with mouldable rheology, which in itself reduces fired strength because sintering is restricted. Good core formulations are more of an art than a science because again there are so many variables to contend with. A good way to limit the effect of the phase change is to dilute the silica with another material such as alumina or zircon powder. These are not soluble in the core leach process but will separate as a sludge as the silica dissolves. This mitigates the disruption of the cristobalite and provides a useful addition to the armoury of the core producer. Another phenomenon, already discussed relative to the silica in the mould, is that certain ceramic materials or impurities promote crystallisation and others may retard it. Silica as supplied is usually much purer than some of the other ceramic materials used for investment casting, mainly because of the natural purity of the silica sand or rock crystal used in its manufacture. Most problems therefore arise from other materials added to the silica-based core. Zircon very slightly degrades the high temperature stiffness because of its own impurities, while alumina may significantly modify and retard the amount of cristobalite formed at the core firing stage. High levels of alumina greatly reduce the resistance to deformation of a silica core. Traces of alkali metals, as in water based binders, promote cristobalite with beneficial effects. Zircon/silica mixes have been generally used as the preferred materials for directionally solidified castings. The silica is exposed to a tem perature of around 1500°C prior to casting at this temperature; with traces of impurity present, the silica converts to cristobalite to form a highly rigid core. As the core is rapidly heated, however, it can be expected to deform grossly above 1200°C, simply under its own weight. The technique developed to produce cored directionally solidified castings, with these much higher process temperatures, is to use tiny metal pins, usually of platinum, located in the moulds to hold the core straight as it heats up through the critical temperature range before crystallisa tion makes it self supporting. By the time the metal is poured, and the pin supports dissolve away, the core has crystallised to a high

120

Investment Casting

cristobalite level. The integrity of the core body is thus preserved, be cause the stable high temperature form of cristobalite is of very similar density to fused silica at these temperatures; the sudden phase change to the low temperature form at 220°C, will only occur after the casting has solidified. With the lower temperatures in the equiaxed casting of turbine blades the main limiting factor to prevent core bow in the blades is the length of the core relative to its chord. Long cored turbine blades, associated with the low pressure stages of an aero engine, may produce more bow in the core than short, stubby high pressure blades. Silica glass tubing, manufactured by a drawing process, has been used to produce long tubes with internal diameters down to 0-5 mm and length to diameter ratios of 100/1. This tubing has the same characteristics as the injection moulded core but, being completely dense rather than porous, it has high intrinsic strength and is sufficiently robust to survive the casting process even with such small diameters. Tubes can be bent using oxyacetylene flames and can therefore be quite useful for producing cast-in holes in steels and superalloys. The tubing is relatively free from impurities and does not rapidly convert to cristobalite, as do injection moulded cores. Cristobalite formation usually initiates on the glass surface owing to impurities and develops rapidly with a powder based core because of the very high surface area available for conversion. Solid tubing only crystallises on its surface (usually owing to traces of alkali from the operator's fingers, which gives it a characteristic bloomed appearance after firing. This is why quartz iodine bulbs used in car head lamps must not be fitted with bare hands, because the high temperature of the silica envelope when the bulb is lit will devitrify or crystallise the surface and weaken it). Silica tubes for cores should either be handled with gloves or thoroughly washed with solvent before they are heated. If the core is to be heated and cast, surface cleaning is not essential; but if it is intended to pre-fire or heat treat the core, with an intermediate cool down, cleaning would be advisable. Surface devitrification on a tube is much more disruptive than the dispersed crystallisation throughout an injection moulded core. After casting, the core must be dissolved away using alkali hydroxides of suitable concentration. Various methods are used to increase the rate of removal from complex areas within the casting. A hot 20% aqueous solu tion of sodium hydroxide is more than sufficient for this purpose. Higher concentrations are sometimes used and potassium hydroxide is an alternative solvent. The main considerations for efficient leaching are the tem perature of the leaching liquid and the degree of agitation at the core/ liquid interface. As the core is progressively dissolved agitation can be come more difficult and in deep pockets the solution may become locally

Investment Materials and Ceramic Shell Manufacture

121

stagnant and spent, creating conditions for the formation of silica gel, which prevents further core removal. To increase the activity of core leaching solutions, high pressures may be employed using special autoclaves, of which one example is shown in Fig. 10. Modem core leaching autoclaves operate at pressures of around 7 bar and introduce a degree of agitation by periodically releasing the pressure and allowing the liquid to boil. High leaching temperatures can also be achieved by increasing the pressure above the solution. A number of designs of high pressure autoclave are available. Pressure levels of 70 bar allow an operational temperature of 250°C, while still higher pressures have been used, up to 100 bar at 350°C; this increase in reaction temperature improves the efficiency of core removal. An interesting side effect of pressure applied to avoid liquid boiling is that as the aqueous caustic alkali is heated it expands, until at 350°C it reaches double its original volume - so that a vessel needs to be only half filled to achieve immersion of all the components. High pressure autoclaves have also been successfully used to remove alumina based cores. Since the late 1970s alumina has been targeted as an

Fig 10 Core leaching autoclave for dissolving fused silica cores after casting. (Courtesy of Leeds and Bradford Boiler Company Ltd., West Yorkshire.)

122

Investment Casting

alternative to the silica core. The reward for a successful core body would be that alumina does not soften or go through any undesirable phase changes as does silica, and is also less reactive to aggressive alloys. However, although alumina is soluble in caustic alkalis, core removal is much too slow for it to be a viable alternative to silica. Developments with more refractory and leachable materials will allow much improved wall sections to be maintained, and a steady flow of publications since the 1970s indicates that research is still proceeding with alumina. The behaviour of silica ceramics, especially in the preformed core operation, is extensively treated in Reference 1, and silica sol technology in Reference 2, which also includes a comprehensive bibliography. Reference 3 contributes to the understanding of the complex reactions and interactions of ceramic mould materials at high temperatures.

CONCLUSION This chapter has concentrated on the material technology of the manufacture of ceramic moulds rather than plant or process details. Silica in its many forms affects the behaviour of the mould throughout the process. In particular, the inherent differences between silica in its amorphous and crystalline states explain many features of ceramic mould behaviour. The quality of investment castings is largely determined by the mould, and many future advances in this industry will surely be made by ceramic mould improvements and innovations.

REFERENCES 1. R.B. Sosman: Phases of Silica, Rutgers University Press, New Brunswick, USA, 1965. 2. R.K.Iler: The Chemistry of Silica, John Wiley and Sons, Chichester, UK, 1979. 3 E.M. Levin, C.R. Robbins and H.F. McMurdie: Phase Diagrams for Ceramicists, Vol. 1, American Ceramic Society, Columbus, Ohio, USA, 1964.

5 Melting and Casting S. M. BOND

INTRODUCTION The central feature of investment casting, as with other casting techniques, is the process of melting metals and pouring them into moulds to manufacture solid forms. The investment casting industry has grown dramatically over the last fifty years and the range of alloys produced is wider than that associated with many other casting processes. Manufactured parts include individual dental castings, turbine blades, thin walled aluminium components and higher volume ferrous parts; some of these products are reviewed in other parts of the present work. Investment foundries demand high quality feedstock to ensure accredited input into their systems, which require stringent checks throughout the entire process. Moulds reach the casting stage with high added value and within well defined quality standards, and the metal melts must be controlled with similar care. Technical facilities of high quality are thus required throughout the melting, pouring and solidification stages. Furnace purchase can be the most important capital expenditure that a foundry has to make and decisions made on a step-by-step basis could lead to an unsuitable pattern of melt flow within the foundry. Criteria such as services available, present and future metal demand, ability to undertake special one-off castings and environmental legislation, must all be taken into account. Thus, although capital costs will always be a major deciding factor, they are by no means the only consideration. Several specialised melting and casting techniques are used within the investment casting industry. Wherever possible air casting is employed on economic grounds, since casting under a vacuum or protective atmosphere adds significantly to unit production costs. The use and value of the final product determine whether vacuum processing is justified and this becomes essential in the melting of high temperature alloys DOI: 10.1201/9781003419228-5

124

Investment Casting

containing reactive elements. Similarly, certain products such as dental or jewellery castings, requiring high definition with small mass, benefit from special techniques such as centrifugal casting. Environmental issues can no longer be ignored by any industrial organisation and foundries are examined in more detail than most. Health considerations as well as legislation demand expenditure in this area, which is also rewarded by a more stable and reliable workforce.12 This is a further consideration in the selection of melting and casting equipment for the investment foundry.

FEEDSTOCK The importance of the feed or remelt stock to the quality of the finished casting cannot be overstated. Investment foundries normally buy stock from reputable suppliers with full analysis and most have in-house analytical facilities to check alloy composition before the metal enters the system. Ferrous alloys can be supplied in either air or vacuum melted form. Air melt ingots are usually in bar or open shell form. The batches are usually induction melted to ensure minimum time at the pouring temperature and close compositional accuracy, and bottom pouring and teapot ladles assist with the high quality requirements of the product. Air melted stock is also shot-blasted to avoid the introduction of heavy oxide into the production melts. Vacuum cast feedstock is essential for investment foundries which require the highest quality standards; vacuum melting reduces detrimental trace elements and allows alloy refining. Customers' returned runners and risers can be recycled in the charges used for the production of the relevant melting stock. To meet the needs of some investment foundries vacuum melt alloy suppliers have extended their technique to produce a superior product, especially for the high temperature alloys. A quiescent bath is established prior to pouring to allow inclusions to rise to the surface of the melt and the cooling rate from the superheat to the pouring temperature is slow and controlled. The metal is poured from the inclusion free region at the base of the teapot or bottom pouring ladle. The resulting lower inclusion levels contribute to higher mechanical properties and fewer casting defects in the end product. Non -ferrous alloys present a more varied picture and feedstocks are sup plied in numerous forms to specifications required by the industry. For the superalloys, covering mainly the nickel-based alloys for high tem perature applications, the position is analogous to that already outlined

Melting and Casting

125

for ferrous materials, with increased emphasis on vacuum processing for the most rigorous control of impurities and inclusions, backed by specialised acceptance tests. Aluminium casting alloys are standardised in two groups, for general engineering and aerospace applications. The first are subject to B.S 1490, Aluminium and Aluminium Alloy Ingots and Castings, and the second to B.S (L) Aerospace and DTD specifications.3 Some of the aerospace alloys are also standardised in B.S 1490, with nearly identical compositions and mechanical properties, but to a higher inspection requirement. There are twenty-one alloys designated in B.S 1490, although for financial and practical reasons most foundries employ only a limited number of these. Wherever possible the investment foundry will use the experience gained over many years to advise the customer on alloy types to meet their specific requirements. Ingots for non-ferrous alloy melting are usually in a standard configuration, independent of the accredited supplier, and are provided with suitable means of identification: those for aluminium alloy production, for example, are colour coded to B.S 1490. MELT REACTIONS Some of the basic principles that apply to all casting techniques will be reviewed before examining details relating specifically to investment casting.4 The reaction of the melt with the surrounding atmosphere is one of the keys to the production of quality castings, whilst the fluidity of the liquid metal and its flow into the mould will affect both shape and internal soundness. Liquid metal is reactive and will attempt to reach equilibrium with its surroundings. These are the gaseous atmosphere, the vessel in which it is contained and any slag that may be on the surface. In practice it is the gasmetal reactions that are of greatest significance. These will always occur when metal is air melted and the removal of hydrogen is a major preoccupation of foundrymen. Moisture from various sources will result in reactions of the form M + H 2 O —> MO + H2 There are several potential sources of moisture. The combustion of hydrocarbon fuels such as gas and oil will produce water vapour. Crucibles are porous to gases and hydrogen will permeate most materials at melting temperatures, so that it can be assumed that hydrogen will enter the melt. Water vapour can also be emitted from refractories, slag forming materials and fluxes, which are hygroscopic.

126

Investment Casting

Moisture can have particularly serious consequences for an aluminium melt, due to the large difference in the solubility of hydrogen in the liquid and solid metal. It should be noted that with the humidity on a normal day, at thirty percent, a melt of aluminium at 750°C will contain just over one millilitre of dissolved hydrogen per kilogramme, which is unacceptable and explains the vital need for degassing treatment. The increase of hydrogen solubility with temperature is well recorded. At 1000°C the solubility of hydrogen in aluminium is forty times greater than at normal casting temperatures, which emphasises the importance of avoiding high superheats. Copper based alloys too undergo a variety of reactions with gaseous atmospheres. Contact with water vapour will increase both hydrogen and oxygen contents of the melt. It is the growth of pores in the presence of hydrogen that creates the main problem in the copper based foundry. Alloys containing zinc present a different situation. The low boiling point of zinc causes evolution of vapour with associated environmental prob lems, although brasses are usually free from porosity, due to the action of the zinc vapour in carrying away other gases. Various techniques can be employed to ensure the quality of the melt; all require systematic and careful practice. As one example granulated charcoal can be applied to a copper alloy melt surface, to effect a reducing atmosphere, so that oxides will be reduced to the metal. Alternatively, oxidizing conditions can be maintained during melting, followed by final deoxidation using additions of elements with strong oxygen affinities; a similar approach is applied in much steel melting. Furnace atmosphere control and proprietary fluxes contribute to the development of the appropriate conditions in the melting of many of these alloys. Superalloys are very reactive and the production of clean castings to closely controlled composition usually requires total exclusion of air by vacuum melting, an approach also used for some alloy steels. Vacuum melting and casting will be further considered in a later section. Effects on fluidity Fluidity is a complex empirical property which is influenced by the physical properties and chemical condition of an alloy, being defined and measured by the comparative distance of liquid metal flow in a stand ardised mould passage. The fluidity of the liquid metal has a major bearing on the quality of the finished casting, since it governs the successful filling of moulds for thinwall castings and the sharpness of cast detail. Pouring temperature will always be the prime factor influencing fluidity, but the quality of the melt is also important. The presence of oxides and other intermetallics within a

Melting and Casting

127

melt can dramatically reduce its fluidity. The extent of the decrease has been evaluated for numerous alloys and conditions and a 20% reduction is not uncommon. Cleaner metal allows lower pouring temperatures, producing a smaller grain size and improved mechanical properties; it is thus regarded as essential for investment castings.

ASPECTS OF MELTING PRACTICE Given the above influences of gases and inclusions on soundness and fluidity, clean metal from the furnace is essential for the production of castings of the highest quality. This will have a major bearing on production flow throughout the process, reduce rework to a minimum and increase the viability of the whole operation. Sound melting practice entails the use of various forms of melt protection, including slags, fluxes and atmosphere control, to minimise adverse reactions, and varies widely with the alloy. Alloys melted and cast under vacuum do not contain significant amounts of dissolved gases, so that their solidification normally proceeds without the dangers of precipitation and pore formation. Air melted iron and copper based alloys do not present major problems in this respect given appropriate melting techniques, as compared with alu minium alloys, for the reasons previously discussed. The special treatments accorded to aluminium to ensure high quality in the cast product will be considered in more detail. Degassing and the removal of inclusions Dissolved gases can be removed from molten alloys by the specialised processes of vacuum melting or vacuum treatment, but the most widely applicable technique employs the passage of a cleansing gas through the melt. This allows the impurity gas to partition into the resulting bubbles and to be swept away, by overthrowing the previous equilibrium between gas dissolved in the metal and the external atmosphere. Suspended inclusions can be reduced by agitation of the melt, for example by stirring, in the presence of a cleansing flux; the inclusions are absorbed into the flux which is then separated from the melt before pouring. Degassing and inclusion treatments can be separate or combined. Much attention has been given to the development of these techniques for aluminium alloys; there is no perfect system although certain techniques owe more to tradition than to maximum effectiveness. Tablet degassing is fast disappearing from investment foundries due to difficulties in observing the stringent health and safety laws now in force.

128

Investment Casting

Traditionally this technique was standard throughout the industry. The degassing tablets are a mixture of hexachloroethane, inorganic fluorides and alkali metal salts, which generate chlorine bubbles when immersed in the melt. Inhalation of dust during handling and fume during use need to be prevented by extraction and suitable face masks. Ingestion must be avoided, so that smoking and the consumption of food and drink need to be prohibited in areas where the tablets are being used. Degassing must never be performed with a rising melt temperature, usually being undertaken as the temperature falls following limited superheating. A well heated, refractory dressed, perforated bell plunger is used to position the tablet near the bottom of the furnace or ladle. This allows the bubbles to disperse and rise freely through the melt. When bubbling ceases the reaction is complete. The molten bath is then lightly stirred and the dross removed. Provided that the humidity is low over the molten metal it is beneficial to allow a standing period after completion of the reaction. Grain refiners are often added in a similar fashion. Inert gas bubbling through the melt is commonly employed as a replacement for chlorine generating tablets. The melt surface is usually covered with a flux to prevent continuing oxidation. A lance is lowered to within a few centimetres of the crucible bottom and nitrogen gas with very low moisture and oxygen contents is bubbled through the melt. Degassing times can, however, be over half an hour for a two hundred kilogramme crucible, a disadvantage which has led to a search for more advanced techniques. Although good degassing results can be obtained, another factor which is often overlooked is the excessive energy consumption. Melts need to be taken to higher temperatures before degassing to allow for the fall during the long purging times. A drop of over 20°C is common with a standard bale-out furnace which, with the excessive holding time, can result in a significant increase in energy costs. Rotary diffusion systems, using an inert gas or a combination of inert gas with a low percentage of chlorine, are now used regularly in some investment foundries.5 A rotation speed of about 500 revolutions per minute, with a gas flow of ten litres per minute, produces a cloud of bubbles which is sufficient to degas 200 kilogrammes of aluminium in five min utes. The bubble cloud also greatly facilitates the removal of non-metallic inclusions by flotation. A baffle board floating on and covering the metal surface can be used to prevent excessive turbulence, and so reduce oxide transfer from the surface to the melt. The process, once the parameters are established, is fully automatic in terms of time and rotation speed. Units can either be mobile, servicing several bale-out furnaces, or fixed to a larger bulk melter.

Melting and Casting

129

Flux feeder systems provide another combination effectively used in the industry. In the flux injection process, flux particles are carried below the melt surface in a stream of nitrogen gas.6 This ensures excellent contact between each particle of flux and the metal and gives the maximum opportunity for absorption of suspended inclusions. The stirring action also reduces the degassing time when compared to the standard inert gas lance and a range of fluxes is available to suit the various aluminium alloys. The equipment basically consists of a holding vessel with a calibrated rotary table feeder, leading to a chamber in which finely divided flux is mixed with the inert gas prior to injection. The parameters for each foundry must be set with care and it is essential that the flux is fully consumed during the degassing, since an excessive residue causes dense fumes during dross removal. Establishment of the optimum criteria normally results in a flux consumption of 0.2% of the charge and degassing times of about five minutes for 200 kilogrammes are achieved. Modification of aluminium -silicon alloys Mechanical properties of aluminium -silicon alloys can be considerably improved by the addition of alkali metals, usually sodium. Modification of the microstructure from a coarse needle-like form into a fine structure has a marked effect on properties, elongation of the eutectic alloy being doubled when the alloy is fully modified. However, the modified state is unstable and tends to revert to the unmodified condition at a rate dependent on alloy composition and melt temperature. Sodium is nowadays added in the form of vacuum sealed aluminium containers of specific weights. The content level is important, with an optimum at 0.012% sodium. The addition of sodium must always be made after the degassing treatment, since strong treatments not only remove hydrogen but also reduce the sodium content. Ingot stock is supplied premodified, but with the use of a controlled proportion of returns the degree of modification would be inadequate without a further limited addition of sodium. There is growing use of strontium for modification, as this is more stable in the molten bath and does not fade so readily as sodium. Grain refinement The mechanical properties of aluminium-silicon alloys are enhanced by modification rather than by specific grain refining treatment. For many other aluminium alloy investment castings the solidification rate is such that grain refinement is not essential. Thick section castings in certain alloys, however, require additions of nucleants to the melt to optimise the microstructure. Whilst degassing itself has a mild grain refining effect, a

130

Investment Casting

specific separate treatment is employed for these alloys and must be undertaken before degassing. Further aspects of the control of micro structure are pursued in Chapters 10 and 12. The pouring operation The actual pouring of the metal is the process most likely to introduce defects into the casting. With hand pouring the skill of the operator is paramount to achieve consistent control of the stream above the mould, the rate of pour and avoidance of oxide, flux or slag carry-over. Greater standardisation of conditions is possible with such equipment as rollover furnaces and vacuum filling, whilst one of the most advanced concepts is the fully automatic pour achieved in the self-tapping induction melting unit embodied in some vacuum casting furnaces used in turbine blade production. FURNACES Investment casting production involves many different alloys, pouring temperatures, casting techniques and throughputs. The diverse nature of the industry thus requires a range of furnaces wider than for most other special casting processes. The furnaces employed vary in atmosphere, temperature control capability and degree of automatic control of melting conditions and handling, depending on the working requirements.7 The function of any furnace is to achieve economic melting of the required volume of the chosen alloy and to attain the specified composition and pouring temperature for the manufacture of castings of the appropriate quality. The types of furnace used for air-melt alloys are determined primarily by the casting temperature. Investment foundries are unlikely to require furnaces with high tonnage capacities, but flexibility of alloy requirements and hence of pouring temperature is a major feature of the industry. Mould size, number of castings per batch and quality standards are also subject to greater variation than in other types of foundry: diecasters, for example, have a requirement for bulk metal, usually of one alloy type. Reverberatory and other large bath furnaces are rarely used in investment foundries, due to their large capacities and difficulties in achieving the metal quality required. These and shaft furnaces are mainly suitable for foundries that require large volumes of a single alloy type and do not meet the flexible requirements of the investment casting industry, where most melting is carried out in crucible or coreless induction units: these types will be examined in more detail.

Melting and Casting

131

Radiant heat furnaces Radiant heat furnaces used in investment casting are normally of the crucible type and the following comments will refer to furnaces other than those classified in the second major group, namely the induction furnaces. A prehistoric dish discovered by an Egyptologist and dating from several thousand years BC gave early evidence of crucible melting. Scientific development of pre-formed crucibles from the basic clay-graphite type has mirrored other advances in the foundry industry during the past century. There are now crucible furnaces suitable for a wide range of alloy types and their associated pouring temperatures. All the modem heat sources, based on electricity, gas and oil are available.8'9 Capacities vary from a few grammes of precious metal to almost two tonnes of copper alloy from lip axis tilting furnaces, although the largest of these are not normally found in the investment casting industry. The main types of radiant furnace which use preformed crucibles are lift-out, baleout, tilting, immersion tube, and immersed crucible. Again, the last two are seldom used in the context of investment casting. The advantages of preformed crucible melting include low melting losses, consistent metal quality, flexibility, ease of installation and, as compared with other melting techniques, modest capital cost. Manufacturers have developed several important features of these furnaces to meet the requirements of the quality-conscious foundry. Crucible technology is now advanced to the stage where life expectancy and degree of contamination are predictable.10 Low thermal mass insulation is now commonplace, providing greater energy efficiency and ease of maintenance.11 Close temperature control is a vital feature and various systems are available depending on the exact requirements of the foundry.12 Some crucibles are supplied with a moulded-in pyrometer pocket and find extensive use in aluminium and copper alloy casting. Although this avoids the use of a floating pyrometer in the melt, it does suffer from a lag in response time. Simple on-off controllers are only found in foundries where temperature control is not critical or on old equipment. Proportional controllers are nowadays invariably specified for furnaces that are to be used in the production of quality castings. Heat sources have been developed over recent years to ensure efficient use of energy and even heating, which facilitate the control of pouring temperature. It is now common for a complete record to be maintained of the melt sequence and pouring temperature. It is important to consider the necessary treatment of crucibles to ensure maximum efficiency. Prior to insertion it is essential that a close inspection be undertaken since many failures have their origins in poor storage. Crucibles must be stored individually, never stacked, in a warm and dry area.

132

Investment Casting

Installation is critical to crucible life. Fuel-fired furnaces require correct alignment of the crucible to the flame, since misalignment can cause premature failure. It is essential that the correct crucible be purchased for the particular heat source; there are, for example, different crucible requirements for electric and fuel-fired heating, the former generating a lower but more even temperature. The respective crucibles require different outer glazes, which if confused would cause rapid failure. Chemical analyses and physical properties of common crucibles are given in Table 1. During the initial heating up and glazing period, thermal shock must be avoided. Eagerness to get back into production may mean that the manufacturer's recommended initial heating cycle is not followed, causing premature failure, whilst mechanical damage can result from careless handling during use; impact cracking, poor charging techniques and the early introduction of low melting point fluxes will all shorten crucible life. Between melts crucibles must be scraped clean of dross whilst still hot, since failure to do so will lead to reduced capacity, inefficient heat transfer and eventually failure due to differential expansion of the hardened dross layers. It is important to remember that all crucibles age with use and that their heat transfer properties deteriorate. This gradually reduces the melting rate of the furnace and increases energy consumption. While this occurs in all types of pre-moulded crucible furnace, it is less likely to be noticed with fuel firing since crucible failure may occur before the effect becomes pronounced, unlike the situation in an electric furnace in which crucible life may be extensive. Replacement may thus become necessary while the crucible is otherwise still sound. Silicon carbide crucibles deteriorate, in terms of heat transfer properties, at a slower rate than traditional clay-

Table 1. Composition and properties of typical crucibles Chem ical analysis

(Wt %)

Silicon carbide

d a y graphite

31.3 41.3 16.5 5.4 3.0 2.5

35.2 12.3 35.0 13.5 4.0 -

2.61 29.0 1.86 4.6 19.9

2.53 28.0 1.80 80.0 21.6

(Wt %)

Carbon SiC Si02 a i 20 3 Fe20 3 B20 3

(Wt %)

Physicalproperties

App. solid density App. porosity, % Bulk density Electrical resistivity (Ohm.cm x 10-3) Thermal conductivity (Watts/mK) Information courtesy of Morgan Thermal Ceramics Ltd.

Melting and Casting

133

graphite crucibles, so that their extra cost may well be justified on the grounds of consistent melt rate and energy savings. Lift-out crucible furnaces have capacities of up to 150 kilogrammes of cop per alloy or 60 kilogrammes of aluminium. They have the advantage of flexibility, both in the varied alloys that can be melted in a single furnace and in the throughput, separate crucibles being kept for each alloy type. Smaller lift-out furnaces are floor mounted, whilst larger furnaces require installation in a pit. These furnaces are usually fuel-fired and are capable of melting alloys up to 1150°C. Crucibles can, in the case of small furnaces, be hand drawn using specially designed tongs, whilst larger crucibles are mechanically handled using hoists. Bale-out furnaces are used throughout the aluminium and copper investment casting industry. They can be employed either as melting units or as holding furnaces. In the melting mode it is possible to change from one aluminium alloy type to another but this requires a rigid quality system since, unlike lift-out furnaces, the crucible is fixed in place until the end of its specified life. Suitable alloy changes require excellent cleaning, as for example when an LM25 follows an L99 melt. Given this care they offer a degree of flexibility at relatively low capital cost. The bale-out furnace, as its name suggests, is emptied using a ladle. The furnace is very simple in concept, consisting of a crucible supported on a stand and surrounded by the heat source. A typical example of a gas fired unit is shown in Fig. 1. Electric resistance furnaces use metallic, wirewound heating elements carried on refractory supports or semiembedded in the hot face lining. Some designs employ silicon carbide elements which are self-supporting. To ensure long life, elements are normally operated at 100 -200°C below their design maximum. Where melt temperatures of 850°C or more are necessary for the production of high silicon aluminium alloys or the lower melting point copper based alloys, silicon carbide elements are employed. These are either conven tional or spirally cut tube sections or high grade open wire coil elements. Foundries using bale-out furnaces as melting units should consider the use of the highest rate non-embedded elements. Element life is usually extensive unless affected by mechanical damage during crucible changes or by the most common cause of failure, liquid metal attack through overfilling and spillage from the crucible. Excessive use of salt tablets for fluxing can have an adverse effect on element life. The development of fuel-fired burners has gained pace over recent years. The expanding flame, high turndown ratio package burner allows some of the freedom enjoyed by electrical resistance heating. To obtain maximum efficiency from these burners it is essential to have the correct

134

Fig 1

Investment Casting

Gas crucible bale-out furnace (Courtesy of Striko UK Ltd.)

rating. A bale -out furnace which is liquid fed from a bulk melting furnace requires a much smaller burner package than a melter/holder. Specifying an over-rated burner on the basis that the furnace may be needed as a reserve melting unit, and then running as a liquid fed holder, is totally uneconomic. A further major factor in the efficient running of fuel-fired furnaces, particularly bale-outs with their relatively small capacities, is the maintenance of the correct air/fuel ratio. For these simple furnaces it is uneconomic to fit the sophisticated electronic combustion control systems used on larger heat input equipment; the same is true of exhaust gas analysis automatic trim systems. Package burners, once set to the optimum air/fuel ratio, maintain this for long periods and a yearly check on the products of combustion for carbon dioxide, oxygen and even for the low levels of carbon monoxide, suffices to ensure efficient energy use.

Melting and Casting

135

These checks will alleviate worries concerning emission of combustion products into the foundry. These simple furnaces have excellent insulation, but regular measurement of the outer shell temperature provides an important check on the state of the insulation and the overall efficiency of the furnace. A quality bale -out furnace, whether electric or fuel-fired, should have an outer shell temperature of less than 35°C at working temperature although this will rise slowly with time and use. Should it reach 100°C, then this would be the equivalent of an extra 7kWh heat loss for a 175kg furnace and melt times will also be adversely affected. The larger the furnace the more noticeable the heat losses. Bale-outs with high shell temperatures are not uncommon within the industry and explain complaints about excessive energy bills and poor melt rates. Foundries running bale-outs containing refractory bricks rather than low thermal mass linings should therefore examine the feasibility of an insulation update. For fuel-fired furnaces the burner may then need to be downrated by 10 to 15%, since a melt rate increase of approximately 15% and an energy saving in excess of 25% can be expected from a conversion to low thermal mass insulation. Similar figures will be obtained for most types of furnace, but the ease of change for a bale-out furnace makes it sound economic sense. The preference of the investment foundry industry has been towards electric resistance as the energy source. The growth of resistance bale-out furnaces can be summed up by the fact that only 3% of all bale -out sales in 1975 were electric, whereas by 1985 this had risen to 66%, although recent developments in fuel-fired furnaces have since produced an even split of sales: this general foundry trend is even more emphatic in the investment casting sector. Crucible tilt furnaces are similar in construction to bale-out furnaces but are larger and more substantial. They are essentially melting units and are rarely used in a holding capacity. The metal is transferred to the mould in a ladle or preheated crucible, except in special cases involving large investment moulds, which may be poured directly from the furnace. Some foundries may still be operating with central axis tilting furnaces, but these have the disadvantage that the pouring point moves during tilting and lip axis pouring has now become standard. Tilting is controlled by twin hydraulic rams to provide a smooth stream of liquid metal, even at low pouring speeds. A key performance criterion for foundries using this type of furnace is the melt rate, in respect of which modern fuel-fired furnaces have an advantage over electric resistance furnaces. Comparisons are, however, often made on the basis of semi-embedded resistance elements as com pared to high velocity burners, whereas these elements are essentially for

136

Investment Casting

holding operations and not for melting. Melt rate increases with crucible capacity: for aluminium the approximate rates for a 200kg crucible are 220kg per hour for a fuel-fired and 145kg for an equivalent electric resistance furnace. For a 550kg crucible the rates are 340 and 250kg per hour respectively. The melt rate advantage is, however, offset by the extensive noise and products of combustion removal problems associated with fuel-fired furnaces of this type. A key factor affecting the melt rate of fuel-fired furnaces is the heat flux from chamber to crucible. The convenience and economics of a robust package burner operating on the simple nozzle mix principle will give a performance adequate for many foundries. Accurate air/fuel ratio control and good sealing at the burner are essential for economic performance; peak flame temperature, and therefore chamber temperature, requires an air/fuel ratio close to stoichiometric and it is essential for high melt rates and energy efficiency that high velocity burners do not entrain significant amounts of cold air into the furnace chamber. Unlike bale-out furnaces, the larger tilting furnaces can benefit from air preheating, either by recuperation or by the use of the regenerator principle. Flue stack recuperators have been shown to improve energy efficiency and require a balance between air preheat and recuperator life. Recuperative burners are seldom used, although several trials have shown interesting results; one problem highlighted in such trials is that under the unusual condition of catastrophic crucible failure metal can enter the burner. Regenerative burner systems are not fully commercially established but final exhaust temperatures are low with high air preheat and trials have shown major energy savings over standard equipment; it has yet to be seen whether this advantage will be accepted by the industry as sufficient to offset the additional capital costs. Other radiant heat furnaces, including fossil fuel fired ceramic immersion tube furnaces,13 have been employed in some aluminium investment foundries. These were designed to meet the requirements of modern foundries for thermal efficiency, low metal loss, accurate temperature control and minimal operator fatigue. By immersing the ceramic tube in the molten metal the heat transfer is by direct conduction instead of radiation and the tube can operate at a temperature close to that of the melt. Wide use was, however, hampered by inconsistent tube life and premature failure, although silicon nitride bonded silicon carbide tubes are now standard and extensive developments are in progress to ensure practicable life, including use of a zircon glaze. These developments may aid the réintroduction of such furnaces, whether with resistance elements or recuperative burners.

Melting and Casting

137

Electric induction furnaces Induction furnaces are widely used throughout the entire spectrum of the investment casting industry. Small units with capacities of up to a few kilogrammes are used for high quality dental and jewellery castings, whilst tilting, roll-over and push-out furnaces are used for numerous alloys and in many foundries.14 Although coreless furnaces now range in capacity up to many tonnes, melting for investment casting is mostly carried out in units below 1 tonne. Vacuum melting furnaces too normally use induction heating in the foundry application. The principle of induction melting has been known for over one hundred years. When an electrically conducting material is placed in an alternating magnetic field, eddy currents induced in the material generate heat. This effect was first employed for metal melting at the end of the nineteenth century, when a primitive channel furnace was developed. This, like its successors, relied on maintaining an unbroken circuit of metal around an iron core, so that such furnaces tend to be of large volume and to be used for single alloys.15 They are thus unsuitable for foundries which need flexibility and batch production. Coreless induction furnaces as used by investment casters do not have this disadvantage. They rely on the same principle but have a simple cylindrical configuration with the features illustrated in Fig. 2. A suitably supported water-cooled copper coil surrounds a refractory lining or cruc-

Fig 2 Section of coreless induction furnace (Courtesy of EA Technology)

138

Investment Casting

ible which contains the charge. Coreless furnaces can be powered at almost any frequency, although most operate at between 200 and 3000 Hz. Optimisation of design, allied with the use of medium frequency, has allowed the power input into relatively small furnaces to be developed dramatically.16'17 A 500 Hz furnace will have the same output as a mains frequency furnace three times its size. Specific power ratings have been increased over the past few years to levels approaching 1000 kW per tonne. Induction furnaces have two distinct features when compared with fuel-fired and resistance heated furnaces. With the latter types the heating rate is proportional to the temperature difference between the heat source and the metallic charge, so that as the charge gets hotter the heating rate is reduced; the opposite occurs in induction melting. In an induction furnace the major variables are the electrical resistivities of the charge and the coil. Resistivity of metals increases with temperature, whilst the furnace coil is kept at a relatively low temperature by the cooling water. Thus as the charge heats up, melting efficiency increases, and the solid metal is at a high temperature for only a short time before it melts. This accounts for the very low metal losses recorded in induction melting. The fast melting times enable single mould melting systems to be econom ically viable.18 The second distinctive feature of induction melting is the beneficial stirring action in the molten metal as typified in Fig. 2. The vigour of this action is directly proportional to the power supplied to the furnace and inversely proportional to the square root of the operating frequency. Modem power supplies to induction furnaces are generated by solid state electronic inverters or converters of various designs. Progress in the manufacture and utilisation of power semiconductors, which are the critical components of converters, is continuing. Motor generator converters as once supplied were at best seventy-five percent efficient, whereas solid state converters are now over ninety-seven percent efficient. The progress in converters has resulted in the ability to supply constant power from the initial heating of the cold charge until the metal is ready to pour. The converter is so regulated that power input to the furnace is kept constant by automatic adjustment of voltage, current and frequency according to load conditions. A nominal 1000Hz supply will produce power at frequencies between about 750 and 1400Hz throughout the melting cycle. Foundries have even found it possible to melt steels and aluminium and copper alloys in the same furnace using the same power supply. This may offer a useful capability to meet urgent customer requirements or to evaluate new markets for castings but is not the most efficient method of operation and brings a risk of cross contamination between melts. Many foundries need to produce castings in dissimilar alloys on a regular basis and so use different furnace

Melting and Casting

139

bodies for each group of alloys. These can still be powered from a single supply, switched from one to the other as needed. The flexibility of the induction coil leads to numerous furnace designs covering the range from a few grammes to several tonnes. Essentially the furnace bodies are of two sorts; those employing pre-formed and removable crucibles and those with fixed linings. Removable crucibles are used when a foundry needs to melt small batches of different alloys which pose problems of cross contamination if melted in the same linings. They also enable the most suitable refractory to be selected for each type of alloy, and allow crucibles to be kept for the melting of individual or closely similar compositions. Fixed linings tend to be preferred for medium to larger batch melting and are normally used where tilt or rollover pouring is employed. Push-out furnaces use crucibles which sit within the induction coil, on top of a hydraulic ram. Once the melt is complete the ram raises the crucible clear of the coil, whence it can be carried by hoist or handshank to the pouring position. Push-out furnaces can be single melting units, but a more usual arrangement has two melting coils fed from a single power supply. As the first crucible completes its melt cycle and the moulds are being poured, the second can be heating the next batch of metal. Capacities are usually 100kg or less but some furnaces take 300kg crucibles. A drawback of this type of furnace is the excavation required to sink the hydraulic system into the foundry floor. The furnace is mounted at a height at which the raised crucible can be most conveniently handled. For high quality investment cast jewellery small single station push -out furnaces are sometimes used. The crucible is in this case at waist height with the hydraulic ram below. Lift-coil furnaces are used to overcome the need for excavation, the crucible being kept stationary at floor level and the induction coil removed. The simplest system is a manually operated hoist to enable the furnace coil to be transferred from one melting station to another. An alternative is the lift and swing system, where two stations are serviced by a coil mounted on a pillar, the coil being hydraulically raised, swung from one station to another and lowered as required. An electrical interlock automatically disconnects the power when the coil is raised and machined guide slots position the coil accurately over the crucible. Capacities are typically up to 100kg. For precious metals small lift coils have handles to facilitate manual lift-off with gloves. Drop-coil furnaces have the crucible on top of a pedestal and the induction coil assembly is raised over it to melt the charge. These are single station

140

Investment Casting

furnaces, coil movement being usually controlled by a hand valve operating pneumatic or hydraulic cylinders within the support posts. When the charge is fully molten the coil assembly is lowered and the crucible removed with tongs; power is automatically disconnected when the coil is lowered. These furnaces are often used to melt small quantities of high temperature alloys or precious metals. They offer excellent melt rates and thus very low metal losses. The melt rates for silver and gold are 1.5 and 2.6kg per minute respectively in a typical five kilogramme furnace. Rollover furnaces are the only type specifically designed for the investment casting industry and an example is illustrated in Fig. 3. The induction melting furnace is in this case mounted on a rotating frame. When the pre-weighed charge is fully molten, a heated mould is inverted and clam ped over the open top and the entire furnace/mould assembly is rolled over. The filled mould is then in the correct position for feeding and can be easily removed. The assembly is returned to the start position to repeat the cycle. Mould filling is enhanced by the added force obtained during rollover. Rollover speed is adjustable down to one second with an appropriate hydraulic power unit. Hand operated rollover systems provide reproducible results with a maximum speed of two seconds. Acceleration and deceleration cycles are automatic and provide cushioned stops in each direction. Moulds of different sizes can be accommodated by adjustment of the hydraulic clamping bar and the use of a ceramic fibre pad between the clamping bar and the hot mould limits the breakages that can occur with this technique. These furnaces are constructed to give protection to the induction coil and to provide a body of sufficient strength to withstand mould clamping stresses. A preformed crucible of alumina or magnesia is firmly mounted inside the coil with a rammed backup refractory; the crucible material needs to be discussed with the refractory supplier to suit the particular alloy type. The crucibles have a limited campaign life, much shorter than that of a rammed monolithic lining, but relining can be undertaken quickly and without highly skilled staff. The crucible is preheated to some degree before use and can be subsequently scraped clean between melts. Most rollover furnaces have capacities of 7 to 30kg but major manufacturers offer furnaces of up to 50kg. Melt rates are sufficiently fast to make the melting and pouring of charges for individual clusters, or trees, of castings economic. Hand pour furnaces are used by manufacturers of jewellery and dental castings who require frequent and rapid melting of small quantities of alloy. These are fitted with a fixed integral crucible and the furnace case is fitted with side handles for one man pouring. For larger volumes a forked shank is attached to the furnace for easier pouring.

Melting and Casting

Fig 3 t5kg rollover induction furnace (Courtesy of Inductotherm Europe Ltd.)

141

142

Investment Casting

Tilting furnaces are manufactured in sizes up to five tonnes. The heavier capacity is typified in the vacuum furnace illustrated in Fig. 4 and used for bulk superalloy production. For precious metal and copper alloys a trunnion tilt-pour furnace of a few kilogrammes capacity can be hand tilted from a stanchion table as shown in Fig. 5, whilst small hydraulic or pneumatic tilting furnaces are also available from many suppliers, with capacities ranging from seven to thirty kilogrammes of steel or equivalent volumes of non-ferrous alloys. A 500kW power supply at 1000Hz, melting steel to 1650°C, will give a typical melting rate of 800kg/h in a 500kg furnace. A type of tilting furnace which meets the requirements of precise pouring directly into an investment mould is the double axis tilt furnace. In operation the furnace body tilts about the first axis, which gives forward reach of the spout to the pouring position. When pouring commences further tilting is about the second axis, so keeping the spout in the same position to give a true lip axis pour. These furnaces are primarily confined to smaller outputs. Standard lip axis tilters are intended for use as melting furnaces to feed ladles rather than for pouring directly into moulds. Refractory linings are normally of the monolithic type, for which there are several types of refractories available depending on the alloy to be melted.19

Fig 4 3 tonne,

1750 kW vacuurn induction furnace for superalloy production (Courtesy

of Induct other m Europe Ltd.)

Melting and Casting

143

Fig 5 Pouring from 25kg hand table induction furnace (Courtesy of Inductotherm Europe Ltd.)

144

Investment Casting Table 2. Typical lining refractories for coreless induction furnaces

Silica Grain size (mm) Up to 4 Bulk density (kgnrr3) 2200 Chemical analysis (Wt %) 99.2 Si02 a i 2o 3 0.8 MgO 0 Fritiing temperature Up to 1600 (°C)

Alumina

AluminaMagnesia

MagnesiaAlumina

Up to 5 2950

Up to 5 2900

Up to 4 2550

4.1 95.9 0 Depends on grade; 650 1000 for Al

0.1 84.4 15.5 1680-1720

2.0 18.0 80.0 1550-1650

Information courtesy of Hepworth Refractories Ltd.

Silica linings can be used with most irons, and for carbon and low alloy steels as well as copper based alloys. Although this type of lining is dense, strong, and resistant to metal attack and thermal shock, its use is declining because of the need for cleaner castings. Alumina linings can be used to melt iron, copper and aluminium alloys. They are more expensive than silica linings but produce cleaner castings and usually give longer life. Basic linings are preferred for some types of alloy steel but are less suited to intermittent use involving frequent cooling to ambient temperature. Table 2 compares some of the physical and chemical properties of the main types of linings. Installation techniques are mechanised for larger furnaces, with hand ramming still commonplace for furnaces below 500kg. This method is very operator sensitive, with the possibility of variations in rammed lining density and laminations; the campaign life can thus be seriously affected. The use of hand held pneumatic rammers and more recently electric/ pneumatic former vibrators has significantly improved lining life. The fritting procedure depends on many factors, including furnace size, power rating and the type of lining and bonding. If long campaign lives are to be achieved, the lining must be fritted-in correctly and to the manufacturer's specification. The recommended fritting cycle for silica linings demonstrates the time required: heat at 110°C per hour up to 1600°C (or 30 - 50°C above the normal operating temperature), then hold for a minimum of one hour before proceeding further with production. SPECIALISED CASTING TECHNIQUES Diverse requirements for investment castings have led to the development of various special melting and casting techniques beyond the simple

Melting and Casting

145

air melting, gravity pour system as inherited from the traditional sand foundry. Some are aimed at the protection of the melt from atmospheric contamination and others at minimising turbulence by control of metal transfer from furnace to mould. Further systems are designed to assist feeding or to control the development of microstructure during solidification. Some of the special techniques and associated equipment combine these aims to establish exceptional product quality; although the principles are not exclusive to the investment casting field it is often here that the associated cost premiums can be most readily absorbed. Vacuum and centrifugal casting are two techniques embodying these principles and both have strong associations with investment casting. Vacuum melting and casting The enclosure of the melting unit, and in some cases the whole casting operation, in a sealed chamber permits these procedures to be carried out in a closely controlled environment.20'21 The process usually employed for investment casting is vacuum induction melting, in which the power is transmitted to a specially insulated melting coil designed to prevent short circuit discharges under the low pressure conditions. Vacuum can be maintained throughout the process or the chamber may be backfilled with inert gas for part of the cycle. The principal advantage of vacuum melting lies in the prevention of oxidation losses of reactive alloying elements, and of contamination by oxygen and nitrogen, whilst hydrogen from water vapour is similarly excluded. The process is extensively used in the production of gas turbine components, for engines ranging from ultra-small units for remotely piloted vehicles, through aero and marine applications, to very large landbased units for power generation. These all require high integrity blades and other components in heat resisting superalloys, for which clean conditions are ensured by both melting and casting under vacuum. Manufacturers have developed various forms of furnace for this purpose, with similar basic features. The vacuum sealed melting chamber is separate from the mould chamber, as exemplified in Fig. 6. Melt stock can thus be added to the bath without destroying the vacuum, and moulds can be similarly loaded for casting without breaking the main vacuum. Pressure is an important operating variable, to be controlled at all stages through comprehensive pump and gauge systems: typical operating pressures are below ICh3 mbar. A typical melt cycle involves the three stages of melting, superheating and pouring, the sequence being carefully determined for each alloy and component type. Heating to the pouring temperature is undertaken un der vacuum to prevent surface oxidation of the charge, a rapid melting

146

Fig 6

Investment Casting

Schematic of large precision vacuum casting furnace for components up to 1200mm diameter (Courtesy of Ley bold Durferrit GmbH.)

Melting and Casting

147

rate being desirable; a typical rate of 10kg per minute is commercially feasible with induction melting. Control is critical at the pouring stage since the molten metal is then at its most reactive and a large surface area is exposed. Human error is reduced wherever possible by control and automation of the main operating variables. Initial heating is usually by a programmed power/ time function up to about 900°C, at which point the control switches to measured temperature of the melt. Several stabilization steps may be programmed to allow pouring lip heating. The temperature is raised to a superheat level and allowed to homogenize; after a controlled soak the power is turned off, allowing the melt to cool to the pouring temperature. Pouring within the furnace can employ a simple remote controlled tilt system or self tapping through the crucible base, whilst in some cases the casting arrangement also incorporates differential heating of the mould and a mould withdrawal mechanism for controlled directional solidification; the use of this principle for the production of columnar or single crystal structures is fully examined in Chapter 12. Such furnaces are fully automatic, with computer control of the main parameters. The scale of the comprehensive melting and casting facilities now available in the advanced field of turbine blade production is indicated by the illustration in Fig. 7. Pouring and feeding; centrifugal casting Both in air and vacuum casting various unorthodox systems are employed to achieve favourable conditions of mould filling and feeding. Apart from the bottom pour arrangement already mentioned, the rollover principle is used to assist filling, whilst direct pressurisation of the system is also used in some cases to enhance both definition and soundness. The CLA process, using vacuum to induce upward filling of the mould cavity, is a further example of the departure from reliance on gravity. Centrifugal casting is a long-established technique for assisted mould cavity filling and enhancement of feeding. Molten metal introduced into a mould cavity which is spun about an external axis of rotation is subjected to very large forces, proportional to the radius of rotation and the square of the angular velocity. These drive the metal into the fine details of the mould and maintain the casting under high pressure during solidification. Centrifugal casting has been used in conjunction with most types of investment casting, but is most strongly associated with dental and jew ellery applications with their requirements for the rapid pressurised filling of small mould cavities with thin passages and intricate features.

148

Investment Casting

Fig 7

Production line with vacuum furnaces for cast turbine blades (Courtesy of RollsRoyce pic.)

Horizontal and vertical axis machines with integral melting units are designed for specific functions: these are fully discussed and illustrated in the specialised sections of Chapter 12.

Melting and Casting

149

REFERENCES 1. B. Drinkall: Foundryman, 84, Jan. 1991, 7-11. 2. G. Morley: Natural Gas, July/Aug. 1991,12. 3. The properties and characteristics of aluminium casting alloys, 1980, Alcan Enfield Alloys Ltd, St. Albans. 4. V. Kondic: Metallurgical Principles of Founding , 1968, 17-32, Edward Arnold, London. 5. S. Sibley and R. Dean Foundry Practice (217), April 1989,20 -21, Foseco International Ltd, Birmingham. 6. A Flux Degasser on an Aluminium Bale-Out Furnace, ED /62/94, Sept. 1985, Energy Technology Support Unit, Harwell. 7. 'Melting/Refractories', Section B, Foundry Management and Technology, Dec. 1990,1 -48. 8. C. Edgerley et al.: Cast Metals, 1, No. 4,1989, 216-222. 9. D. Rachwal and D. Pennington: Foundryman, 85, Feb. 1992, 51-60. 10. Your Crucible, Morganite Thermal Ceramics Ltd., Ref. No. IC099. 11. Improving Bale-Out furnace Performance with Low Thermal Mass Insulation, Energy Efficiency Demonstration Scheme, Jan. 1985, Energy Technology Support Unit, Harwell. 12. R. Atkins: Metals and Materials, 7(1), Jan. 1991,19 -23. 13. 'Melting and Holding Furnaces', Foundry Trade Journal, 165, Jan. 11/25 1991, 45. 14. H. Hellerling et al.: International ABB-Conference on Induction Furnaces, Dortmund, Germany, April 1991,13 -45. 15. Energy Efficiency in the Foundry Sector, Proceedings of European Seminar, San Sebastian, Spain, May/June 1990. 16. 'Induction Feature', Electricity Business News, Summer 1992, 20-27. 17. H. Heine: Foundry Management and Technology, Feb. 1990, 28-33. 18. Guidance Notes for the Efficient Operation of Coreless induction Furnaces, Good Practice Guide, 1992, Energy Technology Support Unit, Harwell. 19. S. Thorpe: The Installation and Selection of Refractories for the Lining of Coreless Induction Furnaces: Hepworth Minerals and Chemicals Ltd. 20. D. Pratt: Material Science and Technology, 2, May 1986, 2, 426-435. 21. G. Bouse and J. Mihalsin: Superalloys, Supercomposites and Superceramics, 1989, Chapter 4, Academic Press Inc., USA.

6

Gating and Feeding Investment Castings T.S. PIWONKA

Success in investment casting is largely dependent on the ability to make good castings without making scrap. To do this the foundryman must design the casting process — the arrangement of the patterns on the sprue, gates and feeders, pouring temperature and pouring speed, and mould preheat temperature — with skill. This design activity is called 'methoding,' or 'gating' the casting. The mould geometry and pouring parameters must be selected so that the metal can enter the mould and fill it rapidly before freezing. The metal must not react with oxygen in the air, to form non-metallic inclusions, while it is filling the mould. Finally, the metal must solidify without forming voids, so that the casting is 'sound.' The subject of this chapter is the methods used to design the mould so that these conditions are met, employing principles of solidification, heat flow, and fluid flow. PRINCIPLES OF SOLIDIFICATION Castings are made by solidifying liquid metal in a mould. The control of the solidification process is achieved by the design of the mould. Before discussing gating and feeding principles, it will be helpful to review some of the aspects of solidification which gating design can influence. During solidification liquid metal is cooled in the mould and freezes, forming a solid casting. When this happens the metal atoms, which are arranged nearly randomly in the liquid, take up regular positions in the solid lattice. As they do so, they give off energy, which appears as the latent heat of fusion. They also occupy less space in the solid than in the liquid, causing the casting to shrink if this difference in volume is not compensated (this shrinkage is metal shrinkage, which can result in voids in the casting, not 'patternmakers shrinkage', which refers to the contracDOI: 10.1201/9781003419228-6

Gating and Feeding Investment Castings

151

tion of the solid casting during its further cooling and can be compensated for by the use of an oversized pattern). For a pure metal, solidification takes place at a specific temperature. Figure 1 shows a typical cooling curve for aluminium. The liquid alumin ium cools until solidification takes place at 660°C. In fact, the liquid un dercools a few degrees, then returns to the melting point and the temperature does not change further until all the liquid has solidified. During this time crystals of aluminium grow from the walls of the crucible or mould in to its centre. These crystals are called dendrites. When the metal is solid, the temperature will continue to fall to ambient. As the casting solidifies, the level of the metal in the container falls and when the casting is solid a cavity will be found in the centre of the metal, showing dramatically that the solid is denser than the liquid and that the metal shrinks on solidification. Low carbon steels, which are nearly pure iron, solidify in this manner. Most castings are made not of pure or nearly pure metals but of alloys which form solutions in the liquid and solid states. When alloys solidify they do so over a range of temperatures. Thus, if a cooling curve is determined for an aluminium - 4.5% copper alloy (Fig. 2) the alloy will

Fig 1 Cooling curve (temperature vs. time) of pure aluminium.

152

Fig 2

Investment Casting

Cooling curve (temperature vs. time) of Al - 4.5% Cu alloy.

start to freeze at 640°C and, under commercial casting conditions, will not be completely frozen until the eutectic temperature of 548°C is reached. In this case, instead of forming dendrites which grow in from the sides of the container, the metal solidifies by forming many small crystals throughout the liquid. These crystals grow into the liquid, which surrounds them as they cool. Both liquid and solid therefore exist as a mixture, which is mushy or pasty in this temperature region and is called the 'mushy zone'. Most non-ferrous alloys, stainless steels, and superalloys solidify in this manner. The individual grains, shown in Fig. 3, are also dendritic in structure. The dendritic structure consists of a primary arm, which has secondary arms growing from it (there may also be tertiary arms which have grown from the secondary arms). The average distance between these secondary arms is known as the 'dendrite arm spacing'. The closer the spacing the easier it is to heat treat the casting and the better, in general, the mechan ical properties of the casting. As the grains grow, their chemistry changes. The first metal to solidify, which is in the interior of the grains, is the purest (it begins to solidify at

Gating and Feeding investment Castings Interdendritic Liquid

153

Secondary Arms

Primary Arm

(a)

(b)

Fig 3 (a) grains solidifyingfrom liquid metal; (b) enlarged view of primary dendrite arm. the highest temperature). In aluminium - 4.5% copper alloy, the first solid metal has a composition of only about 1% copper in solid solution. The liquid which surrounds it is thus depleted in aluminium atoms, and therefore enriched in solute copper atoms. The solid which forms around the first metal to solidify is increasingly rich in solute (copper). In most cases, however, as the end of solidification approaches, the liquid remain ing is so rich in solute (for this alloy, the last liquid to freeze contains 33% copper) that it freezes by forming a eutectic solid — which is a mixture (not a solution) of two separate phases, each of which is a solid solution (for aluminium -copper alloys, one phase contains 5.7% copper and the other 47% copper). Commonly encountered eutectic phases are: graphite in cast iron, silicon in aluminium -silicon alloys, and gamma prime in superalloys. This progressive solute enrichment of the liquid which freezes as the temperature falls is what causes chemical segregation in alloy castings. There are local small changes in the composition, which can add to the difficulty in heat treatment, because time is needed to reach homogeneity by high temperature diffusion of atoms. Molten alloys also dissolve gas readily. When the alloys freeze, this gas is rejected from solution because gas is much less soluble in solids than in liquids. Figure 4 illustrates the case of hydrogen in aluminium, and similar curves may be obtained for hydrogen in iron and nitrogen in iron. The gas may appear as bubbles if the casting solidifies progressively from one end to the other and the gas may float out of the casting, or it may appear as voids in the solid casting if the bubbles cannot escape. A common

154

Investment Casting

Fig 4 Solubility of hydrogen in aluminium. Note how solubility decreases at freezing point, 660°C. source of gas is the moisture in the atmosphere, which is why it is harder to produce sound castings on a hot, humid day in summer than on a clear cold day in winter. One consequence of the mushy or pasty nature of alloy solidification is that the liquid which remains is present in channels between the solid dendrites. As solidification progresses, these channels become smaller. As the metal continues to shrink (as it solidifies) the liquid in the channels must feed the shrinkage. However, as the channels become narrower, movement in them becomes increasingly difficult and some channels become completely blocked by solid, so that liquid in them can no longer feed the shrinkage. This leads to small, dispersed pores (microporosity) in the casting and reduces mechanical properties. Solidification is a nucléation and growth process. Metal crystals, or grains, which may be dendritic, are nucleated, often with the help of nucleating agents added to the melt or the mould, and then grow until they meet other growing crystals, run out of liquid, or are stopped by a mould wall. The foundryman can exercise a certain amount of control over the degree of nucléation, and therefore over the number of grains in the

Gating and Feeding Investment Castings

155

casting. (In practice, it is the size of the grains, not their number, which is usually specified.) Grain refining agents can be added to the melt (for example titanium -boron in aluminium alloys, or ferro -silicon in cast irons), or applied to the mould wall, e.g. the addition of cobalt aluminate to the prime dip in nickel based superalloys. The actual geometry of the eutectic phases can be controlled by adding components which modify the shape during solidification. Examples of such additives are sodium or strontium in aluminium alloys, and magnesium in spheroidal graphite iron. The addition of these grain refiners and eutectic modifiers affects the course of solidification, and the rate at which the latent heat is released. The effect is not great, but it can have subtle effects on the casting structure. The objective of the foundry engineer is to control the nucléation and growth process. Of particular importance is the control of points at which solidification begins, and of how the solidification 'front' (the dividing line between liquid and solid) moves through the casting. This is done by controlling the way in which metal enters the mould cavity and the way heat is removed from the casting. By these means both the solidification process and the quality of the casting are controlled. FLUID FLOW AND GATING DESIGN Gating system design depends on understanding fluid flow. This is quite difficult to describe mathematically because it is three dimensional, and often transient. It involves velocities (which are vector quantities) and, in the case of molten metal, a basic material property, viscosity, which changes as the metal cools. However, most foundry engineers have a good understanding of fluid flow, gained from observing water flow in a stream, or snow fly over a snow fence. Thus most are acquainted with such concepts as turbulence, eddies and laminar flow. In considering fluid flow, it is important to differentiate between steady state conditions, in which everything is constant with time, and transient, or unsteady state conditions, in which many variables change with time. In metal casting, conditions are almost always transient — the runner and sprue are not full at the beginning of pouring, and when they do fill up, the level of metal in the casting cavity changes continuously until it fills and fluid flow stops. Unfortunately, transient conditions are even more difficult to analyse than steady state conditions, and well beyond the capabilities of all but the most powerful and sophisticated computers. In designing gating systems, therefore, it is customary to make the following simplifying assumptions about the mould design and the conditions during pouring:

156

Investment Casting

1. that the runners are full; 2. that the runners are reasonably long; 3. that the metal is discharging into an empty cavity. These assumptions make it possible to treat flow as steady state, but because of the assumptions the calculations are only approximate. An investment casting mould is sketched in Fig. 5a, with its parts labelled. The gating system must attempt to satisfy six requirements. It should: 1. allow the metal to fill the mould quickly, smoothly and with a min imum of turbulence; 2. establish thermal gradients in the mould, which promote casting soundness; 3. remove slag and dross; 4. avoid reoxidation of the metal as the mould is filled; 5. be easy to attach and remove; 6. not distort the casting during solidification. To analyse fluid flow it is necessary to remember two principles: • energy is always conserved; • material is always conserved. For a simple pouring system consisting only of a pour cup, a downsprue, a single horizontal runner, and a casting cavity (as shown in

Fig 5

(a) sketch of an investment casting mould showing its parts.

Gating and Feeding Investment Castings

157

Fig 5 (b) simplified gating system. Fig. 5b), the first principle can be illustrated by balancing the energy at selected points in the runner. The energy terms which are important are: 1. the potential energy term wh. This is the weight of the metal w multiplied by its height above a reference plane h. For simplicity, the reference plane is usually taken to be the plane of the lowest runner, or the lowest point in the casting. 2. The kinetic energy term wV^/lg, where V is the velocity of the metal and g the acceleration due to gravity. 3. The pressure energy term, wP/d or wPv, where P is the pressure exerted by the metal, d the density of the metal and v its specific volume (1/d). 4. The frictional energy, wLF where I F is the sum of the loss coefficients in the system. The liquid metal will lose energy just by friction against the walls of the runner, as well as when it turns the corner into the runner from the sprue, and into the casting from the gate. Because of energy conservation, the energy in the stream at any point is equal to the energy at any other point in the system. Another way to state this is that the sum of all of the energy terms at any point in the system is constant: wh + wPv

+ w V ^ /l g +

w IF = K

where K' is a constant for a given gating design. When the equation is divided by w, Bernoulli's Theorem is obtained: h + Pv + W / 2 g

+ ZF

= K

where K is also a constant. This equation is used extensively to estimate the velocity of metal in a gating system after it has been poured down a sprue of known height. However, in order to make the necessary calculations it is necessary to know that the volumetric rate of flow Q is also the same everywhere in

158

Investment Casting

3 =^*1

Or VS

Fig 6 The law of continuity states that the volume of flow, Q, will be the same everywhere in the pipe. As the area of the pipe is less at plane 2 than at plane 1, the velocity will be greater at plane 2 than at plane 2. the system. The volumetric rate of flow is the velocity of the metal at a given point in a gating system multiplied by the cross sectional area of the channel (runner, sprue, gate, etc.), as shown in Fig. 6: Qi = Q2 = Q3 = ^ 1^ 1 = ^2^2 =

=•••

This is known as the equation of continuity. The term ZF in Bernoulli's Theorem can be expanded to ZF = Z (F, V f / l g + f (L/D) Vav2/2 g } Strictly speaking, the frictional losses are determined for each constriction, change in direction or other discontinuous feature in the gating system, by multiplying each F; by the actual velocity V) of the metal at that point and adding all these terms together. However, for simplicity, it is often more convenient in making gating calculations to assume that an average velocity Vav (usually that which would result if all the potential energy of the metal as it entered the pour cup were completely converted to kinetic energy at the bottom of the sprue) is used for each of the V, terms. In that case it is necessary merely to sum all the F, terms, and multiply them by Vav. The second term in the brackets is used to find the energy loss from flow down the runner, where f is the wall friction factor, L the length of the runner, and D its diameter, if it is round in cross section. Values of F, are given in Table l . 1 The greater the value for F ,, the less efficient is that component of the gating system, and the worse the control

Gating and Feeding Investment Castings

159

Table 11. Values of loss and friction coefficients Gating feature

Loss coefficient Sharp

Streamlined

Sprue entry from pour cup

0.75

0.2

Junction of sprue and runner Right angle bend in runner Square cross section Round cross section

2.0

1.0

2.0 1.5

1.5 1.0

Junction of gate to runner at a right angle Runner choke at base of sprue (where choke is 1/3 of runner) Wall friction losses Round runner Square runner

L = runner length D = runner diameter or side)

4.0 - 6.0 13.0

0.02 0.06

L/D UD

(where and

of the flow will be. Sharp corners in the system are generally bad because they slow down flow and cause stream separation and turbulence. Round runners are more efficient than square or rectangular runners and also hold their heat better. One example of the application of the above equations is their use in designing a sprue. Figure 7 shows the relationships in a free-falling stream of steel. From the values which are given, and Bernoulli's Theorem, the velocity of the stream at any point in the downsprue can be calculated,1 as the figure illustrates. From this, and the law of continuity, the cross sectional area of the stream at any point is obtained. Because the metal stream is accelerating, its cross sectional area is decreasing, regard less of the diameter of the sprue at that point. To confirm that this is the case for all fluids, it is only necessary to turn on a water tap and note the shape of the stream which falls vertically from it. If an energy balance is prepared at each of the points in the sprue, it is found that the pressure at point 3 is less than the pressure at points 1 or 2. This means that air may be aspirated, or drawn into the sprue through the sprue wall, if the mould is porous (as investment casting moulds are). The air can then react with the metal to form oxide inclusions. For this reason, downsprues should be tapered with the small end down so that they conform to the shape which the metal will naturally take. When they are thus tapered, the pressure is balanced, and no air will be aspirated into the stream. If the sprue is not tapered and the combined cross-sectional areas of the runners which leave the sprue well are such that the metal can flow out of the sprue faster than it can be poured into it, the sprue will not fill until

160

Investment Casting v =>/2gh

2

V., * 55.6 in/sec A1» 3.14 in2

V2 » 68.1 in/sec 2.57 in2

V3 = 96.3 in/sec A^= 1.8 in2

Fig 7 Because the liquid in a sprue accelerates owing to gravity, the law of continuity requires the stream diameter to decrease as the liquidfalls. Theflow rate at any given point is 174.6 cu.in./sec.'1 the casting is filled. This means that it will always contain air as well as metal, and this air will react with and oxidize the metal, forming inclusions. To be certain that the sprue fills, the gating system is choked at the bottom of the sprue, or in the runner just beyond its junction with the sprue, by locally reducing the sprue or runner cross-section. This min imum cross-section controls the flow rate of the metal in the runner. What happens at a change in the cross-section of the runner can be illustrated by seeing the effect of an abrupt change, as shown in Fig. 8 (the flow lines of the liquid have also been drawn in). It will be noted that there is little fluid movement in the corners of the large section — the fluid forms eddies and recirculates instead of moving down the runner. In the reduced cross-section, the acceleration and momentum of the stream as it goes through the smaller cross-section actually constrict the flow more than the cross section constriction. Again, this will be an area of low pressure in the runner, and air may be aspirated to react with the metal. Figure 9 summarises the flow patterns at corners in runners, and a pictorial representation of some of the information in Table 1 is given in Fig. 10.

Gating and Feeding Investment Castings

161

Fig 8 Development of low-pressure area downstream from a sharp decrease in the crosssection of a runner. The momentum of the liquid causes the stream diameter to decrease more than the runner diameter, and Bernoulli's theorem shows that the pressure will drop there. The discussion above gives the fundamentals for calculating the dimen sions of a gating system. However, these will only be approximations and it will still be necessary to apply common sense; points to remember are: 1. If the sprue or runner is not full, air will be aspirated and oxidise the metal in the downsprue. 2. Sharp bends and abrupt changes in direction or cross-section size in the runner will cause turbulence and air aspiration, and will risk the formation of inclusions, as well as slowing the velocity of the metal in the runner, risking misrunning. 3. The sprue should be tapered — the metal will behave as if it is tapered whether it is or not and air will be aspirated. On the second point, turbulence is mentioned as something to be avoided. It is important to understand that the flow of fluids can be divided into two regimes, laminar and turbulent. In laminar flow, all the molecules of the fluid flow in a straight line. At the face of the channel wall, where the fluid is in contact with the wall, the fluid molecules are actually not moving at all. However as the distance from the wall in creases, the velocity of the fluid increases, as shown in Fig. 11, and it reaches a maximum in the middle of the stream. Turbulent flow exists when the fluid, instead of flowing in a series of straight parallel lines, breaks up and forms eddies and swirls as it goes downstream. There are two types of turbulent flow, one with a stable

162

Fig 9

Investment Casting

(c)

Flow at corners in runner systems. Note that abrupt changes in direction cause aspiration of air, while a streamlined runner (b) prevents it.1

boundary layer, the other with an unstable boundary layer; these are shown schematically in Fig. 12. With a stable boundary layer, the velocity of the metal at the channel wall is still zero, although within the channel the fluid is moving in a turbulent manner. However, at higher velocities the turbulence in the stream breaks up the boundary layer and sweeps it into the flowing stream. In the case of molten metal in investment moulds, the fluid in contact with the wall of the mould is often oxidized by reacting with the mould material, or as a result of the diffusion of air through the porous mould. When this oxidised metal in the boundary layer is swept into the runner, it can come to rest in the casting, forming slag or dross inclusions. Whether flow will be laminar or turbulent can be calculated by determining the Reynolds number NRe of the fluid. The Reynolds number is a dimensionless number, which means that it is valid regardless of the size of the runner:

Gating and Feeding Investment Castings

163

fX ^ i

Fig 10 Loss coefficients for varions gating system features.

Fig 11 Liquid motion in runner during laminar flow. Note that liquid in contact with the mould wall is stationary. ^Re = 4 R H V

In this equation, V is the velocity of the metal in the runner, R H the hydraulic radius of the runner, and \ i / d l the kinematic viscosity of the

164 Investment Casting I

Boundary Layer

(a)

(b)

Fig 12

Liquid motion in runner during turbulent flow (a) stable boundary layer (Reynolds number below 20,000) (b) no stable boundary layer (Reynolds number above

20, 000).

liquid. R h is equal to the ratio of the cross sectional area of the channel to its perimeter (for a circular channel, RH = 4). Table 22 gives values of kinematic viscosities for various liquid metals. It will be noted that that the kinematic viscosity for water is similar to that of many molten metals. This is why water flow models, using transparent plastic moulds, are used to study metal flow in tundishes and runner systems.

Table 22. Values of kinematic viscosities for molten metals Liquid

Liquid density lb/in3

Viscosity lb/in. sec

Kinematic viscosity in2/sec

0.0361 0.0866 0.0578 0.288 0.220 0.254

0.000056 0.000173 0.000073 0.000179 0.000353 0.000353

0.0016 0.0020 0.0013 0.0006 0.0016 0.0014

Water Aluminium Magnesium Copper Cast iron Steel

1lb/in3 = 2.768 x 104 kg/m3 11b/in. sec = 1.786 x 10 Pas 1 in2/sec = 6.452 x 1 m2/s

Gating and Feeding Investment Castings

165

It has been found experimentally that flow is laminar when the Reynolds number is below 2000. A stable boundary layer will exist when the Reynolds number is below 20,000. Above 20,000 the severe turbulence causes the boundary layer to become unstable and break up. This means that the gating designs which produce the cleanest castings will be those where the Reynolds number is kept below 20,000. Com pletely laminar flow, where the Reynolds number is below 2000, would be better, but the slow pour speeds required for this would cause the metal to freeze off prematurely in the runner. Since dx and p, are properties of the alloy, they cannot be changed. But if V or RH is reduced, then the Reynolds number is reduced. For instance, a rectangular channel having the same cross section as a circular channel will have a lower RH. In theory, if the type of flow, laminar or turbulent, could be controlled precisely at each point in the gating system, not only could re-oxidation of the metal be avoided, but inclusions which were present in the metal from the melting operation could be separated by flotation. Inclusions can be separated by floating or settling under the influence of gravity because they have different densities from that of the molten metal, as shown in Table 3.3 In real gating systems, however, velocities are too high to allow most inclusions to float out. This can be shown by reference to Stokes' Law, which gives the velocity with which an inclusion can float (or settle): V = 2 g i * ( d , - d p) 9p

Table 33. Densities of oxide phases Alloy family/oxide phase Aluminium a i 2o 3 3AI20 3.2Si02 Magnesium MgO Copper CuO ZnO SnO BeO Iron/steel Cast iron Low carbon steel FeO Fe20 3 Fe30 4 Fe2S i04 MnO Cr20 3 S i02

Density, g rrr3 2.41 3.96 3.15 1.57 3.58 8.00 6.00 5.61 6.45 3.01 6.97 7.81 5.70 5.24 5.18 4.34 5.45 5.21 2.65

166

Investment Casting

Here dp is the inclusion density, r its radius, and |x the viscosity of the liquid metal. This shows that separation will be more efficient as the density difference between the alloy and the inclusion increases, and particularly as the size of the inclusion increases. Stokes' Law is, strictly, only valid for Reynolds numbers around 1 or 2 — virtually stagnant liquid — but the general principle is the same at higher values. How important is this size effect? Calculations made for an iron silicate inclusion in liquid steel poured from a height of 200 mm show that if it were 1 mm in diameter it would float out in a runner 128 mm long. However, if it were 100 microns in diameter, it would require a runner nearly 13 metres long. On the other hand, if the diameter of the inclusion were 1 cm, it would float out in only 1 mm.4 The conclusion is obvious: large inclusions can be separated more easily than small ones, and because there are few large inclusions in carefully melted metal it is clear that gating systems will remove few inclusions. The ideal gating system would have a region where small inclusions could come together and grow or 'agglomerate', followed by a section where stable boundary layer or laminar flow would occur, so that the inclusions could float out. In very turbulent flow with an unstable boundary layer, the existing oxide particles, as well as any oxide which forms at the interface between the metal and the mould, are violently mixed into the stream, where they do, in fact, come into contact with each other and agglomerate. In turbulent flow with a stable boundary layer, the oxides formed at the mould/metal interface remain there. Existing oxides from the melting operation undergo turbulent mixing, but this is not as vigorous as in very turbulent flow. Inclusions will float, but at the rate indicated in the example given above, which is far too slow for most of them to be removed in practical gating systems. Despite this, if the mould refractory is one to which oxide phases adhere, such as silica, some inclusions will stick to the runner walls, and will not reach the casting cavity. In designing a gating system for maximum cleanliness, however, some guidelines can be given (see Fig. 5a). Conditions for very turbulent flow are most often found at the base of the downsprue, where the metal has maximum velocity and is forced to change direction. In this area there will be good mixing of the metal and any inclusions, giving the inclusions the best opportunity to coalesce and agglomerate. The cross runners should provide a flow regime of turbulent flow with a stable boundary layer (NRe < 20,000). Then, as the metal enters the up-runners, it should go through a region of laminar flow for a final cleaning. This laminar flow region can be established by placing filters at the base of the up-runners, as shown in Fig. 13. Filters work by forcing the metal to undergo laminar flow. It should be noted that the openings in filters are bigger than most of the inclusions which are removed, so it is clear that

Gating and Feeding Investment Castings

Fig 13

167

Recommended filter placement in investment casting.

the primary action of the filter is not simply a mechanical straining out of inclusions. Straining is accomplished partly by slowing down the flow rate (there is a substantial head loss as the metal goes through the filter), but the most important effect is that the hydraulic radius becomes very small, because all the cells walls increase the effective wetted perimeter. Because the filters are made of materials to which the inclusions adhere, they are able to trap them effectively and keep them out of the casting cavity. Filters should not be placed in the pour cup in castings poured in air, because the metal dribbles through them and forms droplets, which rain down and react with the oxygen in the air in the sprue to form inclusions. In vacuum, of course, there is too little oxygen in the mould to cause an inclusion problem during pouring, although inclusions may be present in the original charge, or as a result of reactions with the crucible, or from a poorly made mould. However, filters in the pour cup often slow the pouring velocity, and can cause misrunning. In vacuum casting, gating systems are designed ignoring the possibility of oxidation during pouring, and the gating system is designed to fill the casting as quickly as possible to avoid misrunning. Pour cups may be larger than in air melting, to catch the metal which is poured very rapidly from the ladle.

168

Investment Casting

Experiments have shown that there is an optimum pour rate for each cluster design. If the mould is poured too slowly, misrunning results; if poured too quickly, reoxidation and inclusions will form and cause scrap. Moulds should be adequately vented, so that when metal enters the mould, the air that it displaces can escape. This is necessary because investment moulds are not permeable enough to allow the air that fills them to escape through the walls as fast as the metal enters, so that vents are often necessary. Investment casters are often faced with the problem of filling thin sections. The ability to do this increases as the pouring temperature of the metal is increased, and as the preheat temperature of the mould increases — because it takes longer for the molten metal to freeze (see below) in a hot mould than in a cold one. The ability to fill thin sections is also a property of the alloy composition, and is called the 'fluidity'. In general, as the difference between the temperatures at which melting starts and ends increases, the fluidity of the alloy decreases. Selection of an alloy is rarely left to the foundryman, but when it is, he should take into account the fluidity of the alloy. Fluidity can also be decreased by the formation of an oxide skin on the alloy, and as the cleanliness of the melt decreases as a result of reactions with the mould material or atmosphere. In designing a gating system, a primary objective should be to make sure that all cavities in all moulds fill and solidify identically. Only in this way can sufficient control be exercised over the process to assure reproducibility. One method which can be used to determine whether this condition is being achieved on a particular cluster is to record the position on the cluster of all scrap and reworked castings. If scrap is occurring at a higher than average rate at one or more positions on the cluster, the gating system should be revised. Gating systems should be symmetrical about the downsprue, with castings placed at the same distance from it. In multilayer clusters, the castings on the top and bottom layers solidify faster than those between, which are insulated to some extent by the top and bottom layers. In investment casting in air an unpressurised gating system is preferred ('unpressurised' means that the ratio of the cross-sectional area of the sprue, to that of the runners, to that of the gates increases, e.g., 1:2:2 or 1:1.25:1.25). This keeps the metal from squirting from the sprue into the runners, and from the gates into the casting cavity. If a pressurised system is used, the metal will break up into tiny droplets, which will react with the air in the casting cavity and form inclusions. Changes in metal direction should be kept as gentle as possible and care should be taken that all runners and gates have generous radii. The use of simple conical pour cups is recommended and the pour cup should be kept full. The system should be choked at the bottom of the sprue rather than at the filters, if they are used.

Gating and Feeding Investment Castings

169

HEAT FLOW AND FEEDER DESIGN Most small investment castings are made without feeders, because they are small enough and their geometry, with high surface area to volume ratios, is such that they solidify quickly and can be fed from the runner and gates. Feeders, which are expensive to attach and remove and which lower the pouring yield, are frequently unnecessary. However, a casting configuration or alloy may be encountered which does require a feeder to prevent shrinkage. In designing a gating and feeding system for a casting, the object is to control solidification so that it begins in one section of the casting, and moves progressively toward the gate or feeder, so that any shrinkage occurs in the gate or the feeder, and the casting itself is sound. This can be done by controlling the heat flow as the casting solidifies. Heat is transferred by the three mechanisms of conduction, radiation and convection. As in fluid flow, the rigorous analysis of heat transfer is difficult — because the mathematics are complex, the problem is threedimensional, and non -steady state conditions exist. Therefore, unless computer programs which solve these complex equations numerically are used (and there are a number of commercial programs available) many simplifications are made to ease the design calculations for gating and feeding systems. The most important heat transfer mechanism in metal casting is by conduction, which occurs when a hot object is placed in contact with a cold object; the hot object loses heat to the cold one, which heats up until there is no temperature difference between the two. If the hot and cold objects are maintained at their original temperatures, a steady-state thermal gradient is established between them. This is illustrated in Fig. 14. The block is placed against a hot surface. Heat flow in the block, q, is dependent on the temperature difference between the two sides of the block (T - T0), and is inversely proportional to the width of the block, x: n

x

The constant of proportionality, k, is the thermal conductivity. The higher the value of k, the more quickly heat is transferred. Metals and alloys have high values of k, while refractories and insulators have low values of k. Thermal conductivity is not constant for a material, but varies with temperature. Metals and mould materials vary in their ability to conduct heat. The more quickly a mould can transfer heat from a hot to a cold region, the faster the casting will solidify. Thermal conductivities for a number of

170

Investment Casting

(a)

(b)

(c)

Fig 14 Temperature distribution in a block placed against a hot surface at left (a) just after contact (b) later (c) at steady state.

Table 4s. Thermal properties of some moulds and metals and alloys Density gm/cc

Material

Thermal conductivity*

Specific heat cal/gm °C

Thermal diffusivity cm2/sec

Temperature °C

Quartz

2.6

0.0009 0.0013

0.269 0.281

0.0013 0.0018

400 1000

Olivine sand

1.8

0.0018 0.0018

0.238 0.281

0.0042 0.0035

400 1000

Zircon sand

2.78

0.0020 0.0023

0.161 0.192

0.0045 0.0043

400 1000

Chromite sand

2.75

0.0017 0.0021

0.190 0.224

0.0033 0.0034

400 1000

0.23 C steel

7.86

0.1018 0.0681 0.0710

0.142 0.239 0.158

400 800 1200

1.0 Cr, 0.3 C steel

7.84

0.0920 0.0619 0.0719

0.142 0.206 0.146

400 800 1200

18Cr8Ni stainless steel

8.00

0.0497 0.0629 0.0762

0.136 0.154 0.159

400 800 1200

Aluminium

2.70

0.566 0.526 0.222

0.239 0.271 0.260

227 527 727t

AI-4.5CU

2.80

0.450

0.232

250

60Cu-40Ni

8.90

0.218

0.119

* The units are cal/cm sec °C. t The given values are for liquid aluminium. 1 gm/cc = 103 kg/m3 1 cal/cm.sec °C 4.187 x 102 W/mK 1 cal/gm °C 4.187 x 103 J/kgK 1 cm2/sec = 10“4 m2/s =

=

722

Gating and Feeding Investment Castings

171

metals and mould materials are given in Table 4.5 Also given are values of the specific heat, which is the amount of energy it takes to raise the temperature of the material by one degree, and the thermal diffusivity, which is the thermal conductivity divided by the product of the density and the specific heat. The thermal diffusivity is a measure of how rapidly heat is absorbed by a mould; in other words, it indicates the ability of the mould to extract heat from the casting. When hot liquid metal is poured into a cold mould, the mould is immediately heated by the metal, which is itself cooled by the mould. If the metal temperature is not high enough, the liquid may cool so much that it freezes in the runner or before the casting is completely filled, causing misrunning. The metal must be sufficiently heated (given the proper amount of superheat above the melting temperature) to stay liquid long enough to fill the mould. It is not always possible to superheat the liquid metal sufficiently to fill the mould; this may occur in a steel casting, where the refractories may not be able to withstand the increased melting temperature. One method used to keep the metal from freezing prematurely is to preheat the mould to a high temperature. The metal does not then cool as fast, and can run further, and into thinner sections, before freezing. If the temperatures of the metal and the mould are measured locally after the metal has been poured, it is found that the temperature profile looks like that shown in Fig. 15. The metal temperature at the mould interface is not equal to the mould temperature. This is because the mould and metal are not in intimate contact after solidification starts. The metal shrinks away from the mould wall, and the interface now resists heat transfer. Interface resistance will be less on the bottom of the casting because gravity will keep the bottom surface in contact with the mould. Radiation heat transfer is proportional to the fourth power of the abso lute temperature. The hotter a casting the greater is its rate of heat transfer by radiation, and as it gets still hotter it transfers heat even more efficiently. The equation which describes radiation heat transfer between two bodies is q = a e , (T,4 - T24) where a is a constant and e1 the emissivity of the hotter body. Calculating the heat exchange is complicated by the fact that the two bodies may be at angles to each other, so that not all of the radiation from the hotter body falls directly on the cooler body. In this case, the equation must be corrected by calculating the view factor between the bodies. Radiation is especially important in the ferrous metals. Because radiation heat transfer is so efficient at high temperatures, the geometry of the

172

Investment Casting MOULD

CASTING

Fig 15 Temperature profile in casting and mould during solidification. Temperature is nearly constant in the liquid, falls off in the mushy zone, and falls rapidly in the solid. But, because the solid shrinks away from the mould, the mould and casting are not at the same temperature at the interface. casting can influence its soldification in ways not expected with simple conduction. For instance, if a cluster is designed so that there are parts surrounded by other parts, the surrounding parts will be 'shadowed' and will cool substantially more slowly than parts which are on the outside of the cluster, which can radiate heat to the surroundings. Similarly, on large castings, some parts of the casting may be shadowed by others, thus cooling much more slowly than would otherwise be expected. Generally, radiation heat transfer is not specifically considered in designing gating systems for investment castings, except in the case of directionally solidified and monocrystalline superalloy castings, where the control of radiation heat transfer is crucial. Convection heat transfer, which occurs in liquids and gases, can be observed in induction furnaces where the mixing is driven by the electromagnetic field, and is called 'forced convection', or by observing the top of a large ladle of steel where a flow pattern is clearly visible; this is

Gating and Feeding Investment Castings

173

caused by metal which is cooling on the top and sides of the melt sinking (as the metal cools, it becomes heavier) and being replaced by hotter, lighter metal. This is called 'natural convection'. Convection heat transfer is not normally considered in investment castings, as they solidify too fast for convective currents to be established. The fact that liquid metal loses heat to the gating system during pouring means that the liquid metal in the mould just after pouring is not at a uniform temperature. The metal that has travelled furthest from the pouring cup has been in contact with the cold mould longest, and has lost the most heat, while the metal that is closest to the pouring cup is hottest, having lost the least heat. The heat lost by the first metal poured has also heated the gating system. Because most metal has flowed through the part of the casting nearest the gate, this region of both the mould and the casting is hottest. Solidification will start with the coldest metal, and will move progressively towards the hottest. Placing gates at the heaviest sections of the casting, so that they freeze last, will encourage this process. Most investment castings are gated without feeders to take advantage of this effect. However, it is occasionally necessary to feed a section of a casting which cannot be reached from a gate, or to provide feed metal at the gate itself. For this purpose a feeder or riser is employed; an example is shown in Fig. 16. Feeders may be attached to the top or the side of a casting. Feeders are reservoirs of molten metal, which are designed to freeze after the section they feed, so that molten metal can continue to flow into the section until its solidification is complete. To design a feeder which freezes after the casting, it is necessary to note that large, heavy chunky sections stay molten longer than thin sections. The greater the surface area for a given volume of casting, the greater the

Fig 16

Feeders attached to a casting.

174

Investment Casting

amount of material in the casting section in contact with the mould, and the more rapidly heat will be able to flow to the mould. Conversely, the more compact the casting the more concentrated is the heat in a limited volume, and the harder it is for the heat to flow out. Chvorinov observed this and, following careful experimentation with a variety of shapes, devised Chvorinov's rule,6 which states that the freezing time of any section of a casting is proportional to the square of the volume divided by the surface area, or tf = C{ V/ SA¥ This shows that as long as the volume to surface area ratio of the feeder is greater than that of the section of the casting which it is to feed, the feeder will freeze after the casting section. Because the last metal to freeze will be in the feeder the shrinkage will be localised there and the casting section will be sound. To determine in which order different parts of a casting will freeze, it is necessary only to compare their volume/surface area ratios. The volume/surface area ratio is also known as the 'modulus' of the casting. A number of highly successful foundry computer programs are based on comparisons of casting moduli with those of feeders. Such programs are successful in most, but not all, casting feeding problems dealing with the elimination of large porosity (macroporosity). They are limited only if it is important to know the thermal gradient, or the cooling rate, or the rate of solid/liquid interface advance. Chvorinov's rule is important in guiding thinking about heat flow from castings. It is clear that solidification must move progressively towards either a feeder or a gate as a source of fresh metal. A heavy section of casting cannot be fed by placing a feeder or a gate on a section separated from it by a thin section, because the thin section will freeze first, choking off the flow of metal to the heavy section, resulting in shrinkage. Very thin sections which protrude from a casting (e.g. an isolated fin) will remove heat locally from that area of the casting at an accelerated rate, causing it to freeze before it might be expected to, as shown in Fig. 177 A right-angled section will freeze in such a way that the effective centreline, where metal is liquid longest, will move away from the exte rior corner (which has the greatest surface area) to the interior corner (see Fig. 18). The rules of heat flow and the requirement that castings freeze progressively from thin sections, where solidification starts, to heavy sections, which provide a reservoir of feed metal, mean that it is best to gate into heavy sections. As already mentioned, in many investment castings, this is all that is necessary to produce sound castings. When feeders are necessary, it is advisable to gate into them, to ensure that the metal in

Gating and Feeding Investment Castings

175

Fig 17 Cross-section of casting with fins during solidification. Because of the high surface-area to volume ratio of the fins they cool faster, and metal behind them freezes in the pattern shoion.7 them is hotter than that in the casting cavity. It is also advantageous to orient castings so that their heavy sections are on top of the cluster when it is poured, so that gravity can help drain these sections and encourage them to act as feeders. Feeders must then, of course, be placed to feed the heavy sections in their turn. Feeders are generally designed to be compact in order to conserve heat. They are spherical when possible, and otherwise have circular cross sections. They are more efficient when gates enter at the base than when they are placed on sections which are not gated. In ceramic shell investment casting the efficiency of feeders can be improved by wrapping an insulating blanket around them. An exothermic compound may be added to the pouring cup after pouring: this compound burns, giving off heat and keeping the pouring cup molten long after it would normally solidify. Chvorinov's rule is most helpful in avoiding gross shrinkage porosity. However, as mentioned at the beginning of this chapter, many investment casting alloys freeze in a 'pasty' manner and form dispersed microshrinkage. Chvorinov's rule is of little help in this case. The minimisation or removal of dispersed microshrinkage requires other techniques. The first is to de-gas heats thoroughly. When an alloy solidifies, gas dissolved in the liquid cannot remain dissolved in the solid, and must come out of solution, forming gas bubbles. These bubbles, or pores, are found at the location of the last liquid to solidify, which will be in the liquid channels present between the dendrites; unless the pores are completely spherical, they will resemble shrinkage cavities. Melts should be completely degassed before pouring.

176

Investment Casting

Fig 18

Comparison of solidification patterns in (a) flat plate and (b) around a corner of the same thickness. At the corner the outside cools faster than the inside, so that the liquid moves towards the inside.

Dispersed microshrinkage can also be minimised by establishing high thermal gradients in the casting (see Fig. 14). These reduce porosity by causing the local area of the casting to solidify so quickly that the lengths of the channels between the dendrites are substantially reduced, and no channels of feed metal are blocked. The steeper the thermal gradient the sounder is the casting. The control of thermal gradients is crucial to designing good gating and feeding systems. It is therefore essential that this should be thoroughly understood. Figure 15 shows a temperature profile through a solidifying casting and an adjacent mould. The thermal gradient is the slope of the temperature profile at any place in the casting, and varies from the liquid (where it is very flat) through the mushy zone (where it increases) to the

Gating and Feeding Investment Castings

177

solid next to the mould (where it is steepest). The gradient just above the temperature where solidification ends is the most important in determin ing the casting properties which are of interest. Fortunately, this is similar to the gradient throughout the solidification zone. High thermal gradients are beneficial to casting quality, as long as they move through the casting to gates or feeders. They decrease porosity and refine the dendrite arm spacing. Consequently, a number of techniques have been developed to increase thermal gradients. Thermal gradients can be increased by raising the maximum temperature (e.g. by increasing the pouring temperature), decreasing the lower temperature (e.g. by lowering the mould preheat temperature), or by increasing the rate at which heat is withdrawn from the mould (e.g. by increasing the thermal conductivity and thermal diffusivity of the mould material). One way to achieve the latter effect is to use chills. These are pieces of metal, such as copper or aluminium, which have high thermal conductivity and diffusivity, and can remove heat from the area of the casting with which they are in contact at a higher rate than does the mould. Chills are placed in the mould so that they contact that part of the casting surface where solidification is intended to begin. It is, however, difficult to use local chills in investment casting unless solid moulds are used, because the chills must be supported by the mould material. Another way that thermal gradients can be established within a ceramic shell mould is by surrounding the part of the mould which is to solidify first with a backing material which has high thermal conductivity. Such material might be steel or copper shot. The rest of the mould can be wrapped with an insulating blanket of refractory fibre. Solidification will then start in that part of the mould which cools the fastest, (the part surrounded by shot) and move towards the part which cools slowest (the part wrapped in insulation). As metal solidifies it gives off its latent heat of fusion, which represents the energy liberated as the atoms assume the lower energy solid state. The latent heat varies from one metal to another. The rate of solidification slows as the latent heat is given off. This can make the establishment of thermal gradients difficult, especially in metals and alloys which have very high latent heats, such as aluminium. The thermal conductivity of the mould can also be manipulated by controlling its porosity and density. As the mould becomes more porous, its thermal conductivity drops. This implies that control of the mould quality and reproducibility is necessary to provide consistent heat transfer conditions for casting solidification. Inclusions also play a role in forming pores. It has been observed that as the cleanliness of the metal improves, porosity decreases. It is difficult for a pore to nucleate because energy must be provided to form the pore

178

Investment Casting

surface. The amount of energy needed for this is substantially reduced when the non -metallic surface of an inclusion is present. This is why pores are frequently found associated with inclusions, and is another reason for taking particular care to use melting and pouring techniques which give clean metal.

COMPUTER MODELS FOR DESIGNING GATING AND FEEDING SYSTEMS Traditionally the mathematics involved in solving the heat and fluid flow equations required for designing gating systems have been so complex that designers have relied almost exclusively on experience and rules of thumb. The only method of checking the design was to pour the casting and inspect it. Today, however, it is possible to check the gating design by using any one of a number of commercially available computer based models. These range in complexity and power from modulus based (Chvorinov rule) models to finite difference and finite element method programs. The advantages of the modulus based models are that they are reason ably simple to program, can be run on personal computers, and give results fairly quickly. They usually ignore fluid flow and the fact that liquid metal is not at a uniform temperature in the mould when solidification starts. They are generally more useful for sand foundries than for investment foundries, as the thin sections typical of investment castings usually freeze off so quickly that unless there is an accurate way of determining the initial temperature of the metal in the mould, significant errors can arise. These programs have had little success in predicting the presence of dispersed microshrinkage. More complex models are based on finite element or finite difference numerical methods for solving the differential equations which describe heat transfer. These methods rely on repetitive recalculations of values in a series of very small regions of the casting, a tedious job for a person but ideally suited for a computer. A 'mesh' must first be constructed to describe the positions and shapes of these small regions in a casting, as shown in Fig. 19.8 This requires considerable skill, and at present is time consuming and difficult. Automatic mesh generators are only now being developed in order to speed up and simplify this step. The programs run on work stations, and may take as long as one week to mesh, and then a day to run. However, they include the effects of fluid flow and heat flow, including radiation heat flow (another very tedious calculation) and today are capable of predicting not only areas of the casting where shrinkage may be expected, but also areas where mechan ical properties may be inadequate. Some programs are able to predict

Gating and Feeding Investment Castings

179

Fig 19 Three-dimensional computer mesh of a casting, for numerical solution of heat flow and solidification.8 grain size and eutectic shape, including the effects of grain refinement and melt modification treatments. Efforts are under way in a number of countries to use sophisticated systems and artificial intelligence to produce initial gating designs for analysis by solidification software programs. When these programs are in use, it will be possible to design foundry gating designs automatically, and engineers will be freed for more challenging tasks. GENERAL RULES FOR GATING AND FEEDING Investment casting is a precision manufacturing process. It is very im portant that the highest level of control be exercised over the process at all times. In devising a method for a casting, a gating and feeding system has to be designed, and the same care must be applied to this design problem as to the design of any other engineered product or process. The final design must give the foundryman the maximum amount of control over pouring and solidification, without incurring excessive cost. A general sequence for designing a gating system may be outlined as follows: 1. Carefully inspect the part drawing to be sure that all dimensions and tolerances are clear and fully understood. Be sure that all quality inspection criteria are clear. Note the location of datum surfaces and locating points for subsequent machining. Do not attempt to method a casting until it is absolutely certain what the acceptance

180

Investment Casting

requirements are. If the drawings or specifications are not clear, consult the customer for clarification. 2. Verify that the shape of the part is clear. This may be obvious from the part drawing; it may be necessary, however, to defer the actual design of the gating system until a wax pattern has been injected and can be inspected. Today this practice is giving way to constructing a solid model of the part on a computer, and using this to visualize where gates are to be placed. 3. If the casting is complex, mentally divide it into smaller segments. Note the arrangement of heavy sections and thin sections. Gates will be generally placed on heavy sections, so will feeders, if used. Solid ification will naturally begin in thin sections, and move to the heavy sections. 4. Visualize the natural flow path of metal into the casting cavity from the gates. Fluids naturally flow down, not up, and flow more easily in thick sections than in thin. Using the equations given earlier, calculate the dimensions of the gating system. A number of general rules for gating system design are given below. These rules apply regardless of the metal being poured, or the shape of the casting. They deal, for the most part, with common sense considerations which must be observed in designing an economical gating and feeding system. 1. Small- or medium -sized castings which are to be cast on trees or in clusters should be arranged so that all castings solidify identically. For multi-piece clusters, symmetrical placement of castings is recom mended. Circular symmetry, as shown in Fig. 20, is preferable to box-type designs in which the castings on the end experience different thermal and flow conditions from those in the middle of the box. Pouring parameters which give good corner castings may also cause scrap in interior castings. In a circular cluster the flow and thermal conditions are identical everywhere. 2. Gates and feeders should be positioned so that they can be attached and removed easily. Look for flat surfaces which facilitate gate removal, and consider the gate removal method to be used; if a cut-off wheel is used it may cut into another casting if the cluster is not properly designed. Care is needed not to locate gates on datum surfaces or locating points, because it is impossible to grind gate stubs off the casting accurately enough to re -establish these features. 3. Drain tubes for wax should be added to positions in the cluster where wax will not naturally drain. This prevents mould damage during de-waxing, and the tubes also serve as vents during pouring.

Gating and Feeding Investment Castings

(

l

Pour Cup

181

J \

(b) Fig 20 Circular symmetry of castings around downsprue (a) is recommended over boxtype designs (b) so that all castings solidify identically. 4. Cluster design should take into consideration the rigidity of the cluster and its support during dipping and storage. If the cluster is not properly supported, the patterns can warp under the weight of the dipcoats, and yield castings which do not meet dimensional specifications. 5. Consider how the gating system will affect dipping and drying. Gates which interfere with airflow over the cluster can cause mould defects. Blind holes should be oriented downward so that they drain easily during dipping.

182

Investment Casting

6. Orientate parts and arrange gating to minimize distortion and warp ing on solidification. Note that the runners may freeze after casting, and can distort the cluster and castings as they shrink. 7. Do not place gates where the metal stream will impinge on thin or fragile cores at right angles. The force of the metal stream can fracture the core. It is better to gate so that the stream is parallel to the core. 8. Use conical pouring cups, because other shapes, such as half spheres or rectangular boxes, cause splashing. Pouring cups, downsprues, and runners should be chosen from standard shapes and sizes to minimize tool and component inventory.

CONCLUSION The design of gating and feeding systems for investment castings remains a matter of judgment, tempered by the application of physical laws governing solidification, fluid flow and heat flow. These laws are complex and are best analysed by using computer modelling programs. However, there is a body of rules, which, if followed, can yield successful gating systems. These rules are not arbitrary, but are based on sound physical principles that apply to all metals cast.

REFERENCES 1. N. Wukovich and G. Metevelis: American Foundrymen's Society Transactions, 97, 285-302,1989. 2. R.W. Ruddle: The Running and Gating of Sand Castings, 20, Institute of Metals, London. 1956. 3. D.L. Twarog: Gating and Feeding of Investment Castings, 7-4. American Foundrymen's Society, Des Plaines, IL. 4. N.H. El-Kaddah and T.S. Piwonka: American Foundrymen's Society Transactions, 98, 295-300, 1990. 5. R.D. Pehlke, A. Jeyarajan and H. Wada: Summary of Thermal Properties for Casting Alloys and Mold Materials, National Science Foundation Report NSF/MEA 82028, Washington, DC, Dec. 1980. 6. N. Chvorinov: in Proc. 30th International Foundry Congress, 357-382, Prague, 1963. 7. M.H. Kim and J.T. Berry: American Foundrymen's Society Transactions, 97, 329 334,1989. 8. Georgia Institute of Technology: A Computer-Aided Design System for Castings Progress Report No. 1, National Science Foundation (Grant DAR 78-24301), Washington, DC, 1980.

7

Finishing Investment Castings H.T. BIDWELL

The finishing department is one of the most labour intensive areas of the investment casting foundry. Usually, about twenty-seven per cent of hourly employees work in this area, which includes knock-out and clean ing, casting cut-off and general grinding and finishing operations. The finishing department is essentially concerned with processing the cast moulds. A typical sequence is as follows: (a) (b) (c) (d) (e) (f)

Remove the bulk of the ceramic shell, usually mechanically. Remove the castings from the running system. Remove remaining refractory by mechanical and/or chemical means. Remove gates from the castings. Remove or blend positive metal from the casting surface. Final abrasive blast cleaning.

In some cases the work of the finishing department will include heat treatment, secondary machining operations and bubble packing the castings ready for shipment. This chapter will be mainly devoted to a review of items (a) -(e) as listed above. The precise sequence of operations may vary from plant to plant but the sequence presented here may be regarded as fairly typical. CASTING KNOCK -OUT In most instances, the bulk of the ceramic shell is removed mechanically. Vibrating hammers shake loose the ceramic, leaving only a relatively small amount adhering to the castings and runner systems. This operation was once one of the dustiest and noisiest in the investment casting plant. Over recent years soundproof, dustproof knock-out cabinets have been developed to contain the knock-out hammer and anvil. A typical knock-out cabinet is illustrated in Fig. 1. DOI: 10.1201/9781003419228-7

184

Fig 1

Investment Casting

Vibratory knock-out cabinet. (Courtesy of Investment Casting Resource Interna -

tional, U SA .)

This first post-casting operation has changed very little over the past fifty years, with the possible exception of a very sophisticated electric shock discharge system developed in Russia, evidently with mass pro duction applications in mind. This system was expensive to engineer, noisy, and costly to operate, but there was no dust. High pressure water blasting systems are now capable of completely removing the ceramic shell. This development is discussed later in the chapter. CUT -OFF OF CASTINGS In most cases castings are removed from the running system with abrasive cut-off wheels or friction saws. Aluminium castings are removed with bandsaws and sometimes with abrasive cut-off wheels. Abrasive Cut - off Machines and Wheels There are two basic types of cut-off machine: (a) Fixed wheel machines, where the work is fed to the wheel (see Fig.

2).

Finishing Investment Castings

185

Fig 2 Abrasive cut-off machine. (Courtesy of Investment Casting Resource International, USA.) (b) Chop stroke machines, where the wheel is applied to the clamped workpiece. Irrespective of the difference in method, there are a few basic rules that apply to the operation of cut-off machines. Perhaps one of the most sim ple and important relates to the horsepower rating of the equipment. For all practical purposes a cut-off machine should have a minimum of 300 watts per centimetre, or one horsepower per inch, of cut-off wheel diame ter. If the machine is under-powered all sorts of problems can occur, including stalling, wheel glazing, and overheated and burned parts. These problems can be learned the 'hard way' by trying to cut cast tool steel without enough power; the overheating will send cracks through every carbide network in the casting. Other general rules are: • Use all available power, and cut as fast as possible. • Run the wheel close to, but do not exceed, the maximum recom mended wheel speed. • Clamp or fix the workpiece as securely as possible. • Maintain the machine in optimum condition, check and re-check spindles and spindle bearings; make sure that all moving parts move easily. The cutting off process has gained considerable attention from the viewpoint of mechanisation and automation. Clearly, if production

186

Investment Casting

quantities are high enough, then it is worthwhile to address the econ omics of automation of cut-off. Completely enclosed and computer con trolled cut-off machines have been used since the early 1980s. These machines are programmed according to the casting part number, and the system is so sophisticated that cut-off wheel wear is measured auto matically after each stroke of the machine and the stroke is automatically changed to compensate for this wear. There are also a number of less sophisticated but equally useful enclosed cut-off machines currently available. Such machines are in many cases more efficient than manual systems and produce a uniform gate height, which is a major advantage when mechanising gate grinding operations. These new machines are a logical progression in improving the cost efficiency of the process. Abrasive cut-off wheels are manufactured to a number of specifications for different alloys. The options within the specifications include the size, hardness or grade of abrasive grit, structure, bond, side patterns and reinforcement. As a general rule, aluminium oxide is the chosen abrasive for use on metals. The grit is available in a range of sizes to suit the particular needs of the investment caster. The hardness or grade of the wheel can be varied. The harder the wheel the slower the wheel wear, but also the slower the rate of cut, which can produce burrs on the edge of the casting being removed. Softer wheels cut more efficiently but wear away more rapidly. The type of bond and the side pattern offer more choices. There are three main types of bond: rubber for wet cutting, resin for wet and dry cutting and shellac for top quality cutting.1 The side pattern of the wheels is important to maintain relief in the cut and to minimize wheel stall. There has been constant research on the design of cut-off wheels and their application. Wheel manufacturers have optimized the type of abrasive and the wheel structures for particular applications. Current technology has reduced abrasive costs by as much as 33%; this has been made possible by the use of better abrasives, new bond systems and improved manufacturing techniques. The specialized nature of many cut-off wheel applications makes it imperative for the investment caster to seek the advice of the wheel and machine manufacturers. The use of the wrong wheel in a given operation results in high costs and poor safety. Cut - Off Using Friction Saws Friction sawing is essentially a melting -burning operation, in which heat is generated by the saw teeth sliding over the metal surface, rather than mechanically cutting into the workpiece as with a conventional bandsaw. Friction bandsaws operate at high speeds, often well over 2500 surface metres per minute. The high speeds apply a large number of blade teeth

Finishing Investment Castings

187

to the workpiece, generating great heat at the tooth/workpiece interface. This is enough to melt the steel momentarily; the molten metal is removed by the belt and the substrate is then subjected to the fast moving teeth and the operation repeated. This is a simplified explanation and for all practical purposes the metal is continuously removed at the blade/ metal interface. The individual teeth of the blade are only in contact with the metal for a fraction of a second and are thus unaffected by the lo calised heat generated. Friction saw manufacturers give clear advice on how to operate the equipment at maximum efficiency. Band speed is critical for efficient cut-off. If the band speed is too low, or if the belt is not held at the correct tension, excessive tooth wear can occur. Incorrect placing and distancing of the band or blade guides can lead to tension problems caused by bowing of the band at higher pressures. The high speeds associated with friction cut-off machines make good maintenance of guides, bearings, supports and other components very important, while poor maintenance can lead to rapid band failure.2

CHEMICAL CLEANING METHODS Although silica refractories can be removed using hydrofluoric acid, the most common chemical cleaning methods employ caustic salts, either in the molten condition or as aqueous solutions. The removal of ceramic material by chemical means is normally carried out after the bulk of the ceramic shell has been removed by other means. It is of course possible to remove the entire shell chemically, but this would not be cost effective.

Molten Salt Baths The molten salt is sodium hydroxide, with or without buffering additives and catalysts. The salt is melted in a steel pot and the bath operates at temperatures in the range 475 -600°C. The molten salt will readily dissolve residual silica based refractories from the surface of the castings. The primary chemical reactions in removing silica shell materials are as follows: Si02 + 2NaOH -> Na2Si03 + H20 Silica + Sodium Hydroxide —> Sodium Silicate + Water 2NaOH + C 02 -> Na2C 03 + H20 Sodium Hydroxide + Carbon Dioxide -> Sodium Carbonate + Water3 The solid products of the reaction deposit in the bath as sludge and are regularly removed from the bath to ensure maximum efficiency.

188

Investment Casting

Immersion times as short as 20 minutes are generally adequate to clear the castings of any adhering siliceous refractory materials. After the salt bath treatment, the castings are subjected to thorough washing to remove excess or carried over salts, and they can then be subjected to acid neutralisation and scale removal. The layout shown in Fig. 3 has been developed by the Kolene Corporation. This system allows for almost continuous operation of the plant; it incorporates a sludge removal system and acid neutralizing and cleaning baths. It is reported that one of these units processes over 5 tonnes of ferrous castings per 8 hour operating day. During this period the bath is desludged every 4 hours and used salt is replenished at the start of each shift. These salt additions are made on the basis of one kilogramme of salt for every fifty kilogrammes of ferrous castings cleaned. The layout of a typical desludging system is shown in Fig. 4. Disposal of the sludge from salt baths has received considerable attention, with increasingly stringent environmental control requirements. The Kolene Corporation has developed a sludge processing system whereby

Fig 3

Molten salt cleaning station. (Courtesy of Kolene Corporation, USA.)

(a) Finishing Investment Castings 189

Fig 4 A typical desludging system as used in salt bath cleaning, (a) Sludge pan in place in salt bath; (b) Sludge pan raised from the salt bath; (c) Sludge pan dump mechanism activated; (d) Sludge pan returned to the salt bath.

190

Investment Casting

the quench/rinse water from the cleaning system is used to dissolve the cooled sludge products. Most of the entrained salts, alkali silicates and other soluble residues are dissolved, while the so-called inert ceramics such as alumino-silicates and zircon are not. The sludge solution is treated to adjust its pH value and any reduction/precipitation required for heavy metals removal is carried out. The treated solution, with suspended solids, is clarified through a filter press and the water is usually clean enough at this stage to be discharged into the sewer or recyled in the system. Molten salt bath cleaning is rapid and efficient in removing accessible refractory materials. The simple system can be modified to include a degree of cathodic protection to eliminate the chances of intergranular corrosion or oxidation of high temperature nickel based alloys. It should be noted, however, that excess immersion time will inevitably lead to intergranular attack. Similarly, if the operating temperature of the molten salt bath is set high in the hope of speeding up the cleaning process, rapid intergranular attack can occur in a wide range of alloys. Hot Aqueous Caustic Cleaning Baths These baths are operated with alkali concentrations, usually of potassium hydroxide, ranging from 5 - 30%. The operating temperature is about 80°C and the castings are immersed in the solution for several hours to remove residual refractory material. After cleaning, the castings are thoroughly rinsed in hot water and dried. Ceramic Core Removal Molten salt baths, or hot aqueous caustic solutions, will remove most accessible core materials provided that the core refractory is leachable. For difficult, inaccessible cores, an autoclave system of leaching can be used. High pressure and intermittent pressure autoclave systems have been successfully used to remove even the most complex cores. Low concentrations of aqueous caustic or hydrofluoric acid have been used very successfully as the leachant.

ABRASIVE BLAST CLEANING METHODS Cleaning by particulate impact is so common that there is little to add to the large amount of information available in the general literature. Blast cleaning is readily divided into two broad and distinct concepts, pressure blasting and airless blasting.

Finishing Investment Castings

191

Pressure Blasting Pressure blasting employs a carrier medium, usually air or water, to project the abrasive at high velocity on to the surface to be cleaned. Shotblasting and vapour blasting machines are familiar examples of each type. Pressure blasting machines are either direct pressure or suction units. The direct pressure system consists of a container, holding the abrasive medium, which is pressurized by compressed air. When the exit or blast nozzle of the equipment is opened, the abrasive is forced out at high velocity. Suction blasting machines operate by a simple venturi effect. The abrasive medium flows through the blasting nozzle owing to the venturi effect generated by an air jet introduced just behind the nozzle. The socalled vapour-blasting machines operate in a similar manner by introducing high pressure water just behind the nozzle. The abrasive, which is in suspension in water, is directed at high velocity on to the workpiece. Both types of machines operate at pressures often in excess of 5 bar. As a general rule the greater the air pressure the faster the cleaning cycle. The speed of cleaning and the type of finish obtained depends on the abrasive used, of which there are two main categories, metal shot and abrasive ceramic grains (grit). The action of metal shot is essentially im pact cleaning. In grit blasting, however, the cleaning action is more of a cutting and scouring operation. The shot is either steel or chilled cast iron and is available in finely graded sizes; the coarser the shot the rougher the surface finish. Ceramic grit is usually either alumina, alumina -zirconia, or silicon carbide. The main criteria defining blast cleaning operations are the type and size of the abrasive, the blasting pressure, the blasting angle and the working distance. At one time the so-called 'sand blasting' machines used sand as the abrasive, but because of the health hazard associated with silica dust, sand has been replaced by metal shot or ceramic grits which have very low free silica contents. Maintaining the operational efficiency of pressure blasting equipment is a matter of common sense. All operating systems should be checked daily for optimum efficiency. Parts subjected to wear should be replaced promptly and according to the recommendations of the manufacturer. For example, cleaning time will rapidly increase if pressure is too low, if the blast nozzle is worn or if the abrasive medium is spent. For reproducible results and optimum productivity, operators should be trained to carry out simple maintenance tasks. They must be made aware of the need to maintain an optimum blasting angle for the particular job or type of cleaning required. For example, at a blasting angle of 45° the finish obtained will be rough, whereas with a blasting angle of 90° a smooth,

192

Investment Casting

uniform finish is obtained. The distance from the blast nozzle to the workpiece should be standardized, because the velocity of the blasting particle will decrease rapidly with increasing distance.

Airless blast cleaners As the name implies, airless cleaners do not use compressed air to direct the abrasive on to the work surface. Instead, the abrasive particles, which are steel shot or ceramic grit, are thrown at the workpiece by the vanes of a centrifugal wheel rotating at high speed. Shot impellers may be used in all types of cleaning machine. Equipment varies from simple barrel tumbling cabinets to sophisticated, continuous operation, multivane cabinets. Figure 5 is a schematic of the impeller and impeller feed assembly which is the basic element of this type of equipment. The steel shot travels from the storage hopper into the wheel hub and along the blade of the wheel, where it is accelerated to speeds in excess of 73 m/s at a flow rate of about 60 Mg/hr.4 Airless blast cleaning machines are not only efficient for cleaning castings, but can be automated to minimize labour costs. They are also more

Fig 5 Airless blast cleaning wheel layout. (Courtesy of Ervin Industries Inc., USA.)

Finishing Investment Castings

193

economical in the use of power than air blast cabinets with the same cleaning capacity. Figure 6 shows a typical airless blast machine using a Goff spinner hanger for automatic processing of castings. To obtain designed capacity, it is essential to ensure that the machine is operated correctly.4 Blast patterns should be tested on a regular basis to ensure that all the shot pellets hit the work surface. The shot mixture should be monitored to ensure the correct shot size distribution for the type of work being processed. The operator should ensure that the feed hopper remains at least two-thirds full at maximum shot throughput. All nonmetallics should be screened from the re-used shot. These recommendations may be self-evident, yet a survey has shown that a high percentage of these systems are operated at less than 50% efficiency because one or other recommendation is not adhered to. For example, if the blast stream is 10% off target, cleaning efficiency will drop by 25%. Good maintenance of the blast wheel and impeller is essential and high wear areas should be checked after every ten hours of operation. Typical wear areas are the impeller, the control cage and the wheel blades. Good record keeping will enable an efficient preventative maintenance programme to be introduced, whereby high-wear components are replaced within a specified period.

WATER BLAST CLEANING Water blast cleaning can be divided into two broad categories: Open -bay water blast cleaning This technique uses water at high pressure for removing most of the shell material and is a carry-over from cleaning sand castings. The operator (suitably attired) projects a high pressure water jet on to the mould material and breaks up the main bulk of the shell. This method is analogous to removing the bulk of the ceramic shell by vibratory knock-out. Water blast cabinet cleaning Water blast cleaning systems use very high water pressures and suitable nozzle materials to produce a coherent stream of water at high velocity which impacts the ceramic material. The technique was originally developed to clean off residual material from the casting assembly. The procedure has now been improved so that all the ceramic is removed in one operation.

194

Fig 6

Investment Casting

Automated airless blast cleaning cabinet showing the casting rotation mechanism.

(Courtesy of GOFF, A Division of George Fischer Foundry Systems, Inc., USA.)

Finishing Investment Castings

195

A great deal of work has been carried out to determine the most im portant features controlling the efficiency of water blast cleaning equipment. There appear to be three main criteria. 1. There is a threshold pressure value below which the ceramic removal rate is too slow for practical purposes. 2. Once the threshold pressure has been reached the removal efficiency increases with the flow of water at high pressure, not with increase in water pressure. 3. When flow rate and water pressure are optimized, the type of nozzle material becomes critical to maintain efficiency. It is now claimed that the integrity or coherence of the water jet is critical to maintain efficient ceramic removal. This can only be achieved on a consistent basis by using nozzle materials that are resistant to wear by water at high pressure and high flow rate, and for this sapphire nozzles are recommended.5 Their use ensures trouble free operation and the production of a reliable, coherent stream of water at high pressure. To remove ceramic material from investment castings water pressures of 300 - 650 bar are required at the highest flow rate attainable without 'down loading' the water through the safety by-pass in the system. The casting to be cleaned is usually placed inside a blast cabinet and clamped to a vertically rotating table. The water blast cleans the surface of the casting as it is rotated within the stream. An interesting variation is to clamp the casting assembly in a horizontal plane (Fig. 7). The assembly is gripped by a special patented clamp; a three-pronged, hydraulically activated claw firmly grips irregularly shaped assemblies and enables them to be rotated about a horizontal axis, while a track-mounted nozzle moves along the axis cleaning the rotating parts. This permits the design of much more compact water blast cleaning units. (Fig. 8). High pressure water blast cleaning cabinets are environmentally acceptable, being virtually dust free and quiet in operation. This cleaning method is rated as being four times faster than traditional shotblasting methods, and the flexibility of the process allows intricate coring to be removed much more rapidly than by chemical leaching.

GRINDING AND FINISHING CASTINGS The investment casting industry employs a wide range of grinding enquipment to remove gates, get rid of blemishes, and clean and polish castings. Typical items of grinding equipment are swing frame, back stand, plunge, and horizontal and vertical platen grinders, and of course

196 Investment Casting

Fig 7 Inside view of a water blast machine using the Tebbe Claw clamping device in the horizontal plane (US Patent 50444). (Courtesy of Triplex Systems, Inc., USA.)

Fig 8 A compact 'horizontal clamp' water blast cabinet. (Courtesy of Triplex Systems, Inc., USA.)

Finishing Investment Castings

197

a wide selection of hand held grinders, burrs and grinding points. This section deals mainly with the use of abrasive belts for finishing castings, because grinding wheels have largely been replaced by abrasive belts for most finishing operations. For many years gates have been removed by hand grinding, whereby the gates were manually presented to the wheel or abrasive belt. This labour-intensive method is inefficient both in time of operation and in the economic use of abrasive. For example, when the abrasive belt is new the removal rate of gate material is high; however, as the abrasive grit dulls, metal removal rates decrease rapidly. A characteristic of the abrasive grains is that under high pressure the grain will fracture and produce new cutting edges, which results in increased belt life. The pressure at which the grain fracture occurs is far above the pressure that can be manually exerted at the belt/metal interface. The development of better abrasive belt structures enables increasingly higher pressures to be used, thus taking advantage of the regeneration of the cutting edges of the abrasive under high pressure. The high pressure load is applied rapidly, hence the term 'plunge grinding'. Two basic plunge grinding concepts have been developed: fixed feed grinding and fixed force grinding. With fixed feed grinding the workpiece is advanced on to the abrasive belt at a constant rate, and to accom modate this the power of the grinding machine has to increase during the grinding operation. Figure 9 shows a typical force curve for a medium

4140 Alloy 50 grit belt

Fig 9 Characteristics offixed feed grinding. (Courtesy of Norton Company, USA.)

198

Investment Casting Table 1. Optimum feed rates for 4140 steel and superalloy PWA 1480

Alloy AMS 4140 steel PWA 1480 alloy

Feed rate

Cut increment/cycle

230 mm/min. 305 mm/min.

6.5 mm 3.2 mm

carbon steel casting. The data, generated by the Norton Company,6 show a continuous increase in force to offset abrasive deterioration with use. Tests have established optimum feed rates for specific types of alloy for a given depth of cut. Table 1 gives optimum feed rates for a 4140 steel and a superalloy PWA 1480. With fixed force grinding the rate of metal removal decreases with the number of cutting cycles. Figure 10 shows how horsepower and cut rate decrease with a constant force over a given number of cycles. Tables 2 and 3 define the basic characteristics of fixed force and fixed feed grinding systems. Grinding machine makers, often in conjunction with abrasive manufacturers, have developed machines and power assisted grinding tables (power packs) which help to optimize metal removal rates with the lowest possible operational costs. These high pressure systems force the gates into the belts at pressures of 3.5 - 5 MPa. The stock removal rate is high, the heat generated is rapidly dissipated and belt life is extended by

Fig 10 Characteristics of fixed force grinding. (Courtesy of Norton Company, USA.)

Finishing Investment Castings

199

Table 2. Characteristics of fixed force grinding

1. 2. 3. 4. 5. 6.

Least expensive Application force is constant Normally straight plunge Normally pneumatic Grinding time varies (increases with each grind) Energy requirement (HP) varied considerably during belt life. Initial HP requirement is very high and diminishes as belt dulls and coefficient of friction between belt and part decreases 7. High pressures cause grain fracture and fresh cutting edges are developed Table 3. Characteristics of fixed feed grinding

1. 2. 3. 4.

More controllable process with more consistent performance In-feed rate is constant Grinding time remains constant. The force increases as the belt dulls Energy requirements (HP) are much lower than in fixed force grinding and remain fairly constant 5. Although very high pressure can be generated, the increase is gradual and can cause dulling of the abrasive before the grain fractures to produce new cutting edges. Proper infeed rates for different alloys are very critical

the breakdown of the abrasive grains under pressure as previously discussed. The removal of gates by plunge grinding makes use of a unit, mounted in front of the contact wheel, which moves the workpiece into the abrasive belt at high pressures. Straight plunge grinding can generate a concave ground surface; this is overcome by incorporating a rise and fall mechanism in the power pack, which is actuated as needed to flatten the ground face. Ground machines have been developed to surface grind and contour grind the castings. Figure 11 shows a double end grinder, one side with a reciprocating bed and the other with a contour grinding assembly. These systems can be fully automated and can handle a high volume of parts. Castings can be loaded in multiples to take advantage of the flexibility of the grinding system. The machines can be fitted with all kinds of automatic and semi-automatic casting feed systems. A typical indexing turntable is shown in Fig. 12 illustrating a six station fixture which enables the castings to be loaded, ground and unloaded in a semi-automatic sequence. All the operator does is to physically load and unload the casting fixtures; the machine does the rest. Offhand or manual grinding systems are widely used for a number of operations. There are some general recommendations to consider when

200

Investment Casting

Fig 11 Double end grinder with horizontal feed (left) and contour grind (right) assem blies. (Courtesy o f G & P . Machinery Corporation, U SA.)

selecting new equipment. It should be ensured that the drive motor is large enough for the application; as a rough guide a minimum of 300 watts per centimetre or 1 HP per inch of belt width is required. For dual belt grinding systems, a separate motor should be used for each belt and the belt tensioning devices need to be more than adequate for the type of work being performed. A major factor in the design of grinding machines is the type of contact wheel used. Contact wheels can be smooth or serrated, hard, mediumhard or soft depending upon the application. They can be of steel, alumin ium, rubber or plastic and smooth or serrated. The selection of the right machine depends on the type of work being processed; in general, the harder wheels are used for more aggressive grinding and the softer wheels for finer finish applications. The contact wheels may be serrated to various degrees to optimize and control the grinding or stock removal rates for a given material and abrasive grit size. When all basic equipment selections have been made, there are three other operational variables to consider; these are belt speed, belt tension and idler wheel design. Tensioning devices can be simple mechanical

Finishing Investment Castings

201

Fig 12 Multi-station gate grinding fixture. (Courtesy of Investment Casting Institute, USA.) springs, hydraulic cylinders or pneumatic cylinders. Hydraulic tensioning is ideal for heavy duty machines, but pneumatic tensioners are most widely used because the constant pressure control associated with air tensioning will accommodate operating variables such as belt stretching. The design of the idler wheel is very important in the operation and performance of abrasive belt equipment. The idler can be tracked automatically so as to ensure that wider belts are uniformly loaded (i.e. that wear is not concentrated in a narrow strip of the belt). Because of the large selection of abrasives available, investment casters are advised to seek specific recommendations from the manufacturers. Tables 4 and 5 give some general guidance on the effect of belt tension and speed on grinding performance and Table 6 lists typical belt speeds for a range of alloys.

Table 4.

Tension range Optimum Tension too low Tension too high

Effect of belt tension on grinding performance

80-350 kPa 140-170 kPa Tracking and belt slip problems Excessive wear on spindle bearings. Premature belt failure. Distortion of contact wheel.

202

Investment Casting Table 5. Effect of belt speed on grinding performance

Belt speed High >23 m/s (4500 FPM) Low 00. 1976

BS970 : 19 72

EN

Carbon Steels

XC 18 XC 32 XC 42M

G S 45 G S 52 G S 60

1 .0443 1 .0551 1 .0553

C1020/1/2/3 C1030 C1040

A1 A2 A3

050A22 060A32 060A42

3 5 8

C LA 2

IVrtfeMn Steel

20 M 6

20 M 6

1.5060

C10 27

A4

150M 19

14A

C LA 3

(45 - 55 Ton) 70 0 - 85 0 N/mm? Alloy Steel

40 NC D 6

34 Cr NI M o 6

1 .6582

9840

BT1

816M40

24

C LA 4

(55 - 65 Ton) 8 5 0 - 1 OOON/mm ’ Alloy Steel

32 NC O 14 M

30 Cr Ni Mo 8

1 . 6580

4337

BT2

823M30

25

High Tensile Steels

40 NC D 10

C LA 7

3 % Cr Mo Steel

35 CO 12 . M

24 Cr M o 10

1 . 7 2 73

CLA 8

Carbon Steel Sudace Hardening

XC 42 . TS . M

Ck 45

1 . 1 19 1

C LA 9

Carbon Steel Case Hardening

X C 15

Ck 15

1 .1 1 4 1

C LA 10

3 % Ni Case Hardening Steel

20 NC D 1 2 .T S .M

10 Ni 14

1 .5637

C LA 11

3% Cr Mo Nttriding Steel

20 C O 12

27 Cr Mo 13 .5

1 . 7365

A&B

1 % Cr Abrasion

50 C 4

50 Cr M o 4

1 . 7228

5 14 7

C

Resisting Steels

60 C D 5

60 Cr Mo 4

1 . 7229

4150

Ni Mo Steel

15 N D B

C LA 1

C LA 5

C LA 12

C LA 13

A B C

A B

5328

8 26M 31(Z)

B4

722M 24

29

C1040

AW 2

060A40

8

C10 16

AW 1

080A15

32

33

B4

4 6 17

722M 24

40

665H17

34

BW2 ; BW3 : BW4

BS 3146, CLA 3, 4 and specifies S and P only under chemical composition: the comparable specifications given are similar materials: however, other alloys will give the mechanical properties required.

Design for Investment Casting 359 Table 7. (cont.) BS 3146

Type of Steel

Speritication ANC 1

A B C

A NC 2

ANC 3

Pari II : 1975

A B

Comparable Specifications

Afnor (see N F A 3 2 . 0 5 6 )

Work staff

A .I .S .I .

A .C .I .

A .M . S .

G X 1 2 Cr 14 G X 20 Cr 14 G X 22 Cr 14

1 .4008 1 . 4 0 27

403 420 420

CA 15

Z 28 C 13 M

18 % C r 2 % Ni Martensitic Steel

Z22 C N 18 .02

GX22 Cr Ni 1 7

1 .4059

431

CB 30

5353

Austenitic 18% Cr 8% Ni Sleets

Z6CN18.10 - M

CX10 Cr Ni 18.8

1.4312

304

CF8

5358:5341

26CN Nb 18.10 - M

GX7 Cr Ni Nb 18.9

1.4552

347

CF8C

Z6 CND 18.12-M Z6CN D Nb 18.12-M

GX6 Cr Ni Mo 18.10 GX6 Cr Ni Mo 18.10 GX7 Cr Ni Mo Nb 18.10

1.4408 1.4408 1.4581

317 316 318

CG8M CFBM

13 % Cr Martensitic Steels

Z12 C 13 M

D IN

N O T ES e .g . Trade Names

_

539:53500

CA 40

Brilish/USA es iteo nr» • 410C21 420C29

S80

8S 970 : 1970 4 10 S21 420 S29 420 S37

56A 56B 56C

431 S29

57

es )ioo ten 304C15

302 S25

58A

5362E

347C17

347 S17

58F

5524C

317C16 316C16 318C17

317 S16 316 S16 320 S17

58J 58H 58H

3 1X45 331C60 334C11

310 S24

ANC 4

A B C

Austenitic 18% Cr 10% Ni 3% Mo Steels

ANC 5

A 8 C

Nickel Chromium Steels

Z12 CNS 25.21 Fe N37C18S NC 15 Fe

Ni Cr 25.20 GX40 Ni Cr Si 36.16 Ni Cr 60.15

1.4843 1.4865 2.4867

310 330

CK 20;K HU MW

ANC 6

A 8 C

Chromium Nicket Steels

Z20 CNS 25.12 Z25 CNS W22 Z15 CNW S22.13

GX35 Cr Ni Si 25.12

1.4837

309

CH20.HF

Nickel 20% Cr 0.4% Ti Alloy

NC 20 T

NI Cl 20 Ti

2.4630

ANC 9

Nickel 20% Cr 2.5% Ti 1 .2% At Alloy

NC 20 TA

Ni C i 20 Ti Al

2.4631

Nimocasl 60* Nimonie 80*

,

ANC 10

Nickel 20% Cr 16.5% Co 2.4% Ti 1.3% Al Alloy

NC 20 K17 TA

N< Cr 20 Co 18 Ti

2.4632

Nimocasl 90* Nimonie 90*

,

ANC 11

Nickel 21% C 10% Mo 10% Co Alloy

NC 21 DK 10

ANC 13

Cobalt 26% Cr 10% Ni 7% W Alloy

KC 25 NW

Co Cr 25 Ni W

2.4966

5382E

X40» Stemte 3U

ANC 14

Coball 27% Cr 5.5% Mo 2 .7% Ni Alloy

KD 270 N

Co Cr 28 Mo

2.4979

53850

Stellite 8'

ANC 15

Nickel 28% Mo Alloy

Ni Mo 28

Ni Mo 30

2.4482

5396

HaseMoy 8» i

ANC 16

Nickel 17% Mo 16.5% Cr 4.5% W Alloy

Ni Mo 15 Cr 15

Ni Mo 16 Cr W

2.4537

S388C

Hastelloy C»

ANC 17

Nickel 9% Si 3% Cu Alloy

Ni Si 10 Cu

2.4566

Ni Cu 30 Fe*

2.4360'

ANC 8

ANC 18

A B C

ANC 19

ANC 20

A 6

Nu 30 Fe

PH Nickel-Cr Nb Mo FeW Alloy

NC 20 Nb OW

A B C

53668

30X30 30X30

Nimocasl 75* Nimonie 75*

C r -N iCu Steel

55

, ’

C242t

CW 12M

Hastelloy D»

4544*

Monel* i Monel H*» Monel H ' t P.E.10/1 M.C.102>

F.V.520 ’ F.V.520»

PH . Cr Ni Cu Mo Steels Cr Ni Cu Mo Steel

ANC 21

ANC 22

Nickel 31% Cu 1% - 4 % Si Alloys

EN

CD 4MCu

5342 5343 5344

17/4 pH

tRegistered trade mark and/or proprietary alloy * Similar materials

360

Investment Casting

The former arose from the observation that fatigue failure in aircraft turbine rotor blades originated at transverse grain boundaries. The use of DS techniques (e.g. in CM247 LR) overcame this limitation and the con tinued development, leading to single crystal blades, gave the opportunity to optimise alloy composition and to eliminate additives which had been necessary to strengthen the grain boundaries of equiaxed alloys but which reduce the melting point of the alloy. This led to modem, more efficient design of turbine blading. Cobalt superalloys, although less extensively used than their nickel counterparts, fill specific and important niche markets (e.g. for implant prostheses). Since they do not have the Ni3Al or strengthening mechan ism of the nickel alloys, they tend to be less strong. In addition to the creep-resistant alloys, castable nickel superalloys have also been developed specifically with enhanced corrosion resistance, an example being Hastelloy C-22; a similar trend with cobalt superalloys has led to the introduction of Ultimet. Tables 8 and 9 list the compositions of some of the currently used superalloys; most of these are known by trade names or designations introduced by the developers of the materials. Fig 8 illustrates the great improvement in temperature capability achieved by the superalloys over the years: in practical terms, the performance can be still further im proved by the judicious use of advanced blade cooling techniques. Aluminium Alloys Apart from the superalloys, aluminium alloys are the most widely used for non-ferrous investment castings. In the UK, the routine alloys tend to be selected from the well-established BS1490: LM series or the BS (L)/DTD aerospace specifications.11' 12 In addition, the industry makes use of US specifications, especially for high strength premium investment castings. Within the BS1490 series, four alloys find main use (Table 10). LM6, the binary aluminium - silicon eutectic alloy, offers excellent casting characteristics with high fluidity that allows the manufacture of intricate, thinsection castings. LM16, a 5% silicon - 10% copper alloy, has good strength up to about 200°C, also coupled with good foundry characteristics. LM5, a 5% magnesium alloy, finds limited use, with the ability to give a polished surface, but it can cause foundry problems due to the magnesium content; the same comments can be applied to the now obsolete LM10, an alloy with 10% magnesium. The greatest number of aluminium alloy investment castings are made in LM25 or its variants. LM25, a 7% silicon -0.5% magnesium alloy, is selected where good properties are required in castings of a shape or of a complexity demanding an alloy with excellent castability and a high level

Design for Investment Casting 361

7001__________________I__________________I__________________I-----------------------------1— 1940 1950 1960 1970 1980 A P P R O X IM A T E Y E A R O F IN T R O D U C T IO N

Fig 8 Temperature capability of alloys versus year of introduction (from S.W.K. Shaw, Materials Science and Techn., 1987 , 3, 682-694).

W. Betteridge

and

of soundness. The alloy also offers corrosion resistance coupled with weldability. Of the aerospace specifications, mention may be made of BS L99, a 7% silicon alloy which has a high level of strength together with castability. The draft European specification prEN 132 lists six aluminium alloys used in investment casting.13 These are shown in Table 11, which indicates the compositions together with the estimated percentage use. There is considerable similarity between this and the UK LM series and at least four of the six are essentially equivalent materials. It may be noted that most types of aluminium alloys can be heat treated to develop enhanced mechanical properties, particularly tensile strength, and such treatment can offer significant improvements. Heat treatment is based on solution treatment just below the solidus, rapid cooling to retain a supersaturated aluminium-based solid solution and precipitation

362

Investment Casting Table 8. Cast nickel base superalloys (after Quigg10)

Alloy

Co

Cr

Mo

W

Ta

Nò

Al

Ti

Hi

C

Others Ni

_ 12.5 4.2 _ _ IN 713 C 2 6.1 0.8 _ .12 B 1900* 10 8 6 4.3 6 .10 1 Rene 80* 4 9.5 14 4 .17 3 5 IN 792* 9 12.7 2 3.2 4.2 3.9 3.9 .21 IN 100 15 10 3 5.5 4.7 .18 1 V IN 738 LC 8.5 16 1.7 2.6 1.7 0.9 3.4 3.4 .11 Rene 125 7 3.8 10 9 2 4.8 2.6 1.6 .10 MAR M 200* 10 19 12.5 1.8 5 2 .15 MAR M 247 10 8.5 0.6 10 3 5.6 1.0 1.4 .16 4 12 5 10 PWA 1480f 5.0 1.5 8 9 2 6 4 0.5 3.7 4.2 G E N - 4t 5 8 0.6 8 6 5.6 1.0 CMSX - 2®t 5 8 0.6 8 6 5.6 1.0 0.1 CMSX -3®t 5.6 1.0 0.1 3 Re CMSX -4®t 10 6.5 0.6 6 6 5 8 10 SRR 99| 5.5 2.2 * Later versions of B 1900, Rene 80, IN 792 and MAR 200 had hafnium added t Single crystal alloy -

-

-

-

BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL BAL

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

Table 9. Cast cobalt base superalloys (after Quigg10) AHoy

ASTM F -75 ST-21 ST-21 WI-52 ST-31 Ultimet

Cr

Mo

28 27 20 25.5 25 26

5.5 5 -

-

5

W

Ni

C

Fe

_

_

0.25 0.27 0.10 0.45 0.5 0.06

3 2 2 1.5 3

-

15 11 7.5 2

3 10 10 9 -

Others

Co

BAL 2 Nb

BAL BAL BAL BAL BAL

_

thereafter by heating at a moderately elevated temperature. Many variations are practised, depending upon the relative importance of maximum strength, ductility or structural stability.14 European heat treatment designations are as follows: F 0 T1 T4 T5 T6 T64 T7

As cast Annealed Controlled cooling fromcasting andnaturally aged Solution heat treated andnaturallyaged where applicable Controlled cooling from casting and artificially aged or overaged - Solution heat treated and fully artificially aged - Solution heat treated and artificially underaged - Solution heat treated and artificially overaged (stabilised)

-

The correct heat treatment for a particular part is an important aspect of design and should be considered at the outset.

Table 10. Typical aluminium alloys for investment casting (from Mills12) Specification BS 1490

Chemical Com position0/*) Cu Mg Si Fe Mn Ni Zn m in max m in max m in max m in max m in max m in max m in max 0.1

3.0

6.0 -

LM 6F

-

0.1

-

0.1

LM 16™ T6 F

1.0

1.5

0.4

0.6 4.5

5.5

0.6

0.2

0.20 0.6

6.5

7.5

0.5

LM 5F

LM

Comparable Specifications

25Ilb5

0.3 -

10.0 13.0

-

Others

USA AA

0.6

0.3

0.7

-

0.1

-

0.1

514.0

0.6

-

0.5

-

0.1

-

0.1

-

0.5

0.25 -

0.1

0.3

0.1

0.1

-

355.0 355.0 356.0 356.0 356.0 356.0

T7

Mechanical Properties

USA Canada France Com m ercial Com m ercial A 57 - 702/3

Germany DIN1725/2

UTS 0.2% PS N/m m 2 N/m m 2 EL % m in min min

213

B320

AG 3T

G - AIM g5

140

90

355 355 356 356 356 356

46020 125 125 135 135 135 135

AS 13 AS 7GY20 AS 7GY23 AS 7GY30 AS 7GY33

G - AISi 12 G - AISiM g -

160 170 230 130 150 160 230

60 120 220 80 120 80 200 125 220 165 200 185

2 -

-

G - AISi7M g

3 5 2 2 1 2.5 -

Aerospace Alloys 0.1 0.8 0.5 0.10 -

1.7 0.25 0.10 -

0.1 0.10 0.10

7.5

0.20

0.10 -

0.10 -

0.10

0.10 -

0.25 -

0.40 0.20 0.30 1.3

1.7 -

-

-

0.10 1.0

1.5

0.25

0.1

0.5

0.75 6.5

7.5 -

0.8 2.0 0.05 0.20 1.5 2.8 0.8 1.0 1.5 0.4 0.6 4.5 5.5 0.10 0.25 4.0 5.0 -

BSL 119

4.7

5.5

-

BSL 154 BSL 155 BSL 169 DTD 716 DTD 722 DTD 727 DTD 735

3.8

4.5

-

0.1 0.1

0.3

0.8 3.5

6.0

-

-

0.20 0.6

0.1

-

0.10 0.5

-

0.05 0.5 -

0.1 0.10 0.1

Zr 10/.30 Co 10/.30 Sb 10/.30

-

355 A356

125 226 226 B135

A - S 2U A - U5GT A - U5GT -

G - AICu4Ti G - AICu4Ti G - AISi7M g

160 250 220 280 230

-

-

-

-

-

215

190

1

-

-

-

-

_

357

-

-

-

-

-

215 280 290 125 145 165 230

160 200 240 77 124 93 200

7 4 3 2 1 2.5 -

-

-

-

-

-

B116 B116 B116 B116

A - S4G A - S4G A - S4G A - S4G

G - AISi5M g G - AISi5M g G - AlSi5M g G - AISi5M g

7 4 2

363

obsolescent

0.10 0.20 0.45 6.7

-

355.0 A356.0

Design for Investment Casting

1.4 0.6 0.25 -

BSL 51* BSL 78* BSL 91* BSL 92* BSL 99

364

Investment Casting Table 11. European specifications for aluminium alloys for investment casting (from reference 14)

Alloy group AlCu AISi7Mg AlSi AISi5Cu AIMg

CEN number 21000 42000 42100 42200 44100 45200 51300

Designation Al Cu4MgTi Al Si7Mg Al Si7Mg0.3 Al Si7Mg0.6 Al S ii2(b) Al Si5Cu3Mn Al Mg5

In recent years there has been increased emphasis on premium quality aluminium investment castings, with high levels of strength, which can be maintained at temperature. Typical of the alloys used for these are A356, A357 and A201.15The first two of these are similar to LM25 but with the use of high purity constituents to reduce deleterious effects on ductility. The addition of beryllium gives A357 which, while offering inferior casting characteristics to A356, can give 0.2% proof stress levels of 275 MPa, ultimate tensile stress of 345 MPa and 5% elongation. Following this line of approach still further, A201 (formerly known as KO-1) offers an unusual combination of high strength with high ductility and toughness and, of all the aluminium cast alloys, can most closely approach the strength of equivalent wrought material; it can hence be designed as a direct replacement for forgings and assemblies, with projected cost savings. The alloy contains a nominal 4% copper and 0.5% silver, the presence of the latter being considered to affect beneficially the Al-Cu-Mg-Si precipitation sequence. Using premium quality casting techniques, guaranteed properties from castings in the T6 (fully heat treated) condition can be as high as 345 MPa proof stress and 415 MPa ultimate tensile stress, with a minimum elongation of 5%; actual strengths as much as 3040% higher have been claimed.15 The alloy is fairly widely used in the investment casting field although its use has been limited by rather poor casting characteristics. In particular, it is susceptible to hot tearing and this restricts the design options. Recently reported work16 claims that hot tearing can be minimised by selecting a 'safe' composition range within the overall composition limits. Copper Alloys Investment castings in copper-base alloys form a small but important part of the market.12 Typical of the alloys used are those of the BS1400 specification shown in Table 12.

Table 12. Typical copper base alloys for investment casting (from Mills12) __________________________________________ Chemical Com position %_________ _________________________________________________Mechanical Properties Cu BS 1400

Phosphor Bronze Leaded Gunmetal Alum inium Bronze High Tensile Brass

m in

max

REM REM REM REM REM 55.0 -

Zn max

m in

10.0 _ 11.0 13.0 4.0 6.0 _ 0.10 0.10 1.0 -

_ 4.0 _ -

min

Pb max

0.05 0.30 6.0 0.50 0.50 REM

min -

4.0 -

max

min

max

min

max

min

max

m in

max

UTS N /m m * m in

-

-

0.10 0.50 2.0 1.0 5.5 1.0

-

0.10 0.15

8.5 8.8 0.5

0.01 0.01 0.01 10.5 10.0 2.5

-

1.0 1.5 3.0

220 220 200 500 640 470

P max

m in

0.25 0.50 0.50 0.15 6.0 0.05 0.05 0.50 -

Ni

4.0 -

Fe

1.5 16.0 0.7

AI

3.5 5.5 2.0

Mn

0.2% PS N/m m * m in

El % min

130 130 100 170 250 170

3 5 13 18 13 18

Design for Investment Casting

PB1 PB2 LG2 AB1 AB2 HTB1

Material

Sn

365

366

Typical Analysis

Mechanical Properties UTS

0.2% Proof

H.T. Con. F T4

N/m m 2 140

F T4 T7

200 200

13 13

Alloy MAG 1

Al 8.5

Zn 0.7

Mn 0.25

Zr

Rare Earth

-

-

MAG 3 [AZ91]

9.5

0.7

0.25

-

-

ZE63

-

5.7

0.25

0.7

2.5

T7

275

MAG 5 [ZE41]

-

4.25

-

0.7

1.25

T1

200

200 125

N/m m 2 85 80

tsi 5.5 5.2

4 ~

95 85 130

5.5 8.5

17.8

5

170

11

13.0

3

135

8.8

tsi 9 13

8

El. %

2 6

-

6.2

Characteristics

General purpose alloy having good strength, ductility, and shock resistance. Good castability General purpose alloy fo r intricate castings, having good strength and pressure tightness. Good castability High strength alloy, having good castability, but expensive due to the necessity fo r a hydrogen gas heat - treatm ent General purpose alloy, having good uniform strength properties and good castability

Investment Casting

Table 13. Typical magnesium base alloys for investment casting (from Mills12)

Design for Investment Casting 367 Leaded gunmetal (LG2C) contains 85% copper and 5% each of tin, zinc and lead. It is the most corrosion resistant of the brasses and consequently is widely used for marine and environmental applications. Being easy to cast and offering moderate strength, it is also widely used for general engineering applications. The phosphor bronzes PB1 and PB2 also give good bearing and wear properties, coupled with good foundry characteristics. PB2 is little different in strength from PB1 but has a significantly better elongation figure. Two grades of aluminium bronze are available and these have different iron contents. They are high strength, corrosion resistant materials used in pumps and similar applications. However, due to rather poor founding characteristics, they tend to be restricted to simple design configurations. HTB1, a high tensile brass, is used for lever arms, brackets and similarly highly stressed parts. An alloy not included in BS1400 is beryllium -copper (beryllium bronze) a heat treatable high strength alloy with 2.5% beryllium and up to 0.55% cobalt. It has the highest electrical conductivity of any alloy of comparable strength and it can be hardened to as much as 400 HV. Because of the beryllium content, however, the alloy poses safety and environmental problems in the foundry. Magnesium Alloys There are very few investment castings produced in magnesium alloy but the material is cast in a limited number of foundries which have installed specialised equipment to deal with it.13 The alloys used tend to fall into the BS2970 or BSL series of alloys and these are listed in Table 13; a number of the alloys are heat treatable, so offering a wide range of mechanical properties. Titanium Alloys Titanium alloys have many engineering and economic advantages but their development has been retarded because of severe problems in handling the molten metal or alloy. However, in recent years there have been marked advances in the science of titanium casting and precision sand and investment casting methods are now both used; the latter process gives better dimensional tolerances, better surface finish and the ability to deal with more complex designs. Where castings are for aerospace applications (e.g. airframe or aero engine parts), the main reason for the use of titanium is the excellent strength/weight ratio and the usual alloy chosen in this case is a Ti-6A1-4V material of mixed (alpha + beta) structure. For general

368

Investment Casting

engineering applications, corrosion resistance is usually the primary requirement and this is met by using commercially pure titanium (CPTi) or titanium with 0.2% palladium. Most of the investment casting applications are in fact in the aerospace field and about 86% of castings in titanium are made from the Ti-6A1-4V alloy. To ensure full soundness, hot isostatic pressing is routine for such parts.18 PURCHASE OF INVESTMENT CASTINGS The logical completion of the design process is the ordering of the investment castings and the designer or buyer should, at this stage, follow certain simple procedures to ensure best results. 1. The buyer should be familiar with the capabilities of individual investment casting foundries. Most investment casters are well used to the production of high quality work - as the possession of their various quality approvals (see Table 14) testifies - but some specialise in one type of work while others tend to deal with other types; again, some are more familiar with ferrous casting, others with non-ferrous, while others deal with both. Yet again, some foundries are particularly adept at producing very thin wall castings. Consideration of these factors should lead the buyer to select the most appropriate foundry for his particular application. 2. The buyer should state, succinctly but thoroughly (and in writing), his precise requirements. These will be clearly based on the functional requirements foreseen for the part. It should be appreciated that different castings may, depending on their application, demand different quality levels. Casting quality can be assessed on various scales, typically from Class 1 to Class 4, the former being considerably more critical than the latter (see Table 15). The choice of the appropriate class can affect component price, by its effect on quality control and testing procedures. 3. The need for the designer to establish a technical liaison with the producing foundry has been emphasised, and this liaison should take account of design modifications that may be recommended by the foundry to ease production and promote casting soundness. Obviously, any such changes must not impair the functional aspects of the design, but if able to be accepted, they are likely to reduce production costs. Nadin19 has given two examples of this approach. The knuckle joint shown in Fig. 9a is a type 316 stainless steel part for a marine application which was originally machined from the solid. All dimensions are castable except the diameter of 0.860-0.865 inch (21.84-21.97 mm); by widening the

Design for Investment Casting

369

Table 14. Typical casting quality approvals (from reference 3) D E S IG N A T IO N

A P P L IC A B IL IT Y

B S 5750 P a r ti (IS O 9001)

Q u a lity sys te m s fo r D e s ig n , M a n u fa c tu re a n d In s ta lla tio n .

BS 5750 P art 2 (IS O 9002)

Q u a lity s y s te m s fo r M a n u fa c tu re a n d In s ta lla tio n .

O p e n to a ll s u p p lie rs in a ll in d u s trie s . A s s e s s m e n t is c a rrie d o u t b y a p p ro v e d in d e p e n d e n t b o d ie s .

A Q A P -4

N A T O In s p e c tio n s ys te m R e q u ire m e n ts fo r In d u s try . (T h is re p la c e s th e U .K . D e fe n c e S p e c ific a tio n D E F - S T A N - 0 5 /2 4 ).

R e q u ire d o f d ire c t c o n tra c to rs to M O D , A s s e s s m e n t is c a rrie d o u t b y th e M O D - D i re c to ra te o f D e fe n c e Q u a lity A s s u ra n c e . (D Q A , s u c c e s s o r to th e A Q D ) .

CAA B ritis h C iv il A ir ­ w o r th in e s s R e q u ire m e n ts (BC AR )

U K C iv il A v ia tio n A u th o r ity A ir N a v ig a tio n O r d e r A p p ro v a ls .

R e q u ire d o f a ll s u p p lie rs to th e B ritis h C iv il A irc ra ft In d u s try . A s s e s s m e n t c a rrie d o u t b y th e C A A .

BCS

B ritis h C a lib ra tio n S e rv ic e .

A c c re d ita tio n s ys te m fo r C a lib ra tio n L a b o ra to rie s .

NATLAS

N a tio n a l T e s tin g L a b o ra to ry A c c re d ita tio n sys te m A c c re d ita tio n S c h e m e . fo r te s t la b o ra to rie s .

NAMAS

N a tio n a l M e a s u re m e n t A c c re d ita tio n S e rv ice .

STANDARD

T h e e x e c u tiv e b o d y , b a s e d at th e N a tio n a l P h y s ic a l L a b o ra to ry , c o - o r d in a tin g BCS a n d N A T L A S .

tolerance very slightly to 0.859-0.867 inch (21.82-22.02 mm), together with the inclusion of a 0.030 inch (0.762 mm) radius at the bottom of the slot, the part could be cast directly to size with no subsequent machining. A cost saving of 30% resulted. In the second example, the T-handle component (Fig. 9b) is a more complex design but it is readily made in mild steel by machining and welding. While normal investment casting tolerances are acceptable, the only change was an agreed increase from 2 mm wall thickness to 3 mm, making for easier casting and thereby allowing the part to be cast as a single piece, with a cost saving in this case of some 50%. 4. The buyer should require the submission of a sample casting, with production to start only after approval of this sample; this should apply to new work and where major design changes have been made. This practice is sometimes omitted, to speed up completion of the order, but such a practice can be dangerous. Approval of a sample serves three distinct

370

Investment Casting Table 15. Examples of casting quality classification (from reference 3) Class 1

A casting, the single failure of which would endanger the lives of operating personnel, or cause the loss of a missile, aircraft or other vehicle. Class 2

A casting, the single failure of which would result in a significant operational penalty. In the case of missiles, aircraft and other vehicles, this includes loss of major components, unintentional release or inability to release armament stores, or failure of weapon installation components. Class 3

Castings not included in Class 1 or 2 and having a margin of safety of 200% or less.

Class 4

Castings not included in Class 1 or 2 and having a margin of safety greater than 200%. Grades

Castings shall be of grades A, B, C or D as shown in the appropriate tables of the standard.

purposes. It provides a check on the design of the component; tools made for secondary operations can be checked to make sure they will be satisfactory when production commences; and it offers the buyer and foundry an opportunity to determine acceptable quality or permissible dimensional deviations at least cost and loss of production time.

CONCLUSION This chapter has presented the basic principles that should govern the design of investment castings, to obtain the most efficient and cost effective product. The design process should be based on an accurate perception of the properties and characteristics needed to ensure adequate performance and the temptation to over-design, simply because the process is technically capable of such refinement, should be avoided. It was Ruskin who said that good quality is never an accident; it is the result of a conscious desire for a better product. The same may be said of design, and routine acceptance of this philosophy, coupled with detailed consultation between those involved, will lead to the best use of investment castings and the full capitalisation of their many advantages.

Design for Investment Casting

371

(a) Casting dimensions for a knuckle joint in type 316 stainless steel, originally machined from solid for marine applications. Dimensions shown in inches.

(b) A T-handle in mild steel. Dimensions shown in millimetres. Fig 9

Design modification for investment casting (from Nadin19).

REFERENCES 1. H.T. Bidwell: Investment Casting, The Machinery Publishing Co Ltd, London, 1969. 2. P.R. Beeley: Foundry Technology, Butterworth, London, 1972. 3. Designers' Handbook for Investment Casting, British Investment Casting Trade Association, Birmingham UK, 1990. 4. Investment Casting Handbook, Investment Casting Institute, Dallas, Texas, US, 1979.

372

Investment Casting

5. J. Campbell: Castings, Butterworth Heinemann, Oxford, 1991. 6. J. Hocking: 9th BICTA Conference, London, May 1968, Paper No 4. 7. VDG Reference Sheet P690 (Investment Casting), VDG, Dusseldorf, Germany, 1992. 8. Guide to Alloy Selection, vol. 1, British Investment Casting Trade Association, Birmingham UK, 1980. 9. Investment Casting Specifications: (a) Carbon and Low Alloy Steels; (b)Corrosion and Heat Resistant Steels, Nickel and Cobalt Base Alloys; (c) Cross Reference Specifications, British Investment Casting Trade Association, Birmingham,

UK, 1978. 10. R.J. Quigg: 8th World Conference on Investment Casting, London, June 1993, Paper No. 25. 11. The Properties of Aluminium and Its Alloys, Aluminium Federation (8th Edition), Birmingham UK, 1981. 12. D.F. Mills: Investment Casting for the 1990's, BICTA, Birmingham UK, September 1991, Paper No. 4. 13. Aluminium and Aluminium Alloys: Casting - chemical composition and mechanical properties, CEN Pr En 132/100 (5th Draft), 1982. 14. British and European Casting Alloys - Their Properties and Characteristics, (R. Hartley Ed., Association of Light Alloy Refiners, Birmingham UK 1992. 15. M. Randall: Foundry Trade )., 1987,161 (3360), 948-951. 16. H. Chadwick: Ibid., 1993,167, 3484, 642-644. 17. P.J. Bridges and F. Hauzeur: Cast Metals, 1991, 4, (3), 152-154. 18. D. Eylon, J.R. Newman and J.K. Thorne: Metals Handbook, 10th Edn, Vol. 2, pp634-646. 19. G. Nadin: private communication.

12. REVIEW OF APPLICATIONS

INTRODUCTION This chapter, surveying applications of investment casting, is arranged in four separate sections, each designed to stress the process and product characteristics associated with a well-defined commercial field. The sections are differently structured, reflecting their respective key features. In Section 12.1, the main theme is the progressive metallurgical evolution of what is essentially a single type of product, the gas turbine blade, and the distinctive role of investment casting in its development. Section 12.2, by contrast, covers a wide and diverse product spectrum ranging across the whole engineering landscape and encompassing many different alloys and cast shapes; this highlights the versatility of the process as demonstrated in its varied illustrated examples. Many producers are, of course, active in both these fields. Sections 12.3 and 12.4 cover more specialised sectors relating to jewellery and to surgical and dental products. These areas are largely, although not wholly, separate from the mainstream of investment casting activity and the respective sections have been written by specialist authors involved with them at first hand. Since both entail aspects of materials, equipment and manufacturing technique falling outside the scope of other chapters, these longer sections are designed to provide selfcontained accounts of both production and application. DOI: 10.1201/9781003419228-12

1 2 .1

Application to Aerospace P.R. BEELEY

Investment casting has been closely identified with aerospace products since World War 2 saw the advent of jet propulsion and the gas turbine. The requirement for turbine blades caused the process to emerge into the engineering arena, after a long history in which dental, jewellery and art casting had provided its principal outlets. The development of investment casting technology since that initial step continued for a long period in parallel with progressively more exacting requirements for turbine blades, imposed by the need for aircraft engines with increasing thrust, thrust-to-weight ratio and operating efficiency. This progress is highlighted by the fact that the available thrust has grown by a factor of about 50 over the same period. The operating factor most closely involved in the advance of both performance and efficiency in the gas turbine is temperature. Since high stresses are also encountered, alloy development has been aimed at maximizing resistance to high temperature creep and oxidation. Heat resisting steels were used in the early stages, but these soon gave way to the nickel and cobalt based superalloys which best combine these qualities. Early aerofoil castings were confined to fixed rather than moving parts of the engine, for which wrought material was preferred, in line with the concurrent perception of its greater integrity and reliability. Initially both individual and multi-aerofoil nozzle segments were produced as investment castings, these being the stationary elements in the engine. Later development extended the use of castings to rotating aerofoils, or blades, since the investment casting process offered wide freedom of design and the considerable economic advantage of near net shape products. In addition the cast structure gave greatly enhanced scope for extension of the crucial properties to progressively higher temperatures. The rotating turbine blade is a highly stressed, safety-critical component, providing arguably the most challenging application for any engin eered product and requiring total integrity and reliability; the adoption of investment casting was thus a highly significant development in the DOI: 10.1201/9781003419228-13

Application to Aerospace

375

history of the process. Other aerospace applications have exploited the same rigorous standards. There are two routes by which the designer can pursue the aim of producing blades able to withstand higher operating temperatures within the engine. The first is alloy development, to improve creep, fatigue and oxidation resistance and so to enable the blade to attain higher temperatures without fracture, degradation or excessive deformation. The second is the modification of blade design to embody cooling systems which give a lower blade temperature for a given gas path temperature. As will later be shown, the direct approach of alloy property extension to higher temperatures is itself facilitated by developments in casting technology, whilst investment casting techniques have made a major contribution to the blade cooling objective. High temperature alloys The first of the nickel-base alloys was derived from a simple nickelchromium alloy originally used for resistance heating elements, where stress levels are low and working life largely determined by resistance to oxidation and embrittlement. The 80Ni-20Cr Nimonic 75 material, which was considered suitable for operating temperatures up to 750°C, engen dered a series of alloys in which the temperature capability was gradually enhanced with the introduction of other elements, primarily designed to increase creep resistance by providing obstacles to plastic deformation. The face-centred cubic gamma matrix structure of the nickel-base alloys can be strengthened both by solid solution elements such as cobalt and molybdenum, and by the introduction of dispersions of a finely divided intermetallic phase, gamma prime (y'). The essential elements for the formation of this phase are titanium and aluminium, producing the compound Ni3 (Ti Al). One particular advantage of the investment casting route is the capacity of the cast structure to carry high volume fractions of this strengthening phase, such as would impair the high temperature deformation properties required for the forging of a wrought blade. Further strengthening is achieved by the formation of carbides of the general composition M23C6 which have a beneficial effect on the creep properties of the grain boundaries; this is an important consideration in a normal polycrystalline structure since the boundaries become a source of weakness at high temperature. Elements such as chromium, titanium and zirconium form such carbides. The progressive development of properties in the nickel-base alloys is treated in detail in Reference 1. The introduction of the various elements involved in the enhancement of creep resistance does have the effect of lowering the solidus temperature as well as raising the process temperature required to dissolve the gamma prime phase, so that the forging

376

Investment Casting

temperature range becomes progressively narrower and casting the more practicable alternative, since creep resistance can then be pursued to tem peratures very near the solidus. The approach to enhanced creep resistance through large volume fractions of y', based on high aluminium and titanium contents, gives alloys of lower density, since smaller additions of the high density solid solution strengthening elements can then be used. Such compositions are, however, prone to the formation of sigma and other embrittling phases at intermediate temperatures and so require close control of composition using phase computation techniques based on alloying theory; this represents one of the more recent advances in the field of high temperature alloys. The use of cast alloys in this application is also greatly facilitated by the availability of vacuum melting and casting plant. Vacuum processing gives close control of composition by reducing losses of reactive alloying elements, especially titanium and aluminium, through preferential oxida tion. Gas porosity is eliminated and the incidence of non-metallic inclu sions is reduced to low levels in keeping with the high quality standards. The clean, protected melting conditions enable furnace charges to be based on pre-alloyed melting stock which has been produced in bulk and checked for specification and quality. A further important contribution to the use of investment cast blades has been made by rigorous quality assurance procedures and non destructive testing techniques, particularly dye and fluorescent penetrant testing and radiography. Microstructure control Apart from the role of alloy composition, casting conditions too can be controlled to develop microstructures with the appropriate high temperature properties. One technique successfully used in turbine blade production has been grain refinement by the incorporation of reagents in the primary investment coating applied to the wax patterns. Cobalt alumínate has been used for this purpose and the resulting microstructures are fine and homogeneous, avoiding local concentrations of low melting point segregates and porosity associated with coarser dendritic structures. Castings subjected to this treatment have greatly improved fatigue resistance. A more radical form of structure control is the deliberate use of directional solidification for the production of aligned microstructures, to be subsequently examined in detail. Blade cooling Before consideration of the further process and alloy developments in the turbine blade field, it will be useful to mention the second approach to

Application to Aerospace

377

enhanced performance, that of blade cooling, in which gas flow through passages within the blade is used to reduce the component temperature for a given gas temperature within the engine. Investment casting provides a highly efficient means of forming the internal cooling passages. Pre-formed ceramic cores produced by specialist manufacturers are in serted in the dies before wax injection. The wax patterns embodying the cores are slurry coated and the shell thickness is built up in the normal way. After the standard dewaxing and firing sequences the cores remain locked in the shells as an integral mould feature ready for casting. The cores are finally removed from the confined passages in the casting by chemical leaching. The coring technique permits the shaping of passages of great complexity. The progressive early evolution of cooling performance in gas turbine blades is demonstrated in Fig. 1 and made a major contribution to the increasing temperature capability of the engine. An equally important advance in the potential of investment cast blades came, however, with the development of the directional solidification technique. Its role in the enhancement of blade performance has been fully reviewed in References 3 and 4, and is subject to detailed treatment in Reference 5. Directional solidification The concept of directional solidification as an aid to feeding for the im provement of internal soundness in castings has long been understood

Fig 1

Evolution ofblade cooîing principle (after H.E. Gresham2).

378

Investment Casting

and applied. Freezing in a temperature gradient for the intentional pro duction of aligned microstructures came much later with the manufacture of all-columnar cast permanent magnets, noted for their outstanding properties in the direction of the grain structure. A further application was in the production of high purity metals by zone-refining, in which quantitative control of the freezing process was established through basic studies of the main solidification parameters. This approach was also adopted in the manufacture of investment cast blades. The conditions for directional solidification were referred to in Chap ter 10, where Fig. 15 shows an arrangement embodying a water-cooled chill and an exothermically lined mould, designed to ensure that growth will be confined to crystals nucleated on the chill surface. For full con trol, however, it is necessary to maintain a constant freezing rate along the whole length of the casting. This requires a special furnace operating on the principle shown in Fig. 2. A ceramic shell mould with an open base is mounted on the water-cooled chill and after pouring the whole assembly is gradually withdrawn from the induction heating coil.

Fig 2 Main principle of casting equipment employed for controlled directional solidification.

Application to Aerospace

379

Fig 3 Examples of multi-crystalline, directionally solidified precision cast aero-engine turbine blades, incorporating complex internal cooling passageways form ed in situ by the use of ceramic cores. Blades etched to display grain structures. (Courtesy of A E Turbine Components Ltd).

Solidification again begins at the chill surface and proceeds upwards, but a steady state is reached in which the solid-liquid interface is maintained at a fixed position in a steep temperature gradient, suppressing further nucleation ahead of the interface and inducing longitudinal growth of columnar grains until the whole blade has solidified. Fig. 3 shows examples of cast turbine blades etched to reveal the all-columnar structures. This type of aligned structure gives a major improvement in creep properties, mainly due to the elimination of most of the transverse grain boundaries; these represent a source of weakness at high temperature, especially under the centrifugal stresses generated in high speed rotation of the turbine. The directionally solidified structure is also associated with a high standard of feeding, minimizing porosity in the blade section. The preferred crystallographic orientation developed in the longitudinal direction enhances fatigue life: this results from a reduction in the

380

Investment Casting 30

o \° z < tr h-

cn

a. LU

LU

oc o

0

0

100 TIME

HR.

Fig 4

Comparative creep behaviour of ( V conventional]]) cast (equiaxed), (2) D S colum nar and (3) single crystal nickel-base alloy MarMlOO, at 980 ° C approx and 20 7 MPa (after F.L. VerSnyder and M .E. Shank6).

Young's Modulus, which causes any strain due to differential expansion to generate a lower stress in the blade. The typical influence on creep behaviour is illustrated in the comparative deformation-time curves shown in Fig. 4. Stress rupture life and creep ductility are enhanced relative to the random cast equiaxed structure, and blade castings of this type were widely adopted in aircraft turbines. The principle was later extended to the production of single crystal castings, so eliminating the longitudinal boundaries as well. This requires an investment casting mould shaped with constrictions and changes of direction in the gate, designed to eliminate grains by competitive growth during the approach of the solidification front towards the casting itself, until only a single grain remains. Various mould designs have been used to achieve this condition; one of these is shown in Fig. 5. Seed crystals can also be used, to develop selected orientations relative to the blade profile. The further improvement in creep resistance made possible by the single crystal technique is shown in Fig. 4. Full three-dimensional control of orientation, with homogeneity, also maximizes mechanical fatigue resistance and tensile strength. One production furnace as used for the directional solidification of investment cast blades is illustrated in Fig. 6. It embodies a self-tapping vacuum induction melting crucible placed above a mould chamber, from

Application to Aerospace

381

Fig 5 Arrangement fo r production of single - cn/stal blade casting, embodying spiral constriction (Courtesy of Rolls Royce PLC).

which the mould assembly can be withdrawn at a controlled rate without breaking the protective vacuum until solidification is complete. Heat flow conditions and detailed aspects of the equipment and techniques used for directional solidification are more fully described in Reference 5.

382

Investment Casting

Fig 6 Main features of Rolls Royce production unit for manufacture of directionally solidified turbine blades (from M.]. Goulette et al/j. The ability to produce single crystal castings opened a new approach to the composition of the high temperature alloys, removing the need for grain boundary strengthening elements such as carbon, boron and zirconium. Without these elements the solidus temperature can be raised and other elements substituted to enhance the properties of the crystal itself after solution treatment. Great care is required in the processing of single crystal blade castings in the modified compositions. Unwanted grain boundaries constitute a major hazard and can, for example, result from mechanical damage during stripping and cleaning, followed by recrystallization on subsequent

Application to Aerospace

383

Fig 7 Large aero -engine gas turbine blade in nickel-base alloy, with general luall thickness o f 1.5 mm (Courtesy of A E Turbine Components Ltd).

heat treatment. Rogue boundaries could cause catastrophic failure in the absence of the strengthening elements. This necessitates advanced techniques of automatic non-destructive orientation assessment, coupled with rigorous process and quality control at all stages. Process control features introduced in the search for consistency in this critical application of investment castings include the previously mentioned self-tapping crucible system employed by some producers. Rapid melting of the charge is terminated by the final melting of a pre-formed retaining disc of the alloy, positioned to seal the bottom taphole; repro ducible pouring temperatures are achieved on a power input-time basis, which eliminates the need for direct measurement. Ceramic filters can be embodied in the gating systems to trap any oxide inclusions. Castings in high temperature alloys are subject to freckling, a defective condition arising from the movement of alloy-segregated liquid of low melting point through pipes or 'channels' in the solidification zone. The

384

Investment Casting

driving force for this is ' therm osolu tal ' convection caused by local d if ferences in liquid density, due to both tem perature and com position. The liquid phase becom es enriched in the lighter alu m inium and titanium alloy com ponents relative to the dendrites, and the convection nucleates equiaxed grains. Steep temperature gradients minimize this effect by reducing the extent of the semi- solid zone and promoting fine dendrite arm spacings, but attention has also been given to adjustments of composition to reduce the density differences and convective flow. A heavy element, tantalum, can be sub stituted for part of the tungsten content, and itself partitions preferentially to offset the segregation of the lighter components. This exam ple demonstrates how the control of properties in single crystal castings requires a full understanding of the alloy constitution.

(a) Fig 8 Investment cast nickel-base alloy blades for large land-based gas turbines used in electrical power generation up to 200 Mw. Cooling passageways cast in situ using ceramic cores.(a) (above) equiaxed structure, in aerofoils up to 800 mm and 25 kg (b) (opposite) multicrystalline DS columnar structure, in aerofoils up to 500 mm and 20 kg (Courtesy of AE Turbine Components Ltd).

Application to Aerospace

385

This must include the partitioning characteristics of elements between the matrix and the y' phase, and prediction of their effects on solid solution strengthening and precipitation hardening. Computer modelling techniques can be employed to simplify these problems. Modem advances in the application of investment casting have frequently been associated with the special thin-wall capability of the process and this applies in the aerospace as well as in the general engineering field. Fig. 7 shows a large aero-engine gas turbine blade, vacuum cast in IN 713LC nickel-base alloy. In this case the component is cast hollow for lightening purposes to leave a general section thickness of only 1.5 mm, so that the aerofoil section resembles a sheet-metal product; similar characteristics are seen in the much larger blades cast for land-based gas turbines and exemplified in Fig. 8.

Fig 8 (b)

386 Investment Casting

Fig 9 Cross section of investment cast turbine wheel in nickel-base alloy, showing fine structure through hub section (Courtesy of AE Turbine Components Ltd.)

Further aerospace applications Although any treatment of investment castings for the aerospace field must necessarily focus on gas turbine blades, other important components are produced in a variety of alloys. These share the same high integrity requirement and similarly high inspection standards. They include large and complex thin wall engine carcase parts, such as housings, to replace sheet metal fabrications, whilst integrally bladed turbine wheels for smaller engines have provided major cost savings. One of the latter components is shown sectioned in Fig. 9. In this case special solidification conditions were required to produce a uniform fine-grained

Application to Aerospace

387

Fig 10 One-piece investment casting of hollow nozzle guide vane aerofoils for helicopter gas turbine; diameter 300 mm (Courtesy ofAE Turbine Components Ltd). microstructure in the relatively thick hub section, thus ensuring maximum resistance to low cycle fatigue failure. Other large castings include one-piece nozzle guide vane aerofoils, integrally cast to enhance rigidity and eliminate assembly costs: the example shown in Fig. 10 has a diameter of 300 mm and forms one complete stage of a helicopter gas turbine. Even larger investment castings are produced in titanium alloy. Fig. 11 shows integrally cast stators for an aircraft gas turbine, whilst Fig. 12 shows a large and complex intermediate case with hollow struts, also in titanium alloy. This casting, 102 cm in diameter and weighing 100 kg, is the main frame supporting the front of the aircraft engine. A notable example of an aircraft structural component is the light alloy wingtip investment casting used on the A340 Airbus and illustrated in Fig. 13. This unit is approximately 500 mm in length, with a wall thickness of 1.6 mm, and considerable cost savings were made as compared with the previous fabricated version. Greater stress on complexity is seen in the examples illustrated in Figs 14 and 15. The former is a main gearbox housing for the auxiliary engine

388

Investment Casting

Fig 11 Integrally cast titanium stators fo r aircraft gas turbine (Courtesy of Howmet Corporation).

Fig 12

Intermediate case casting with hollow struts , in titanium alloy. Diameter 1020

mm, iveight 100 kg (Courtesy of Howmet Corporation).

Application to Aerospace

389

Fig 13 Aircraft wingtip casting in aluminium alloy. Length 500 mm, wall thickness 1.6 mm (Courtesy of P.L Castings Ltd, AICAL Division).

and starter motor in a military aircraft, cast in a heat resisting steel; the unitary construction dispenses with the need for sub-assemblies as would be required in the absence of the exceptional capabilities of the investment casting process for rigid and precisely aligned design features in multi-membered components. The complex missile launcher component

Fig 14 Investment cast gearbox housing for auxillianj aircraft engine and starter motor, in 20 C r 10 Ni 3W steel to BS 3146 A N C 6B (Courtesy ofCronite Precision Castings Ltd).

390

Investment Casting

illustrated in Fig. 15 enabled maximum weight savings to be combined with a substantially lower cost than that of any alternative manufacturing route. A different type of mechanical function is seen in the critical aluminium alloy pulley system shown in Fig. 16. Again produced to aircraft stand ards, the castings govern the movement of control wires positioning com ponents such as tailplanes and wing elevators and are therefore crucial to the safe operation of the aircraft. Minimal machine finishing of bores and mating surfaces is all that is required to achieve the final dimensions. Such examples indicate the diminishing distinction between aerospace and general engineering applications, at least in terms of component geometry: the former has, however, exerted a major beneficial influence on design and quality standards over a much wider range of uses.

Fig 15 Investment cast load bearing component in 17/4 high strength precipitation hardening stainless steel (Courtesy of PA. Castings Ltd).

Application to Aerospace 391

Fig 16

Investment cast aluminium alloy pulleys fo r aircraft control wires, to BS 2L 99

and Class 1 inspection conditions (Courtesy of Stone Foundries Ltd).

12.2 General Applications of Investment Castings R.F. SMART

While it is true that turbine engine components such as blades and vanes constitute the biggest single market for investment castings, nevertheless there are a very great diversity of other applications — both in the released sector (including defence applications) and in general commercial outlets. In many cases, the choice of the investment casting route is dictated by technical considerations — property requirements, design complexity — but in more and more cases the process is being selected because it offers the cheapest manufacturing option in terms of total costs. Table 1, prepared by R. Palmer of Trucast Ltd, lists a range of typical applications for the process, while Fig. 17 illustrates this diversity.

Table 1. Typical applications of investment castings

Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Aerospace Automotive products Fighting vehicles Field guns Automatic weapons Rifles/sporting guns Missiles Rocketry Satellites Hovercraft Medical implants Medical instruments Orthopaedic appliances Nuclear power

Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Power generation (land based gas turbines) Food processing Petrochemical equipment Computers Pumps Safety equipment Yachting chandlery Bicycle parts Roulette Wheels Equestrian products Artwork Golf clubs Musical instruments Cigarette making machinery

KEY TO FIGURE 17 (opposite) 1: Marine equipment, boat fittings; 2: gun and rifle parts; 3: vehicle components; 4: instrumentation parts; 5: agricultural machinery; 6: chemical and processing industry parts; 7: oilwell and mining equipment components; 8: surgical and orthopaedic parts; 9: business machines and computers; 10: sports equipment castings; 11: machine tool components; 12: textile, knitting and sewing machinery; 13: electrical components; 14: parts for pneumatic and hydraulic equipment. DOI: 10.1201/9781003419228-14

General Applications of Investment Castings

Fig 17

Typical applications of investment casting (Courtesy ofP.I. Castings Ltd).

393

394

Fig 18

Investment Casting

Defence applications of steel investment castings (Courtesy of Deritend Precision

Castings Ltd).

Fig 19

Investment cast optical housing parts for Starstreak missile systems , the components being cast in type IT90 armour quality steel (Courtesy of Deritend Precision Castings Ltd).

General Applications of Investment Castings

395

1. Black Cap motor fin attachments 2. Roll type boost nozzle 3. Rear rail casting 4. Actuator housing parts 5. Three parts for Rapier launching system 6. Support stem nozzle for Blowpipe 7. Sidewinder missile part (not referred to in text - steel, composition unknown)

Fig 20

Selection of steel missile components investment cast (Courtesy of Deritend

Precision Castings Ltd).

Fig 21

Knuckle joint in 316 stainless steel; cast T -handle in mild steel, both manufac tured by investment casting (Courtesy o f S & T Precision Castings Ltd).

Steel investment castings account for about one-third of the total output by value. Many are for the defence industries and Fig. 18 shows typical applications. The two parts for the alternator for the Chieftain Battle Tank (shown top left) are manufactured in BS3146 CLA1 A, a plain carbon steel offering a range of tensile properties after suitable heat treatment. Shown centrally lower in the photograph is a gimbal housing (a conical ring) for a ship's compass and this has been manufactured in stainless steel. The other components are parts for the thermal observation gunnery sight for

396

Fig 22

Investment Casting

Selection of 316 stainless steel castings (Courtesy o/B SA Precision Castings Ltd).

a battle tank and these include a scanner cover, yoke and cradle in BS3146 CLA5A, a typical high tensile steel. Fig. 19 illustrates investment cast optical housing parts for the Starstreak missile systems. All are cast from a special armour quality steel (type IT90) and the parts cover a range of sizes; the circular mount is 330 mm in diameter, 200 mm high and weighs 21 kg, while the largest of the parts shown measures 430 mm in length, 400 mm in height and 330 mm in width, with a weight of 47 kg. These applications correct the common erroneous belief that investment castings are only used for very small components.

General Applications of Investment Castings

Fig 23

397

Various examples of stainless steel investment castings, mainly in type 316 alloy

(Courtesy of BSA Precision Castings Ltd).

398 Investment Casting

Fig 24

316 type stainless steel, investment casting valve parts (Courtesy Deritend Precision Castings Ltd).

General Applications of Investment Castings

Fig 25

399

Investment cast golf club heads (Courtesy of Dunlop Slazenger International

Ltd).

Other missile applications for steel investment castings feature in Fig. 20. The two fin attachments for the Black Cap Motor which powers the Sea Wolf Missile are made from RS200 steel, with very low sulphur and phosphorus

400

Investment Casting

Fig 26 Selection of wheels and rotors made by investment casting (Courtesy ofD eritend Precision Castings Ltd). levels. There are also show n a num ber of other applications, cast in a variety of steels — a roll type boost nozzle (A60 alloy), a rear rail casting (H C102 alloy), an actuator housing for a M cD onnel Douglas H arpoon M issile (BA M S 5343), three parts for a Rapier Launching System (BS3146 A & C2) and the support stem nozzle for a Blowpipe missile (DTD 5172, HC7). Investm ent casting can be the preferred m anufacturing route in m any quite m undane applications, and Fig. 21 show s tw o such castings, a knuckle joint in type 316 stainless steel and a T - handle in mild steel. The form er is for a m arine application and was originally m achined from solid; by changing to the cast route a reported cost saving of 30% resulted. The T - handle, a little m ore com plex in design, w as originally m ade by m achin ing and w elding the parts. The com ponent was m ade by investm ent cast ing at a cost saving of 50%. Both of these parts were subject to slight re design to optim ise the change and this is discussed in Chapter 11. Stainless steel investm ent castings are w idely used, a typical selection of parts being show n in Fig. 22. Fig. 23 show s a stainless steel im peller investm ent casting (centre low er) and, in 316 alloy, from the left bottom corner clockw ise, a spool valve, a box cover, a butterfly disc, an im peller/ rotor, a boss for a heat exchanger and a pum p cover for a brew ing application. In fact, the use of steel investm ent castings is com m on in this area and Fig. 24 show s a range of investm ent cast 316 stainless steel valve

General Applications of Investment Castings

Fig T I

401

Investment casting parts fo r operational and model gun s (Courtesy of Deritend

Precision Castings Ltd).

parts, made to fit aluminium alloy lager casks; Fig. 25 illustrates the ubiquitous investment cast golf club head which has been one of the commercial outlets for the process for very many years. Fig. 26 displays a selection of wheels and rotors made by the process. These include such components as power rotors for compressed air turbines, up to 168 kg in weight, produced in 14/4 PH steel, nozzle ring castings for turbo chargers, compressor wheels, stainless steel (316L) wheels for a water pump and impellers in AWC3B. These applications illustrate the wealth and diversity of parts made in this application area. Fig. 27 shows the wide range of investment castings made in a variety of alloys for both operational and model guns; these range from triggers, sights and hammers to breech blocks, barrels and bolts. Much work has been carried out by the industry in recent years to produce castings of high integrity, fully sound and therefore suitable for fatigue-rated applications. The cartridge extractor for a 30 mm naval cannon, shown in Fig. 28, is an example of this type of approach. By attention to the process to prevent the occurrence of inclusions, together with a post-casting hipping (hot isostatic pressing) operation, the steel investment castings can be produced with fatigue strength as good as that of similar forgings. In life proof trials of the cartridge extractor shown, this critical component performed satisfactorily and this led to the specification of investment casting as the preferred manufacturing option.

402

(a)

Investment Casting

(b)

Fig 28 30 mm naval cannon (a) in which the cartridge extractor (b) is now specified as an investment casting (Courtesy of N EL — Croivn copyright reserved).

One company using such high integrity castings in critical applications has estimated total manufacturing cost savings ranging from 30 -70%; the cartridge extractor already mentioned, for example, was about 38% cheaper to manufacture as an investment casting than as a forging. It is not sugested that the use of hipping be incorporated routinely in the manufacture of components that are already performing adequately as castings; the new developments are, in the main, intended to open up an extended market for castings offering exceptional fatigue performance under extreme conditions. It is believed that very considerable potential exists in this field. Among the non-ferrous alloys, there is a wide range of applications of aluminium and its alloys. These have, in fact, had a very long history and the statue of Eros in Piccadilly — which celebrated its centenary in 1993

General Applications of Investment Castings

403

— is comprised of a series of investment castings, one of the earliest uses of aluminium castings. In present day outlets, one of the major application areas for aluminium investment castings is in electronic boxes and chassis. Fig. 29 shows a typical chassis casting for an electronic application and Fig. 30 a selection of box type castings. Fig. 31 comprises a selection of aluminium alloy investment castings used in a TV camera. These were produced in aluminium alloy BS1490 LM25 in the fully heat treated (T6) condition, and the casting route was developed by close cooperation between customer and supplier as an alternative to the use of sheet metal fabrication, to give a major cost saving.

Fig 29

Alum inium alloy investment cast chassis casting (Courtesy of Deritend Preci-

sion Castings Ltd).

404

Investment Casting

Fig 30 Investment cast aluminium alloy boxes (Courtesy of Deritend Precision Castings Ltd).

Fig 31 Selection of LM25 alloy investment castings used in a TV camera (Courtesy of Finecast (Maidenhead) Ltd).

General Applications of Investment Castings Group of fo u r L M 25 aluminium alloy investment castings used in the construction of a hand -held thermal imager (Courtesy Finecast (Maidenhead) Ltd).

405

Fig 32

406

Investment Casting

Fig 33 A range of titanium alloy investment castings used in aerospace (Courtesy o flE P Structures SE T TA S Division).

The use of LM25 (or its derivatives in the aluminium -silicon magnesium system, such as 2L99) is widespread in the industry and the next example (Fig. 32) shows a group of four castings used in the con struction of a hand-held thermal imager. Again, this represents a change from another manufacturing method and was the outcome of cooperation between foundry and user. The main housing and the cover are thin walled castings that reduce weight, while the internal castings were designed to offer maximum rigidity. It is interesting that the equipment had originally come into service three years earlier and at that stage had weighed three times as much as it does now. This dramatic improvement in the performance/weight ratio was due partly to advances made during that period in thermal imaging technology and partly to the development of an optimum design of investment casting. Titanium investment castings are used in a range of applications and Fig. 33 shows a selection of such castings for aerospace applications. As in the case of the steel castings mentioned earlier, these castings are normally hot isostatically pressed to close up internal porosity and to give ascast mechanical properties equal to those obtainable from equivalent wrought material. Other applications for titanium investment castings

General Applications of Investment Castings

407

include boat fittings, offshore installations, chemical plant, general industrial machinery and prostheses. Fig. 34 shows one of the larger investment castings that are produced within Europe; the limit here is around 1 tonne - although heavier castings can be produced in the USA - and this is determined by the pour capacity of the available vacuum arc-melting furnace.

Fig 34

Investment cast titanium component for gas turbine, weighing 70 kg (Courtesy

o flE P Structures S E T T A S Division).

12.3 Jewellery Investment Casting P.E. GA1NSBURY

INTRODUCTION AND HISTORICAL DEVELOPMENT Over the second half of the twentieth century investment casting has grown to become probably the most important fabrication process in the quantity production of precious metal jewellery. It has also become widely used in fine jewellery, particularly for repeated motifs in large pieces, and in designer jewellery, often using hand made wax patterns. The silversmithing industry has a long history of the use of sand casting for such items as handles, spouts, feet and other decorative parts on holloware, and for figure, bird and animal models. Unlike jewellers, however, silversmiths were slow to adopt the investment process. Over recent years the traditional silver foundries have all but disappeared and the specialised skills involved in silver sand casting have been lost. During this time there were improvements in the quality of relatively large castings available from the investment process and silversmiths have now turned to these products. The techniques of investment casting developed for jewellery and silver have recently also become increasingly used in the production of art medals in gold and silver alloys and bronze. This has greatly widened the design possibilities for medals and has established medal design as a distinct art form. The casting of jewellery by lost wax techniques goes back to the ancient origins of the process. However, the modern technique for quantity production was only developed in the twentieth century from the techniques of dental casting, which originated in the latter years of the nineteenth century. The starting point of the modern process was the development of a liquid investment slurry in the 1890s by Philbrook.8 It is to him and his successors in the dental profession9 that the whole modern investment casting industry owes its foundation. Further aspects of dental casting development and practice are treated elsewhere in the present chapter. DOI: 10.1201/9781003419228-15

Jewellery Investment Casting

409

As early dental castings were invariably in precious metal alloys, it was a small step to use the same techniques to produce jewellery castings. These would usually have been one-offs, using hand made wax patterns or natural objects that could be burnt out, such as leaves, insects and nutshells. Almost certainly the first true investment castings for jewellery were made by dental technicians. Flexible wax dies The key step in the development of investment casting as a full-scale jewellery production process was the technique of making flexible vulcanised rubber dies for the production of wax patterns, developed by Jungersen10 in Canada in 1935. The dies are produced from positive master patterns made by normal jewellery fabrication techniques and enable jewellers to make complex and undercut wax patterns rapidly and cheaply. The process does not require the services of skilled die sinkers and allows far more complex pieces to be produced than are possible by stamping or pressing in traditional dies. The development of the process was retarded by World War 2 and it did not significantly reach Europe until the late nineteen forties. Meanwhile, in view of the strong patent held by Jungersen, alternatives were sought in the USA and processes for making multipart dies in low melting point alloys were developed. These proved satisfactory for simple non -undercut patterns and could be used with the same simple wax injection equipment as rubber dies. They were widely employed by manufacturers to avoid licence fees. However, once the rubber die technique became freely available around 1950, the use of fusible alloy dies rapidly declined, except in a few specialised applications requiring high precision patterns. Wax injection When multiple jewellery castings became a requirement, wax injection techniques were needed. Initially, simple syringe injection was used, or even centrifugal casting of wax, using spring driven machines to be described later. Neither technique was satisfactory and low-pressure air injectors with thermostatically controlled, electrically heated wax pots became commercially available after the war, first in the USA and later in the UK. Today, however, high quality wax injectors for jewellery casting mainly originate from Germany and Japan. Casting waxes Dental waxes were far too soft for injection purposes. Some of the injection waxes for industrial casting were suitable for jewellery patterns, but being brittle were difficult to remove from complex dies. As the process

410

Investment Casting

developed in the jewellery industry, however, waxes with a satisfactory balance of properties were developed by the suppliers. Investments Initially dental plaster/silica investments were the only ones available to jewellery casters. These, being formulated to fulfil the difficult expansion requirements of dentistry, were expensive and were made to use as very thick, quick-setting slurries unsuitable for jewellery investing. In the USA, however, some of the large dental supply houses began to market plaster/silica investments tailored to jewellery requirements. These, being free of the rigorous requirements of dentistry, were cheaper and setting times were controlled at around 5 to 10 minutes. They had sufficient fluidity at the correct powder-water ratios to allow the investing of large moulds of complex patterns to be comfortably completed. In the USA these suppliers still dominate the field. In Europe the only significant jewellery investment manufacturers are in the UK and none is of dental origin. Plaster/silica investments were, and still are, the preferred materials for casting gold and silver alloys at temperatures up to about 1100°C. Above this temperature, however, breakdown of the calcium sulphate content begins. This can result in heavy sulphur contamination of some white golds and palladium- and platinum-rich alloys with casting temperatures between 1100 and 2000°C. The most satisfactory investments for such alloys are based on finely ground, extremely pure quartz. When mixed with a small volume of water this gives very free flowing slurries. These set by water separation on standing and form block moulds that will withstand casting temperatures of around 2000°C. Platinum castings from this type of investment probably have the smoothest finish of any commercial investment castings. This is attributed to slight glazing of the refractory and the complete absence of oxide formation on the platinum surface. This type of investment was developed in the USA where suitable silicas were available. In the UK in the late 1940s it was found that less pure indigenous silica powders would behave similarly if a small quan tity of a sequestering agent such as sodium hexametaphosphate was added to the mixing water11. For a time such investments were popular in the UK for all types of jewellery casting, but plaster/silica investments subsequently regained pride of place. Today the pure silica investments are only used by a few specialist platinum casters. Proprietary chemically setting, phosphate-bonded high temperature investments may also be used with high melting point jewellery alloys. Jewellery is almost always cast in cylindrical block moulds. Ceramic shell casting has been used but has never generated significant interest among jewellery casters.

Jewellery Investment Casting 411 Investing technique It was here that the jewellery process began to diverge from dental techniques. Dental investment slurries are thick and quick setting. To avoid the formation of air bubbles at the pattern surface the investment is often applied with a small brush while the pattern and its support are held on a small vibrating table. The assembly is then set up in a metal ring, into which further investment is vibrated to form the complete mould. With the greater complexity of some jewellery patterns and of multiple pattern assemblies, the greater size of castings, and the need for quantity production of moulds, hand investing became impracticable. Vibration of moulds after filling with investment was ineffective in dislodging entrapped air from complex patterns and, with the brittle waxes initially used, frequently resulted in broken patterns. The problem was solved by vacuum treatment of the investment slurry immediately after mixing, and again after pouring into the mould ring. Once vacuum de-aeration had been established there was little subsequent change other than scaling up to use larger moulds and the introduction of automation. The most sig nificant progress has been the understanding of the mechanism of setting of plaster investments and the recognition that close control of mixing water temperature, mixing time and mould handling was necessary for the highest mould quality. Dewaxing and burnout The only major development in this area has been a limited introduction of steam dewaxing. Instrumentation and automatic control of burnout furnaces have become virtually universal. Melting Initially, following dental practice, gas/air or gas/oxygen torch melting was used for melting all jewellery alloys. In skilled hands this is fast and efficient, although only visual temperature control is possible. However, the weight of charge is limited to around 500 grammes and oxidation of melts, compositional changes and contamination can occur if there is any lapse in technique. These problems were first addressed by the scaling up of small dental centrifugal casting machines incorporating platinum wound or carbon resistor melting furnaces, and machines with induction melting were later widely adopted. As requirements for ever larger melts arose, particularly in the Italian industry, separate gas, resistance or induction melting furnaces for pouring into static, vacuum -assisted casting machines became desirable, allowing simpler atmosphere and temperature control than was possible with melting units integrated into casting machines.

412

Investment Casting

Casting machines Jewellery and dental casting moulds cannot be filled by simple gravity pouring. Reproduction of fine surface detail and thin sections from unvented moulds of low permeability with small volumes of metal, demands some form of pressure casting. In the early days of dental casting, various devices had been developed to exploit centrifugal force, air pressure or steam pressure. The most effective of these was centrifugal casting. In its simplest form a melting basin was formed in the top of the mould and connected to the cavity by a narrow sprue. The sprue diameter was such that the molten metal could not pass through by gravity alone. The mould was held in an open-top container, to which was connected a short length of chain fitted with a swivel and a handle at the other end. When the metal was ready for casting the mould was swung round by the operator until the metal had solidified. This technique was still in use in dentistry in the fifties and was effective if hazardous. It was unsuitable, however, for the larger melts required for jewellery casting. Simple mechanisation brought spring driven casting machines. In prin ciple these consist of a horizontal bar on which the mould is secured at one end, in front of a melting hearth with a pouring spout butted up to the open end of the mould. The weight of this assembly and the metal charge is counterbalanced by adjustable weights at other end of the bar. At its centre the bar is mounted on a vertical spindle connected to a clock spring. When the spring is wound and released, molten metal in the hearth is rapidly thrown into the mould. A freewheel permits the arm to rotate for sufficient time to allow the casting to solidify. The principle is illustrated in Fig. 35 and an example of an actual machine in Fig. 36. Machines of this type, scaled up from the original dental models, were almost exclusively used for precious metal jewellery casting up to the 1960s. With increased demand for castings, the size of centrifugal machines increased and power driven machines became available that could cast up to 500 grammes of 18 carat gold from a single melt, near the limit for torch melting. Machines were also produced, notably by Jelraus in the USA, employing resistance melting in platinum wound furnaces with graphite crucibles. Induction melting with a spark gap generator featured in a smallcapacity centrifugal casting machine introduced by Ecco in the USA in the late 1930s. Ecco machines were still in use in the post war period and a few had reached Europe. In the late 1950s larger induction machines using valve generators became available in Europe, initially from two Italian manufacturers, Aseg Galloni and Manfredi, and eventually machines capable of melting and casting up to 7 kg of 18ct gold were

Mould

Jewellery Investment Casting

Sectional illustration of the principle of horizontal, spring driven casting machines using torch melting.

413

Fig 35

414

Investment Casting

Fig 36

A Kerr horizontal centrifugal casting machine using gas torch melting (Courtesy of Hoben Davis Ltd).

developed. The general arrangement of such machines is illustrated in Fig- 37- . In centrifugal casting machines with induction melting molten metal transfer is inefficient. The melting crucible is normally of conventional form with a pouring lip, or a hole if hooded. Melting takes place in a vertical induction coil that is retracted immediately before the start of rotation of the casting arm. As the speed increases, centrifugal force causes the molten metal to climb the sloping side of the crucible opposite the centre of rotation. When the metal reaches the lip it is thrown into the mould, which is fitted horizontally with its open end close to the lip. The metal thus enters the mould as a thin, fast moving stream. As the whole assembly is rapidly rotating in the horizontal plane, the stream when it leaves the crucible lip curves back against the direction of rotation and may strike the side of the entrance cone to the mould.12 This can cause turbulence, aggravated by the high rotation speeds required to lift small melts from deep crucibles, leading to air entrapment in the moulds and porosity in the castings. Static casting using simple vacuum assistance had been in use in the USA for small scale jewellery casting, and on a larger scale for base metal

Melting configuration Optical pyrometer

jewellery Investment Casting 415

Fig 37 Sectional illustration o f the principle of horizontal centrifugal casting machines with induction melting. (a) (above) M elting coiifiguration (b) (see over) Casting configuration after 180 degree revolution.

416

Casting configuration after 180 degree revolution

Investment Casting

Fig 37(b)

Jewellery Investment Casting 417

Fig 38

Sectional illustration o f the principle of flanged flask, vacuum assisted casting machines.

fashion jewellery casting. Moulds are placed, sprue up, on a heat resisting gasket over a hole in a vacuum-tight box connected to a vacuum pump or tank. The vacuum is opened to the box just before pouring and is maintained until solidification is complete. The reduction of pressure in the mould cavity is sufficient to allow atmospheric pressure to force the metal into the mould. For very small scale work a plain vacuum chamber, connected by flexible tubing to a separate stepped mould support plate, has been successfully used for high quality work. The vacuum chamber can be evacuated by a water pump and can double as a de-aeration facility. Given a good water pressure this British manufactured equipment provides the individual jeweller with a complete casting facility at very low cost. In 1970 an improvement to the vacuum-assisted casting process, originated by the Italian manufacturer Di Maio, started a revolution in jewellery investment casting that broke the monopoly of centrifugal techniques in large scale production casting. Efficiency was increased by employing

418

Investment Casting

Fig 39 Sectional illustration of the principle of fully integrated, static, vacuum assisted casting machines.

perforated moulding flasks fitted with heavy flanges at the metal entry end. Moulds were made by the normal investing technique, with the holes sealed by paper tape or a rubber sleeve. There can only be a single sprue entry to the mould cavity and no patterns can be located above the flange level. The machine is provided with an open-top cylindrical chamber to contain the mould, which is supported by a ring with a gasket

Jewellery Investment Casting 419 corresponding to the flange on the flask, on the general principle as shown in Fig. 38. In the original machine the casting chamber was located in the top of a larger vacuum chamber, with a quick-opening valve con necting the two. This assembly was mounted on a rotary vacuum pump, making a compact, inexpensive machine of high casting capacity and considerably more efficient than the earlier technique. The machine has been much copied, with minor variations, and the technique is still giving excellent service over the whole range of jewellery and silversmith casting production. The only exception is its unsuitability for high melting point platinum and palladium alloys. Despite the developments it must be said that vacuum -assisted casting techniques, though most effective in skilled hands, are less forgiving than centrifugal casting, which remains the most popular method for small scale operations. Following the success of the closed chamber vacuum -assisted casting machines, manufacturers in Germany began to develop fully integrated vacuum -assisted machines. These aimed to increase efficiency by com bining induction melting with bottom pouring, the casting chamber being sealed to the botton of the crucible assembly. Some of these machines have enclosed melting chambers as illustrated in Fig. 39, allowing vacuum or controlled atmosphere melting, with accurate temperature control by thermocouples fitted in the crucible stopper rods. They may be used with fully automatic operation once the metal charge and mould have been placed in the machine. In parallel with the latter development in static casting equipment several manufacturers produced centrifugal machines in which the melting coil and rotor arm were enclosed in a vacuum chamber. These expensive hybrids were slow in operation and were not widely used; many gold alloys contain zinc and thus cannot be melted under vacuum, and there was little evidence of improved castings in other alloys. A curious spin -off from the success of vacuum -assisted casting has been the application of vacuum to the bottoms of moulds being cast in otherwise ordinary induction melting, centrifugal casting machines. This technique, with its further complications, was claimed to draw harmful gases out of the mould cavity before the metal was thrown in. The fact that the only gas likely to be present in a properly fired mould would be beneficial carbon dioxide was overlooked. Since vacuum accumulators were not fitted and only small-diameter vacuum connection tubes were used, any significant evacuation of the mould seems unlikely. This idea had originally been applied in the USA in the 1930's to the Torit torch melting, spring driven, centrifugal casting machine. This was unique in that it spun in the vertical plane and gave an

420

Investment Casting

exceptionally fast take-off compared to conventional horizontal axis machines. The vacuum was provided by a small pump driven from the rotation spindle. The same machine is still made today but without the vacuum pump. It and British machines developed from it remain, in some authorities' opinions, the most satisfactory for the casting of jew ellery platinum alloys. (Figure 40)

Fig 40 A Nesor vertical, spring driven casting machine using torch melting (Courtesy of Hoben Davis Ltd).

Jewellery Investment Casting

421

CURRENT PRACTICE

Applications Jewellery With its refinement and improved productivity over recent years, investment casting has largely replaced stamping and pressing in the production of jewellery. It is estimated that in the United States the process accounts for 90% of all production. Economics are an important factor here. Die making costs for press work are much greater than for rubber dies, and high production is needed to offset these costs. On the other hand, when high production is required, stamping and pressing is much cheaper in unit cost once die costs have been amortised. Large production runs are, however, the exception rather than the rule in jewellery production and even more so in silversmithing. Mechanical forming cannot com pete with the flexibility of casting for the production of the multiplicity of short runs usually required. Stamping can produce simple components at lower weights than casting, so that when gold prices escalated a few years ago, some manufacturers reverted to stamping to reduce product weight. However, such light gold jewellery did not greatly appeal to buyers and there was soon a change back to more substantial cast work. A major advantage of investment casting in jewellery manufacture is that complex forms with piercing and multiple undercuts can readily be produced without expensive hand making and assembly. The list of com ponents produced is virtually endless. They range from single stone settings to complete pieces with dozens of settings incorporated. It is possible to cast gold and silver alloys directly on to some precious stones without damaging them. Stones can be placed in rubber dies and wax patterns incorporating settings injected on to them. After casting the stones are firmly set, eliminating hand setting. This is done by making the rubber die from a pattern that already has the stones in place. Thus an impression of the stones is formed in the die and provided that calibrated stones are used, they can be accurately placed in the die before the wax is injected. Diamonds must be coated with a flux to prevent surface deterio ration during burnout, but some precious stones, particularly synthetics, can be directly cast-on without damage. An example of this practice is shown in Fig. 41. Apart from mainstream production, investment casting is also used in the making of one-off fine or art jewellery pieces. In fine jewellery, repeated motifs are often used to assemble complex pieces or suites of jewellery. By casting these motifs much repetition hand work is elimi-

422

Investment Casting

.-

i 1£ #

*£>

-

.5 V * C

1 «

*3 ^

1 Fig 41

Sterling silver rings cast directly on to cut , synthetic spinel stones (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

nated, though individual castings may be modified to give a hand-made appearance. In art jewellery, designers may wish to incorporate sculptural forms such as figures and animals or abstract shapes that would be difficult and expensive to make by hand in metal. Hand-made waxes may be cast as individual pieces, or an initial casting may be used as a master pattern to make a rubber die for multiple pieces. In recent years techniques for the electroforming of carat golds have become available, and for the production of light hollow pieces this is an alternative to casting. Hollow forms can be produced with much lower wall thickness than is possible by casting, considerably reducing material costs. Equipment and operating costs are however, very high, as contin uous computer control is needed to ensure maintenance of carat gold composition and colour. As a result electroforming is probably only a viable alternative to casting for large pieces, where reduced weight is desirable on wearability grounds rather than gold cost. Ironically, the most effective electroforming technique uses wax patterns made in rubber dies identical to those used in investment casting. Silversmithing Silversmiths had always used castings and the universal change to the investment process provided a great improvement in quality. Actual applications have little changed, although contemporary silver designers have exploited the enhanced capabilities of the process (see Fig. 42), as have modem jewellery designers. One application in silversmithing is in the modelling of objects such as buildings, aircraft, cars, boats, military items, birds and animals. Sometimes plastic model kits for such items are directly

jewellery Investment Casting 423

Fig 42 Typical silversmiths ' castings (Courtesy of Chris Walton, The Worshipful Com pany of Goldsmiths, London).

usable as expendable patterns. Animal and bird models are also frequently created in wax by specialist artists. Small objects of this type may be cast using the wax model itself as an expendable pattern. With larger, more expensive models, moulds may first be produced in a room temperature vulcanised (RTV) material from the wax, to enable expendable wax patterns to be reproduced in the numbers required while preserving the original. The one competitor to investment casting in silversmithing is electroforming, which does not require expensive computer control as does carat gold forming. The technique is older than the modern investment casting process but had fallen into disuse in the silverware industry and has only become re-established in parallel with the adoption of investment casting. Currently only fine silver, rather than the 92.5% silver 7.5% copper Sterling alloy, can be used, but here the cost difference is not a major factor as it is between fine and carat golds. For highly detailed, thin section, single face repeat or long motifs such as building friezes, or for applied decoration and hollow forms, electroforming is a viable proposition. In general, however, it offers no real threat to the position of investment casting in silversmithing. Alloys Unlike any other application of metals, the compositions of precious jew ellery and silverware alloys are constrained primarily by legal requirements relating to their precious metal contents, and secondly by aesthetic requirements, rather than by mechanical, physical or chemical properties. Gold alloys are usually designated by the carat system, 24 carats being 100% gold, or in some countries by the millesimal system, ie parts per thousand. The most common gold alloys are 8, 9, 10, 14, 18 and 22 carat. Silver alloys are usually expressed in millesimals, the most widely used

424

Investment Casting

containing 985 (Britannia), 925 (Sterling), 900, 850 and 800 parts per thousand. Platinum and palladium alloys are expressed in conventional percentages of the precious metal, usually 95% minimum. In most developed countries the precious metal contents are fixed by law and the products subject to official testing and marking. In the United Kingdom for example, the legal standards are Britannia and Sterling silver, 9, 14, 18 and 22 carat gold and 95% platinum. There is no UK standard for palladium alloys. Precious metal alloys used for casting are usually identical in composition with corresponding wrought alloys. With gold, colour match of wrought and cast parts used in the same piece is important and, as the traditional gold alloy compositions are readily castable, there has been little impetus to develop specialised casting alloys. Coloured golds are normally alloyed with silver and copper and sometimes zinc, particularly in the lower-carat alloys. Colour is controlled by the relative proportions of silver and copper. Silver-rich alloys have a greenish colour, whereas copper rich alloys are red. Zinc in the lower carat alloys enhances the richness of the yellows and acts as a deoxidiser. The one area where traditional coloured gold alloys were not com pletely satisfactory for investment casting was in 9 and 10 carat yellow golds. These, with gold contents of 37.5 and 41.6% respectively, are high in copper and, to obtain a reasonable yellow gold colour, normally contain around 7.5% zinc. With this type of composition zinc may be volatilised during melting, with consequent fuming and compositional changes. In the 1970s it was found that additions of up to 0.5% silicon inhibited fuming during melting. As a bonus, silicon-containing low-carat gold alloys exhibit a better yellow colour than their silicon-free equivalents. They are more subject to grain growth on overheating than the silicon free alloys, but have come into wide use for casting. White gold alloys contain up to 20% of palladium and/or nickel. The high-palladium alloys are rather soft and the high nickel alloys hard. Zinc in nickel-containing alloys reduces hardness and improves whiteness. Loss of zinc during melting of these relatively high melting point alloys restricts the re-use of scrap metal. Silicon is a highly deleterious impurity so cannot be used to overcome the problem as for the yellow alloys, and great care is needed if previously cast metal is re-used. Silver alloys for casting are normally alloyed solely with copper, and identical compositions are used both for casting and wrought production. With platinum and palladium only 5% of alloying additions is permissible. Some difficulties were met when the traditional compositions came to be used for investment casting and improved alloys have been de veloped. For platinum the 5% cobalt alloy has proved the most satisfactory for casting, as has the 5% nickel alloy for the notoriously difficult-tomelt palladium.

Jewellery Investment Casting

425

Pattern Making Master patterns The starting point for most jewellery and silverware castings is a master pattern, normally in metal, from which the wax injection die is made. The pattern may be an existing or prototype piece and it is unfortunate that the rubber die system makes the illegal copying of jewellery all too easy. Ideally the master pattern is made in precious metal, commonly silver, and is finished to the highest standard. If precise dimensions are required allowance for shrinkage must be made, although this is not normally a critical factor. Abrupt section changes, sharp angles and heavy sections isolated from the main body of the pattern are avoided where possible, but hot tears are an occasional problem in the closed rings that inevitably feature in jewellery. Complex textured and pierced surfaces are very well reproduced. For components having complex or delicate surfaces, consideration must be given at the design stage to the provision of areas for sprue attachment, so that patterned areas are not damaged in fettling. Ideally, sprues are tailored into the pattern to provide smooth metal flow and progressive solidification, but unfortunately they are often attached for convenience of cutting off rather than on feeding grounds, and shrinkage porosity is the most common defect found in jewellery castings. Whatever the form of the sprueing system on any individual pattern it is normally finished with an approximately 25 mm length of 3 to 5 mm rod. This provides support during vulcanisation of the rubber, and subsequently forms the wax injection channel in the die. Where applicable, master patterns are highly polished. If made in silver or other metal liable to attack by sulphur, they are preferably rhodium or chromium plated to avoid slight roughening during vulcanising. Corresponding roughening of the rubber surface otherwise resists wax pattern removal from the die. A typical example of a sprued master pattern is shown in Fig. 43.

Fig 43

A jewellery master pattern, sprued ready fo r rubber die making (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

426

Investment Casting

Rubber dies Once a master pattern is available, a rubber die can be made and in production in a matter of an hour or two. If modifications are subsequently required these can be carried out rapidly and at low cost. Rubber dies are made in simple rectangular aluminium frames that are sometimes split horizontally into two equal halves. Specially compounded raw sheet rubber is packed into one half of the mould frame. The pattern is positioned on the rubber, centrally in the frame, with the end of the injection sprue rod in a hole drilled in the centre of one narrow side. Depending on the form of the pattern, the rubber may be cut away to accommodate it and hollow areas may be roughly filled with pieces of rubber. In ideal practice, a shaped former corresponding to the wax injector nozzle is slipped on to the injection sprue and butted up to the end of the frame. The second half of the frame is then filled with rubber with a slight excess. The mould assembly is placed in a screw or hydraulic press with electrically heated, thermostatically controlled platens and pressure is gradually increased as the assembly heats up. When the vulcanising temperature of 150°C is reached, the press is closed down hard on the frame and left for 30 minutes to complete vulcanisation. The assembly is then removed from the press, and when cool the solid rubber block is removed from the frame and any flash trimmed off. Using a surgical scalpel, the block is cut into two parts, along the centre line of the narrow faces starting at the protruding sprue rod. During cutting the rubber is stretched and registration locks cut inside each corner. When the pattern is reached it is released by cutting along a predetermined parting line and wherever else required. Die cutting is a skilled operation, though quickly learnt, and most jewellery patterns can be released leaving a two-part die with a perfect negative impression. Occasionally it is necessary to leave a separate piece of rubber in a partly enclosed area of the pattern. This is cut with appropriate registration to enable it to be replaced accurately in the finished die. It is removed from the pattern, with appropriate further cutting, after the main die halves have been separated. Once the die is complete, trial wax injections are made. If necessary, cuts are made in the rubber to vent any areas where air entrapment occurs or to release the wax pattern without distortion or breakage. If the rubber is well stretched during cutting none of the cuts should show on the injected wax surface. Many refinements and variations to the technique have been developed over the years, to deal with ever more complex patterns and to simplify the production of routine dies, but the basic technique has remained as described. The versatility, low cost and speed

Jewellery Investment Casting

427

of production of rubber dies has been a major factor in the enormous growth of investment casting as a jewellery production process. Rubber dies have an excellent life and with plain patterns many thousands of waxes may be obtained from a single die. The most common cause of die failure is the breaking off of thin rubber sections in complex dies. However, it should be borne in mind that the cost of a rubber die is very low and given the availability of the master pattern a new die can be produced and in use in as little as one hour. RTV dies There is a limit to the size in which vulcanised rubber dies can conven iently be made and the process is unsuitable for pattern materials that could be damaged by the heat and pressure of the vulcanising process. Many silversmiths' patterns are too large for vulcanising and dies for these and delicate and temperature-sensitive patterns are made in RTV materials such as silicone or polysulphide rubbers. After catalytic setting these are cut in the same way as vulcanised rubber. Such dies have limited lives due to the much lower tear strength as compared with ordinary rubber, and are much more expensive. Wax pattern making Multiple wax pattern production is invariably by some form of air pressure injection, injectors ranging in capacity from about 500 ml to several litres. All have thermostatically controlled electric heating, and injection nozzles normally operate in the horizontal plane. Wax injection temperature is about 60°C, with an air pressure of up to 80kN/m2. For hand injection the die is firmly held between two flat plates and the injection opening is pressed against the spring loaded nozzle. A few seconds dwell is allowed, depending on the volume of the pattern, then the pressure is released and the die put aside for the wax to set. Operators work with several dies at once, these being filled and the patterns removed in sequence, (see Fig. 44) Mechanical and pneumatic clamps to hold dies for injection are available but experienced operators can work faster by hand and these devices are little used with simple air pressure injectors. Properly made rubber dies require little treatment to aid wax release. Normally an occasional light dusting with talc from a pouncing bag is all that is required. Silicone release fluids are sometimes used in difficult dies, but there is danger of build-up of silicone in the die, resulting in pattern defects. Automatic vacuum injectors that evacuate the die before injecting the wax are coming into increasing use in jewellery casting. These may be operated with hand presentation of the die, though the more sophisticated machines automatically clamp and present the die to the injection

428

Investment Casting

(a)

(b) Fig 44

(a) and (b) The removal of a wax pattern from a rubber die after injection (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

nozzle. Some models have facilities for automatic air pressure adjustment to suit individual dies. These machines are very expensive compared with simple air pressure injectors but reduce dependence on operator skill, give increased production rates and will fill delicate patterns not possible with simple injectors. Although most quantity production is based on the use of rubber or RTV dies, there are many applications in jewellery as well as in silversmithing for original casting patterns shaped directly in wax, whether for one-off pieces or for the production of the cast metal master patterns for diemaking. Hard waxes provided for this purpose can be sawn, filled, carved or even gently machined and are available in a variety of sections, including tubes with offset bores, in sizes suitable for making rings. Their use can give big savings in precious metal scrap, and of time in the production of one-off or prototype pieces. In another technique, flexible wax wires may be used to produce freehand decoration on sheet or carved wax surfaces, often as the first stage in the fabrication of a cast master pattern for subsequent quantity casting. Softened sheet wax can be used to form smooth convoluted shapes that

jewellery Investment Casting

429

are difficult or impossible to make directly in metal, but can be readily cast once a master has been produced. The range of finely detailed models that can be produced in wax is virtually unlimited and in this area modern investment casting is very close to the ancient process. To make large hollow castings, for example for animal model bodies, slush casting of wax into RTV dies made from solid models has also been successfully used (see Figs 45 and 46). In such cases it is usually necessary to insert core support wires of the alloy being cast into the interior space.

Fig 45

A sectioned hollow wax animal pattern made by slush casting in an R T V die.

One of the halves has been filled with investment to show the wall thickness (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

Fig 46 An 18 ct gold, one-piece hollow casting, average wall thickness 1mm. 'The Flying Horse of Kansu', a direct copy of the original (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

430

Investment Casting

Fig 47

A wax tree of ring patterns, ready fo r investing (Courtesy of Chris Walton, The Worshipful Company of Goldsmiths, London).

Pattern setting up Setting up for almost all production gold and silver jewellery casting is by the tree method, in which the patterns are attached radially to a vertical feeder. The length of the feeder, commonly about 150 mm, depends on production requirements and casting machine capacity. The individual patterns are wax welded to the feeder with a heated tool, with care to ensure a sound joint and smooth fillet (see Fig 47). Patterns are normally prepared for setting up with a single short sprue connected directly to the pattern, or to a sprue system when multiple feeding is employed. Ideally the wax injection sprue should be made with a suitable section to serve as the casting sprue. Patterns are attached to the main feeder at an angle of about 20° above the horizontal and disposed in a spiral arrangement to avoid, as far as possible, lines of heavy sections falling together. Whilst not metallurgically ideal, this method is the most practicable, having regard to ease of setting up and fettling, economy of flask space and maximum numbers of castings per mould. Ideally the central sprue should be tapered from the pouring end to the top to maximise progressive solidification, although few casters use this refinement. It involves more molten metal but fewer scrap castings so no extra cost is likely. It is impracticable to set up large silversmiths' and jewellery patterns on tree sprues and these are usually cast in small numbers direct from a heavy pouring basin. For platinum casting tree sprueing is sometimes used, but as casts are usually quite small direct sprueing from a flat topped or conical feeder that forms the pouring basin is to be preferred. It should be remembered that these alloys cost around £14 per gramme, have a relative density of about 20 and are cast at temperatures approaching 2000°C. Under these

Jewellery Investment Casting 431 conditions centrifugal casting with its rapid metal transfer is the most technically satisfactory method. However, in view of the very dense, high value metal and the greater likelihood of mould breakage it is wise to keep melt weights reasonably low. Because of the pressure casting techniques employed, venting is not used in moulds for precious metals, whilst risers, which would have to be blind, are rarely used either.

Mould Making Investments Modern plaster investments are based on mixtures of the allotropic forms of silica, cristobalite and quartz, generally bound with the alpha form of calcium sulphate hemihydrate, hydrocal, rather than the beta hemihydrate, plaster of Paris. The proportion of silica to bonding plaster is about 3 to 1. Investment powders are formulated by their manufacturers to give the required combinations of setting time, strength, expansion characteristics and fineness, allied to price. Chemical additions are made to the basic hydrocal/silica mixtures to control setting time, setting expansion and the viscosity of the mixed slurry, and to aid de-aeration. High temperature investments form only a tiny proportion of investments used in precious metal casting and will not be further discussed in this context. Mould making technique Pattern assemblies, whether tree-type or direct feeding, are normally mounted on circular rubber mouldings known as sprue bases. These incorporate a truncated conical or hemispherical central portion that forms the pouring basin in the finished mould. A raised rim around the edge of the sprue base forms a seal to the cylindrical stainless steel flask. Flask diameters range from about 75 to 250 mm, those used with tree sprueing being 100 or 150 mm. Length depends on the shape and number of pieces required and the physical capacity of the casting machine. Most centrifugal machines are limited to flask dimensions of up to 150 mm diameter by 200 mm in length. Many vacuum-assisted machines can accommodate much longer flasks. Flasks for centrifugal casting are plain lengths of tube. For some vacuum -assisted casting techniques, particularly with large-diameter moulds, flask walls are perforated with 10 to 15 mm diameter holes, with about 50 mm at each end left unperforated. Before investing, the holes are sealed, generally with gummed paper tape, to contain the investment slurry.

432

investment Casting

Investment mixing For optimum results in casting it is important that mixing and investing are carefully timed so that the whole process is completed just before the investment begins to set. Finishing too soon can result in rough surfaces on castings towards the bottom of the mould. Carrying out de-aeration once setting has commenced can result in entrapment of steam bubbles on pattern surfaces, causing large excrescences on castings. Moving moulds during setting can cause investment cracks, resulting in fins on castings. For small scale production investment may be hand mixed in a flexible bowl, but some form of power mixing is mostly used. The investment powder is weighed and the water addition measured, investments being formulated to commence setting at a fixed time controlled by the water temperature. The user can thus vary the time to suit the technique being employed. Industrial kitchen mixers are widely used, but for large scale mould making, purpose designed integrated machines are available, in which the whole mixing and mould filling process is carried out under vacuum. Investment de-aeration Bubbles formed from dissolved air in the mixing water can adhere to patterns during investing. This results in beads of extraneous metal on the surfaces of castings, unless steps are taken to de-aerate the investment slurry as the final stage in the investing process. Except when vacuum mixing machines are used de-aeration is carried out by subjecting the mixed investment, and then the filled flasks, to a vacuum sufficient to cause the water to boil. Large steam bubbles then sweep away any entrapped air. Dewaxing and firing Dewaxing of plaster bound investments is normally commenced about one hour after the investment has set. It is usually carried out in the burnout furnace as the first stage of the firing programme. This is not ideal as considerable fume is produced, causing environmental problems. Some burnout furnaces have facilities for draining the molten wax from the chamber before it reaches ignition temperature, whilst in other cases dewaxing is carried out in a separate low temperature oven. Much more satisfactory is steam dewaxing, in which moulds are heated over boiling water, the wax quickly melting out and collecting in the water. Steam dewaxing saturates the set investment with water, which prevents the molten wax from penetrating the investment, a potential cause of inferior surface finish. Steam dewaxing also reduces the min imum firing time necessary to complete the burnout of carbonaceous

jewellery Investment Casting

433

residues. Many jewellery casters, however, prefer to use a single stage dewax and burnout cycle. In quantity production this permits overnight firing and the completion of the active stages of the process, from casting one day's moulds to investing the next, in a normal working day. Burnout furnaces for jewellery casting are usually conventional gas or electrically fired units based on pottery kilns. Small rotary hearth furnaces promote even heating and are very convenient for locating and retrieving individual moulds but are expensive and little used. Most modern furnaces are fitted with mechanical or electronic programme controllers, permitting overnight burnout under automatic control. The normal maximum burnout temperature for plaster bound investments is around 730°C since at higher temperatures reaction between calcium sulphate and silica causes breakdown of the investment. After progressive heating, moulds are soaked at the burnout temperature for sufficient time to remove all carbonaceous residues. The temperature is then reduced to the level required for casting, which depends on the casting technique and the nature of the patterns. Statically cast moulds require a higher temperature than similar centrifugally cast moulds, but a more important factor is pattern section and shape. Moulds containing thin delicate patterns or requiring long metal runs are cast with mould temperatures of around 700°C, whereas for heavy section castings the level may be as low as 350°C. Casting into moulds with temperatures below 350°C can cause severe mould cracking due to thermal expansion effects, resulting from the temperature inversions that occur in both quartz and cristobalite between 250° and 350°C.

Melting for Casting Gold and silver alloys These present few problems in melting. Silver alloys are prone to oxygen pick-up that can result in internal oxidation, but simple melting atmosphere control readily prevents this. In torch melting silver can be kept free of oxidation by correct flame control, whilst in furnace melting reducing or neutral atmospheres ensure oxide free melts. The copper content of jewellery silver alloys ensures that any oxygen absorbed forms copper oxide, and gas porosity is very rare in silver castings. Coloured gold alloys are even more forgiving than silver, and provided that the melt surface is kept bright by fluxing or atmosphere control, clean castings are easily obtained. White gold alloys do sometimes give difficulties. Except for low-carat, silver-whitened alloys, white golds contain nickel and/or palladium and melt at higher temperatures than the coloured alloys. Under reducing conditions and with their higher melting points they are liable to pick up

434

Investment Casting

silicon from the melting crucibles, risking hot shortness and cracked castings. The white golds, particularly those containing palladium, should therefore be melted under neutral or slightly oxidising conditions. Platinum rich alloys Apart from its high casting temperature, approaching 2000°C, which requires the use of high temperature investments, platinum is probably the easiest metal to cast. Molten platinum dissolves neither oxygen, nitrogen nor hydrogen and has no oxide, although it can promote the reduction of deleterious impurities from refractories when melted under reducing conditions and will rapidly pick up carbon if melted in contact with carbonaceous materials. However, when melted in air in oxide-type refractories, the metal remains clean without fluxes or protective atmospheres and its high density assists complete mould filling. The high casting temperature, with relatively small melts, requires rapid metal transfer to the mould, favouring torch melting with a strongly oxidising flame, with casting in a spring driven centrifugal machine rotating on a horizontal axis. These machines, often confusingly described as Vertical', are driven by coil springs as opposed to the clock springs in the more common vertical axis machines. This type of drive gives a much faster take-off than any other type of centrifugal machine, ensuring very rapid metal transfer. Many casters do use vertical axis casting machines, with induction melting, for platinum, but the relatively slow metal transfer requires a higher superheat than for horizontal axis machines. This, allied with the difficulty of accurate temperature measurement, can lead to overheating, resulting in poor crucible life, poor surface finish and danger of metal breakout, which can be catastrophic at around 2000°C. Naturally with torch melting there cannot be any instrumental tem perature control, but experience has shown that a skilled platinum melter can readily judge the correct casting temperature in the shallow hearths used. Judgement by eye is much more unreliable with induction melts in conventional crucibles, where the surface temperature may not be representative of the body of metal. Torch melting For small production with simple centrifugal casting machines, natural gas or propane/compressed air torches are used for melting gold and silver alloys. A purpose-designed torch for melting with natural gas and compressed air has proved considerably more efficient than the town's gas torches previously used. Early problems with natural gas had caused many casters to change to oxygen -propane torches, but whilst these had adequate melting power,

Jewellery Investment Casting 435 the increased flame temperature brought the danger of overheating. They did, however, prove very satisfactory as replacements for the oxygentown's gas torches formerly used for platinum. Furnace melting Small, natural gas/low pressure air melting crucible furnaces are very satisfactory for the separate melting of gold and silver alloys for static vacuum -assisted casting, particularly for large moulds. For smaller scale work resistance furnaces are sometimes used but are slow in melting compared to modem gas fired furnaces. Most integrated casting machines use medium frequency induction melting, with automatic melt temperature control in the more sophisticated machines. Power ratings range between 3 and 18kVA, giving melting capacities of up to around 7 kg of 18ct gold. Casting Centrifugal casting machines Spring driven machines derived from dental equipment, used with torch melting, are still employed in small scale casting operations and, as discussed previously, are particularly satisfactory for the casting of high melting point alloys. For quantity production, power driven machines are widely used with torch melting and invariably with induction melting. These rotate at speeds of up to 300 rpm and most have variable speed or torque controls. Power driven machines do not have such a high take-off speed as spring driven machines but as melts are generally larger this is less important. Power drives maintain rotation until the metal has com pletely solidified, eliminating the danger of run-back that can occur with large melts on spring driven machines. Only one type of centrifugal casting machine employing resistance melting is now in wide use, this having been scaled up from a dental machine. It employs a carbon resistance furnace and has a unique casting system using Durville pouring in conjunction with centrifugal force to ensure complete mould filling (Figure 48). Although this has proved popular in some countries, particularly Germany, centrifugal machines with induction melting have been preferred in the large Italian and American jewellery industries. Despite the need to retract the melting coil before rotation, and the inefficiency of the metal transfer method, these have greater flexibility, faster melting and larger available capacity. Such machines are now available from manufacturers in the UK, Italy, France, Germany, Spain and the USA. All are basically similar, varying in capacity, degree of automation and method of melt temperature indication, if any.

436

Investment Casting

Mould clamp

Turntable

To drive motor Fig 48

Sectional illustration of the principle of the Linder roll-over, centrifugal casting machine with carbon resistance melting.

To overcome the need for retraction of thermocouples from the melt before casting, some manufacturers fit radiation pyrometers. These have not always proved satisfactory in jewellery workshop conditions, so that thermocouples are often still preferred. Many jewellery casters, however, still depend more on visual judgment of melt temperatures than on instruments. Following growing penetration of sophisticated static vacuum-assisted casting machines incorporating vacuum melting, some manufacturers are

Jewellery Investment Casting

437

again producing centrifugal machines with the melting and casting facilities enclosed in a vacuum chamber. The disadvantages of the centrifugal machine remain and, as many jewellery alloys contain zinc and cannot be melted in a vacuum, this facility is frequently redundant. Such a feature does enable all air to be removed from the melting chamber, to be replaced by a neutral atmosphere, but the advantage gained is probably illusory for most gold and silver alloys and warrants neither the slower procedure nor the high cost. Static casting machines Simple vacuum -assisted casting machines, as described earlier, are still used with separate furnace melting, particularly in non-precious metal jewellery casting and small scale operations. Machines with enclosed casting chambers are more popular for large scale work. Some types permit the use of plain rather than flanged and perforated flasks, which are expensive and awkward to handle in the larger sizes. Small-scale machines of this type are integrated into complete casting centres for small workshops. Investment mixing facilities, burnout furnace and casting head are built into a single casing, using the same vacuum pump for investment de-aeration and casting. The major development in recent years has been in the area of automatic, fully integrated vacuum -assisted machines. The seminal machine of this type was the German Tnresa' which appeared in the seventies. This has now developed into a range of highly sophisticated machines, including models that employ a single induction generator for melting for investment casting and for the continuous casting of strip, rod and tube. Besides these induction melting machines some manufacturers have produced modest capacity machines on the same principle, but with resistance melting. Some have discarded full vacuum melting in favour of open melting with bottom pouring into evacuated moulds. Another approach to vacuum melting and casting has been to use induction melting with conventional pouring in a closed chamber which can be evacuated if required. With either vacuum or controlled atmosphere melting, vacuum -assisted mould filling is used. The first com mercial machines of this type appeared in Denmark in the 1970s but did not penetrate the international market. In the late 1980s two major Italian manufacturers introduced machines on a similar principle. The Di Maio machine employs resistance melting with automatic pouring by power tilting the crucible when preset conditions of temperature and atmosphere are attained. The Galloni machine has induction melting and also pours automatically; before pouring, the mould, initially in the vertical position, is automatically rotated so that its axis is at 90° to the crucible, with the entry cone close to its lip. The whole assembly is then power

438

Investment Casting

tilted to give smooth semi-Durville pouring. With both machines, atmospheric pressure is admitted to the chamber at the moment of pouring while vacuum is maintained on the mould body. One of these recent machines and some associated castings are shown in Figs 49 and 50. Cleaning and Finishing of Castings After casting it is usual to plunge moulds into water when the visible metal surface has cooled to black heat. This disintegrates the bulk of investment but leaves some adhering to the cast surface. The most effective method of removing this is by high pressure water jet or glass bead blasting, leaving the metal with an oxide-free, fine satin finish. Individual castings are then cut from their sprues by hand or power shears or saw ing, leaving a minimum of metal to be removed by filing or grinding to restore contours. For fine work, castings are finished by normal jewellery or silver polishing techniques. Sometimes they may be considerably worked by chas-

Fig 49 The Galloni 'Robocnstroll-over, automatic, vacuum assisted casting machine. (Reproduced by courtesy of Aseg Cationi S.p.A. San Colombano al Lambra, Milan.)

Jewellery Investment Casting 439

Fig 50 A selection of castings in gold and silver alloys made in a Galloni 'Robocast' machine. (Reproduced by courtesy of Aseg Galloni S.p.A., San Colombano al Lambra, Milan.) ing or other means to sharpen detail or give the appearance of hand making. Barrel polishing is widely used for finishing jewellery castings for general production, and modern multi-stage techniques have eliminated much hand work. Future Trends Investment casting is now fully established as the major production process for quantity jewellery production, and as an important tool in fine and art jewellery making and silversmithing, and there is no sign of any alternative process which could displace it. No doubt the suppliers of equipment and consumables for the process will continue to introduce improvements and innovations to their product ranges but it would seem unlikely that these will be as rapid or as radical as those of the last forty years.

440

Investment Casting

Computer aided design and machining CAD/CAM has been in use for some years in the highly specialised manufacture of American school rings. Wax or plastics injection dies are made in metal on computer controlled die sinking machines, to designs drawn from computer stored libraries of the various customary symbols. The silversmith Stuart Devlin has demonstrated how jewellery and silverware can be designed with the aid of advanced computer graphics.13 The digital design data obtained could be used for the automatic production of complex casting patterns for direct investing and casting using stereo-lithography. It is considered feasible to produce complete trees of multiple patterns by this technique, so eliminating master patterns, dies, wax injection and setting up. Prohibitively expensive equipment would at present be required, but the rapid pace of computer development and applications suggests that such techniques may well be available to manufacturing jewellers and silversmiths within a few years.

12.4 Investment Casting in Surgery and Dentistry M.F. LECLERC

SURGICAL IMPLANT INVESTMENT CASTINGS Historical Background The repair and restoration of function of the skeletal system affected by injury, disease or a congenital defect can often be facilitated by the use of implants made of non -living materials. It is difficult to determine when foreign materials were first buried in the body. Archaeological evidence clearly indicates that surgical procedures were performed in several ancient civilisations. Most cases involved the use of a noble metal to repair a localised structural defect. For example, in 1546, Ambroise Paré described the use of gold plates to repair traumatic defects in the skull and gold wire to repair abdominal hernias. By the early 19th century, many different metals including iron, gold, silver, lead, bronze, steel and platinum had been employed, usually in the form of wires or pins, to treat bone fractures. The incidence of infection due to surgical procedures was very high, leading to generally poor results. Progress in surgery was slow and mixed liberally with superstition until the latter part of the nineteenth century. Pasteur's and Lister's asep tic surgical techniques, developed around 1883, and shortly thereafter Roentgen's discovery of X-rays in 1895, added a new dimension to orthopaedic surgery. As the occurrence of infection was brought under control, the relationship between material properties and the success of implant surgery became more clearly apparent. Tissue compatibility, corrosion, fatigue and wear resistance, together with tensile strength, were identified as the critical characteristics. The noble metals, gold and silver, met the first two criteria but lacked strength for applications involving high stress. Metals such as brass, copper and steel had adequate strength DOI: 10.1201/9781003419228-16

442

Investment Casting

for many applications but exhibited poor tissue compatibility and corro sion resistance. In the beginning of the twentieth century, surgical techniques were developed for the fixation of bone fractures with a plate and screw combination. Sherman-type bone plates were fabricated from steels containing such alloying elements as chromium and vanadium. By the 1920's, the use of these steels became questionable because of poor tissue compatibility. At that time, however, no other material was available combining the necessary strength with adequate corrosion resistance for the exacting conditions. In 1926, when Strauss patented the 18-8 SMo stainless steel, with 2 to 4% molybdenum and a reduced carbon content of 0.08%, a material was created which promised improved resistance to acid and chloride containing environments. This had far superior corrosion resistant properties to anything that had been available up to that time and immediately attracted the interest of orthopaedic surgeons. Bone plates, screws and other fixation devices were fabricated and used as surgical implants. This material formed the basis for the Type 316L alloy in common use today. The Co-Cr-Mo casting alloy most commonly used in producing investment cast surgical implants was first employed by C.S. Venable and W.G. Stuck in 1936 for dental implants. After the alloy had proved to be excep tionally corrosion resistant and compatible with the bodily environment, it was used as early as 1939 by M.N. Smith Peterson for the manufacture of cast hip joint cups. Because of its exceptional abrasion resistance the Co-Cr-Mo casting alloy is still being used in the manufacture of investment cast prostheses.14"16 When titanium became commercially developed in the late 1940's, it was very soon evaluated as a surgical implant material. The metal possessed a good combination of mechanical and corrosion resistant properties and also demonstrated outstanding tissue compatibility. Although a few internal fixation devices were also used in the United States in the 1950's and 1960's, the most extensive clinical use of titanium was in Great Britain, the material being generally used in its wrought condition. Interest in the Ti-6A1-4V alloy and Extra Low Interstitial (ELI) versions of this alloy for surgical implants surged in the United States in the late 1970's. This alloy now finds wide application in its forged and wrought conditions in orthopaedic surgery. During the 1980's a high nitrogen austenitic stainless steel was intro duced into the orthopaedic implant market place. This exhibited extremely good mechanical properties and has excellent biocompatibility and corrosion resistant properties. Although the alloy is commonly used in the forged and cold worked conditions, several investment casting houses have begun to use it for investment cast surgical implants, with varying success.

Investment Casting in Surgery and Dentistry

443

With the availability from 1945 onwards of apparently biologically acceptable polymers and improved metallic implant materials, coupled with advances in other branches of medicine such as antibiotics and anaesthesia, a new era dawned. Almost all joint replacement bearing systems now consist of metal and plastic combinations and are often cemented into position using polymethylmethacrylate bone cement. Although the first plastic material tried, ie. PTFE, failed in service because it wore very badly in the body,17 the high density and now the ultra high molecular weight polyethylenes seem to meet all the requirements for implant materials and are currently giving excellent service in man. Figure 51 shows a radiograph of a patient's leg which has been subjected to Total Knee Joint replacement surgery. During the 1980's both titanium alloys and further new types of cobalt based alloy have been used to produce surgical implants. Although these are mostly based on wrought bar stock or forgings, several orthopaedic companies are currently pursuing the investment casting of these alloys for implants.

Basic Requirements of Surgical Implants Before considering metallurgical aspects mention should be made of the exacting special requirements of implants. The primary problem lies in human variability. Although people are all broadly similar, each individ ual is unique in body chemistry, physical dimensions and responses. For an implant to be accepted in the body and fulfill its intended function it needs to be the correct size, shape and design and to have been placed in the patient using good surgical techniques. The materials used must be non-toxic, biocompatible, tissue compatible, should have high yield, fatigue and torsional fatigue strengths, and should also exhibit good corrosion, wear and fretting resistance.

Alloys Used in Modem Surgical Implant Production The main alloys now used in the manufacture of orthopaedic implants are molybdenum bearing austenitic stainless steels, cobalt-chromium molybdenum casting alloys, and titanium and its alloys. Stainless steel and titanium alloy implants are frequently produced as forgings or machined from wrought bar. Over recent years, however, many manufacturers have experimented with their production as investment castings, with varying success, since it proved difficult to achieve consistent composition, microstructure and soundness. Cobalt based alloys are finding wide application, in some cases as castings. These are often subjected to post-casting heat treatments for

444

Investment Casting

Fig 51 An X-ray of a typical total knee joint replacement. The upper (femoral) and lower (tibial) components have been produced from heat treated Co-Cr-Mo alloy investment castings and are separated by an ultra high molecular weight polyethylene bearing surface which has been moulded into the tibial component. The patella has been resurfaced using a wrought Ti-6Al~4V alloy metal-backed, ultra high molecular weight, polyethylene component. The plastic components in this radiograph are radio-translucent and the metallic components are radio-opaque. homogenization and to improve ductility. Some are also subjected to hot isostatic pressing to reduce microporosity and improve mechanical prop erties. Other cobalt based alloys are supplied in the hot and cold worked conditions. The chemical compositions of the basic alloy types used to produce modern orthopaedic implants are shown in Table 2.18~22 Their mechanical

Investment Casting in Surgery and Dentistry

445

Table 2. The compositions of the basic alloys used for surgical implant production

Alloy type -> Composition i (wt %) Carbon Manganese Phosphorus Sulphur Silicon Tungsten Cobalt Chromium Nickel Molybdenum Iron Aluminium Vanadium Titanium Nitrogen Copper Hydrogen Oxygen

Stainless steel18 to BS 7252: Parti (wrought)

Co -Cr- Mo19

0.03 max 2.0 max 0.025 max 0.010 max 1.0 max 17.0-19.0 13.0-15.0 2.25-3.5 Balance

0.35 max 1.0 max 1.0 max Balance 26.5-30.0 2.5 max 4.5-7.0 1.0 max

-

-

0.10 max 0.50 max -

-

to BS 7252: Part 4 (as cast)

-

-

-

-

-

Unalloyed titanium20 to BS 7252: Part 2 (wrought) 0.10 max -

TÌ-6AI-4V21 TÌ-6AI-4V22 to BS 7252: Part 3 (wrought) 0.08 max -

-

-

-

-

0.20 max -

-

Balance 0.03 max 0.015 max 0.18 max

0.30 max 5.50-6.75 3.50-4.50 Balance 0.05 max 0.015 max 0.20 max

to ASTM F1108 (cast and hipped) 0.10 max -

0.20 max 5.50-6.75 3.50-4.50 Balance 0.05 max 0.015 max 0.20 max

properties, together with those of human bone and some of the other materials finding use in orthopaedics, are shown in Table 3. These data confirm the suitability of investment cast alloys for the requirements of surgical implant manufacturers. Investment Casting Techniques Most orthopaedic implants are destined to be used for either accident surgery, including bone plates, screws, pins and wires, or for cold surgery, as in hip, knee, shoulder, elbow, wrist and ankle joint replacements. The manufacturer often prefers the investment casting route when the complex three-dimensional geometry frequently required in joint replacements makes fabrication, forging or conventional machining uneconomic. The castings require a homogeneous microstructure having a high degree of microcleanness and very low microporosity, as well as the appropriate composition, mechanical properties and dimensional accuracy. British Standard BS 7252: Part 4 (Dual numbered with the International Standard ISO 5832/IV)19 and British Standard 7254: Part 5,23 define the chemical and mechanical requirements for the production of Co-Cr-Mo

Type and condition

Typical standard for implant application

Ultimate tensile strength (M Pa)

0.2% tensile yield stress (M Pa)

Stainless steel

316 L annealed bar

490-690

190

200

40

-

183

245-300

Stainless steel

316 L cold worked bar

505-605

195-295

200

35

-

320

300

High nitrogen stainless steel High nitrogen stainless steel Cobalt-chromium molybdenum alloy Cobalt-chromium molybdenum alloy Cobalt-chromium tungsten-nickel alloy Cobalt-nickel-chromium molybdenum alloy Titanium, commercially pure

Annealed bar Cold worked prosthesis As cast

740 1150 700

430 810 500

200 200 200

30 15 8

-

269 365 300

460 640 235-340

665

445

200

10

-

300

235-340

860

310

230

30

-

450

480

1000

650

230

20

-

350

400-450

240

170

127

24

-

240

250-280

Titanium alloy

Ti-6Al-4V annealed bar

860

780

111

10

-

350

400-440

Titanium alloy

Ti-6AI-4V investment cast and hipped A t 03

BS 7252 Type D; ISO 5832/1; ASTM F138 BS 7252 Type D; ISO 5832/1; ASTM F138 BS 7252/9; ISO 5832/9 BS 7252/9; ISO 5832/9 BS 7252/4; ISO 5832/4; ASTM F75 BS 7252/4; ISO 5832/4; ASTM F75 BS 7252/5; ISO 5832/5; ASTM F90 BS 7252/6; ISO 5832/6; ASTM F562 BS 7252/2 Grade 1; ISO 5832/2; ASTM F67 BS 7252/3 Type TA1; ISO 5832/3; ASTM F136 ASTM F1108

860

758

111

8

-

350

420

BS 7253/2; ISO 6474; ASTM F603 BS 7253/5; ISO 5834/2; ASTM F648 BS 7253/1; ISO F451; ISO 5833/1 -

270

N/A

380

0

4000

2000

N/A

35

21

0.5

350

20

-

N/A

25

N/A

2

5

70

-