Although acid-base cements have been known since the mid 19th century, and have a wide variety of applications, there ha...
164 downloads
716 Views
14MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Although acid-base cements have been known since the mid 19th century, and have a wide variety of applications, there has been a failure to recognize them as constituting a single, well-defined class of material. This book remedies the situation by unifying the subject and treating this range of materials as a single class. These cements are defined as materials that are formed by mixing a basic powder with an acidic liquid, and offer an alternative to polymerization as a method for forming solid substances. They are quick-setting materials, with unusual properties, which find diverse applications as biomaterials and in industry.
Chemistry of Solid State Materials Acid-base cements Their biomedical and industrial applications
Chemistry of Solid State Materials Series Editors A. R. West, Department of Chemistry, University of Aberdeen H. Baxter, formerly at the Laboratory of the Government Chemist, London 1 Segal: Chemical synthesis of advanced ceramic materials 2 Colomban: Proton conductors 3 Wilson & Nicholson: Acid-base cements
Acid-base cements Their biomedical and industrial applications Alan D. Wilson formerly Head, Materials Technology, Laboratory of the Government Chemist Senior Research Fellow, Eastman Dental Hospital
John W. Nicholson Head, Materials Research, Laboratory of the Government Chemist
m CAMBRIDGE
0
UNIVERSITY PRESS
CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 2RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521372220 © Cambridge University Press 1993 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1993 This digitally printed first paperback version 2005 A catalogue recordfor this publication is available from the British Library Library of Congress Cataloguing in Publication data Wilson, Alan D. Acid—base cements: their biomedical and industrial applications /Alan D. Wilson, John W. Nicholson p. cm. - (Chemistry of solid state materials; 3) Includes bibliographical references and index. ISBN 0-521-37222-4 1. Adhesives. 2. Dental cements. I. Nicholson, John W. II. Title. III. Series. TP968.W54 1993 620.1'35-dc20 91-38946 CIP ISBN-13 978-0-521-37222-0 hardback ISBN-10 0-521-37222-4 hardback ISBN-13 978-0-521-67549-9 paperback ISBN-10 0-521-67549-9 paperback
Dedicated to the past and present members of the Materials Technology Group at the Laboratory of the Government Chemist
Contents
1
Preface Acknowledgements Introduction References
xvii xix 1 4
2
Theory of acid-base cements 2.1 General 2.2 The formation of cements 2.2.1 Classification 2.2.2 Requirements for cementitious bonding 2.2.3 Gelation 2.3 Acid-base concepts 2.3.1 General 2.3.2 History of acid-base concepts 2.3.3 Acid-base concepts in AB cement chemistry 2.3.4 Relevance of acid-base theories to AB cements 2.3.5 Acid-base strength 2.3.6 Acid-base classification 2.3.7 Hard and soft acids and bases (HSAB) References
5 5 7 7 8 10 12 12 12 14 19 20 22 24 26
3
Water 3.1 3.1.1 3.1.2 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.3.3 3.4
30 30 30 30 31 31 33 34 34 35 36 40
and acid-base cements Introduction Water as a solvent Water as a component Water Constitution Water compared with other hydrides The structure of water The concept of structure in the liquid state The structures of ice Liquid water Water as a solvent
IX
Contents 3.4.1 Hydrophobic interactions 3.4.2 Dissolution of salts 3.4.3 Ion-ion interactions in water 3.4.4 Dissolution of polymers 3.5 Hydration in the solid state 3.5.1 Coordination of water to ions 3.6 The role of water in acid-base cements 3.6.1 Water as a solvent in AB cements 3.6.2 Water as a component of AB cements 3.6.3 Water as plasticizer References
40 41 44 45 47 47 48 48 48 51 52
4
Polyelectrolytes, ion binding and gelation 4.1 Polyelectrolytes 4.1.1 General 4.1.2 Polyion conformation 4.2 Ion binding 4.2.1 Counterion binding 4.2.2 The distribution of counterions 4.2.3 Counterion condensation 4.2.4 Effect of valence and size on counterion binding 4.2.5 Site binding - general considerations 4.2.6 Effect of complex formation 4.2.7 Effect of the polymer characteristics on ion binding 4.2.8 Solvation (hydration) effects 4.2.9 Hydration of the polyion 4.2.10 Hydration and ion binding 4.2.11 Desolvation and precipitation 4.2.12 Conformational changes in polyions 4.2.13 Interactions between polyions 4.2.14 Polyion extensions, interactions and precipitation 4.3 Gelation References
56 56 56 58 59 59 59 63 65 67 69 70 72 73 76 77 79 82 82 83 85
5
Polyalkenoate cements 5.1 Introduction 5.2 Adhesion 5.2.1 New attitudes 5.2.2 The need for adhesive materials 5.2.3 Acid-etching 5.2.4 Obstacles to adhesion 5.2.5 The nature of the adhesion of polyalkenoates to tooth material 5.3 Preparation of poly(alkenoic acid)s 5.4 Setting reactions
90 90 92 92 92 93 93 94 97 98
Contents 5.5 Molecular structures 5.6 Metal oxide polyelectrolyte cements 5.7 Zinc polycarboxylate cement 5.7.1 Historical 5.7.2 Composition 5.7.3 Setting and structure 5.7.4 Properties 5.7.5 Modified materials 5.7.6 Conclusions 5.8 Mineral ionomer cements 5.9 Glass polyalkenoate (glass-ionomer) cement 5.9.1 Introduction 5.9.2 Glasses 5.9.3 Poly(alkenoic acid)s 5.9.4 Reaction-controlling additives 5.9.5 Setting 5.9.6 Structure 5.9.7 General characteristics 5.9.8 Physical properties 5.9.9 Adhesion 5.9.10 Erosion, ion release and water absorption 5.9.11 Biocompatibility 5.9.12 Modified and improved materials 5.9.13 Applications 5.10 Resin glass polyalkenoate cements 5.10.1 General 5.10.2 Class I hybrids 5.10.3 Class II hybrids 5.10.4 Properties References
99 101 103 103 103 104 106 112 113 113 116 116 117 131 133 134 143 146 147 152 156 159 162 166 169 169 170 171 173 175
Phosphate bonded cements
197
6.1 6.1.1 6.1.2 6.1.3
197 197 198
6.1.4 6.1.5 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5
General Orthophosphoric acid solutions Cations in phosphoric acid solutions Reactions between oxides and phosphoric acid solutions Effect of cations in phosphoric acid solutions Important cement-formers Zinc phosphate cement General History Composition Cement-forming reaction Structure
201 203 204 204 204 204 205 207 212 XI
Contents 6.2.6 6.2.7 6.2.8 6.2.9 6.2.10 6.3 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.4.5 6.4.6
xn
Properties Factors affecting properties Biological effects Modified zinc phosphate cements Hydrophosphate cements Transition-metal phosphate cements Magnesium phosphate cements General Composition Types Cement formation and properties Cement formation with phosphoric acid Cement formation with ammonium dihydrogen phosphate 6.4.7 Cement formation with diammonium hydrogen phosphate 6.4.8 Cement formation with ammonium polyphosphate 6.4.9 Cement formation with aluminium acid phosphate 6.4.10 Cements formed from magnesium titanates 6.5 Dental silicate cement 6.5.1 Historical 6.5.2 Glasses 6.5.3 Liquid 6.5.4 Cement-forming reaction 6.5.5 Structure 6.5.6 Physical properties 6.5.7 Dissolution and ion release 6.5.8 Biological aspects 6.5.9 Conclusions 6.5.10 Modified materials 6.6 Silicophosphate cement 6.7 Mineral phosphate cements References
214 218 219 219 220 220 222 222 222 222 223 223
Oxysalt bonded cements 7.1 Introduction 7.1.1 Components of oxysalt bonded cements 7.1.2 Setting of oxysalt bonded cements 7.2 Zinc oxychloride cements 7.2.1 History 7.2.2 Recent studies 7.3 Magnesium oxy chloride cements 7.3.1 Uses 7.3.2 Calcination of oxide 7.3.3 Setting chemistry
283 283 284 284 285 285 286 290 290 290 291
223 231 232 232 235 235 235 237 241 243 249 253 255 260 261 262 263 265 265
Contents
8
9
7.3.4 Kinetics of cementation 7.3.5 Phase relationships in the MgO-MgCl2-H2O system 7.3.6 Consequences for practical magnesium oxychloride cements 7.3.7 Impregnation with sulphur 7.4 Magnesium oxy sulphate cements 7.4.1 Setting chemistry 7.4.2 Phase relationships in the MgO-MgSO4-H2O system 7.4.3 Mechanical properties of magnesium oxysulphate cements 7.5 Other oxy salt bonded cements References
293 294
Miscellaneous aqueous cements 8.1 General 8.2 Miscellaneous aluminosilicate glass cements 8.3 Phytic acid cements 8.4 Poly(vinylphosphonic acid) cements 8.4.1 Metal oxide polyphosphonate cements 8.4.2 Glass polyphosphonate cements 8.5 Miscellaneous copper oxide and cobalt hydroxide cements References
307 307 307 309 310 311 314
Non-aqueous cements
318
9.1 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.2.5 9.2.6 9.2.7 9.2.8 9.2.9 9.2.10 9.2.11 9.3 9.3.1 9.3.2 9.4 9.4.1 9.4.2 9.4.3
General Zinc oxide eugenol (ZOE) cements Introduction and history Eugenol Zinc oxide Cement formation Setting Structure Physical properties Biological properties Modified cements Impression pastes Conclusions Improved ZOE cements General Reinforced cements 2-ethoxybenzoic acid eugenol (EBA) cements General Development Setting and structure
295 297 299 299 300 302 304 305
315 316
318 320 320 321 321 322 323 331 333 334 334 335 335 336 336 336 337 337 337 339 Xlll
Contents
10
xiv
9.4.4 Properties 9.5 EBA-methoxyhydroxybenzoate cements 9.5.1 EBA-vanillate and EBA-syringate cements 9.5.2 EBA-divanillate and polymerized vanillate cements 9.5.3 EBA-HV polymer cements 9.5.4 Conclusions 9.5.5 Other zinc oxide cements 9.6 Calcium hydroxide chelate cements 9.6.1 Introduction 9.6.2 Composition 9.6.3 Setting 9.6.4 Physical properties 9.6.5 Biological properties 9.6.6 The calcium hydroxide dimer cement References
340 342 342 344 345 346 347 347 347 348 348 350 350 351 352
Experimental techniques for the study of acid-base cements 10.1 Introduction 10.2 Chemical methods 10.2.1 Studies of cement formation 10.2.2 Degradative studies 10.3 Infrared spectroscopic analysis 10.3.1 Basic principles 10.3.2 Applications to AB cements 10.3.3 Fourier transform infrared spectroscopy 10.4 Nuclear magnetic resonance spectroscopy 10.4.1 Basic principles 10.4.2 Applications to AB cements 10.5 Electrical methods 10.6 X-ray diffraction 10.6.1 Basic principles 10.6.2 Applications to AB cements 10.7 Electron probe microanalysis 10.7.1 Basic principles 10.7.2 Applications to dental silicate cements 10.7.3 Applications to glass-ionomer cements 10.8 Measurement of mechanical properties 10.8.1 Compressive strength 10.8.2 Diametral compressive strength 10.8.3 Flexural strength 10.8.4 Fracture toughness 10.9 Setting and rheological properties 10.9.1 Problems of measurement 10.9.2 Methods of measurement
359 359 360 360 361 361 361 362 364 364 364 365 366 367 367 368 369 369 369 369 370 371 372 372 373 374 375 375
Contents 10.10 Erosion and leaching 10.10.1 Importance in dentistry 10.10.2 Studies of erosion 10.11 Optical properties 10.11.1 Importance in dentistry 10.11.2 Measurement of opacity 10.12 Temperature measurement 10.13 Other test methods References
Index
378 378 379 379 379 380 380 381 382
386
xv
Preface
The senior author first became interested in acid-base cements in 1964 when he undertook to examine the deficiencies of the dental silicate cement with a view to improving performance. At that time there was much concern by both dental surgeon and patient at the failure of this aesthetic material which was used to restore front teeth. Indeed, at the time, one correspondent commenting on this problem to a newspaper remarked that although mankind had solved the problem of nuclear energy the same could not be said of the restoration of front teeth. At the time it was supposed that the dental silicate cement was, as its name implied, a silicate cement which set by the formation of silica gel. Structural studies at the Laboratory of the Government Chemist (LGC) soon proved that this view was incorrect and that the cement set by formation of an amorphous aluminium phosphate salt. Thus we became aware of and intrigued by a class of materials that set by an acid-base reaction. It appeared that there was endless scope for the formulation of novel materials based on this concept. And so it proved. Over the years, from 1964 to date, a team at the LGC, with its expertise in Materials Chemistry, has studied many of the materials described in this book, elucidating structures, setting reactions and behaviour. This experience has formed a strong experimental background against which the book was written. In addition we have maintained contact with leaders in this field throughout the world. We should mention Professor Dennis Smith of Toronto University, who amongst his many achievements invented the adhesive zinc polycarboxylate cement (Chapter 5); Dr G. M. Brauer, who was for many years at the Institute for Materials Research, National Bureau of Standards, Washington, D.C., and is the acknowledged authority on cements formed by the reaction between zinc oxide and phenolic bodies (Chapter 9); and Dr J. H. Sharp of the University of Sheffield, who has developed magnesium phosphate cements (Chapter 6). xvu
Preface
In particular we thank Dr J. H. Sharp for supplying original photographs for use in the section on magnesium phosphate cements and for critically reading the draft manuscript and making constructive suggestions. On clinical matters we have benefited from a 20-year collaboration with Dr J. W. McLean OBE. Our own research at the LGC, while not confined to, has centred on, cements formed by the reactions between acid-decomposable glasses and various cement-forming acids (Chapters 5, 6, 8, 9). One of these materials invented at the LGC, the glass polyalkenoate or glass-ionomer cement, has proved of immense importance. Indeed, so successful has this material been in general dentistry, that the Materials Technology Group earned the Queen's Award for Technology in 1988. This material illustrates the useful combination of properties that can be found in the acid-base cements, for it has the aesthetic appearance of porcelain, the ability to adhere to teeth, and also the ability to releasefluoridewith its beneficial effect of reducing caries. We hope that this work will encourage, stimulate and assist others choosing to work in this interesting field. Alan D. Wilson John W. Nicholson
xvin
Acknowledgements
We make a particular acknowledgement to the late Dr John Longwell CBE, Deputy Government Chemist in 1964, who encouraged the Laboratory to enter the field, and to the line of Government Chemists who supported the work over the long years; the late Dr David Lewis CB, the late Dr Harold Egan, Dr Ron Coleman CB (who became Chief Scientist of the Department of Trade and Industry), Mr Alex Williams CB and Dr Richard Worswick. We note the particular contributions of Brian Kent, present Head of the Materials Technology Group, as co-inventor of the glass polyalkenoate cement way back in 1968, and of Dr John McLean OBE in developing clinical applications. It was Surgeon Rear Admiral Holgate CB, OBE, Chief Dental Officer at the Ministry of Health in 1964, who introduced Dr McLean to the Laboratory of the Government Chemist (LGC) to initiate a collaboration that proved so fruitful. Since then there has been constant support from the Department of Health and its various officers and also from the British Technology Group, particularly from G. M. Blunt and R. A. Lane. Most importantly we acknowledge the contribution of those who worked at that essential place, the laboratory bench, on which everything depends. Our colleagues in the Materials Technology Group (formerly the Dental Materials Group) who have worked with one or other of us since 1964 are: R. F. Batchelor, B. G. Lewis, Mrs B. G. Scott, J. M. Paddon, G. Abel, Dr S. Crisp, A. J. Ferner, Dr H. J. Prosser, M. A. Jennings, Mrs S. A. Merson, M. Ambersley, D. M. Groffman, S. M. Jerome, D. R. Powis, Mrs P. J. Brookman (nee Brant), R. P. Scott, J. C. Skinner, Dr R. G. Hill, G. S. Sayers, Dr C. P. Warrens, Miss A. M. Jackson, Dr J. Ellis, Miss E. A. Wasson, Miss H. M. Anstice, Dr J. H. Braybrook, Miss S. J. Hawkins and A. D. Akinmade. xix
Acknowledgements In addition we have received support from members of other divisions at the LGC: Dr R. J. Mesley, M. A. Priguer, D. Wardleworth, Dr I. K. O'Neill, B. Stuart, R. A. Gilhooley, Dr C. P. Richards, Dr O. M. Lacy and Dr S. L. R. Ellison. Guest workers to the Materials Technology Group who have contributed include Professor P. Hotz (Klinik fur Zahnerhaltung der Universitat, Bern), Ms T. Folleras (NIOM, Scandinavian Institute of Dental Materials). Workers in other Government Research Stations and the Universities who have collaborated with us are: R. P. Miller, D. Clinton, Dr T. I. Barry, Dr I. Seed (National Physical Laboratory); K. E. Fletcher (Buildings Research Station); Miss D. Poynter (Warren Spring Laboratory); Professor L. Holliday, Dr J.H.Elliott, Dr P. R. Hornsby, Dr K. A. Hodd, Dr A. L. Reader (Brunei University); R. Manston, Dr B. F. Sanson, Dr W. M. Allen, P. J. Gleed (Institute for Research on Animal Diseases); Professor Braden (London Hospital); A. C. Shorthall (Birmingham University), I. M. Brook (University of Sheffield); and R. Billington (Institute of Dental Surgery, London). We thank Dr L. J. Pluim of the Rijksuniversiteit te Groningen for drawing our attention to the early and neglected work of E. van Dalen on zinc phosphate cements. We thank Mrs Margaret Wilson for her help in checking the proofs and the indexing. We acknowledge the stoic forbearance of our wives in putting up with the disturbances and neglect of domestic routines and duties occasioned by the writing of a book. Alan D. Wilson John W. Nicholson
xx
1
Introduction
Acid-base (AB) cements have been known since the mid 19th century. They are formed by the interaction of an acid and a base, a reaction which yields a cementitious salt hydrogel (Wilson, 1978) and offers an alternative route to that of polymerization for the formation of macromolecular materials. They are quick-setting materials, some of which have unusual properties for cements, such as adhesion and translucency. They find diverse applications, ranging from the biomedical to the industrial. Despite all this there has been a failure to recognize AB cements as constituting a single, well-defined class of material. Compared with organic polymers, Portland cement and metal alloys, they have been neglected and, except in specializedfields,awareness of them is minimal. In this book we attempt to remedy the situation by unifying the subject and treating this range of materials as a single class. Human interest in materials stretches back into palaeolithic times when materials taken from nature, such as wood and stone, were fashioned into tools, weapons and other artifacts. Carving or grinding of a material is a slow and time-consuming process so the discovery of pottery, which does away with the need for these laborious processes, was of the greatest significance. Here, a soft plastic body, potter's clay, is moulded into the desired shape before being converted into a rigid substance by firing. Pottery is but one of a group of materials which are formed by the physical or chemical conversion of a liquid or plastic body, which can be easily shaped by casting or moulding, into a solid substance. Other examples of this common method of fabrication are the casting of metals and the injection moulding of plastics. Into this category come the water-based plasters, mortars, cements and concretes which set at room temperature as the result of a chemical reaction between water and a powder. Some of these have been known 1
Introduction since antiquity. The AB cements are related to these materials except that water is replaced by an acidic liquid. ThefirstAB cement, the zinc oxychloride cement, was reported by Sorel in 1855. It was prepared by mixing zinc oxide powder with a concentrated solution of zinc chloride. Its use in dentistry was recommended by Feichtinger in 1858 but it did not prove to be a success (Mellor, 1929). However, other AB cements have proved to be of the utmost value to dentistry, and their subsequent development has been closely associated with this art (Wilson, 1978). The AB cements, developed against the backcloth of the severe demands of dentistry, have interesting properties. Some possess aesthetic appeal and the ability to adhere to base metals and other reactive substrates. Most have superior properties to plasters, mortars, and Portland cements, being quick-setting, stronger and more resistant to erosion. These advantageous properties make them strong candidates for other applications. In fact, one of these cements, the magnesium oxychloride cement of Sorel (1867), is still used to surface walls and floors on account of its marble-like appearance (Chapter 7). In the 1870s more effective liquid cement-formers were found: orthophosphoric acid and eugenol (Wilson, 1978). It was also found that an aluminosilicate glass could replace zinc oxide, a discovery which led to the first translucent cement. Thereafter the subject stagnated until the late 1960s when the polyelectrolyte cements were discovered by Smith (1968) and Wilson & Kent (1971). In recent years Sharp and his colleagues have developed the magnesium phosphate cements - Sharp prefers the term magnesia phosphate cement - as a material for the rapid repair of concrete runways and motorways (Chapter 6). These applications exploit the rapid development of strength in AB cements. This cement can also be used for flooring in refrigerated stores where Portland cements do not set. Interestingly, this material appears to have started life as an investment for the casting of dental alloys. The glass polyalkenoate, a polyelectrolyte cement, of Wilson & Kent (Chapter 5), was originally developed as a dental material but has since found other applications. First it was used as a splint bandage material possessing early high-strength and resistance to water. Currently, it is being used, as a biocompatible bone cement, with a low exothermicity on setting and the ability to adhere to bone, for the cementation of prostheses (Jonck, Grobbelaar & Strating, 1989). Outside thefieldof biomaterials it has been patented for use as a cement for underwater pipelines, as a foundry sand and as a substitute for plaster
Introduction in the slip casting of pottery. Quite often it appears as a substitute for plaster of Paris, for it is stronger, less brittle and more resistant to water. There are other possibilities. Its translucent nature suggests that it could be used for the production of porcelain-like ceramics at room temperature. Phosphate and polyelectrolyte AB cements are resistant to attack by boiling water, steam and mild acids and this suggests that they could be employed in technologies where these properties are important. The ability of the polyelectrolyte-based AB cements (Chapter 5) to bond to a variety of substrates, combined with their rapid development of strength - they can become load-bearing within minutes of preparation suggests that they have applications as rapid-repair and handyman materials. A current area of interest is the use of AB cements as devices for the controlled release of biologically active species (Allen et aL, 1984). AB cements can be formulated to be degradable and to release bioactive elements when placed in appropriate environments. These elements can be incorporated into the cement matrix as either the cation or the anion cement former. Special copper/cobalt phosphates/selenates have been prepared which, when placed as boluses in the rumens of cattle and sheep, have the ability to decompose and release the essential trace elements copper, cobalt and selenium in a sustained fashion over many months (Chapter 6). Although practical examples are confined to phosphate cements, others are known which are based on a variety of anions: polyacrylate (Chapter 5), oxychlorides and oxysulphates (Chapter 7) and a variety of organic chelating anions (Chapter 9). The number of cements available for this purpose is very great. A recent development has been the incorporation of a bioactive organic component into the AB cement during preparation. Since AB cements are prepared at room temperature, this can be done without causing degradation of the organic compound. In this case, the AB cement may merely act as a carrier for the sustained release of the added bioactive compound. Another development has been the advent of the dual-cure resin cements. These are hybrids of glass polyalkenoate cements and methacrylates that set both by an acid-base cementation reaction and by vinyl polymerization (which may be initiated by light-curing). In these materials, the solvent is not water but a mixture of water and hydroxyethylmethacrylate which is capable of taking dimethacrylates and poly(acrylic acid)-containing vinyl groups into solution. In the absence of light these materials set slowly and
Introduction have extended working times, but they set in seconds when illuminated with an intense beam of visible light. These hybrids are in their infancy but have created great interest. From this account we are to expect diversification of these AB cements both for biomedical and for industrial usages. There should be further developments of the glass polyalkenoate cements both as bone substitutes and as bioadhesives. We also expect more types of AB cements to be formulated as devices for the sustained release of bioactive species. These materials would have applications in agriculture, horticulture, animal husbandry and human health care. In industrialfieldswe expect that there will be continued interest in developing AB cements as materials for the rapid repair of constructural concrete, as materials for the surfacing of floors and walls, and as adhesives and lutes for cementation in aqueous environments. The hybrid light-cured cements also appear to be a promising new line of development which may give us entirely novel classes of materials. References Allen, W. M., Sansom, B. F., Wilson, A. D., Prosser, H. J. & Groffman, D. M. (1984). Release cements. British Patent GB 2,123,693 B. Jonck, L. M., Grobbelaar, C. J. & Strating, H. (1989). The biocompatibility of glass-ionomer cement in joint replacement: bulk testing. Clinical Materials, 4, 85-107. Mellor, J. W. (1929). A Comprehensive Treatise on Inorganic and Theoretical Chemistry, vol. IV, p. 546. London: Longman. Sorel, S. (1855). Procede pour la formation d'un ciment tres-solide par 1'action d'un chlorure sur l'oxyde de zinc. Comptes rendus hebdomadaires des seances de T Academie des sciences, 41, 784-5. Sorel, S. (1867). On a new magnesium cement. Comptes rendus hebdomadaires des seances de VAcademie des sciences, 65, 102—4. Wilson, A. D. (1978). The chemistry of dental cements. Chemical Society Review, 7, 265-96.
2
Theory of acid-base cements
2.1
General
From the chemical point of view AB cements occupy a place in the vast range of acid-base phenomena which occur throughout both inorganic and organic chemistry. Like Portland cement they are prepared by mixing a powder with a liquid. However, this liquid is not water but an acid, while the powder, a metal oxide or silicate, is a base. Not surprisingly, the cement-forming reaction between them is extremely rapid and a hardened mass is formed within minutes of mixing. AB cements may be represented by the defining equation Base + Acid = Salt + Water (powder) (liquid) (cement matrix) The product of the reaction, the binding agent, is a complex salt, and powder in excess of that required for the reaction acts as the filler. Each cement system is a particular combination of acid and base. The number of potential cement systems is considerable since it is a permutation of all possible combinations of suitable acids and bases. Cement-forming liquids are strongly hydrogen-bonded and viscous. According to Wilson (1968), they must (1) have sufficient acidity to decompose the basic powder and liberate cement-forming cations, (2) contain an acid anion which forms stable complexes with these cations and (3) act as a medium for the reaction and (4) solvate the reaction products. Generally, cement-forming liquids are aqueous solutions of inorganic or organic acids. These acids include phosphoric acid, multifunctional carboxylic acids, phenolic bodies and certain metal halides and sulphates (Table 2.1). There are also non-aqueous cement-forming liquids which are multidentate acids with the ability to form complexes. Potential cement-forming bases are oxides and hydroxides of di- and
Theory of acid—base cements Table 2.1. Examples of acids used for cement formation Protonic acids (used in aqueous solution)
Aprotic acids (used in aqueous solution)
Phosphoric acid Poly(acrylic acid) Malic acid Tricarballylic acid Pyruvic acid Tartaric acid Mellitic acid Gallic acid Tannic acid
Magnesium chloride Zinc chloride Copper(II) chloride Cobalt(II) chloride Magnesium sulphate Zinc sulphate Copper(II) sulphate Cobalt(II) sulphate Magnesium selenate Zinc selenate Copper(II) selenate Cobalt(II) selenate
Protonic acids (liquid non-aqueous) Eugenol 2-ethoxybenzoic acid
Table 2.2. Examples of bases used for cement formation Copper(II) oxide Zinc(II) oxide Magnesium oxide Cobalt(II) hydroxide Cobalt(II) carbonate Calcium aluminosilicate glasses Gelatinizing minerals
trivalent metals, silicate minerals and aluminosilicate glasses (Table 2.2). All cement-forming bases must be capable of releasing cations into acid solution. The best oxides for cement formation are amphoteric (Kingery, 1950a,b) and the most versatile cement former is zinc oxide, which can react with a wide range of aqueous solutions of acids, both inorganic and organic, and liquid organic chelating agents. Gelatinizing minerals, that is minerals that are decomposed by acids, can act as cement formers, as can the acid-decomposable aluminosilicate glasses. In this chapter the nature of the cementitious bond and the acid-base reaction will be discussed.
The formation of cements 2.2
The formation of cements
2.2.1 Classification Before proceeding further it is well to consider the term cement, for its definition can be the source of some confusion. Both the Oxford English Dictionary and Webster give two alternative definitions. One defines a cement as a paste, prepared by mixing a powder with water, that sets to a hard mass. In the other a cement is described as a bonding agent. These two definitions are quite different. The first leads to a classification of cements in terms of the setting process, while the second lays emphasis on a property. In this book the term cement follows the sense of thefirstof these definitions. Cements can be classified into three broad categories: (1) Hydraulic cements. These cements are formed from two constituents one of which is water. Setting comprises a hydration and precipitation process. Into this category fall Portland cement and plaster of Paris. (2) Condensation cements. Here, cement formation involves a loss of water and the condensation of two hydroxyl groups to form a bridging oxygen: R-OH + HO-R = R-O-R + H2O One example is silicate cement where orthosilicic acid, chemically generated in solution, condenses to form a silicic acid gel. Another is refractory cement where a cementitious product is formed by the heat treatment of an acid orthophosphate, a process which again involves condensation to form a polyphosphate. (3) Acid-base cements. Cement formation involves both acid-base and hydration reactions (Wilson, Paddon & Crisp, 1979). These cements form the subject of this book. This classification differs from that given by Wygant (1958), who subdivides cements into hydraulic, precipitation and reaction cements. The advantage of the present classification is that it clearly differentiates phosphate cements formed by condensation from those formed by an acid-base reaction (Kingery, 1950a). Wygant includes these in the same category, which can be confusing. Moreover, he puts silicate cements and the heat-treated acid phosphate cements into separate categories, although both are condensation cements.
Theory of acid—base cements 2.2.2
Requirements for cementitious bonding
The essential property of a cementitious material is that it is cohesive. Cohesion is characteristic of a continuous structure, which in the case of a cement implies an isotropic three-dimensional network. Moreover, the network bonds must be attributed to attractions on the molecular level. Increasingly, recent research tends to show that cements are not bonded by interlocking crystallites and that the formation of crystallites is incidental (Steinke et al., 1988; Crisp et al., 1978). The reason is that it is difficult to form rapidly a mass which is both cohesive and highly ordered. Cement formation requires a continuous structure to be formed in situ from a large number of nuclei. Moreover, this structure must be maintained despite changes in the character of the bonds. These criteria are, obviously, more easily satisfied by aflexiblerandom structure than by one which is highly-ordered and rigid. Crystallinity implies well-satisfied and rigidlydirected chemical bonds, exact stoichiometry and a highly ordered structure. So unless crystal growth is very slow a continuous molecular structure cannot be formed. In random structures, stoichiometry need not be exact and adventitious ions can be incorporated without causing disruption. Bonds are not highly directed, and neighbouring regions of precipitation, formed around different nuclei, can be accommodated within the structure. Continuous networks can be formed rapidly. Thus, random structures are conducive to cement formation and, in fact, most AB cements are essentially amorphous. Indeed, it often appears that the development of crystallinity is detrimental to cement formation. The matrices of AB cements are gel-like, but these gels differ from the tobermorite gel of Portland cement. In AB cements, setting is the result of gelation by salt formation, and the cations, which cause gelation, are extracted from an oxide or silicate by a polyacid solution. The conversion of the sol to a gel is rapid and the cements set in 3 to 5 minutes. Two basic processes are involved in cement formation: the release of cations from the oxide or silicate and their interaction with polyacid. This interaction involves ion binding and changes in the hydration state which are associated with gelation and structure formation (Section 4.3). Thus, there are two reaction rates to be considered: the rate of release of cations and the rate of structure formation. These two reaction rates must be balanced. If the rate of release of cations is too fast a non-coherent precipitate of crystallites is formed. If too slow the gel formed will lack strength.
The formation of cements During cement formation, domains are formed about numerous nuclei and there must be bonding between the domains as well as within them. In AB cements bonding within the domains is mainly ionic, with a degree of covalency. The attractive forces between domains are those of a colloidal type. In random structures, residual forcefieldsexist which act in a similar fashion to polar forces and serve to bond domains. These forces must include hydrogen bonds, for the addition offluorideions always enhances cement strength and the fluoride-hydrogen bond is a strong one. The structures of cement gels bear some relationship to the structure of glasses. Spatially, the O2~ ion is dominant. The matrices are based on a coordinated polyhedron of oxygen ions about a central glass-forming cation (Pauling, 1945). In effect, these are anionic complexes where the cations are small, highly charged, and capable of coordinating with oxygen or hydroxyl ions. Examples of these polyhedra are [SiOJ, [POJ and [A1OJ. Thus, wefindthat there are silicate, phosphate and aluminosilicate glasses and gels. There are, however, differences which are best illustrated by reference to the simple example of silica glass and silica gel. In silica glass, Si4+ is fourcoordinate and the polymeric links are of the bridging type: -«>Si—O—Si66
52-66 45-52 , as = Z/r
(2.1)
where Z is the charge on the central atom and r its ionic radius. Cartledge (1928b) then used values of ^ 0 5 to define acidity and basicity of a species.
f5 value >3-2 2-2-3-2 Cu2+ > Zn2+ > Mg2+ > Ca2+ The first three form amphoteric oxides and are distinctly superior, as cement-formers, to the latter two which form weakly basic oxides. Data from Table 2.3b indicate that optimum cement formation occurs with cations that have In values lying between 18 and 29. 2.3.6
Acid-base classification
The strength of a Lewis acid or base depends on the particular reaction, and for this reason there is no absolute scale for the strengths of Lewis acids and bases. However, certain qualitative features have been observed. Ahrland, Chatt & Davies (1958) divided metal ions (which are Lewis acids), on the basis of the stability of their complexes, into what they termed class (a) and class (b) acceptors (Table 2.4). They stated that class (a) acceptors form their most stable complexes with ligands of the lightest member of a non-metal group. By contrast, class (b) acceptors form their most stable complexes with heavier members of each group. Thus, complex stability can be ranked according to the ligand as follows. For class (a) acceptors O P S and for class (b) acceptors O c Ti
a/b border
N
0
F
Al
Si
P
s
Cl
Zn
Ga
Ge
As
Se
Br
Ru
Rh Pd Ag
Cd
In
Sn
Sb
Te
I
Os
Ir
Pt Au
Hg
TI
Pb
Bi
Po
At
Mn Fe
Rb Sr ^ft Zr Nb
Mo
Tc
Cs Ba 1^a Hf Ta
W
Re
Class (a)
C
Ni Cu
Cr
V
B
a/b border
Co
Class (b)
a/b border
Theory of acid-base cements 2.3.7
Hard and soft acids and bases (HSAB)
This concept of Chatt and his coworkers was developed further by Pearson (1963, 1966, 1968a,b) in his theory of hard and soft acids and bases. Hard acids correspond with class (a) acceptors and soft acids with class (b) acceptors. Hard acids prefer to react with hard bases and soft acids prefer to react with soft bases. Hard acids are characterized by small size, high positive charge and absence of outer electrons which are easily excited to higher states; they are thus of low polarizability. In this class are the common protonic acids, HA, the hydrogen-bonding molecules in the Lewis scheme and Mg 2+ , which are all acids of relevance to AB cements. The soft acids have low or zero positive charge, large size and several easily excited outer electrons (often d orbital electrons). These properties lead to high polarizability. The division between these two classes is not sharp; amongst the intermediate class are Zn 2+ and Cu 2+ . Pearson (1966) defines a soft base as 'one in which the donor atom is of high polarizability and low electronegativity and is easily oxidized or associated with empty, low-lying orbitals'. A hard base has opposite properties. 'The donor atom is of low polarizability and high electronegativity, is hard to reduce, and is associated with empty orbitals of high energy.' The underlying theory for hard-hard and soft-soft preferences is obscure and no one factor is responsible (Pearson, 1966). Pearson (1963, 1968b) advanced several explanations. He stated that the ionic-covalent theory provides the most obvious explanation. Hard acids are assumed to bind bases primarily by ionic forces and soft acids by covalent bonds. High positive charge and small size favour strong ionic bonding, and bases of large negative charge and small size would be most strongly held. Soft acids bind to bases by covalent bonding, and the atoms should be of similar size and electronegativity for good bonding. The classification of Lewis acids and bases relevant to AB cements is shown below. Hard acids: Borderline acids: Hard bases: 24
H A , H + , Ca 2 + , Mg 2 + , Al 3 + , Si 4+ Zn 2+ , Cu2+ H 2 O, O H , F", POJ", SO2", RCOO"
Acid-base concepts Table 2.5. YatsimirskiVs hardness indices {Yatsimirskii, 1970) Base
Indices
Acid
Indices
OHF-
6-3 1-7 1-7 0-8 0-5
H+
9-0 1-2 10 0-2 01
HPOJCH3COO-
sor
H2O
In 3 + Cu 2 + Zn 2 + La 3 +
zero
Extension of HSAB theory Yatsimirskii (1970) attempted to quantify HSAB theory and produced hardness indices (S) for acids and bases. These indices were obtained by plotting the logarithms of the equilibrium constants for the reactions of bases with the proton (the hardest acid) against similar values for the reactions with CH 3 Hg + (one of the softest acids). For acids, the hydroxyl ion (the hardest base) and the chloride ion (a soft base) were chosen. These S indices for cations and anions relevant to AB cements are shown in Table 2.5. Bases which add on through F or O and do not form Tr-bonds have similar hardness values; they are hard bases. Soft bases form dative 7r-bonds with many cations. They have high-energy-level occupied orbitals with unshared electron pairs. Yatsimirskii considered that the hard and soft classification was too general and proposed instead a more detailed approach. He classified Lewis acids and bases into six groups, based on the nature of the adduct bonding. Group (1) Cations and anions which are incapable of donor-acceptor interactions. These are the large univalent ions. Bonding is purely by Coulomb and Madelung electrostatic interactions. From the Lewis point of view these are not acids or bases. They have no cement-forming potential. Group (2) Strong a-acceptor acids and donor bases. Included here are protonic acids, which are relevant to AB cements. Their adducts can only contain one coordinate bond. Group (3) G- and n-acceptor acids and donor bases with o-interactions predominating. In this group acceptors are capable of adding on electron pairs of donors in both types of interactions. Includes cations with stable closed electron shells: Al 3+ , Mg 2+ , Ca2+ and 25
Theory of acid-base cements Zn 2+ . Donors are ligands coordinated through oxygen atoms or fluoride ions: RCOO", PO*~, OH", F" and H 2 O. These acceptors and donors are of relevance to AB cements. Group (4) Strong a- and n-acceptor acids and donor bases. Bi3+, In 3+ and Sn2+ are of some relevance to AB cements. Group (5) Acids that are o-acceptors but capable of n-donation in backbonding. This group includes cations with mobile d electrons e.g. Cuw+, Co w+ , Fe w+ . Group (6) Bases that are a-donors but n-acceptors. According to Yatsimirskii, group (2) and (3) species are equivalent to Pearson's hard acids and bases, and group (4), (5) and (6) species correspond to Pearson's soft acids and bases. This classification is of more value than HSAB theory to our subject. It can be seen that all cementforming anions come from group (3) and cations from groups (3), (4) and (5). Thus, the bonding in cement matrices formed from cation-anion combinations is not purely a but contains some n character.
References Ahrland, S., Chatt, J. & Davies, N. R. (1958). The relative affinities of ligand atoms for acceptor molecules and ions. Quarterly Reviews, 12, 265-76. Baes, C. F. & Mesmer, R. E. (1976). The Hydrolysis of Cations. New York: John Wiley. Bell, R. P. (1947). The use of the terms 'acid' and 'base'. Quarterly Reviews, 1, 113-25. Bell, R. P. (1973). The Proton in Chemistry. Ithaca, New York: Cornell University Press. Bjerrum, J. (1951). Die Entwickhmgsgeschichte des Saure-Basenbegriffes und iiber die ZweckmaBigkeit der Einfuhrung eines besonderen Antibasenbegriffes neben dem Saurebegriff. Naturwissenschaften, 38, 461-4. Boyle, R. (1661). The Sceptical Chymist. Everyman Library Edition, 1911. Brensted, J. N. (1923). Einige Bemerkungen iiber den Begriff der Sauren und Basen. Recueil des Travaux chimiques des Pays-Bas et de la Belgique, 42, 718-28. Bronsted, J. N. (1926). The acid-base function of molecules and its dependency on the electronic charge type. Journal of Physical Chemistry, 30, 777-90. Bungenberg de Jong, H. G. (1949). In Kruyt, H. R. (ed.) Colloid Science II, p. 2. Amsterdam: Elsevier Publishing Co. Inc. Cady, H. P. & Elsey, H. M. (1922). A general conception of acids, bases and salts. Science, 56, 27 (Lecture abstract). Cady, H. P. & Elsey, H. M. (1928). A general definition of acids, bases and salts. Journal of Chemical Education, 5, 1425-8. 26
References Cartledge, G. H. (1928a). Studies on the periodic system. I. The ionic potential as a periodic function. Journal of the American Chemical Society, 50, 2855-63. Cartledge, G. H. (1928b). Studies on the periodic system. II. The ionic potential and related properties. Journal of the American Chemical Society, 50, 2863-72. Chatt, J. (1958). The stabilisation of low valent states of the transition metals. Journal of Inorganic & Nuclear Chemistry, 8, 515-31. Cotton, F. A. & Wilkinson, G. (1966). Advanced Inorganic Chemistry, 2nd edn. New York, London & Sydney: Wiley Inter science. Crisp, S., O'Neill, I. K., Prosser, H. J., Stuart, B. & Wilson, A. D. (1978). Infrared spectroscopic studies on the development of crystallinity in dental zinc phosphate cements. Journal of Dental Research, 57, 245-54. Crosland, M. P. (1962). Historical Studies in the Language of Chemistry. London: Heinemann. Crosland, M. (1973). Lavoisier's theory of acidity. Isis, 64, 306-25. Day, M. C. & Selbin, J. (1969). Theoretical Inorganic Chemistry. New York: Reinhold. Finston, H. L. & Rychtman, A. C. (1982). A New View of Current Acid-Base Theories. New York: John Wiley & Sons. Flood, H. & Forland, T. (1947a). The acidic and basic properties of oxides. Ada Chemica Scandinavica, 1, 592—604. Flood, H. & Forland, T. (1947b). The acidic and basic properties of oxides. II. The thermal decomposition of pyrosulphates. Acta Chemica Scandinavica, 1, 781-9. Flood, H., Forland, T. & Roald, B. (1947). The acidic and basic properties of oxides. III. Relative acid-base strengths of some polyacids. Acta Chemica Scandinavica, 1, 790-8. Flory, P. J. (1953). Principles of Polymer Chemistry, Chapter 11. Ithaca, New York: Cornell University Press. Flory, P. J. (1974). Introductory lecture. In Gels and Gelling Processes. Faraday Discussions of the Chemical Society, No. 57, pp. 7-18. Franklin, E. C. (1905). Reactions in liquid ammonia. Journal of the American Chemical Society, 27, 820-51. Franklin, E. C. (1924). Systems of acids, bases and salts. Journal of the American Chemical Society, 46, 2137-51. Germann, A. F. O. (1925a). What is an acid? Science, 61, 71. Germann, A. F. O. (1925b). A general theory of solvent systems. Journal of the American Chemical Society, 47, 2461-8. Hall, N. F. (1940). Systems of acids and bases. Journal of Chemical Education, 17, 124^8. Hodd, K. A. & Reader, A. L. (1976). The formation and hydrolytic stability of metal ion-polyacid gels. British Polymer Journal, 8, 131-9. Jensen, W. B. (1978). The Lewis acid-base definitions: a status report. Chemical Reviews, 78, 1-22. Kingery, W. D. (1950a). Fundamental study of phosphate bonding in refractories. I. Literature review. Journal of the American Ceramic Society, 33, 239-41. 27
Theory of acid-base cements Kingery, W. D. (1950b). Fundamental study of phosphate bonding in refractories. II. Cold setting properties. Journal of the American Ceramic Society, 33, 242-7. Kolthoff, I. M. (1944). The Lewis and Bronsted-Lowry definitions of acids and bases. Journal of Physical Chemistry, 48, 51-7. Lewis, G. N. (1916). The atom and the molecule. Journal of the American Chemical Society, 38, 762-85. Lewis, G. N. (1923). Valence and the Structure of Atoms and Molecules. New York: Chemical Catalog Co. Lewis, G. N. (1938). Acids and bases. Journal of the Franklin Institute, 226, 293-337. Lowry, T. M. (1923a). The uniqueness of hydrogen. Chemistry & Industry, 42, 43. Lowry, T. M. (1923b). Co-ordination and acidity. Chemistry & Industry, 42, 1048-52. Luder, W. F. (1940). The electronic theory of acids and bases. Chemical Reviews, 27, 547-83. Luder, W. F. (1948). Contemporary acid-base theory. Journal of Chemical Education, 25, 555-8. Lux, H. (1939). 'Sauren' und 'Basen' im Schelzfluss: Die Bestimmung der Sauerstoffionen-Konzentration. Zeitschrift fur Elektrochemie, 45, 303-9. Pattison Muir, M. M. (1883). Heroes of Science-Chemists, Chapter IV, pp. 171-89. London: Society for Promoting Christian Knowledge. Pauling, L. (1945). The Nature of the Chemical Bond. Ithaca, New York: Cornell University Press. Pearson, R. G. (1963). Hard and soft acids and bases. Journal of the American Chemical Society, 85, 3533-9. Pearson, R. G. (1966). Acids and bases. Science, 151, 172-7. Pearson, R. G. (1968a). Hard and soft acids and bases, HSAB. Part I. Fundamental principles. Journal of Chemical Education, 45, 581-7. Pearson, R. G. (1968b). Hard and soft acids and bases, HSAB. Part II. Underlying theories. Journal of Chemical Education, 45, 643-8. Read, H. H. (1948). Rutle/s Elements of Mineralogy, 24th edn. London: Thomas Murby & Co. Smith, D. C. (1968). A new dental cement. British Dental Journal, 125, 381-4. Steinke, R., Newcomer, P., Komarneni, S. & Roy, R. (1988). Dental cements: investigation of chemical bonding. Materials Research Bulletin, 23, 13-22. Usanovich, M. I. (1939). On acids and bases. Journal of General Chemistry (USSR), 9, 182-92. Vander Werf, A. (1961). Acids, Bases, and the Chemistry of the Covalent Bond. New York: Reinhold. Walden, P. (1929). Salts, Acids and Bases: Electrolytes, Stereochemistry. New York: McGraw-Hill. Wilson, A. D. (1968). Dental silicate cements: VII. Alternative liquid cement formers. Journal of Dental Research, 47, 1133-6. Wilson, A. D. & Kent, B. E. (1971). The glass-ionomer cement: a new
28
References translucent cement for dentistry. Journal of Applied Chemistry and Biotechnology, 21, 313. Wilson, A. D., Paddon, J. M. & Crisp, S. (1979). The hydration of dental cements. Journal of Dental Research, 58, 1065-71. Wygant, J. F. (1958). Cementitious bonding in ceramic fabrication. In Kingery, W. D. (ed.) Ceramic Fabrication Processes, pp. 171-88. New York: John Wiley & Sons. Yatsimirskii, K. (1970). Acid-base and donor-acceptor properties of ions and molecules. Theoretical and Experimental Chemistry (USSR), 6, 376-80.
29
3
Water and acid-base cements
3.1
Introduction
The setting reaction for the great majority of acid-base cements takes place in water. (The exceptions based on o-phenols are described in Chapter 9.) This reaction does not usually proceed with formation of a precipitate but rather yields a substance which entrains all of the water used to prepare the original cement paste. Water thus acts as both solvent and component in the formation of these cements. It is also one of the reaction products, being formed in the acid-base reaction as the cements set. 3.1.1
Water as a solvent
It is widely recognized that the solvent in which any chemical reaction takes place is not merely a passive medium in which relevant molecules perform: the solvent itself makes an essential contribution to the reaction. The character of the solvent will determine which chemical species are soluble enough to enter solution and hence to react, and which species are insoluble, and thus precipitate out of solution, thereby being prevented from undergoing further chemical change. In the case of water, as will be seen, polar and ionic species are the ones that most readily dissolve. But even so, mere polarity or ionic character is not sufficient to ensure solubility. Solubility depends on a number of subtle energetic factors, and the possible interactions between water and silver chloride, for example, do not fulfil the requirements despite the ionic nature of the silver salt. Hence silver chloride is almost completely insoluble in water. 3.1.2
Water as a component
In AB cements water does not merely act as solvent for the setting reaction. It also acts as an important component of the set cement. For example, 30
Water glass-ionomer dental cements as generally formulated include at least 15% by mass of water, all of which becomes incorporated into the complete cement (Wilson & McLean, 1988). Indeed, great importance is attached to the retention of water by these cements, since if they are allowed to dry out by storage under conditions of low humidity, they shrink significantly, and develop cracks and crazes. Another class of AB cement, the oxychloride cements of zinc and magnesium, are also formulated in aqueous solution and retain substantial amounts of water on setting (Sorrell & Armstrong, 1976; Sorrell, 1977). Water may have a number of roles in the set versions of these cements. It is capable of solvating the cement-forming ions, such as Ca2+ or Zn2+, depending on the cement. It also contributes a sheath of solvating molecules around polyelectrolytes such as poly(acrylic acid) in glassionomer and zinc polycarboxylate cements. Significant amounts of water are known to be retained by metal polyacrylate salts at equilibrium and this water contributes to reducing the glass transition temperature of such materials by acting as a plasticizer (Yokoyama & Hiraoko, 1979). These various aspects of water in AB cements are covered in the present chapter. Its solvent character, structure and hydration behaviour are described, and the chapter concludes with a more thorough consideration of the precise role of water in the various AB cements.
3.2 3.2.1
Water Constitution
Water has a deceptively simple chemical constitution, consisting as it does of molecules containing two atoms of hydrogen and one of oxygen. It was viewed by the ancients as one of the four 'elements', following Aristotle's classification, the others being air, fire and earth. The modern view that it is a compound composed of hydrogen and oxygen was first established in 1789 by two amateur chemists, Adriaan Paets van Troostwijk (1752-1837), a merchant, and Jan Rudolph Deiman (1743-1808), a pharmacist (Hall, 1985). They were able to show by synthesis which elements combine to make water, forming it from reaction of hydrogen gas with oxygen. Their work was important historically for the part it played in undermining the phlogiston theory of combustion. It was left to the great Swedish chemist J. J. Berzelius (1779-1848) to determine that the ratio of hydrogen to oxygen is 2:1. 31
Water and acid-base cements Table 3.1. Molecular dimensions of normal and isotopic water in the vapour phase {Benedict, Gailar & Plyler, 1956) Molecule
Bond length, pm
Bond angle, degrees
D2O H2O HDO
95-75 95-718 95-71
104-474 104-523 104-529
As a compound water is remarkable. It is the only inorganic liquid to occur naturally on earth, and it is the only substance found in nature in all three physical states, solid, liquid and vapour (Franks, 1983). It is the most readily available solvent and plays a vital role in the continuation of life on earth. Water circulates continuously in the environment by evaporation from the hydrosphere and subsequent precipitation from the atmosphere. This overall process is known as the hydrologic cycle. Reports estimate that the atmosphere contains about 6 x 1015 litres of water, and this is cycled some 37 times a year to give an annual total precipitation of 224 x 1015 litres (Franks, 1983; Nicholson, 1985). The bond lengths and bond angle for the water molecule are known very precisely following studies of the rotation-vibration spectra of water vapour, and also the vapour of the deuterated analogues of water, D2 O and HDO (Eisenberg & Kauzmann, 1969). The data for these compounds are shown in Table 3.1. The nuclei of the water molecule, regardless of the isotopes involved, form an isosceles triangle having a slightly obtuse angle at the oxygen atom. All of the data in Table 3.1 refer to the equilibrium state of the water molecules, which is formally acceptable, but is actually a hypothetical state, since it assumes neither rotation nor vibration in the molecule. The equilibrium bond lengths and bond angles can be seen to differ little between the different isotopic molecules. Such a finding agrees with the predictions of the Born-Oppenheimer approximation, that the electronic structure of a molecule is independent of the mass of its nuclei, it being the electronic structure of a molecule alone which determines the geometry. The bond angle in water is slightly less than the ideal tetrahedral angle of 109-5°.This is attributed to the presence of lone pairs of electrons on the oxygen atom which repel more strongly than the bonding pairs of electrons between the oxygen and hydrogen nuclei (Speakman, 1975). The valence32
Water Table 3.2. Properties of hydrides of first row elements (Weast, 1985-6)
Compound
Relative molar mass
Melting point, °C
Boiling point, °C
Gas phase dipole moment, debyes
CH 4 NH 3 H2O HF
16 17 18 19
-182-0 -11-1 00 -834
-1640 -33-4 1000 19-5
000 1-47 1-85 1-82
shell electron-pair repulsion concepts of Gillespie & Nyholm (1957) show that such increased repulsion by lone pairs closes the angle between the bonding pairs slightly but significantly for the water molecule. The O-H bond energy of water is taken as half the energy of formation of the molecule, since water has two such bonds. This gives a value of 458-55 kJ mol"1 at 0 K (Eisenberg & Kauzmann, 1969). Related to the bond energy is the dissociation energy, i.e. the energy required to break the bond at 0 K. Neither of the O-H bonds in water has a dissociation energy equal to the O-H bond energy. Instead, the first O-H dissociation energy has been found experimentally to be 424-27 kJ mol"1. From conservation of energy considerations which lead to the requirement that the sum of the two dissociation energies must equal the energy of formation, it is found that the second O-H dissociation energy has to take a value of 492-83 kJ mol"1. This has been explained (Pauling, 1960) by postulating an electronic rearrangement on the oxygen atom of the O-H fragment left behind after scission of the first O-H bond, and that breaking the bond between oxygen in this new electronic configuration and the remaining hydrogen requires greater energy. 3.2.2
Water compared with other hydrides
Water shows properties that are interestingly different compared with hydrides of the neighbouring elements of thefirstrow of the periodic table. Some of these properties are given in Table 3.2. From this table, water can be seen to have a very high melting point and a very high boiling point for its relative molar mass. Indeed, it is the only one of the hydrides of the 33
Water and acid-base cements elements from this portion of the periodic table to be liquid at room temperature and atmospheric pressure. In the gas phase it has a dipole moment that, while only slightly greater than that of hydrogenfluoride,is the highest for this group of hydrides. All of these properties point to water having a structure in which its constituent molecules are more highly associated and interact more strongly than the molecules of the closely related hydrides.
3.3
The structure of water
At first sight the concept of a 'structure' for liquid water appears strange. In the solid state atoms are relatively fixed in space, albeit with some vibrational motion about equilibrium positions, and no difficulty is associated with the idea of locating these equilibrium positions by some appropriate physical technique, and thereby assigning a structure to the solid.
3.3.1
The concept of structure in the liquid state
With water or any other liquid, molecules do not occupy even reasonably fixed locations but have considerably more freedom for movement than in the solid state. What then do we mean by the term structure applied to a liquid? To answer this question we need to consider the kind of physical techniques that are used to study the solid state. The main ones are based on diffraction, which may be of electrons, neutrons or X-rays (Moore, 1972; Franks, 1983). In all cases exposure of a crystalline solid to a beam of the particular type gives rise to a well-defined diffraction pattern, which by appropriate mathematical techniques can be interpreted to give information about the structure of the solid. When a liquid such as water is exposed to X-rays, electrons or neutrons, diffraction patterns are produced, though they have much less regularity and detail; it is also more difficult to interpret them than for solids. Such results are taken to show that liquids do, in fact, have some kind of long-range order which can justifiably be referred to as a 'structure'. In considering the structure of a liquid, two possible conceptual approaches exist. One is to begin from an understanding of the gaseous 34
The structure of water state, characterized as it is by gross translational movement of the constituent molecules and substantial disorder. The liquid is then viewed as a gas that has been condensed and in which translational motion has become constrained. Alternatively, consideration can start from the solid state, with its well-characterized structure, having little or no translational motion, but some vibrational motion of the constituent atoms or molecules. The liquid state is then viewed as a solid in which some degree of translational motion has become allowed, but with a structure still recognizable as being derived from that existing in the solid state (Franks, 1983). With the growth in application of the techniques of X-ray and neutron diffraction to the study of the liquid state, the latter approach has become increasingly favoured in recent years. In this section, rather than give a detailed account of theories of the liquid state, a more qualitative approach is adopted. What follows includes first a description of the structure of ice; then from that starting-point, ideas concerning the structure of liquid water are explained. 3.3.2
The structures of ice
Water is capable of solidifying into a number of different structural states or polymorphs depending, for example, on the external pressure applied during solidification. The simplest and most common of these polymorphs is known as ice I, whose structure was first determined by W. H. Bragg (1922). In this structure, every oxygen atom occupies the centre of a tetrahedron formed by four oxygen atoms, each about 0-276 nm away. The water molecules are connected together by hydrogen bonds, each molecule being bonded to its four nearest neighbours. The O-H bonds of a given molecule are oriented towards the lone pairs on two of these neighbouring molecules, and in turn, each of its lone pairs is directed towards an O-H bond of one of the other neighbours. This arrangement gives an open lattice in which intermolecular cohesion is large. The arrangement of oxygen atoms in ice I is isomorphous with the wurtzite form of zinc sulphide, and also with the silicon atoms in the tridymite form of silicon dioxide. Hence, ice I is sometimes referred to as the wurtzite or tridymite form of ice (Eisenberg & Kauzmann, 1969). Location of the hydrogen atoms in ice I has caused more problems. This is because hydrogen is less effective at scattering X-rays or electrons than oxygen. For a long time, arguments about the position of hydrogen were based on indirect evidence, such as vibrational spectra or estimates of 35
Water and acid-base cements residual entropy at 0 K (Eisenberg & Kauzmann, 1969). Since the advent of neutron diffraction the positions of the hydrogen atoms have become clearer. These studies have shown that the water molecules have very similar dimensions in ice I to those in the isolated molecule: the O-H bond length is 0-101 nm and the bond angle 104*5°. Ice I is one of at least nine polymorphic forms of ice. Ices II to VII are crystalline modifications of various types, formed at high pressures; ice VIII is a low-temperature modification of ice VII. Many of these polymorphs exist metastably at liquid nitrogen temperature and atmospheric pressure, and hence it has been possible to study their structures without undue difficulty. In addition to these crystalline polymorphs, socalled vitreous ice has been found within the low-temperaturefieldof ice I. It is not a polymorph, however, since it is a glass, i.e. a highly supercooled liquid. It is formed when water vapour condenses on surfaces cooled to below -160°C. It is not appropriate in this chapter to give a detailed review of the solidstate behaviour of water in its various crystalline modifications. However, there are some general structures which are relevant and worth highlighting. Firstly, water molecules in these various solids have dimensions and bond angles which do not differ much from those of an isolated water molecule. Secondly, the number of nearest neighbours to which each individual molecule is hydrogen-bonded remains four, regardless of the ice polymorph. The differences in structure between the polymorphs, particularly the high-pressure ones, lie in (a) the distances between the non-hydrogen bonded molecules, and hence the amount of' free volume' in the structure, (b) the angles of the hydrogen bonds, which may differ markedly from the 180° of ice I, and (c) the distance between nearest neighbouring oxygen atoms, which may fall to well below the 0-276 nm value in ice I. All of these are consistent with closer packing of the water molecules, and a closing up of the cage structure by comparison with that found for ice I. 3.3.3
Liquid water
Before considering the details of the structure of liquid water, it is important to define precisely what is meant by the term structure as applied to this liquid. If we start from ice I, in which molecules are vibrating about mean positions in a lattice, and apply heat, the molecules vibrate with greater energy. Gradually they become free to move from their original 36
The structure of water lattice sites and acquire significant translational energy. However, translational energy is not confined to molecules in the liquid state. There is a finite possibility of any molecule in ice I moving from its lattice site, thus acquiring translational energy. In principle, a given molecule can move through the solid structure in a process that is essentially diffusion. From this model of ice I we derive three meanings of the term structure for the solid. We may refer to the positions of the molecules at an instant of time. We may allow some averaging of the positions, i.e. we may have a vibrationally averaged structure, considered over a short time-period, during which molecules have time to undergo only minor vibrational reorientations. Finally we may have a diffusionally averaged structure, considered over longer time-periods, in which the minor translational motion has been allowed to proceed to such an extent as to be significant. These three possible structures, the instantaneous, the vibrationally averaged and the diffusionally averaged, are referred to as I-, V- and Dstructures respectively. Let us now turn our attention to liquid water. Just as in ice I, molecular motions may be divided into rapid vibrations and slower diffusional motions. In the liquid, however, vibrations are not centred on essentially fixed lattice sites, but around temporary equilibrium positions that are themselves subject to movement. Water at any instant may thus be considered to have an I-structure. An instant later, this I-structure will be modified as a result of vibrations, but not by any additional displacements of the molecules. This, together with the first I-structure, is one of the structures that may be averaged to allow for vibration, thereby contributing to the V-structure. Lastly, if we consider the structure around an individual water molecule over a long time-period, and realize that there is always some order in the arrangement of adjacent molecules in a liquid even over a reasonable duration, then we have the diffusionally averaged D-structure. No experimental technique exists for determining I-structures in either the liquid or the solid state. Techniques do exist for obtaining information on both the V- and D-structures of liquid water; the results of applying these techniques are considered next. Spectroscopic studies have established that for liquid water, the Vstructure has the following features. (a) Considerable local variation between the environments of the individual water molecules, compared with the relatively uniform 37
Water and acid-base cements molecular environments in a crystal of ice I. The frequency spans of the uncoupled O-H and O-D spectral bands indicate that some nearest neighbours are as close as 0-275 nm, while others are separated by 0.310 nm or more. The most probable equilibrium separation is about 0.285 nm (Eisenberg & Kauzmann, 1969). (b) The differences between the various molecular environments are continuous. In other words, the V-structure does not contain discrete types of molecular environment. (c) The frequency of the stretching band indicates that hydrogen bonds in the V-structure are weaker than those in ice I, though still distinctly present. Ideas about the D-structure have come mainly from two sources, namely a consideration of the underlying reasons for the values of certain physical properties, such as heat capacity or compressibility, and a study of radial distribution functions that arise from X-ray diffraction work on liquid water. The D-structure represents the average arrangement of molecules around an arbitrary central water molecule. This average is either the 'space average' for several central molecules in different V-structures, or the 'time average' for a single molecule over very long periods of time. Near the freezing point, the D-structure is found to have relatively high concentrations of neighbours at distances 0-29, 0-50 and 0*70 nm from the central water molecule. This suggests that a substantial hydrogen-bonded network is discernible, even in the liquid state. As the temperature is raised, so the distinct concentrations at 0-50 and 0-70 nm disappear. Thermal agitation thus distorts or destroys the hydrogen-bonded networks, and the amount of observable long-range order decreases significantly. Structural studies on liquid water reveal that the majority of molecules are effectively tetrahedral, since the O-H bonds and the lone pairs are used in hydrogen-bonding. Questions remain about the nature of these hydrogen-bonds (Symons, 1989). Specifically: on average, how many such hydrogen bonds are formed per molecule, how strong and how linear are they, and what is their lifetime? One recent approach has been to consider the possibility that, because of their weakness, some of the hydrogen bonds in liquid water will break. This then gives concentrations of free O-H bonds, OHfree, and free lone pairs, LPfree, on certain molecules which are bonded to only three others (Symons, 1989). Symons (1989) also suggests that the chemical properties of liquid water depend on the relative concentrations of these species. Fully hydrogen-bonded water can be 38
The structure of water considered as inert; reactions requiring attack by O-H depend on the concentration of OHfree molecules; and those requiring nucleophilic attack by lone pairs depend on the concentration of LPfree molecules. Evidence for OHfree and LPfree molecules has been obtained spectroscopically using monomeric deuterated water, HOD, in inert solvents such as dimethyl sulphoxide, though debate continues over interpretation of the results obtained in such studies. An important phenomenon when considering the differences between ice I and liquid water is that water achieves its maximum density not in the solid state, but at 4 °C, i.e. in the liquid state. The reasons for this were first discussed by Bernal & Fowler (1933). They noted that the separation of molecules in ice I is about 0-28 nm, corresponding to an effective molecular radius of 014 nm. Close packing of molecules of such radius would yield a substance of density 1*84 g cm"3. To account for the observed density of 10 g cm"3, it was necessary to postulate that the arrangement of molecules was very open compared with the disordered, close-packed structures of simple liquids such as argon and neon. The increase in density on melting is assumed to arise from two competing effects that occur as water is heated. First, increasing translational freedom for the water molecules weakens the hydrogen-bonded network that exists in ice I. This network thus collapses, and reduces the volume. Second, increased vibrational energy for the molecules causes an effective increase in the volume occupied by any one molecule, thus enlarging the overall volume of the liquid. The first effect is considered to predominate below 4 °C, the second above 4 °C. Overall, the main conclusions that are to be drawn concerning the structure of liquid water are as follows. (a) Water has a degree of long-range order that is appropriately described as structure; it is possible to measure detailed parameters for either a vibrationally averaged or a diffusionally averaged structure. (b) The force between molecules that sustains this order in the liquid state is the hydrogen bond. (c) The bond lengths and angles of individual water molecules are almost independent of whether they occur in ice I or liquid water.
39
Water and acid-base cements 3.4
Water as a solvent
The general criterion for solubility is the rule that 'like dissolves like'. In other words polar solvents dissolve polar and ionic solutes, non-polar solvents dissolve non-polar solutes. In the case of water, this means that ionic compounds such as sodium chloride and polar compounds such as sucrose are soluble, but non-polar compounds such as paraffin wax are not. In general, solubility depends on the relative magnitudes of three pairs of interactions, namely solute-solute, solvent-solvent and solute-solvent (Robb, 1983). For a substance to be soluble in a given liquid, the solute-solvent interactions must be greater than or equal to the other two interactions. Insolubility does not only result from the kind of energetic considerations outlined above. It can also be the result of essentially kinetic barriers. For example, the naturally occurring macromolecule cellulose is not soluble in water, yet its monomer, D( + )-glucose is extremely watersoluble (Morrison & Boyd, 1973). This is because cellulose adopts a wellordered structure, in which individual hydroxyl groups are aligned via hydrogen bonds; the overall structure simply has too great an integrity to allow water molecules to enter and hydrate the individual molecules in order to carry them off into solution. 3.4.1
Hydrophobic interactions
The qualitative discussion of solubility has focussed so far on the attractive forces in solute-solvent interactions. However, where water is concerned, it is also important to consider the forces of repulsion due to the so-called 'hydrophobic' interactions that may arise in certain cases (Franks, 1975). These hydrophobic interactions may be explained in terms of thermodynamic concepts. Measuring enthalpy changes for the dissolution of hydrocarbons, such as alkanes, in water shows that heat is evolved, i.e., AH is negative and energetically water and alkanes attract each other. However, such attraction does not make alkanes soluble in water to any appreciable extent. This is because the free energy change AGsolution opposes the process and is positive. From the Gibbs equation, Absolution — Absolution
40
1 Absolution
Water as a solvent it follows that the ^ASsolution term (and hence A*Ssolution itself) must be negative. This means that the proposed solution has lower entropy and is more ordered than pure water, which is a striking conclusion, since entropy is usually increased by mixing. It occurs because the relatively ordered structure of the liquid water, based as it is on a hydrogen-bonded array of water molecules, actually becomes more ordered when alkane molecules enter it. This result is attributed to the formation of a 'cage' structure of water molecules around the non-polar alkane molecule, in which water has less vibrational and translational freedom than in the pure liquid (Franks, 1983). In cases where the solvation energies are large, as for example when ionic compounds dissolve in water, these hydrophobic effects, based on adverse changes in entropy, are swamped. Dissolving such compounds can be readily accomplished due to the very large energies released when the ions become hydrated.
3.4.2
Dissolution of salts
Salts dissolve in water with dissociation of the constituent ions, this concept having been proposed originally by S. Arrhenius in 1887. His first idea was that all salts, including those of what would now be regarded as weak acids or bases, are completely dissociated at extreme dilution (Hall, 1985). It was eventually realized that substances such as NaCl, KC1, etc, are effectively completely dissociated at all concentrations. Dissolution of an ionic salt is essentially a separation process carried out by the interaction of the salt with water molecules. The separation is relatively easy in water because of its high dielectric constant. Comparison of the energies needed to separate ions of NaCl from 0-2 nm to infinity shows that it takes 692-89 kJ mol"1 in vacuum, but only 8-82 kJ mol"1 in aqueous solution (Moore, 1972). Similar arguments have been used to try to estimate solvation energies of ions in aqueous solution, but there are difficulties caused by the variations in dielectric constant in the immediate vicinity of individual ions. In order to dissolve ionic solutes so readily, water molecules must solvate the ions as they enter solution. Consequently, water molecules lose their translational degrees of freedom as a result of their association with specific ions. It is possible to estimate the number of water molecules in clusters of the type M+ (H2O)W using mass spectrometry (Kebarle, 1977). 41
Water and acid-base cements The number of water molecules in such a cluster, the hydration number, varies with ionic size; it is four for Li+, three for Na+, but only one for Rb+. Mass spectrometry has been used to study the energetics of solvation and has shown that the enthalpies of attachment of successive water molecules to either alkali metal or halide ions become less exothermic as the number of water molecules increases (Kebarle, 1977). The Gibbs free energies of attachment for water molecules have also been found to be negative. The different hydration numbers can have important effects on the solution behaviour of ions. For example, the sodium ion in ionic crystals has a mean radius of 0-095 nm, whereas the potassium ion has a mean radius of 0133 nm. In aqueous solution, these relative sizes are reversed, since the three water molecules clustered around the Na+ ion give it a radius of 0-24 nm, while the two water molecules around K+ give it a radius of only 0-17 nm (Moore, 1972). The presence of ions dissolved in water alters the translational freedom of certain molecules and has the effect of considerably modifying both the properties and structure of water in these solutions (Robinson & Stokes, 1955). The precise orientation of water molecules around cations is not clear, though two models have been proposed for the possible structures that occur (Vaslow, 1963). In one, the water molecules are arranged so that the dipole moments are aligned with the centres of the ions. In the other, water molecules are arranged so that interaction between the lone-pair orbitals on the oxygen atoms and orbitals on the cation is maximized. This latter model is supported by molecular dynamics calculations (Heinzinger & Palinkas, 1987; Heinje, Luck & Heinzinger, 1987). Less uncertainty surrounds the structure of hydrated anions: the hydrogen atoms are almost collinear with the oxygen atoms and the centres of the ions (Briant & Burton, 1976). Monte-Carlo calculations have shown that F~ is surrounded by four hydrogen atoms each 017 nm away (Watts, Clementi & Fromm, 1974). Neutron scattering has been used for studying the state of solvation of ions in aqueous solution (Enderby et ai, 1987; Salmon, Neilson & Enderby, 1988). These studies have shown that a distinct shell of water molecules of characteristic size surrounds each ion in solution. This immediate hydration shell was called zone A by Frank & Wen (1957); they also postulated the existence of a zone B, an outer sphere of molecules, less firmly attached, but forming part of the hydration layer around a given ion. The evidence for the existence of zone B lies in the thermodynamics of 42
Water as a solvent
the hydration process, and may be appreciated by considering the isoelectronic species KCl and two moles of argon. The standard enthalpy of hydration is significantly more exothermic for KCl than for the two moles of argon; however, the corresponding entropy of hydration is less for KCl than for argon. The results for the values of enthalpy can be readily understood in terms of the greater intensity of the interactions between water molecules and the ions of KCl than between those of water molecules and the uncharged argon atoms. At first sight the greater loss of freedom in the water molecules involved in hydrating the ions of KCl would be expected to reduce disorder in such solutions. In other words the entropy of hydration for KCl ought to be greater than the entropy of hydration for two moles of argon. To explain the fact that the opposite is found experimentally, Frank & Evans (1945) suggested that there is a compensating gain in entropy which can be attributed to disruptions in the water-water interactions within zone B. As a result of these electrostatic effects aqueous solutions of electrolytes behave in a way that is non-ideal. This non-ideality has been accounted for successfully in dilute solutions by application of the Debye-Huckel theory, which introduces the concept of ionic activity. The Debye-Huckel limiting law states that the mean ionic activity coefficient y± can be related to the charges on the ions, z+ and z_, by the equation Iog 10 y ± =-0-509z + z_ Ionic activity essentially represents the concentration of a particular type of ion in aqueous solution and is important in the accurate formulation of thermodynamic equations relating to aqueous solutions of electrolytes (Barrow, 1979). It replaces concentration because a given ion tends not to behave as a discrete entity but to gather a diffuse group of oppositely charged ions around it, a so-called ionic atmosphere. This means that the effective concentration of the original ion is less than its actual concentration, a fact which is reflected in the magnitude of the ionic activity coefficient. Debye-Hiickel theory assumes complete dissociation of electrolytes into solvated ions, and attributes ionic atmosphere formation to long-range physical forces of electrostatic attraction. The theory is adequate for describing the behaviour of strong 1:1 electrolytes in dilute aqueous solution but breaks down at higher concentrations. This is due to a chemical effect, namely that short-range electrostatic attraction occurs 43
Water and acid-base cements either between ion-pairs or between solvent-separated ion-pairs (Russo & Hanania, 1989); this effect becomes important in concentrated aqueous solutions of the type used to form AB cements. 3.4.3
Ion-ion interactions in water
Two ions are particularly important in the chemistry of water, namely H+ and OH~ (Clever, 1963). Hydrogen ions do not exist as discrete entities. This is because ionization of the hydrogen atom leaves behind a proton, which is very small compared with a typical ion. Thus the local charge density developed around the proton is very high. The polarizing effect of such a high charge density is such that the resulting system is simply too unstable to form to any detectable extent in aqueous solutions. When water undergoes self-ionization, a range of cationic species are formed, the simplest of which is the hydronium ion, H3O+ (Clever, 1963). This ion has been detected experimentally by a range of techniques including mass spectrometry (Cunningham, Payzant & Kebarle, 1972), as have ions of the type H+ (H2O)W with values of n up to 8. Monte-Carlo calculations show that H3O+ ions exist in hydrated clusters surrounded by three or four water molecules in the hydration shell (Kochanski, 1985). These ions have only a short lifetime, since the proton is highly mobile and may be readily transferred from one water molecule to another. The time taken for such a transfer is typically of the order of 10~14 s provided that the receiving molecule of water is correctly oriented. Several other discrete species have been found to arise from the selfionization of water. These include H5Og (Kearley, Pressman & Slade, 1986), H4O2+ (Bollinger et al., 1987), H9O+ (Robinson, Thistlethwaite & Lee, 1986) and hydrated electrons (Hart & Anbar, 1970). Intense ion-ion interactions which are characteristic of salt solutions occur in the concentrated aqueous solutions from which AB cements are prepared. As we have seen, in such solutions the simple Debye-Huckel limiting law that describes the strength goes up so the repulsive force between the ions becomes increasingly important. This is taken account of in the full Debye-Hiickel equation by the inclusion of a parameter related to ionic size and hence distance of closest approach (Marcus, 1988). For concentrated solutions, there are approaches that are more sophisticated than that of Debye & Hiickel. A particularly successful method of describing such solutions is that due to McMillan & Mayer (1945) which has subsequently been developed by Ramanathan & 44
Water as a solvent Friedman (1971). This approach is described as the hypernetted chain procedure and in it ion-ion pair potentials are expressed as the sum of four terms. These are: COUL U , the charge-charge interactions between two ions, / and j , as a function of their separation, COR U , a repulsive core potential term, CAVU, arising from the dielectric cavity effect, and GUR U , the so-called Gurney potential (Gurney, 1953), which describes the effect of co-sphere overlap. Using this approach, calculations can be made of volumetric, entropic and energy parameters taking account of the effect of overlapping cospheres. Some indication of the organization in the solution is also possible. The properties of a number of concentrated salt solutions have been analysed by this procedure, including simple 1:1 salts, alkaline earth salts and alkylammonium salts. A number of other attempts have been made to account for the properties of concentrated aqueous solutions of ionic compounds by procedures that represent further improvements on the simple DebyeHiickel approach. However, they lie outside the scope of the present chapter. The important point to emphasize is that the concentrated aqueous solutions that are generally employed in the preparation of AB cements tend to exhibit significant ion-ion interactions; such interactions lead to significant deviations from ideality which may be accounted for by substantial extension of the ideas of simple dilute solution theory. 3.4.4
Dissolution of polymers
AB cements are not only formulated from relatively small ions with well defined hydration numbers. They may also be prepared from macromolecules which dissolve in water to give multiply charged species known as polyelectrolytes. Cements which fall into this category are the zinc polycarboxylates and the glass-ionomers, the polyelectrolytes being poly(acrylic acid) or acrylic acid copolymers. The interaction of such polymers is a complicated topic, and one which is of wide importance to a number of scientific disciplines. Molyneux (1975) has highlighted the fact that these substances form the focal point of 'three complex and contentious territories of science', namely aqueous systems, ionic systems and polymeric systems. 45
Water and acid-base cements In polyelectrolytes, the ionic charges are carried by groups which are themselves attached covalently to the macromolecular backbone. When all of the groups are negatively charged, as with polyacrylate, the polyelectrolyte that results is referred to as a polyanion. Polyelectrolytes are of high solubility in water, especially when compared to most organic macromolecules. This increased solubility may be attributed not only to the more favourable interactions between the charged groups and the water molecules, but also to the fact that entropy strongly favours dissolution and dissociation of these molecules (Molyneux, 1975). The conformations adopted by polyelectrolytes under different conditions in aqueous solution have been the subject of much study. It is known, for example, that at low charge densities or at high ionic strengths polyelectrolytes have more or less randomly coiled conformations. As neutralization proceeds, with concomitant increase in charge density, so the polyelectrolyte chain uncoils due to electrostatic repulsion. Eventually at full neutralization such molecules have conformations that are essentially rod-like (Kitano et ai, 1980). This rod-like conformation for poly(acrylic acid) neutralized with sodium hydroxide in aqueous solution is not due to an increase in stiffness of the polymer, but to an increase in the so-called excluded volume, i.e. that region around an individual polymer molecule that cannot be entered by another molecule. The excluded volume itself increases due to an increase in electrostatic charge density (Kitano et al.9 1980). In a study of the transition in conformation from random coil to stiff rod by poly(acrylic acid), it was found that the point of transition depended on a number of factors, including the nature of the solvent, the temperature, the particular counterion used and the degree of dissociation (Klooster, van der Trouw & Mandel, 1984). Methanol was used in this study, though in terms of Flory-Huggins solution theory it is not a good solvent for poly(acrylic acid) at room temperature. In other words, the polymer adopts a tightly coiled conformation that excludes solvent, thereby approaching the point of precipitation. This phenomenon may be responsible for the observation that the addition of methanol to poly(acrylic acid) solutions intended for use in glass-ionomer cements prevents gelation (Wilson & Crisp, 1974). This was originally attributed to methylation of the polymer, leading to a reduction in the stereoregularity of the poly(acrylic acid) and hence to a lessening of the readiness with which stable hydrogen-bonded links could be formed. However, there is the alternative possibility that the presence of 46
Hydration in the solid state methanol altered the conformation of the polymer and that this conformational change prevented the development of the hydrogen-bonded network necessary for gelation to occur. 5.5
Hydration in the solid state
Many ionic compounds contain what used to be referred to as 'water of crystallization'. For example, magnesium chloride can exist as a fully hydrated salt which was formerly written MgCl2. 6H2O, but is more appropriately written Mg(OH2)6Cl2, since the water molecules occupy coordination sites around the magnesium ions. This is typical. In most compounds that contain water of crystallization, the water molecules are bound to the cation in an aquo complex in the manner originally proposed by Alfred Werner (1866-1919) in 1893 (Kauffman, 1981). Such an arrangement has been confirmed in numerous cases by X-ray diffraction techniques. 3.5.1
Coordination of water to ions
The ions that tend to be involved in AB cements include such species as Al3+, Mg2+, Ca2+ and Zn2+. These are all capable of developing a coordination number of six, and hexaquo cations are known to be formed by each of these metal ions (Hiickel, 1950). The typical requirements for an ion to develop such coordination characteristics are that the ion should exist in the + 2 or +3 oxidation state, and in this state should be of small ionic radius (Greenwood & Earnshaw, 1984). Another feature of the metal ions that are typically involved in cementitious bonding in AB cements is that most of them fall into the category of hard in Pearson's Hard and Soft Acids and Bases scheme (Pearson, 1963). The underlying principle of this classification is that bases may be divided into two categories, namely those that are polarizable or soft, and those that are non-polarizable or hard. Lewis acids too may be essentially divided into hard and soft, depending on polarizability. From these classifications emerges the useful generalization that hard acids prefer to associate with hajd bases and soft acids prefer to associate with soft bases (see Section 2.3.7). Of the ions most often implicated in cementitious bonding in AB cements, Ca2+, Mg2+ and Al3+ are classified as clearly hard; Zn2+ by contrast falls into the category that Pearson designated 'borderline', as 47
Water and acid-base cements does Cu2+ (Pearson, 1963). This means that most of these ions form particularly stable complexes with hard bases, i.e. those which are not readily polarizable. This requirement demands that the bases be those of first row elements, such as oxygen or nitrogen. Water is thus a hard base, and the complex that it forms with these ions involves coordination by the oxygen atoms. As predicted by the Hard and Soft Acids and Bases concept, the aquo complexes of the cement-forming metal ions are extremely stable and do not readily lose their coordinated water. Hence, one of the functions of water in fully set AB cements is coordination to the metal ions. 3.6
The role of water in acid-base cements
Water has three possible roles in acid-base cements. First, it acts as the medium for the setting reaction of these materials, and second, it is one of the components of the set cement, actually becoming incorporated into the cement as it hardens. Third, water may act as plasticizer in these cements. All of these roles are reviewed here. 3.6.1
Water as solvent in AB cements
Water as the solvent is essential for the acid-base setting reaction to occur. Indeed, as was shown in Chapter 2, our very understanding of the terms ' acid' and' base', at least as established by the Bronsted-Lowry definition, requires that water be the medium of reaction. Water is needed so that the acids may dissociate, in principle to yield protons, thereby enabling the property of acidity to be manifested. The polarity of water enables the various metal ions to enter the liquid phase and thus react. The solubility and extent of hydration of the various species change as the reaction proceeds, and these changes contribute to the setting of the cement. 3.6.2
Water as a component of AB cements
Water is also a component of set AB cements. In glass-ionomer cements, for example, it may serve to coordinate to certain sites around the metal ions. It also hydrates the siliceous hydrogel that is formed from the glass after acid attack has liberated the various metal ions (Wilson & McLean, 1988). Such reactions continue long after the initial hardening of the cement is complete, and for this reason water must be retained as far as possible during the first hours and days after formation of the cement. If water is lost from the cement and desiccation occurs, these post-hardening 48
The role of water in acid-base cements reactions stop, and the cement will not achieve maximum possible strength. Moreover, if desiccation is excessive, the material will also shrink and crack. Water occurs in glass-ionomer and related cements in at least two different states (Wilson & McLean, 1988; Prosser & Wilson, 1979). These states have been classified as evaporable and non-evaporable, depending on whether the water can be removed by vacuum desiccation over silica gel or whether it remains firmly bound in the cement when subjected to such treatment (Wilson & Crisp, 1975). The alternative descriptions 'loosely bound' and 'tightly bound' have also been applied to these different states of water combination. In the glass-poly(acrylic acid) system the evaporable water is up to 5 % by weight of the total cement, while the bound water is 18-28 % (Prosser & Wilson, 1979). This amount of tightly bound water is equivalent to five or six molecules of water for each acid group and associated metal cation. Hence at least ten molecules of water are involved in the hydration of each coordinated metal ion at a carboxylate site. It has been suggested by Ikegami (1968) that the carboxylate groups of a polyacrylate chain are each surrounded by a primary local sphere of oriented water molecules, and that the polyacrylate chain itself is surrounded by a secondary sheath of water molecules. This secondary sheath is maintained as a result of the cooperative action of the charged functional groups on the backbone of the molecule. The monovalent ions Li+, Na+ and K+ are able to penetrate only this secondary hydration sheath, and thereby form a solvent-separated ion-pair, rather than a contact ion-pair. Divalent ions, such as Mg2+ or Ba2+, cause a much greater disruption to the secondary hydration sheath. The effectiveness with which divalent ions cause gelation of poly(acrylic acid) has been found to follow the order Ba2+ > Sr2+ > Ca2+ (Wall & Drenan, 1951) and this has been attributed to the formation of salt-like crosslinks. Gelation has been assumed to arise in part from dehydration of the ion-pairs (Ikegami & Imai, 1962), and certainly correlates with precipitation in fairly dilute systems. Indeed, the term precipitation has sometimes been applied to the setting of AB cements derived from poly(acrylic acid) as they undergo the transition from soft manipulable paste to hard brittle solid. At the molecular level, a number of features are associated with the phenomenon of gelation or precipitation. In particular the disruption of the secondary hydration sheaths around the polyacrylate chains appears 49
Water and acid-base cements important (Prosser & Wilson, 1979). In glass-ionomer cements the bound water has been assumed to be associated with the intrinsic water spheres around the carboxylate anion-metal cation units, while evaporable water is associated with the secondary hydration sheath around the polyacrylate chain. As these cements age, the ratio of tightly bound to loosely bound water increases. This is accompanied by an increase in strength and modulus of the cement and by a decrease in plasticity (Paddon & Wilson, 1976; Wilson, Crisp & Paddon, 1981). The loosely bound water in glass-ionomer cements is labile, and is easily lost or gained. Indeed, such cements are stable only in an atmosphere of 80% relative humidity (Hornsby, 1980). In higher humidities the cements absorb water and the resulting hydroscopic expansion can exceed the shrinkage that usually accompanies setting, which is a distinct clinical advantage for the use of these cements in dentistry. By contrast, the cement can lose water under drying conditions leading to shrinking, crazing and failure to develop full strength. Glass-ionomer cements become less susceptible to desiccation as they age, because a greater proportion of the water in older cements has become 'tightly bound'. Early contact with moisture is also damaging, and this problem is overcome clinically to some extent by using some sort of protection such as clear nail varnish to seal the cement during its early life (Wilson & McLean, 1988). However, this does not give perfect results, and as yet there is no ideal barrier material for this purpose (Earl, Hume & Mount, 1985). The role of water in dental silicate cements was studied by Wilson et al. (1970), and they found that the properties of these materials including setting time, compressive strength and resistance to attack by water and acids were markedly affected by the amount of water in the original cement paste. Water in these materials was also found to fall into the two categories of evaporable and non-evaporable. In this case non-evaporable water was defined as that water remaining in the cement after heating at 105 °C for 24 hours. With increasing acid concentration (i.e. decreasing amounts of water in the initial cement paste), the amount of nonevaporable water went down, until at 75 % phosphoric acid concentration nearly all of the water in the cement was found to be evaporable. Moreover the cements containing almost no non-evaporable water were found to be extremely weak. Hence the non-evaporable (bound) water could be equated with bonding water. Infrared analysis had previously shown that non-bonding water was associated with the water-soluble hydrated salt 50
The role of water in acid-base cements sodium dihydrogen phosphate, NaH 2 PO 4 .H 2 O (Wilson & Mesley, 1968). The presence of this compound was known to have a deleterious effect on cement properties and the water associated with it was known to be readily removed. Hence in cements prepared from aqueous solutions having high concentrations of phosphoric acid this salt was assumed to be present in quantity and to be responsible for the relatively high levels of evaporable water (Wilson et aL, 1970). A series of AB cements can be prepared from aqueous solutions of oxides and halides (or sulphates) of magnesium or zinc. These cements are described in detail in Chapter 7. For the moment we will confine our discussion to a consideration of the role of water in these cements. In the cements of this type a number of phases are known to be present. For example, in the zinc oxychloride cement two discrete phases, corresponding to the composition ZnO. ZnCl 2 . H 2 O in the ratios 4:1:5 and 1:1:2 respectively, are known to occur (Sorrell, 1977). Similarly, in the magnesium oxychloride cement, phases corresponding to Mg(OH) 2 . MgCl 2 . H 2 O in the ratios 5:1:8 and 3:1:8 have been shown to exist and have been studied by X-ray diffractometry (Sorrell & Armstrong, 1976). The precise structural role played by the water molecules in these cements is not clear. In the zinc oxychloride cement, water is known to be thermally labile. The 1:1:2 phase will lose half of its constituent water at about 230 °C, and the 4:1:5 phase will lose water at approximately 160 °C to yield a mixture of zinc oxide and the 1:1:2 phase. Water clearly occurs in these cements as discrete molecules, which presumably coordinate to the metal ions in the cements in the way described previously. However, the possible complexities of structure for these systems, which may include chlorine atoms in bridging positions between pairs of metal atoms, make it impossible to suggest with any degree of confidence which chemical species or what structural units are likely to be present in such cements. One is left with the rather inadequate chemical descriptions of the phases used in even the relatively recent original literature on these materials, from which no clear information on the role of water can be deduced. 3.6.3
Water as plasticizer
An additional possible role for water in AB cements is plasticization. Water is known to act as plasticizer for a number of polymeric materials, whether synthetic or natural, and whether or not they are predominantly 51
Water and acid-base cements polar. Thus water has been found to affect the properties of poly(methyl methacrylate) (Turner, 1982) and of alkyd resins used in surface coatings (Mayne & Mills, 1982). As plasticizer, the principal effect of water in these systems is to reduce the glass transition temperature, Tg, and this in turn affects a number of other properties of the materials, including rigidity, dimensional stability and diffusion coefficients within the bulk. Given the polar nature of the components of AB cements, the known water content of set cements, and the fact that water has been shown to act as plasticizer in pure metal poly(acrylate) salts (Yokoyama & Hiraoko, 1979) it seems probable that one of the roles of water in the solid state of AB cements is that of plasticizer. References Barrow, G. M. (1979). Physical Chemistry, 4th edn. Tokyo: McGraw-Hill Kogakusha Ltd. Benedict, W. S., Gailar, N. & Plyler, E. K. (1956). Rotation-vibration spectra of deuterated water vapour. Journal of Chemical Physics, 24, 1139-65. Bernal, J. D. & Fowler, R. H. (1933). A theory of water and ionic solutions with particular reference to hydrogen and hydroxyl ions. Journal of Chemical Physics, 1, 515-48. Bollinger, J. C , Faure, R., Yvernault, T. & Stahl, D. (1987). On the existence of the protonated dication H4 O2+ in sulfolane solution. Chemical Physics Letters, 140, 579-81. Bragg, W. H. (1922). The crystal structure of ice. Proceedings of the Physical Society of London, 34, 98-103. Briant, C. L. & Burton, J. J. (1976). Molecular dynamics study of the effects of ions on water microclusters. Journal of Chemical Physics, 64, 2888-95. Clever, H. L. (1963). The hydrated hydronium ion. Journal of Chemical Education, 40, 637^41. Cunningham, A. J. C , Payzant, J. D. & Kebarle, P. (1972). A kinetic study of the proton hydrate H+(H2O) and equilibria in the gas phase. Journal of the American Chemical Society, 94, 7627-32. Earl, M. S. A., Hume, W. R. & Mount, G. J. (1985). Effect of varnishes and other surface treatments on water movement across the glass-ionomer cement surface. Australian Dental Journal, 30, 298-301. Eisenberg, D. & Kauzmann, W. (1969). The Structure and Properties of Water. Oxford: Oxford University Press. Enderby, J. E., Cummings, S., Herdman, G. H., Neilson, G. W., Salmon, P. S. & Skipper, N. (1987). Diffraction and study of aqua ions. Journal of Physical Chemistry, 91, 5851-8. Frank, H. S. & Evans, M. W. (1945). Entropy in binary liquid mixtures; partial molal entropy in dilute solutions; structure and thermodynamics in aqueous electrolytes. Journal of Chemical Physics, 13, 507-32. 52
References Frank, H. S. & Wen, W-Y. (1957). Structural aspects of ion-solvent interactions in aqueous solutions-water structure. Discussions of the Faraday Society, 24, 133^0. Franks, F. (1975). The hydrophobic interaction. In Franks, F. Water. A Comprehensive Treatise, vol. 4, Chapter 1. London and New York: Plenum Press. Franks, F. (1983). Water. London: Royal Society of Chemistry. Gillespie, R. J. & Nyholm, R. S. (1957). Inorganic stereochemistry. Quarterly Reviews of the Chemical Society, 11, 339-80. Greenwood, N. N. & Earnshaw, A. (1984). The Chemistry of the Elements. Oxford: Pergamon Press. Gurney, R. W. (1953). Ionic Processes in Solution. New York: McGraw-Hill. Hall, V. M. D. (1985). In Russell, C. A. (ed.) Recent Developments in the History of Chemistry. London: Royal Society of Chemistry. Hart, E. J. & Anbar, M. (1970). The HydratedElectron. New York: Wiley. Heinzinger, K. & Palinkas, G. (1987). In Kleeberg, H. (ed.) Interactions of Water in Non-ionic Hydrates. Berlin: Springer-Verlag. Heinje, G., Luck, W. A. P. & Heinzinger, K. (1987). Molecular dynamics simulation of an aqueous sodium perchlorate solution. Journal of Physical Chemistry, 91, 331-8. Hornsby, P. R. (1980). Dimensional stability of glass-ionomer cements. Journal of Chemical Technology and Biotechnology, 30, 595-601. Hiickel, W. (1950). Structural Chemistry of Inorganic Compounds, vol. 1. New York: Elsevier. Ikegami, A. (1968). Hydration of polyacids. Biopolymers, 6, 431-40. Ikegami, A. & Imai, N. (1962). Precipitation of polyelectrolytes by salts. Journal of Polymer Science, 56, 133-52. Kauffman, G. B. (1981). Coordination Chemistry. New York: Heyden. Kearley, G. J., Pressman, H. A. & Slade, R. C. T. (1986). The geometry of the H5OJ ion in dodecatungstophosphoric acid hexahydrate, (H5O£)3 (PW^O^), studied by inelastic neutron scattering vibrational spectroscopy. Journal of the Chemical Society Chemical Communications, 1801-2. Kebarle, P. (1977). Ion thermochemistry and solvation from gas phase ion equilibria. Annual Reviews in Physical Chemistry, 28, 445-76. Kitano, T., Taguchi, A., Noda, I. & Nagasawa, M. (1980). Conformation of polyelectrolytes in aqueous solution. Macromolecules, 13, 57-63. Klooster, N. Th. M., van der Trouw, F. & Mandel, M. (1984). Solvent effects in polyelectrolyte solutions. 3. Spectropho tome trie results with (partially) neutralised poly(acrylic acid) in methanol and general conclusions regarding these systems. Macromolecules, 17, 2087-93. Kochanski, E. (1985). Theoretical studies of the system H3O+(H2O)n for n = 1—9. Journal of the American Chemical Society, 107, 7869-73.
Marcus, Y. (1988). Ionic radii in aqueous solution. Chemical Reviews, 88, 1475-98. Mayne, J. E. O. & Mills, D. J. (1982). Structural changes in polymer films. Part 1. The influence of the transition temperature on the electrolytic resistance 53
Water and acid-base cements and water uptake. Journal of the Oil and Colour Chemists' Association, 65, 138^2. McMillan, W. G. & Mayer, J. E. (1945). The statistical thermodynamics of multicomponent systems. Journal of Chemical Physics, 13, 276-305. Molyneux, P. (1975). Synthetic polymers. In Franks, F. Water. A Comprehensive Treatise, vol. 4, Chapter 7. London and New York: Plenum Press. Moore, W. J. (1972). Physical Chemistry, 5th edn. London: Longman Group Ltd. Morrison, R. T. & Boyd, R. T. (1973). Organic Chemistry, 3rd edn. New York: Allyn and Bacon. Neilson, G. W., Schioberg, D. & Luck, W. A. P. (1985). The structure around the perchlorate ion in concentrated aqueous solutions. Chemical Physics Letters, 111, 475-9. Nicholson, J. W. (1985). Waterborne Coatings. OCCA Monograph No. 3. London: Oil and Colour Chemists' Association. Paddon, J. M. & Wilson, A. D. (1976). Stress relaxation studies on dental materials. 1. Dental cements. Journal of Dentistry, 4, 183-9. Pauling, L. (1960). The Nature of the Chemical Bond, 3rd edn. Ithaca, New York: Cornell University Press. Pearson, R. G. (1963). Hard and soft acids and bases. Journal of the American Chemical Society, 85, 3533-9. Prosser, H. J. & Wilson, A. D. (1979). Litho-ionomer cements and their technological applications. Journal of Chemical Technology and Biotechnology, 29, 69-87. Ramanathan, P. S. & Friedman, H. L. (1971). Refined model for aqueous 1-1 electrolytes. Journal of Chemical Physics, 54, 1086-99. Robinson, R. A. & Stokes, R. H. (1955). Electrolyte Solutions. London: Butterworth Scientific Publications. Robinson, G. W., Thistlethwaite, P. J. & Lee, J. (1986). Molecular aspects of ionic hydration reactions. Journal of Physical Chemistry, 90, 4224-33. Robb, I. D. (1983). Polymer-small molecule interactions. In Finch, C. A. (ed.) Chemistry and Technology of Water Soluble Polymers. New York: Plenum Press. Russo, S. O. & Hanania, G. I. H. (1989). Ion association solubilities and reduction potentials in aqueous solution. Journal of Chemical Education, 66, 148-53. Salmon, P. S., Neilson, G. W. & Enderby, J. E. (1988). The structure of Cu2+ aqueous solutions. Journal of Physics C, 21, 1335-49. Sorrell, C. A. & Armstrong, C. R. (1976). Reactions and equilibria in magnesium oxychloride cements. Journal of the American Ceramic Society, 59, 51-4. Sorrell, C. A. (1977). Suggested chemistry of zinc oxychloride cements. Journal of the American Ceramic Society, 60, 217-20. Speakman, J. C. (1975). The Hydrogen Bond. London: Chemical Society. Symons, M. C. R. (1989). Liquid water- the story unfolds. Chemistry in Britain, 25, 491-4. 54
References Turner, D. T. (1982). Poly(methyl methacrylate) plus water. Sorption kinetics and volumetric changes. Polymer, 23, 197-202. Vaslow, F. (1963). The orientation of water molecules in the field of an alkali ion. Journal of Physical Chemistry, 67, 2773-6. Wall, F. T. & Drenan, J. W. (1951). Gelation of polyacrylic acid by divalent ions. Journal of Polymer Science, 1, 83-8. Watts, R. O., Clementi, E. & Fromm, J. (1974). Theoretical study of the lithium fluoride molecule in water. Journal of Chemical Physics, 61, 2550-5. Weast, R. C. (ed.) (1985-6). Handbook of Physics and Chemistry. Ohio: Chemical Rubber Company. Wilson, A. D. & Crisp, S. (1974). Unpublished data cited in Wilson & McLean, 1988. Wilson, A. D. & Crisp, S. (1975). Ionomer cements. British Polymer Journal, 1, 279-96. Wilson, A. D., Crisp, S. & Paddon, J. M. (1981). The hydration of a glass-ionomer (ASPA) cement. British Polymer Journal, 13, 66-70. Wilson, A. D., Kent, B. E., Batchelor, R. F., Scott, B. G. & Lewis, B. G. (1970). Dental silicate cements. XII. The role of water. Journal of Dental Research, 49, 307-14. Wilson, A. D. & McLean, J. W. (1988). Glass-ionomer Cement. Chicago, London, etc.: Quintessence Publishers. Wilson, A. D. & Mesley, R. J. (1968). Dental silicate cements. VI. Infrared studies. Journal of Dental Research, 47, 644—52. Yokoyama, T. & Hiraoko, K. (1979). Hydration and thermal transition of poly(acrylic acid) salts. Polymer Preprints of the American Chemical Society, Division of Polymer Chemistry, 20, 511-13.
55
4
Polyelectrolytes, ion binding and gelation
4.1
Polyelectrolytes
4.1.1
General
The setting of AB cements is an example of gelation, and gelation is related to ion binding. A theoretical examination of the various phenomena associated with ion binding and gelation finds its clearest exposition in the field of polyelectrolytes. Moreover, this field may be wider than it seems at first. Polyelectrolytes form the basis of those modern cements which are distinguished by their ability to adhere to reactive surfaces. At present the main use of such cements lies in the medical field, principally in dental surgery. They adhere permanently to biological surfaces where they have to withstand adverse conditions of wetness, chemical attack, the stress of biological activity, and chemical and biological changes within the substrate. Nevertheless, adhesive bonds are maintained. Polyelectrolytes are polymers having a multiplicity of ionizable groups. In solution, they dissociate into polyions (or macroions) and small ions of the opposite charge, known as counterions. The polyelectrolytes of interest in this book are those where the polyion is an anion and the counterions are cations. Some typical anionic polyelectrolytes are depicted in Figure 4.1. Of principal interest are the homopolymers of acrylic acid and its copolymers with e.g. itaconic and maleic acids. These are used in the zinc polycarboxylate cement of Smith (1968) and the glass-ionomer cement of Wilson & Kent (1971). More recently, Wilson & Ellis (1989) and Ellis & Wilson (1990) have described cements based on polyphosphonic acids. There is also the question of whether there are inorganic polyelectrolytes within the field of AB cements. A number of cements are based on orthophosphoric acid, and in the two most important ones aluminium is 56
Polyelectrolytes known to be essential for cement formation. Aluminium forms complex aluminophosphoric acids with orthophosphoric acid. The solutions of these complexes are markedly viscous and there is some NMR spectroscopic evidence that these aluminophosphoric acids form linear polymers based on the Al-O-P linkage (Sveshnikova & Zaitseva, 1964; Akitt, Greenwood & Lester, 1971; O'Neill et al., 1982). Callis, Van Wazer & Arvan (1954), Salmon & Wall (1958) and Wilson et al. (1972) consider that aluminophosphate polymers are formed in which [POJ tetrahedra are linked by aluminium atoms. Polyanion chains containing many linked charged groups exert a considerable electrostatic effect on the orientation of dipolar solvent molecules and on the counterions. The counterions are constrained to remain in the neighbourhood of the charged polymer chains, a phenomenon known as ion binding. This phenomenon is supported by a wealth of experimental evidence (Morawetz, 1975; Wilson & Crisp, 1977; Rymden & Stilbs, 1985a, b) and an early illustration of it is found in the work of Huizenga, Grieger & Wall (1950a, b) who observed that, in an electric field, cations were sometimes transported with the polyanion.
r
CH2
2
CH2
CH—COOH
T
CH—COOH Poly (acrylic acid)
CH2
1
1
CH—SO2OH CH2 CH—SO 2 OH
Poly (vinyl sulphonic acid)
C H — COO" 1
C H — COO~
1
Polyacrylate
1 1 CH—PO(OH)
CH 2
2
T2
CH—PO(OH) 2
Poly (vinyl phosphonic acid)
Figure 4.1 Some typical anionic polyelectrolytes.
57
Polyelectrolytes, ion binding and gelation 4.1.2
Polyion conformation
The shape, configuration or morphology of a polyion is usually known as its conformation. There are very many possible conformations available to a polyion because of the flexibility of the main chain due to the free rotation of bonds. The particular conformation adopted will be the one with the lowest free energy. This free energy has two components, one arising from chainflexibilityand the other from electrical interactions. The intrinsic free energy of rotation is a function of the relative position of neighbouring bonds. There are energy minima, at the trans position, which corresponds to the stretched form of the chain, and at the two gauche positions, corresponding to the contracted form. The difference in energy between the trans and gauche positions is one important factor determining theflexibilityof the chain. The other component of free energy arises from interactions between charged groups on the polyion, counterions and solvent molecules. There are two broad kinds of polyion conformation; the random coil and the ordered helix. In a helix there are regularly repeated structures along the coil; there are none in the case of a random coil. In this book we are concerned with the latter where there are often several conformations with approximately equal free energies and, thus, conformational changes occur readily. Random coil conformations can range from the spherical contracted state to the fully extended cylindrical or rod-like form. The conformation adopted depends on the charge on the polyion and the effect of the counterions. When the charge is low the conformation is that of a contracted random coil. As the charge increases the chains extend under the influence of mutually repulsive forces to a rod-like form (Jacobsen, 1962). Thus, as a weak polyelectrolyte acid is neutralized, its conformation changes from that of a compact random coil to an extended chain. For example poly(acrylic acid), degree of polymerization 1000, adopts a spherical form with a radius of 20 nm at low pH. As neutralization proceeds the polyion first extends spherically and then becomes rod-like with a maximum extension of 250 nm (Oosawa, 1971). These pHdependent conformational changes are important to the chemistry of polyelectrolyte cements. The situation is more complex in the reactions found in AB cements because neutralization is accompanied by ion binding. Although a polyion chain extends as the number of ionized groups increases, the binding of 58
Ion binding counterions has the reverse effect because intrachain repulsive forces are decreased. An increase in the concentration of polyions in solution has the same effect, for an increase in interchain repulsion will inhibit the unwinding of polymer chains. Thus, the effects predicted by dilute solution theory will be much less in the concentrated conditions found in AB cements. From this it can be seen that the effect of ion binding on conformation change is complex. Conversely, conformation affects the binding of counterions to polyions (Jacobsen, 1962). In the compact spherical conformation some ionized groups on polymer chains will be inaccessible for ion binding. 4.2
Ion binding
4.2.1
Counterion binding
Oppositely charged ions are attracted to each other by electrostatic forces and so will not be distributed uniformly in solution. Around each ion or polyion there is a predominance of ions of the opposite charge, the counterions. This cloud of counterions is the ionic atmosphere of the polyion. In a dynamic situation, the distribution of counterions depends on competition between the electrostatic binding forces and the opposing, disruptive effects of thermal agitation. The phenomenon has been studied by a number of techniques: titration (Gregor & Frederick, 1957; Kagawa & Gregor, 1957); viscosity and electrical conductance measurements (Gregor, Gold & Frederick, 1957; Bratko et al., 1983); determination of counterion activity (Kagawa & Katsuura, 1955); measurement of transference (Ferry & Gill, 1962); dilatometry (Strauss & Leung, 1965; Begala & Strauss, 1972); and NMR spectroscopy (Rymden & Stilbs, 1985a, b). Ion binding is affected by the size and charge of the counterion, the charge and conformation of the polyion, and states of hydration. We will examine these effects in some detail. 4.2.2
The distribution of counterions
The potential distribution around the polyion is important to any discussion of counterion binding and hydration effects. Oosawa (1971) has distinguished four regions of potential about a polyion (Figure 4.2): (1) a localized potential hole around each charged group, (2) a cylindrical potential valley or tube along the polyion chain, (3) a spherical trough in 59
Poly electrolytes, ion binding and gelation
the apparent volume occupied by the whole of the coiled chain, and (4) the region outside the polyion. Counterions are distributed between these four potential regions and may be classified as free, bound but mobile (atmospheric) and localized {site-bound). Free ions remain outside the volume of the polyion (in region 4); the remaining ions are bound to the polyion. Of the bound ions, the mobile atmospheric ions occupy the potential trough or valley around each polyion (regions 2 and 3). Localized binding occurs in the potential holes at the sites of the individual charged groups of the polyion, and ion-pairs are formed. Oosawa (1971) used a simple calculation to illustrate the effect of a highly charged polyion on the binding counterions. The distribution between free ions and bound ions depends on the ratio between potential energy and kinetic energy. In the case of a random coil, containing n ionized groups of charge — e0, and of spherical conformation, radius/?, the potential drop, Si//, for a counterion of charge + e0 at the edge of the polyion is given by neo/eop
\ \
r
Region of free counterions
Figure 4.2 The four regions of potential about a polyion. Based on Oosawa (1971).
60
(4.1)
Ion binding
where the dielectric constant of the solvent is equal to e0. Hence the potential energy is nel/eop
(4.2)
The ratio of the potential energy to the kinetic energy, kT, is ne*/eopkT
(4.3)
This ratio is related linearly to the degree of polymerization n. In the case of a poly(acrylic acid) where n = 1000 and p = 20 nm, this ratio works out at 35. Thus, many of the counterions must enter the region of the polyion. Even when 90 % of the counterions are within the polyion this ratio is still high with a value of 3-5. A similar calculation for the rod-like random coil gives an energy ratio of 26 and similar arguments apply. Oosawa (1971) developed a simple mathematical model, using an approximate treatment, to describe the distribution of counterions. We shall use it here as it offers a clear qualitative description of the phenomenon, uncluttered by heavy mathematics associated with the Poisson-Boltzmann equation. Oosawa assumed that there were two phases, one occupied by the polyions, and the other external to them. He also assumed that each contained a uniform distribution of counterions. This is an approximation to the situation where distribution is governed by the Poisson distribution (Atkins, 1978). If the proportion of site-bound ions is negligible, the distribution of counterions between these phases is then given by the Boltzmann distribution, which relates the population ratio of two groups of atoms or ions to the energy difference between them. Thus, for monovalent counterions nJK = (nJV^xpi-e^/kT)
(4.4)
where Sy/ is the average potential difference between the two phases, nh is the number of bound ions contained in a total volume Vh9 nt is the number of free ions contained in a total volume Vt and n is the total number of counterions in a total volume V. This case can be rewritten In {(1-/?)/# = ln{0/(l-0)}-e o l for Q ^ 1
(4.12)
£-> \/Q and y-> \/Q for Q ^ 1
(4.13)
where y is the activity coefficient and equals /?/(l — 0). The consequences of these solutions are shown in Figure 4.4. The abscissa is n, the total number of counterions or charged groups on the polyion, and is proportional to Q. Along the ordinate are the numbers of counterions bound, nh, and free, nt, equal to n(\—fi) and n^respectively. The increase of counterion binding with the charge on the polyion has 63
Polyelectrolytes, ion binding and gelation
been termed counterion condensation as it is analogous to the condensation of a vapour. This point is illustrated by Figure 4.4. As the charge on the polyion increases from zero there is a proportional increase in the number of counterions. At first, all the counterions are free and none are bound. This continues until a plateau is reached at a critical value of Q = 1. Above this point, all additional counterions are bound to the polyion and the number of free ions remains constant. Thus, PQ remains constant and /? decreases as Q increases. The simple situation depicted in Figure 4.4 is a limiting one, and the discontinuity does not occur when 6 > 0; however, for large values of g, PQ increases only slowly. Nor does the discontinuity appear in the case of the spherical conformation, but again, for large values of P, PP increases only logarithmically. Thus, the situation is similar to that for the rod-like configuration although there is no specific critical value for P. The above treatment is based on the assumption that 9 is small. However, as Figure 4.3 shows, /? does not greatly change with concentration so that counterion condensation is probably insensitive to concentration. The delayed binding of counterions is of some importance to the onset of gelation.
bound / counterions, / n
b
a
7
H
C r—t
a c
/ free / counterions, / n f / / / / :Q = i / / / / / : / • / ;
/
Total number of counterions, n[Q] Figure 4.4 The abscissa is n, the total number of counterions or charged groups on the polyion, and is proportional to Q. Along the ordinate are the number of counterions bound, nh, and free, «f, equal to n(\ —f$) and «/? respectively. Based on Oosawa (1971).
64
Ion binding Theories of counterion condensation have been reviewed by Manning (1979, 1981); and Satoh, Komiyama & Iijima (1984) have extended the theory.
4.2.4
Effect of valence and size on counterion binding
Cations of small ionic radius and high charge are more firmly bound than monovalent ions of large ionic radius (Ikegami & Imai, 1962; Strauss & Leung, 1965; Begala & Strauss, 1972). Divalent ions are more strongly bound than monovalent ions and the interaction is often localized. This can be examined theoretically by applying Oosawa's two-phase model to counterions with a valence z and charge of + e0, and a polyion with a total charge — ne. Equations (4.5) and (4.8), which were developed for univalent ions, can be rewritten, thus: In {(1-/?)//?} = \n{e/(\-e)}-zeQ5¥/kT
(4.14) (4.15)
The critical value for Q is 1/z. There is a proportional increase in the number of free counterions, njz, as Q increases from zero, reaching a plateau when Q = 1/z. Also, below this value the degree of dissociation, /?, increases as the concentration decreases, and tends to unity as v tends to zero. When Q > 1/z, f5 decreases with 9 and tends to 1/zQ as 6 tends to zero. The number of free ions cannot exceed n/z2Q. Note that this number is inversely proportional to the square of the valence. The condensation of ions is thus very sensitive to valence; for multivalent counterions it takes place at a lower value of Q and the number of free ions is much smaller (l/* a ). Imai (1961) has observed that multivalent counterions are more strongly bound than are monovalent ones. This phenomenon can be demonstrated theoretically by considering equilibrium conditions for two counterions with valencies zx and z2 (z2 > zx) and degrees of dissociation fix and /?2. For a cylindrical model the equilibrium equations are ^-^/.A)^
(4.16)
fflwJi+fMW (4.17) where fx and/ 2 represent the proportions of the total charge carried by the counterions, i.e./i+/ 2 = 1. 65
Polyelectrolytes, ion binding and gelation As 9^0 then the solutions to these equations fall into four groups depending on the value of Q. A->1, &^1,
Px-> 1, P1->l/f1z1Q,
&"•!
forg0 02^O
for l z
(4.18) l
z
/ 2 < Q < /fi 2
(4.19)
for I/ft z2 ^ Q ^ I/ftz x (4.20) for I / f t z ^ e (4.21)
These equations are represented graphically in Figure 4.5. Increase in the binding of counterions as Q increases is reflected as a decrease inftvalues. No binding occurs until Q reaches l/z 2 , when the binding of the highervalence ions begins. This process is complete when Q attains a value of l/ftz 2 . There is no further binding of counterions until Q reaches l// 1 z 1 when the binding of the lower-valence ions commences. Figure 4.5 shows that when there is a mixture of counterions then those of the higher valence are preferentially bound. Lower-valence ions can completely suppress the dissociation of those of higher valence.
Figure 4.5 The effect of Q on the dissociation {fix /?2) of ions of two valencies. Note the suppression of the dissociation ( / y of ions of higher valence zx by those of lower valence z2. Based on Oosawa (1957).
66
Ion binding The selective binding of cations is not as sensitive to size as to valence. The value of Q for the condensation of counterions of the same valence is unaffected. In the case of monovalent cations, the dissociation of all counterions is complete at infinite dilution, when Q ^ 1. When Q ^ 1 the dissociation of the smaller counterion is always greater than that of the larger one and increases relatively as Q increases. A number of workers have observed that the strength of binding of monovalent counterions depends on ionic radius. However, the effect of ionic radius is somewhat obscure as it depends on hydration phenomena and whether the size of the bare ion or that of the hydrated ion is the significant parameter (Wilson & Crisp, 1977). 4.2.5
Site binding - general considerations
Not all ions are mobile within the ionic atmosphere of the polyion. A proportion are localized and site-bound-a concept apparently first suggested by Harris & Rice (1954). Localized ion binding is equivalent to the formation of an ion-pair in simple electrolytes. Experimental evidence comes mainly from studies on monovalent counterions. This concept is due to Bjerrum, who in 1926 suggested that in simple electrolytes ions of the opposite charge could associate to form ion-pairs (Szwarc, 1965; Robinson & Stokes, 1959). This concept of Bjerrum arose from problems with the Debye-Hiickel theory, when the assumption that the electrostatic interaction was small compared with kT was not justified. Bjerrum considered the case of spherical ions in a solvent of dielectric constant e. The probability of finding two ions of opposite charge at a distance A from each other is calculated from the number of ions surrounding a central ion of opposite charge in a spherical shell of thickness dA and radius A. This probability, W{A), is given by W(A) dA = (4nA2 dA/v) exp (e2/sAkT)
(4.22)
for monovalent ions. This distribution has a minimum, Am, at Am = e'/2ekT
(4.23)
For a cation of charge z+ and anion of charge z_, this minimum becomes Am = z+z_e2/2ekT
(4.24)
When A > Am the ions are free and the Debye-Hxickel theory applies. When A < Am the two ions tend to approach each other and form an ionpair, and there is no contribution to the electrostatic energy from the interaction between an ion and its atmosphere. 67
Poly'electrolytes, ion binding and gelation
The high dielectric constant of water normally militates against the formation of ion-pairs for simple salts because a high dielectric constant reduces the strength of the electrostatic forces. The phenomenon is more readily observed in solvents of low dielectric constant; for a typical monomonovalent salt, ion-pair formation takes place only when the dielectric constant is less than 41 (Fuoss & Kraus, 1933). The fraction of all ions forming ion-pairs is W(A)dA
(4.25)
J2c
where a is the radius of the central ion. This distribution has some inconsistencies - for example it diverges when R is large - and was modified by Fuoss (1934); see Figure 4.6. These arguments for simple electrolytes can be extended to the relationship between the two types of bound counterion in polyelectrolytes: the bound but mobile (atmospheric) and the localized (site-bound). Under equilibrium conditions, the relationship between sitebound and atmospheric ions is (4.26)
2
I
Con tact distance
Ion-pair range Inter-ion distance
Figure 4.6 The Fuoss (1934) distribution function.
68
Ion binding
where ns is the number of site-bound ions, «a is the number of atmospheric ions and K is the equilibrium constant. For monovalent cations in dilute solution (0-1 M) the degree of localized ion binding is negligible; in more concentrated solutions some site binding does occur. In general, localized ion binding can be expected only with multivalent cations. When site binding occurs, the equations which relate the numbers of free and bound ions require some modification. The relationship is then between free ions and those bound ions that are mobile. The equations are similar to equations (4-8), (4-9), (4-14), and (4-15), but the number of sitebound ions has to be discounted in all calculations for P, Q, /? etc. 4.2.6
Effect of complex formation
In a discussion of papers by Rice & Harris (1954) and Harris & Rice (1954), Van Wazer (1954) suggested that there could be covalent binding as well as electrostatic interaction and that cations could be held at specific sites by complex formation. This is a reasonable inference, because site binding is significant only with multivalent cations and strong electrostatic interactions. Under these conditions ion polarization occurs and bonds have some covalent character (Cotton & Wilkinson, 1966). This is illustrated by the data of Gregor, Luttinger & Loebl (1955a,b). They measured the complexation constants of poly(acrylic acid), 0-06 N in aqueous solution, with various divalent metals, which, as it so happens, are of interest to AB cements (Table 4.1). The order of stability was found to be Mg < Ca < Co < Zn < Mn < Cu Mandel & Leyte (1964) found a similar order for the complexes of poly(methacrylic acid): Mg < Co < Ni < Zn < Cd < Cu Some of these divalent cations form part of the Irving-Williams series: Mn, Fe, Co, Ni, Cu and Zn. Irving & Williams (1953) examined the stability constants of complexes of a number of divalent ions and found that the order Mn < Fe < Co < Ni < Cu > Zn held for the stability of most complexes irrespective of the nature of the coordinated ligand. The stability constants of metal-poly(alkenoic acid) 69
Polyelectrolytes, ion binding and gelation Table 4.1. Metal PAA complexes (Gregor, Luttinger & Loebl, 195 5 a,b) Metal ion
Crystal ionic radius A
Complexation constant
Cu 2+ Mn 2+ Zn 2+ Co 2+ Ca 2+ Mg 2+
0-72 0-80 0-80 0-72 0-99 0-66
6-0 xlO 3 2-3 x 103 2-1 x 103 4-0 x 102 1-OxlO2 6-0 x 101
complexes, for the most part, follow the Irving-Williams series as do the stabilities and strengths of poly(alkenoic acid) cements. The complexation constant for copper(II) is particularly high and Wall & Gill (1954) have suggested that chelate formation takes place with two carboxyl groups:
From all of this discussion it is apparent that, as Manning (1979) said, the binding between counterion and polyion can range from atmospheric to covalent site binding. 4.2.7
Effect of the polymer characteristics on ion binding
The extent of ion binding depends on a number of characteristics of the polyion: degree of dissociation, acid strength, conformation, distribution of ionizable groups and cooperative action between these groups (Wilson & Crisp, 1977; Oosawa, 1971; Harris & Rice, 1954, 1957). The hydration state of the macromolecule, which is in turn dependent on conformation, also affects ion binding (Begala, 1971). 70
Ion binding There are differences in ion binding between different polyacids. Thus, alkali metal ions are bound more strongly to poly(acrylic acid) than to the weaker poly(methacrylic acid) (Wilson & Crisp, 1977). Again, the ranking order for the binding strength of alkali metal ions depends on the nature of the polyanion, and the order is different for poly(acrylic acid) than for poly(maleic acid) or poly(itaconic acid). Thus, for poly(acrylic acid) the binding strength increases in descending order of the ionic radius of the bare cation: K+ < Na+ < Li+ For poly(maleic acid) and poly(itaconic acid), the binding strength increases in descending order of the size of the hydrated metal ion, which is the reverse of that for the bare ion (Muto, Komatsu & Nakagawa, 1973; Muto, 1974). This observation has been explained by postulating the formation of a stable ring structure with a hydrogen bridge between ionized and non-ionized carboxylate groups. The strength of ion binding is enhanced when the arrangements of the functional groups permit chelate formation (Begala & Strauss, 1972). Thus, magnesium is more firmly bound to poly(vinyl methyl ether-maleic acid) than to either poly(acrylic acid) or poly(ethylene maleic acid). The charge or number of dissociated groups on a poly acid chain depends on the degree of neutralization and is reflected by the pH of the solution. Behaviour is determined by the site binding of hydrogen ions; in the case of a weak polyacid the number of free hydrogen ions may be neglected. It follows that decrease of site binding of hydrogen ions is directly proportional to the amount of added alkali. In the case of poly(acrylic acid) its polymer chain can be regarded as a copolymer containing pendant COOH and COO" groups, the relative amounts of each depending on the degree of neutralization. When the degree of neutralization is small the charge on the polyion and the number of counterions will also be small and the majority of counterions will be free. As the degree of neutralization, a, increases, the polyion charge, Q, will increase. This observation follows from the following equations: Q = nel/eokTl
(4.27)
where n = number of ionized groups on the polyion. It follows that Q = anoel/sokTl
(4.28) 71
Poly electrolytes, ion binding and gelation
where n0 = the potential number of ionizable groups on the polyion. When Q is low, most of the counterions are free, but as neutralization increases a point is reached at which the counterions condense; above this point, additional counterions are bound. This follows from the discussion in Section 4.2.3.
4.2.8
Solvation (hydration) effects
The solvation (hydration) and desolvation of ions is important to the gelation process in AB cement chemistry. The large dipole moment of ionpairs causes them to interact with polar molecules, including those of the solvent. This interaction can be appreciable. Much depends on whether the solvent molecule or molecules can intrude themselves between the two ions of the ion-pair. Thus, hydration states can affect the magnitude of the interaction. The process leading to separation of ions by solvent molecules was perceived by Winstein et al. (1954) and Grunwald (1954). Consider two ions in contact. As they are pulled apart the potential energy of the two ions increases. At some critical point the separation becomes sufficient for a polar solvent molecule to occupy the space between them, which reduces the energy of the system. Further separation increases the energy of the system again. These changes demonstrate that two types of ion-pair exist: contact and solvent-separated. This distinction is meaningful if the resultant distribution function is of the type shown in Figure 4.7 (Szwarc, 1965). This figure shows that there is a high probability that the cation and anion are either in contact, separated by a solvent molecule or far apart (Szwarc, 1965). Intermediate positions are improbable. The structure of solvated ion-pairs has been studied by Grunwald (1979) using dipole measurements. Winstein & Robinson (1958) used this concept to account for the kinetics of the salt effects on solvolysis reactions. They considered that carbonium ions (cations) and carbanions could exist as contact ion-pairs, solvated ion-pairs and as free ions and that all these forms participated in the reactions and were in equilibrium with each other. These equilibria can be represented, thus: X : Y = X+Y" = X+SY~ = X+ contact solvent-separated free ions ion-pair ion-pair 72
Ion binding
where X is a carbonium radical and X+ the carbonium ion, Y~ the carbanion and Y the associated radical. S represents a solvent molecule. Eigen & Tamm (1962a,b) and Atkinson & Kor (1965, 1967) envisage a more complex situation and consider that there are two kinds of solventseparated ion-pairs: those with one intervening molecule of solvent and others where the ion-pair is fully solvated (Wilson & Crisp, 1977). 4.2.9
Hydration of the polyion
The electric potential around a polyion can aflfect the structure of water. There are three regions of potential about a polyion to consider: the potential holes at the site of the individual charged groups, cylindrical regions along the polymer chain, and the outlying region. In the outlying region the potential is small and the water molecules have a normal structure. In the other two regions there are strong electricfields,and water molecules are oriented and have special structures. Oriented water is denser and has a higher refractive index than normal water (Begala & Strauss, 1972; Ikegami, 1964, 1968). The structure of this water can be affected by ion binding. If the counterions are tightly bound at the sites of individual charged groups, the
Solvent ion-pair Inter-ion distance
Figure 4.7 Distribution function for contact and solvent-separated ion-pairs.
73
Poly electrolytes, ion binding and gelation structure of the water around them will be profoundly modified. If the counterions are not localized but mobile, the influence on water structure may be small. Thus, the state of the binding of a counterion will be reflected in changes in water structure which in turn can be measured by changes in refractive index or density. The effect has been studied experimentally by Ikegami (1964, 1968) who measured changes in refractive index, Asai (1961) employing an ultrasonic method, Begala & Strauss (1972) who measured changes in molar volume, and Grunwald (1979) using dipole moment measurements. When an acid is neutralized by a base the refractive index of the salt solution formed is less than the weighted mean of the refractive indices of the acid and base solutions from which it is formed. Likewise, the density increases. By these means, the progress of neutralization may be followed. At low degrees of neutralization, the average distance between ionized groups is great, so that the rearrangement of neighbouring water molecules induced by the ionization of a carboxyl group is solely due to the charge on that individual group. Individual hydration spheres of oriented water, intrinsic water, are formed at each charged site. In the case of poly(acrylic acid) when the degree of neutralization, a, is 0-3 the radius of these spheres is 031 nm (Ikegami, 1964). As a increases, the average distance between ionized groups decreases so that these neighbouring groups begin to have an effect. When a exceeds 0-3, individual water spheres begin to overlap and eventually coalesce into a cylindrical form. With further increases in a, a second outer cylindrical sheath of water appears in which water molecules are oriented by the cooperative effect of two or more carboxyl groups. When neutralization is complete, the inner layer of intrinsic water assumes a cylindrical form along the length of the polyion with a diameter of 0-5-0-7 nm (Ikegami, 1964). The outer second cylindrical hydration region has a diameter of 0-9-1-3 nm (Figure 4.8). The explanation for this volume increase is as follows. For a cylindrical model of uniform charge density the electric field around the cylinder is 2cm0e0/e0rl
(4.29)
where a is the degree of neutralization, r the radius of the cylinder and /its length. When the magnitude of the electric field exceeds a certain value, water molecules are reoriented; the above expression shows that the radius, r, of the cylinder increases as the degree of neutralization, a, increases. 74
Ion binding
According to Ikegami (1968) the presence of hydrophobic groups, for example the methyl group in poly(methacrylic acid), can induce an additional hydration region around neighbouring charged groups. The arrangement of carboxyl groups on the polyacid is also important. Thus, poly(ethylene maleic acid), PEMA, which is an 'isomer' of poly(acrylic acid), PAA, has a different hydration structure. Whereas in PAA the COOH groups are pendant on alternate chain C atoms, those in PEMA are paired on adjacent chain C atoms. These structural differences affect hydration (Begala, 1971). The separation of the hydrophilic carboxyl groups by a pair of hydrophobic chain C atoms effectively prevents the cooperative effect between ionizable groups. Thus, by contrast with PAA, as the degree of ionization increases, the hydration regions around PEMA never coalesce to form a cylindrical sheath. In the fully ionized state there is a spherical region of intrinsic water around each carboxyl group and an outer spherical region of water which encloses each pair of carboxyls. The formation of a stable hydrogen-bonded ring structure as in poly(itaconic acid) and in poly(maleic acid) has also been shown to affect hydration states (Muto, Komatsu & Nakagawa, 1973; Muto, 1974). 0.9-1.3nm 0.5-0.7 nm
al,PQ=l
Osmotic force oc 1 — 1 /Q and the Coulombic force is constant. At high Q values the contribution from osmotic pressure predominates. This is shown by the ratio of the two forces which is given by Coulombic force: Osmotic force oc /?2/(l —/?) These considerations apply to dilute solutions. In concentrated solutions the extensive forces will be diminished. Also if the bound counterions become site-bound then both extensive forces are diminished. These are important factors to consider in the theory of acid-base gelation in AB cements, where solutions are concentrated and many counterions are sitebound.
Osmotic force
Figure 4.12 The effect of Q on the extensive forces, coulombic and osmotic, acting on a cylindrical and a coiled polyion. Based on Oosawa (1971).
81
Polyelectrolytes, ion binding and gelation 4.2.13 Interactions between poly ions Repulsive coulombic forces exist between charged polyions. These are attenuated by the bound counterions; conversely they are stronger for polyions having a higher concentration of free counterions. When the charge along the polyion, Q, is small the forces involved are purely coulombic repulsion forces. However, when Q exceeds a certain value, counterions condense on the polyions and reduce the repulsive forces. Attractive forces arise from dipole interaction, a result of the fluctuations in the cloud of counterions. Although the mean distribution of counterions is uniform along the length of the polyion, there are fluctuations in the cloud of counterions which induce transient dipoles. When two polyions approach each other counterionfluctuationsbecome coupled and enhance the attractive force. Since polyions have a high polarizability these attractive forces can be considerable. The repulsive force between polyions, calculated for the mean equilibrium distribution of the counterions, is ell/eQz*
(4.30)
where / is the length of the polyion. The attractive forces are kTl/D2
(4.31)
where D is the average distance between polyions. If D7~ (Q1); ring and chain silicates, (SiO3)2w~ (Q2). Certain sheet and three-dimensional silicates can also yield gels with acids if they contain sites vulnerable to acid attack. This occurs with aluminosilicates provided the Al/Si ratio is at least 2:3 when attack occurs at Al sites, with scission of the network (Murata, 1943). Mase (1961) explains the reactivity of silicate minerals towards acids in terms of the polarizing ability of the cation and the bond energy of the mineral. Clearly, too, the reactivity of minerals towards acids is connected with their basicity, and one notes that the orthosilicates are basic minerals. According to the ideas of Flood & Forland (1947a,b) (Section 2.3.3) basicity is related to the residual polarizability of the oxygen atom. If this is large, as will be the case if the associated cation has little polarizing power, then the oxygen atom is basic. Thus, sodium silicates are the most basic of the silicates, and silica itself is acidic and so resistant to acid attack. Basicity is also related to the silicate structure: orthosilicates are more basic than pyrosilicates which, in turn, are more basic than ring and chain silicates. Using this information, Crisp et al. (1977, 1979) and Hornsby et al. (1982) selected candidate minerals for cement formation with poly (acrylic acid) and found a number of minerals that formed cements (Table 5.4). 114
Table 5.4. Properties of mineral ionomer cements {Crisp et al, 1979; Hornsby et aL, 1982)
Powder: liquid, gem" 3
Setting time, min
Compressive strength (24 h), MPa Humid
Sohibil
Water
Orthosilicates Willemite Gadolinite
Zn 2 [SiOJ Be 3 Fe(YO) 2 [SiOJ 2
2-0 2-1
26 140
19 40
21 40
0-35 002
Pyrosilicates Gehlenite Hardystonite
Ca2Al[AlSiO7] Ca 2Zn[Si 2O7]
1-0 10
48 6
1 9
6 12
2-4 0-4
Chain silicate Wollastonite
Ca3[(SiO3)3]
1-0
14
18
3
3-4
Sheet silicate Thuringite
(Fe(II), Fe(III), Mg, Al)12[(Si, Al)8O20](O, OH, F ) r
1-5
29
28
35
10
Zeolite Scolecite
Ca[Al2Si3O10]3H2O
0-5
20
160
30
1-35
Ultramarine Hackmanite
Na 8 [Al 6 Si 6 O 2 J(Cl 2 ,S)
10
13
89
37
2-4
Feldspar Labradorite
(Ca,Na)[Al1_2Si2_3O8]
1-0
59
134
2
3-6
Polyalkenoate cements While some formed hard, rigid cements that were stable in water, others yielded rubbery or plastic masses that were hydrolytically unstable. Minerals with cement-forming capability were found in the following classes: (1) Island silicates containing discrete ions. Orthosilicates, SiO4~ (Q°); pyrosilicates, Si2O*~ (Q 1 ); and ring silicates, Si3OJ-, S&.O"- (Q2). Most orthosilicates reacted completely with poly(acrylic acid) solution; an exception was andradite, Ca 3 Fe 2 [SiO4]3. Even so, the cements of gehlenite and hardystonite were very weak and affected by water. Only gadolinite and willemite formed cements of some strength which were unaffected by water, probably because one contained beryllium and iron and the other zinc. (2) Chain silicates, consisting of connected metasilicate units, (SiO3)2w~ (Q2), and of an open structure. Wollastonite Ca(SiO3) reacted completely with poly(acrylic acid), but the cement was much affected by water. (3) Sheet silicates (Q3) with significant isomorphic replacement of Si4+ by Al3+ or Fe 3+ . These were decomposed by poly (aery lie acid) to silica gel. The chlorite, thuringite, formed a strong cement but was much affected by water. (4) Aluminosilicates with a three-dimensional network (Q 3 and Q4) where the Al/Si ratio was 2:3. These reacted with poly(acrylic acid), but none reacted completely. The zeolite, scolecite, the feldspar, labradorite, and the ultramarine, hackmanite, gave high-strength cements but all were much affected by water - the strength of the labradorite cement disappeared almost entirely - possibly because of the presence of free acid. 5.9
Glass polyalkenoate (glass-ionomer) cement
5.9.1
Introduction
The glass polyalkenoate cement, formerly known as the glass-ionomer cement, was invented by Wilson and Kent in 1969 (Wilson & Kent, 1973) and is now well established as a material that has an important role in clinical dentistry. It has proved to have considerable development potential and has been subjected to continuous development, improvement and 116
Glass polyalkenoate (glass-ionomer) cement diversification. It is the most versatile of all dental cements and currently accounts for most of the research and development on them. There are other applications of the cement as a biomaterial. It is used as a splint bandage material and as a bone cement. Glass polyalkenoate cement has a unique combination of properties. It adheres to tooth material and base metals. It releases fluoride over a long period and is a cariostat. In addition it is translucent and so can be colourmatched to enamel. New clinical techniques have been devised to exploit the unique characteristics of the material. The material originated from the general dissatisfaction with the clinical performance of the dental silicate cement. Wilson and his coworkers made extensive studies on the dental silicate cement (Section 6.5) and drew the conclusion that this cement could not be further improved. Wilson (1968) examined several alternatives to orthophosphoric acid, including organic chelating agents, as a liquid cement-former, but none of these were successful. Finally, after considerable research, the glass polyalkenoate cement was developed (Wilson & Kent, 1971, 1972, 1973). The cement is formed by mixing an ion-leachable glass powder with an aqueous solution of a poly(alkenoic acid). The glass is generally a fluoride-containing calcium aluminosilicate but calcium may be replaced by strontium or lanthanum. The cement was originally known as ASPA, an acronym of Aluminosilicate Polyacrylic Acid. The term ASPA is now applied to materials developed by the Laboratory of the Government Chemist in the UK, and was once also the brand name of an early commercial material. For many years it was known as the glass-ionomer cement - indeed, that is still the term in common use-but the International Standards Organization officially adopted the name glass polyalkenoate cement. The term glass-ionomer cement is now used as a generic term to cover these cements and the new glass polyphosphonate cements invented by Ellis and Wilson in 1987-9 (Ellis & Wilson, 1990). 5.9.2
Glasses
General The powders used in glass polyalkenoate cement formulations are prepared from glasses and not opaque sintered masses. In this they resemble the traditional dental silicate cement from which they are descended. The glass plays several roles in the chemistry and physics of the glass polyalkenoate 117
Polyalkenoate cements cement. It acts as a source of ions for the cement-forming reaction, controls the setting rate and strength of the cement and imparts the property, unusual in a cement, of translucency. Chemically these are special aluminosilicate glasses. Until quite recently, all were calcium aluminosilicates, but now calcium is sometimes wholly or partly replaced by strontium and lanthanum. Most glasses also contain fluorides which, besides lowering the temperature of glass fusion, play a role in cement formation and affect cement properties. Provided the Al/Si ratio is high enough, these glasses are decomposed by acids to release cement-forming ions (Wilson & Kent, 1973, 1974; Crisp & Wilson, 1978a,b, 1979; Kent, Lewis & Wilson, 1979; Wilson et al., 1980; Hill & Wilson, 1988a). They are similar to the glasses used for dental silicate cements, although the Al/Si ratio is higher. Types of glass There are a great number of potential glasses and some can be extremely complex. All contain silica and alumina and an alkaline earth or rare earth oxide or fluoride. The two essential glass types are SiO2-Al2O3-CaO and SiO2-Al2O3-CaF2, from which all others are derived. Oxide glasses have been reported by Crisp & Wilson (1978a,b, 1979), Wilson et al. (1980), and Hill & Wilson (1988a). The fusion mixtures contain silica, alumina and calcium carbonate to which sodium carbonate or calcium orthophosphate may be added. They may be represented thus, with fusion temperature given in parentheses: SiO2-Al2O3-CaO (1350-1550 °C) SiO2-Al2O3-CaO-P2O5 (1370-1450 °C) SiO2-Al2O3-CaO-Na2O (1200-1350 °C) Influorideglasses, calciumfluorideis an essential constituent, but generally cryolite, Na3AlF6, is also added as a flux to lower the temperature of fusion. Aluminium orthophosphate is also generally added to the fusion mixture for various reasons. Of course, the various elements may be added in different ways. Thus, calcium orthophosphate, aluminium fluoride and sodium carbonate are often used in the preparation of fluoride glasses. Apart from lowering the temperature of glass fusion, fluoride improves the handling qualities of the cement paste, increases cement strength and translucency, and has a therapeutic quality when used as a dental filling material. In fluoride glasses the ratio of alumina to silica controls the setting time of the cement; fluoride tends to slow setting while aluminium 118
Glass polyalkenoate (glass-ionomer) cement orthophosphate improves the mixing of the paste. Sodium in the glass improves the translucency of the cement but can affect its hydrolytic stability. In addition, glasses have been reported where calcium is replaced by strontium or lanthanum (Akahane, Tosaki & Hirota, 1988) which impart radio-opacity to the cement. Fluoride glasses are difficult to classify because the various constituents can be added to the fusion mixture in several ways. However, glasses of the Laboratory of the Government Chemist (Wilson & Kent, 1973; Kent, Lewis & Wilson, 1979; Wilson et ai, 1980; Hill & Wilson, 1988a), which form the basis of many commercial cements, can be represented as SiO2-Al2O3-CaF2 (1150-1350 °C) SiO2-Al2O3-CaO-CaF2 (1320-1450 °C) SiO2-Al2O3-CaF2-AlPO4 (1150-1300 °C) SiO2-Al2O3-CaF2-AlPO4-Na3AlF6-AlF3 (1100-1300 °C) where again the temperature of fusion is given in parentheses. After fusion the molten glass is shock-cooled by pouring it onto a metal plate and then into water. The glass fragments are then finely ground to pass either a 45-|im sieve for afillingmaterial, or a 15-jim sieve for a finegrained luting agent. The glass powders may be annealed after preparation by heating at 400 to 600 °C; in general, the effect is to slow down the setting reaction. Sometimes the powder is acid-washed to improve the mixing qualities of the cement. Structure of aluminosilicate glasses The formation of a cement is dependent on the ability of the glass to release cations to acid solutions. It is not sufficient for network-modifying cations to be exchanged for hydrogen ions as this would restrict the attack to the glass surface only. It is required that the glass structure itself be completely decomposed if all the glass ions are to be available for release. Aluminosilicate glasses have this property. To discuss why this is so it is useful to have an appropriate conceptual framework, and one can be developed from the Random Network model of Zachariasen (1932). Zachariasen (1932) conceived of a glass structure as a random assembly of oxygen polyhedra, these polyhedra consisting of a central glass-forming cation surrounded by a small number of oxygen atoms, e.g. [SiOJ tetrahedra. These polyhedra were considered to be linked at corners only, via 2-coordinate oxygen atoms. This concept amounts to regarding a glass as a type of highly crosslinked polymer based on -Si-O-Si- linkages. This 119
Polyalkenoate cements idea, although much criticized (Rawson, 1967), has proved to be a fruitful one. Zachariasen (1932) added another criterion, namely that the random network was three-dimensional, and, therefore, in modern terminology, composed only of Q4 and Q3 units. Hagg (1935) considered that this requirement was not always necessary and a glass might contain large irregular anionic groups. The work of Trap & Stevals (1959) supported this view, for they prepared so-called invert glasses containing only Q2 and Q1 units, that is glasses with no crosslinking. In these glasses at least half the oxygen atoms are non-bridging -O" groups, so the -Si-O-Si- chains are anionic and are held together by network-modifying cations (these do not form part of the glass structure). Today, following Ray (1975, 1983) we would call these ionic polymers. We are now in a position to discuss requirements for ionomer glasses further. Consider the case of the simple silica (SiO2) glass where we can represent the network diagrammatically thus: O
I
O
I
O
I
— Si—O—Si—O—Si—O—Si — O
I
I
I
O
I
I
This infinite three-dimensional network is electrically neutral and impervious to acid attack. If so-called network-modifying cations are introduced then this network must acquire a negative charge leading to the breaking of an Si-O-Si bridge to form non-bridging oxygens: \ / — Si—O—Si— / \
Ca2+ \ „ • — S i — Or Ca 2+ "O — S i — /
/ \
This is a type of ionic polymer where the negative charge on the network is balanced by the positively charged network modifier. Statistically all types of [SiOJ tetrahedra, Q1, Q2, Q3 and Q4, will be present in varying proportions, depending on the ratio of bridging to non-bridging oxygens. Aluminosilicates are more complex as aluminium can be either a network modifier in sixfold coordination or a network former in fourfold coordination. In the latter case, Al3+ is able to replace Si4+ in the glass network because it has a similar ionic radius, but the network then acquires a negative charge. If this charge becomes sufficiently high then the network becomes susceptible to acid attack. Again this charge on the network has to be balanced by positively charged network-modifying cations. Thus, we 120
Glass polyalkenoate (glass-ionomer) cement can regard an aluminosilicate glass structure as consisting of linked [SiOJ and [A1OJ~ tetrahedra. There are restrictions on the replacement of Si4+ by Al 3+ . The Al/Si ratio apparently cannot exceed 1:1 (Lowenstein, 1954). Nor can all the aluminium go into the network if there are insufficient network-modifying cations to balance the network charge. Under such conditions aluminium adopts a sixfold coordination. We must note that recently Ellison & Warrens (1987), using 27 A1NMR spectroscopy, have found evidence for the existence of aluminium in pentacoordination in asymmetric or distorted sites using previously established assignments (Kirkpatrick et al. 1986; Risbud et al., 1987; Cruikshank et al., 1986). A negatively charged network of non-bridging oxygens and aluminium sites renders these glasses susceptible to acid attack. Overall the introduction of network-modifying cations and aluminium ions increases the polarizability of oxygen ions and, therefore, vulnerability to acid attack. The mode of acid decomposition of an aluminosilicate glass is depicted in Figure 5.5. It can be seen that attack by hydrogen ions involves exchange of network-modifying cations (Ca2+, Na + ) and rupture of the aluminosilicate network at aluminium sites to yield silicic acid and aluminium ions (Wilson, 1978b; Prosser & Wilson, 1979). Glasses used in glass polyalkenoate cements have been observed to release cations, fluoride if present and silicic acid (Crisp & Wilson, 1974a; Wasson & Nicholson, 1990,1991). Similar observations have been made for the related dental silicate cement
polymerizes silica gel Figure 5.5 The mode of acid decomposition of an aluminosilicate glass.
121
Polyalkenoate cements Table 5.5. Glass compositions and acid extracts {Crisp & Wilson, 1974a; Wasson & Nicholson, 1990) G-200
G-338
Mole ratio
Glass
Cement extract
Glass
Si:Al Ca:Al Na:Al F:A1
0-98 0-89 014 2-46
1-17 0-25 0-97
0-67 0-26 0-44 1-67
Cement extract 0-63 0-36 0-65
Table 5.6. Composition of oxide glasses used in studies on polyalkenoate cements, parts by mass {Wilson et al., 1980; Crisp, Merson & Wilson, 1980)
SiO2 A12O3 CaO Ca 3 (PO 4 ) 2 Appearance Crystallites
G-273
G-275
G-287
G-255
G-247
120 102 168 —
240 102 112 —
180 102 56 —
120 102 56 —
160 100 — 140
clear
clear
clear
opaque
opal
—
—
—
An
—
Properties
Powder .liquid, g cm"3 Setting time (37 °C), min Strength (24 h), MPa
20 2-75
95
30 40 35
30 8-25
29
30 40 56
— 5-25
72
An = Anorthite
(Wilson & Kent, 1970). The release of each glass species is roughly governed by the amount contained in the glass and so varies with glass type (Table 5.5). The reactivity of a glass towards acids depends on its acid-base properties and both the Bronsted-Lowry and Lewis theories have been applied to oxide glasses (Volf, 1984). Basic components of a glass are the metal oxides, and acidic ones are silicon, boron or phosphorus oxides. The important factor is the state of the oxygen atoms. In purely oxide glasses the basicity of a glass depends on the ability of the oxygen atoms to give up 122
Glass polyalkenoate (glass-ionomer) cement electrons. This is greatest when the oxygen atoms are associated with cations of low electrostatic field strength, for example Na+ and Ca2+, and least when the cations have a high electrostatic field strength, for example the highly charged, small Si4+ ion. Lux (1939) introduced the symbol pO (note it is not an exponent like pH) to quantify the acid-base balance in a glass, and various attempts have been made to obtain values for this parameter. All are based on the electronegativity of the cation or a related characteristic, such as electrostatic field strength (Volf, 1984). SiO2-Al2O3-CaO glasses In these glasses (Table 5.6) the coordination state of aluminium depends on its chemical environment and can only be entirely fourfold when the Ca/Al SiO * D A o 9 1:3
•
Non-setting Slow setting Moderate setting Fast setting Fast setting crystalline mass Ultra-fast setting
1:2 Al /Si mole ratio
C9S
C,S
CaO
A1 2 O 3
Figure 5.6 Triangular composition diagram for SiO2-Al2 O3-CaO glasses, showing that glasses with cement-forming ability fall within the gehlenite and anorthite composition region, and that only glasses with less than 61 to 62 % by mass of silica have the potential to form a cement (Hill & Wilson, 1988a).
123
Polyalkenoate cements Table 5.7. Properties of cements formed from glasses corresponding to the generic formula xSi02.Al203.Ca0 Cement properties
moles(x)
mass %
setting time, minutes
strength, MPa
10 20 40 60
21-9 35-9 52-8 62-7
3-5 2-25 40 non-setting
104 74 35 zero
SiO * D • o •
Zero strength Unworkable Weak Low strength Moderate strength
1:3 1:2 Al /Si mole ratio
C2S C,S
CaO
A12O3
Figure 5.7 Triangular composition diagram for SiO2-Al2 O3-CaO glasses. Glasses in the gehlenite region yield stronger cements (95 to 104 MPa in compression) than those in the anorthite region (29 to 56 MPa) (Hill & Wilson, 1988a).
124
Glass polyalkenoate (glass-ionomer) cement Table 5.8. Composition of fluoride glasses used in studies on polyalkenoate cements, parts by mass (Kent, Lewis & Wilson, 1979; Wilson et al., 1980)
SiO2 A12O3 CaF 2 Na 3 AlF 6 A1F3 A1PO4 SiO 2 :Al 2 O 3 bymass Appearance Crystallites
G-200
G-307
175 100 207 30 32 60 1-75 opaque Fl
133 67 100 100 100 100 — — — — — — 1-33 0-67 clear opaque Fl,Co
175 100 166 — — 60 1-75 opal Fl
175 100 117 — — 60 1-75 opaque Co
175 100 90 135 32 170 1-75 opal Ap
30 3-5 107
3-5 3-0 149
30 30 199
1-8 3-75 149
Properties Powder: liquid, g cm"3 30 Setting time (37 °C), min 5-2 Strength (24 h), MPa 185
G-309
30 6-5 166
G-235 G-237
G-338
Fl = Fluorite, Co = Corundum, Ap = Apatite
ratio > 1:2 and Al/Si ratio < 1:1 (Isard, 1959; Lowenstein, 1954). Thus, aluminium is in fourfold coordination in anorthite glass (Ca: Al = 1:2, Al:Si = 1:1); the glass is composed of Q4 and Q 3 units, i.e. it is threedimensional. Gehlenite glass must contain some aluminium in sixfold coordination (Ca: Al > 1:2, Al:Si > 1:1) and is composed of paired Q 1 units, i.e. [AlSiO7]. A study by Ellison & Warrens (1987) on two of these glasses, using 27A1 and 29Si NMR, produced results not too dissimilar from theoretical predictions. In glass G-273, Ca3Al2Si2O9 (Table 5.6), aluminium was found to be mainly in tetrahedral coordination with a minor amount in octahedral coordination. Similar results were found for glass G-275, Ca2Al2Si4O13 (Table 5.6), but, in addition, some aluminium was found to be pentacoordinate. Possible structural units were considered to be Q 3 (1A1) and Q 2 (0A1) with some Q 4 (3A1). The number of Al replacing Si in the second coordination sphere is given in parentheses. Glasses that have cement-forming ability fall within the gehlenite and anorthite composition regions of this system, and only glasses with less than 61 to 62 % by mass of silica have potential to form a cement (Figure 5.6). Cements are not formed if the Si/Al mole ratio exceeds 3:1. When the 125
Polyalkenoate cements ratio is less than 2:1 fast-setting cements are obtained with setting time of 2 to 10 minutes. There is only a very small region for slow-setting cements and as Table 5.7 shows there is a critical region between setting and nonsetting. Glasses in the gehlenite region yield stronger cements (95 to 104 MPa in compression) than those in the anorthite region (29 to 56 MPa) (Figure 5.7). SiO2-Al2O3-CaF2
glasses
These (Table 5.8) are the basic type from which most biomedical glass polyalkenoate cements are derived. Although thefluoridecontent is high, many of these shock-cooled glasses are clear. Clear glasses are confined to a narrow central compositional range at the centre of the phase diagram where the Al2O3/CaF2 ratio is around 1:1 by mass and the SiO2/Al2O3 ratio exceeds 1-33:1 by mass (Figure 5.8). Outside this regionfluorite,and sometimes corundum, phase-separate. Even the clear glasses can be SiO2 • • A • o •
Clear, non-setting Opal, non-setting Clear, slow setting, low strength Opal, slow setting, low strength Clear, fast setting, high strength Opal, fast setting, high strength
1:3
CaF 2 /Si0 2 ratio by mass
ratio by mass
3:1
CaF2
AI2O3
ratio by mass
Figure 5.8 Clear glasses are confined to a narrow central compositional range at the centre of the phase diagram where the Al 2 O 3 /CaF 2 ratio lies in the region 1:1 by mass and the SiO 2 /Al 2 O 3 ratio exceeds 1-33:1 (Hill & Wilson, 1988a).
126
Glass polyalkenoate (glass-ionomer) cement induced to phase-separate when heated to 450 °C, and this reduces their reactivity. The ability of a glass to form a cement is governed by the SiO 2 /Al 2 O 3 ratio which represents the acid-base balance in these glasses. If this ratio is 3-0 or more by mass then the glass will not form a cement. If it is below 2-0 then the cements formed are rapid-setting (2-5 to 5-0 minutes). Glasses in a very narrow band around a ratio of about 2-0 are slower-setting (6-5 to 18 minutes). The critical ratio for non-setting lies somewhere between 2-0 and 3-0. The effect of SiO 2 /Al 2 O 3 ratio on setting time and compressive strength is shown in Figure 5.9. Note that compressive strength increases steadily as the SiO 2 /Al 2 O 3 ratio decreases. Setting time decreases as the SiO 2 /Al 2 O 3 ratio decreases until a point is reached when phase separation 200 - | 1 -
10.0i
0.8
8.0 150
|6.0
o> cb