IUCr NIONOGRAPHS ON CR.Ys'1‘A1,L()C1mP1—n'
`
`- 14
`
`IOEL BERNSTEIN
`
`INTERNATIONAL UNION OF CRYSTALLOGRAPHY
`
`Page 1 of 37
`
`Grunenthal GmbH Exhibit 2003
`
`Rosellini v. Grunenthal GmbH
`
`IPR2016—00471
`
`

`
`Polymorphism in
`Molecular Crystals
`
`JOEL BERNSTEIN
`
`Department of Chemistry
`Ben-Gurion University of the Negev
`
`CLARENDON PRESS o OXFORD
`2002
`
`Page 2 of 37
`
`

`
`OXFORD
`UNIVERSITY pruass
`
`Great Clarendon Street, Oxford OX2 6DP
`
`Oxford University Press is a department of the University of Oxford.
`it furthers the University’s objective ofcxcellence in research, scholarship.
`and education by publishing worldwide in
`Oxford New York
`
`Auckland CapeTown Dares Salaam Hong Kong Karachi
`KualaLumpur Madrid Melbourne Mexico City Nairobi
`New Delhi Shanghai Taipei Toronto
`With offices in
`
`Argentina Austria Brazil Chile Czech Republic France Greece
`Guatemala Hungary Italy Japan Poland Portugal Singapore
`South Korea Switzerland Thailand Turkey Ukraine Vietnam
`
`Oxford is a registered trade mark of Oxford University Press
`in the UK and in certain other countries
`
`Published in the United States
`by Oxford University Press lnc.. New York
`
`© Oxford University Press, 2002
`
`The moral rights of the author have been assorted
`Database right Oxford University Press (maker)
`First published 2002
`First published in paperback 2008
`
`All rights reserved. No part of this publication may be reproduced.
`stored in a retrieval system, or transmitted. in any form or by any means,
`without the prior pemiission in writing of Oxford University Press,
`or as expressly permitted by law, or under terms agreed with the appropriate
`reprographics rights organization, Enquiries concerning reproduction
`outside the scope of the above should be sent to the Rights Department.
`Oxford University Press. at the address above
`
`You must not circulate this book in any other binding or cover
`and you must impose this same condition on any acquirer
`British Library Cataloguing in Publication Data
`Data available
`
`Library of Congress Cataloging in Publication Data
`Bernstein. Joel.
`Polymorphism in molecular crystals / Jocl Bernstein.
`(IUCr monographs on crystallography; I4)
`lncludes index.
`
`1. Polymorphism (Crystallography) 2. Molecular crystals. I. Title. 1]. International
`Union of Crystallography monographs on crystallography; l4.
`QD951 .B57 2002
`548’.3—dc2l
`2001047556
`
`Typeset by
`Newgen Imaging Systems (P) Ltd., Chennai, India
`Printed in Great Britain
`by
`Biddles l.td., King's Lynn, Norfolk
`
`ISBN 978——0— I 9—85060S—8 (l lbk.)
`ISBN 978-0-19-—9236S6—5(Pbk.)
`
`M98765-132]
`
`Page 3 of 37
`
`

`
`8
`
`INTRODUCTION AND HISTORICAL BACKGROUND
`
`McCrone (1957, 1965) has also given detailed descriptions of the microscopic
`examinations and phenomena that can be used to distinguish polymorphism from other
`phenomena that sometimes have been mistakenly labelled as pseudopolymorphismz
`mesomorphism (i.e. liquid crystals). grain growth (boundary migration and recrys-
`tallization), and lattice strain.
`
`1.2.3 Conventions for naming polymorphs
`
`Part of the difficulty encountered in searching and interpreting the literature on poly-
`morphic behaviour of materials is due to the inconsistent labelling of polymorphs.
`In many cases, the inconsistency arises from lack of an accepted standard notation.
`However, often, and perhaps more important, it is due to the lack of various authors’
`awareness of previous work or lack of attempts to reconcile their own work with
`earlier studies (see, for instance, Bar and Bernstein 1985). While many polymor-
`phic minerals and inorganic compounds actually have different names (e.g. calcite,
`aragonite and vaterite for calcium carbonate or rutile, brookite, and anatase for tita-
`
`nium dioxide) this has not been the practice for molecular crystals, which have been
`labelled with Arabic (1, 2, 3, . . .) or Roman (I, II, III, . . .) numerals, lower or upper
`case Latin (a, b, c, .
`.
`. or A, B. C, .
`. .) or lower case Greek (or, ,8, y. .
`. .) letters, or by
`names descriptive of properties (red form, low-temperature polymorph, metastable
`modification, me).
`As Threlfall (1995) and Whitaker (1995) have commented, arbitrary systems for
`naming polymorphs should be discouraged to avoid confusion surrounding the num-
`ber and identity of polymorphs for any compound. Relative stability and/or order of
`melting point, as well as a specification of the monotropic or enantiotropic nature
`of the polymorphic form (see Section 2.2.4) have also been suggested as a basis for
`labelling (Herbslein 2001), but these do not allow for the discovery of forms with
`intermediate values, in addition to the fact that small differences in stability or melt-
`ing point might lead to different order and different labelling by different workers.
`McCrone (1965) proposed using Roman numerals for the polymorphs in the order of
`their discovery, with the numeral I specifying the most stable form at room tempera-
`ture. By Ostwald‘s Rule (Ostwa1d 1897) (Section 2.3) the order of discovery should
`in general follow the order of stability the least stable appearing first. McCrone also
`supported the suggestion by the Koflers (Kofler and Keller, 1954) that the Roman
`numeral be followed by the melting point in parentheses. In fact, the successors of
`the Koflers at the Innsbruck school have very much followed this practice (Kuhncrt-
`Brandstatter 1971), although in general it has not been adopted by others. The. use of
`melting points is complicated by the fact that while this datum has a clear thermo-
`dynamic definition. a number of techniques are employed to determine the melting
`point (or melting point range, in many cases) so that real or apparent inconsistencies
`may arise from such a designation (see Sections 4.2 and 4.3).
`In view of the body of literature already ex tant and the questions surrounding the
`definition ofa polymorph it does not appear to be practical to define hard and fast rules
`for labelling polymorphs. The Kofler method has clear ad vantages, since the melting
`point designation may eliminate some questions of identity; hence its use should be
`
`Page 4 of 37
`
`

`
`IS THIS MATERIAL POLYMORPHIC?
`
`9
`
`encouraged. For those studying (and naming) polymorphic systems it is important to
`be fully aware of previous work, to try to identify the correspondence between their
`own polymorphic discoveries and those of earlier workers, and to avoid llippancy in
`the use of nomenclature in the naming of truly new polytnorphs.
`
`1.3
`
`Is this material polymorphic?
`
`1.3.1 Occurrence ofpolymorphism
`
`Perhaps the most well-known statement. about the occurrence of polymorphism is
`that of McCrone (I965): ‘It is at least this author’s opinion that every compound has
`different polymorphic forms and that, in general, the number of forms known for a
`given compound is proportional to the time and money spent in research on that com-
`pound.’ As a corollary to this rather sweeping, even provocative, statement. McCrone
`noted that ‘all the common compounds (and elements) show polymorphism’, and he
`cited many common organic and inorganic examples.
`These echo similar statements by Findlay (1951) p. 35. ‘[polymorphisml is now
`recognized as a very frequent occurrence indeed‘, Buerger and Bloom (1937), ‘poly-
`morphism is an inherent property of the solid state and that it fails to appear only
`under special conditions’, and Sirota (1982), ‘[polymorphismJ is now believed to be
`characteristic of all substances, its actual non-occurrence arising from the fact that. a
`polymorphic transition lies above the melting point of the substance or in the area of
`yet unattainable values of external equilibrium factors or other conditions providing
`for the transition.‘
`Such statements tend to give the impression that polymorphism is the rule rather
`than the exception. The body of literature in fact indicates that caution should be
`exercised in making them. It appears to be true that instances of polymorphism are
`not uncommon in those industries where the preparation and characterization ofsolid
`materials are integral aspects of the development and manufacturing of products (i.e.
`those on which a great deal of time and money is spent): silica, iron, calcium sili-
`cate, sulphur, soap. pharmaceutical products, dyes, and explosives. Such materials,
`unlike the vast majority of compounds that are isolated, are prepared not just once, but
`repeatedly. under conditions that may vary slightly (even unintentionally) from time to
`time. Similarly, in the attempt to grow crystals of biomolecular compounds, much time
`and effort is invested in attempts to crystallize proteins under carefully controlled and
`slightly varying conditions. and polymorphism is frequently observed (Bernstein et al.
`1977; McPherson 1982). Even with the growing awareness and economic importance
`of polymorphism, most documented cases have been discovered by serendipity rather
`than through systematic searches. Some very common materials, such as sucrose and
`naphthalene, which certainly have been crystallized inn umerablc times. have not been
`reported to be polymorphic. The possibility of polymorphism may exist for any par—
`ticular compound. but the conditions required to prepare as yet unknown polymorphs
`are by no means obvious. There are as yet no comprehensive systematic methods
`for feasibly determining those conditions. Moreover, we are almost totally ignorant
`about the properties to be expected from any new polymorphs that might be obtained.
`
`L
`
`Page 5 of 37
`
`

`
`3
`
`Controlling the polymorphic form obtained
`
`Crystal growth is a science and an art. The scientist's role in the crystal growth process is that
`of an assistant who helps molecules to crystallize. Most molecules, after all, are very good at
`growing crystals. The scientific challenge is to learn how to intervene in the process in order
`to improve the final product. (Fitter 1991)
`
`I am very sorry, that to the many . . . difficulties which you meet with, and must therefore
`surmount, in the serious and effectual prosecution of Experimental Philosophy, I must add one
`discouragement more, which will perhaps as much surprise you as dishearten you; and it is.
`that besides that you will find. .
`. many of the experiments published by Authors, or related
`to you by the persons you converse with, false or unsuccessful, .
`.
`. you will meet with several
`Observations and Experiments, which though communicated for true by Candid Authors or
`undistrusted Eye-witnesses, or perhaps recommended to you by your own experience, may
`upon further tried disappoint your expectation, either not at all succeeding constantly, or at
`least varying much from what you expected.
`This Advertisement may seem of so discouraging a nature that I should much scruple the
`giving it to you, but that I suppose the trouble at that unsuccessfulncssc which you may meet
`with Experiments, may be somewhat lessened. by your being forewarned of such contingencies.
`And that you should have the luck to make an Experiment once, without being able to make
`the same thing again, you might opt to look upon such disappointments as the effect of and
`unfriendliness in Nature or Fortune to your particular attempts, as proceeding from a secret
`contingency incident to some Experiments, by whomever they be tried. (Boyle l66l)
`
`Control nevertheless i_s_. important in science—tremendously so—not as an end but as a
`component of proof. The ability to control is the strongest possible demonstration of true
`understanding. Many doubted whether Becquerel, Cutie, Bohr, Oppenheimer and the rest
`really understood what causes what inside the atom. But after July 16, 1945, when the day
`dawned prematurely to the northwest of Alamagordo, at White Sands, New Mexico, no one
`could possibly doubt any more, for the atom bomb was plainer than the sun. With a demon-
`strated ability to control, the good scientist may sign off like the mathematician at the end of a
`proof: Quad era! demrmstrandum. (Huber I991)
`
`3.1 General considerations
`
`Crystallization is a process that has fascinated both scientists and casual observers
`throughout the ages. It is indeed remarkable that upwards of 102° molecules or ions,
`distributed essentially randomly throughout some fluid medium (gas, liquid, or solu-
`tion) coalesce, very often spontaneously, to form a regular solid with a well-defined
`
`Page 6 of 37
`
`

`
`AGGREGATION AND NUCLEATION
`
`67
`
`structure, or in the case of polymorphs, with a limited number of well-defined struc-
`tures. Those structures are invariant across a wide variety of conditions, in some cases
`almost. under any conditions for which crystals form. Two of the principal questions
`t.o be asked for such a process is how it begins and how it proceeds, especially in the
`context of polymorphic systems. A great deal of work has been devoted to attempts
`to answer these questions, and in spite of considerable progress especially on exper-
`imental and empirical fronts. there is still much to be learnt in developing current
`models. Historical treatments of the classic notions of crystallization and recrystal-
`lization, including many important references, have been given by Tipson (1956) and
`van Hook (1961). A more recent thorough account may be found in Mullin‘s book
`(1993).
`For any substance it is possible in principle to define experimentally the solvents.
`temperature range, rate of evaporation or cooling, and many of the other conditions
`under which it will crystallize. This collection of conditions has been called the
`accurrence domain (Sato and Boistelle 1984). That domain exists for any substance,
`but rarely, if ever, are its contents completely known. The contents of the occurrence
`domain for any material——in the present context, any pol ymorph——are not necessarily
`unique. In regions in which there is an intersection of domains, one may expect that
`two or more polymorphs would crystallize under essentially identical conditions. On
`the other hand, determining which regions of the domain are unique to a particular
`polymorph can be advantageous in determining crystallization strategy. This chapter
`deals with a number of the factors which should be considered in making such a
`determination. along with examples of the phenomena associated with competitive
`polymorphic crystallizations.
`
`3.2 Aggregation and nucleation
`
`The thermodynamics and kinetics outlined in Chapter 2 attempt to treat the question
`of crystallization on the macroscopic scale. On the microscopic scale we would like
`to be able to answer questions about the critical size and structure of a collection of
`molecules that will grow into the eventual crystal. In particular. how and when will
`polymorphs be obtained or be prevented from forming? Classically, the first stage of
`crystallization is viewed as nucleation, the spontaneous formation or introduction ofa
`nucleus, or centre of crystallization, in the crystallization medium from which crystals
`may grow, although nuclei may also be destroyed before growing into larger crystals.
`The size of such nuclei has been a matter of considerable discussion. On the one
`hand, for instance, Ostwald (1902) claimed that particles containing between 103 and
`I0” molecules are not sufficiently large to induce crystallization from supersaturated
`solutions. but later work indicated that much more modest numbers (e.g. 10-105)
`may be considered a critical size to generate crystals (McIntosh 1919; Tamman and
`von Gronow 1931). This amounts to a cube of approximately 100 A on an edge and a
`‘crystal’ nucleus weighing as little as 10”” g. Additional aspects of the question of
`the size of a crystal nucleus are discussed by Mullin (1993).
`
`Page 7 of 37
`
`

`
`68
`
`CONTROLLING THE POLYMORPHIC FORM OBTAINED
`
`In an attempt to avoid some of the confusion extant in the current literature on the
`nature of nucleation, Mullin has provided a useful schematic classification for various
`terms in use:
`
`NUCLEATION
`
`
`
`HOMOGENEOUS
`(spontaneous)
`
`
`
`
`
`HE'l‘EROGENEOUS
`(induced by foreign particles)
`
`SECONDARY
`(induced by crystals)
`
`
`
`
`
`Primary nucleation refers to those systems that do not contain crystalline matter.
`When no foreign bodies are present (i.e. the crystallization results from the spon-
`taneous fonnation of nuclei of the crystallizing material) the process is referred to
`as homogeneous. The presence (intentional or unintentional) of foreign particles can
`also induce nucleation, which is then tenned heterogeneous.
`Secondary nucleation deals with the situation in which nuclei are generated in
`the vicinity of crystals of the solute already present in a supersaturated solution.
`The solute crystals may have resulted from primary nucleation or may be deliberately
`added. This subject has also been covered by Mullin, as well as in a number of reviews
`(Strickland—Constable 1968; Botsaris I976; DeJong 1979; Garside and Davey 1980;
`Garside 1985; Nyvlt et al. I985).
`Mullin has argued that the minimum number of molecules in a stable crystal nucleus
`can vary from about ten to several thousand. A model based on the simultaneous
`collision of this number of molecules with the degree of order required for it to
`be recognized by additional molecules as a crystal is highly unlikely. A more likely
`scenario is that the nucleus would be generated by a sequence of bimolecular additions
`in which the so-called critical cluster would be built up stepwise:
`
`A + A x——‘ 2A
`
`A+2A :—‘ 3A
`
`A + (n — l)A .——* nA (critical cluster A1)
`
`In Mullin’s model. further molecular additions to the critical cluster results in
`nucleation.
`
`, A,,, in a system of
`. .
`A solution or melt can contain a variety of clusters Al, .
`competing equilbria. Each cluster in turn is a potential critical cluster for the nucleus
`of one or more polymorphic crystal modifications. In the context. of polymorphic
`structures, in particular those which crystallize under similar conditions, there must
`be a number of processes of this type, all involved in competing equilibria. This
`
`Page 8 of 37
`
`

`
`AGGREGATION AND NUCLEATION
`
`59
`
`is the idea behind Etter’s (1991) extension of this model, describing the clusters as
`aggregates,
`
`Polymorphic form 1
`
`/\
`/ \ Polymorphic form 2
`
`Aggregate l
`
`Molecules
`
`/
`
`Aggregate 2
`
`/\
`
`Polymorphic form 3
`
`Polymorphic form 4
`
`which must contain the structural essence of the eventual crystal structure(s), and are
`therefore likely to be dominated by the same intermolecular interactions. Because
`such a system involves multiple equilibria, once nucleation occurs for one of the
`polymorphic forms, the equilibrium will be displaced in favour of that form at the
`expense of other forms. On a qualitative basis, this demonstrates the competition
`between kinetic and themiodynamic factors. For instance, even if Polymorphic Form
`1 were the thermodynamically most stable one, Polymorph 3 might be the only one
`obtained if Aggregate 2 nucleated crystal growth faster than Aggregate 1. When these
`factors are equal, or very nearly so, then two or more modifications may result from
`the same aggregate or from different aggregates, leading to concomitant polymorphs
`(see Section 3.5).
`Twenty years before Etter’s model for competing aggregate structures in the forma-
`tion of polyrnorphs, Powers (1971) clearly stated the fundamental question regarding
`the challenge of understanding the nucleation process:
`
`It would appear almost certain that the development of these at least transient aggregates are
`the precursors to the development of phase transitions to the ordered solid nucleus, in harmony
`with the local thermodynamic conditions. Yet though a vast amount of study——conferences.
`books, etc.—have been devoted to nucleation, it is still uncertain how this last transition takes
`place.
`
`Some experimental evidence for the presence of different aggregates in solu-
`tion leading to polymorphic structures has been presented recently by Niither et al.
`(1996a,b), and there have been attempts to relate nucleation rates with proposed struc-
`tures in solution based on molecular modelling (Petit et al. 1994). The number of
`studies of this nature is sure to increase with increasing sensitivity and sophistication
`of both experimental and computational tools; those are the kinds of investigations
`that can provide answers to Powers’ challenge.
`From a practical point of view, control over nucleation, and in cases of polymorphic
`systems, control over the polymorph obtained as a result of nucleation, has been
`
`Page 9 of 37
`
`

`
`70
`
`CONTROLLING THE POLYMORPHIC FORM OBTAINED
`
`the concern of those industries for which crystallization is a crucial or final step
`in the production process: for example, sugars, amino acids. pharmaceuticals, and
`fatty acids. A number of examples of studies regarding polymorphic variation and
`preferences for nucleation may be found in the literature from those disciplines.
`The initiation of crystal growth has been a problem for the sugar industry since
`its infancy. Aqueous solutions of sugars often tend to form syrups—indeed, that has
`become one form of marketing, although clearly not the preferred one. As noted above,
`Powers has reviewed the role of nucleation in the sugar industry, including much of
`the accumulated experience involving sucrose. As in many industries, successful
`techniques developed over decades or centuries and were considered trade secrets or
`even commonplace practice without being scientifically recognized or understood.
`Thus, traditionally, sucrose crystallizations, carried out from the huge copper vats in
`which solutions were concentrated, were initiated (i.e. nucleated) by the mechanical
`shock of hammering on the vat (Fig. 3.1). Yet it was only in 1912 that Young (1911)
`described mechanical shock as a factor in nucleating supersaturated solutions (Powers
`1971). Two other curiosities relating to nucleation of sugars demonstrate some of
`the difficulties encountered. Turanose was long considered to be a liquid at room
`temperature (Powers 1971), until it spontaneously crystallized; following that event
`fresh batches of the material always crystallized. In another case more closely related
`to polymorphism, cx—D—mannose had been prepared routinely until the appearance of
`[3-D—mannose, following which the on form could not be induced to crystallize in the
`same laboratory (Levene 1935; Dunitz and Bernstein 1995). As Powers noted, both
`of these cases can be attributed to unintentional seeding, a topic treated in more detail
`in Section 3.6.
`
`Black and Davey (1988) describe a number of the interrelationships and practical
`aspects of the control of nucleation, crystal growth. and polymorphic transforma-
`tion of amino acids. The factors described and demonstrated for primary nucleation
`of L-glutamic acid include temperature, critical nucleus, relationship of interfaeial
`tension to solubility, thermal history, induction time, agitation, and effect of additive.
`Kitamura (I989) studied many of these nucleation factors in the competitive crys-
`tallization of the at and fl forms of L—glutamic acid. He found that at 25 “C only the or
`modification nucleates and grows. In this system, at least, the effect of temperature
`on the relative nucleation rates of the two polymorphs is more ‘remarkable’ than the
`effect of the supersaturation ratio: as the temperature is increased with a constant
`supersaturation ratio, the amount of 0: decreases. He also reported that the ,6 form
`tends to nucleate in stagnant solutions, while at 25 “C essentially only 0! nucleated
`homogeneously.
`In an example from the pharmaceutical industry Sudo et al. (1991) studied the rel-
`ative nucleation propertics of forms A and B ofcimetidine, which is reported to have
`four polymorphic non—solvated forms and three polymorphic monohydrates. Modi-
`lication A is preferred for pharmaceutical formulations. The ‘waiting time method’
`was used to study the primary nucleation process (Harano and Oota 1978), mainly for
`competitive crystallization of the A and B modifications. A is a therrnodynarnically
`metastable form and is more soluble than B in any solvent. At high supersaturation
`
`Page 10 of 37
`
`

`
`AGGREGATION AND NUCLEATION
`
`7|
`
`‘:I...'.;‘ll_Jlll.-l
`'
`-.z
`‘up
`;"'v‘~‘
`".*u‘|:s:lI‘,-I--.'-A »-/
`,',ll{lIi."‘i Slut" » /~ "
`
`W -'
`Wlli'
`
`ll
`tit‘:-‘ll *
`
`0
`
`»CRYST At_LlZll\lCi PAl\lS
`
`Fig. 3.1 Detail from an 1850 drawing of the London sugar refineries of Messrs. Fairrie and
`Co.. showing a copper erystallizing pan For sugar. The worker to the right of the pan is holding
`a mallet which was used to hang on the pan to induce nucleation of the crystallization process.
`(Reproduced from Fairrie l925, with permission.)
`
`in the presence or absence of seeds. At
`(SA 2 4.5) modification A is obtained.
`SA 2 3.6, A was obtained regardless of the form of seed. At SA 5 2.0 the form of
`the seed determined the form obtained.
`Stcaric acid is often considered a prototype for the long chain acids used in many
`processes and applications. Sato and Boistelle (1984) studied the occurrence and
`crystallization behaviour of three of the polymorphic modifications (A. B, and C) by
`varying conditions such as temperature. supersaturation, and solvent from which they
`determined occurrence domains for the existence of the three forms. Polymorph A
`is thermodynamically unstable at all temperatures studied; below '30 “(I form B is
`most stable, while form. C is more stable above 30°C. Forms A and C nucleate
`
`L
`
`Page 11 of 37
`
`

`
`72
`
`CONTROLLING THE. POLYMORPHIC FORM OBTAINED
`
`preferentially from non-polar solvents at high supersaturation, whereas pol ymorph B
`nucleates more readily at lower supersaturations. Nucleation of Form B is preferred
`at higher supcrsaturations from polar solvents. This solvent effect could be influenced
`by the rate of stirring. lf stirring is sufficiently ‘violent’ it increases nucleation, which
`enhances the formation of Form B.
`
`The mere existence of polymorphic structures can be used as a probe of the
`nucleation process. For instance in considering the aggregation process in super-
`saturated solutions of 2,6—dihydroxybenzoic acid, Davey et al. (2000) found a direct
`link between the relative occurrence of two polymorphic forms (from toluene and
`chloroform solutions) and the solvent-reduced self-assembly (aggregation) of the
`molecule.
`
`Nucleation from the melt has been studied for palm oil, composed of triglycerides
`of palmitic and oleic acids, and exhibiting at least three polymorphs (van Putte and
`Bakker 1987). Nucleation curves (induction time t vs temperature T) of palm oil and
`palm stearin show discontinuities at 297 and 306 °C respectively, indicating the onset
`of nucleation, and the demarcation of the occurrence of the polymorphs, as confirmed
`by isothermal Differential Scanning Calorimetry (DSC) studies (Ng l990a,b).
`
`3.3 Thermodynamic vs kinetic crystallization conditions
`
`Physical organic chemists have long been accustomed to making the distinction
`between ‘thermodynamic’ and ‘kinetic’ conditions when referring to reactions and
`reaction mechanisms. In chemical parlance, thermodynamic conditions essentially
`means those conditions under which thermodynamic equilibrium is maintained or
`very nearly maintained. Kinetic conditions refer to situations that are far from
`equilibrium (van Hook 1961).
`In terms of crystallization (Ward 1997), thermodynamic conditions might refer to
`a slow evaporation, a very slow cooling, a slow crystallization from the melt at a
`constant temperature only slightly below the melting point, a slow sublimation for
`which there is only a small difference between the temperature of the solid and that
`of the cold finger on which the sublimate is crystallizing, etc. On the other hand,
`kinetic conditions might refer to a high degree of supersaturation, rapid cooling of a
`solution or melt, rapid evaporation of solvent, large temperature difference between
`the sample and cold finger in a sublimation, etc. A number of examples will serve to
`demonstrate how these principles have been applied to the crystallization of different
`modifications in polymorphic systems.
`
`p-chlorobenzylidene—N—p—chloroaniline 3-1 is dimorphic (Bernstein and Schmidt
`1972; Bernstein and lzak l976). The thermodynamically more-stable orthorhombic
`
`3-I
`
`Page 12 of 37
`
`

`
`THERMODYNAMIC VS KlNETIC CRYSTALLIZATION CONDITIONS
`
`73
`
`form may be obtained by slow evaporation of a solution in which ethanol or methylc.y-
`elohexane is the solvent. Prismatic crystals usually grow in hours to days, depending
`on the initial concentration of the solution. The metastable Iriclinic form is obtained by
`dissolving the maximum amount of substance in a boiling ethanolic solution, which
`is then immediately placed in a desiccator freshly charged with calcium chloride.
`Needle-like crystals appear within minutes, and the desiccant aids in accelerating
`the rate of evaporation. The crystals are metastable and may begin to spontaneously
`transform to the orthorhombic modification in periods ranging from hours to days.
`Berman et al. (1968) noted that ‘mannitol is unusual among carbohydrates in that
`exists in several polymorphic forms’, indicating that a number of these are often
`obtained simultaneously. They describe the preparation of a number of these mod-
`ifications. The or form is obtained by slow crystallization from 96 per cent ethanol,
`the or’ form by evaporation from 100 per cent ethanol and the fl form from aqueous
`ethanolic solutions, all apparently under thermodynamic conditions. On the other
`hand the 1/ form is obtained kinetically by rapid cooling ofa l:l water—ethanol solu-
`tion. An additional Ic form was obtained (unexpectedly) upon evaporation of a boric
`acid/methanol solution (Kim et al. 1968).
`Bock has studied a number of systems in which different polymorphs were obtained
`under thermodynamic and kinetic conditions. (2-pyridyl)(2—pyrimidyl)amine 3-11 is
`dimorphic. Modification I is readily crystallized thermodynamically ‘from any sol-
`vent’ (toluene was actually used) while modification II is obtained kinetically by fast
`evaporation of an ethereal solution or by resolidification of the melt (Bock et al. 1997).
`
`H I
`
`N
`
`N
`\‘/ \l
`/
`Q’ Q ..
`
`N
`
`HC
`3
`
`EGCII
`
`ocrg
`
`0....
`
`3-"
`
`3-"!
`
`In 2,3,7,8—tetramethoxythianthrene 3-11], the less stable (lower in both density and
`absolute value of lattice energy) monoclinic modification is obtained under kinetic
`conditions: rapid crystallization from polar diisopropyl ether, whereas the more stable
`(higher density and lattice energy) orthorhombic modification is thermodynamically
`obtained from a non—polar hydrocarbon solvent.
`In pharmaceutical applications the choice of polymorphic modification for formu-
`lation depends very much on the robustness of the crystallization process as well
`as the properties and characteristics of the preferred modification. Hence, consider-
`able effort is expended in gaining control over the polymorphic form obtained under
`various conditions. As noted above, up to four polymorphic modifications and three
`monohydrates have been reported for cimetidine (SmithKline Beeeham's Tagamet®)
`(Bavin et al. 1979; Prodic—Kojic er al. 1979; Shibata et al. i983; Hegedus and Gorog
`1985). In experiments to selectively crystallize the A form in preference to the more
`stable B congener it was found that with isopropanol as a solvent, A crystallizes
`exclusively at high supersaturation, in the presence or absence of seeds (Sudo et al.
`1991).
`
`L
`Page 13 of 37
`
`

`
`74
`
`CONTROLLING THE POLYMORPHIC FORM OBTAINED
`
`In another example, an antiarrhythmic under development (McCauley er al. 1993)
`was shown to exist in two anhydrous polymorphs,
`two dihydrated enantiotropic
`polymorphs, a monohydrate, and the solvates of several organic solvents. Following
`characterization of all of these modifications it was desired to selectively obtain one
`of the dihydrates, termed modification A, which is thermodynamically less stable at
`room temperature than another dihydrate, D, in contact with aqueous solutions, but A
`is more stable over a wider range of relative humidities. The enantiotropic transition
`point between these two crystal modifications is 37 °C. Procedures were developed for
`obtaining/\ preferentially. Above the transition point a themiodynamic crystallization
`is carried out at 50 °C. using typeA seeds as an added precaution to force the crystal-
`lization to type A. The desired typeA can also be obtained under kinetic conditions by
`spontaneous crystallization below the transition point followed by rapid filtration and
`removal of excess water. The latter procedure prevents a transformation from the A
`state (metastable below the transition temperature) to the D form in the crystallization
`medium. Similar considerations were applied to develop procedures for the selective
`crystallization of the oz and ,3 modifications of glutamic acid (Kitamura 1989).
`
`3.4 Monotropism, enantiotropism, and crystallization strategy
`
`The