Analyst, October 1995, Vol. 120
`
`2435
`
`Analysis of Organic Polymorphs
`A Review
`
`Terence L. Threlfall,
`Chemistry Department, University of York, Heslington, York, UK YO1 5DD
`
`Summary of Contents
`Introduction and Definition of Polymorphism
`Significance of Polymorphism
`Distinction From Related Phenomena
`Stability of Polymorphs
`Methods for the Examination of Polymorphs
`Microscopy
`Infrared Spectroscopy
`Raman Spectroscopy
`Ultraviolet and Fluorescence Spectroscopy
`Solid-state Nuclear Magnetic Resonance and Nuclear
`Quadrupole Resonance Spectroscopy
`X-ray Crystallography
`Thermal Analysis
`Solubility and Density Measurement
`Solvates
`Quantitative Aspects
`Amorphous and Crystalline Solids
`References
`Keywords: Polymorphism; phase transitions; amorphous
`materials; solvates; microscopy; thermal analysis; infrared
`spectroscopy; Raman spectroscopy; solid-state nuclear
`magnetic resonance spectroscopy; X-ray diflraction
`
`Introduction and Definition of Polymorphism
`
`Polymorphism1-7 in the chemical sense of the word* is a
`phenomenon of the solid state, associated with the structure of
`the solid. It has proved difficult to define precisely although the
`basic concept is readily understood. The definitions which have
`been offered vary in breadth but the implication of all of them
`is that polymorphs involve different packings of the same
`molecules in the solid.4 The question of how similar the same
`molecules must be and of how dissimilar the different packing
`arrangements must be in order to qualify as polymorphs is more
`than a matter of semantics but goes to the root of our
`understanding of the organic molecular solid state.
`McCrone has defined a polymorph as ‘a solid crystalline
`phase of a given compound resulting from the possibility of at
`least two crystalline arrangements of the molecules of that
`compound in the solid state’ and has listed those types of solid
`phenomena which are excluded from this definition. Later
`writers who have accepted this definition have tended to
`substitute their own list of exclusions,5 if they have addressed
`the matter at all. Buerger ’s tentative definition3 ‘ideally, two
`polymorphs are different forms of the same chemical compound
`which have distinctive properties’ is broader and appears not to
`
`* An on-line search of Chemical Abstracts will reveal more than 47000 entries under
`‘polymorphism’. Over 90% of these relate to genetic polymorphism, which at least in
`its origins can claim the true etymology of the word. Some selectivity between
`biological and chemical uses can be achieved, but there is no certain searching strategy.
`Searching under ‘phase transition’ and related concepts will generate a further 44000
`entries, most of which refer to inorganic systems, and cannot be easily disentangled.
`Nevertheless, these represent only a proportion of the papers containing information on
`polymorphs and polymorphism. Hence it is not possible to state how many
`publications relate to those aspects of polymorphism described here.
`
`accept. the need for separate phases and to include amorphous
`forms. The nature of the amorphous state899 will be discussed
`later.
`Polytypismlo is one-dimensional polymorphism, referring to
`different stacking of the same layers. It is most familiar in
`inorganic systems, particularly silicon carbide, but has been
`recognized in organic crystals, both as orderedll-13 and as
`disordered stacking.14 There is no special term for two-
`dimensional polymorphism, although some liquid crystal
`systems display it. Liquid crystals are notorious for their ability
`to exist in different phases both in the mesomorphic and in the
`solid state15-17 and this has led to the suggestion that the term
`polymorphism should apply to liquids as well as solids,’* but it
`is only the solid dimensions of liquid crystals which can adopt
`distinct packing arrangements. Liquid-crystal polymorphism
`will not be dealt with specifically in this review except where it
`is related to the polymorphism of solids. The long standing
`questionlg of whether allotropy and polymorphism are dis-
`tinct20 is not an issue in the case of organic compounds.
`Inorganic polymorphs have been excluded because the ex-
`tended structures of which most inorganic crystals are com-
`posed raise concepts not discussed here.21.22 Protein polymor-
`phism usually refers to minor molecular sequence changes23924
`rather than to packing, but different crystal packing of protein
`molecules is also known? Polymorphism of thin films26327 and
`polymers, both of biologica128,29 and of synthetic30 origin,
`although of the same nature as the concept of polymorphism
`considered here, will not be discussed.
`There is a profusion of words in the English language for the
`phenomena discussed in this review, yet not enough because of
`the overlapping usage. ‘Polymorph’ (dimorph, trimorph) ‘form’
`and ‘modification’ are all used to describe polymorphic phases,
`but ‘form’ and ‘modification’ are also used in reference to
`crystal habit. ‘Polymorph’ and ‘form’ have been used to
`describe solvates, whilst ‘pseudopolymorph’ doubles for both
`solvates and for those solids which are otherwise not considered
`true polymorphic forms. The term ‘pseudopolymorphic solvate’
`applied to crystals losing solvent molecules without change of
`crystalline form offers yet another source of confusion in
`terminology. Genetic polymorphism which is now the major use
`of the term is often described as ‘polymorphisms’ but this is
`occasionally seen also in chemical senses. In view of the almost
`universal use of ‘polymorphic’ as the appropriate adjective, the
`word ‘polymorphous ’ seems superfluous despite dictionary
`support. There is an urgent need for consistent usages so as to be
`able to delineate the phenomena under consideration.
`There is no clear choice as to the best method of designating
`polymorphs. Arbitrary systems are to be discouraged, but
`numbering based either on order of melting point or of room
`temperature stability have been recommended. Both are
`susceptible to change as a result of later identification of new
`polymorphic forms. Numbering based on order of discovery is
`unchangeable, but requires a knowledge of the history of the
`compound. The addition of the crystal class, as has been
`suggested for minerals31 is not very practicable, since crystal-
`lographic classes are rarely determined from optical micro-
`scopic or X-ray powder diffraction studies for organic com-
`pounds. The assignment of a space group is even less realistic.
`
`Argentum EX1042
`
`Published on 01 January 1995. Purchased by rosalie.beard@kmob.com on 09 December 2013.
`
`View Article Online
`
` / Journal Homepage
`
` / Table of Contents for this issue
`
`Page 1
`
`

`

`2436
`
`Analyst, October 1995, Vol. 120
`
`In any case the distribution of organic molecules amongst
`crystal classes and space groups is extremely limited, as is
`discussed later.32.33 The addition of a melting or upper transition
`point to a Roman numeral is probably the best compromise,l
`although care must be taken to distinguish the melting point of
`the polymorph and that of the transformed product.
`
`Significance of Polymorphism
`The continuing investigation of polymorphism by the Innsbruck
`school (Kofler, Kuhnert-Brandstatter, Burger) over more than
`half a century has shown that around one-third of organic
`substances show crystalline polymorphism under normal pres-
`sure c0nditions.3~,35 A further third are capable of forming
`hydrates and other solvates.
`Much of the literature on the polymorphism of organic
`compounds relates to pharmaceutical products.l,3~0 The
`incentive for this interest in polymorphism began with the need
`to satisfy regulatory authorities in various countries as to the
`bioavailability of formulations of new chemical entities.36.37 Of
`the several contributory factors to the bioavailability of finished
`products, the inherent solubility and rate of dissolution of the
`drug substance itself are of major importance. The solubility is
`dependent on the polymorphic state, as different polymorphs
`have different energies and therefore different solubilities.40 It
`has been pointed out, particularly by Burger,36 that the
`difference in solubility between polymorphs is likely to result in
`significant bioavailability differences, in practice, only in
`exceptional cases. Although some may think that this represents
`an extreme view, the consequences of polymorphism on
`bioavailability are commonly overstated. Chloramphenicol
`is
`palmitate, over which the original concerns were
`unique in that the solubility is related to the rate of enzymic
`attack on the s0lid.4~ This and novobiocin,43 which involves
`consideration of the amorphous state, are among the handful of
`examples of marketed products showing major bioavailability
`differences as a result of polymorphism.
`As formulations have become more sophisticated and as the
`tolerances on products have become tighter, the need to identify
`polymorphic behaviour at an early stage of development has
`become important in the pharmaceutical industry as a means of
`ensuring reliable and robust processes44 and conformity with
`good manufacturing practice. The aim is to avoid, inter alia,
`
`tabletting problems and subsequent tablet f a i l ~ r e , ~ ~ , ~ ~ crystal
`growth in suspension^^^^^^ and resultant caking, precipitation
`from solutions and problems with ~uppositories,~9 as well as
`chemical production problems such as filtrability and to ensure
`analytical reproducibility. By extension such considerations
`relate to the control of quality in manufacture and product
`reliability in any industry by ensuring that the processes are well
`understood and under control so that unpleasant surprises do not
` occur.^^ This point is most dramatically illustrated in the
`explosives industry, where the wrong polymorph can have
`greatly increased sensitivity to detonation.51.52 Pigment colour
`and solubility are polymorph dependent,53-5’ as are photo-
`graphic and photolithographic sensitizers.6O The performance of
`industrial products, particularly those based on natural fats and
`waxes61Jj2 and derived soaps,63 and on petroleum produ~ts6~365
`is in many cases related to polymorphic composition and degree
`of crystallinity. The same is true of the processing, acceptability
`and deterioration of foods and confectionery containing
`fats,66,67 sugars,68-7* polysaccharides73 and other constitu-
`ents.74-75 A comprehensive summary of the solid-state proper-
`ties of lipids has recently appeared.76
`It is also worth establishing the polymorphic behaviour of a
`compound for the sake of good order in documentation so that
`reference works, for example, pharmacopoeias, do not contain
`conflicting data34.77 such as a spectrum of one polymorph, but
`the melting point of another.
`
`A major incentive to the study of polymorphism in the
`pharmaceutical industry during development has become
`strikingly apparent recently in the use of subsidiary patents on
`desirable polymorphic
`to prolong the patent life of
`major products. Much recent pharmaceutical patent litigation
`has concerned polymorphs and particular interest has been
`taken in Glaxo’s patent on the polymorph of ranitidine79
`(Zantac) which if held valid will extend the patent protection
`from 1995 to 2002 in many countries.80 For a compound with
`annual sales of over 2 400 million pounds sterling,gl the
`financial incentives to investigate polymorphs are obvious.
`Finally, the very existence of polymorphism tells us some-
`thing about the solid-state. Investigation of polymorphic
`systems, especially those with a large number of forms can help
`in understanding solid-state and molecular behaviour and
`intermolecular interactions82 and the relationship between
`crystal structure, crystal growth and crystal habits3 and their
`influence on bulk properties. Apart from knowledge for its own
`sake, this is of clear application in the development of organic
`electronicS4,85 and other specialty productsgcg8 and in under-
`standing the function of biological membranes.89
`
`Distinction From Related Phenomena
`At one time polymorphism was regarded only as different
`arrangements of rigid molecules in the solid ~tate.gO,~l* A clear
`dichotomy existed between this and arrangements of molecules
`in different forms, such as could be imagined would occur with
`isomeric, tautomeric, zwitterionic and chiral structures and later
`with different conformers.92 The early crystallographic studies
`on rigid aromatic molecules tended to reinforce the distinction.
`This simple division could only be maintained whilst details of
`the rich variety of solid-state structures were inaccessible. The
`early examples of dynamic isomerism and tautomerism were
`f e ~ 9 3 - 9 ~ and the proposition that they could not be part of
`polymorphism was copied by reviewers until even the examples
`were f0rgotten.9~ A quoted example of a tautomeric solid-state
`structure, that of 3,5-dichloro-2,6-dihydroxy dimethyl tere-
`phthalic acid was shown in 1972 not to be tautomeric, but to
`involve conformational change with hydrogen bonding differ-
`ences.96 One would have expected examples of tautomerically
`related solid structures to be exceedingly numerous, since the
`molecular energetic requirements can easily be fulfilled as is
`shown by the widespread occurrence of tautomerism in
`solution.97 Tautomeric polymorphism is surprisingly rare, but a
`well investigated example is now known, that of 2-amino-
`3 -hydroxy-6-phenylazopyridine .98
`There are a few papers in the literature either where
`tautomeric polymorphism is invoked99-105 or where examina-
`tion of the IR spectra is suggestive of forms whose difference
`resides in transfer of hydrogen between one part of the molecule
`and another. 106 The instances of 1,3-~yclohexadienone and
`squaric acid (3,4-dihydroxy-3-cyclobutene- 1,2-dione are more
`difficult to place unambiguously in the category of tautomeric
`polymorphism. Proton transfer between donor and acceptor
`oxygen sites results in little change in over-all structure. lo7
`Both tautomeric equilibrium and the neutral ++ zwitterionic
`equilibrium formally involve such an intramolecular hydrogen
`transfer. The nominal difference is that a charge separation is
`produced in zwitterions which cannot be extinguished intra-
`molecularly by a double-bond rearrangement cascade. The
`difference may be even smaller in practice because charge
`stabilization of zwitterions can occur intermolecularly, for
`example, in solution through solvation, whilst tautomeric
`structures can retain a substantial part of their charge as shown
`by dipole moment and IR spectroscopic studies.108JO9 Anthra-
`
`* Earlier literature can be accessed Ilia references I , 2 and 10.
`
`Published on 01 January 1995. Purchased by rosalie.beard@kmob.com on 09 December 2013.
`
`View Article Online
`
`Page 2
`
`

`

`nilic acid exists as two metastable forms containing only
`uncharged molecules and a form stable at room temperature,
`half the molecules of which have been shown from crystallo-
`graphic studies to be zwitterionic and half uncharged.110 A
`related phenomenon is the changing of allegiance of hydrogen-
`bonded hydrogens between electron donor atoms, which is a
`prolific source of polymorphism. I 11 The role of hydrogen-
`bonding networks in determining crystal structure has been
`discussed extensively. 1 12 Conformational differences between
`molecules of different structures have been admitted, perhaps
`reluctantly, and distinguished by the title conformational
`polymorphism.113 The original examples form one extremity
`where molecules in distinctive conformations pack similarly,9*
`but it is now obvious from the plethora of crystal structures, as
`could always have been deduced from elementary considera-
`tions of energy minimization, that any change of packing will
`cause geometrical change in molecules and conversely that any
`change in geometry will invite different packing of the
`molecules.82 The extent will depend on the rigidity of the
`molecules. Although some floppy ring systems maintain their
`shape in different forms] 14,115 even nominally rigid structures
`such as the ring systems of steroids116 can show substantially
`different conformations in different polymorphs. Heteroaro-
`matic117-121* and benzoquinone122 planes are frequently bent
`and even benzene rings123 may be. Thus it seems pragmatic to
`accept conformational polymorphism as a normal sub-set of
`polymorphism and the term will only be used here when it is
`necessary to distinguish cases of substantial conformational
`change.
`The distinction between polymorphism and chirality is made
`in most accounts of polymorphism; yet it has recently been
`pointed out that if conformational polymorphism is accepted,
`then racemates and conglomerates of rapidly interconverting
`chiral systems are in fact polymorphs.5 Such systems are
`generally ones with an easy conformational change where the
`trivial distinguishing feature from other conformational poly-
`morphism is that the result of such a change is a reflection of an
`asymmetrical structure across a mirror plane. Although this
`seems difficult to accept, the dextrorotatory and laevorotatory
`forms of such systems are then equally p01ymorphs.l~~ The
`narrow line of demarkation between polymorphism, conforma-
`tional polymorphism and chirality first seems to have been
`recognized by Eistert et al..l*5 Examples of rapidly inter-
`changing enantiomers in solution capable of independent
`existence in the solid state are known126-127 but uncommon.
`A further extension of the concept of conformational
`polymorphism is to be found where there is rapid interconver-
`sion between isomers.l28 As in the chiral examples, one
`molecular species or the other becomes exclusively incorpor-
`ated in the crystal because the mechanism of crystal growth acts
`as such an exquisitely discriminatory process.
`Since a hydrate and an anhydrous form are constitutionally
`distinct, they cannot bear a strictly polymorphic relationship on
`the basis of any definition. However, the observation of material
`of different melting point or other properties during re-
`crystallization may be due (apart from chemical reaction with
`solvent or decomposition) to solvation or polymorphism and the
`methods of examination are similar in either case. Hence the
`term 'pseudopolymorphism' has become common 3o particu-
`larly in the pharmaceutical industry. The term seems un-
`necessary and could lead to confusion131 with its use to describe
`all other phenomena related to polymorphism] and so will not
`be used here. It must be emphasized, however, that the
`distinction between solvates and polymorphs is not as clear-cut
`as might be imagined, either conceptually or practically.
`
`* In the case of phenothiazines'zl the point of interest is not that the ring system is bent,
`but that the heteroatoms are out of the plane of the aromatic rings and in the opposite
`sense to expectation.
`
`Analyst, October 1995, Vol. 120
`
`2437
`
`The traditional narrow view of polymorphism, rigidly
`excluding chirality and isomerism, has caused considerable
`difficulty128 to the investigators of the systems described above
`and it is suggested that the way to avoid these problems is to
`adopt the gloss originally proposed by McCrone and co-
`workers1.37 on his definition of polymorphism, namely that the
`criterion is that the component molecules must have the same
`structure in solution irrespective of the polymorph from which
`they were derived; but, as has been suggested by D ~ n i t z , ~
`without excluding tautomerism, isomerism or conformers per
`se. Thus, rapidly interconverting species would be accepted,
`whilst slowly interconverting species would be excluded, as
`was surely within the original contemplation. Despite appear-
`ances, this proposal is likely to multiply examples of poly-
`morphism very little and it avoids what otherwise must be
`artificial situations of accepting phases as polymorphs based on
`impeccable polymorph behaviour until their crystal structure
`reveals excluded molecular forms.98.l 10~132 If, as asserted, the
`transition between polymorph I and polymorph I1 of 1,3-cyclo-
`hexadiene occurs by transfer of hydrogen from one oxygen to
`another, then this is nominally an example of tautomeric
`polymorphism.107 If, on the other hand, the same change occurs
`or can be made to occur by a movement of the whole molecule
`then it is an example of regular polymorphism. The boundaries
`between the various alternative solid structural concepts are too
`subtle and too vague to be used to define polymorphism.
`Although the requirement of the same structure in solution
`has been canvassed above, one-component phase diagrams are
`constructed on the basis of equilibrium with vapour, rather than
`liquid. It is just in the instance of conformational, configura-
`tional or hydrogen mobility that molecular differences between
`vapour,133,134 melt, solutionl26,135 and solid are found. The
`mobilities are inevitably of different magnitudes in different
`states. We shall be increasingly obliged to decide where to draw
`the boundaries of polymorphism as more comparative studies
`involving polymorphs and molecular structure in different
`states are undertaken.
`One negative consequence of accepting the wider view of
`polymorphism should be noted, namely that the thermodynamic
`relationships discussed later are likely to be less certain for the
`wider polymorphic farnily.9O
`
`Stability of Polymorphs
`Polymorphs, or strictly dimorphs where only two forms are
`under consideration, may be in an enantiotropic or monotropic
`relationship.19.136 An enantiotropic relationship implies that
`each form has a range of temperature over which it is stable with
`respect to the other and a transition point at which the forms are
`equistable and in principle interconvertible. 137 Above that
`temperature the thermodynamic tendency is to the formation
`exclusively of the form stable at the higher temperature. Below
`the transition temperature the low- temperature form is the only
`stable one with respect to the other, although there is usually a
`greater tendency for the high temperature form to become
`frozen-in than for a low- temperature form to persist beyond its
`stability range.8 Forms outside their range of stability are
`described here as metastablel38. In the case of a monotropic
`relationship one form is metastable with respect to another at all
`temperatures. There is no observable transition point, although
`the thermodynamic description implies a theoretical transition
`point above the melting point which is therefore unattainable. 139
`The use of the terms enantiotropic or monotropic in reference to
`a phase, as opposed to a transition, is ambiguous and likely to
`lead to confusion, since a polymorph can have a monotropic
`relationship to a second polymorph, but be enantiotropic in
`relation to a third polymorph. Flufenamic acid provides such an
`example. 140 The distinction between thermodynamic and
`kinetic transition points also needs to be drawn.141
`
`Published on 01 January 1995. Purchased by rosalie.beard@kmob.com on 09 December 2013.
`
`View Article Online
`
`Page 3
`
`

`

`2438
`
`Analyst, October 1995, Vol. 120
`
`Polymorphs only exist in the solid state: melting or
`dissolution destroys any distinctions. It is therefore important in
`examining polymorphs analytically not to submit them to
`conditions under which they melt, dissolve or are rendered more
`likely to interconvert. Heating and gri11ding142-1~4 are obviously
`potentially hazardous operations in this context, but often
`cannot be avoided. The presence of solvent, even one in which
`the substance appears insoluble, will speed up the inter-
`conversion.145 Trace moisture, acid or alkali on vessels can be
`similarly effective in interconverting polymorphs or in catalys-
`ing competing and confusing phenomena such as ring-opening
`reactions, for example, in 3,5-dihydroxy-3-methylvaleric acid
`derivatives,l46, or group transfer reactions. 147
`It might be supposed that a transition during grinding would
`always be from less stable polymorph to the polymorph more
`stable at that temperature, but in our experience, as well as from
`the literature,145 this is not always true, presumably because the
`transformation takes place at a local temperature generated by
`the grinding and the unstable form becomes frozen-in by rapid
`cooling outside the immediate area of grinding.148 This can only
`occur in cases in which the transition temperature does not lie
`too far above ambient. There may be alternative explanations,
`namely interconversion via amorphization or that a less stable
`polymorph may become the more stable one when in the form
`of small crystallites, as a result of surface effects. The latter
`phenomenon has been observed and investigated theoretically
`in the case of phthalocyanine pigments.149 The possibility of
`growing unstable forms in microdrop conditions has been
`known for some time,34 but recently the value of emulsions for
`this purpose has been suggested.150 Although it would be
`desirable to have more compelling evidence than that obtained
`by differential scanning calorimetry (DSC) alone to establish
`the relationship between forms grown in this way, it does appear
`that new forms can be produced as well as metastable ones
`which are otherwise only accessible via the melt. The product of
`a polymorphic transition can also depend on particle
`size. 15 1,152
`Mnyukh and Petropavlov, in extensive studies of the
`transformation of individual crystals, observed that strict
`orientation of axes between mother and daughter phases was not
`preserved upon transformation.153 They have concluded that
`only reconstructive transitions, i.e., those involving the growth
`of new crystals in place of the old, take place for organic
`compounds. Even rapid transitions, described as atypical, were
`observed to follow the same patterns. No displacive (marten-
`sitic, co-operative) mechanism involving concerted structural
`change is therefore possible for organic compounds in
`Mnyukh’s scheme. Whilst it would now appear that the
`reconstructive mechanism is the usual one, there are many
`examples involving preservation of axial orientation at phase
`transitions4 some of which appear to be topotactic rather than
`only epitaxial. 154-157.
`Irrespective of the mechanism and the rate of conversion at
`the point of transition, the stability in practice of a metastable
`polymorph at room temperature varies enormously, 158 from
`examples where the transformation is so rapid that the only
`evidence of the transient existence of a polymorph is its
`pseudomorphic outline, 1 to those which can be kept indefinitely
`and indeed refuse to transform in the absence of heat, high
`humidity or solvents.152 The majority of systems are in fact
`quite robust to handling. It may therefore be thought that some
`of the present work presents over-concern with the possibility of
`transforming polymorphs during analytical examination. How-
`ever, the modifications of some compounds show extraordinary
`sensitivity to handling in so many different ways. For example,
`with octakisphenylthionaphthalene, pressure on a cover-slip
`causes the yellow form to change to red;lsY with ethylenedia-
`mine hydrochloride, mere contact with KBr is stated to cause
`transformation;160 with D,l-pantolactone 2,4-dihydroxy-3,3-di-
`
`methylbutyric acid y-lactone, absorption of IR radiation in the
`spectrometer is sufficient for transformation; 161 and with
`meprobamate, high humidity may rapidly transform an other-
`wise indefinitely stable polymorph.162 The problem is that this
`sensitivity may not be apparent until after the measurements
`have been made and then only if the analyst is alert, so that it is
`not possible to be too careful at the outset. Three of the
`commonest methods, IR spectroscopy, X-ray powder diffrac-
`tion and differential scanning microscopy are unreliable for
`comparison of identity unless the sample is examined as a fine
`powder, but grinding can mislead into belief of identity if it
`induces transformation. This is why optical microscopy is so
`valuable for the initial examination. On the other hand, where
`transformation is sluggish, solubility determinations will be of
`more value than instrumental measurements for establishing the
`stability relation~hips.3~
`The existence of enantiotropically related polymorphs is
`indicative of the fact that the relative stabilities and therefore the
`Gibbs energies of the forms are very similar.163Jw For this
`reason the empirical forecasting of polymorphism of a given
`compound is unlikely to be reliable.88J65 Despite this, groups of
`compounds such as sulfonamides, barbiturates and steroids are
`known to be extraordinarily susceptible to polymorph forma-
`tion.39 Around 70% of these are now known to be polymorphic.
`Other examples include theophylline derivative^,^^ coumar-
`ins,87 alkanes,64,65 fatty acids and their derivatives61362 mol-
`ecules which form liquid crystals, l~~~ and molecules which
`pack badly.166 With the advent of molecular modelling
`techniques for crystal growth prediction, interest has been
`generated in the computer prediction of polymorphism.87 The
`task is difficult because of the lacunae in our understanding of
`polymorph structure.
`
`Methods for the Examination of Polymorphs
`Polymorphs can be sought deliberately by cooling or quenching
`of melts, by condensation of vapour, or by crystallization under
`different conditions, although they are often encountered by
`chance. In the process of crystallization from solution, the
`expected effect of crystallization temperature may be overshad-
`owed by other factors, particularly deliberate or adventitious
`seeds.167 The importance of crystallization control during
`process development and the attitudes when unexpected
`polymorphic forms are encountered has been described by
`Bavin? ‘the process of crystallization is taken for granted by
`most chemists and it takes a reaction vessel clogged with an
`unstirrable mass to provoke serious thought’.
`All the solid-state properties of the different polymorphic
`modifications of a compound will be different, but often only
`marginally so, to the point of instrumental indistinguishability.
`For this reason, it is important to look at potentially poly-
`morphic systems by a variety of techniques to avoid erroneous
`conclusions. Failure to recognize a polymorph is the more
`obvious situation but it is also possible to identify polymorphs
`where none exist, if reliance is placed on too few techniques.168
`Substances with multiple forms can require substantial effort for
`their complete elucidation, especially when previous studies
`have characterized the forms inadequately. 142~48,15191697170
`The techniques which have been available for a long time for
`the examination of polymorphs include those listed in Table 1.
`Which are the commonest methods depends to some extent on
`the area of interest, but in industrial practice, microscopy, IR
`spectroscopy, DSC, X-ray powder diffraction, solubility and
`density measurements have been the most widely used
`techniques. Within the past decade several new techniques and
`instrumental accessories have become widely available. These
`ease the manipulation of polymorphs and so lessen the danger of
`interconversion, or enable new properties to be investigated and
`allow measurements to be made which would have formerly
`
`Published on 01 January 1995. Purchased by rosalie.beard@kmob.com on 09 December 2013.
`
`View Article Online
`
`Page 4
`
`

`

`been impossible on the specimen under examination because of
`its size or microcrystallinity, for example. These developments
`are listed in Table 2. In general, the application of these newer
`techniques to polymorphism has not been adequately reviewed.
`Much of this article will therefore be devoted to a description of
`these methods in relation to examples taken from the literature
`on polymorphism. Some attention will also be devoted to
`aspects of the traditional techniques which have been given
`surprisingly little coverage in the reviews. Apart fom the
`techniques discussed below, there have of course been many
`other methods applied to particular aspects of polymorphism
`and solid-solid phase transitions. Examples include scanning
`
`tunnelling micr0scopy,6~ electron diffra~tion,~~ atomic force
`microscopy,l71 crystal etching,17* electron microscopy64J73
`and thermobarometric measurements. 174
`The analytical strategy in approaching