1148
`
`Chem. Mater. 1994, 6, 1148-1158
`
`Solid-State Pharmaceutical Chemistry
`
`S. R. Byrn,*,t R. R. Pfeiffer,* G. StephensonJ D. J. W. Grant,~ and W. B. Gleason~
`
`Department of Medicinal Chemistry and Pharmacognosy, Purdue University,
`West Lafayette, Indiana, 47907 and Department of Pharmaceutics and Department of
`Laboratory Medicine and Pathology, Biomedical Engineering Center, University of Minnesota,
`Minneapolis, Minnesota 55455
`
`Received February 8, 1994. Revised Manuscript Received June 24, 1994®
`
`Solid-state pharmaceutical chemistry encompasses a wide range of studies on pharmaceutical
`solids including (1) determination of the physical properties of polymorphs and solvates, (2)
`physical transformations between polymorphs and solvates, (3) chemical reactions in the solid
`state, and (4) solid-solid reactions which occur in pharmaceutical preparations. Recent advances
`in this field include improved understanding of crystallization processes, improved understanding
`of the need for characterization of polymorphs and solvates for both control and regulatory
`purposes, and a better understanding of the mechanisms of solid-state degradations and solid-
`solid reactions. This review will briefly describe recent advances in the following areas: (1)
`crystallization and the properties of crystals of pharmaceutical solids; (2) characterizations of
`crystal forms of drugs using solid-state NMR spectroscopy.
`
`The study of the solid-state chemistry of drugs not only
`encompasses many scientific disciplines but also impinges
`on virtually all phases of the pharmaceutical industry,
`from discovery to successful marketing. It is clear that an
`understanding of the molecular structure of the solid state
`can lead to better design and control of drug performance.
`The mission of those working in the field of solid-state
`pharmaceutical chemistry is to provide each drug in a solid
`form that has optimum performance in a given application.
`Pursuit of this mission requires recognition of several
`general, interrelated points: (1) Drugs can exist in a
`number of solid forms, each having different properties
`of pharmaceutical importance, including stability and
`bioavailability; the number and properties of these forms
`are largely unpredictable and vary considerably from case
`to case. (2) The forms of a drug may interconvert under
`various conditions. (3) Once a solid form is chosen for a
`product, methods for analysis and control of the form must
`be devised.
`Let us briefly review each of these points with regard
`to their status in current practice and to some associated
`scientific challenges that remain.
`(1) The most common solid forms that may be found
`for a given drug substance are as follows: crystalline
`polymorphs, forms having the same chemical composition
`but different crystal structures and therefore different
`densities, melting points, solubilities, and many other
`important properties; solvates, forms containing solvent
`molecules within the crystal structure, giving rise to unique
`differences in solubility, response to atmospheric moisture,
`loss of solvent, etc. Sometimes a drug product may be a
`desolvated solvate, formed when solvent is removed from
`a specific solvate while the crystal structure is essentially
`retained--again, many important properties are unique
`to such a form; finally, amorphous solid forms that have
`no long-range molecular order (i.e., no crystallinity) and
`
`t Purdue University.
`t Department of Pharmaceutics, University of Minnesota.
`! Department of Laboratory Medicine and Pathology, University of
`Minnesota.
`¯ Abstract published in Advance ACS Abstracts, August 15, 1994.
`
`which tend to be more soluble, more prone to moisture
`uptake, and less chemically stable than their crystalline
`counterparts (pharmaceutical processing operations may
`produce solids of low crystallinity intermediate between
`that of a crystalline solid and an amorphous solid).
`To be sure, differentiating among the various solid forms
`of a substance is generally a routine matter. A number
`of analytical methods, used together, make this possible.
`Various familiar methods reveal the chemical composition
`ofthe solid and reveal the presence ofsolvent. The physical
`form is further characterized by X-ray powder diffraction,
`infrared and solid-state NMR spectroscopy, differential
`scanning calorimetry, and microscopy. The single most
`valuable piece of information about a crystalline solid,
`although not always available, is the molecular structure,
`determined by single-crystal X-ray diffraction.
`Perhaps the chief challenge in managing the phenom-
`enon of multiple solid forms of drugs is our inability to
`predict how many forms can be expected in a given case:
`too often costly delays are encountered when a less soluble
`solid form suddenly appears late in a development
`program. Progress along these lines awaits Analysis and
`quantification of the myriad intermolecular forces within
`any proposed crystal structure as well as the ability to
`postulate the likely packing modes for a given molecule
`in all its configurations. Further research similar to that
`of Margaret Etter and co-workers, reviewed in section A,
`will doubtless lead to better success in predicting alterna-
`tive solid forms of new drugs.
`A second challenge relates to the strikingly different
`reactivity of different solid forms of many substances,
`whether it be oxidation, dehydration, decarboxylation, or
`other chemical reactions. Kinetics involving the solid state,
`in which specific contacts (or noncontacts) between
`reactive groups are dictated by the structure of a given
`solid form, are apt to be more complex than kinetics in
`solution, where the corresponding molecular encounters
`are much more random. Much the same can be said about
`the relative difficulty of elucidating mechanisms in these
`two states. Moreover, reactions in the solid state may be
`further complicated by other, often unknown factors such
`
`0897-4756/94/2806-1148504.50/0
`
`© 1994 American Chemical Society
`
`Lupin Ex. 1040 (Page 1 of 11)
`
`

`

`Reviews
`
`Chem. Mater., Vol. 6, No. 8, 1994 1149
`
`as nucleation of a reaction product phase, presence of
`various amounts of amorphous component, strain or
`disorder in the crystal structure, and multiplicity of
`reaction products. As a restdt, there is very Httle
`information available to guide in any general way the
`prediction of the stability of compounds in the solid state,
`let alone in their different solid forms or different
`formulations. Thus, the stability of all bot the most rugged
`products is generally characterized empirically at actual
`storage temperatures, and the value of studies at elevated
`temperatures must be assessed case by case. Elucidation
`of the kinetics and mechanisms of solid-state reactions is
`pursued in only the most pressing cases.
`(2) A dictate of formulation technology is that the
`physical form of the drug substance, after being defined
`and verified, should not change once the product has been
`manufactured. Therefore, in addition to identifying
`various solid forms of a drug, an understanding of the
`specific factors that may bring about transformations
`between the forms is also essential.
`Rates of solid-to-solid transformations in drugs are
`affected by one or more of the following variables:
`temperature, solubility in a given liquid phase, and vapor
`pressure of solvent. At a minimum, the investigator must
`therefore first identify and obtain certain physicochemical
`data on each relevant solid form:
`¯ transition temperature(s) between polymorphs
`¯ solubility of the drug in all solvents and solvent
`mixtures used in preparation of the final drug substance
`and throughout the formulation process
`¯ equilibrium water vapor pressure vs composition
`isotherm
`This information reveals which solid form is the most
`physically stable (least soluble) under specific conditions
`of temperature and composition; thus, if that form is
`already the one present, no transformation will occur. If,
`however, a less physically stable (more soluble) form is
`present, the direction of any transformation can now be
`predicted for given conditions. One might well ask why
`the most stable form is not always selected and, indeed,
`that is usually the case. There are however, situations
`where the peculiar properties of a less stable solid form
`are required for a product’s performance--achieving higher
`solubility, for example. In such cases, the manufacturer
`has the added burden of demonstrating the lack of
`transformation to a more stable solid form throughout
`the life ofthe product. As discussed in the previous section,
`this task requires dealing with the problems of solid-state
`kinetics and must usually be approached empirically.
`(3) Two main challenges to the analysis of pharmaceu-
`tical solids are dealing quantitatively with mixtures of
`forms in the drug substance and identifying the solid form
`of the active ingredient in the formulated product,
`particularly when the drug is a minor component in the
`presence of numerous other materials (excipients).
`With this background on the current status of solid-
`state pharmaceutical chemistry, we can now turn to a
`number of recent advances in this field.
`The work of Margaret Etter and co-workers, treated in
`section A, is an example of how present research addresses
`some of the issues of structure prediction raised above.
`The topics discussed in sections B-D have significant
`bearing on many of the issues discussed in all three of the
`introductory subsections.
`
`A. Forces Holding Crystals Together
`
`Two main types of forces are responsible for holding
`drug crystals together: nonbonded interactions and
`hydrogen bonding. Nonbonded interactions occur in all
`crystals while hydrogen bonding is important in many
`compounds, especially pharmaceuticals. Etter has re-
`viewed the extent and types of hydrogen bonding which
`can exist in organic solids.1,2
`Carboxylic acids have been a particularly useful class
`of compounds for investigating alternative hydrogen
`bonding possibilities. For example, in o-alkoxybenzoic
`acids both dimerization and formation of intramolecular
`hydrogen bonding are observed,a In o-anisic acid, dimers
`are observed in the solid state while intramolecular
`hydrogen bonds are observed in dilute solution. However,
`
`0
`
`I
`0..-H
`
`CH3 CH3
`
`Solid State Solution
`
`in o-ethoxybenzoic acid, only intramolecular hydrogen
`bonds are observed in the solid state and in solution.
`
`C~H~
`
`Solution and Solid State
`
`The Etter group studied the hydrogen bonding in
`salicylamide derivatives and pointed out that two types
`of hydrogen bonding are possible in these compounds.4
`One type involves an intramolecular NH...O hydrogen bond
`and an intermolecular C----O...HO-- hydrogen bond and
`the other an intramolecular CffifO...HO-- hydrogen bond
`and an intermolecular NH...O hydrogen bond.
`
`Intra: HN"" O
`
`Inter: Cm 0"" HO
`
`Intra: Cm O--" HO
`
`Inter:
`
`HN--- O
`
`(1) Goerbitz, C. H.; Etter, M. C. Int. J. Pept. Protein Res. 1992, 39,
`93-110.
`(2) Etter, M. C. A¢¢. Chem. Ree. 1~0, 23, 120-6.
`(3) Etter, M. C.; Urbanczyk-Lipkowska, Z.; Fish, P. A.; Panunto, T.
`W.; Baures, P. W.; Frye, J. S. J. Crystallogr. Spectros¢. Ree. 1988, 18,
`311-25.
`
`Lupin Ex. 1040 (Page 2 of 11)
`
`

`

`1150 Chem. Mater., Vol. 6, No. 8, 1994
`
`Reviews
`
`Table I. Reliable and Occasional Hydrogen-Bond Donors
`and Acceptors
`
`type
`
`reliable donor
`
`occasional donor
`reliable acceptors
`
`occasional acceptors
`
`Not COH.
`
`functional g~oup involved
`
`-OH,= -NH2, -NHR, -CONH2, -CONHR,
`-C00H
`-COH, -XH, -SH, -CH
`-C00H, -CONHC0-, -NHCONH-,
`-CON< (1-3=), >P~O, >S~O, -OH
`>O, -NO~, -CN, -CO, -COOR, -N<,-C1
`
`Etter, MacDonald, and Bernstein developed a graph-
`theory-based approach to classifying and symbolically
`representing the different types of hydrogen bonds that
`can be formed.5
`Etter also developed rules governing hydrogen bonding
`in solids. These rules require a classification of hydrogen
`bond donors and acceptors into "reliable" hydrogen-bond
`donors and acceptors and "occasional~ donors and accep-
`tots (Table 1). Using these classifications, three rules were
`devised: (1) All (or as many as possible) good proton donors
`and acceptors are used in hydrogen bonding. (2) Six-
`membered ring intramolecular hydrogen bonds form in
`preference to intermolecular hydrogen bonds. (3) The
`best proton donors and acceptors remaining after intra-
`molecular hydrogen bond formation will form intermo-
`lecular hydrogen bonds. These rules apply quite well to
`hydrogen bonding of small molecules. However, in some
`larger molecules, e.g., erythromycins, steric factors make
`it impossible to satisfy all of the possible hydrogen bonded
`interactions, and some donors and acceptors are not
`involved in any hydrogen bonds.
`Cocrystals. An important aspect of research into
`hydrogen bonding involves the realization that cocrystals
`can be obtained from certain solutions containing more
`than one molecular species. Cocrystals can also be formed
`by mixing or grinding two solids together. Cocrystals are
`usually formed between a hydrogen-bond donor molecule
`and a hydrogen-bond acceptor molecule. The nature of
`hydrogen bonding in cocrystals can also be described using
`the above rules. The cocrystals observed by Etter’s group
`include numerous ureas with ketones, carboxylic acids with
`2-aminopyridine, and also adenine or cytosine combined
`with many acidic organic compounds6 such as carboxylic
`acids and N-acylamino acids. Other classes of cocrystals
`investigated by Etter’s group are
`
`pyrimidine, pyridines: carboxylic acids
`
`pyridine-N-oxides: acids, alcohols, amines
`
`phosphine oxides: acids, amides, alcohols,
`ureas sulfonamides, amines, water
`
`carboxylic acids: other carboxylic acide, amides
`
`m-dinitroureas: acids, ethers, phosphine oxides,
`sulfoxides, nitroanilines
`
`imides: other imides, amides
`The formation of cocrystals may be quite important in
`explaining solid--solid interactions in many fields including
`those of pharmaceuticals.
`
`(4) Etter, M. C.; Reutzel, S. M. J. Am. Chem. Soc. 1991,113, 2586-98.
`(5) Etter, M. C.; MacDonald, J. C.; Bernstein, J. Acta Crystallogr.,
`Sect. B 1990, B46, 256-62.
`(6) Etter, M. C.; Reutzel, S. M. J. Am. Chem. Soc. 1991,113, 2586-98.
`
`This elegant work of Etter and co-workers has greatly
`increased our understanding of the hydrogen-bonding
`interactions of molecules in both the solid state and in
`solution.
`
`B. How Crystals Form
`Crystallization and the factors controlling the formation
`of crystals is an extremely important area in solid.state
`pharmaceutical chemistry. Dr. Margaret Etter made an
`extremely important observation when she pointed out
`that molecules in solution often tend to form different
`types of hydrogen-bonded aggregates and hypothesized
`that these aggregate precursors are related to the crystal
`structures that form from the supersatttrated solution.2
`This concept helps to explain the many different hydrogen-
`bonding motifs seen in different solids.
`A number of factors can affect the crystal formed either
`by influencing the hydrogen-bonded aggregate precursors
`in solution or influencing one of the many other factors
`involved in crystallization. These include (I) solvent
`composition or polarity, (2) concentration or degree of
`supersaturation, (3) temperature including cooling rate
`and the cooling profile, (4) additives, (5) seeds and the
`presence of seeds, (6) pH (pH is important for crystal-
`lization of salts), and (7) agitation.
`The composition of the solvent used is known to
`influence crystallizations either directly or by influencing
`the temperature at which the crystallization is initiated.
`For example, in the mannitol system, the a-polymorph is
`formed by evaporation of 100 % ethanol while the B.poly-
`morph is formed by crystallization from aqueous ethanol.7
`In a study of inosine, Suzuki showed that crystallization
`from water gave the a-form whereas crystallization from
`70 % DMSO gave the ~-form.s
`A second important factor influencing crystallization is
`the degree of supersaturation--the ratio of the concentra-
`tion of the solution to that of a saturated solution. In his
`study of the polymorphism of cimetidine, Sudo showed9,1°
`that in isopropyl alcohol at high supersaturation ratios
`(greater than 3.6) form A crystallized spontaneously or in
`the presence of seeds of either form A or form B, whereas
`in lower supersaturation ratios (less than 2) form A
`crystallized if there were A seeds and B crystallized if
`there were B seeds.
`Temperature can have a very significant effect on the
`polymorph produced. Studies by Kitamura11 on the
`crystallization of L-glutamic acid showed that at 45° the
`a-form nucleates slowly resulting in B-form growth,
`whereas at 25 °C the a-form nucleates rapidly causing
`a-form growth.
`The effect of additives on crystallization has been of
`interest for many years. Early work by Simonelli indicated
`that polymeric additives could prevent the crystallization
`of certain phases.12 Significantly, studies in recent years
`by Lahav and co-workers18 have shown that additives (as
`
`(7) Berman, H. M.; Jeffrey, G. A.; Rosenstein, R. D. Acta Crystallogr.
`1968, B24, 442-449.
`(8) Suzuki, Y. Bull. Chem. Soc. Jpn. 1974, 47, 2551-2552.
`(9) Sudo, S.; Sato, K.; Harano, Y. J. Chem. Eng. Jpn. 1991, 24, 237-
`242.
`(10) Sudo, S.; Sato, K.; Harano, Y. J. Chem. Eng. Jpn. 1991, 24, 628-
`632.
`(11) Kitamura, M. J. Cryst. Growth 1989, 96, 541-546.
`(12) Simonelli, A. P.; Mehta, S. C.; Higuchi, W. I. J. Pharm. Sci. 1970,
`59, 633.
`(13) Weissbuch, I.; Addadi, L.; Lahav, M.; Leiserowitz, L. Science
`1991, 253, 637-646.
`
`Lupin Ex. 1040 (Page 3 of 11)
`
`

`

`Reviews
`
`Chem. Mater., Vol. 6, No. 8, 1994 1151
`
`little as 0.03 %) can inhibit nucleation and crystal growth
`of a stable polymorph, thus favoring the growth of a
`metsstable polymorph. They also showed that it is possible
`to design crystal nucleation inhibitors to control poly-
`morphism.
`In addition, Grant et al. have shown the effects of
`additives on the properties of adipic acid, acetaminophen
`(paracetamol), and (R,S)- (-)-ephedrinium-2-naphthalene-
`sulfonate, a chiral drug. The effects on each of these
`crystalline solids will now be considered in turn.
`When adipic acid is crystallized from water containing
`traces of n-alkanoic acids14 or oleic acid,15 the additive is
`taken up by the crystals, while the very small water content
`is not significantly affected. Changes in crystal habit and
`in the crystal growth kinetics are consistent with a
`structural model in which the added n-alkanoic acid
`molecules occupy lattice sites at the crystal surfaces,is The
`incorporated additive also changes the thermodynamic
`properties of the crystal. Notably, the crystal energy and
`entropy are increased, as measured by reductions in the
`enthalpies of fusion and solution, while the melting point
`is little affected.14a5 In addition, the Gibbs free energy of
`the crystal increases, as measured by the dissolution rate
`and the specific surface area, and the density of the crystals
`is changed. Furthermore, the compaction properties, as
`measured by Hiestand’s indices of tableting performance
`of the crystals are also modified.17 Low levels of incor-
`porated n-octanoic acid produce an increase in lattice
`strain, reducing the energy required for plastic deformation
`leading to improved tableting performance. At higher
`levels of incorporated n-octanoic acid, the tableting
`performance again approaches that of the pure crystal,
`indicating a reduction in plasticity. The above results
`with adipic acid may be explained by the impurity defects
`and attendant dislocations introduced by incorporated
`additive, resulting in increased lattice strain at low
`concentrations of additive and in reversal of the effects at
`higher concentrations. These proposed changes in lattice
`strain have recently been confirmed by corresponding
`increases and decreases in the mosaic spread of the Laue
`diffraction pattern when single crystals are irradiated by
`white X-radiation from a synchrotron source,is
`When acetaminophen is crystallized from water con-
`taining the structurally related synthetic impurity, p-
`acetoxyacetanilide (PAA), the additive is incorporated and
`the crystals become acicular (needle-shaped). Increasing
`concentrations of PAA lead to increasing uptake of PAA
`to a maximum constant value (suggesting a saturated solid
`solution), to a decrease in water content, and to an increase
`in the length/width ratio.15 Higher concentrations of PAA
`cause the length/width ratio and the water content to
`return to nearly the initial values. The maximum length/
`width ratio and the minimum in water content correspond
`approximately to maxima in the enthalpy and entropy of
`fusion and in the intrinsic dissolution rate of the crystals,
`the melting point being little affected. The physical
`properties of the acetaminophen crystals mentioned above
`
`(14) Chow, K. Y.; Go,J.; Mehdiadeh, M.; Grant, D. J. W. Int. J. Pharm.
`1984, 20, 3-24.
`(15) Chow, A. H. L.; Chow, P. K. K.; Wang, Z.; Grant, D. J. W. Int.
`J. Pharm. 1985, 25, 41-55.
`(16) Davey, R. J.; Black, S. N.; Logan, D.; Maginn, S. J.; Fairbrother,
`J. E.; Grant, D. J. W. J. Chem. Soc., Faraday Trans. 1992, 88, 3461-3466.
`(17) Law, D.; Grant, D. J. W. Pharm. Res. 1993, 10, S-152, PT-6102.
`(18) Grant, D. J. W.; Law, D.; Ristic, R.; Shekunov, B.; Sherwood, J.
`N. 24~h Annual Meeting of the Fine Particle Society, August 24-28,
`Chicago, IL, 1993.
`
`were measured under 29 different crystallization condi-
`tions, defined by the initial concentration of PAA, the
`initial supersaturation of acetarninophen, and the rate of
`stirring of the crystallization solution.~9 Statistical analysis
`of the properties of these crystals, crystallized and analyzed
`in triplicate, showed strong correlations between the
`length/width ratio and the concentration of PAA taken
`up by the crystal and between the intrinsic dissolution
`rate and the length/width ratio, the thermodynamic
`quantities playing a minor role.2° These results demon-
`strate the significance of additive-induced changes in
`crystal habit in influencing the intrinsic dissolution rate
`of acetaminophen crystals. Thus, whereas the behavior
`of doped crystals of adipic acid may be attributed to
`differences in lattice strain, the behavior of doped acet-
`aminophen crystals may be attributed to differences in
`crystal habit.
`In the above examples, adipic acid and acetaminophen
`exist as achiral molecules and crystallize in achiral space
`groups. The question arises as to the effect of traces of
`the opposite enantiomer on the crystal properties of a
`chiral drug. To help answer this question, the chiral drug
`(R,S)-(-)-ephedrinium 2-naphthalenesulfonate [(-)-EN]
`was crystallized from aqueous solutions containing traces
`of the opposite enantiomer (S,R)-(+)-ephedrinium 2-naph-
`thalanesulfonate [(+)-EN].21 Crystals of (-)-EN took up
`the opposite enantiomer with an appreciable segregation
`coefficient (0.153), while the water content and melting
`point of the crystals remained constant. Uptake of the
`opposite enantiomer led to changes in the thermodynamic
`properties and intrinsic dissolution rate. These changes
`are similar to those observed when adipic acid incorporated
`n-alkanoic acid from the crystallization solution. The
`similar behavior suggests an analogous molecular mecha-
`nism in the solid state, implying that doping with the
`opposite enantiomer produces changes in crystalproperties
`that are analogous to doping of a foreign molecule of
`different, but related structure. Interestingly, (-)-EN and
`(+)-EN when heated together give a phase equilibrium
`diagram with a eutectic (together with limited solid
`solution formation) between each enantiomer and the
`racemic compound. This observation suggests that the
`observed changes in the thermodynamic properties and
`intrinsic dissolution rate of (-)-EN may be attributed to
`doping of the crystals with [(+)-EN/(-)-EN] molecular
`pairs or clusters rather than simply with the added (+)-
`EN molecules.
`Seeding is used extensively to control crystal form and
`also to control the extent of nucleation (i.e., final particle
`size). In almost all cases, seeding can be used to control
`the desired crystal form. For example, Suzuki et al.s
`showed that the a-form of inosine could be obtained by
`crystallization from water whereas to obtain the ~-form,
`seeds of the ~-form must be used. An interesting study
`of the effect of seeding was reported by Konapudi, and
`the mechanism of crystallization was elucidated by
`McBride and Carter.22 Sodium chlorate crystallizes in
`both a chiral and a racemic form. Since sodium chlorate
`is not chiral in solution, crystallization from aqueous
`solution produces equal numbers of L and D crystals.
`
`(19) Chow, A. H. L.; Grant, D. J. W. Int. J. Pharm. 1988, 42, 123-133.
`(20) Chow, A. H. L.; Grant, D. J. W. Int. J. Pharm. 1989, 51,129-135.
`(21) Duddu, S. P.; Fung, F. K. Y.; Grant, D. J. W. Int. J. Pharm. 1993,
`94, 171-179.
`(22) McBride, J. M.; Carter, R. L. Angew. Chem., Int. Ed. Engl. 1991,
`30, 293-295.
`
`Lupin Ex. 1040 (Page 4 of 11)
`
`

`

`1152 Chem. Mater., Vol, 6, No. 8, 1994
`
`Reviews
`
`Surprisingly, crystallization of an aqueous solution with
`stirring gives mostly crystals of one chirality, either L or
`D. investigation showed that this effect was due to the
`fact that once a particular crystal (either L or D) has fi)rmed,
`when that initial crystal collides with the stirrer and is
`broken into many small seeds which then nucleate further
`crystalIization of that hand crystal (either L Or D). This
`observation supports the idea that one nucleating seed
`can produce a single crystalline form. It also suggests
`that in this particular crystallization, if one added a seed
`of either L or D crystals then one would obtain entirely
`that crystal form. This is one of the best examples of the
`concept: one seed, one crystal.
`
`McCrone, in a letter to the editor of the Journal of
`Applied Crystallography, suggested that nucleation by
`seeding was the best explanation for the situation in which
`one unexpectedly obtains a new, more stable crystal form
`and then finds it is difficult or impossible to obtain the
`older less stable crystal form.2a In response to this letter
`Jacewicz et al.~ suggested that it is not impossible to
`produce the earlier form. They stated that the original
`crystal form should be capable of being produced but that
`the selection of the right conditions may require some
`time and trouble,
`
`C. Disorder in Crystals of Pharmaceutical Solids
`
`Another solid^state phenomenon which has important
`implications for pharmaceutical solids is disorder. Just
`as crystals of chiral and racemic drugs may have different
`physical properties, the presence of disorder in a crystal
`may also affect the physical properties of a crystal. One
`method of classifying disorder is to distinguish between
`cases involving static and dynamic disorder. Because
`single-crystal X-ray techniques give results which are
`averages over a long time period relative to molecular
`motion, X-ray crystallography at one temperature is
`usually unable to distinguish between static and dynamic
`disorder. Static and dynamic disorder can often be
`distinguished by comparing X-ray thermal parameters
`(ADPs) for two different temperatures. If the disorder
`results in observable resonance peaks in solid-state NMR
`spectra, then clearly this technique can be used to study
`this disorder.
`In the crystallographic literature there are numerous
`examples of disorder which are caused by a molecule
`crystallizing in a single conformation but in a different
`orientation relative to other molecules in the crystal, There
`are also examples where the disorder is due to molecules
`crystallizing in different conformations.
`
`Two such examples of conformational disorder are the
`crystal strt~ctt~res of nonsteroidal antiinflammatory
`(NSAID) drug salicylsalicylic acid (SALSALATE)ss,~ and
`the antiarrhythmic compound ftecalrtide (TAMBOCOR).~7
`
`(:28) McCrone, W. C, J, AppL Crystallogr. 1~75, 8, ~42.
`(24) J~cewicz, V. W.; Nayler, J, N. C. d, AppL Crystallogr. 197~, 12,
`396-397,
`
`Figure I. The two geometries of the SALSALATE molecule
`which 8re incorporated into the disordered crystal.
`
`Figure 2. The two geometries of the TAMBOCOR which are
`i~corporated into the disordered crystal.
`
`The salsalate molecule can crystallize in two different
`conformations, each of which is intramolecularly hydrogen-
`bonded. Presumably, these two different intramoIecular
`:modes of hydrogen bonding are roughly equivalent
`energetically and the gross shape of the molecule in both
`conformations is similar so that both conformations can
`be ir~corporated into a single crystal as shown in Figure
`1.
`
`The other interesting example is the antiarrhythmic
`compound flecainide which is disordered in the solid state.
`The form of the compound which exhibits this disorder
`is actually the acetate salt, which was originally used for
`therapeutic purposes. Here the piperidinium ring takes
`on two alternate chair conformations as shown in Figure
`2o
`
`What are the consequences of such disorder? At the
`level of processing, the difficulty in reproducing material
`with the same properties may be responsible for different
`
`(25) Prink, N.; Gieason, W. B.; Sweeting, L. M, In IVth Midwest
`Organic Solid-State Chemistry Symposium; Lincoln~ NE, 1992.
`(26) Frhak, N,; Beaur!ine, J.; Glew~on, W. B. In Am. Cryst. Assoc.
`Annual Meeting; Pittsburgh, PA, 1992,
`(2q) Gleas~n, W. B.; Bannit, E. H. In Am. Cryst. Assoc., Annual
`Meeting, ~987; p PE].8.
`
`Lupin Ex. 1040 (Page 5 of 11)
`
`

`

`Reviews
`
`Chem. Mater., Vol. 6, No. 8, 1994 1153
`
`A
`
`60
`
`~0
`
`S
`
`~ 8C
`
`v/era-~
`Figure 3. IR spectra of Nujol mulls of the two polymorphs of
`ranitidine hydrochloride: A, form 1; B, form 2.~
`
`2000
`
`I i I
`1800
`1600
`t&O0
`
`I
`1200
`
`I
`1000
`
`i
`800
`
`I
`600
`
`batches having dissimilar behavior. Batch-to-batch vari-
`ability may be caused by differences in the ratio of the
`different conformations of molecule in the bulk sample.
`Although disorder is a well-known phenomenon in the
`crystallography of small organics, the consequences of
`disorder in drug molecules appears not to have been widely
`considered. Just as crystals of chiral and racemic drugs
`may have different physical properties, the presence of
`disorder in a crystal may also affect the physical properties
`of the crystal. The amount of disorder (overall ratio of
`high occupancy to low occupancy conformations in the
`batch sample) found in a material may be highly dependent
`on the exact crystallization conditions. In turnthe physical
`properties of individual batches may range from free-
`flowing powders to flocculent precipitates suffering badly
`from the effects of static electricity.
`In addition, dynamic disorder reflects enhanced mo-
`lecular mobility in the solid state. This enhanced mo-
`lecular mobility may lead to enhanced chemical reactivity
`as suggested by the four-step mechanism of solid-state
`reactions first advanced by Paul and Curtin.2s In this
`mechanism the first step is molecular loosening. Obvi-
`ously, crystals possessing dynamic disorder are relatively
`loosely packed (at least near the site of disorder) and thus
`may be expected to be more reactive. At Purdue Uni-
`versity, we are presently investigating cases in which the
`disorder of the system may relate to enhanced reactivity.
`Are disordered samples with different amounts of
`disorder distinguishable? One method to distinguish this
`is discussed later in this review, solid-state NMR spec-
`troscopy. Another approach is powder diffraction.
`Whether a particular sample will give a powder diffraction
`line which will change as a function of the disorder is a
`function of the scattering power of atoms in the alternate
`conformations and their contribution to scattering for a
`particular scattering plane. Computer programs are
`available for the calculation of powder diffraction patterns
`from single-crystal X-ray diffraction data. In one case we
`
`(28) Paul, I. C.; Curtin, D. Y. Acc. Chem. Res. 1973, 7, 223.
`
`have examined, that of salicylsalicylic acid, our calculations
`indicate that certain diffraction lines are sensitive to the
`ratio of high occupancy to low occupancy conformations
`present in the salicylsalicylate sa