Dev. Chem. Eng. Mineral Process. 11(S/4), pp. 579-602,2003. Control of Polymorphism in Crystallization of Amino Acid Mitsuta ka Kitamu ra* Department of Chemical Engineering, Hiroshima University 1-4-1, Kagamiyama, Higashi-Hiroshima City, 739-852 7, Japan The crystallization mechanism and the controlling factors in the crystallization of amino acid polymorphs were investigated. It was demonstrated that the effect of supersaturation (Ostwald's step rule) was not observed for systems of either L-Glutamic acid (a, B) or L-Histidine (A, B). The intensive effect of temperature was shown in the crystallization of L-Glutamic acid (L-Glu) polymorphs, however, no temperature effect was observed in the case of L-Histidine (L-Hisopolymorphs. It was suggested that such a difference could be related to the difference of the molecular conformation between the polymorphs, and the contribution of the conformers in the solutions to the nucleation process of L-Glu polymorphs. of both L-Glu polymorphs (a and nuclei-above-nuclei mechanism. B) The growth appeared to be due to the With these results, temperature and supersaturation appeared to be the basic important controlling factors (the primary factors) of polymorphism. Solvents and additives (the secondary factors) also have an influence on polymorphism. The influence of ethanol compositions on the crystallization of L-His polymorphs (A and B) was observed, and demonstrated that the solvent * Email: email@example.com 5 79 M.Kitamura effect is very kinetic, i.e. the nucleation and growth rate of the A form are suppressed with an increase in the ethanol composition. The effects of L-Phenylalanine as an additive on the crystallization and transformation rates were shown in batch crystallization. The effect of L-phenylalanine on the growth rate and morphologV of L-Glu polymorphs was also examined using the single crystal method in a flow system. The relationship between the additive concentration and the growth rate of each polymorph was expressed by the proposed equation. The importance of the kinetics of the crystallization for control of the polymorphic crystallization was demonstrated. The supersaturation needed for control of the selective crystallization of the polymorph was also calculated. From these results the controlling factors for the crystallization of polymorphs have been classified in order to clarifL their complex interrelationship. Introduction The control of crystallization of polymorphs is eagerly demanded in industry because the functionality and properties (e.g. bioavailability, morphology and purity) of many kinds of materials depend on the molecular arrangement in crystals. However, the precipitation behavior of polymorphous crystals is usually complex and the crystallization mechanism is unclear. The crystallization process of polymorphs is composed of the competitive nucleation and crystal growth of the polymorphs, and the transformation from metastable forms to stable forms. To control the polymorphism, the dependence of the elemental steps in the crystallization upon the operational factors should be determined [ 1, 21. Temperature and supersaturation are fundamental conditions in normal crystallization, and these factors also have an influence on the polymorphic crystallization . The addition of polymorphous seed crystals sometimes affects the crystallization behavior [4-51, furthermore the crystallization of the polymorphs is frequently influenced by additives [S-71. For example, it is known that the other amino acids present as an impurity (additive) in solutions frequently have an influence on the crystallization of the amino acid as a tailor-made additive. 580 Control of Polymorphism in C?ystallization of Amino Acid The effect of amino acids as additives appears in the crystallization of amino acid polymorphs . The solvent is also a very important factor in industrial crystallization [8-101 and is effective on the polymorphism of amino acid 191. It was also reported that various interfaces have influences on the crystallization of polymorphs [ 11, 121. Recently the effect of additives and solvents on crystal growth have been explained by analyzing the conformation of the molecule in crystals, and the interactions of the additive and solvent molecule with the crystal surface. However, it has been found that the dynamic behavior of polymorphs in the crystallization is very dependent on the various operational conditions. For the selective crystallization of the polymorphs, the key-controlling factors in crystallization should be known and, therefore, the crystallization mechanism of the polymorphs may become clear. Hence we have investigated the controlling factors in the polymorphic crystallization of amino acids. Experimental Details (i) Crystallization of L-glutamic acid L-Glutamic acid was crystallized by rapidly cooling the heated solution of various concentrations to a set temperature . The concentration change of solutions was measured, and the slurry was sampled and filtrated during the crystallization process. The transformation behavior was examined by measuring the composition change of polymorphs by XRD. The morphology of the precipitated crystals was observed through a microscope. (ii) Growth rate measurement using single crystal in aflow system A single crystal of each /3 and a polymorphs prepared by batch crystallization were mounted in the growth cell and the growth rates were measured at 298 K (f0.05 K) in flowing solutions with and without L-Phe .The sizes of the crystals and their morphology were measured using a microscope and video-TV system, and growth rates calculated from the crystal size change. M.Kitamura (iii) Crpstallizatlon of L-histidine Crystallization was carried out in the aqueous solution and the mixed solution of water and ethanol in various ratios by the same method as that used in the L-Glu system [ 9 ] . Concerning the growth rate measurement, the pure polymorphous crystals were sieved and a cut between 350 and 420 pm was used. The crystals were suspended in the supersaturated solutions and growth rates measured from the mass change of the crystals. The polymorphous composition of the crystals was determined by XRD. Results and Discussion Effect of temperature and supersaturation on the crystallization of L-glutamic acid (L-Glu) polymorphs L-Glutamic acid (L-Glu) has two polymorphs of a and f3 (both are and only the lattice parameters orthorhombic with a space group of P212121 are different) (31. Differential crystallization was carried out to examine the controlling factors, i.e. the unsaturated solution(s) was rapidly cooled to the set temperature and then a small amount of crystal was precipitated under stirring conditions. The polymorphous compositions were analyzed by XRD. In the crystallization of L-glutamic acid (u and f3 forms) it appeared that the supersaturation ratio had very little effect on the crystallization ratio of the L-Glu polymorphs at each temperature between 298 K and 323 K . For example, at 298 K with the supersaturation ratio range for p crystal (Sg) between 1.7-4.8, and at 318 K with SB between 1.3-3.0, the influence of the supersaturation ratio on the polymorphous composition could not be observed. This may indicate that the “Ostwald’s step rule”  cannot be observed in this crystallization process. Alternatively, the temperature intensively influenced the crystallization behavior, i.e. at a lower temperature (e.g. 293 K) only the u form precipitates and with a temperature increase the proportion of p in the crystals increased as shown in Figure 1. We presume that such a dependence of the crystallization behavior on temperature is due to the change in the relative concentration of conformers [141 with temperature. I 582 Control of Polymorphism in Crystallization of Amino Acid 0 285 290 295 300 305 310 315 320 325 Temp.M Figure 1. Dependence of a composition in precipitated crystals (X a ) on crystallization temperature. Temp. Conformer a 11 1 . L Conformer p Cluster a Nucleation Nuclei 1 Crystal a Growth 11 Cluster p 1 Nucleation Nuclei 1 Growth Crystal p Figure 2. Conformers and crystallization process. 583 M.Kitamura The conformers corresponding to each polymorph may aggregate and form different clusters as shown in Figure 2. The relative concentration of these clusters may be related to the precipitation ratio of the polymorphs. In relation to the change of the structure of the cluster, Nyvlt [ 151 measured the metastable zone width and explained the crystallization behavior of various hydrates of citric acid, ferrous sulphate and sodium hydrogen phosphate with the structural change of the clusters in solutions versus temperature. In the crystallization, the transformation from a to j3 proceeds by the “solution-mediated mechanism”. The transformation rate is controlled by the growth rate of the p crystals and also depends on temperature, i.e. the rate at 298 K is much slower than that at 3 18 K. Even when the p crystals were added to the solution at 298 K, the transformation rate was hardly affected. This phenomenon indicates that the growth rate of p crystal is very dependent on temperature (the activation energy of the growth of p crystal is large), and at 298 K the growth rate is very slow. The crystallization behavior is also influenced by another significant factor. For example, from stagnant solutions the crystal tends to crystallize. This may indicate that the two-dimensional nucleation on the surface of the crystallizer is advantageous for the nucleation of the p crystal. Furthermore, at the interface between the air and the solution the p crystal tends to crystallize preferentially. This may be caused by the template of the adsorption layer of L-Glu molecule on the interface [ 161. II Growth mechanism and morphological change of L-Glu polymorphous crystals The growth kinetics and morphological change of a and p crystals of L-Glu were investigated using the single crystal method . The seed crystals of L-Glu polymorphs (a and p) shown in Figure 3 were prepared through batch crystallization. The seed crystals were fixed in the flow cell and the linear growth rate in each direction was measured at 298 K by a microscope. It was demonstrated that the growth rates of a crystal in the ,[OIO]and [OOl] 584 Control of Polymorphism in Crystallization of Amino Acid directions, and that of $ crystal in the [loo] direction appeared not to be influenced by the diffusion of the solute and were controlled by the surface reaction process. From the dependence of the growth rate on the relative supersaturation (a), the mechanism of growth in all directions of the a crystal appeared to be a “nuclei-above-nuclei” mechanism as given by Equation (1) below [ 181, rather than a BCF mechanism [ 191. where X and Y are constants which depend upon the system. This result was also supported by the observation of the wavy front of the bunched steps with the atomic force microscope . The morphological changes of a seed crystals were observed in the growth process as shown in Figure 4, i.e. the (011) face developed. The morphological change of the a crystal is due to the difference in the dependence of the growth rate (G)on the supersaturation (a) between the (111) and (110) faces as shown in Figure 5 . This indicates the important contribution of the kinetic process to the morphological change. Dependence of the growth rate of the (1 0 1) face of flcrystal on the relative supersaturation (a) is shown in Figure 6. From the kinetic data, the growth mechanism of the (101) face of the p crystals is also considered to be the “nuclei-above-nuclei” mechanism. The growth rate of the (1 11) face of the a crystal is several times that of the growth rate of the (101) face of at the same supersaturation. This observation also corresponds to the preferential crystallization of a at 298 K. Furthermore, the ratio of the growth rate of p to that of a tends to increase with an increase in the supersaturation as determined by: G(lOl)~/G(lll),= , 0.22 exp(-0.14/0) (2) 585 M.Kitamura (111) Flow A1 A3 1 D2 (a) Figure 3. Seed crystals of L-Glu polymorphs: (a) a ;(b) p Figure 4. Morphological change of a crystal in pure system. 586 Control of Polymorphism in Crysallization of Amino Acid 0 0.1 0.3 0.2 0.5 0.b 0.1 0.6 u c-I Figure 5. Comparison of the growth rate (G) of a crystal between ( I I I ) and (011) face. 0 ? O . 0.0 O L . 0.1 ' 0.2 . ' . 0.3 ~ . 0.4 ~ . 0.5 B . 0.1 S . 0.1 u [-I Figure 6. Dependence of the growth rate of (101) face of Bcrystal on the relative supersaturation (u). 58 7 M.Kitamura The parameters of the “nuclei-above-nuclei” model for the growth rate in each direction of both polymorphs were given by the least squares method. The edge free energy (y) was also calculated from the parameter Y given by: Y = (Z/ 3 ) ( y /kT)’ (3 1 The edge free energy of the (1 0 1) face of fl is relatively larger than that of a crystals (( 111) and (001) faces). This observation is related to the solubility difference, i.e. the edge free energy of the stable form fl with a low solubility is higher than the metastable form a with a high solubility. The higher growth rate of a may be due to the large kinetic effects and the low edge free energy of a. III Crystaliization of L-histidine (L-His) polymorphs compared to L-Glu L-Histidine (L-His) has two polymorphs of orthorhombic (P212121)and monoclinic (P21)which are denoted as the A and B forms, respectively . The solubility measurement (at 293-330K) indicated that the stable phase is the A form .It was observed that the polymorphs crystallize with the same proportion (fraction of A, XA= 0.4-0.6) regardless of the concentration range (C)from 0.34 to 0.48 mol/l, corresponding to the supersaturation ratio range in respect to the A form (SA) from 1.34 to 1.90 as shown in Figure 7. This result also indicates that Ostwald’s step rule cannot be observed in the crystallization of L-His. However, the dependence of XA on temperature was found to be very small as shown in Figure 8, i.e. the tendency is very different from that in the case of L-Glu. Furthermore, the seed effect on crystallization behavior was examined. The seed crystals of each polymorph (1 mg, 105-149pm) were added to the solution (50 ml) just after the solution attained the set temperature during the crystallization before the primary nucleation occurred. It appeared that neither the A nor B seeds had any effect on the crystallization behavior of the polymorphs (see Figure 7). For the crystallization of L-Glu, the seed effect was observed. These differences of the seed effect between the L-Glu and L-His polymorphs may be related to the 588 Control of Polymorphism in Crystallizationof Amino Acid 1.o Temp. 293K 0 : 0.8 0 : no seeds A B ) seeds n A 0.6 4 x 0.4 0.2 0 0.35 0.45 0.40 C Emo1/11 Figure 7. Dependence of composition of A form in L-His crystals (Xd on the concentration (C) and seed efsect. -1 (b) 0.8 0.6 - 0.4 - 0.2 - 280 GHls T -7 1 285 J 290 295 300 Temp. [Kl 305 310 315 Figure 8. Dependence of A composition in precipitated crystals (XJ on crystallization temperature. 589 M.Kitamura t [mid --*: 240 ---o-: 480 -. : 1000 n E 0.4 a \ 3 U 0.2 0 200 400 600 800 bml Figure 9. Change in crystal size distribution in the transformation process by solution-mediated mechanism. VOL I-] Figure 10. Dependence of A composition (XJ in crystals on ethanol composition (VUL). 590 Control of Polymorphism in Cytallization of Amino Acid different physical properties between these amino acid molecules. When the solubility difference between the polymorphs of each amino acid is compared, it is much smaller in the case of L-histidine (44%) than in the case of L-glutamic acid (2548%). This is the cause of the slow transformation rate of L-His polymorphs when compared to that of L-Glu polymorphs. Such behavior may be attributed to the differences of the free energy of the molecule and the crystal structures between each of the polymorphs. L-His has a bulky imidazole group and the number of carbons is smaller than L-Glu, while the conformational difference between A and B polymorphs is very small , in comparison with that for L-Glu polymorphs. Hence in the case of L-His, the activation energy for the exchange between the conformer should be very small, and the conformers corresponding to A and B forms in solutions may be present at about the same probability. The transformation by the “solution-mediated mechanism” also occurs between the L-His polymorphs. During crystallization a change in the crystal size distribution (CSD) was observed as shown in Figure 9 . For example, as shown in Figure 8, two peaks were observed in CSD and the peak for the small size crystals increased with time, whereas the peak of the large size crystals decreased. This change in CSD is due to the “solution-mediated transformation”, i.e. the metastable B form with a large size dissolves and the stable A form nucleates and grows. IV Effect of solvent composition on L-His crystallization When L-His was crystallized from the mixture of ethanol and water, the effect of the solvent composition on the crystallization of the polymorphs was observed . Figure 10 shows the relationship between the fraction of A form in a precipitate (X,) obtained by the differential crystallization and the volume fraction of ethanol (VOL) at each temperature. It can be seen that the value of XA is not influenced by the ethanol composition at a low VOL, however, it decreases rapidly at a VOL larger than about 0.2 and finally pure B crystals precipitate. With a temperature decrease, the increase of the amount of B form due to ethanol seems to be accelerated. Dependence of the solubility 591 IU. Kitamura (C') on the solution composition at each temperature is shown in Figure 1 1. It can be seen that the solubility difference decreases with the increase of the ethanol fraction (VOL). However, this tendency cannot explain the crystallization behavior seen in Figure 10. The effect of the solvent composition on the growth rate was examined at 293 K by measuring the mass change of the polymorphous seed crystals which were suspended in the supersaturated solutions. The growth rate (G) was calculated from the mass change of the seed crystals, assuming a spherical shape. It appeared that the growth rate of B form is larger than that of A at the same supersaturation level for every VOL. However, when the growth rate of A and B form is compared at the same concentration, they appear to take very similar values at zero of VOL as shown in Figure 12(a). With the increase of VOL up to 0.4, the difference of the growth rate increased as seen in Figure 12(b). This result indicates that ethanol suppresses the growth rates of A selectively at an ethanol composition larger than 0.2 . In comparison with the crystallization behavior in Figure 10, it is expected that not only the growth but also the nucleation of A is hindered by ethanol. It is presumed that the increase of VOL may make the concentration of the conformer corresponding to B increase, while the desolvation energy of B may also get smaller in comparison with A. This may result in the preferential nucleation and growth of B. The solvent effect seems to increase with a decreasing temperature as shown in Figure 10. This may be due to the decrease of the molecular motion with temperature. V Effect of additives on L-Glu crystallization Various amino acids influence the Crystallization of L-Glu polyrnorphs, e.g. L-valine and L-leucine isomers inhibit the crystallization rate of L-Glu and are also included in L-Glu crystals in the course of the crystallization process 1231. The effect of L-Phe on the nucleation and the growth of L-Glu is relatively large. Figure 13 shows the composition change of a crystal in the course of the crystallization at 318 K . It can be seen that with an increase 592 Control of Polymorphism in Cgstallization of Amino Acid Temp Cryrtols A 313K (+ : B - & : 0 0.1 0.2 0.3 293K (zj 283K ( gi 01; 0.5 0.6 $ 0.7 VOL 1-3 Figure 11. Dependence of solubility of L-His polymorphs on ethanol composition. Temp. 293M voL= 0 4 4 :A 12 C Cmolill C tmollll Figure 12. Relationship between growth rate of L-His polymorphs and concentrations at: (a) 0 of VOL; and (b) 0.4 of VOL. 593 M. Kitamura 1.0 I Y ’0.5 4: 0 0.5 +: 10 20 30 00 50 60 70 80 90 1 [hl Figure 13. Effect of L-Phe concentration on the a composition in precipitated crystals (y) at 318 K. . . L c d.9X10 dmovl ,o ‘0° P -5 0:Al 0:Al 9 ‘ 9‘ O:AZ @:A3 A:B1 I I ,A 0’ I Y / / @-*A .O ,o‘ , 0 I t [mln] t [min] Figure 14. Crystal size change (AL) with time (t) at L-Phe concentrations o j (a) 7 . 7 x 10-4mol/l; and (b) 1.3 x lo” mol/l. 594 Control of Polymorphism in Crystallization of Amino Acid of L-Phe concentration, the nucleation of the rate, i.e. the growth rate of p crystal and the transformation p crystal, is suppressed. This result also means that the growth of a crystals is hindered by L-Phe, because the crystallization rate is retarded by the increasing L-Phe concentration. To examine the difference in the effect of L-Phe on the growth kinetics and the morphology of L-Glu polymorphs, the single crystal method in a flow system was used and the effect of L-Phe on the growth kinetics was compared between the polymorphs . The seed crystals of L-Glu polymorphs (a and p) are shown in Figures 3(a) and (b). It was observed in the growth rate measurement of a crystals that the growth rate in the [OOl]direction (D) is not influenced by L-Phe. However growth in the direction parallel to the (001) main face (Al,A2, A3 and Bl) is retarded by L-Phe. In pure solutions and at low concentrations of L-Phe (less than 7.7 x mol/l), the growth of the seed crystals in each direction (Al,A2, A3 and B1) was very linear in time for both polymorphs (see Figure 14a). However, with an increase of the L-Phe concentration, the growth became irregular and the growth rate tended to fluctuate with time (see Figure 14b). Furthermore, with the addition of L-Phe in the solution, the intensive morphological change appeared in the case of the a crystal (see Figure 15), i.e. the growth rate in the [ 1 1 11 direction was suppressed and the new (110) face appeared, while the small (011) face disappeared. This means that L-Phe adsorbs selectively on the (1 10) face and inhibits the growth of that face in preference to the (OOl), (111) and (011) faces under this condition. From the crystallographic data it can be seen that the carboxylic acid group is rich in the (001) face (see Figure 16). Thus, it may be assumed that L-Phe cannot adsorb on the (001) face because of repulsion from the carboxylic acid group. However, discrimination of the interaction with L-Phe among the (110), (111) and (011) faces is difficult. Davey [ 5 ] investigated the effect of different 1,5 dicarboxylic acid compounds on the morphological change of L-Glu with a conformational analysis. He observed that the (1 1 1) face of L-Glu is predominant with glutaric acid while the (1 10) face appears with transglutaconic and trimesic acids. In our previous work with batch crystallization, it was observed that even when using the same additive, L-Phe, the (1 11) face becomes predominant. This indicates the 595 M Kitamura Figure 15. Morphological change Figure 16. Molecular arrangement in of a crystal in the presence the surface of (OOI), (I I I ) and (I 10) of L-Phe. faces of a crystal. 1.4 -= h ! I 1.2 - 0.8 - 0.4 - 0.2 0.2 C = 0.003 ml/l (u-O.SO0) - Calculatrd 0.4 0.6 ~ ~ ~ 1 tn~i/ii 0 3 0.8 cp' Figure 17. Growth rate dependence of L-Glu polymorphs (G) on concentrations (Cp) of L-Phe; a. a-crystal (I 10) face; b. /I-crystal (I 01) face. 596 Control of Polymorphism in Crystallization of Amino Acid importance of the kinetics and the crystallization conditions on the morphological change. Figure 17(a) shows the relationship between the growth rate of the (1 10) face and L-Phe concentration. With L-Phe concentration (Cp), the increase of the growth rate of the (1 10) face of the a crystal declined, and finally stopped at the critical concentration, Cp* (1.7 x lo-’ mol/l). However, the growth of p crystal, which is the fast growing face of the p crystal, was also hindered and the growth rate of the (101) face of p decreased with the (101) face of the the L-Phe concentration as shown in Figure 17(b). It appears that the critical concentration of L-Phe, Cp* for the (1 10) face of a is about twice that for the (101) face of p (0.7 x lo” mol/l) at almost the same L-Glu concentrations (C). These results demonstrate the preferential inhibition of the growth of p. The dependence of the growth rates on the concentration of the additive (Cp) can be expressed by Equation (4) below for both polymorphs, using the growth model of a curved step and the adsorption density of the additive . The observation of a ‘cluster of grapes’ on the crystal surface with an atomic force microscopy supports the model assumption that the adsorption of L-Phe defines the advancing steps [ 181. G = G,[l- 2pCd”’] Pc (4) == P where Go is growth rate in the pure solution; p c is the critical radius; d is the adsorption density; y is the edge free energy; a is the lattice constant; k is the Boltzmann constant; and T the temperature (K). To analyze the adsorption behavior of the additives, the Langmuir-type model was applied. However, it was clear that in both cases of a and p polymorphs the Langmuir-type isotherm could not be applied to the experimental results, and the Freundlich-type isotherm equation was applicable. Therefore, the adsorption density (d) is expressed by: d = KCpn where K and n are the constants in the Freundlich-type equation. 597 M.Kitamura The calculated values of the growth rates by the least squares method using Equations (4) and (6) are shown in Figures 17(a) and (b) by the solid lines. These results indicate that the effect of the additives is not only due to the molecular interaction between the additive molecule and the crystal surface, but the kinetic process also contributes significantly to the effect. The adsorption density can be correlated with the distance between the neighboring adsorbed additive molecules ( e ) by: d=F 1 (7) The dependence of the growth rate of the (1 10) face of a crystals on the relative supersaturation ( 0 ) and adsorption density is given by: Values of (Go-G)/Go are plotted against l/a at various additive concentrations for a form in Figure 18, showing an approximately linear relationship. For the p form, a similar relationship was obtained. The slope of a line corresponds to the value of B&, and the linear relationship implies that the density (d) is almost a constant and is independent of supersaturation. For the practical use of additives in the selective crystallization of polymorphs, then the supersaturation and the concentration of additives should be controlled simultaneously. If the growth of only p form is completely inhibited, then the value of 2p, should be between the t values of each polymorph (!. and .!B ) and the following equations apply: B B -<u<la e, Hence the supersaturation (a) in the crystallization can be determined for each Cp value from these equations. 598 Control of Polymorphism in Crystallization of Amino Acid I 2 4 6 8 l/u 10 12 14 [-I Figure 18. Plot of (Go-G)/Goagainst 1 / tfor ~ (110) face of a crystal. I Solubility of polymorphs 1 I 0 Control of crystallivltion of polymorphs Primarv factor Supersaturation *Temperature *Coolingrate Stirring rate Mixing rate of reactant solutions 'Seed crystals Nucleation 0 Growth 0 Transformation I Secondarv factor -*Additives Solvent *Interface PH I Figure 19. Schematic diagram of controlling factors in crystallization of polymorphs. 599 M.Kitamura VI Controllingfactors in crystallization of polymorphs Using the preceding results we classified the controlling factors for the crystallization of polymorphs as shown in Figure 19. The most influential factors were classified as the primary factors, these were both supersaturation and temperature. In addition, the cooling rate may be part of the primary factors. However, in the cooling crystallization, changes in the temperature and supersaturation occur simultaneously. Therefore, the effect of the supersaturation and temperature should be examined separately. In the reactive crystallization, the mixing rates of the solutions appeared to be very important . Factors due to the external effects such as solvents, additives, interface and pH are grouped in the secondary factors. The effects of the additive is quite complex and the kinetic effect should be considered as well as the molecular structures . The solubility of each polymorph is not the main operational factor, however, the supersaturation of each polymorph is based upon the solubility. Furthermore, the relative thermodynamic stability and the direction of the transformation between the polymorphs can be expected from the solubility of polymorphs. Conclusions We have shown that the effects of supersaturation (Ostwald's step rule) were rarely observed in the systems (L-giutamic acid (a, p) and L-histidine (A, B)). The temperature effect was significant for L-glutamic acid polymorphs, however, no temperature effect was observed in the crystallization of L-histidine polymorphs. These different temperature effects may be related to differences in the molecular conformation between the polymorphs. The growth mechanism of L-glutamic acid polymorphs appeared to be almost identical, however, the dependence of the growth rate on the degree of supersaturation was different. The influential factors, such as temperature and supersaturation, can be regarded as the primary factors in polymorphic crystallization. External factors such as solvents and additives also have an influence on polymorphic crystallization. The solvent effect was studied by 600 Control of Polymorphism in Crystallization of Amino Acid the influence of ethanol composition on the crystallization of L-His polymorphs (A and B). It was shown that the effect is due to the suppression of the nucleation and the growth rate of A form with an increase in the ethanol composition. The effect of amino acids as an additive on the crystallization of L-glutamic acid was examined. In particular, the effects of L-Phenylalanine on the growth rate and morphology of L-glutamic acid polymorphs were examined using the single crystal method. Equations are presented which describe the relationship between the additive concentration and the growth rate of each polymorph. The importance of the kinetics of the crystallization for the control of the polymorphic crystallization was demonstrated. The calculation of the supersaturation needed for the control of the selective crystallization of the polymorph was also determined. From these results the controlling factors for the crystallization of polymorphs have been classified in order to clarify their complex interrelationship. References 1. 2. 3. 4. 5. 6. 7. 8. 9. Kitamura, M. 1989. Crystallization of amino acids. J. Japan. Assoc. Crystal Growth, 16, 61-66. Kitamura, M. 1995. Crysral Growth Hundbook. Japanese Association for Crystal Growth, Kyoritsu-Shuppan, 547-55 1. Kitamura, M. 1989. Polymorphism in the crystallization of L-glutamic Acid. J. Crystal Growth, 96, 541-546. Beckmann, W. 2000. Seeding the designed polymorph: Background, possibilities, limitations, and case studies. Orgunic Process Research and Development, 4, 372-383. Davey, R.J., Blagden, N., Potts, G.D., and Docherty, R. 1997. Polymorphism in molecular crystals: Stabilization of a metastable form by conformational mimicry. J. Am. Chem. Soc., 119, 1767-1772. Yu, L., M.Reutzel-Edens, S., and Mitchell, C.A. 2000. Crystallization and polymorphism of conformationally flexible molecules: problems, patterns, and strategies. Organic Process Research and Development, 4,396402. Kitamura, M., and Ishizu, T. 1998. Kinetic effect of L-phenylalanine on growth process of L-glutamic acid polymorph. J. Crystal Growth, 192, 225-235. Blagden, N., Davey, R.J., Lieberman, H.F., Williams, L., Payne, R., Roberts, R., Rowe, R., and Docherty, R. 1998. Crystal chemistry and solvent effects in polymorphic systems sulphathiazole. J. Chem. Soc. Furuday Trans., 94(8), 1035-1044. Kitamura, M., Frukawa, H., and Asaeda, M. 1994. Solvent effect of ethanol on crystallization and growth process of L-histidine polymorphs. J. Crystal Growth, 141, 601 M.Kitamura 193-199. 10. Threlfall, T. 2000. Crystallization of polymorphs: Thermodynamic insight into the role of solvent. Organic Process Research and Development, 4,384-390. 11. Addadi, L., Berkovitch-Yellin, Z., Weissbuch, I., Mil, J.V.. Shimon, L.J., Lahav, M., and Leiserowitz, L. 1985. Growth and dissolution of organic crystals with "tailor-made" inhibitors, implications in stereochemistry and materials science. Angew. Chem. Int. Ed. Engl., 24,466-485. 12. Chen, B.D., Cilliers, J., Davey, R.J.,Garside, J., and Woodburn, E.T. 1998. Templated nucleation in a dynamic environment: Crystallization in foam lamellae. J. Am. Chem. Soc., 120, 1625-1626. 13. Ostwald, W. 1897. Studies on formation and transformation of solid materials. Z. Phys. Chern.. 22,289-330. 14. Ham, N.S.1974. NMR studies of solution conformations of physiologically active amino acids, in Molecular and Quantum Pharmacology, D. Reidel Publishing Company, New York, pp.261-268. 15. Nyvlt. J. 1995. The Ostwald rule of stages. Crystal Res. Technol., 30,443-449. 16. Weissbuch, I., Frolow, F., Addadi, L., Lahav, M., and Leiserowitz, L. 1990. Oriented crystallization as a tool for detecting ordered aggregates of water-soluble hydrophobic a-amino acids at the air-solution interface. J. Am. Chem. Soc., 112, 7718-7724 17. Kitamura, M., and Ishizu, T., 2000. Growth kinetics and morphological change of polymorphs of L-glutamic acid. J. Crystal Growth, 209, 138-145. 18. Gilmer, G.H., and Bennema, P. 1972. Computer simulation of crystal surface structure and growth kinetics. J. Crystal Growth, 13/14, 148-153. 19. Burton, W.K., Cabrera, N.. and Frank, F.C. 1951. The growth of crystals and the equilibrium structure of their surfaces. Phil. Trans. Royal Soc. London, 243, 299-358. 20. Kitamura, M.. and Onuma, K. 2000. In situ observation of growth process of a-L-glutamic acid with atomic force microscope. J. Colloid Interface Sci., 224, 31 1-3 16. 21. Maddin, J.J.. McGandy, E.L., Seeman, N.C., Harding, M.M., and Hoy, A. 1972. The crystal structure of the orthorhombic form of L-(+)-histidine. Acra Cryst., B28, 2377-2382. 22. Kitamura, M. 1993. Crystallization behavior and transformation kinetics of L-histidine polymorphs. J. Chem. Eng. Japan, 26,303-307. 23. Kitamura, M., and Nakamura, T. 2001. Inclusion of amino acids and the effect on growth kinetics of L-glutamic acid. Powder Technol., 121, 39-45. 24. Kitamura, M., and Funahara, H. 1994. Effect of L-and D-phenylalanine on crystallization of L-glutamic acid polymorphs. J. Chem. Eng. Japan, 27, 124-126. 25. Kitamura, M., Konno, H., Yasui. A., and Masuoka, H. 2002. Controlling factors and mechanism of reactive crystallization of calcium carbonate polymorphs from calcium hydroxide suspensions. J. Crystal Growth, 236,323-332. Received: 25 July 2002; Accepted after revision: 20 July 2003.