The effect of additive aggregation state on controlling crystallization. Crystallization of L-asparagine monohydrate in the presence of carboxylic acid functionality additives

---

Abstract

L-Asparagine monohydrate crystals grown from 100–300% supersaturated solutions in the presence of 1–20 mg ml⁻¹ polyacrylic acid exhibit a needle-like morphology, instead of their normal prismatic morphology. A similar effect is observed for crystals grown in octanoic acid emulsions. In contrast, no growth modification is observed for L-asparagine monohydrate crystals grown in 20 mg ml⁻¹ sodium polyacrylate, sodium octanoate and propanoic acid. Polyamido amine (PAMAM) dendrimers (generations −0.5, 1.5, 3.5 and 4.5) end-capped with carboxylate groups also exhibit no growth inhibitory effect at 20 mg ml⁻¹, even after acidification to pH 2.8. The ability of the carboxylic acid functionality to potentially act as a growth inhibitor for L-asparagine when in solution, or emulsion droplets, is contrasted with the L-asparagine crystallization promotion produced by carboxylic acid films spread at the air–water interface.

---

Introduction

The control of crystallization is crucial to many diverse biological and industrial processes. For example, many ocean fish rely on macromolecules to prevent ice crystallizing in their blood,¹ the production of a particular polymorph is essential in drug formulations,² and impeding crystallization in oil and gas pipes prevents pipe blockage.³ Control may be affected by promoting crystallization of a particular polymorph or morphology or conversely by inhibiting crystallization. The former is achieved through promoting nucleation, whereas the latter occurs through inhibiting crystal growth. In both cases, an additive may achieve the desired effect provided that a sufficiently strong interaction exists between part of the additive and the crystalline species. Consequently, a particular functional group can act as either a nucleation promoter or crystal growth inhibitor, depending upon its aggregation state. If the additive aggregates to form a planar surface, then nucleation is promoted provided that the crystallizing material adsorbs onto this surface. However, if the additive is molecularly dispersed, then it may adsorb onto a growing crystal and inhibit further growth by impeding the attachment of additional crystalline material. The question then arises, at what aggregation state will a species switch from a crystal growth inhibitor to a nucleation promoter in a particular system, i.e. what range of emulsion/microemulsion droplet sizes will promote crystallization? What is the role of surface curvature and how do the dynamics of the system (i.e. the rate at which the micelles/emulsion droplets move and the molecular residence time in the micelles/emulsion droplets) affect the process? Previous studies have shown that size and polymorphic form can be controlled through crystallization in micellar, microemulsion and emulsion systems.^4–9 However, a systematic study of the role of surface curvature and aggregation state upon promoting nucleation and inhibiting crystal growth as a function of supersaturation has not been conducted to our knowledge. This manuscript details preliminary results obtained in investigating these issues.

We have investigated the effect of carboxylic acids and their sodium salts on the crystallization of L-asparagine monohydrate. This system was chosen because the carboxylic acids and their sodium salts are readily available from standard chemical suppliers in a range of different molecular weights, and with increasing length of hydrocarbon tails, and range from water soluble, through to micellar species, to insoluble species. In addition, different architectures are also available for the polymeric species, viz dendritic species and copolymers. Asparagine was chosen because this material forms a conglomerate, and so the effect of ionic strength can be decoupled from supersaturation by comparing the behaviour of supersaturated L-asparagine and an equivalent concentration of undersaturated DL-asparagine. In addition, previous studies^10,11 have shown that the habit can be affected by the addition of tailor-made additives at concentrations of ca. 2–20 mg ml⁻¹. L-Asparagine crystallizes as the monohydrate with an orthorhombic unit cell of P2₁2₁2₁ symmetry and unit cell parameters of a = 5.593 Å, b = 9.827 Å, c = 11.808 Å, Z = 4.¹² The crystal is held together by a network of hydrogen bonds. The typical aqueous morphology is prismatic, being slightly elongated along the a-direction and bounded by {012} and {011} faces together with smaller {101}, {111}, {020} and {002} faces.

Experimental

All the materials were purchased from Aldrich in their purest available forms, and were used without further purification. The polyacrylic acid and sodium polyacrylate had molecular weights of ca. 2000 and 2100, respectively. The supersaturated asparagine solutions were prepared by adding the required amount of material to ultra-pure water and heating to ca. 20 °C above the saturation temperature for 1 h so that the material had fully dissolved. The solution was then hot-filtered through 0.22 μm filters to remove any dust particles, and the filtrate was heated for another 2 h at ca. 20 °C above the saturation temperature before allowing the solution to cool slowly to room temperature. Oil in aqueous emulsions were prepared by placing oil (hexanoic acid, octanoic acid and toluene) drops directly onto the asparagine solutions containing 20 mg ml⁻¹ added sodium octanoate, and then shaking. The hexanoic acid and octanoic acid emulsions were typically stable for months, whereas the toluene emulsions separated into oil and aqueous phases after a few weeks. Emulsions of aqueous droplets in oil, stable for a few days, were prepared by shaking the aqueous phase with a 15% polyacrylic acid polyethylene copolymer dissolved in 1∶3 phenol∶toluene.

Results and discussion

Polyacrylic acid was found to inhibit the growth of L-asparagine monohydrate specifically on the {012} and {011} faces so that needle-like crystals were produced, instead of the normal prismatic morphology observed in aqueous solutions (see Fig. 1). At relative supersaturation levels of 100–300%¹³ at 20 °C, the effect was achieved using concentrations of 1–20 mg ml⁻¹, with lower supersaturations requiring less additive to obtain the needle morphology. In particular, the needle morphology was evident at a polyacrylic acid concentration of ca. 1, 2 and 10 mg ml⁻¹ at relative supersaturations of 100, 200 and 300%, respectively. Crystals grown at lower supersaturations (50–80%) were often bounded by only the {012} faces parallel to the a-direction, with the {011} faces developing under the faster growth rates arising from higher supersaturations. With increasing polyacrylic acid concentration, secondary nucleation caused aggregated crystals to develop, with the individual needles radiating from a central core to give a characteristic floret appearance (see Figs. 1e and 2b). This morphology may have been induced by the increased polyacrylic acid concentrations decreasing the growth rate to such an extent that secondary nucleation is favoured and/or by the increasing quantity of polyacrylic acid adsorbed onto the crystal surfaces acting as a nucleation centre.

Fig. 1 a) Normal prismatic morphology of L-asparagine monohydrate crystals grown from aqueous solution. b–e) Growth from 200% supersaturated L-asparagine solution in the presence of 1, 2, 5 and 10 mg ml⁻¹ polyacrylic acid, respectively. Note that the needle-like crystals become increasingly aggregated at polyacrylic acid concentrations of 5 and 10 mg ml⁻¹.

Higher magnification optical micrographs of crystals grown in the presence of polyacrylic acid revealed the presence of striations running approximately parallel to the a-direction (see Fig. 2). The observed change in crystal habit in the presence of polyacrylic acid is not attributable to the reduced pH (pH ≈ 4.1, 3.9 and 3.7 at 2, 5 and 10 mg ml⁻¹ of polyacrylic acid, respectively) of these solutions: acidification of 200% supersaturated L-asparagine monohydrate solutions to pH 3 by concentrated hydrochloric acid did not induce any noticeable morphology change. HPLC analysis could not detect the presence of the polymer in the crystals, indicating that the inhibitory effect is achieved by adsorption of polyacrylic acid onto kink and/or surface terrace sites¹⁴ on the {012} and {011} faces, impeding further crystal growth until the additive is desorbed.

Fig. 2 Higher magnification micrographs of L-asparagine monohydrate crystals grown from 200% supersaturated solutions in the presence of a) 2 mg ml⁻¹ polyacrylic acid, and b) 10 mg ml⁻¹ polyacrylic acid. Striations running approximately parallel to the needle length can be seen in a) and b).

In contrast, sodium polyacrylate and sodium octanoate had no effect on the crystal morphology in this supersaturation range at concentrations of 20 mg ml⁻¹. The pH of these solutions is ca. 7–8, so the additives remain predominantly in their anionic form. Even at 100 mg ml⁻¹ salt concentrations, only a slight elongation in the a-direction was observed for crystals grown in 100% supersaturated solutions, whereas at 200% supersaturation no effect was discernible. Furthermore, propionic acid had no effect on the asparagine monohydrate growth morphology at a concentration of 100 mg ml⁻¹ in the 200% supersaturated system, and only a slight elongation in the a-direction occurred in the 100% supersaturated system; at 20 mg ml⁻¹ no growth modification was observed.

Complexation of multivalent cations by polyacrylic acid is known to occur, however, no such interaction exists between polyacrylic acid and monovalent cations such as sodium.¹⁵ Consequently we would expect similar behaviour with polyacrylic acid salts of other monovalent cations. However, results with divalent cations would be mediated by the reduction in the effective carboxylate charge that would arise from complexation, and at sufficiently high concentrations, by the precipitation of the polyacrylate salt.¹⁵ We also expect little or no complexation of asparagine by polyacrylic acid, particularly since we have no evidence for a reduction in the extent of crystallization that would be indicative of a reduced supersaturation due to complexation.

Polyamido amine (PAMAM) dendrimers, generations −0.5, 1.5, 3.5 and 4.5, which are end-capped with 4, 16, 64 and 128 carboxylate groups, respectively, also had no discernible effect on the typical asparagine monohydrate aqueous morphology at a concentration of 20 mg ml⁻¹, even after acidification of the solution to pH 2.8 with concentrated HCl. Previous studies on calcium carbonate crystallization have shown that growth inhibition increases with the number of absorbing groups per molecule.¹⁶ This study also shows that the molecular architecture is critical. The generation 4.5 dendrimer would be expected to be predominantly in its free acid form at pH 2.8, and thus would contain >100 adsorbing carboxylic acid groups. However, no growth inhibition effect was observed, presumably because the (expected) spherical architecture in solution prevents the simultaneous adsorption of many of these groups upon the asparagine monohydrate crystal surface. In contrast, the linear architecture of the polyacrylic acid enables multiple adsorption of the carboxylic acid groups upon the crystal surface, and hence it is an effective growth inhibitor despite its lower average number of adsorbing groups (ca. 30).

We found that hexanoic acid and octanoic acid are insufficiently soluble (with solubilities of 9.68 and 0.68 mg ml⁻¹, respectively, in water) to produce any growth modification in L-asparagine monohydrate crystallization. However, aqueous emulsions of octanoic acid, stabilized through the addition of 20 mg ml⁻¹ sodium octanoate, also produced a needle-like morphology in the 100% supersaturated L-asparagine system; in the 200% supersaturated system, the effect was less marked. This demonstrates that a particular species does not need to be soluble in the same phase as the growing crystallite to be an effective growth inhibitor; emulsions can deliver the required inhibitory effect. Optical micrographs shown in Fig. 3 reveal that the emulsion droplets (of μm size) adhere to the growing crystal interface, facilitating the transfer of the acid to the crystal interface, so that a larger concentration of acid blocks the {011} and {012} crystal faces. In comparison, emulsions of hexanoic acid produced a slight elongation of the normal prismatic aqueous morphology in 100% supersaturated systems, with little discernible growth modification in the 200% supersaturated systems.

Fig. 3 Micrographs of L-asparagine monohydrate crystals grown from octanoic acid emulsions. a) Immature crystal grown in situ from an evaporating emulsion. b) Mature crystal, partially covered with solution, extracted from an octanoic acid emulsion initially containing 200% supersaturated L-asparagine. Note the emulsion droplets adhering to the crystal surfaces in a) and b), some of which are arrowed in b).

Conversely, spread films of stearic acid and a 15% copolymer of polyacrylic acid with polyethylene at the air–aqueous interface result in nucleation promotion of L-asparagine monohydrate in the pH range 3–8. Analysis of the faces developing beneath the films by in situ optical microscopy and external reflection FTIR show that the {012}, {020}, {101} and {011} faces may all develop at surface pressures around 10 mN m⁻¹, with the range of faces developing being dependent upon the solution pH.¹⁷ These combined studies demonstrate that a particular functional group can act as a nucleation promoter when present at a planar interface, but may inhibit crystal growth if molecularly dispersed, or aggregated into emulsion droplets. The growth inhibitory effect of the octanoic acid emulsions indicates that emulsion systems may potentially act as both nucleation promoters and crystal growth inhibitors, depending upon the supersaturation of the system. Computational simulations are in progress to investigate the differing nucleation promotion and crystal growth inhibition effects of the carboxylic acid and carboxylate functionalities, and the effects of additive architecture and conformational flexibility. Further experimental studies are also required to determine the extent to which the change from nucleation promotion to growth inhibition is orchestrated by the alteration in surface curvature of these systems, and by the more rigid confinement of the spread films at the planar interface. We have noted, however, that in supersaturated L-asparagine systems consisting of emulsions and excess dispersed phase, the crystallization occurs preferentially at the planar interface between the two phases, irrespective of whether the oil or aqueous phase comprises the dispersed phase.

Acknowledgements

S. J. C. is grateful to ICI for the kind provision of her lectureship.

Notes and references

1. A. D. J. Haymet, L. G. Ward and M. M. Harding, J. Am. Chem. Soc., 1999, 121, 941.
2. N. Blagden, R. J. Davey, R. Rowe and R. Roberts, Int. J. Pharm., 1998, 172, 169.
3. A. J. Hennessy, A. Neville and K. J. Roberts, J. Cryst. Growth, 1999, 198/199, 830.
4. T. Hirai, S. Hariguchi, I. Komasawa and R. J. Davey, Langmuir, 1997, 13, 6650.
5. R. J. Davey, A. M. Hilton, J. Garside, M. de la Fuente, M. Edmondson and P. Rainsford, J. Chem. Soc., Faraday Trans., 1996, 92, 1927.
6. J. D. Hopwood and S. Mann, Chem. Mater., 1997, 9, 1819.
7. G. K. Lim, J. Wang, S. C. Ng and L. M. Gan, Langmuir, 1999, 15, 7472.
8. H. Furedi-Milhofer, N. Garti and A. Kamyshny, J. Cryst. Growth, 1999, 198/199, 1365.
9. J. Yano, H. Furedi-Milhofer, E. Wachtel and N. Garti, Langmuir, 2000, 16, 10 005.
10. L. Addadi, Z. Berkovitch-Yellin, N. Domb, E. Gati, M. Lahav and L. Leiserowitz, Nature, 1982, 296, 21.
11. S. N. Black, R. J. Davey and M. Halcrow, J. Cryst. Growth, 1986, 79, 765.
12. J. J. Verbist, M. S. Lehmann, T. F. Koetzle and W. C. Hamilton, Acta Crystallogr., Sect. B, 1972, 28, 3006.
13. The relative supersaturation is given by [(c − c₀)/c₀] × 100%, where c and c₀ are the actual and saturation concentrations, respectively.
14. K. Sangwal, J. Cryst. Growth, 1999, 203, 197.
15. J. P. Boisvert, A. Malgat, I. Pochard and C. Daneault, Polymer, 2002, 43, 141.
16. N. Wada, K. Kanamura and T. Umegaki, J. Colloid Interface Sci., 2001, 233, 65.
17. S. J. Cooper, submitted to Langmuir.

---