Polymorphism is the property of a compound to exist in more than one crystalline form. Different polymorphic forms of the same substance can have different properties such as solubility, density, melting point, stability or bioavailability. Monitoring and controlling the polymorphic form during batch or continuous crystallizations is very important to ensure the purity and quality of the final product. Significant research effort is being currently devoted to new technologies to detect nucleation of different polymorphs and to develop new control strategies to eliminate the undesired form [1,2]. The type of molecular bonding in solution is very important in the determination of the polymorphic form of the nucleated structure [3-5]. Many experimental and computational studies are focused on understanding how the solute-solvent interactions influence polymorphism [6-10]. The most used technique to detect changes in the liquid state just before nucleation is ATR-FTIR, although Kulkarni et al.  have successfully used Raman to understand the type of bonding between molecules of isonicotinamide in different solvents and its effect on the polymorphism of this compound. Furthermore, the presence of additives or impurities interacting with the solute can define the polymorphic outcome of a crystallization process. Additives can inhibit nucleation and/or growth of a specific polymorph favouring another form (kinetic effect). They can also have a thermodynamic effect and favour nucleation of a specific polymorph by lowering its free energy of nucleation. The effect of structurally related additives is of particular interest since they can affect both polymorphism and morphology of the crystals by incorporation in the crystal lattice during nucleation and growth [11-14].
Anthranilic acid (2-aminobenzoic acid or ortho-aminobenzoic acid, OABA) is mainly used for the synthesis of dyes and intermediates for the production of some pharmaceuticals. It can exist in three different polymorphic forms: I, II and III. Form I is stable at ambient temperature while form III is the most stable form at higher temperature. Form II is enantiotropically related with form I and it is always metastable . Form I is the only one that presents zwitterions, together with neutral molecules in the crystal structure while both form III and II are characterized by hydrogen bonded carboxylic acid dimers. The maximum amount of zwitterions for OABA is in water while the addition of an organic solvent to the solution usually generates a decrease in the percentage of zwitterions and favours the formation of dimers [16-18]. Because dimers are typically present in form II and III and zwitterions in form I, solutions at high content of an organic solvent should theoretically favour metastable form II or III while water-rich solutions should nucleate preferentially form I.
In this work, the effect of the solvent composition in on the polymorphic outcome of OABA cooling crystallizations was studied: a relation between the ATR-UV/Vis and Raman signal of clear solutions and the polymorphic outcome of cooling crystallizations was found. Additionally, the effect of the presence of a structurally related additive (benzoic acid, BA) was tested in different solvents.
Material and Methodology
In order to determine the solvent effect a first set of cooling crystallizations was performed using the same kinetic conditions on different organic solvent/water solutions (0 to 30% of organic solvent w/w) of OABA in order to find a correlation between polymorphic outcome and position of specific UV and Raman peak maxima. The used conditions for the crystallization studies were: starting temperature at 50 °C, cooling rate of -1 °C/min, initial saturation temperature of around 40-45°C. Isopropyl alcohol (IPA), ethanol, methanol and acetonitrile were used as organic solvents. ATR-UV/Vis and Raman spectra were recorded during all crystallization experiments.
The second set of experiments was conducted in the presence of benzoic acid in three different solvents. Mixtures of IPA and water (0%, 10% and 20% IPA w/w) were used. Saturation temperatures similar to the first set of experiments were used for each solvent. Solid OABA was added to the solvent and the temperature was raised to 50°C in order to dissolve all the particles. The solution was kept at high temperature for 20-30 minutes and then cooled down at a rate of -1°C/min to 10 or 20°C depending on the initial saturation temperature. After that a small amount of additive was added to the slurry and the solution was heated up again and then cooled down with the previous cooling rate. More consecutive additions of additive in the same solution were performed maintaining the same kinetic conditions (initial temperature and cooling rate) but increasing the ratio of additive and OABA in solution. Particle vision and measurement (PVM), focused beam reflectance measurement (FBRM) and UV spectroscopy were used to detect nucleation and monitor OABA concentration and crystals during the experiment while online Raman spectroscopy was used to determine the polymorphic form nucleated. On and off-line images of the crystals were also analysed in order to determine the effect of BA on their size, shape, agglomeration and breakage. Additional off-line techniques such as SEM, XRD and DSC were also used.
Results and Conclusions
The study on the solvent effect showed a link between polymorphic outcome of the cooling crystallization experiments and the position of UV and Raman peaks maxima in clear undersaturated solutions. Only form I and II could be nucleated with the chosen solvents. Limit positions for a typical UV/Vis and Raman peak of OABA, over which only form II was nucleated and below which only form I resulted, were determined. The limit value for the UV peak position was found to be 331 nm at cooling rates of 0.5 °C/min and saturation temperatures of 30 °C and 40 °C. For a cooling rate of -1 °C/min and saturation temperature of 30 °C, 331 nm is still the limit peak position to obtain pure form I, whereas for higher OABA concentration it was found that pure form II is always obtained above 331 nm and pure form I always nucleates below 328 nm. For Raman spectroscopy the limiting value of the studied peak was found to be 1037 cm-1 for all the experiments. These results can be used to predict nucleation of zwitterionic polymorphs for UV and Raman sensitive substances and to design suitable solvent systems that favour the consistent production of the desired polymorphic form. In the second part of the work the structurally related additive (benzoic acid) was found to promote nucleation of form III in solvents where normally only form I or II nucleate. This structurally related additive can be easily incorporated in the crystal structure of both form II and III but not in form I and it was found to increase the time of polymorphic conversion of both the metastable forms into the stable form I. At BA/OABA ratios smaller than 0.1 no change in the polymorphic form nucleated was observed: form I was always nucleated from water while form II or mixture of II and I were nucleated in water and 10% IPA. At IPA concentration of 20% w/w in solution and BA/OABA<0.1 pure form II was consistently nucleated. The morphology of form II was not strongly affected by the presence of small concentrations of benzoic acid while a higher aspect ratio of form I was observed. However, a higher tendency to agglomeration was noticed for both polymorphs in the presence of this additive. For BA/OABA ratios higher than 0.1 form III started nucleating in all the tested solvents. Pure form III was obtained for BA/OABA ratios higher than 0.3-0.4 (depending on the solvent), while mixtures of form III and I or form III and II were nucleated for ratios between 0.1 and 0.3.
 G. Févotte, Chemical Engineering Research and Design, 85, 7 (2007).
 Z.K. Nagy, R. Braatz, Annual Review of Chemical and Biomolecular Engineering, 3, (2012).
 J. Bernstein, Crystal Growth and Design, 11, (2011).
 R.J. Davey, S.L.M. Schroeder, J.H. ter Horst, Angewandte Reviews, 52,(2013).
 P.G. Vekilov, Crystal Growth and Design, 10,(2010).
 R.A. Chiarella, A.L. Gillon, R.C. Burton, R.J. Davey, G. Sadiq, A. Auffret, M. Cioffi, C. Hunter, Faraday Discussions, 136, (2007).
 R.J. Davey, N. Blagden, S. Righini, H. Alison, M.J. Quayle, S. Fuller, Crystal Growth and Design, 1, 1 (2001).
 R.J. Davey,G. Dent, R.K. Mughal, S. Parveen, Crystal Growth and Design, 6, 8(2006).
 S.A. Kulkarni, E.S. McGarrity, H. Meekes, J. H. ter Horst, Chemical Communications, 48, (2012).
 S. Parveen, R.J. Davey, G. Dent, R.G. Pritchard, Chemical Communications, 12, (2005).
Fevotte, G., Gherras, N., Moutte, J., Crystal Growth and Design, 13, (2013).
 Garnier, S., Petit, S., Coquerel, G., Journal of Crystal Growth, 234,(2002).
Gu, C.H., Chatterjee, K., Young, V.J. , Grant, D.J.W., Journal of Crystal Growth, 235,( 2002).
He, X., Stowell, J.G., Morris, K.R., Pfeiffer, R.R., Li, H., Stahly, G.P., Byrn, S.R., Crystal Growth and Design, 1, 4(2001).
 S. Jiang, J.H. ter Horst, P.J. Jansens, Crystal Growth and Design, 10, (2010a)
 L. Zapala, J. Kalembkiewcz, E. Sitarz-Palczak, Biophysical Chemistry, 140, (2009).
 O. Abou-Zied, B.Y. Al-Busaidi, J. Husband, The journal of physical chemistry, 10, (2013) .
See more of this Group/Topical: Separations Division