Supersaturated aqueous solutions of glycine exposed to intense pulses of plane-polarized laser light at 1.06 μm unexpectedly crystallized into the polar γ-polymorph of glycine. Control solutions not exposed to the laser always produced crystals of α-glycine, the expected and most stable form. This result suggests a new approach to polymorph control and possibly a means to produce new polymorphs.

In this study, the application of Raman spectroscopy to quantitative in-line monitoring of the crystallization from solution of an enantiotropic polymorphic system, flufenamic acid, is explored. Separate feasibility studies were performed to evaluate the utility of Raman spectroscopy to quantitatively monitor changes in solution concentration and to measure polymorphic composition. The crystallization of flufenamic acid was then monitored as a function of temperature and time using an immersion probe coupled to a Raman spectrometer. The spectral data contained information about both the solution phase and the crystallized solids. Using standard data analysis techniques, solution and solids information could be deconvoluted. Changes in solute concentration as a function of time were quantified from the Raman data and showed good agreement with results from other techniques used for corroboration. The polymorphic forms initially crystallizing from solution, in the presence and absence of seeds, could be clearly identified. Furthermore, the kinetics of the conversion of the metastable form to the stable form could be readily followed. This study illustrates the complex nature of the crystallization process, the ability of Raman spectroscopy to monitor this process, and the potential for this technique to aid in process optimization and control.

longer random numbers since they were proven to be following a certain definite pattern in a simple rectangular plot of h¡a vs. Gf (Rowe, 1974). Literature CitedNo wonder he could not deduce our values from the data presented in Table II! It was needless to say that our raw data were first plotted as /rfa vs. Gf and found to be independent of Gf. Bradshaw and Myers (1963), Chang and Wen (1966), and Balakrishnan and Pei (1974, 1975 also had similar observations and used the same dimensionless groups.Admittedly the range of our data was quite short, as pointed out by Dr. Rowe. However, it is very gratifying to note that other workers (Balakrishnan and Pei, 1974; 1975) have subsequently extended the experimental range to 0.02 < Ar/ReSm2 < 10.0 and vindicated our results.Dr. Rowe would be glad to notice that our data were no