(164b) Model-Based Versus Calibration-Based Application of in Situ ATR-FTIR and Raman Spectroscopy

Cornel, J. - Presenter, Institute of Process Engineering, ETH Zurich, Switzerland

Crystallization is one of the most important separation processes to obtain solid material with a high purity and is widely applied in the production of fine chemicals and pharmaceuticals. Identification and characterization of the different kinetic phenomena in crystallization and precipitation, i.e., nucleation, growth and dissolution, are essential for process understanding, allow process modeling and enable process design, optimization and control. Apart from particle size characterization techniques, such as in situ microscopy and laser (back) scattering techniques, spectroscopic tools provide detailed time-resolved information about a given crystallization process. Among others, in situ ATR-FTIR and Raman spectroscopy are both recognized as essential techniques enabling the estimation of solid as well as liquid phase compositions thereby allowing process understanding and enabling process modeling.

The quantitative application of both spectroscopic techniques usually involves multivariate calibration approaches such as principal component regression (PCR) and partial least squares regression (PLSR) to overcome numerical problems due to collinearity [1]. This makes that the quantitative use of both techniques requires extensive calibration efforts which slows down and hampers their application in academia as well as in industry.

Recently, we proposed a calibration-free approach to apply in situ Raman spectroscopy in a quantitative manner to monitor and model a solvent-mediated polymorph transformation [2, 3]. The secondary nucleation rate parameters of the thermodynamically stable polymorph were obtained by fitting the measured time-resolved Raman spectra directly using a detailed process model. A comparison to parameters obtained by fitting the measured solid-phase composition profiles demonstrated that essentially the same process model was obtained. However, by fitting the measured time-resolved Raman spectra directly, lengthy calibration procedures could be avoided.

In this work, a similar approach is presented to characterize the crystal growth rate of paracetamol in water using ATR-FTIR and Raman spectroscopy without any calibration efforts. The classical approach to estimate these growth rate parameters is to measure the desupersaturation profile during seeded batch experiments using for instance ATR-FTIR spectroscopy combined with multivariate calibration techniques. The growth rate parameters are then obtained by fitting the desupersaturation profile using an optimization routine and a population balance equation based model [4]. Using the same population balance model and a similar optimization routine, it was shown that by fitting the measured time-resolved ATR-FTIR spectra directly and by fitting the desupersaturation profiles obtained from these spectra combined with a calibration model, essentially the same crystal growth rate parameters can be estimated. Very similar crystal growth rate parameters could be obtained by fitting the measured time-resolved Raman spectra. The quality of the obtained crystal growth rate was assessed by comparison with literature data. These results show that crystal growth rates can be estimated in a calibration-free manner using both in situ spectroscopic techniques, thereby avoiding lengthy calibration procedures [5].


[1] Cornel, J.; Lindenberg, C.; Mazzotti, M. Quantitative application of in situ ATR-FTIR and Raman spectroscopy in crystallization processes. Ind. Eng. Chem. Res. 2008, 47, 4870.

[2] Cornel, J.; Lindenberg, C.; Mazzotti, M. Experimental characterization and population balance modeling of the polymorph transformation of L-Glutamic acid. Cryst. Growth Des. 2009, 9, 243.

[3] Cornel, J.; Mazzotti, M. Calibration-free quantitative application of in situ Raman spectroscopy to a crystallization process. Anal. Chem. 2008, 80, 9240.

[4] Schöll, J.; Lindenberg, C.; Vicum, L.; Brozio, J.; Mazzotti, M. Precipitation of alpha L-Glutamic acid: determination of growth kinetics. Faraday Discuss. 2007, 136, 247.

[5] Cornel, J.; Mazzotti, M. Estimating the crystal growth rate of paracetamol using in situ ATR-FTIR and in situ Raman spectroscopy in a calibration-free quantitative manner Manuscript in preparation.