Journal
INTERNATIONAL JOURNAL OF PHARMACEUTICS
Volume 385, Issue 1-2, Pages 29-36Publisher
ELSEVIER
DOI: 10.1016/j.ijpharm.2009.10.019
Keywords
FT-Raman spectroscopy; D-Mannitol; Crystal polymorphism; Multivariate calibration; PLS regression; Artificial neural networks; Chemometry; Quantitative analysis
Categories
Funding
- University of Innsbruck
Ask authors/readers for more resources
D-Mannitol is known to exist in five solid-state forms, a hemihydrate, an amorphous form and three polymorphic forms (I degrees, II and III), which tend to crystallize concomitantly. Therefore, a fast and simple method for the simultaneous quantification of these polymorphs in powder mixtures was developed on the basis of FT-Raman spectroscopic data. partial least-squares (PLS) regression and artificial neural networks (ANNs). A combination of the first derivative and orthogonal signal correction (OSC) was found to be the optimal data pretreatment that significantly increased the predictive performance of the models. The RMSEPs (root-mean-squared errors of prediction) obtained by PLS for the modifications (mods) I degrees, II and III were 0.44%, 0.34% and 0.36% respectively The estimated limits of detection are similar to 0.5% (mod. I degrees) and <1% (mods II and III). The ANNs model yielded slightly higher RMSEP values of 0.51%, 0.39% and 0.41%. In contrast to related previous studies. calibration was performed with carefully prepared ternary mixtures of all polymorphs, which is one of the reasons for the high precision and accuracy of the presented multivariate models. (C) 2009 Elsevier B V. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available