The identification. characterization and quantification of crystal forms ar
e becoming increasingly important within the pharmaceutical industry. A com
bination of different physical analytical techniques is usually necessary f
or this task. In this work solid-state techniques, diffuse reflectance infr
ared Fourier transform spectroscopy (DRIFTS) and X-ray powder diffractometr
y (XRPD) were combined to analyze polymorphic purity of crystalline ranitid
ine-HCl. an antiulcer drug, H2 receptor antagonists. A series of 12 differe
nt mixtures of Form 1 and 2 was prepared by geometric mixing and their DRIF
T spectra and XRD powder patterns were obtained and analyzed, either alone
or combined together, using Artificial Neural Networks (ANNs). A standard f
eed-forward network. with back-propagation rule and with multi layer precep
tron architecture (MPL) was chosen. A working range of 1.0-100%: (w/w) of c
rystal Form 2 in Form 1 was established with a minimum quantifiable level (
MQL) of 5.2% and limit of detection of 1.5% (w/w). The results demonstrate
that DRIFTS combined with XRPD may be successfully used to distinguish betw
een the ranitidine HCl polymorphs and to quantify the composition of binary
mixtures of the two. (C) 2001 Elsevier Science B.V. All rights reserved.