Analysis of granulation mechanism in a high-shear wet granulation method using near-infrared spectroscopy and stirring power consumption

• Published in: Colloid and polymer science Vol. 298; no. 8; pp. 977 - 987
• Main Authors: Koyanagi, Keita, Ueno, Akinori, Hattori, Yusuke, Sasaki, Tetsuo, Sakamoto, Tomoaki, Otsuka, Makoto
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.08.2020; Springer Nature B.V
• Subjects: Beads; Binding; Calibration; Characterization and Evaluation of Materials; Chemistry; Chemistry and Materials Science; Complex Fluids and Microfluidics; Food Science; Glass; Granulation; Granulators; Hydroxypropyl cellulose; Infrared spectra; Infrared spectroscopy; Moisture content; Molecular interactions; Nanotechnology and Microengineering; Near infrared radiation; Original Contribution; Physical Chemistry; Polymer Sciences; Power consumption; Power management; Regression analysis; Soft and Granular Matter; Spectrum analysis; Water purification; In-line near-infrared spectroscopy; Process analytical technology; High-shear wet granulation; Agitation power consumption; Partial least-squares regression analysis; Chemometrics
• ISSN: 0303-402X; 1435-1536
• DOI: 10.1007/s00396-020-04655-y

The dynamic granulation process of high-speed shear wet granulation (HSWG) was measured by in-line near-infrared spectroscopy (NIRS) and agitation power consumption (APC) methods. Molecular interactions between powder particles and the binding liquid were analyzed based on both NIRS and APC data by multivariable regression analysis. The granulated sample used glass beads ( d 50  = 46 μm) with or without hydroxypropyl cellulose, and the binder solution used purified water. The HSWG granulator (2-L volume) with APC device and NIRS was used, and the agitator was rotated at 600 min −1 and the chopper at 2000 min −1 with glass beads to be granulated being 920 g (0.6 L), and a total of 360 mL of purified water was added at 10 mL/min. In order to establish calibration models to predict APC and amount of binding water of the granular formulations, NIRS spectra of the granular samples were recorded every 10 s for 40 min. The calibration models to predict moisture content and APC were constructed based on the corrected NIRS spectral data by partial least-squares regression (PLSR) analysis. The relationships between actual and predicted values for moisture content and APC produced a straight line, respectively. The regression vector (RV) of the PLS model to predict the water content showed the presence of free water between the bead powder particles. On the other hand, the RV for the APC showed the presence of bound water between the particles.