Abstract
Twin-screw wet granulation is a crucial unit operation in shifting from pharmaceutical batch to continuous processes, but granulation kinetics as well as residence times are yet poorly understood. Experimental findings are highly dependent on screw configuration as well as formulation, and thus have limited universal validity. In this study, an experimental design with a repetitive screw setup was conducted to measure the effect of specific feed load (SFL), liquid-to-solid ratio (L/S), and inclusion of a distributive feed screw on particle size distribution (PSD) and shape as well as residence time distribution of a hydrophilic lactose/microcrystalline cellulose based formulation. An intermediate sampling point was obtained by changing inlet ports along the screw axis. Camera-based particle size analysis (QICPIC) indicated no significant change of PSD between the first and second kneading section, except for low L/S and low SFL where fines increase. Mean residence time was approximated as a bilinear fit of L/S and SFL. Moreover, large mass flow pulsations were observed by continuous camera measurements of residence time distribution and correlated to hold-up of the twin-screw granulator. These findings indicate fast granulation kinetics and process instabilities for high mean residence times, questioning current standards of two kneading compartments for wet granulation. The present study further underlines the necessity of developing a multiscale simulation approach including particle dynamics in the future.
Highlights
Industrial wet granulation is presently mainly carried out by batch processing
It is of note that a significant influence of distributive feed screw (DFS) on particle size was strongly dependent on the chosen evaluation parameters, e.g., evaluation with partial least squares regression or multiple linear regression, inclusion of quadratic interactions, and alpha = 0.05 versus alpha = 0.01
liquid-to-solid ratio (L/S) had the greatest influence on d50 and L/S, whereas specific feed load (SFL) and screw length had the greatest influence on Mean residence time (MRT)
Summary
Industrial wet granulation is presently mainly carried out by batch processing. Recently, the possibility of inline quality testing, real-time release, batch-size flexibility, and reduced footprint, among other advantages, has sparked a growing interest in continuous processing [1,2], especially in the pharmaceutical and bio-pharmaceutical domains [3]. TSG simulations were primarily developed based on these specific DoEs. Alongside, neural-network [16,17] and computational fluid dynamic [18] approaches, as well as population balance modeling (PBM)—see, e.g., [19,20,21]—have been utilized, mainly to model high-shear granulation processes [22]. Neural-network [16,17] and computational fluid dynamic [18] approaches, as well as population balance modeling (PBM)—see, e.g., [19,20,21]—have been utilized, mainly to model high-shear granulation processes [22] Most of these developments have been successful in predicting critical quality attributes (CQAs), like the resulting particle size distribution (PSD). The setup allows to collect data at an intermediate point by moving the inlet ports for liquid and powder to reduce the screw length, and will help modeling different sections independently of each other to monitor granulation kinetics. Results, and new findings of the conducted DoE will help to calibrate and validate the simulation framework in a later stage of research
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
