Examining particle size growth in twin screw granulation up to steady state with acoustic emissions

Abstract

The transient evolution of granules was studied to learn new details about the underlying mechanism for continuous wet granulation in a twin-screw extruder. Sieving and a new inline PAT for particle size development was used to gain these insights. The onset for steady state was established based on observing a consistent PSD, which occurred at five times the mean residence time of the process, over a range of degrees of fill (DF; 12–30 %). The early stages of startup for granulation were captured by the inline PAT, showing different stages of granule growth for particle sizes ranging from 102 to 2230 μm. The analysis found that conveying elements have a stronger influence on granule growth at a low DF whereas the kneading zone had a stronger influence on granule growth at a higher DF. This study presents new details on this black-box process while highlighting the unique value of PAT to twin-screw granulation.

Highlights

Granulation at a higher degree of fill (DF) becoming increasingly dependent on the kneading zone.

Steady state will be reached by 5 times the mean residence time independent of DF.

Inline PAT can output a representative particle size distribution with only 5 s sampling.

Introduction

In the pharmaceutical industry, most global regulatory bodies are encouraging the development of advanced quality control strategies around the manufacturing of drug products, with specific focus on continuous processes. Twin screw granulation has gained significant interest from the pharmaceutical industry as one such process over this past decade [1]. Twin screw granulation (TSG) offers significant advantages over comparable high shear batch processing, including systematic integration with other unit operations, increased throughput, shorter process (residence) time, and higher product consistency [2, 3].

The transition to steady state in TSG has gone unstudied to date because the process operates as a black-box and most of our available characterization tools need substantial sample sizes to draw conclusions and yet this period of time may reveal significant details about the granulation mechanism. Understanding how granule sizes evolve during this early period up till a steady particle size distribution is obtained, should better highlight the mechanism inside this machine, which is often described using the classical stages of nucleation, propagation and granule breakage seen in batch processes [4, 5] without evidence that they accurately describe what is taking place inside the extruder (especially its compression zone). The transient period may be too short to fully address our gaps in knowledge about this process, but its study has the potential to add new information on a continuous processing method that has just begun to produce commercial drugs. The complexity of the process is attributed to the confounding influences of its process variables, including formulation properties [6, 7], screw design [7, 8], screw speed [9], mass flowrate [4, 10], and liquid-to-solids (L/S) ratio [9, 11]. These variables are strongly affected by the degree of fill (DF), with the granulation mechanism seemingly changed as a result. By the process mapping done to date it has been observed that at lower DF, granules are thought to experience less frictional forces and are able to move freely in the barrel, which allows for increased growth and consolidation [6, 12]. Conversely, at higher DF, granules are more constrained, moving in a plug-like manner, which is believed to reduce the number of particles being wetted and yielding more fines produced [[13], [14], [15]]. Shi [16] confirmed this transition in the granulation mechanism was around a DF of 20 %, observed in both 18 mm and 27 mm twin-screw granulators, where the impact of screw speed on the PSD below this DF is more prominent than the mass flow rate, and vice versa above this value. Seem et al. [14] also mentioned in their review that screw speed does have an impact on granulation at very high DF (we estimate at 60 % DF and onwards), which indicates that effects of screw speed may only be prominent at the extremes (very low or high DF) from a process mapping standpoint.The only studies known to have been explicitly concerned with whether the process in a TSG was steady, relied on torque measurements and then only in a GEA ConSigma® machine [9, 11, 14, 17]. While torque is a helpful variable to monitor operational consistency and will have some indirect correspondence to the state of the compression zone where binder spreading and particle agglomeration will primarily be occurring, it is not a direct descriptor of output performance such as end product responses like fracture strength and dissolution rate, or the more universally required information like particle size distribution (PSD). In fact, Ryckaret et al. [18], who provided some of the most thorough studies on the effects of torque in twin-screw granulation, concluded that its inline measurement cannot be used to monitor particle size. Additionally, none of these studies examined what was happening inside the TSG during the transition period in their machines.

The purpose of this study was to explore the evolution of granulation during the transient start-up period after first establishing when steady state occurred for the PSD. The work compares the capabilities of off-line sieve classification versus a recently developed inline PAT based on acoustic emission (AE) [19] to understand how PSD evolves in a twin-screw granulator. The term PAT in this paper refers to an inline acoustic sensor coupled with a trained artificial intelligence (AI) model. The goals of the research were to assess the value of inline systems on their capacity to observe phenomena in a continuous process on a very short timescale. This study was also the first opportunity to relate mean residence time (MRT) to the time to reach steady state in a twin-screw granulator, which is commonly understood information for other chemical processes [[20], [21], [22], [23]].

Download the full article as PDF here Examining particle size growth in twin screw granulation up to steady state with acoustic emissions

or read it here

Materials

All trials were conducted with a placebo formulation of 60 wt% Flowlac® 100 lactose monohydrate (Meggle Pharma; Germany), 20 wt% Avicel® PH-102 microcrystalline cellulose (International Flavors and Fragrances; USA), and 20 wt% Kollidon SR® polyvinyl acetate/povidone (BASF; USA). The formulation was pre-mixed and then dried at 50 for 24 h before use. The binder solution consisted of 2 wt% METHOCEL E3PLV (International Flavors and Fragrances; USA) dissolved in distilled water.

H.A. Abdulhussain, M.R. Thompson, Examining particle size growth in twin screw granulation up to steady state with acoustic emissions, Powder Technology, Volume 448, 2024, 120294, ISSN 0032-5910, https://doi.org/10.1016/j.powtec.2024.120294.