Discrete element analysis of the effect of particle cohesion on die filling behaviour of pharmaceutical powders in a rotary tablet press

Abstract

This study investigates the effect of particle cohesion on die filling behaviour in a rotary tablet press using a combined Discrete Element Method (DEM) and experimental approach. Microcrystalline Cellulose (MCC) spheres were characterised experimentally and used as the model powder. A validation study is conducted where tablet weight and its variability, obtained from DEM simulations, and corresponding experiments are compared. In addition, Residence Time Distribution (RTD), with both pulse and step change inputs, are obtained numerically and experimentally to further validate the DEM models. A satisfactory agreement between the DEM and experimental results was obtained. Systematic DEM simulations are then performed to explore the influence of powder cohesion on die filling behaviour. The results revealed that increasing powder cohesion reduces tablet weight (uniformity) considerably under constant fill depth. Additionally, more cohesive powders have a propensity to remain in the feed frame for a longer time under similar process conditions. It is also shown that optimal tablet weight with minimal variability could be obtained at intermediate turret speeds, whereas high turret speeds (90 rpm) are associated with lower tablet weight uniformity. Additionally, at low turret speeds, increasing paddle speed to an intermediate level improve process efficiency, notably through increases in mean tablet weight and tablet weight uniformity when fill depth is kept constant. These simulations elucidate the critical factors affecting tablet manufacturing, allowing for the optimisation of process parameters to maximise weight uniformity and performance while minimising experimental burden.

Introduction

Rotary tablet presses are commonly used to manufacture pharmaceutical tablets at high throughputs. However, in early-stage drug product development, it can be challenging to conduct full-scale experiments on a rotary tablet press due to the limited availability of active pharmaceutical ingredient (API). Moreover, during the research and development stage, APIs can be very costly. Although experiments can be performed using small-scale equipment, such as single-punch tablet compression machines, the powder flow dynamics and die filling of a rotary press cannot be investigated. For this reason, many efforts have been made to mimic and simplify die filling systems of rotary tablet presses (Jackson et al., 2007, Schneider et al., 2007, Wu, 2008, Wu and Guo, 2012, Zheng et al., 2022). In these cases, the die filling process is normally analysed with one stationary or moving die in combination with a moving or stationary feeder, respectively.

Depending on process parameters and material properties, three distinct die-filling mechanisms can be identified: gravity feeding, forced feeding, and suction filling (Baserinia and Sinka, 2019). In rotary presses, the specific filling mechanism is mainly determined by the design of the fill cam and its alignment with the feed frame opening. Thus, depending on the positioning and movement of the lower punch, both gravity filling and suction filling can be considered. Gravity filling occurs when the lower punch is at the overfilling position, allowing powder to enter the die by nose flow, bulk flow, or intermittent flow. Suction filling occurs when the lower punch moves down during the filling process to create a partial vacuum that draws the powder into the die. A combined gravity and suction filling can also be realised when the lower punch is positioned at an intermediate height between the die table and overfilling positions. (Wu and Guo, 2012).

More complex die-filling systems employed include a die table with multiple dies and a single chamber feed frame equipped with a paddle (Ketterhagen, 2015, Li et al., 2023, Mateo-Ortiz and Méndez, 2016). In the study of Mateo-Ortiz and Méndez (2016), die-filling process in a single chamber feed frame was simulated to investigate the particle dynamics in a confined space by considering various operating parameters and paddle wheel design. It was demonstrated that feed frame design could highly impact tablet weight variability, attrition, and segregation, all of which can alter tablet quality. Tabletting performance was assessed using parameters, such as tablet weight, tablet weight variability, hold-up mass, and Residence Time Distribution (RTD) (Goh et al., 2017, Pauli et al., 2020, Sinka et al., 2009).

Recent advancements in computing hardware and software, especially the usage of the GPU in accelerating DEM simulations, have made it possible to simulate full-scale die-filling systems (Forgber et al., 2022, Mateo-Ortiz and Méndez, 2015, Mateo-Ortiz et al., 2014, Siegmann et al., 2020). For example, Forgber et al. (2022) evaluated the RTD for a production-scale rotary press machine using DEM simulations and validated the RTD results against experiments. In addition to the evaluation of the tablet weight and its variability, the RTD provides further insights into the mixing dynamics of the feed frame. This knowledge enables manufacturers to optimise operating conditions and formulations, enhance product consistency, and ensure the reliability of oral solid dosage forms (Furukawa et al., 2020, Pauli et al., 2020). Typically, it involves conducting labour-intensive experiments using a tracer material. Two major addition methods can be identified to introduce a tracer into the system, namely pulse/impulse and a step-change (Escotet Espinoza, 2018, Jansson, 1980). In the case of a pulse input, a precise and instantaneous introduction of a tracer is allowed to the system. A step-change experiment involves the sudden, continuous introduction of a tracer into the system at a constant rate. After introducing the tracer to the system, the concentration of the tracer as a function of time is recorded at the outlet of the system and offers critical information about process dynamics.

Although several previous studies have compared DEM simulations with experimental measurements in feed frames or die-filling systems, comprehensive quantitative validation remains scarce. Specifically, most prior works have focused on validating a single metric, such as Residence Time Distribution (RTD) or particle velocity, often using qualitative or semi-quantitative comparisons. In contrast, validation linking both intermediate process dynamics (e.g., RTD) and final product attributes (e.g., tablet weight and weight variability) within a single, consistent DEM framework has not yet been reported. Furthermore, while previous experimental studies have demonstrated that powder cohesion affects die filling, a systematic and quantitative understanding of how the degree of cohesion influences die-filling behaviour remains limited. This gap is particularly pronounced under dynamic conditions relevant to high-speed rotary tablet presses, where the interplay between cohesion, turret speed, and paddle rotation direction governs filling performance.

Therefore, the objective of this study is to assess die-filling quality in a rotary tablet press using a fully validated DEM approach. The novelty of this work lies in three key aspects: (i) a comprehensive, quantitative validation against both RTD (process dynamics) and tablet weight/variability (final product attribute) using identical experimental conditions and particle properties; (ii) a systematic investigation of how the degree of powder cohesion affects die-filling performance across a range of turret speeds and paddle rotation directions; and (iii) the use of multiple performance metrics, including tablet weight, weight variability, RTD, and tracer particle trajectories, to provide a mechanistic understanding of die filling under cohesive conditions. By directly comparing DEM predictions with experimental measurements obtained via an NIR probe inserted into the feed frame, this study aims to establish a validated modelling approach that can be used to guide process design and troubleshooting in high-speed rotary tablet pressing.

Continue reading here

Particle Characteristics

Spherical particles of Microcrystalline Cellulose (MCC), VIVAPUR Spheres 350 (JRS Pharma GmbH & Co. KG, Rosenberg, Germany) and VIVAPHARM Sugar Spheres, were used in this study. Sugar spheres were selected as tracer material based on their similarity with MCC spheres and ability to measure its concentration in the powder blend using NIR.

M. Alizadeh Behjani, L. De Souter, C. Zheng, B.J. Nitert, T. De Beer, C.-Y. Wu, Discrete element analysis of the effect of particle cohesion on die filling behaviour of pharmaceutical powders in a rotary tablet press, International Journal of Pharmaceutics, 2026, 126986, ISSN 0378-5173, https://doi.org/10.1016/j.ijpharm.2026.126986.