Roller compaction: Ribbon splitting and sticking

Abstract

Roller compaction is the main technique employed in dry granulation. Ribbon sticking and splitting are among the major factors that can hinder the use of this process for some formulations. Ribbon splitting can occur either transversally (through the ribbon thickness) or longitudinally (through the ribbon width). It was observed that transverse splitting is commonly associated with sticking of the split ribbons to the rollers and results in an inferior performance of the process. Longitudinal splitting is associated with an across-width distribution of the ribbon density so that there may an adverse effect on the mechanical strength and dissolution properties of the tablets formed from the milled granules. The aim of the current work was to elucidate the mechanisms of splitting by an experimental study involving single component powders with a range of yield strengths, including those that are commonly used as excipients. Both smooth and knurled rollers were employed without and with lubrication by applying magnesium stearate to the rollers. The minimum gap was fixed and the maximum roll stress was varied. The observed trends for the smooth rollers were rationalised in terms of a splitting index, which is a measure of the residual stresses driving crack growth relative to the tensile strength of the ribbons. There was a lower limit at which splitting was observed but the occurrence of transverse splitting decreased and that for longitudinal splitting increased with increasing values of the index, which was accompanied by an increase in mixed transverse-longitudinal splitting. Transverse splitting was always associated with sticking to the rollers and was prevented by external lubrication. The main difference with the knurled rollers was that in some cases transverse splitting occurred without sticking to the rollers. A detailed discussion of the mechanisms involved is presented.

Introduction

Roller compaction is a pressure mediated dry granulation process by which materials are agglomerated without a liquid binder. A powder is fed between two counter-rotating rollers by gravity from either a hopper or a screw feeder. Initially, the powder slips against the rollers but eventually the increase in the wall shear stress is sufficient to prevent such wall slip; this occurs at the so-called nip angle, which is the downstream angular difference relative to that at the minimum gap between the rollers. In this compaction zone, high stresses are applied to produce a ribbon, which is subsequently milled into granules of a specified particle size range (Kleinebudde, 2004, Omar et al., 2015).

Despite roller compaction being well established, there remain intrinsic challenges to ensuring that the design and operation is sufficiently optimal. Leakage of relatively high amounts of uncompacted powder between the seals of the rollers can greatly affect the process efficiency. The heterogeneity of the ribbons in terms of the density distribution across the width has a negative impact on the uniformity of downstream granules (Gamble et al., 2010, Guigon and Simon, 2003). Ribbon-roller adhesion can also cause serious difficulties. Hamdan et al. (2010) observed that variations in the roll gap, and hence ribbon density, could be attributed to such adhesion. Dawes et al. (2012a), concluded that the sticking of an adhesive powder formulation can be inhibited by external lubrication of the rollers using magnesium stearate (MgSt). They also investigated the effects of internal lubrication by incorporating MgSt in the powder feed. It was observed that without MgSt, sticking occurred 30–60 s after start up and was more problematic when operating in the automated roll gap control mode, which controls the set gap by adjusting the speed of the screw feeder. The extent to which the speed of the auger changes, and hence the mass throughput, depends on the thickness of the adhered layer (Dawes et al., 2012b). However, ribbons did not stick when MgSt was added to the formulation. When the powder adheres to the rollers, it disturbs the integrity of the ribbon and so affecting the continuity of the process.

There are a considerable number of studies that have investigated the phenomenon of powder-metal adhesion in the field of tabletting. The adhesion to the punches is mainly attributed to either the formulation or the process conditions. Hygroscopicity, cohesiveness, low melting point, insufficient lubrication and/or unsuitable powder characteristics such as particle size and polymorphism are the main formulation based factors (Lam and Newton, 1991, Paul et al., 2017, Podczeck et al., 1996). The use of an unsuitable surface configuration, excessively small compaction pressures, and/or rough surfaces are the main process based causes (Saniocki, 2014). However, sticking to a roller surface has not been investigated in detail where a different combination of shear and normal stress are involved.

Ribbon splitting is one of the main problems (Osborne, 2013), which has the potential to occur in different directions because of the complexity of the stress field imposed on the powder feed. Guigon et al. (1996)observed different types of ribbon splitting as one of the common defects that could occur when using fine powders. Wu et al. (2010) investigated the effects of the moisture content of microcrystalline cellulose (MCC) as a feed powder, and noted that the ribbons produced at high values (>11.44%) tended to split across the width. It was ascribed to the resulting reduction in the tensile strength of the ribbons. Ende et al. (2007) reported that, for a range of powder feeds, the occurrence of ribbon splitting (the exact location of the splitting was not mentioned) was minimised when the gap was <2.6 mm. Cunningham et al. (2010) suggested that the reversal of the direction of the shear stress at the neutral angle could be a contributory factor given that particle compacts are weak in shear. However, it would be difficult to evaluate experimentally. Ribbon splitting is analogous to the capping and lamination phenomena observed in a tabletting process, which is ascribed in the literature to various process and formulation parameters (Hiestandx et al., 1977, Paul and Sun, 2017, Wu et al., 2008). Critical factors are the stored elastic strains, non-uniform density distribution, sticking to the punches and air entrainment.

Thus, previous research has focussed mainly on the consequences of ribbon splitting and some of the mitigation approaches. The aim of the current work was improve the mechanistic understanding of this phenomenon in terms of the ribbon-roller interactions in both the bulk and at the roller walls. Feed powders with a range of yield strengths were roller compacted and the nature of the cracks in the ribbons was observed with and without MgSt applied directly to the rollers. Depending on the applied pressure of the rollers, there were two types of cracks that occurred in isolation or in combination: through-width and through-thickness, which will be referred to as longitudinal and transverse cracks. It was possible rationalise this behaviour by measurement of the density distribution across the ribbons, their tensile strength and their adhesion to the rollers.

The excipients were selected to produce a gradual change in the mechanical properties, i.e. mechanical properties that ranged from CaCO₃ as hard particles and a high yield strength value to a more plastically deformable particles of MCC. Seven different powders with a range of mechanical properties were selected: calcium carbonate (CaCO₃) (Longcliffe Quarries Ltd., UK), maltodextrin (Glucidex 6, Roquette, France), α-lactose monohydrate 200 M (GranuLac®200, MEGGLE Germany), microcrystalline cellulose (MCC) (VIVAPUR®101, JRS PHARMA, Germany), anhydrous lactose (NF-DT, Kerry bioscience, USA), mannitol C160 (Pearltol®C160, Roquette, France), and pregelatinized maize starch (Starch 1500®, Colorcon, UK). The effects of externally lubricating the rollers were investigated by applying 0.25 g magnesium stearate (MgSt) (Peter-Greven). Scanning electron micrographs of the particles are shown in Fig. A1

The powders were equilibrated for three days at a relative humidity of 40% and a temperature of 25 °C in an environmental chamber (Binder KMF 240 climatic chamber, Binder, UK). Their particle size distributions were measured using a Camsizer XT (Retsch Technology GmbH, Germany) as shown in Table 1 together with their true densities as supplied by the manufacturer.