Asymptotic limits on tablet coating variability based on cap-to-band thickness distributions: A discrete element model (DEM) study

摘要

Graphical abstract Display Omitted Highlights ? The relationship of cap-to-band ratios in tablet coating is mathematically analysed. ? The spray coating process is modelled using DEM combined with post-process methods. ? Asymptotic limits based cap-to-band ratios are computed and compared with DEM data. ? The spray angle can influence the cap-to-band coating thickness distribution. Abstract The uniformity of the coating thickness distribution is an important quality metric in the manufacture of pharmaceutical tablets during the spray coating process. An investigation of the asymptotic limits of coating thickness variability of tablets of different shapes was carried out based on their cap-to-band coating thickness distributions. A theoretical analysis shows that the cap-to-band coating thickness ratio is expected to be equal to the cap-to-band area ratio projected onto the spray direction divided by the actual cap-to-band surface area ratio. When the cap-to-band projected area ratio is larger (or smaller) than the cap-to-band surface area ratio, the mean coating thickness on the cap is larger (or smaller) than that on the band. To verify this, the dynamics of tablets in a rotating pan was modelled using discrete element method (DEM) simulations, while an image analysis technique based on the output of DEM simulations was applied to model the spray coating process and analyse the cap-to-band coating thickness ratio. A ray-tracing sampling method was further used to obtain the cap-to-band sample ratio. It was also found that a smaller spray angle with respect to the horizontal direction can decrease or even invert the cap-to-band coating thickness ratio, leading to a larger coating thickness on the band than the cap. Nevertheless, an asymptotic value of cap-to-band relative standard deviation can be reached once the cap-to-band coating thickness ratio becomes constant during the coating process. This asymptotic limit is within the range predicted based on the cap-to-band projected area ratio and surface area ratio.