The application of percolation threshold theory to predict compaction behaviour of pharmaceutical powder blends

Percolation theory provides a statistical model which can be used to predict the behaviour of powder blends based on particle-particle interactions. The aim of this study was to investigate if percolation theory could be used to predict the drug loading concentration of pharmaceutical tablets, and the relative density of a blend, above which tablet tensile strength is reduced, resulting in the production of unsatisfactory products. The model blend studied contained ibuprofen as the API, which exhibits poor flow and compressibility, and microcrystalline cellulose (MCC) as the excipient, which exhibits good flowability and compressibility. Two MCC grades with differing physical properties were investigated, Vivapur (R) 102 (air streamed dried quality), and Emcocel (R) 90 (spray dried quality) to test the theory. Blends containing 2.5 to 40% w/w of ibuprofen were compacted at a range of pressures and the values of the powder true density, compaction pressure, tablet envelope density, and tablet tensile strength were used to calculate the percolation thresholds mathematically. The drug loading threshold values predicted with the model (19.08% w/w and 17.76% w/w respectively for Vivapur (R) 102 and Emcocel (R) 90) were found to be in good agreement when compared to experimental data and the infinite cluster of drug was visually confirmed on the surface of tablets using Raman imaging. The capability of multivariate analysis to predict the drug loading threshold was also tested. Principal component analysis was unable to identify the threshold, but provided an overview of the changes of the analysed properties as ibuprofen drug loading increased. It was also able to identify differences between blends containing Vivapur (R) or Emcocel (R). In conclusion, percolation theory was able to predict the maximum acceptable drug loading for this binary system of API and excipient. This methodology could be employed for other binary systems to predict maximum drug loading potential without the need for time consuming and expensive tablet production.