Correlation between compactibility values and excipient cluster size using an in silico approach

Abstract

Background: In silico simulation and percolation theory are important tools in the study of physical and mechanical behavior of pharmaceutical compacts. The aim was to generate a new in silico simulation program that describes the mechanical structure of binary compacts formed from an excipient with excellent compactibility and a drug with null compactibility.

Materials and methods: Paracetamol and microcrystalline cellulose powders were compressed under different pressures. Values for the indentation hardness and tensile strength were measured and fitted to the Leuenberger’s model. On the other hand, compacts with different composition were in silico simulated. In each system, the biggest excipient cluster was identified and quantified using the Hoshen−Kopelman algorithm. Then, the size of the biggest in silico cluster was correlated with experimental compactibility values.

Results and discussion: The Leuenberger’s model resulted in good fit to the experimental data for all formulations over 40% of excipient load. Formulations with high drug load (≥0.8) had reduced range for forming compacts and gave low compactibility values. The excipient percolation threshold for the simulated system was 0.3395, indicating that over this excipient fraction, a compact with defined mechanical properties will be formed. The compactibility values presented a change in the range of 0.3–0.4 of excipient fraction load, just where the in silico excipient percolation threshold was found.

Conclusion: Physical measurements of the binary compacts showed good agreement with computational measurements. Subsequently, this in silico approach may be used for the optimization of pharmaceutical powder formulations used in tablet compression.

Keywords::

• Compaction
• modeling
• percolation theory
• indentation hardness
• tensile strength
• percolation threshold
• in silico simulation

Acknowledgments

To Andrés Granados for his supporting in the development of the computer simulation program and SEM Department of the University of Basel.

Declaration of interest

L. Martinez received a Grant from CONACyT (211351) and R. Villalobos received a fellowship from DGAPA-UNAM (Mexico) during his sabbatical at University of Sydney, Australia. The authors report no conflicts of interest.