Prediction Of Drucker-Prager-Cap Model Parameters For Mixtures From Single Component Data

head_personhead_person

John Strong

AbbVie
head_personhead_person

Sean Garner

AbbVie

 

Abstract:

In the evolution of drug product development, formulation design has progressed from more of a paint-by-numbers approach to a materials science exercise that considers the properties that individual components add to the blend. The ability to predict at an early stage the manufacturability (e.g., tabletability) and performance of a given blend greatly facilitates the robust and efficient development of formulation drug products. In pursuit of this goal, pharmaceutical scientists have attempted, and to some extent achieved, blend models which utilize individual component data to predict blend behavior. These blend models have spanned a range of complexity from simple lever rules to physicochemical mechanistic approaches. An example of a semi-empirical phenomenological model that has shown some utility is a tablet tensile strength prediction based on individual component tensile strengths. Another popular modeling approach is Finite Element Analysis using the Drucker-Prager/Cap (DPC) model to predict tablet properties such as density distribution. In this work, we hypothesized that the extraction of the DPC model parameters for binary mixtures could be characterized from the individual components of the mixture. This study explored the use of a simple geometric mixing rule approach to predict the DPC model parameters for binary mixtures from the single components. The validity of the geometric mixing rule approach was verified by comparing the loading-unloading behavior for the compaction of cylindrical flat-faced tablets at various weight-to-weight concentrations from both experimental and finite element simulation results. Further verification of the approach was achieved by comparing density distributions results from micro-computed tomography (μCT) and the FEM. The results show that the proposed mixture rule approach can be used to adequately predict the stress and density distributions of binary mixtures based upon the properties of the single component powders.