(335g) Pharmaceutical powder compaction; an accurate finite-element representation and prediction of binary mixtures

Despite tablets being one of the most commonly used drug dosage forms, the design of a successful tablet production processes still involves a substantial element of trial and error. Tablets are typically made by the direct compression of pharmaceutical powders. However, if this process is not carefully controlled, the resulting tablets can be too weak, too strong, or show defects. This complex problem arises because the tablet’s properties and occurrence of defects all depend on the 1) tablet geometry, 3) powder formulation, and 3) compression settings such as the maximum force.[1] New tools that can predict tablet quality depending on each of these production parameters are highly desirable. In this way, the time-to-market and development costs of new drugs could be reduced significantly.

The finite element method (FEM) is a widely used tool to predict tablet quality as a function of production parameters.[2] Although FEM simulation provides detailed information, it requires detailed parametrisation of the constitutive material model, which is typically done on a case-by-case basis. Moreover, the accuracy of the results is typically limited to a narrow range of pressures.

In this work, we present an automated and highly accurate FEM parametrisation framework for powder compaction.[3] The framework takes experimental data from a compaction simulator and tablet-crushing tests to parametrise the density-dependent Drucker-Prager cap (dDPC) constitutive model for the powder. By using several physical constraints and extrapolations, the framework is highly accurate and valid over a broad range of pressures. Validation test were performed on micro-crystalline cellulose (MCC), dibasic calcium phosphate dihydrate (DCPD), and three mixtures thereof. Simulated and experimental compaction curves showed a mean deviation of 2.5% of the maximum compaction pressure, only slightly more than the experimental variability of 1.2%, over a range from 0 to 500 MPa. Shear stress profiles from simulations also provided a plausible mechanism for the experimentally observed chipping defects.

The newly proposed framework already allows the use of FEM to anticipate the effect of tablet geometry and compaction settings on tablet quality. However, investigating the influence of the powder formulation remains challenging. The current framework requires experimental data of each and every powder mixture, making the entire process incredibly cumbersome for even a small number of components.

In this work, we also propose a set of mixing rules using the isostress assumption (i.e. that the stresses on all the individual components of the powder are the same). The mixing rules were then be used to determine the FEM parametrisation of several mixtures using only the FEM parameters of the individual pure components. Three powder mixtures of MCC and DCPD were used for validation. The predicted and experimental compaction curves showed a mean deviation of 4.8% of the maximum compaction pressure. The predicted shear stress profiles were similar to those previously simulated, providing the same mechanism for the chipping defects. The mixing rules can thus be used to estimate the influence of the powder formulation on the tablet quality.

Finally, the effect of mixture homogeneity was investigated. Because the mixing rules using the isostress assumption gave accurate predictions, a powder distribution that gives a similar stress distribution for all materials should also result in accurate predictions. Simulating tablets in which MCC and DCPD were separated into two horizontally stacked layers resulted in a mean 4.5% error with respect to the experimental homogeneous mixtures. This again justifies the use of the isostress assumption. A further comparison of spatial arrangements of the powder (vertical layers and a staircase pattern) showed that the residual radial pressure, a key indicator of tablet defects, was lower when the softer material (MCC) was placed near the die wall.

To conclude, the framework can be used to obtain accurate FEM parameterisations for powder compaction simulations. The framework also was extended with a number of mixing rules. Using either experimental data or mixing rules resulted in accurate predictions of the compaction curve and a similar chipping defect mechanism. In the future, this framework can be used to systematically asses the influences of tablet geometry, compaction force, and powder formulation on relative tablet quality, using only experimental data of the constituent single-component powders.