An exact model for predicting tablet and blend content uniformity based on the theory of fluctuations in mixtures

The content uniformity (CU) of blend and tablet formulations is a critical property that needs to be well controlled in order to produce an acceptable pharmaceutical product. Methods that allow the formulations scientist to predict the CU accurately can greatly help in reducing the development efforts. This article presents a new statistical mechanical framework for predicting CU based on first principles at the molecular level. The tablet is modeled as an open system which can be treated as a grand canonical ensemble to calculate fluctuations in the number of granules and thus the CU. Exact analytical solutions to hard sphere mixture systems available in the literature are applied to derive an expression for the CU and elucidate the different factors that impact CU. It is shown that there is a single ratio, {\lambda}\equiv<w^2.f^2>/<w.f>; that completely characterizes “granule quality” with respect to impact on CU. Here w and f denote the weight of granule and the fractional (w/w) assay of API in it. This ratio should be as small as possible to obtain best CU. We also derive analytical expressions which show how the granule loading impacts the CU through the excluded volume, which has been largely ignored in the literature to date. The model was tested against literature data and a large set of tablet formulations specifically made and analyzed for CU using a model API. The formulations covered the effect of granule size, percentage loading, and tablet weight on the CU. The model is able to predict the mean experimental coefficient of variation (CV) with good success and captures all the elements that impact the CU. The predictions of the model serve as a theoretical lower limit for the mean CV (for infinite batches or tablets) that can be expected during manufacturing assuming the best processing conditions.

💡 Research Summary

The paper introduces a rigorous statistical‑mechanical framework for predicting content uniformity (CU) of pharmaceutical blends and tablets. Recognizing that existing CU models largely ignore excluded‑volume effects and granule loading, the authors treat a tablet as an open system described by the grand canonical ensemble. Granules are modeled as impenetrable hard spheres characterized by weight (w) and fractional API assay (f). Within this ensemble, the number of granules of each type fluctuates according to the chemical potentials μ_i, and the variance of these fluctuations can be derived analytically.

The authors leverage the exact solutions of the Percus‑Yevick (PY) integral equation for single‑component hard‑sphere fluids (Wertheim, Thiele) and the Lebowitz solution for multicomponent hard‑sphere mixtures. These solutions provide the radial distribution function g(r) and, crucially, the matrix of number‑fluctuation covariances for all species. By differentiating the grand‑canonical partition function, they obtain expressions for ⟨ΔN_i²⟩ and cross‑terms ⟨ΔN_iΔN_j⟩ in terms of the derivatives of the chemical potentials.

From these results, a compact expression for the coefficient of variation (CV) of the API dose emerges:

 CV² = (1/D²)