Investigating the Influence of Cosolvents on the Solubility of Febuxostat in Supercritical CO2: Experimental Analysis and Artificial Intelligence Study

Abstract

Assessing the solubility of pharmaceutical compounds in supercritical carbon dioxide (scCO₂) is a fundamental prerequisite for advancing the application of supercritical technologies in the pharmaceutical sector. The supercritical solubility of Febuxostat, a drug with limited solubility often prescribed for gout patients to lower blood uric acid levels, has previously been studied within the pressure and temperature ranges of 12–27 MPa and 308–338 K, exhibiting a mole fraction range of 0.05 × 10^–4–7.42 × 10^–4. This study explores the influence of various cosolvents, namely, ethanol, acetone, and dimethyl sulfoxide (DMSO), on the supercritical solubility of Febuxostat under comparable conditions. Results indicate solubility within the mole fraction range of 0.180 × 10^–4–26.658 × 10^–4 for ethanol, 0.120 × 10^–4–14.810 × 10^–4 for acetone, and 0.108 × 10^–4–14.366 × 10^–4 for DMSO, respectively. Ethanol exhibits the most substantial impact, enhancing the supercritical solubility of Febuxostat by approximately 2.4–3.8 times, while acetone and DMSO contribute to an increase of approximately 2–2.5 times. Furthermore, the study employs empirical models and an artificial neural network (ANN) approach to theoretically investigate the supercritical solubility of Febuxostat with these cosolvents. Among the empirical models, the Jouyban model demonstrates the most accurate correlation for the solubility data of all cosolvents. Moreover, the ANN model demonstrates exceptional accuracy in forecasting the solubility of Febuxostat, with a mean AARD % of 3.207% and R_adj of 0.993 across all experimental measurements for Febuxostat solubility in scCO₂ + ethanol, scCO₂ + acetone, and scCO₂ + DMSO solvents.