In a significant stride towards optimizing pharmaceutical processes, researchers have developed new models to predict the solubility of antihypertensive drugs in supercritical carbon dioxide (SCCO2). This breakthrough, led by Randa H. Atwan from the Faculty of Pharmacy at Al-Ahliyya Amman University in Jordan, promises to revolutionize drug formulation and process design, with potential ripple effects across the energy sector.
The study, published in the Journal of Carbon Dioxide Utilization, focuses on nine antihypertensive drugs from diverse pharmacological classes. These include thiazide diuretics, ACE inhibitors, calcium channel blockers, alpha- and beta-blockers, and vasodilators. The research team compiled a dataset of 196 experimental data points, spanning a broad range of temperatures (308–343 K) and pressures (9–30 MPa).
Atwan and her team initially evaluated nine existing empirical correlations from the literature. While these correlations performed well for specific compounds or limited operating ranges, they lacked consistency across all antihypertensive drugs. “Existing models often fell short when applied broadly,” Atwan explained. “We needed a more robust solution that could handle the diversity of these drugs and the range of conditions.”
To address this, the researchers developed a new six-parameter empirical correlation based on physically meaningful variables. This model achieved an impressive overall Mean Absolute Percentage Error (MAPE) of 5.79%, an R² of 99.71%, and the lowest Akaike Information Criterion (AIC) among all tested models. “Our new model outperformed all existing correlations, maintaining a MAPE below 10% for every drug,” Atwan noted.
But the innovation didn’t stop there. To eliminate the need for drug-specific fitting, the team designed a generalized Gaussian Process Regression (GPR) model. This model used the same thermodynamic inputs, along with molecular weight and melting temperature, achieving an excellent testing MAPE of 4.47% and an overall R² of 99.90%. “The GPR model is particularly exciting because it doesn’t require re-adjustment for each drug,” Atwan said. “It’s a more scalable and practical solution.”
The implications of this research extend beyond the pharmaceutical industry. Supercritical CO2 is a versatile solvent used in various industrial processes, including decaffeination, polymer production, and more recently, in the energy sector for enhanced oil recovery and carbon capture and storage. Accurate solubility predictions can enhance the efficiency and cost-effectiveness of these processes.
The study also confirmed that the dataset spans both below- and above-crossover regions, allowing the models to capture dual temperature–pressure effects. Trend and sensitivity analyses further confirmed the physical consistency and interpretability of both models.
As the world seeks more sustainable and efficient industrial processes, this research offers a robust and scalable tool for predicting drug solubility in SCCO2. It underscores the potential of machine learning and empirical modeling in driving innovation and improving process design.
“Our framework provides a solid foundation for future developments,” Atwan concluded. “It’s not just about antihypertensive drugs; it’s about creating a versatile tool that can be adapted for various applications, including those in the energy sector.”
This research is a testament to the power of interdisciplinary collaboration and the potential of advanced modeling techniques to shape the future of industrial processes. As the world continues to grapple with the challenges of sustainability and efficiency, such innovations will be crucial in driving progress and achieving our goals.