Engineers develop green extraction technology for essential oils

As part of efforts to promote sustainable manufacturing, an international team of chemical engineers has developed a high-efficiency green extraction technology that could redefine the future of the global essential oil industry.

The innovation was led by a Nigerian researcher, Obiora Muojama of the University of Alabama, Tuscaloosa, AL, USA, whose contributions helped deliver one of the most energy-efficient and environmentally responsible extraction systems.

According to the research published in the Elsevier Journal, the innovation was developed by a team comprising Muojama; Professor Heri Septya Kusuma of Universitas Pembangunan Nasional “Veteran” Yogyakarta, Indonesia; and Professor Andrew Amenaghawon, a leading researcher in sustainable process engineering at the University of Benin.

It noted that a peer-reviewed study introduced a solventless microwave hydrodistillation process powered by advanced machine learning and metaheuristic optimisation techniques, producing nearly three times more oil than conventional methods while cutting energy consumption and carbon emissions by more than 80 per cent.

“For decades, essential oils used in perfumes, skincare products, aromatherapy, wellness goods, and pharmaceuticals have been made through steam distillation, a slow, energy-heavy, and resource-intensive approach. This new research overturns the long-standing paradigm with a system that delivers an oil recovery rate of 2.949 per cent, one of the highest yields ever recorded for patchouli, in just 85 minutes – far less than the three to four hours typical in factory settings. The method uses zero chemical solvents, produces no wastewater, and is directly aligned with circular-economy principles and global climate priorities.

“What sets this breakthrough apart is not just the experimental method but the sophisticated computational framework layered on top of it. The research team combined Box-Behnken statistical design with three machine-learning models: Artificial Neural Networks, Kernel Ridge Regression, and Extreme Gradient Boosting, supported by SHAP interpretability analysis and a powerful biological optimisation algorithm known as Manta Ray Foraging Optimisation,” the study mentioned.

It explained that they built a digital twin of the extraction system, enabling them to evaluate thousands of conditions in silico and identify the optimal extraction parameters: 570 W microwave power, a 1.0 g/mL feed-to-solvent ratio, and 85.41 minutes extraction time.