Modeling of reduction kinetics of Cr<sub>2</sub>O<sub>7</sub><SUP>-2</SUP> in FeSO<sub>4</sub> solution via artificial intelligence methods

Creative Commons License

SEVİM F., AYDIN T., IRMAK M. C.

SCIENTIFIC REPORTS, cilt.15, sa.1, 2025 (SCI-Expanded) identifier identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 15 Sayı: 1
  • Basım Tarihi: 2025
  • Doi Numarası: 10.1038/s41598-025-13392-7
  • Dergi Adı: SCIENTIFIC REPORTS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, BIOSIS, Chemical Abstracts Core, MEDLINE, Veterinary Science Database, Directory of Open Access Journals
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • Atatürk Üniversitesi Adresli: Evet

Özet

This study aims to model the reduction kinetics of potassium dichromate (K2Cr2O7) by ferrous ions (Fe2+) in sulfuric acid (H2SO4) solutions using artificial intelligence-based regression models. The reaction was monitored potentiometrically under controlled hydrodynamic conditions, and an experimental dataset was generated by varying key parameters including temperature, stirring speed, grain size, and Fe2+ and H+ concentrations. The dataset contains 263 data points representing the conversion rates at different time intervals and experimental conditions. To explore the predictive capabilities of AI in modeling complex chemical kinetics, we applied and compared several regression models: Gradient Boosting, Random Forest, Decision Tree, K Nearest Neighbors, Linear, Ridge, and Polynomial Regression. Hyperparameter tuning was performed using random search to optimize each model's performance. Among these, the Gradient Boosting Regression model demonstrated the best accuracy with an R2 value of 0.975 and RMSE of 0.046. Feature importance analysis revealed that stirring speed and temperature were the most influential parameters. To our knowledge, this is the first study to model the Cr(VI) reduction kinetics using a broad range of AI-based regression methods applied to experimentally derived data. The findings highlight the potential of artificial intelligence in replacing conventional kinetic modeling with faster, cost-effective, and highly accurate approaches.