ML-Based Corrosion Inhibition Prediction of 3- Substituted Chromone Derivatives
Abstract
This study assessed eleven 3-substituted chromone derivatives for their effectiveness as corrosion inhibitors in acidic environments. Eight compounds were tested in the lab, and a Random Forest (RF) model was built using electrochemical parameters (Rct, Cdl, Icorr, Ecorr) to predict how well they would work as inhibitors. The model had a high correlation between predicted and actual efficiency, with strong predictive performance (R² = 0.7984, RMSE = 2.3073, MAE = 2.0172). Compounds modified with hydroxyl (–OH) exhibit enhanced inhibitory activity owing to improved adsorption and electron-donating properties. The validated model was subsequently applied to three chromone acrylic ester compounds that had not been previously examined. These compounds exhibited a high predicted inhibitory efficiency (~94–95%) with a pronounced concentration-dependent trend. The results show that machine learning can speed up the process of testing effective corrosion inhibitors by a lot while also cutting down on the amount of work needed.References
[1] T. W. Quadri, L. O. Olasunkanmi, O. E. Fayemi, M. M. Solomon, and E. E. Ebenso, “Zinc Oxide Nanocomposites of Selected Polymers: Synthesis, Characterization, and Corrosion Inhibition Studies on Mild Steel in HCl Solution,” ACS Omega, vol. 2, no. 11, pp. 8421–8437, Nov. 2017, doi: 10.1021/acsomega.7b01385.
[2] M. Sharma et al., “Enhanced corrosion resistance of low alloy steel in NaCl environment using 3-substituted chromone derivatives in epoxy coatings,” J. Mol. Struct., vol. 1327, p. 141151, 2025, doi: https://doi.org/10.1016/j.molstruc.2024.141151.
[3] D. A. Winkler, M. Breedon, P. White, A. E. Hughes, E. D. Sapper, and I. Cole, “Using high throughput experimental data and in silico models to discover alternatives to toxic chromate corrosion inhibitors,” Corros. Sci., vol. 106, pp. 229–235, 2016.
[4] X. Luo et al., “4-aminoazobenzene modified natural glucomannan as a green eco-friendly inhibitor for the mild steel in 0.5 M HCl solution,” Corros. Sci., vol. 151, pp. 132–142, 2019, doi: https://doi.org/10.1016/j.corsci.2019.02.027.
[5] M. Sharma, R. K. Tiwari, and S. Kumar, “Characterization and inhibition efficiency of HCAE as a corrosion inhibitor in saline and acidic environments,” J. Integr. Sci. Technol., vol. 12, no. 6, p. 840, 2024.
[6] V. Uskoković and V. M. Wu, Calcium phosphate as a key material for socially responsible tissue engineering, vol. 9, no. 6. 2016. doi: 10.3390/ma9060434.
[7] R. Kumar, S. Chahal, S. Dahiya, N. Dahiya, S. Kumar, and S. Lata, “Experimental and theoretical approach to exploit the corrosion inhibition activity of 3-formyl chromone derivatives on mild steel in 1 m H2SO4,” Corros. Rev., vol. 35, no. 2, pp. 95–110, 2017.
[8] N. U. S. Riyaz, M. Khaled, A. Alshami, and I. A. Hussein, “Machine learning-driven prediction of corrosion inhibitor efficiency: Emerging algorithms, challenges, and future outlooks,” Arab. J. Sci. Eng., pp. 1–16, 2025.
[9] C. I. Ekeocha, I. N. Uzochukwu, I.-I. N. Etim, B. I. Onyeachu, and E. E. Oguzie, “Machine learning models and computational simulation techniques for prediction of anti-corrosion properties of novel benzimidazole derivatives,” Mater. Today Commun., vol. 41, p. 110156, 2024, doi: https://doi.org/10.1016/j.mtcomm.2024.110156.
[10] Z. Rajabi, F. Afshar Taromi, S. Pourmahdian, and H. Eivaz Mohammadloo, “Maximizing Corrosion Resistance of HA+ Ce Coated Mg Implants Using Random Forest and Whale Optimization Algorithm,” Processes, vol. 12, no. 3, p. 490, 2024.
[11] Q. Zhu et al., “Synergistic effect of polypyrrole functionalized graphene oxide and zinc phosphate for enhanced anticorrosion performance of epoxy coatings,” Compos. Part A Appl. Sci. Manuf., vol. 130, p. 105752, 2020.
[12] D. Wang, S. Li, Y. Ying, M. Wang, H. Xiao, and Z. Chen, “Theoretical and experimental studies of structure and inhibition efficiency of imidazoline derivatives,” Corros. Sci., vol. 41, no. 10, pp. 1911–1919, 1999.
[13] R. A. Saunders, Molecular surface area measures of polarity and hydrogen bonding for QSAR. Cardiff University (United Kingdom), 2004.
[14] L. Breiman, “Random forests,” Mach. Learn., vol. 45, no. 1, pp. 5–32, 2001.
[15] G. Landrum, “Rdkit: Open-source cheminformatics software. 2016,” URL http//www. rdkit. org/, https//github. com/rdkit/rdkit, vol. 149, no. 150, p. 650, 2016.
[16] D. P. Mishra, H. K. Gupta, G. Saajith, and R. Bag, “Optimizing heart disease prediction model with GridsearchCV for hyperparameter tuning,” in 2024 1st International Conference on Cognitive, Green and Ubiquitous Computing (IC-CGU), 2024, pp. 1–6.
[17] D. Chicco, M. J. Warrens, and G. Jurman, “The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation,” Peerj Comput. Sci., vol. 7, p. e623, 2021.
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
