Deep learning-aided prediction and mechanistic analysis of reaction kinetics in biochar-catalyzed antibiotic degradation

Abstract

The proliferation of antibiotics in the environment poses a severe threat to public health and ecosystems. While biochar-based catalytic systems offer a promising remediation strategy, their design is complicated by nonlinear interactions between material properties and reaction conditions. To overcome this, we present an interpretable machine learning framework for the accurate prediction of reaction rate constants (k) for antibiotic degradation. A comprehensive dataset was compiled from the literature, encompassing 16 features across three domains including biochar catalyst properties, elemental composition, and reaction conditions. Among the six machine learning (ML) algorithms evaluated, the transformer-based deep learning model TabPFN (Tabular Prior-data Fitted Network) demonstrated superior performance (Test R² = 0.91, RMSE = 0.021), outperforming tree-based, kernel-based, and neural network models. Model interpretability analyses revealed that catalyst properties contributed the largest share of predictive capability (59.3%), followed by reaction conditions (25.9%) and elemental composition (14.8%). Persistent free radicals (formed at 450–550 °C) and high total pore volume (> 0.23 cm³ g⁻¹) were identified as key drivers of reaction kinetics, along with optimal oxidant (0.5–5.5 mg L⁻¹) and pollutant concentrations (< 22 mg L⁻¹). These insights were embedded into a user-friendly web-based GUI, enabling rapid k prediction for new biochar catalysts with prediction errors below 20% on external validation samples. This work provides both a robust predictive tool and a generalizable, data-driven methodology for understanding and optimizing complex environmental catalytic processes.

Graphical Abstract

Keywords

Machine learning / TabPFN / Biochar / Antibiotics / Graphical user interface

References

[1] Balaban NQ, Helaine S, Lewis K, et al. . Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol, 2019, 17: 441-448

[2] Breiman L. Random Forests. Mach Learn, 2001, 45: 5-32

[3] Brillas E. A critical review on ibuprofen removal from synthetic waters, natural waters, and real wastewaters by advanced oxidation processes. Chemosphere, 2022, 286 131849

[4] Chicco D, Warrens MJ, Jurman G. The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput Sci, 2021, 7 e623

[5] Fang G, Gao J, Liu C, et al. . Key role of persistent free radicals in hydrogen peroxide activation by biochar: implications to organic contaminant degradation. Environ Sci Technol, 2014, 48: 1902-1910

[6] Hawryluk-Sidoruk M, Raczkiewicz M, Krasucka P, et al. . Effect of biochar chemical modification (acid, base and hydrogen peroxide) on contaminants content depending on feedstock and pyrolysis conditions. Chem Eng J, 2024, 481 148329

[7] Hollmann N, Müller S, Purucker L, et al. . Accurate predictions on small data with a tabular foundation model. Nature, 2025, 637: 319-326

[8] Hron K, Templ M, Filzmoser P. Imputation of missing values for compositional data using classical and robust methods. Comput Stat Data Anal, 2010, 54: 3095-3107

[9] Hu B, Ai Y, Jin J, et al. . Efficient elimination of organic and inorganic pollutants by biochar and biochar-based materials. Biochar, 2020, 2: 47-64

[10] Huang D, Luo H, Zhang C, et al. . Nonnegligible role of biomass types and its compositions on the formation of persistent free radicals in biochar: Insight into the influences on Fenton-like process. Chem Eng J, 2019, 361: 353-363

[11] Jaffari ZH, Jeong H, Shin J, et al. . Machine-learning-based prediction and optimization of emerging contaminants’ adsorption capacity on biochar materials. Chem Eng J, 2023, 466 143073

[12] Jia H, Zhao S, Shi Y, et al. . Formation of environmentally persistent free radicals during the transformation of anthracene in different soils: roles of soil characteristics and ambient conditions. J Hazard Mater, 2019, 362: 214-223

[13] Jiang S, Hou Y, Man Z, et al. . Guiding experiment with machine learning: a case study of biochar adsorption of Ciprofloxacin. Sep Purif Technol, 2024, 334 126023

[14] Jordan MI, Mitchell TM. Machine learning: trends, perspectives, and prospects. Science, 2015, 349: 255-260

[15] Klüpfel L, Keiluweit M, Kleber M, Sander M. Redox properties of plant biomass-derived black carbon (biochar). Environ Sci Technol, 2014, 48: 5601-5611

[16] Latif J, Chen N, Saleem A, et al. . Machine learning for persistent free radicals in biochar: dual prediction of contents and types using regression and classification models. Carbon Res, 2024, 3 39

[17] Leng L, Xiong Q, Yang L, et al. . An overview on engineering the surface area and porosity of biochar. Sci Total Environ, 2021, 763 144204

[18] Li Y, Li J, Pan Y, et al. . Peroxymonosulfate activation on FeCo2S4 modified g-C3N4 (FeCo2S4-CN): mechanism of singlet oxygen evolution for nonradical efficient degradation of Sulfamethoxazole. Chem Eng J, 2020, 384 123361

[19] Li Z, Sun Y, Yang Y, et al. . Biochar-supported nanoscale zero-valent iron as an efficient catalyst for organic degradation in groundwater. J Hazard Mater, 2020, 383 121240

[20] Li N, Li R, Duan X, et al. . Correlation of active sites to generated reactive species and degradation routes of organics in peroxymonosulfate activation by Co-loaded carbon. Environ Sci Technol, 2021, 55: 16163-16174

[21] Liang J, Xu X, Qamar Zaman W, et al. . Different mechanisms between biochar and activated carbon for the persulfate catalytic degradation of sulfamethoxazole: roles of radicals in solution or solid phase. Chem Eng J, 2019, 375 121908

[22] Liu G, Zhang Y, Yu H, et al. . Acceleration of goethite-catalyzed Fenton-like oxidation of ofloxacin by biochar. J Hazard Mater, 2020, 397 122783

[23] Liu B, Xi F, Zhang H, et al. . Coupling machine learning and theoretical models to compare key properties of biochar in adsorption kinetics rate and maximum adsorption capacity for emerging contaminants. Bioresour Technol, 2024, 402 130776

[24] Lu P, Huang Q, Chi Y, Yan J. Preparation of high catalytic activity biochar from biomass waste for tar conversion. J Anal Appl Pyrolysis, 2017, 127: 47-56

[25] Lundberg SM, Lee S-I (2017) A unified approach to interpreting model predictions. Adv Neural Inf Process Syst. 30. https://arxiv.org/abs/1705.07874

[26] Luo K, Yang Q, Pang Y, et al. . Unveiling the mechanism of biochar-activated hydrogen peroxide on the degradation of ciprofloxacin. Chem Eng J, 2019, 374: 520-530

[27] Luo X, Shen M, Liu J, et al. . Resource utilization of piggery sludge to prepare recyclable magnetic biochar for highly efficient degradation of tetracycline through peroxymonosulfate activation. J Clean Prod, 2021, 294 126372

[28] Muzyka R, Misztal E, Hrabak J, et al. . Various biomass pyrolysis conditions influence the porosity and pore size distribution of biochar. Energy Oxf, 2023, 263 126128

[29] Nathan C, Cars O. Antibiotic resistance—problems, progress, and prospects. N Engl J Med, 2014, 371: 1761-1763

[30] Ohore OE, Addo FG, Zhang S, et al. . Distribution and relationship between antimicrobial resistance genes and heavy metals in surface sediments of Taihu Lake, China. J Environ Sci China, 2019, 77: 323-335

[31] Ouyang D, Chen Y, Yan J, et al. . Activation mechanism of peroxymonosulfate by biochar for catalytic degradation of 1, 4-dioxane: important role of biochar defect structures. Chem Eng J, 2019, 370: 614-624

[32] Padarian J, Minasny B, McBratney A. Using deep learning to predict soil properties from regional spectral data. Geoderma Reg, 2019, 16 e00198

[33] Qin L, Yang L, Liu X, et al. . Formation of environmentally persistent free radicals from thermochemical reactions of catechol. Sci Total Environ, 2021, 772 145313

[34] Ruan X, Sun Y, Du W, et al. . Formation, characteristics, and applications of environmentally persistent free radicals in biochars: a review. Bioresour Technol, 2019, 281: 457-468

[35] Shan H, Wang W, Wang Z, et al. . Unlocking the superior flexibility and enhanced catalytic oxidation performance of CoTiO3 nanofibrous membranes through zirconium doping. Sep Purif Technol, 2024, 348 127714

[36] Sun Z, Sun S, Wang S, et al. . Prediction of uranium adsorption performance by machine learning for sustainable seawater extraction. Sci CHINA Chem, 2025,

[37] Tan J, Chen X, Shang M, et al. . N-doped biochar mediated peroxydisulfate activation for selective degradation of bisphenol A: The key role of potential difference-driven electron transfer mechanism. Chem Eng J, 2023, 468 143476

[38] Wang J, Zhuan R. Degradation of antibiotics by advanced oxidation processes: an overview. Sci Total Environ, 2020, 701 135023

[39] Wang R, Chen H, He Z, et al. . Discovery of an end-to-end pattern for contaminant-oriented advanced oxidation processes catalyzed by biochar with explainable machine learning. Environ Sci Technol, 2024, 58: 16867-16876

[40] Xie J, Latif J, Yang K, et al. . A state-of-art review on the redox activity of persistent free radicals in biochar. Water Res, 2024, 255 121516

[41] Xu L, Wu C, Liu P, et al. . Peroxymonosulfate activation by nitrogen-doped biochar from sawdust for the efficient degradation of organic pollutants. Chem Eng J, 2020, 387 124065

[42] Yang J, Zhang Z, Wang J, et al. . Pyrolysis and hydrothermal carbonization of biowaste: a comparative review on the conversion pathways and potential applications of char product. Sustain Chem Pharm, 2023, 33 101106

[43] Yi X, Lin C, Ong EJL, et al. . Occurrence and distribution of trace levels of antibiotics in surface waters and soils driven by non-point source pollution and anthropogenic pressure. Chemosphere, 2019, 216: 213-223

[44] Yu J, Zhu Z, Zhang H, et al. . Persistent free radicals on N-doped hydrochar for degradation of endocrine disrupting compounds. Chem Eng J, 2020, 398 125538

[45] Yu T, Rajasekar A, Zhang S. A decennial study of the trend of antibiotic studies in China. Environ Sci Pollut Res, 2023, 30: 121338-121353

[46] Zhang M, Bai X, Liu D, et al. . Enhanced catalytic activity of potassium-doped graphitic carbon nitride induced by lower valence position. Appl Catal B Environ, 2015, 164: 77-81

[47] Zhang S, Hu B, Zhang L, Xiong Y. Effects of torrefaction on yield and quality of pyrolysis char and its application on preparation of activated carbon. J Anal Appl Pyrolysis, 2016, 119: 217-223

[48] Zhang Y, Yin M, Sun X, Zhao J. Implication for adsorption and degradation of dyes by humic acid: light driven of environmentally persistent free radicals to activate reactive oxygen species. Bioresour Technol, 2020, 307 123183

[49] Zhang L, Cheng H, Pan D, et al. . One-pot pyrolysis of a typical invasive plant into nitrogen-doped biochars for efficient sorption of phthalate esters from aqueous solution. Chemosphere, 2021, 280 130712

[50] Zhang L, Zhang C, Lian K, Liu C. Effects of chronic exposure of antibiotics on microbial community structure and functions in hyporheic zone sediments. J Hazard Mater, 2021, 416 126141

[51] Zhang W, He Y, Li C, et al. . Persulfate activation using Co/AC particle electrodes and synergistic effects on humic acid degradation. Appl Catal B Environ, 2021, 285 119848

[52] Zhang R, Zhang R, Zimmerman AR, et al. . Applications, impacts, and management of biochar persistent free radicals: a review. Environ Pollut, 2023,

[53] Zhang P, Yang Y, Duan X, Wang S. Oxidative polymerization versus degradation of organic pollutants in heterogeneous catalytic persulfate chemistry. Water Res, 2024, 255 121485

[54] Zhong S, Zhang K, Bagheri M, et al. . Machine learning: new ideas and tools in environmental science and engineering. Environ Sci Technol, 2021, 55: 12741-12754

[55] Zhu X, Li Y, Wang X. Machine learning prediction of biochar yield and carbon contents in biochar based on biomass characteristics and pyrolysis conditions. Bioresour Technol, 2019, 288 121527

[56] Zhu G, Zhu J, Fu X, et al. . Co nanoparticle-embedded N, O-codoped porous carbon nanospheres as an efficient peroxymonosulfate activator: singlet oxygen dominated catalytic degradation of organic pollutants. Phys Chem Chem Phys, 2020, 22: 15340-15353

[57] Zhu X, Wan Z, Tsang DC, et al. . Machine learning for the selection of carbon-based materials for tetracycline and sulfamethoxazole adsorption. Chem Eng J, 2021, 406 126782

[58] Zou J, Yu J, Tang L, et al. . Analysis of reaction pathways and catalytic sites on metal-free porous biochar for persulfate activation process. Chemosphere, 2020, 261 127747

Funding

Qin Chuang Yuan” Innovation and Entrepreneurship Talent Program of Shaanxi Province(QCYRCXM-2023-051)

National Natural Science Foundation of China(Grant No. 42307334)

Key Research and Development Projects of Shaanxi Province(Grant No. 2019ZDLNY01-02-01)

“One Hundred Talents” program of Shaanxi Province (Grant No. SXBR9171)

Shaanxi Science Fund for Distinguished Young Scholars(Grant No. 2019JC-18)

Open Fund of State Key Laboratory of Soil Erosion and Dryland farming on Loess Plateau(Grant No. A314021402-2021012)

China Postdoctoral Science Foundation(Grant No. 2023M732875)

RIGHTS & PERMISSIONS

The Author(s)