Acid-modified biochar, a material known for its high specific surface area (SSA) and rich pore structure, holds promise in applications such as heavy metal remediation, soil amendments, and catalyst carriers. Key properties like SSA, average pore size (APS), and total pore volume (TPV) determine its adsorption capacity, reactivity, and water holding capacityWater holding capacity is the amount of water that soil can retain. Biochar can significantly increase the water holding capacity of soil, improving its ability to withstand drought conditions and support plant growth.   More. Optimizing these properties is crucial but complicated due to the interactions among various preparation conditions.

In a recent study published in Science of The Total Environment, researchers Fangzhou Zhao, Lingyi Tang, Wenjing Song, Hanfeng Jiang, Yiping Liu, and Haoming Chen employed machine learning and bibliometric analysis to address this challenge. They developed four machine learning models to predict the SSA, APS, and TPV of acid-modified biochar, utilizing a comprehensive dataset. Among these models, the extreme gradient boosting (XGB) model showed the best performance, with high accuracy in predicting SSA (R²=0.92), APS (R²=0.87), and TPV (R²=0.96).

The study found that modification conditions primarily influenced SSA and TPV, while pyrolysis conditionsThe conditions under which pyrolysis takes place, such as temperature, heating rate, and residence time, can significantly affect the properties of the biochar produced.   More were critical for APS. Using the XGB model, researchers optimized the modification conditions to achieve ideal preparation parameters, producing biochar with an SSA of 727.02 m²/g, APS of 5.34 nm, and TPV of 0.68 cm³/g. The model’s predictions were verified through experimental results, demonstrating a strong generalization ability (R²=0.99, RMSE=12.355).

This research provides valuable insights for refining biochar preparation strategies, enhancing its industrial applicability.