It is an effective removal method to adsorb heavy metals by adsorption materials, so as to realize their separation from environmental media. Optimizing the structure of the materials and adsorption conditions are essential for improving the adsorption performance. Amino-functionalized magnetic humic acid nanoparticles (AF-MHA) was synthesized using a co-precipitation method. And the adsorbing material showed remarkable lead (Pb(II)) adsorption capacity under neutral pH conditions and a maximum adsorption capacity of 119 mg/g. The adsorption process was attributed to complexation with functional groups like amine, carboxyl, and phenol hydroxyl. In order to enhance the optimization of adsorption parameters, machine learning (ML) models including Artificial Neural Network (ANN), Random Forest (RF), Support Vector Regression (SVR), and CatBoost were employed. After comparative study we find the CatBoost model was found to be the most accurate predictor with a correlation coefficient of R2=0.95. ML also facilitated the identification of key factors influencing adsorption capability by assessing the importance of input features. With ML-assisted optimization parameters offering a strategic advantage in optimizing adsorption conditions and enhancing performance, AF-MHA is a promising adsorbent for treating divalent metal-contaminated water.
Read full abstract