Optimisation led energy-efficient arsenite and arsenate adsorption on various materials with machine learning

Area of Science:

• Environmental Science and Engineering
• Materials Science
• Computational Chemistry

Background:

• Arsenic contamination in water is a global environmental and health concern.
• Predicting arsenic adsorption on materials is vital for water remediation but challenging.
• Energy consumption is a critical factor in optimizing adsorption processes.

Purpose of the Study:

• To develop accurate machine learning models for predicting arsenite (As(III)) and arsenate (As(V)) adsorption capacities.
• To identify optimal materials and conditions for energy-efficient arsenic removal from water.
• To create a user-friendly web application for estimating arsenic adsorption.

Main Methods:

• Collected literature data on arsenic adsorption across diverse materials.
• Trained machine learning models (CatBoost, XGBoost, LGBoost) using material properties and reaction parameters.
• Employed genetic optimization to determine maximum adsorption capacities with low energy consumption.

Main Results:

• CatBoost model achieved high accuracy (R²=0.99) for both As(III) and As(V) adsorption prediction.
• Initial arsenic concentrations were key influencing factors for adsorption.
• Identified optimal adsorption capacities of 291.66 mg/g for As(III) and 271.56 mg/g for As(V) using specific composite materials.

Conclusions:

• Machine learning effectively predicts and optimizes energy-efficient arsenic adsorption for water treatment.
• Developed models and a web application facilitate practical design for arsenic removal.
• This approach is crucial for advancing inorganic arsenic treatment in aquatic environments.