Long-chain perfluoroalkyl carboxylic acids removal by biochar: Experimental study and uncertainty based data-driven...

Area of Science:

• Environmental Chemistry
• Materials Science

Background:

• Long-chain perfluoroalkyl carboxylic acids (PFCAs) are persistent, toxic, and increasing in environmental concentrations.
• Effective removal strategies for PFCAs are urgently needed to mitigate environmental and health risks.

Purpose of the Study:

• To investigate the adsorptive removal of perfluorononanoic acid (PFNA) and perfluorodecanoic acid (PFDA) using wood and compost-derived biochar.
• To analyze the influence of biochar properties and initial contaminant concentrations on adsorption efficiency.
• To model PFCA adsorption behavior using deep neural networks.

Main Methods:

• Adsorption experiments using wood and compost biochar with varying PFCA concentrations.
• Analysis of biochar size, weight, and PFCA concentration impacts on removal efficiency.
• Application of deep neural networks with noise injection for adsorption modeling.

Main Results:

• Compost-derived biochar exhibited superior PFCA adsorption compared to wood biochar due to higher aromaticity and functional groups.
• High adsorption efficiencies were achieved: 90.13% for PFDA and 85.8% for PFNA at 500 μg/L.
• PFDA showed higher adsorption than PFNA, attributed to its longer perfluorinated chain length.
• Machine learning models accurately predicted adsorption behavior, accounting for data uncertainties.

Conclusions:

• Biochar, particularly from compost, is a promising adsorbent for removing persistent PFCAs from contaminated water.
• Biochar properties and contaminant characteristics significantly influence adsorption efficacy.
• Advanced modeling techniques can effectively predict and optimize adsorption processes for environmental remediation.