Machine learning – Driven surface grafting of thin-film composite reverse osmosis (TFC-RO) membrane

Abstract

Modifying reverse-osmosis (RO) membrane performance is challenging and time-consuming due to the complex interplay of various factors that influence the membrane's performance. To address this challenge, we have explored the potential of using machine-learning (ML) to graft the polyamide (PA) surface of an RO membrane to increase water permeability and overcome the limitations of the permeability/selectivity tradeoff. We identified moieties with positive and negative contributions toward water permeability by applying Shapley-Additive-exPlanations (SHAP) analysis to our model as an explainable artificial intelligence (XAI) method. We attempted to improve the subunits of the PA's structure with positive Shapley values and graft the polyamide RO membrane layer of a commercial membrane, Dupont XLE, resulting in a substantial increase in water permeability. The membranes were characterized using FTIR, EDS, and SEM analysis, and their performance was evaluated for water permeance and NaCl rejection using a dead-end stirred cell. The modified membrane exhibited a significant improvement in the commercial membrane's water flux, increasing from 2.45 LMH to 4.9 LMH. Our results demonstrate the potential of using ML to replace traditional trial-and-error methods for modifying PA-RO membrane polyamide layers and advancing the development of higher efficient and sustainable RO membranes for water treatment and purification applications.