A low-cost, scalable aerogel evaporator via machine learning-assisted steel needle templating for solar desalination

Abstract

Solar-driven interfacial evaporators have emerged as a promising, efficient, and sustainable technology for seawater desalination. Key performance indicators for such systems include high evaporation rates, excellent photothermal conversion efficiency, and long-term operational stability in saline environments. Although evaporators fabricated via unidirectional freezing exhibit significant advantages in enhancing evaporation rates, their complex fabrication procedures and high associated costs severely hinder their large-scale practical application in seawater desalination. In this study, we propose a low-cost and scalable steel needle array templating strategy that enables the precise construction of highly ordered and continuous vertically aligned channels within solar evaporators. By employing this method with graphene oxide (GO) and sodium alginate (SA) as building blocks, we fabricate a composite aerogel (DN-GSA) evaporator to significantly enhance water transport and photothermal conversion efficiency. The evaporator with D1.4N8-GSA achieves an evaporation rate of 2.605 kg m⁻² h⁻¹ and a solar-to-vapor conversion efficiency of 92.91% under 1 kW m⁻² irradiation. Benefiting from millimeter-scale vertically aligned channels created by the steel needles, D1.4N8-GSA exhibits excellent salt resistance and maintains stable operation for 12 h in 3.5 wt% NaCl solution without salt crystallization. To validate the reliability of parameter screening and quantify their influence, five machine learning models were trained using 20 experimental samples. Among them, XGBoost exhibited the best predictive performance, yielding an optimal evaporation rate of 2.581 kg m⁻² h⁻¹ with a relative error of less than 1%. Feature importance analysis indicated that the number of needles accounted for 67.63% of the total feature importance. This work integrates systematic experiments with machine learning-assisted optimization, offering a low-cost and scalable strategy for high-performance solar desalination while addressing the challenges of efficiency and practicality in real-world applications.