Nasrollahpour Sepideh, Tanhadoust Amin, Pulicharla Rama, Brar Satinder Kaur
Department of Civil Engineering, Lassonde School of Engineering, York University, Toronto, ON M3J 1P3, Canada.
Department of Civil Engineering, Isfahan University of Technology (IUT), Isfahan, Iran.
iScience. 2024 Oct 10;27(11):111140. doi: 10.1016/j.isci.2024.111140. eCollection 2024 Nov 15.
Given the persistence and toxicity of long-chain perfluoroalkyl carboxylic acids (PFCAs) and their rising concentrations, there is an urgent need for effective removal strategies. This study investigated the adsorptive removal of PFCAs, specifically perfluorononanoic acid (PFNA) and perfluorodecanoic acid (PFDA), using biochar derived from wood and compost. Factors such as biochar size, weight, and initial PFCA concentrations were analyzed to assess their impact on adsorption efficiency over time. The adsorption of PFDA and PFNA reached 90.13% and 85.8%, respectively, at an initial concentration of 500 μg/L. Advanced machine learning techniques, specifically deep neural networks, were employed to model adsorption behavior, incorporating noise injection to account for data uncertainties and preventing overfitting. Results demonstrated the superior performance of compost-derived biochar due to its higher aromaticity and functional group availability. The longer chain length of PFDA contributed to its higher adsorption efficiency compared to PFNA.
鉴于长链全氟烷基羧酸（PFCA）的持久性和毒性及其浓度不断上升，迫切需要有效的去除策略。本研究调查了使用木材和堆肥衍生的生物炭对PFCA，特别是全氟壬酸（PFNA）和全氟癸酸（PFDA）的吸附去除情况。分析了生物炭尺寸、重量和初始PFCA浓度等因素，以评估它们随时间对吸附效率的影响。在初始浓度为500μg/L时，PFDA和PFNA的吸附率分别达到90.13%和85.8%。采用先进的机器学习技术，特别是深度神经网络，对吸附行为进行建模，并通过注入噪声来考虑数据不确定性并防止过拟合。结果表明，堆肥衍生的生物炭性能更优，因为其具有更高的芳香性和官能团可用性。与PFNA相比，PFDA较长的链长导致其吸附效率更高。
