Jin Changsheng, Lu Jingjing, Gao Yin, Hu Baowei, Liu Yuxi
School of Life and Environmental Sciences, Shaoxing University, Huancheng West Road 508, Shaoxing, 312000, PR China.
School of Business, Shaoxing University, Huancheng West Road 508, Shaoxing, 312000, PR China.
Sci Rep. 2025 Jun 4;15(1):19502. doi: 10.1038/s41598-025-04527-x.
Cadmium (Cd) and antimony (Sb) coexistence in industrial effluents poses significant threats to environmental safety and human health. Consequently, developing effective methods for the simultaneous removal of Cd(II) and Sb(V) from aqueous solutions is critically important. In this study, the adsorption performance of a birnessite (BS) and fulvic acid (FA) composite (BS-FA) for the simultaneous removal of Cd(II) and Sb(V) was optimized using response surface methodology (RSM) in combination with machine learning (ML) techniques, including the genetic algorithm-back propagation neural network (GABP) and random forest (RF) models. The RF model demonstrated superior predictive accuracy (R² = 0.8037, RMSE = 0.0625) compared to the RSM and GABP models. Under the optimized conditions (pH = 6, adsorbent dosage = 0.87 g L, adsorption time = 4 h, ionic strength = 0.01 mol L⁻¹, initial concentration = 25.5 mg L⁻¹), the removal efficiencies of Cd(II) and Sb(V) were 96.9% and 70.2%, respectively. Microscopic and mechanistic analyses revealed that Cd(II) and Sb(V) interacted with the Mn-O bonds in BS and the oxygen-containing functional groups (C-OH and -COOH) in FA, forming stable complexes within the Cd-Sb coexistence system. This study successfully integrates ML models and RSM to optimize and predict the adsorption process, offering valuable insights for mitigating the environmental and health risks associated with Cd and Sb contamination in water treatment.
工业废水中镉（Cd）和锑（Sb）的共存对环境安全和人类健康构成重大威胁。因此，开发从水溶液中同时去除Cd(II)和Sb(V)的有效方法至关重要。在本研究中，利用响应面法（RSM）结合机器学习（ML）技术，包括遗传算法-反向传播神经网络（GABP）和随机森林（RF）模型，优化了水钠锰矿（BS）和富里酸（FA）复合材料（BS-FA）对Cd(II)和Sb(V)的同时去除吸附性能。与RSM和GABP模型相比，RF模型表现出更高的预测准确性（R² = 0.8037，RMSE = 0.0625）。在优化条件下（pH = 6，吸附剂用量 = 0.87 g L，吸附时间 = 4 h，离子强度 = 0.01 mol L⁻¹，初始浓度 = 25.5 mg L⁻¹），Cd(II)和Sb(V)的去除效率分别为96.9%和70.2%。微观和机理分析表明，Cd(II)和Sb(V)与BS中的Mn-O键以及FA中的含氧官能团（C-OH和-COOH）相互作用，在Cd-Sb共存体系中形成稳定的络合物。本研究成功整合了ML模型和RSM，以优化和预测吸附过程，为减轻水处理中与Cd和Sb污染相关的环境和健康风险提供了有价值的见解。
