Waste materials generated from agriculture and fruit-processing industries can provide cost-effective and cost-efficient materials for the removal of contaminants from aqueous medium. This study deals with adsorption of copper ions from aqueous phase by employing orthophosphoric acid–modified biochar derived from coconut (Cocos nucifera) husk. Biochar characteristics demonstrated increased surface area (24 times) and a porous structure with functional groups such as –OH, –N-H, –CH₂, C=O, and –C-N. These were responsible for active adsorption sites. Key process parameters were optimized and maximum removal efficiency was obtained at dose (0.4 g/L), time (60 min), pH (6), and initial concentration (10 mg/L). To predict the adsorptive removal, artificial neural network (ANN) modeling was performed. The mean absolute error (MAE), root mean squared error (RMSE), and coefficient of determination (R²) provided by ANN model at optimized conditions were found to be 2.63, 4.60, and 0.91, respectively. The behavior of Cu(II) adsorption was closely predicted based on a feed-forward ANN (back propagation) learning algorithm with 4–2–1 topological arrangement. The equilibrium data suggested Langmuir isotherm with a maximum monolayer adsorption capacity of 175.44 mg/g and R² = 0.990 to be the best-suited model. Coconut husk, being an easily available agro-waste, can therefore be used efficiently as a low-cost precursor for the development of an economical adsorbent for inorganic and organic pollutants.

农业和水果加工业产生的废料可以提供具有成本效益和成本效益的材料，用于从水性介质中去除污染物。这项研究通过使用正磷酸改性的椰子壳（椰子壳）衍生的生物炭来处理水相中的铜离子吸附。生物炭特性显示出增加的表面积（24倍）和带有–OH，–NH，–CH ₂等官能团的多孔结构，C = O和–CN。这些负责活性吸附位点。优化了关键工艺参数，并在剂量（0.4 g / L），时间（60 min），pH（6）和初始浓度（10 mg / L）下获得了最大去除效率。为了预测吸附去除，进行了人工神经网络（ANN）建模。在优化条件下，由ANN模型提供的平均绝对误差（MAE），均方根误差（RMSE）和确定系数（R ²）分别为2.63、4.60和0.91。根据前馈ANN（反向传播）学习算法（具有4–2–1拓扑排列），可以精确预测Cu（II）的吸附行为。平衡数据表明Langmuir等温线的最大单层吸附容量为175.44 mg / g，R²  = 0.990是最合适的模型。椰子壳是一种易于获得的农业废料，因此可以有效地用作开发无机和有机污染物的经济吸附剂的低成本前体。