Neural network models for simulating adsorptive eviction of metal contaminants from effluent streams using natural materials (NMs)

With the rise in environmental-conscious research, natural materials (NMs) have drawn attention as eco-sustainable solution for removing hazardous pollutants via adsorption. Although adsorption processes are renowned for their simple implementation, the mechanisms involved in the adsorption of toxins can be complex due to the number of variables involved and their nonlinear interaction. Literature unveils numerous modelling procedures to optimize process variables for the successful metal ions adsorption; however, artificial neural networks' (ANN) algorithmic approach has accelerated the adsorption propensity of adsorbents for metals ions in water. This review evaluates the ANN approaches (i.e., feedforward neural networks (FFNNs) and neural networks coupled with global optimizers) to simulate the adsorption of different metal ions ranging from heavy metals to highly toxic contaminants (e.g., Ur, Th, As, Cd, Cr, Co, etc.) on NMs. Further, the relative influence of process parameters (such as contact time, pH, initial metal concentration, and dose of NMs) on adsorption has also been outlined. An outlook for future development in the field is provided.

Automated summary

With the rise in environmental-conscious research, natural materials (NMs) have been considered as an eco-sustainable solution for removing hazardous pollutants via adsorption. Artificial neural networks (ANN) have accelerated the adsorption propensity of adsorbents for metal ions in water, and this review evaluates their approaches in simulating the adsorption of different metal ions on NMs. The relative influence of process parameters on adsorption and future development in the field are also outlined.