Surfactant-Activated pharmaceutical waste biomass for efficient removal of Basic Violet 14: Experimental Investigation, Machine-Learning Optimization, and mechanistic validation by DFT calculations

Abstract

Dyes are extensively employed in the pharmaceutical industry and laboratories and their persistence in wastewaters presents severe environmental and health hazards due to toxicity and resistance to biodegradation. This study examines an efficient method of dye removal by valorization of pharmaceutical waste, Streptomyces rimosus (SR) biomass, raw and impregnated with SDS. The biosorbents were evaluated for removal of Fuchsin dye (Basic Violet 14, BV14) under varying pH, biosorbent dose, contact time, initial dye concentration, and temperature. SDS-SR showed ∼ 98% removal at 2 g/L compared to 4 g/L for raw SR to have similar efficiency. Kinetic data were described by the pseudo-second order model and equilibrium data was described by Hill and Sips isotherms suggesting the presence of heterogeneous surfaces, cooperative adsorption and multilayer formation. Thermodynamic analysis confirmed a spontaneous (ΔG°<0), exothermic (ΔH°<0) physisorption-driven process. DFT calculation results showed that the SDS modification enhanced the BV14 adsorption energy from −1.42 to −2.87 eV, leading to an increase of the electrostatic and hydrophobic interactions which are consistent with experimental observations. Additionally, machine learning models (ANN, linear regression, decision tree, random forest) were trained on experimental data where ANN model resulted in the highest predictive accuracy. Hybrid optimization using ANN couple with GA and PSO was used to find optimal operational conditions for maximum adsorption. This study proves that SDS-modified SR is an effective and eco-friendly biosorbent as it combines waste valorization and molecular-level understanding using DFT with predictive AI modeling. The approach offers a sustainable, circular-economy strategy for pharmaceutical and laboratory wastewater treatment, combining high removal efficiency, mechanistic insight, and data-driven optimization.

Keywords Biosorbent valorization; Surfactant-modified biomass; Dye adsorption; Hybrid ANN optimization; DFT mechanistic validation