Modeling and prediction of cadmium ion biosorption efficiency using neural networks with Hypnea valentiae

Heavy metals in water, particularly cadmium, pose significant risks to environmental and public health due to their toxicity. This study investigates the removal of cadmium ions from aqueous solutions using Hypnea valentiae , a naturally abundant and cost-effective biosorbent. Batch experiments demonstrated that optimal cadmium removal occurred at pH 5, a biosorbent dose of 2 g/L, a contact time of 50 min, and a temperature of 303 K. Characterization using scanning electron microscopy, energy-dispersive X-ray spectroscopy, Fourier-transform infrared spectroscopy, and X-ray diffraction confirmed morphological and chemical changes in the biosorbent, indicating successful binding of cadmium ions to surface functional groups. Kinetic analysis revealed that the adsorption process followed a pseudo-second-order model, highlighting chemisorption as the dominant mechanism. Isotherm modeling showed that the Langmuir model best described the adsorption, with a maximum capacity of 141.24 mg/g. Thermodynamic studies indicated that the process was exothermic and spontaneous, with decreasing spontaneity at higher temperatures. Furthermore, an artificial neural network model demonstrated strong predictive accuracy (R 2 = 0.989), closely aligning experimental and predicted outcomes. Regeneration tests showed that the Hypnea valentiae biosorbent retained 63.4 % of its original efficiency after eight cycles, highlighting its durability and potential for reuse. These findings highlight Hypnea valentiae 's potential as an efficient, sustainable biosorbent for cadmium remediation, offering a viable solution for addressing heavy metal pollution in water systems. • Achieved 99.154 % Cd (II) removal, showing Hypnea valentiae as an efficient biosorbent. • Langmuir model showed 141.24 mg/g capacity, suitable for large-scale cadmium removal. • Adsorption was exothermic and spontaneous, proving its thermodynamic favorability. • Neural network predicted adsorption with R 2 = 0.989, aiding process optimization.