Application of RSM and ANN for optimization and modeling of biosorption of chromium(VI) using cyanobacterial biomass

TLDR: In this paper, the authors used an Artificial Neural Network-based model to predict abatement of heavy metal ion in wastewaters with high accuracy for various operating conditions with reasonably high accuracy.

Abstract: Proper treatment of heavy metal ions present in wastewaters is a major concern. With extensive usage in various industries, Cr(VI) contamination has become threatening for the environment. Biosorption is a favorable technique for heavy metals removal. In the present study, dried cyanobacterial consortium of Dinophysis caudata and Dinophysis acuminata were used to assess its biosorption capability. The surface texture and morphology of the biosorbent were obtained through scanning electron microscopy. The presence of different chemical bonds, namely hydroxyl, C–H and C–N, was confirmed through FTIR study. Pseudo-second-order Mckay-Ho model was found to perform best to fit the kinetic data. Temkin adsorption isotherm model fit best to the equilibrium data. Response surface methodology (RSM) was employed to optimize Cr(VI) abatement. Effect of initial concentration (IC) of metal ion, temperature, pH variation and amount of adsorbent (AD) were studied during batch study. Maximum Cr(VI) abatement after 5 min contact time was 80.77% for an IC of Cr(VI) of 25 mg/L, at pH 11 and 45 °C with the AD of 2.5 g/L. The optimum removal conditions as shown by RSM study were IC of Cr(VI): 15 mg/L, AD: 1 g/L, pH: 11, and the removal was predicted as 81.72%. Artificial neural network-based model was further developed based on experimental points which indicated that the model can predict abatement of Cr(VI) for various operating conditions with reasonably high accuracy.