Applications of Box-Behnken experimental design coupled with artificial neural networks for biosorption of low concentrations of cadmium using Spirulina (Arthrospira) spp.

Abstract

The present study deals with the application of artificial intelligence techniques coupled with Box-Behnken (BB) design to model the process parameters for biosorption of cadmium using live Spirulina (Arthrospira) spp. as adsorbent in open race way pond with Zarrouk medium. The biomass concentration of Spirulina spp. decreased to half at 4 ppm Cd (II) after 8 days. Based on the LCt50 values, 3.69 ppm (8th day), Spirulina (Arthospira) maxima showed maximum tolerance. Considerable growth and bioaccumulation of Spirulina spp. is observed below 1 ppm and tolerant up to 3 ppm. The cadmium adsorption on Spirulina spp. showed good correlation (R2 = 0.99) when applied to Freundlich equation and data fit into pseudo second order kinetics. A four factorial, three blocks and three level Box-Behnken design with initial concentration (1 ppb to 5 ppb), biosorbant dosage (0.1 gdw to 0.2 gdw), agitation speed (12 rpm to 16 rpm) and pH (6 to 8) as independent variables and percentage adsorption as dependent variable were selected for study. The data were further processed using artificial neural network model and DIRECT algorithm for better optimization. The final Cd (II) concentration of 0.5 ppb was achieved with 1 ppb initialconcentration under optimal conditions. A continuous desorption process was also developed for removal of cadmium from Spirulina (Arthrospira) sp.