Caffeine removal using brewery’s spent grain modified with H3PO4: experimental observation and statistical analysis

This work studied the performance and reusability of the biosorbent obtained from a brewery’s spent grain (BSG), applying experimental observation and statistical analysis to caffeine removal from the real matrix. A study of caffeine removal on a batch scale, through kinetic and isotherm models, applying classical and Bayesian statistic criteria, was performed to choose the most adequate model and estimate the maximum amount adsorbed (qmax) and the equilibrium constant (Ke). The biosorbent regeneration through ultrasound and thermal oxidation and the biosorbent characteristics—including the structural changes after the regeneration cycles—were also evaluated. The best operational conditions for caffeine removal (above 80%) were obtained at natural pH and adsorbent dosage of 20 g L−1, with the qmax ranging from 5.86 to 8.83 mg g−1. Results demonstrate that acid sites disappear after adsorption indicating that caffeine binds to the adsorbent surface through hydrogen bonds (chemisorption occurrence hypothesis). The biosorbent was regenerated eight consecutive times and kept the adsorption performance above 60%. Besides, it was efficient in the alkalinity removal (80%), inorganic carbon (95%), and caffeine (98%), which shows the potential for application in the treatment of aqueous matrices.