Etude comparative des modèles d’isothermes linéaires et non linéaires pour l’adsorption de colorants.

Abstract

Many industries, such as, textile, tanning, printing, food, cosmetics, etc...; use organic compounds as dyes to color their final products. The presence of these species in water, ever at very low quantities, is highly visible and undesirable. The decontamination by the conventional techniques proves in certain cases ineffective. The evaluation of the effectiveness of elimination of the dyes present in the effluents of textile was undertaken. In the present study, adsorption batch studies were carried out using a naturally raw material for removal of Supranol Yelow&Nylosane Red and Malachite Green from the aqueous solution, with an explicit objective of comparison of linear and non-linear regression methods for finding variation in isotherm coefficients and fitness of the models. Five linearized isotherm models including four linearized Langmuir models and Freundlich isotherm model were discussed, and their coefficients were estimated. However, an effective usage of the adsorbent necessitates a deeper understanding of the adsorption characteristics. Most of the study of adsorption characteristics are confined to analysis of mono- and bi- parametric isotherm models (and rarely, linearized multi-parametric isotherm models), due to the difficulties in solving higher parametric models, as well as fairly satisfying results by lower-parametric models. Models of Langmuir and Freundlich are respectively introduced to explain the adsorption.The Langmuir model is more representative. This indicates that the sorption might look to the filing of a monolayer of solute molecules on the surface of sorbents. -Adsorption isotherms modeling showed that the interaction of dye with the surface is localized monolayer adsorption. - In order to determine the best-fit isotherm models, three error analysis methods of Residual Mean Square Error and Chi-square statistic (χ2) and Average percentage error API were used to evaluate the data. Although all the studied isotherm models showed fairly good fit to the experimental data, but Langmuir-1 and 2 isotherms was found to be the best representative for JS RN VM adsorbent. Besides, it was observed that to determine the isotherm parameters non-linear isotherm models were found to be the best representative of adsorption characteristics, than their linearized counter-parts.