Rapid and high-capacity ultrasonic assisted adsorption of ternary toxic anionic dyes onto MOF-5-activated carbon: Artificial neural networks, partial least squares, desirability function and isotherm and kinetic study

- Ultrasonically assisted hydrothermal synthesis of activated carbon-MOF-5 hybrid.
- Rapid simultaneous ultrasonic assisted adsorption of dyes.
- Ultrasound assisted decreases the adsorption time.
- Predict the dyes removal behavior by partial least squares and artificial neural network.

The present paper focused on the ultrasonic assisted simultaneous removal of fast green (FG), eosin Y (EY) and quinine yellow (QY) from aqueous media following using MOF-5 as a metal organic framework and activated carbon hybrid (AC-MOF-5). The structure and morphology of AC-MOF-5 was identified by SEM, FTIR and XRD analysis. The interactive and main effects of variables such as pH, initial dyes concentration (mg L−1), adsorbent dosage (mg) and sonication time (min) on removal percentage were studied by central composite design (CCD), subsequent desirability function (DF) permit to achieved real variable experimental condition. Optimized values were found 7.06, 5.68, 7.59 and 5.04 mg L−1, 0.02 g and 2.55 min for pH, FG, EY and QY concentration, adsorbent dosage and sonication time, respectively. Under this conditions removal percentage were obtained 98.1%, 98.1% and 91.91% for FG, EY and QY, respectively. Two models, namely partial least squares (PLS) and multi-layer artificial neural network (ANN) model were used for building up to construct an empirical model to predict the dyes under study removal behavior. The obtained results show that ANN and PLS model is a powerful tool for prediction of under-study dyes adsorption by AC-MOF-5. The evaluation and estimation of equilibrium data from traditional isotherm models display that the Langmuir model indicated the best fit to the equilibrium data with maximum adsorption capacity of 21.230, 20.242 and 18.621 mg g−1, for FG, EY and QY, respectively, while the adsorption rate efficiently follows the pseudo-second-order model.