Application of polymerized graphene oxide for optimization of o-xylene removal from aqueous solutions through response surface methodology

Introduction
Following an accidental oil spill, the removal of surface oil are usually the important concern for decreasing energy loss and damage to the environment. O-xylene is one of the common compounds which exist in the effluent of oil and petroleum industries. Consequently, the removal of o-xylene from groundwater and surface waters is essential. The one of main techniques that are useful for cleaning up the soluble oil are adsorption. Adsorption methods are simple and economical and the removal of toxic chemicals using porous adsorbents has aroused attention. Graphene as a new synthetic 2D allotrope of carbon has advantages as an adsorbent due to its flexible structure. Graphene has a very fine potential as adsorbent material due to its low cost and environmentally friendly product and by the fact that graphene can be reused without any significant change in its adsorption capacity.

Materials and Methods
Preparation of graphene oxide (GO)
GO was prepared from graphite powder by Hummers and Offeman method.
Polymerization of GO
2 g of GO powder were dissolved in 20 ml of dimethylformamide with stirring. Then, allylamine (10 ml) was added to the solution by continuous stirring at 160 rpm for 2 days. Later, the mixture was washed via dimethylformamide and filtered, the product was dried at room temperature.
The grafting of GO with allylamine (GO-AA) was transferred into the flask and was dissolved in 30 ml of methanol. Next, APS (0.75 g) and MVK (30 ml) were rapidly poured to the solution. The mixture was heated at 60°C in a nitrogen atmosphere for 5 h under stirring. The solution was washed with methanol and filtered. At the end, the product was dried at room temperature.
Modification of polymerized GO with methyl vinyl ketone (GO-MVK)
At first, GO-MVK was mixed with ethanol (150 ml) and aniline (50 ml) and added into the flask. The mixture was fitted under atmosphere of nitrogen at 40°C for 6 h, with agitation speed of 180 rpm. After filtering, the final product was washed with ethanol and further dried in vacuum oven at 40°C for 3 days.
Batch adsorption experiments
For batch adsorption experiment, the desired dosage of graphene oxide grafted with poly methyl vinyl ketone and aniline (GO-MVK-ANI) was mixed with 50 ml of aqueous o-xylene solution (20 mg/l) in a 100 ml flask which was shaken using a rotary shaker at 150 rpm at 25 ± 2ºC. Then, the solution was filtered with syringe filter. The concentration of o-xylene was quantified using GC-FID.
RSM design
RSM is the statistical and mathematical method for designing experiments, building models, and estimating the effects of several factors for desirable responses. This technique is suitable for optimizing the effective parameters with a minimum number of experiments. Herein, the effects of 3 variables in adsorption process including contact time, initial pH and adsorbent dose were investigated with a standard central composite design (CCD). The number of experiments is selected 20 basing on a 23 full factorial CCD for the 3 factors. Five levels of factors were investigated according to the obtained experimental data using MINITAB 16 which are presented in Table 1.








Table 1 Experimental range and levels of independent variables
Variable Factor Unit Range and level
-α -1 0 1 +α
Contact time X1 min 10.91 45 95 145 179.09
Initial pH X2 - 4.31 5 6 7 7.68
Adsorbent dose X3 g/L 0.05 0.8 1.9 3 3.75
Finally, the optimum value of parameters has been earned for the selected target (o-xylene removal efficiency) in optimization process of RSM program.

Results and Discussion
The synthesized product was characterized via Fourier transform-infrared resonance (FTIR) spectroscopy, scanning electron microscopy (SEM), Energy-dispersive X-ray (EDX) spectroscopy and Brunauer-Emmett-Teller (BET) analysis. The total pore volume and average pore diameter of the GO are enlarged by polymerization and modification. The effect of three different parameters including the contact time, initial pH and adsorbent dose on adsorption process of o-xylene using GO-MVK-ANI were investigated. The o-xylene adsorption behavior at different contact time was carried out and shown in Fig. 1a. As can be seen, the maximum o-xylene removal was observed in the initial 10 min of adsorption process. Then, the adsorption capacity became constant, approximately. The effects of pH on adsorption capacity are shown in Fig. 1b. There was no significant change in removal efficiency during the increasing of the pH range. According to Fig. 1c, when adsorbent dose increased, the o-xylene removal efficiency improved. The reasons of this enhancement are that the surface area of the adsorbent, available adsorption sites and also active functional groups are improved with increasing of adsorbent dose. Basing on RSM results, the R2-value is found very high for o-xylene removal (R2 = 99.27%) which is a good confirmation between the experimental and the predicted results. Finally, the optimization of adsorption process was applied for the removal of o-xylene. The experimental checking under optimum conditions (the contact time of 11 min, pH of 4.35 and adsorbent dose of 2.43 g/L) was obtained 74.5% for o-xylene removal efficiency which is close to the model result (75 %).


Fig. 1 Main effects of (a) contact time, (b) initial pH and (c) adsorbent dose on o-xylene removal efficiency using GO-MVK-ANI

Conclusions
The aim of this research was to investigate GO-MVK-ANI to adsorb o-xylene from aqueous solutions. The total pore volume and average pore diameter of GO were improved from (0.016 cm3/g, 4.853 nm) to (0.022 cm3/g, 23.187 nm) by polymerization and modification. Based on the results acquired, adsorption contact time was 11 min. It confirms that this adsorbent has high efficiency in the removal of o-xylene for conditions that need rapid treatment. Moreover, GO-MVK-ANI has high stability in the different ranges of solutions pH. According to the ANOVA results, the model presents high R2-value of 99.27% for o-xylene removal and indicates that the accuracy of the polynomial model was successful. The o-xylene removal efficiency under experimental optimum conditions was obtained 74.5% which confirms close to the RSM results.