Optimization and mechanism of Acid Orange 7 removal by powdered activated carbon coupled with persulfate by response surface method

Azo dyes are widely utilized in the textile industry, which play an important role in industry. A great deal of dyeing wastewater is generated by the textile industry, and about 10% of plants discharge into water directly substances that may have carcinogenic or mutagenic effects on human beings (Bakheet et al. 2013; Cai et al. 2015). Acid Orange 7 (AO7), a typical azo dye which account for about 50% of commercial dyes, is widely utilized in the textile industry (Azam & Hamid 2006; Yusuf et al. 2017). Moreover, AO7 has the characteristics of high stability, toxic, and being potentially carcinogenic that make it difficult to dispose of (Brillas & Martinez-Huitle 2015). Thus, the emission of wastewater that contains AO7 will contaminate water and affect the survival of animals continuously without effective treatment. Unfortunately, activated sludge processes and physical adsorption methods (Cai et al. 2014) are ineffective to remove AO7 so that it is necessary to search for an effective and economical method to remove such contaminants.

The experiments of AO7 decolorization were carried out in 250 mL Erlenmeyer flasks with magnetic stirring at 300 rpm. The initial pH was adjusted with H₂SO₄ and KOH firstly, then PDS was added into the reactor, which contained AO7 before the reaction, and a certain amount of PAC was quickly added into the reactor. After a fixed interval of reaction of 20 min, 4 mL of the solution was quickly sampled, and filtered with a 0.45 μm membrane for analysis. The UV-vis absorbance of the samples was analyzed at λ = 484 nm by a 721E Spectrophotometer. UV-vis spectra of AO7 were conducted from 200 to 700 nm using a T6 UV-visible spectrophotometer. The TOC of the samples was quantified by Liqui TDC II analyzer. The morphologies of PAC were examined by scanning electron microscopy (SEM, Tescan Inc., USA). The surface groups of PAC were detected by Boehm titration and Fourier transform infrared spectra (Nicolet iS5 FT-IR spectrometer with KBr Pellet). The key experiments were repeated three times.

Central composite design (CCD) and Box-Behnken design are applied in response surface analysis. In this experiment, a CCD statistical model was applied to optimize these operational factors, which included PDS, PAC dosages, and initial pH, and also studied the combined interaction among them. The choice of operational parameters evaluated by response surface methodology (RSM) is based on our preliminary experiments. Each parameter has three levels designed as +1, 0, −1 for high, middle, and low values, respectively. The model consists of a complete factorial design with three replications at the design center, leading to a total of 17 experiments.

In this work, the decolorization of AO7 was investigated, which considered the adsorption of PAC, the oxidation of PDS, and the oxidation in a PAC/PDS system, which is displayed in Figure 1(a). The results showed that PDS alone had a slight effect on AO7 decolorization and only 8% was oxidized in 120 min, whereas a significant decolorization was observed in the PAC adsorption system, in which 55% was faded in 120 min, which was better than that of other researchers (Zhang et al. 2013; Chen et al. 2016a, 2016b; Huang et al. 2017). Moreover, a removal rate of 98% was observed in the PAC/PDS system, which was clearly better than PDS only and PAC only. Besides, the decolorization of AO7 in these three systems followed first-order kinetics that are presented in Figure 1(b). And the reaction rate constant (k value) in these systems was 6.7 × 10⁻⁴, 0.5 × 10⁻² and 3.1 × 10⁻² min⁻¹, respectively. It could be seen that the k value in the PAC/PDS system was higher than PAC only and PDS only. Therefore, the result obviously indicated the remarkable synergistic effect of PAC and PDS. Moreover, by analyzing the removal of AO7 in PAC and PAC/PDS processes at different temperatures as displayed in Figure S1 and Table S1 (available with the online version of this paper), we found that the adsorption is a spontaneous process and adsorption capacity would decrease with increasing temperature from 288 to 333 K (Abukhadra et al. 2018; Shaban et al. 2018). Besides, the k value increased with the increase of temperature indicating that the reaction is an endothermic reaction in the PAC/PDS system (Abukhadra et al. 2018; Shaban et al. 2018). And the activation energy of the PAC/PDS system was calculated to be 26.68 kJ/mol, which is significantly lower than 81.15 kJ/mol in the base/PMS/AO7 system (Qi et al. 2016) indicating that PAC could be an excellent catalyst to activate PDS to decolorize AO7.

The effect of different PAC dosages on AO7 decolorization in the PAC/PDS system was evaluated and is displayed in Figure 2(b). It illustrated that with the increase of PAC dosages, the AO7 decolorization rate was observed to be faster in the PAC/PDS system. Compared with the AO7 decolorization rate at 0.05 g/L PAC dosages, there was an accelerating effect when PAC dosages increased to 0.3 g/L. The reason might be that higher PAC dosages increased the BET, which increased the amount of active sites for adsorption and the catalytic reaction. The reaction speed had slowed down markedly when PAC dosages further increased from 0.3 g/L (k value = 6.03 × 10⁻² min⁻¹) to 0.4 g/L (k value = 6.10 × 10⁻² min⁻¹). This was probably because the reaction rate was limited by the reaction rate of the radical.

The effect of three significant parameters on the decolorization of AO7 was optimized using RSM as given in Table S2 (available online). Furthermore, their responses were predicted and the optimum conditions for decolorization were also determined. The results of the analysis of variance (ANOVA) of the response surface quadratic analyses and model terms are shown in Table 1. From the ANOVA results, the model F value 140.17 indicated that the linear model with the curvature term was statistically significant. And the values of ‘lack of fit’ (F value) were less than 0.05, which implied that the lack of fit was not significant relative to the pure error. Moreover, the coefficient of determination (R²) of the model was 0.9642, and the adjusted R² was 0.9838. Both of the values are very near to 1.0, indicating a high correlation between the actual and the predicted values as shown in Figure S3 (available online) (Lim et al. 2013; Selvakumar et al. 2013). Thus, it was a good predictor of response and could be used to predict the optimal value. By regression analysis, a quadratic model was predicted as in the following equation: The removal of AO7=20.26681-2.23275 × pH + 47.33064 × PDS + 490.69493 × PAC + 13.90476 × pH × PAC-103.52941 × PDS × PAC-9.02522 × PDS²-1005.27954 × PAC².

By using the quadratic model, a removal of 100% was predicted, the optimum condition for the decolorization rate of AO7 was found to be as follows: PAC (0.19 g/L), PDS (1.64 g/L), and initial pH (4.14). A confirmatory experiment was performed three times in order to test the suitability of the model and a removal of 100% was obtained that illustrated the reliability of the model. Moreover, the interactions of the three parameters were also analyzed by RSM. Figure 4 indicates the Contour plot of percent removal of AO7 as a function of two parameters. The elliptical shape of the curve indicated good interaction between the two variables and the circular shape indicates no interaction (Oladipo & Gazi 2014). From Figure 3, a relatively significant interaction between PDS and PAC was observed. It showed that the decolorization rate of AO7 increased slightly with the increase of PDS and PAC concentrations. A relative interaction between initial pH and PAC was also observed and there was also a relative interaction.

The effect of SO₄²⁻ on AO7 decolorization in the PAC/PDS system was researched as displayed in Figure 4(b). It illustrated that with the increasing dosages of SO₄²⁻, AO7 decolorization accelerated gradually. Moreover, when higher dosages of SO₄²⁻ were added, the k values gradually increased from 2.72 × 10⁻² to 4.20 × 10⁻² min⁻¹. It could be seen that the effect of SO₄²⁻ on AO7 decolorization in the activated PDS process was consistent with that previously reported (Neppolian et al. 2002; Rao et al. 2014). On the one hand, a higher SO₄²⁻ dosage could prevent the quenching and transformation of SO₄^−• as in Equation (6) and (7). On the other hand, it was reported previously that a higher dosage of SO₄²⁻ could increase the ion strength and induce the dye dimerization in the solution (Chen et al. 2016a, 2016b). The aggregation of dye molecules would increase the extent of adsorption of AO7 on the PAC surface. Thus, the addition of SO₄²⁻ could raise the adsorption of AO7 on the PAC surface for AO7 decolorization induced by the radical species activated by the oxygen functional groups onto PAC.

In order to figure out the change of structural characteristics of AO7 during the decolorization process in the PAC/PDS system, samples were taken every 20 minutes and monitored in the spectrum of AO7. AO7 consists of an azo bond, hydrazone form, naphthalene ring, and a benzene ring basically, which accounted for the four apparent peaks at 484, 430, 310, and 229 nm, respectively. The maximum absorption wavelength for AO7 was determined to be 484 nm, which accounts for the orange color of the solution and can be attributed to the azo bond (Bauer et al. 2001). The decolorization of AO7 over 120 min is shown in Figure 5(a). Moreover, the UV-vis spectrum of PDS in the distilled water was also detected. It was apparent that characteristic peaks at 484 nm and 310 nm declined and the removal efficiency of these peaks was about 90% and 30% in 120 min, respectively. While the 229 nm absorption band was still present, a high position proving the highly refractory nature of the benzene ring. Thus, it could be speculated that AO7 might be partly mineralized in PAC/PDS system. In order to demonstrate whether AO7 was mineralized or not, the TOC variation was monitored as shown in Figure 5(b). TOC gradually decreased during AO7 decolorization and the removal efficiency at 60 min and 120 min reaches 44% and 50%, respectively. Above all, it proved that PDS activated by PAC could break up the AO7 molecular structure and mineralize AO7 effectively.

Therefore, both SO₄^−• and •OH are possibly responsible for the decolorization of AO7 in the PAC/PDS system. In order to identify the dominating oxidizing species working in the experiment, methanol, tertbutanol, and phenol, which were commonly used as free radical inhibitors, were utilized. Methanol was highly reactive with both SO₄^−• and •OH with k values of 9.7 × 10⁸ and 3.2 × 10⁶ M⁻¹s⁻¹, respectively. Whereas tertbutanol was also an effective quenching agent for •OH while reacting much more slowly with SO₄^−• such that the k values were (3.8–7.6) × 10⁸ and (4–9.1) × 10⁵ M⁻¹s⁻¹, respectively. Besides, phenol was another strong quencher radical (k_•SO4⁻_/phenol = 8.8 × 10⁹ M⁻¹s⁻¹, k_•OH/phenol = 6.6 × 10⁹ M⁻¹s⁻¹) that showed a strong inhibiting effect on SO₄^−• and •OH (Ziajka & Pasiuk-Bronikowska 2005). As shown in Figure 6, when methanol or tertbutanol were added into the PAC/PDS system, there was only a slight inhibitory effect on AO7 decolorization. When phenol was added into the system, the reaction of AO7 decolorization slowed down and the inhibiting effect was remarkable. Methanol and tertbutanol were relatively hydrophilic that could compete with SO₄^−• and •OH radicals taking place in the aqueous solution even though they couldn't accumulate on the PAC surface to a significant extent. While phenol is more hydrophobic than methanol and tertbutanol, thus, it is easier to approach carbon surface to prevent the activation of PDS by oxygen surface function groups which were distributed in the carbon material surface (Chen et al. 2016a, 2016b). Therefore, the areas of free radicals-induced degradation of AO7 were most likely taking place on the surface of the PAC. Moreover, SO₄^−• was mainly responsible for the AO7 degradation in PAC/PDS system.

A reuse experiment was conducted to detect the recovery performance of PAC four times, as displayed in Figure S7 (available online). The decolorization efficiencies of AO7 removal for four reuse cycles gradually decreased and were 97%, 90%, 79%, 53%, respectively. Also, the morphologies and functional groups of PAC before and after reaction were inspected by SEM, FT-IR, and Boehm titration. As displayed in Figure S8 and Table S3 (available online), the SEM images of virgin PAC appear to be rough and dusty, whereas the oxidation of PDS resulted in a relatively neat surface on the PAC. Moreover, after the reaction, the decrease of the peak at 3,420 cm⁻¹ and the increase of the asymmetrical band at 1,110 cm⁻¹ revealed the increase of acid groups and the C-O-C band, which resulted from the oxidation of the C = O band by PDS (Li et al. 2017) It was consistent with the results of change of the functional groups on the PAC surface by Boehm titration. The results were similar to previous studies reporting that the removal rate gradually decreased with the increase of reuse cycles of AC owing to the decrease in surface area of AC and the consumption of oxygen-containing alkaline functional groups (Li et al. 2017), which indicated that AC may not actually be a catalyst for electron transfer media, but a PS-activated initiator (Liu et al. 2018). The deactivation of PAC in the PAC/PDS system could be due to the exhaustion of electron-donating residues on the PAC, which can trigger PDS into SO₄^−•. Besides, the incomplete removal of AO7 and intermediate products adsorbed on the surface of PAC inhibited the interaction between PDS and PAC. But when PAC was reused three times the removal rate of AO7 was obviously decreased, which was different from other researches. This was mainly because of the difficulty of recovery and serious problems with loss of PAC, which was also a problem that restricted the development of PAC in the field of water treatment.

The present work demonstrates that PAC can not only remove AO7 through adsorption but also catalyzes PDS, forming SO₄^−• and ·OH to remove AO7. The decolorization processes follow first-order kinetics. Moreover, univariate analysis showed that the decolorization efficiency increased as the PDS and PAC dosages increased. The decolorization rate of AO7 is highest in acidic conditions and the maximum removal is found at pH 3. The results of RSM show PDS, PAC dosages, and initial pH have remarkable interaction. The optimum condition through RSM is found to be as follows: PAC (0.19 g/L), PDS (1.64 g/L), and initial pH (4.14). Addition of Cl⁻ and SO₄²⁻ accelerated AO7 decolorization while the presence of HCO₃⁻ slightly retarded AO7 decolorization. A maximum of 100% color removal and 50% TOC reduction confirm the mineralization of AO7 in PAC/PDS system. The efficiency of AO7 removal in four reuse cycles, which gradually decreased, indicates that PAC was partially deactivated in the PAC/PDS system. Nevertheless, this experiment is just under optimal pH but did not consider AO7 removal efficiency under the condition of normal pH and raw water from a textile factory. In addition, the specific degradation process of AO7 also requires further study in the PAC/PDS system. Moreover, this system is expected to be used in the field of printing and dyeing wastewater treatment and emergency water treatment.

This work was financially supported by the Fundamental Research Funds for the Central Universities of China (No. 2018CDYJSY0055), the National Natural Science Foundation of China (No. 51308563), Graduate Research and Innovation Foundation of Chongqing of China (No. CYS18030), and the Frontier Interdisciplinary Training Project of Fundamental Research Funds for the Central Universities of China (No. 2018CDQYCH0053).