Optimized Degradation of Bisphenol A by UV/H2O2 coupled to Microwaves in a Novel Reactional Setup


 In this paper, the UV/H2O2/MW (microwave) process was compared with the UV/H2O2 one, using bisphenol A (BPA) as a model-pollutant. The proposed experimental setup was operated in batch recycle mode and allows for the direct comparison among different processes: UV only, H2O2 only, MW only, UV/H2O2, UV/MW, H2O2/MW, and UV/H2O2/MW, as well as temperature control to minimize thermal effects. The degradation of BPA at near-environmental concentration (100 µg L−1) was optimized by an experimental design approach (Response Surface Methodology) and its residual concentration was measured by HPLC. Approximately 95% of the initial BPA amount could be removed in 30 min at the optimal conditions (CH2O2 = 20 mg L−1, flow rate = 700 mL min−1, and MW power = 245 W). The experiments designed for comparing the UV/H2O2 and the UV/H2O2/MW processes showed that the use of MW doubled the initial pseudo-first-order degradation rate (from 0.046 to 0.10 min−1) and significantly increased the maximum oxidation capacity of the system (from 86 to 100%). Although the reasons behind those results are still unclear, it seems that the existence of non-thermal effects of the MW irradiation should be considered.


INTRODUCTION
The production of plastics has been increasing exponentially since the 1960's, reaching more than 8000 metric tons in 2015. Most of those end up as waste, polluting superficial waters and soil (GEYER et al., 2017). Furthermore, many compounds used in the production of plastics, such as resins and phenols, are hazardous to the biota as a whole. One of those compounds is Bisphenol A (BPA). BPA (4,4'-isopropylidenediphenol; Nº CAS 80-05-7) is largely used in the production of polycarbonate (KRISHNAN et al., 1993). Due to its moderate solubility (0.15 mg mL -1 ) and log D pH=7.4 (4.04), it can be found in surface waters, groundwaters, domestic and industrial sewage, sediments, and organisms (STAPLES et al., 1998;HING-BIU;PEARL, 2000;YAMAMOTO et al., 2001;MOLINA-MOLINA et al., 2019). Its hazardous effects are significant: endocrine disruption, carcinogenic and estrogenic activities, even in human beings (HOWDESHELL et al., 1999;DOUNG et al., 2009;GRINGRICH et al., 2019;BERGER et al., 2018). As it is also highly persistent in the environment, the development of treatment technologies capable of removing BPA from the environment is of utmost importance.
Advanced Oxidation Processes (AOP) are technologies increasingly studied for the removal of many persistent compounds in the environment, mainly in the aqueous medium. They can be defined as processes that generate highly-reactive, non-selective hydroxyl radicals in sufficient amounts for oxidizing the majority of the complex chemicals present in a wastewater (GOGATE; PANDIT, 2004;CHONG et al., 2010;OLLIS et al., 2011;SILANPÄÄ et al., 2018).
Among the various AOP, the UV/H 2 O 2 process can lead to the complete degradation and mineralization (the conversion into CO 2 and H 2 O) of many organic pollutants (LEGRINI et al., 1993;RIBEIRO et al., 2015). Generally, such systems consist of the addition of H 2 O 2 and the use of a UV light source emitting between 200 and 280 nm (AL-KDASH et al., 2004;STASINAKIS, 2008). That combination leads to increased quantum yields, since both H 2 O 2 and the majority of the organic compounds absorb energy within that range (CHU, 2001;YUAN et al., 2009;SHU et al., 2013). The accepted mechanism for the photolysis of hydrogen peroxide is the homolysis of the molecule into hydroxyl radicals, according to Equation 1 (ESPLUGAS et al., 2002).
The UV/H 2 O 2 process has a number of advantages: low waste generation and robustness, as H 2 O 2 is miscible in water and has almost no limitation regarding pH. However, H 2 O 2 in excess acts as a • OH radicals scavenger, decreasing the removal efficiency. Nevertheless, long treatment times are usually necessary for achieving total mineralization of the organic matter present, increasing the power consumption and hindering its applications at larger scales. Therefore, many studies try to combine the UV/H 2 O 2 process to other techniques (BOLTON et al., 1998;KURBUS et al., 2003).
The simultaneous application of microwaves (MW) and UV light leads to better results in photochemical processes. KLÁN et al. (2001) proposed one of the first reactors combining the UV/H 2 O 2 process with the MW. The major advantage of MW is its greater ability to accelerate chemical reactions with improved yields and selectivity (PERREUX;LOUPY, 2001;HORIKOSHI et al., 2008).  (FERRARI et al., 2009;CHEN et al, 2011;REMYA and SWAIN, 2019). This paper proposes a novel reactional setup for the UV/H 2 O 2 /MW process with no EDL as the UV source, allowing for inferring the existence of MW non-thermal effects. Table 1 shows some information about bisphenol-A. A stock solution of bisphenol A (Sigma-Aldrich) 10 mg L −1 was prepared with ultrapure water, with no pH adjustment. Analytical grade H 2 O 2 (29% in weight -Synth) was used in the degradation studies. Ammonium metavanadate (Vetec) and

Reagents
sulfuric acid (Mallinckrodt) were used in the residual peroxide analysis. Ultrapure water, HPLC-grade acetonitrile (Merck) and acetic acid (Merck) composed the mobile phase for BPA quantification.

Reactional System
The proposed reactional system is shown in Figure 1. It consists of an irradiation unit with a medium-pressure mercury vapor lamp (250 W, Philips HPL-N, radiant flux of 108 kJ m −2 s −1 at  > 254 nm) and a magnetic stirrer, connected to an adapted domestic microwave oven (1100 W, 2450 MHz) by a pumping system with silicon hoses. A ventilation system was installed in the microwave to avoid overheating. In order to minimize thermal effects from the microwave heating, a cooling system was placed between the microwave oven and the irradiation unit. The system temperature was monitored at the exit of the microwave oven and at the entrance of the irradiation unit. Flow rate was measured and controlled by a rotameter placed just before the microwave oven. Both reactors (one inside the irradiation unit and another one inside the microwave oven cavity) consisted of borosilicate glass recipients with 250 and 350 mL nominal volume, respectively. The system working volume was fixed at 800 mL.

Fig. 1
Schematics of the reactional system used in this work.

Experimental Procedure and Analytical Methods
In a typical run, the system was filled with 800 mL of 100 µg L −1 aqueous BPA (pH = 6.0). The working volume of the reactor inside the irradiation unit was fixed at 150 mL, so that the distance between the lamp and the surface of the solution was kept constant (20 cm). The reactor inside the microwave oven ( Figure 2) was completely filled. Prior to the experiments, the system was purged in order to eliminate any bubbles, guaranteeing a steady flow. H 2 O 2 concentration, flow rate, and MW power in each experiment were set according to a factorial design. The experiments were started when turning on the lamp and the MW oven (during 30 min), simultaneously. With the help of the cooling system, the temperature was kept at 20 ± 2ºC at the entrance of the irradiation unit. Two digital thermometers were placed at the irradiation unit and at the exit of the MW oven, so that the temperature could be measured. Experiments running only ultrapure water with different combinations of flow rate and MW power were performed in order to measure the respective temperature increase. In order to optimize the system, an experimental design methodology, namely the response surface methodology (RSM), was used. First, a preliminary 2³ full factorial design was performed, with duplicates of each experiment. The studied factors were: initial H 2 O 2 concentration, flow rate, and MW power (the respective levels are shown in Table 2), and the response variable was BPA degradation (BEZERRA et al., 2008).
Second, based on the obtained results, a series of experiments were performed along with the path of steepest ascent, i.e., aiming at increasing degradation, until the optimal region was reached (Table   4). Finally, the optimal degradation condition was obtained with additional experiments that followed an experimental design with more than two levels (BEZERRA et al., 2008). The Statistica ® 11.0 software was used for the necessary calculations.
Due to the low initial concentration of BPA, extraction and concentration were necessary. The The slope of the decay curve at any time can be mathematically determined by taking its derivative in time (Equation 4).
When tends to zero, the slope tends to − 1⁄ (h −1 ) and its physical meaning is the initial (pseudo-first-order) removal rate of BPA. The removal rate here has no concentration units because 0 ⁄ is dimensionless.
On the other hand, when is long enough, tending to infinity, 1⁄ is the theoretical maximum removal fraction, which is equivalent to the maximum oxidation capacity (MOC) of the system (Equation 5).
The characteristic constants ( and ) were estimated with the aid of the Solver package (Excel ® software), using the Generalized Reduced Gradient (GRG) nonlinear solving and least squares methods.

Temperature Increase and H 2 O 2 Decomposition by Microwave Heating
From the exit of the cooling system to the entrance of the MW oven, the temperature was maintained at 20 ± 2ºC. Some experiments evaluating the temperature increase at the exit of the MW oven, depending on flow rate and MW power, are described in Table 3. As expected, the lowest flow rate combined with the highest MW power generated the greatest temperature increase: 15 ± 2ºC. Nonetheless, the observed temperature increases seemed not to significantly interfere with the concentration of hydrogen peroxide at the end of the reactional period (30 min): 601 ± 8 µg L -1 and 616 ± 5 µg L -1 with and without MW irradiation, respectively. Those results show the temperature increase due to the MW radiation did not significantly decompose hydrogen peroxide.  The Pareto chart shown in Figure 3 allows one to assess the statistical significance of each factor and of their respective interactions (bars that cross the red line, 95% confidence interval). All factors and interactions were significant within the studied ranges. The 2 2 factor was the most significant one, as hydrogen peroxide is the source of hydroxyl radicals, responsible for BPA degradation. Its positive effect (positive number next to the respective bar) means that degradation increased when the 2 2 factor was varied from the low to the high level. The flow rate factor also had a positive effect probably because increased flow rates allow for greater volumes of solution to be irradiated during the reaction time.

Experimental Design
On the other hand, MW power presented a negative effect within the studied range. As showed before, the variation of temperature seems to not influence the production of hydroxyl radicals. However, one cannot state whether the presence of MW radiation does or does not influence BPA degradation kinetics and/or the generation of the degradation products. Moreover, that negative effect does not necessarily mean that the MW radiation was detrimental to the system, when compared to another system without it.  According to the Pareto chart (Figure 4), the variation of 2 2 had a positive effect on the reaction system and was statistically significant within the studied range; on the other hand, the increment of flow rate values tends to reduce the removal of BPA, although it was not statistically significant, again within the studied range.  The response surface for the latter experiments ( Figure 5) showed that the optimal condition was not achieved yet. Therefore, a new series of experiments were performed, varying only 2 2 and setting MW power and flow rate factors at 245 W and 700 mL min −1 , respectively. The results are shown in Table 6   Finally, the optimal degradation conditions were set: 2 2 = 20 mg L −1 , flow rate = 700 mL min −1 , and MW power = 245 W. In 30 min, approximately 95% of the initial BPA was removed.

The Effect of MW Radiation
The main feature of the proposed setup is the possibility of performing degradations by several means: UV only, H 2 O 2 only, MW only, UV/H 2 O 2 , UV/MW, H 2 O 2 /MW, and UV/H 2 O 2 /MW. Therefore, the BPA degradation by the UV/H 2 O 2 /MW and the UV/H 2 O 2 combinations were directly compared. At the optimal conditions, no significant difference was observed between them, both achieving 95% BPA degradation in 30 min. This finding contradicts the preliminary factorial design, which demonstrated that MW radiation did have a significant effect.
Then, it was hypothesized that high H 2 O 2 concentrations could "mask" the effect MW have, resulting in similar results. To check that hypothesis, the reaction conditions were changed to the ones of experiments # 5/13 from the preliminary factorial design ( 2 2 = 5 mg L −1 , flow rate = 600 mL min −1 , and MW = 245 W). The corresponding results are presented in Figure 6. The results showed a significant difference between the UV/H 2 O 2 and the UV/H 2 O 2 /MW systems: in 60 min, 65 and 95% BPA degradations were achieved, respectively. Moreover, the MW roughly doubled the initial pseudo-first-order degradation rate (from 0.046 to 0.10 min −1 ) and the MOC increased from 86 to 100%.
The reasons behind those results are still unclear. The temperature increase of 4  2°C inside the MW oven does not seem to be solely responsible for doubling the initial degradation rate. As argued by many researchers, MW radiation may favor some reaction mechanisms among others, which may lead to different degradation rates and oxidation capacities. As discussed before, the real effects of MW radiation on any reactional system are still under debate in the scientific community, remaining the existence of non-thermal MW effects an open question (NÜTCHER et al., 2004;DUDLEY et al., 2015;DÍAZ-ORTIZ et al., 2019;TIAN;LIO;ZU, 2020).

Comparison with the literature
Comparing the present study with others already published is a challenge. Different initial BPA  Table 7 summarizes (and somewhat compares) this work with those other three.
As one can see in Table 7, regarding the UV/H 2 O 2 process, the normalized removal rate, ̅ , obtained by Sharma, Mishra, and Kumar (2015) is one order of magnitude lower than the other two.
Moreover, the use of O 3 by Liu et al. (2018) did not improve BPA degradation. However, two points must be highlighted.
First, the results reported by Liu et al. (2018) are difficult to be compared with the others due to the pH they used (9.0), which is close to the pKa of BPA (9.6) . Certainly, at that pH, the concentration of the conjugated base of BPA is quite significant. Therefore, the reactions involved in the BPA degradation were probably different. Second, as Sharma, Mishra, and Kumar (2015) did not optimize the reaction using a multivariate approach, it is not fair to compare optimized and non-optimized degradations. , increased significantly (from 42.2 to 65.5 g, which is approximately 1.5 times greater), as well as the normalized removal rate (from 9.4 × 10 −2 to 1.5 × 10 −1 ).

CONCLUSIONS
• • The response surface methodology allowed for optimizing the degradation system with a small number of experiments and yet, statistically sound information.
• The effect of microwaves upon BPA degradation was only perceivable at relatively low H 2 O 2 concentrations. It was hypothesized that increased H 2 O 2 concentrations would allow for the generation of great amounts of hydroxyl radical, "masking" effects of smaller magnitude, as it is the case of that of microwaves in this system.
• The effect microwaves exerted upon the system was two-fold: both the pseudo-first-order degradation rate and the maximum oxidation capacity were significantly enhanced.