In our previous work, we studied and optimized an electrochemical reactor for the treatment of bisphenol A (BPA) using the electro-Fenton process at a laboratory scale (Chmayssem et al. 2017). We investigated the effect of the applied current intensity on the removal efficiency of BPA during the treatment process. Our findings revealed that the limiting current intensity under the hydrodynamic conditions of 6 L.min-1 was found to be approximately 0.8 A (0.51 A / 100 g of glassy carbon pellets). Furthermore, the limiting current density was determined to be 0.012 A/cm², taking into account the section of the fixed bed that is exposed to the flow.
In this work, the focus was on optimizing the current intensity for the treatment of bisphenol A (BPA) using the electro-Fenton process at a semi-pilot scale. To achieve this, the variation of the removal efficiency according to time for different applied current intensities ranging from 10 A to 30 A was studied. The results, illustrated in Figure 4, show that the removal efficiency improves with an increase in the applied current intensity, until the range of 15 to 17.5 A is reached. However, for higher intensities, the treatment appears to be less efficient. This suggests that the applied current intensity is higher than the limiting current for the O2 reduction and that a portion of the produced H2O2 is being reduced to H2O (Babić et al. 1993; Noe et al. 2012). Therefore, the limiting current intensity under the hydrodynamic conditions corresponding to 10 L.min-1 (600 L/h) was determined to be near 17.5 A (0.51 A / 100 g of glassy carbon pellets). Theoretically, this current should be divided equally among 5 cathodes and therefore, each cathode should be supplied by 3.5 A as the cathodes were connected electrically in parallel.
To verify the hypothesis that the number of electrodes affects the removal efficiency of BPA during the treatment by electro-Fenton process, we investigated the effect of the electrodes number using the semi-pilot scale reactor. To do this, two experiments were carried out using an initial concentration of 10 mg/L of BPA. The first experiment used the typical reactor configuration of 5 cathodes and 6 anodes, and the second experiment used a configuration of only one 3D-cathode and 2 DSA-anodes (configuration 2). Figure 5 shows the variation of the removal efficiency of BPA over time for both cases. As expected, the typical reactor configuration (5 cathodes and 6 anodes) was found to be more efficient in degrading BPA compared to the configuration of only one cathode (configuration 2). This is likely due to the increased efficiency of reducing oxygen into H2O2 and producing •OH as the number of cathodes in the electrochemical reactor increases.
Furthermore, it is widely accepted that the reaction of •OH with phenolic compounds is primarily considered as a pseudo-first-order reaction, suggesting a direct attack of hydroxyl radicals on BPA (Brillas et al. 2009). In light of this hypothesis, the apparent constant rate (Kapp) of BPA could be calculated using the theoretical equation of a pseudo-first-order reaction. Figure 5 (inset) shows the variation of Ln (C0/Ct) over time during the first 10 minutes of treatment of BPA (10 mg/L) by electro-Fenton process at the semi-pilot scale reactor. As expected, a linear relation is observed with a slope of 0.076 min-1 corresponding to the value of Kapp in the typical reactor configuration (5 cathodes) and 0.376 min-1 corresponding to the value of Kapp in the configuration containing only one cathode. These results indicate that the typical reactor configuration (5 cathodes) is more efficient in degrading BPA compared to the configuration of only one cathode.
The ratio of the obtained slope values is approximately 5, which confirms that the configuration of the EF reactor with 5 cathodes is 5 times more efficient than the reactor configuration that contains only one cathode. As expected, this confirms that the applied current intensity in the electrochemical reactor was equally distributed among the 5 cathodes, as the cathodes were electrically connected in parallel, resulting in a current density of 0.012 A/cm² (Fig. S2). This value is comparable to the current density found in the laboratory scale reactor of the electro-Fenton process (Chmayssem et al. 2017). In summary, these results confirm that our reactor design could be a good candidate for the scale-up of the electro-Fenton process for industrial treatment applications.
On the other hand, the electro-Fenton experiments at the semi-pilot scale reactor were initially conducted using high concentrations of BPA solutions in order to monitor the concentration of BPA during treatment and study the degradation kinetics as a function of treatment time. Herein, we investigated the impact of initial BPA concentration on the removal efficiency of BPA during treatment using the electro-Fenton process (from 10 to 150 mg.L-1). We carried out experiments for 90 minutes at 15 A and found that as the initial concentration of BPA increased, the time required to achieve a certain level of removal efficiency also increased (Fig. 6). For instance, at a BPA concentration of 150 mg.L-1, it took 50 minutes to achieve 90 % removal efficiency. However, as the initial concentration decreased (i.e. for 10 mg.L-1), the time required to achieve the same removal efficiency decreased as well (less than 5 min). This suggest that this semi-pilot reactor configuration is more efficient in treating low concentration of persistent pollutant by a single pass of the solution through the fixed bed cathode.
The effect of flow rate on the removal efficiency of BPA by the EF process at semi-pilot scale was also investigated. Figure 7 illustrates the variation of the removal efficiency over time at 15 A for three different flow rates: 200 L/h, 600 L/h and 1000 L/h (the maximum flow rate at semi-pilot scale reactor). As shown, an improvement in removal efficiency was observed as the flow rate increases. For example, after 15 minutes of treatment, a removal efficiency of 45% was observed at 200 L/h, while the removal efficiency increased to 65% at a flow rate of 1000 L/h. This observation can be attributed to the increase in liquid flow velocity (v) through the fixed bed, which can be calculated as 0.0019 m.s-1 at 200 L/h and as 0.0097 m.s-1 at 1000 L/h, ensuring more favorable hydrodynamic conditions. This results in an increase in the Reynolds number (Re= v*dp/n) and subsequently the mass transfer coefficient (km= A*vb), which leads to an improvement in the degradation efficiency of BPA, where: dp is the particle diameter, n the kinematic viscosity, A and b constants (M. Benzina. S. Gabsi 1991; Sedahmed 1996). For this reason, it is of high importance to provide a reactor design that support high liquid flow rates that allows to apply high current intensities in optimized mode and as a result to reach high removal efficiency of BPA in short treatment time (a few m3.h-1).
Moreover, electro-Fenton process typically involves biodegradability change (Hou et al. 2015). Today, many studies shows that combining EF process prior to the biological treatment could be a particularly interesting solution for water remediation (Mansour et al. 2014; Olvera-Vargas et al. 2016; Aboudalle et al. 2018; Baiju et al. 2018; Arellano et al. 2020). For this reason, we evaluated the biodegradability change during treatment of BPA by EF process at semi-pilot scale reactor. To proceed, we used the BOD5 (Biochemical Oxygen Demand) to COD (Chemical Oxygen Demand) ratio, which is a standardized measure of the biodegradability of a solution, and considered a ratio higher than 0.4 to indicate that the solution is biodegradable (Oller et al. 2011; Rivero et al. 2014). We investigated the biodegradability of BPA for different treatment times (30, 60, and 90 minutes) and for two initial BPA concentrations of 10 and 50 mg.L-1. Additionally, we studied the mineralization of BPA by monitoring the TOC (Total Organic Carbon) versus the time of EF treatment (30, 60, and 90 minutes).
As seen in Figure 8, a significant improvement in biodegradability (BOD5/COD > 0.6) is observed after 90 minutes of treatment, with a 33% reduction in TOC at an initial BPA concentration of 50 mg.L-1. Additionally, good results were obtained for low concentrations of BPA (≤ 10 mg.L-1), where only a few minutes of treatment were needed to improve biodegradability. This enhancement is also reflected in a high TOC abatement (> 70 %) at the end of the treatment (90 min). It can be observed that COD and TOC abatement increases with treatment time and as the initial BPA concentration decreases. Here, the residual TOC is mainly composed of small organic acids which are difficult to oxidize as reported elsewhere (Chmayssem et al. 2017) and typically require many hours of treatment to achieve total mineralization (Ayoub et al. 2011; Ozcan et al. 2013; Mansour et al. 2015).
For the scale-up of EF reactor to an industrial pilot plant, two cases should be considered based on the source and the concentration level of the effluent (Fig. 9). For the treatment of an industrial discharge with high contamination levels (> mg.L-1), a batch reactor should be designed with multiple fixed bed cathodes in series as a pre-treatment process. The treatment duration will depend on the pollution charge and biodegradability level. For low concentration levels of contamination (< mg.L-1), treatment can be directly applied through a single pass of the solution in the electrochemical reactor. Few seconds of treatment should be sufficient and biodegradability can be achieved through the electro-Fenton reactor. The current intensity should be set at 0.51 A per 100 g of glassy carbon pellets (0.012 A.cm-2 for the section of the fixed bed exposed to the flow). Scale-up is limited by the flow rate of the solution, which should be kept within acceptable values (a few m3.h-1). To make scaling up easier, it is recommended to keep the thickness of the 3D cathode bed constant as used in the laboratory and semi-pilot scale reactor (2 cm).