3.1. Formation of ClO3− and ClO4− in EO systems
Cl− can be oxidized in EO systems to form reactive chlorine species that lead to ClO3− and ClO4−. It was shown that the presence of 1.0-mM Cl− resulted in increased HOCl, ClO3−, and ClO4− on BDD and Ti4O7 anodes (Fig. 1). Almost no appreciable HOCl was detected during the 4.0 h electrooxidation process on the BDD anode, while the concentration of HOCl increased continuously in the system with Ti4O7 anode and reached 103.2 µM at 8.0 h. In both systems, ClO3− concentration increased and then plateaued, while the concentration of ClO4− increased continuously. The transformation of Cl− was faster on BDD anode in general. The data (Fig. 1) indicate that the concentration of ClO3− reached 276.2 µM in the first 0.5 h and then decreased on BDD anode, while almost all Cl (about 1000 µM) was transformed to ClO4− within 4 h. The formation of ClO3− on Ti4O7 electrode was slower, reaching a plateau of ~ 350.6 µM in 4.0 h and then decreasing slowly. The formation of ClO4− also appeared to be more slowly on Ti4O7 electrode, taking about 8.0 h to reach 780.0 µM.
Cl− can be transformed in electrooxidation by direct electron transfer (DET) to ClO3− and ClO4− through a pathway of multiple steps (R1 − R3). A previous study showed that the DET of Cl− did not occur on the Ti4O7 anode, thus resulting in slower formation of ClO3− and ClO4− than on BDD anode11. However, indirect routes (R4 − R5) can lead to Cl• generation on both Ti4O7 and BDD anode11, which can further go through the reactions in R2 and R3 to form HOCl and chlorinated by-products. It is noted that the conversion of Cl− to HOCl and the chlorinated byproducts via both DET and indirect routes involves the hydroxyl radicals (HO•) that are formed by water oxidation on anode.
\(\equiv \text{S}\) +C\({\text{l}}^{-}\to \equiv \text{S}\left(\text{C}{\text{l}}_{\text{a}\text{d}\text{s}}^{{\bullet }}\right)\) +\({\text{e}}^{-}\) (R1)
$$\text{C}{\text{l}}^{{\bullet }}+\text{H}{\text{O}}^{{\bullet }}\underrightarrow{\text{-}{\text{H}}^{+}}\text{OC}{\text{l}}^{-}$$
R2
$$\text{OC}{\text{l}}^{-}\underrightarrow{\text{-e,}+\text{H}{\text{O}}^{{\bullet }},\text{-}{\text{H}}^{+}\hspace{1em}}\text{Cl}{\text{O}}_{2}^{-}\underrightarrow{\text{-e,}+\text{ H}{\text{O}}^{{\bullet }},\text{-}{\text{H}}^{+}\hspace{1em}}\text{Cl}{\text{O}}_{3}^{-}\underrightarrow{\text{-e, }+\text{H}{\text{O}}^{{\bullet }},\text{-}{\text{H}}^{+}}\text{Cl}{\text{O}}_{4}^{-}$$
R3
\(\text{H}{\text{O}}^{{\bullet }}+\text{C}{\text{l}}^{-}\leftrightarrow \text{ClH}{\text{O}}^{{\bullet }-}\) k = 4.3 × 109 M− 1‧s− 1 (R4)
\(\text{ClH}{\text{O}}^{{\bullet }-}+{\text{H}}^{+}\to \text{C}{\text{l}}^{{\bullet }}+{H}_{2}O\) k = 2.1 × 1010 M− 1‧s− 1 (R5)
The rate of Cl− conversion to chlorate and perchlorate in EO systems has been simulated using a model of two sequential steps by assuming each step as pseudo-first-order kinetics (R6 − R7)27, 28. The rate constants k1 and k2 in such sequential equations were obtained by fitting the data as shown in Fig. 1a and b using the software Kintecus v6.80 29. The values of k1 and k2 were fitted to be 5.40 × 10− 4 and 7.16 × 10− 4 s− 1, respectively, on the BDD anode, while for Ti4O7 anode the values were 8.59 ×10− 5 and 1.34 × 10− 4 s− 1, respectively. This indicates that Cl− is oxidized to ClO3− and ClO4− more easily on BDD, evidenced by the larger k1 and k2 on the BDD anode than on Ti4O7 anode. Retarded formation of ClO3− and ClO4− makes it advantageous to apply Ti4O7 anodes in water/wastewater treatment.
\(\text{C}{\text{l}}^{-}\underrightarrow{\text{H}{\text{O}}^{\bullet }}\text{C}\text{l}{\text{O}}_{3}^{-}\) k1 = -\(\frac{{dc}_{C{l}^{-}}}{dt}\) (R6)
\(\text{C}\text{l}{\text{O}}_{3}^{-}\underrightarrow{\text{H}{\text{O}}^{\bullet }}\text{C}\text{l}{\text{O}}_{4}^{-}\) k2 = -\(\frac{{dc}_{Cl{O}_{3}^{-}}}{dt}\) (R7)
3.2. Effects of electrolytes on the formation of ClO3− and ClO4−
A set of experiments were performed to evaluate the ClO3− and ClO4− formation by EO with BDD and Ti4O7 anode in solutions containing different supporting electrolytes, including 100-mM Na2SO4, NaNO3, Na2B4O7, and Na2HPO4. The concentrations of ClO3− and ClO4− measured in different electrolyte solutions are summarized in Fig. 2. Overall, the transformation was more rapid on BDD anode in all the supporting electrolyte solutions. As shown in Fig. 2a and b, ClO4− concentration reached 990 µM after 4.0 h with Na2SO4 as supporting electrolyte, which accounts for about 99% of the total Cl− initially included in the solution. Almost no ClO3− was detected. At the same current density, the formation of ClO4− was slower with NaNO3, Na2B4O7, and Na2HPO4 as supporting electrolytes on the BDD electrode. As shown, the total ClO3− and ClO4− concentration was lower when Ti4O7 was used as the anode. For example, the ClO4− concentrations were 212.84 and 276.78 µM, respectively, after 4.0 h with Na2B4O7 and Na2HPO4 as supporting electrolytes. In particular, the ClO4− concentration in BDD system was 572.6 µM after 4.0 h with NaNO3 as the supporting electrolyte, while it was only 92.4 µM at the same condition on the Ti4O7 anode.
3.3 Inhibitory effect of co-existing constituents
Experiments were performed to examine EO in the presence of Cl− as well as a few co-existing constituents, including MeOH, H2O2 and KI, so as to investigate the effect of the coexisting constituents on the formation of ClO3− and ClO4− with Ti4O7 anode.
3.3.1 MeOH
The high MeOH content in still bottoms has been shown playing a role to scavenge the chlorine25. As such, the EO experiment was performed in 100-mM Na2HPO4 solutions containing 1.0 mM Cl− and varying quantities of MeOH. The addition of MeOH appeared to impact the conductivity of the reaction solution slightly. The conductivity dropped from 10.51 mS cm− 1 to 9.79 mS cm− 1, but the anodic potential increased at the same current density (10 mA cm2) (Fig. S1a), from 2.93 V in the absence of MeOH increasing to 3.22 V with 100 mM MeOH. The formation of ClO3− and ClO4− during EO treatment at 10 mA‧cm− 2 is displayed in Fig. 3. In the absence of MeOH, ClO3− reached 117.8 𝜇M in about 1.0 h and then decreased. The value decreased to 17.3 and 0.0 𝜇M containing 10 mM and 100 mM MeOH, respectively. Such a time course profile indicates the further reaction of ClO3−. The formation of ClO4− increased monotonically, reaching 329.0 µM in 8.0 h in the absence of MeOH. When 10 mM and 100 mM MeOH were spiked, almost no ClO4− were formed for the first 2.0 h, after which ClO4− started to increase, reaching 300.0 µM and 251.8 µM in 8.0 h, respectively. The formation of ClO4− was completely inhibited when 1000 mM MeOH was added, indicating that MeOH inhibited the formation of ClO3− and ClO4−. Delayed formation of ClO4− in the presence of lower MeOH dosage (10 and 100 mM) may be caused by MeOH depletion over time. Formation of ClO3− and ClO4− was neither observed in acid or neutral conditions when 1000-mM MeOH was spiked, by respectively using 50-mM NaH2PO4 + 50-mM Na2HPO4 (pH 6–7) or 100-mM H3PO4 (pH 2–3) as electrolytes instead of Na2HPO4 (pH 10–11)
A prior study proved that Cl− was not oxidized to Cl• via DET on the Ti4O7 anode, while Cl• was formed mainly through the indirect pathways (R4 − R5)11. Cl• reacts with another Cl− to form Cl2•−. Cl• and Cl2•− also combine with each other to form free chlorine (Cl2, HClO)9, 30, 31. These chlorine species may accumulate and diffuse away from the anode surface, and finally convert into ClO3− and ClO4−. The reaction rate constant between MeOH and Cl• is 5.7 × 109 M− 1 s− 1. Therefore, MeOH can scavenge Cl• in the bulk solution, leading to inhibited generation of ClO3− and ClO4−.
3.3.2 H2O2
Yang et al. found that the formation of ClO4− during EO with BDD anode can be largely inhibited by adding H2O2 27. Therefore, H2O2, a commonly used quenchers were also investigated in this study. The time-course data of ClO3− and ClO4− formation in the presence of H2O2 are shown in Fig. S2. Using Kintecus v6.80, the data in Fig. S2 were fit to obtain k1 and k2 represented in equation R6 − R7, and they were 9.78 × 10− 5 and 7.09 × 10− 4 s− 1, respectively, in the absence of co-existing constituents (Fig. 4). The values of k1 and k2 decreased to 1.16 × 10− 6 and 1.87 × 10− 4 s− 1 when 1000-mM H2O2 were spiked, respectively. The data shown in Fig. S2 and Fig. 4 also showed that addition of H2O2 at 100 mM also significantly limited ClO3− and ClO4− formation during the EO.
H2O2 is known to be both an oxidant (H2O2/H2O, E0 = 1.76 V) and a reductant (O2/H2O2, E0 = 0.68 V), depending on pH. Earlier studies have demonstrated that HOCl can be reduced back to Cl− by H2O232, 33(R8-R9). In addition to free chlorine, H2O2 can also react with the chlorine radical species directly (R10-R11). Thus, it is presumed that the reduction of HOCl and chlorine radical species by H2O2 outweighed the oxidation of Cl− by H2O2 in the EO system, and thus decreased ClO3− and ClO4− formation.
\(\text{H}\text{O}\text{C}\text{l}+{\text{H}}_{2}{\text{O}}_{2}\to {\text{H}}^{+}+{\text{C}\text{l}}^{-}+{\text{H}}_{2}\text{O}+{\text{O}}_{2}\) k = 1.1 × 104 M−1s−1 (R8)
\(\text{C}{\text{l}}_{2}+{\text{H}}_{2}{\text{O}}_{2}\to {\text{O}}_{2}+2\text{H}\text{C}\text{l}\) k = 1.3 × 104 M−1s−1 (R9)
\({\text{C}\text{l}}^{\bullet }+{\text{H}}_{2}{\text{O}}_{2}\to \text{H}{\text{O}}_{2}^{\bullet }+{\text{C}\text{l}}^{-}+{\text{H}}^{+}\) k = 2.0 × 109 M−1s−1 (R10)
\({\text{C}\text{l}}_{2}^{\bullet -}+{\text{H}}_{2}{\text{O}}_{2}\to \text{H}{\text{O}}_{2}^{\bullet }+{2\text{C}\text{l}}^{-}+{\text{H}}^{+}\) k = 1.4 × 105 M−1s−1 (R11)
3.3.3 KI
Cl− (Cl•/Cl−, 2.41 V) and Br−(Br•/Br−, 1.62 V) can be oxidized by HO• to form carcinogenic chlorate and bromate 34, while I−, having a lower reduction potential of 1.33 V 35_ENREF_25, may be more readily oxidized than Cl− and Br− in theory_ENREF_31. It was also found in our previous studies that NaI may be used as a Cl− free salt to regenerate PFAS-laden ion exchange resin without compromised capability in PFAS recovery 25. To evaluate the impact of I− on the formation of ClO3− and ClO4− during EO process, an EO experiment was performed in the presence of I−. It should be noted that the anodic potential was relatively constant at the same current density (10 mA cm2) with I− at different levels (Fig. S1b). The presence of I− inhibited the formation of ClO3− and ClO4− significantly as shown in Fig. 5. Almost no ClO3− was formed during the first 4.0 h and then increased to 25.5 µM after 8.0 h in the presence of 20-mM KI. Similarly, the formation of ClO4− increased slowly during the first 4.0 h and reached 287.2 µM at 8.0 h. Furthermore, near-complete inhibition of ClO3− and ClO4− formation was achieved when 100 mM KI was spiked, with the values of k1 and k2 decreased to 0 and 4.76 × 10− 6 s− 1, respectively (Fig. 4). This suggests that I− outcompetes CI− for reaction with HO•, leading to a slower generation of HOCl on Ti4O7, and thus inhibiting the formation of ClO3− and ClO4−.