Photoelectrocatalytic treatment and resource utilization of industrial waste salt for chlor-alkali electrolysis

Pesticides, fine chemicals, and many other chemical industries usually produce a large amount of waste solid salt which is detrimental to the environment when treated by burning and rigid landfill. In contrast to traditional disposal strategies, resource utilization of waste salt is beneficial for both the environment and economy. However, the current technique for the resource utilization of waste salt, such as nanofiltration, is high cost and hard to popularize. In this study, the photoelectrocatalytic treatment of waste salt obtained from the glyphosate industry and its utilization as a raw material for chlor-alkali electrolysis are proved to be feasible. The waste salt consists mainly of NaCl, with ~ 1.31 wt% of organic impurities. A TiO2 nanotube electrode was employed for the photoelectrocatalytic treatment of brine with NaCl concentration of 270 g L−1 prepared from waste salt. After preliminary treatment, the total organic carbon content (TOC) of the waste salt brine was reduced to 50 mg L−1, with a removal ratio of 85%. It is able to meet the standard of refined brine in the chlor-alkali industry (TOC < 20 mg L−1) with further treatment. A study on the photoelectrocatalytic mechanism reveals that the main oxidative species contributing to the degradation are holes (h+) and chlorine active substances other than Cl∙ under the condition of high Cl− concentration. The organic impurities in the waste salt are poisonous to both the electrode and membrane in the process of chlor-alkali electrolysis, leading to an increase in the voltage. With photoelectrocatalytic treatment, most of the organic impurities can be removed so that the waste salt can be utilized as a raw material for chlor-alkali electrolysis.


Introduction
As one of the most important economic pillars in China, the chemical industry is still facing safety and pollution problems. Waste salt is a typical class of solid waste extensively produced in major chemical industries including the fine chemical, pesticide, pharmaceutical chemical, soda ash, and caustic soda industries [1]. These waste salts produced by wastewater treatment usually contain residues of highly toxic heavy metals and organic compounds [1,2]. The process to treat wastewater contaminated with salts can be divided into direct treatment (such as high-temperature oxidation) and treatment as concentrated brine wastewater. The latter usually adopts separation and oxidation processes, including physical methods (such as extraction, recrystallization, resin exchange, etc.) [2][3][4][5], chemical methods (such as advanced oxidation processes (AOPs), oxidant oxidation, etc.) [3,4], and biological methods (such as halophilic organisms, etc.) [6]. Among them, AOPs take the advantages of good development and application prospects due to their strong oxidation ability gained through the generation of hydroxyl radicals (•OH) in addition to their low risk for secondary pollution. AOPs generally refer to catalytic wet oxidation (CWO) [7], O 3 oxidation [8], Fenton process [9], ultraviolet photolysis (UV) [8], electrochemical oxidation [10], photocatalytic oxidation and their combined processes such as photoelectrocatalysis (PEC) [11,12]. The application of PEC to electrolyte with high conductivity and outstanding performance in concentrated waste brine treatment has attracted increasing attention from researchers in recent years [13].
PEC, a combined process of photocatalysis (PC) and electrocatalysis (EC), uses an external bias to transfer photogenerated electrons to the circuit, which reduces the recombination rate of photogenerated electrons and h + and improves the wastewater treatment efficiency. The electrolyte concentration can affect circuital current and thus have a significant impact on the photoelectrocatalytic treatment effect [14]. Therefore, the electrolyte is usually required to be conductive. In addition, the operation variables such as the materials and structure of photoanode, the applied current or bias voltage, the type of electrolyte and its pH value, as well as the type and content of pollutants, also have an impact on the treatment efficiency [13,15]. Jingdong Zhang et al. had summarized the application of different light sources in wastewater photoelectrocatalytic treatment [13]. More than 75% of the relevant literatures in 2015 ~ 2020 reported the application of visible light or solar light source in the photoelectrocatalytic treatment of wastewater, while the application of traditional ultraviolet light source was much fewer. This statistic shows that the research focus has transferred to visible light and sunlight light sources [13]. Accordingly, in order to meet the requirement of visible light excitation, traditional photoelectrocatalytic photoanode materials such as TiO 2 and ZnO need to be modified or replaced by narrow band gap semiconductors such as BiVO 4 and Fe 2 O 3 [16,17]. Though the researches on PEC have made some progress, most of them are still limited to synthetic wastewater on a laboratory scale with low pollutant concentration of 5 ~ 50 mg L −1 in general. Some companies report that PEC has low efficiency in actual wastewater treatment [13,18]. At present, photoelectrocatalytic technologies are still far from application for actual wastewater treatment.
Based on the sustainability awareness of 3R (reduction, reuse and recycle), the ideal strategy for waste salt treatment is realizing harmless treatment and resource utilization in an eco-friendly and low-cost manner [18]. The recently developed Hazardous Industrial Salt Waste Innocuous Disposal Integration (HDI) is a typical example [19]. Treated with highly active activity conversion additives such as active silicon and aluminum substances at high temperature condition, waste salt will be transformed into inert glass phase with a stable lattice network structure. Such crystal materials are harmless to the environment and can be used as fracturing proppant and building materials for oil and gas exploitation. HDI exhibits a good application prospect [19]. The products after harmless treatment can be effectively reused to achieve additional economic benefits, thereby reducing the overall treatment cost.
In this study, an idea of recycling waste salt from the chemical industry was proposed and verified to be feasible in the laboratory. The waste salt obtained from the production line of a glyphosate production company in China, was prepared from glyphosate wastewater after preliminary catalytic wet oxidation and evaporation treatments. First, the composition of waste salt was characterized. Then, the waste salt was treated by PEC with a TiO 2 electrode under UV light illumination. Finally, the treated salt was reused as raw material for chlor-alkali electrolysis. Studies on photoelectrocatalytic technologies have indicated the stable and high energy output of UV light source. On the other hand, TiO 2 nanotubes (TiO 2 NTs) have higher pollutant degradation efficiency, while maintaining the advantages of TiO 2 electrodes, namely low cost and high stability [20]. Thus, in order to provide sufficient oxidation capacity for the treatment of the concentrated waste salt brine to meet the standard of chloralkali refined brine, UV lamp and TiO 2 nanotube material are selected as the light source and photoanode, respectively.
Resource utilization is the most ideal strategy for chemical waste salt treatment. For example, NaCl waste salt generally originates from chemical industries such as the glyphosate industry, as a consequence of reaction pH adjustment with NaOH and hydrochloric acid. Besides, the chlor-alkali industry requires refined NaCl as the raw material, and chlor-alkali electrolysis products can be used to prepare raw materials for glyphosate production. Thus, advanced treatment of NaCl waste salt may develop a closed-loop process for the resource utilization of NaCl waste salt to the benefit of both the economy and environment. However, few processes for the resource utilization of waste salt have been applied to practice due to the extremely complex types and compositions of industrial waste salt. Figure 1a shows a successful recycling route for the NaCl waste salt gained from the glyphosate industry [21]. During the CWO process, most of the organic impurities in the glyphosate mother liquor were degraded, with only a small amount of organic residual remained, which could be further removed by the subsequent nanofiltration process. Finally, brine with the total organic carbon content (TOC) lower than 20 mg L −1 was obtained and could be subjected to chlor-alkali electrolysis. But the main drawbacks of this technology are the high maintenance cost of nanofiltration membrane filtration and small space for process improvement.
Based on the high salt concentration of the glyphosate mother liquor, it is believed that electrochemistry-related advanced oxidation processes such as EC or PEC may be suitable for its treatment. In this study, PEC was chosen as a potential alternative to the nanofiltration process for the advanced treatment of glyphosate wastewater, as depicted in Fig. 1b. As a combined process, PEC shows the synergistic effect of PC and EC with higher quantum efficiency and oxidation capability [22,23]. It is regarded as one of the most promising advanced oxidation processes, especially when Glyphosate production Fig. 1 a Nanofiltration process and b photoelectrocatalytic treatment process for the resource utilization of NaCl waste salt from the glyphosate industry employed in specific situations, such as its application to salty wastewater treatment.

Characterization of waste salt
The waste salt was obtained from Hubei Xingfa Chemicals Group Co., Ltd. in Yichang city, Hubei Province, China. Its content of water and organic impurities was calculated by the weight loss upon its heat treatment at 105 °C for 24 h and 600 °C for 4 h at ambient atmosphere, respectively. Thermogravimetric analysis on the sample was performed via a simultaneous thermogravimetry and differential thermal analyzer (TG∕DTA, TG 209 F3, NETZSCH, Germany) with a heating rate of 10 °C per minute under air atmosphere. X-ray diffraction (XRD, PW3040∕60, PANalytical, Netherlands) technique was used to characterize the phase composition of the sample with a scanning speed of 4° per min. X-ray photoelectron spectrometer (XPS, Thermo Scientific, USA) was used to analyze the element types of the waste salt and the binding energy of the spectrum was calibrated with C1s peak at 284.6 eV as a reference. The functional group information of the organic impurities was characterized by Fourier transform infrared spectroscopy (FTIR; Perkin Elmer; USA), with a spectral range from 4000 to 500 cm −1 .
The detailed chemical composition of the organic impurities of the waste salt were studied by high-performance liquid chromatography (HPLC, Agilent, USA) with a ZORBAX ion-exchange column (SAX, Agilent, USA) and a UV detector. The mobile phase was an aqueous solution containing 0.01 mol L −1 KH 2 PO 4 and 3% methanol with a pH range from 2 to 3. The flow rate of the mobile phase was 1.5 mL min −1 , and the wavelength of the UV detector was fixed at 215 nm (Fig. S1a-g, Supporting Information) [24].

Preparation and characterization of TiO 2 nanotube electrode
The TiO 2 nanotube electrode was prepared by the anodic oxidation method. A piece of titanium foil of 13 cm×15 cm was washed in ethanol for 10 min and etched in 20 wt% nitric acid aqueous solution containing 4 wt% HF for 1 min, and flushed in deionized water and dried with nitrogen gas. Then, the clean titanium foil was subjected to anodic oxidation at 60 V DC (GERS, China) potential for 10 h, employing a nickel foam with the same size as the auxiliary electrode. The distance between the two electrodes was controlled at 2 cm. The electrolyte was an organic solution of ethylene glycol containing 12.5 wt% H 2 O and 0.5 wt% NH 4 F. The reaction temperature was kept at 25 °C using a water bath [20,25].
After the anodic oxidation, the titanium foil was taken out and flushed with ultrapure water immediately. Then the asprepared sample was dried at room temperature and calcined at 450 °C for 2 h at ambient atmosphere to obtain the final anatase TiO 2 nanotube electrode.
The cross section and surface morphology of the TiO 2 sample were observed by a scanning electron microscope (SEM, NANOSEM 450, FEI), and the phase composition of the sample was characterized by XRD with a scanning speed of 4° min −1 .

Photoelectrocatalytic degradation process
The photoelectrocatalytic experiment was carried out in a two-electrode system employing an electrochemical workstation (CS310M, Corrtest, China). The electrolytic cell was an enamel water tank with a size of 30 cm×20 cm×4 cm. The prepared TiO 2 nanotube electrode with the size of 13 cm×15 cm (the exposure area of 10 cm×15 cm) was used as the anode and a nickel foam as the cathode. Two peristaltic pumps were adopted to circulate the electrolyte. The schematic diagram of the device is shown in Fig. 2. The electrolyte with the NaCl concentration of 270 g L −1 was prepared by dissolving the waste salt in ultrapure water. The photoelectrocatalytic experiment was carried out at 2.5 V constant voltage, under a UV light illumination (365 nm, 20 W) for 50 h at room temperature. During the photoelectrocatalytic process, the electrolyte was sampled at desired intervals and its change in TOC content was analyzed by a TOC/TN tester (TOC-L CPN, Shimadzu, Japan). For comparison, photocatalytic oxidation experiment was carried out at the same conditions but without the potential supply.

Study on the photoelectrocatalytic mechanism
To have an insight into the PEC mechanism at the situation of high concentration chloride electrolyte, experiments on the detection of reactive species generated during the photoelectrocatalytic process were performed, including hydroxyl radical (•OH), chlorine radical (Cl•), and holes (h + ).
For the detection of •OH, terephthalic acid was used as a probe molecule, which can capture •OH in a rapid and targeted manner to generate a highly fluorescent product 2-hydroxyterephthalic acid [26,27]. The experimental process adopted was similar to the photoelectrocatalytic degradation process except replacing the electrolyte by a solution containing a certain concentration of NaCl, 1 × 10 −3 M terephthalic acid, and 4 × 10 −3 M NaOH. Different NaCl concentrations of 0, 0.5, 1, 2, 3, and 4 M were studied to reveal the impact of Cl − concentration on •OH generation. During photoelectrocatalytic treatment, the electrolyte was sampled for the detection of fluorescence signals at 425 nm, which is the characteristic emission wavelength of 2-hydroxyterephthalic acid. A laser of 312 nm was used for the excitation of the species to generate fluorescence.
To verify whether Cl• can be generated during the photoelectrocatalytic process, 5,5-dimethyl-1-pyrrolidine-N-oxide (DMPO) was used as a probe molecule, which can capture highly active Cl• in a targeted manner to form DMPO-Cl with strong seven-line electron spin resonance (ESR) signals [28,29]. The experimental setup was also similar to the photoelectrocatalytic degradation process apart from the difference in the composition of electrolyte. The electrolyte was prepared by adding 150 µL DMPO into 20 ml 270 g L −1 NaCl solution. After photoelectrocatalytic treatment for 5, 30, and 120 min, the electrolyte was sampled for ESR analysis (Bruker A200, Germany) within 10 min (the lifetime of DMPO-Cl is ~ 15 min only [28]).
To explore the impact of h + on the degradation of contaminants on the condition of high Cl − concentration, sodium oxalate was used as h + inhibitor [26,30]. The methodology adopted was similar to the photoelectrocatalytic degradation process except for a little change in the composition of electrolyte. In this experiment, a mixture of 270 g L −1 analytical reagent NaCl, and 50 mg L −1 methyl orange (MO), with or without the addition of 1.5 g L −1 sodium oxalate was used as electrolyte. Considering there may be diffusion limitation, a series of experiments were conducted at different stirring speeds (see Supporting Information for details).

Chlor-alkali electrolysis
The behavior of the waste salt as a raw material for chloralkali electrolysis before and after photoelectrocatalytic treatment was studied on a home-made electrolytic cell, as schematically shown in Fig. S2 (Supporting Information). In this two-electrode system, a 3 cm×6.5 cm titanium (Ti) mesh coated with ruthenium (Ru) and iridium (Ir) was adopted as the anode, while a 3 cm×6.5 cm nickel mesh coated with Ir (ANCAN, China) was applied as the cathode, and a perfluorosulfonic acid-carboxylic acid bi-layer membrane was used [31]. In order to simulate the operation conditions of the chlor-alkali industry, 270 g L −1 NaCl solution with the pH adjusted to 2 ~ 3 was used as the anode electrolyte, and 0.1 M NaOH solution was used as the cathode electrolyte [32]. Before electrolysis, the anode electrolyte was ion-exchanged through a chelating resin (Seplite LSC-500, SUNRESIN, China) packing column to remove Ca 2+ and Mg 2+ . Two electrolyte storage tanks were connected to the positive and negative electrodes, respectively, and the electrolytes were circulated by peristaltic pumps to keep the total amount of electrolyte at both poles at 1 L. The electrolysis was carried out at a constant current of 100 mA for 10 h. To shed light on the impact of the impurities in the waste salt on the electrolysis, different electrolytes were prepared by mixing the waste salt and analytical reagent NaCl at different ratios, which were noted as 0%, 25%, 50%, 75% and 100% (for example, 25% means the ratio of the waste salt and analytical reagent NaCl is 1:3).
Electrochemical impedance spectroscopy (EIS) technique was performed after 30-min cell electrolytic activation at 100 mA to analyze the poisoning mechanism of the impurities on chlor-alkali electrolysis. The EIS data were recorded at open-circuit potential with a frequency range of 100 kHz to 0.01 Hz and an AC amplitude of 10 mV.

Characterization of waste salt
The water content of the waste salt measured by constant temperature drying experiments is 2.87%; and the content of organic impurities is estimated to be about 1.31% according to the calcination test at 600 °C. The TG∕DTA curve of the waste salt is shown in Fig. S3a (Supporting Information). According to the TG∕DTA analysis, the mass loss of waste salt is about 2.63% when heated to 150 °C, which indicated that the evaporation of water is close to its mass loss at constant temperature. The thermogravimetric mass loss of the waste salt from 150 to 600 °C is 5.9%, revealing a higher content of organic impurities than that measured on the condition of constant temperature treatment at 600 °C. This discrepancy may be resulted by the blowing loss of some sample as a result of air flow density changes in the heating range, which can be proved by the mass loss of about 1.29% of analytical reagent NaCl under the same thermogravimetric test conditions in Fig. S3b (Supporting Information). The DTA curve has no obvious peak, indicating that there is no compound dominating the composition of the organic impurities. The survey and fine XPS spectra of the waste salt are shown in Figs. 3 and 4. The main elements of the waste salt are sodium (Na), chlorine (Cl), oxygen (O), carbon(C), nitrogen (N) and phosphorus (P), and the impurities are mainly organic matters containing nitrogen and phosphorus, which are consistent with the source of the waste salt from glyphosate mother liquor wastewater. The atomic contents of both nitrogen and phosphorus are lower than 3%. Figure 5 shows the results of the XRD test. It is evident that the main phase composition of the waste salt is NaCl, which remains stable under the constant temperature treatment at 600 °C.
The FTIR spectra of the samples are shown in Fig. 6. The spectrum of the waste salt exhibits obvious contour double peaks at 3400 ~ 3500 cm −1 and 1620 ~ 1650 cm −1 , which could be assigned to the telescopic vibration of -OH (ν OH ) and in-plane bending vibration of -NH 2 (δ NH ), respectively, while the peaks at 1050 to 1100 cm −1 are characteristic of C-N stretching vibration (ν CN ). Other peaks with weaker signal strength, such as those located at 450-500 cm −1 , 600-620 cm −1 , and 1900 ~ 2100 cm −1 , may be caused by the carbon chain bending vibration (γ CH ) and telescopic vibration of P-H (ν PH ). Therefore, it can be determined that the main impurity component in the waste salt contains -NH 2 group. The peak at 3400-3450 cm −1 after thermal treatment at 600 ℃ is attributed to -OH of adsorbed water (ν OH ).
The impurity components in the waste salt were further studied by HPLC technique with several glyphosate degradation intermediates for comparison, the test results of which are shown in Fig. S1 (Supporting Information). As shown in the results, the waste salt contains no glyphosate or (Aminomethyl)phosphonic Acid (AMPA), but ~ 0.034 wt% of NO 3 − and a small amount of three or more unknown organic impurities. The retention time of these unknown organic impurities is close to that of glycine and sarcosine, revealing glycine, sarcosine or other components with similar structures are possibly included.
Based on the above characterization, conclusions can be drawn as follows: the waste salt originated from the glyphosate mother liquor wastewater mainly consists of NaCl after preliminary CWO treatment with about 1.31 wt% of organic impurities. The main elements contained in the waste salt are Na, Cl, O, C, N, and P, and a certain amount of NO 3 − was also detected, while the other N element mainly exists in the form of -NH 2 group. Glyphosate and AMPA were excluded from the waste salt. The main impurity components are probably oxygenated organic compounds containing -NH 2 , such as glycine, sarcosine or others with similar structures.

Electrode characterization
Highly ordered TiO 2 nanotube array samples prepared by anodic oxidation is widely used in the field of PEC because of their high regulation, large surface area, high stability, and easy fabrication [33,34]. Thus, in this study, TiO 2 nanotube electrode is employed as a typical photoelectrocatalyst for the treatment of water contaminated by salt. Figure 7 shows the morphology and phase characterization results of the TiO 2 electrode. The electrode consists of highly ordered TiO 2 nanotube arrays grown on a Ti substrate, with the nanotube diameter of about 200 nm and nanotube layer thickness of about 5 μm.

Comparison between PEC and PC
PEC generally has higher quantum efficiency than PC because the applied potential on the electrode can promote the separation of the photo excited electron-holes pairs. But the precondition for photoelectrocatalytic application in wastewater treatment is that the wastewater as electrolyte must be of high conductivity. Therefore, it seems that PEC is quite suitable for brine wastewater treatment. The treatment efficiency of waste brine by PC and PEC is compared in Fig. 8. As expected, the TOC degradation rate by PEC is much higher than that of PC within the same treatment time. After photoelectrocatalytic treatment for 50 h, the degradation ratio of TOC by PEC reaches 85%. In contrast, the value for PC is only 16%. Importantly, upon the treatment by PEC, the TOC of the waste salt was reduced to 50 mg L −1 . Furthermore, after 100 h treatment, the TOC of electrolyte was able to meet the quality standard of refined salt in the chlor-alkali industry (TOC < 20 mg L −1 ), as shown in Fig.  S4 (Supporting Information) [35].

Verification of oxidation mechanism
The mechanism of PEC has been widely considered and studied by researchers [36][37][38][39]. They reveal that generally the photoelectrocatalytic process is an oxidation process dominated by hydroxyl radicals. However, it seems that a different mechanism exists on the condition of high salt concentration. Polcaro et al. found that a large number of halogens were generated in the photoelectrocatalytic treatment of brine wastewater with a high content of chloride ion [39]. The main degradation pathways include direct oxidation on anode holes (h + ) and indirect oxidation by •OH or chlorine active substances (such as Cl•, free chlorine (Cl free ), ClO 2 and other chlorine active substances) produced by the reaction of holes (h + ) with Cl − . The photoelectric synergistic effect in PEC is able to improve the oxidation efficiency.
To have an insight into the degradation mechanism in this study, several experiments were carried out to detect the reactive species generated during the photoelectrocatalytic process. To detect •OH, terephthalic acid was added into the electrolyte as a probe molecule, which can react with •OH forming a highly fluorescent product 2-hydroxyterephthalic acid in a rapid and targeted manner [26]. The fluorescence spectra obtained by capturing •OH with terephthalic acid are shown in Fig. 9. No fluorescence signal was detected in the electrolyte of a high Cl − concentration (Fig. 9a), but the signal became obvious when the salt concentration was reduced to 5.8 g L −1 , which proved that •OH could be generated at low Cl − concentrations and inhibited when the salt concentration was too high. The fluorescence intensity as a function of the salt concentration is shown in Fig. 9c. When the concentration of sodium chloride is higher than 4 mol L −1 , the generation of •OH is almost completely inhibited by Cl − .
DMPO was used to detect Cl• as it is able to capture Cl• forming DMPO-Cl with characteristic ESR signals [28,29]. Unfortunately, no obvious free radical signal was detected after photoelectrocatalytic treatment for 5 min, 30 and 120 min (Fig. S5, Supporting Information). Therefore, it could be concluded that PEC cannot produce Cl• in the electrolyte of a high Cl − concentration. In other words, Cl• is not the final product of Cl − react with •OH or h + . As illustrated in previous researches, Cl• can quickly combine with Cl − to form Cl 2 • − [40,41]. These radicals are highly active and will react rapidly with Cl − forming more stable intermediates such as chorine, ClO − , and organochlorinated compounds [40,41]. After PEC for several minutes, the electrolyte was sampled and set without any treatment. At given intervals, the sample was tested with a potassium iodide starch paper. Impressively, it could turn the color of the paper into blue within ~ 8 h, and its oxidability completely vanished with longer setting time. Based on the long lifetime of the oxidative species, it is likely to be ClO − . By using UV-vis spectroscopy technique [42], the generation of ClO − during the PEC process was proved. As shown in Fig. S6 (Supporting Information), the concentration of generated ClO − gradually increased with the proceeding of the PEC process. This fact proves that chloride was oxidized by h + or •OH during the PEC process, which is responsible for the absent of •OH in the studied system. Sodium oxalate was selected as a hole scavenger to clarify the function of h + during the photoelectrocatalytic process [26,30]. With the addition of sodium oxalate, the concentration change of pollutants in the waste salt brine cannot be monitored by TOC technique. Thus, a synthesized NaCl brine wastewater of 270 g L −1 containing 50 mg L −1 methyl orange (MO) was prepared as the electrolyte with or without the addition of 1.5 g L −1 sodium oxalate, for the sake of detecting the concentration change of MO and sodium oxalate by UV-vis spectroscopy and HPLC, respectively. The experimental results in Fig. S7a (Supporting Information) show that the oxidation of MO is significantly inhibited in the presence of sodium oxalate, implying that h + play a critical role in the photoelectrocatalytic process. To exclude the possibility that this phenomenon was caused by the generated oxidants in the bulk electrolyte preferred to oxidize sodium oxalate rather than MO, the same solution was subjected to directly react with different dosages of sodium hypochlorite. As shown in Fig. S7b (Supporting Information), the degradation efficiency of MO showed no obvious difference with or without the addition of sodium oxalate, which justified sodium oxalate is only a scavenge of h + in the studied system. These facts verify that holes (h + ) still play an important role in the photoelectrocatalytic oxidation process even in the case of high salt content. Fig. S8 (Supporting Information) compares the linear sweep voltammetry (LSV) and EIS curves of the TiO 2 sample during PEC process in high concentration electrolytes of NaCl and NaNO 3 . The independence of LSV current density on stirring speed and the typical semi-arc Nyquist plots reveal the PEC process is kinetically dominated by the charge transfer process. It   is worth noting the current density of LSV curves recorded in NaCl electrolyte is higher than that in NaNO 3 electrolyte (Fig. S8a, Supporting Information), which is in consistent with the lower charge transfer resistance of the former (Fig.  S8b, Supporting Information), because the h + can not only oxidize water but also chloride ions. As the concentrations of water and chloride ions are sufficient high, there is no diffusion limitation for them. Therefore, stirring cannot promote the current density, that is, the quantity of active chlorine substances. If the degradation of pollutants is dominated by active chlorine substances but not the h + , stirring should not affect the degradation efficiency. Actually, the degradation of MO was promoted at elevated stirring speed and suppressed by the addition of hole scavenger even at high stirring speeds (Fig. S9, Supporting Information), which again corroborating that h + play an important role in the degradation process.
With stronger stirring, the contact between h + and pollutants was promoted and thus their reaction was accelerated.
According to the research of Tang et al. and the experimental results in this study, the mechanism of PEC in electrolyte with high Cl − content is depicted in Fig. 9d [43]. The degradation of organic impurities by photoelectrocatalytic treatment is mainly attributed to direct oxidation by holes (h + ) or indirect oxidation by active chlorine substances other than Cl•, while the generation of •OH is extremely suppressed by Cl − .

Chlor-alkali electrolysis
To demonstrate that the photoelectrocatalytic technologies are applicable to the resource utilization of waste salt for the chlor-alkali industry, the waste salt solution obtained after photoelectrocatalytic treatment was subjected to chlor-alkali electrolysis. Electrolytes with NaCl concentration of 270 g L −1 were prepared by mixing different ratios of waste salt∕ analytical reagent salt (WS∕ARS) to study the poisoning effect of pollutants on the chlor-alkali electrolysis system. Figure 10a shows the voltage change with time during the electrolysis. The voltage would rise first and then gradually decreased and finally leveled off after 10-h electrolysis, which was recorded as the initial stable voltage E s . With the increase of the proportion of WS, E s gradually increased, and the value in the case of WS∕ARS=100%∕ 0 was nearly 0.1 V higher than that of WS∕ARS=0∕100%. This phenomenon indicated that the impurities in the WS were harmful for the electrolysis, which would increase the voltage during chlor-alkali electrolysis and reduce electrolytic efficiency.
To shed light on the poisoning mechanism of impurities on the electrolysis, EIS study was performed on these electrolytes with different concentrations of WS. As shown in Fig. 10, the Nyquist plots of each sample start with a large semi-circular arc at the high-frequency region and then approach to a straight line at the low-frequency region. The character of these Nyquist plots indicates that the electrochemical reaction was controlled by both the polarization and diffusion kinetics. It should be noting that the diffusion limitation occurs in the membrane for the ion diffusion in the solid ion-exchange membrane is much slower than that in liquid electrolyte [44], which is proved by the independence of LSV current density on stirring speed (Fig. S10, Supporting Information). The EIS plots could be described by the typical equivalent circuit shown in Fig. 10. In this model, R s is the solution resistance, Z w is called Warburg resistance representing the resistance of ion diffusion∕transport, R ct is the charge transfer resistance at the solid-electrolyte interface, and CPE is the constant phase element (defined as Z = [T(iω) p ] −1 ) generally used to depict a real doublelayer capacitor. With the ratio of WS∕ARS increased, W 1 and R ct increased significantly. This result implied that the WS had a toxic effect and would affect both the electrode and separator. The fitting circuit diagrams and corresponding EIS fitting results of WS/ARS = 100%/0 before and after photoelectrocatalytic treatment in comparison with WS∕ ARS=0∕100% are shown in Fig. 10c and d. The values of each element after fitting are listed in Table 1. The fitting results reveal that both the values of Z w and R ct in the case of WS∕ARS=100%∕ 0 are significantly higher than that of WS∕ARS=0∕100%. Thus, it could be understood that the impurities of the WS were poisonous to both the electrode and membrane. It is speculated that organic impurities may be adsorbed on the electrode surface and would reduce the effective area of the electrode, and penetrate into the gap of the ion membrane, resulting in the local failure of the cation-exchange membrane. As illustrated in Fig. S11 (Supporting Information), the electrochemical active surface area (ECSA) of the electrode measured in WS electrolyte decreased about 7.8% than that in ARS electrolyte. After electrolysis in the WS electrolyte for 30 min, the electrode was taken out and measured again in the ARS electrolyte, and its ECSA was recovered to the initial level. Upon photoelectrocatalytic treatment, most of the organic impurities were degraded, as reflected by the sharp decline in the values of Z w and R ct , which leads to lower polarization potential than that before photoelectrocatalytic treatment when the salts were used as a raw material for chlor-alkali electrolysis (Fig. S12, Supporting Information). These findings demonstrate that the photoelectrocatalytic process is an efficient technology for the resource utilization of WS.

Conclusion
In this study, a strategy of photoelectrocatalytic treatment is proposed for the resource utilization of the waste salt stream released by the glyphosate industry. The main component of the waste salt obtained from glyphosate production mother liquor after catalytic wet oxidation (CWO) treatment is NaCl (NaCl content > 95%), and the impurities are mainly inorganic impurities such as sodium nitrate (NaNO 3 ), as well as some organic impurities whose structure is similar to those of glycine and sarcosine, containing -NH 2 group, but AMPA, glycine, and sarcosine are excluded. Such waste salt cannot be directly used as raw material for chlor-alkali  electrolysis, because the organic impurities with a content of ~ 1.31 wt% still have an obvious toxic effect on the electrode and diaphragm. A self-made TiO 2 nanotube electrode was used for advanced photoelectrocatalytic treatment of the waste salt. The treatment efficiency is significantly higher than that of PC under the same conditions. After deeply treated by PEC, the TOC of the waste salt was reduced to less than 20 mg L −1 , which met the requirement for refined salt in the chlor-alkali industry. Under the condition of high Cl − concentration electrolyte, the photoelectrocatalytic degradation of organic impurities was realized mainly through direct oxidation by holes (h + ) on the surface of TiO 2 nanotubes and indirect oxidation by chlorine active substances such as free chlorine and ClO 2 other than •OH and Cl•. EIS analysis revealed that the organic impurities in the waste salt was poisonous to both the electrode and membrane when the waste salt was used as a raw material for chlor-alkali electrolysis, leading to an increase in the voltage and a decline in electrolysis efficiency. After photoelectrocatalytic treatment, most of the organic impurities were removed so that the waste salt could be utilized for chlor-alkali electrolysis. The current research offers a possible method for the resource utilization of waste salt from chemical industries.