Degradation of 17 Benzodiazepines by the UV/H2O2 Treatment

Benzodiazepines are one group of psychoactive drugs widely detected in water environments, and their persistence during conventional wastewater treatment has raised great concerns. Here we investigated the degradation of 17 benzodiazepines in water by UV/H2O2 treatment. The results showed that the UV/H2O2 treatment significantly increased the degradation of 17 benzodiazepines in phosphate buffer solutions at pH 7.0. This can be attributed to the high reactivity of hydroxyl radicals (·OH) towards benzodiazepines with second-order rate constants of 3.48 × 109 M−1 s−1–2.44 × 1010 M−1 s−1. The degradation of alprazolam, a typical benzodiazepine, during the UV/H2O2 treatment was increased with the increasing H2O2 dosage. The solution pH influenced the alprazolam degradation significantly, with the highest degradation at pH 7.0. Water matrix, such as anions (Cl−, HCO3 −, NO3 −) and humic acid, decreased the degradation of alprazolam by UV/H2O2 treatment. Based on the degradation products identified using quadrupole time-of-flight mass spectrometer, the degradation mechanisms of alprazolam by UV/H2O2 treatment were proposed, and hydroxylation induced by ·OH was the main reaction pathway. The degradation of 17 benzodiazepines by UV/H2O2 treatment in wastewater treatment plant effluent and river water was lower than that in phosphate buffer solutions. The results showed that the benzodiazepine psychoactive drugs in natural water can be effectively removed by the UV/H2O2 treatment.


INTRODUCTION
Pharmaceuticals and personal care products (PPCPs) in water environments are a well-recognized concern due to their widespread occurrence and potential harm to environmental organisms Danner et al., 2019;Patel et al., 2019;Lu et al., 2020Lu et al., , 2022. Benzodiazepines, a main class of psychoactive drugs, have pharmacological effects such as sedative-hypnotic, anxiolytic, anticonvulsant, antiepileptic, and muscle relaxant (Cunha et al., 2017). According to the annual report of International Narcotics Control Board (INCB), benzodiazepines have been used in more than 80 countries, and the global manufacture amount of benzodiazepines was over 195 tons in 2019. Most benzodiazepines, which cannot be completely metabolized in the body, are discharged, resulting in their wide occurrences in municipal wastewater, surface waters and seawaters at concentrations of 24.3-95.9 ng L −1 (Subedi and Kannan, 2015;Wu et al., 2015;Fernandez-Rubio et al., 2019). Existing previous studies have found that the conventional water and wastewater treatment processes, such as biological processes, coagulation, filtration and chlorination, are ineffective to remove benzodiazepines, which can lead to their discharge into the receiving waters (Cunha et al., 2017;Fick et al., 2017;Lei et al., 2021). The persistent nature of benzodiazepines may pose a negative impact on aquatic ecosystem. For example, diazepam at low environmental levels can decrease the growth rate of Daphnia magna and increase mortality rates of the young zebrafish (Danio rerio) (Kalichak et al., 2016;Rivetti et al., 2016). The additive effect of benzodiazepines at low concentrations has been confirmed. Benzodiazepines can alter the social behavior and feeding rates of freshwater fish-European perch (Perca fluviatilis) (Brodin et al., 2013;Cerveny et al., 2020). Consequently, it is essential to eliminate the residual benzodiazepines from water.
Advanced oxidation processes (AOPs) have been reported to be highly efficient processes for the degradation of refractory PPCPs (Miklos et al., 2018;Huang et al., 2020;Lee et al., 2020;Wang and Zhuan, 2020;Wu et al., 2020). Some AOPs, such as photo-Fenton, heterogeneous photocatalysis, UV/TiO 2 and UV/ H 2 O 2 have been explored to remove benzodiazepines in water (Bosio et al., 2019;Cunha et al., 2019;Mitsika et al., 2021). The photocatalytic degradation of the selected benzodiazepines (i.e., alprazolam and diazepam) was proved to be effective with a photo-Fenton reaction system (Mitsika et al., 2021). The heterogeneous photocatalysis using synthesized composites based on TiO 2 and activated carbon (TiO 2 /AC) as catalysts under sunlight-simulated irradiation can remove over 97.5% of bromazepam, clonazepam, and diazepam from water (Cunha et al., 2019). The UV/TiO 2 and UV/H 2 O 2 performed better under acidic conditions and removed 60-80% of alprazolam, clonazepam, diazepam, and lorazepam (Bosio et al., 2019). Limited previous studies examined the degradation of two benzodiazepines (i.e., diazepam and oxazepam) by UV/H 2 O 2 treatments (Huber et al., 2003;Kosjek et al., 2012). Our previous studies also showed that the removal of diazepam was significantly enhanced by UV/chlorine and simulated sunlight/chlorine treatments due to the substantial contribution of ·OH (Yang et al., 2018;Yang et al., 2020). However, the kinetics and transformation products for ·OH reaction with other benzodiazepines are still missing.
This work aimed to investigate the degradation mechanism of 17 benzodiazepines by UV/H 2 O 2 treatment in water. The reaction kinetics for 17 benzodiazepines in UV and UV/H 2 O 2 processes were determined, respectively. The experimental parameters such as dosage of H 2 O 2 , solution pH and coexisting constituents (i.e., Cl − , HCO 3 − , NO 3 − , humic acid) were carefully evaluated. The degradation products and reaction pathways for a typical benzodiazepine (alprazolam) were tentatively revealed based on quadrupole time-of-flight mass spectrometer. Finally, the removal of 17 benzodiazepines by UV/H 2 O 2 treatment was also conducted in wastewater treatment plant (WWTP) effluent and river water for practical application.
The reaction solutions were prepared with Milli-Q water (≥18.2 MΩ cm). River water (RW) and municipal WWTP effluent (WW) used in the irradiation experiments were sampled in Guangzhou, China. The detailed characteristics of RW and WW are presented in Supplementary Table S2.

Irradiation Experiments
The UV/H 2 O 2 experiments were performed in a DS-GHX-V photochemical reactor (Doosi Instrument Corporation, Shanghai, China) with a magnetic stirrer, as shown in Supplementary Figure S1. A 20 W low-pressure mercury lamp with an output wavelength of mainly 254 nm was equipped as UV irradiation source. The UV lamp was located in the center of the photochemical reactor surrounded by a quartz cooling jacket, which maintained the solution temperature at 25 ± 1.0°C. Experimental solutions were added in a series of 50 ml quartz tubes with a Teflon-coated stir bar, which were placed on the top of the magnetic stirrer. The rotation speed was 100 rpm/ min. Based on Canonica et al. (2008), atrazine was used as a chemical actinometer to determine the photon fluence rate under UV irradiation, and the value of photon fluence rate was determined to be 28.74 μE m −2 s −1 (Supplementary Text S1).
The kinetic experiments of 17 benzodiazepines by UV/H 2 O 2 treatment were conducted in the 10 mM phosphate buffer at pH 7.0. The initial concentration of each benzodiazepine, pCBA and H 2 O 2 was set at 100 μg L −1 and 100 μM, respectively. The pCBA was simultaneously added in the reaction solutions to quantify hydroxyl radical (·OH) concentration during the UV/ H 2 O 2 treatment. At predetermined time intervals, 1 ml of the reaction solutions were withdrawn and quenched with TBA (10 mM) to measure the residual concentrations of 17 benzodiazepines and pCBA by ultrahigh-performance liquid chromatograph-triple quadrupole mass spectrometry (UPLC-MS/MS) method. All experiments were performed in triplicate with pH variation of below 0.1 unit.
Experiments were also performed to evaluate the effects of experimental parameters, including H 2 O 2 dosage (0-200 μM), solution pH (3.0-11.0), inorganic ions (Cl − , HCO 3 dissolved organic matter (humic acid, HA). A higher initial concentration of alprazolam (3,088 μg L −1 ) and H 2 O 2 dosage (1 mM) was conducted for UV/H 2 O 2 treatment to facilitate the detection and identification of degradation products by ultrahigh-performance liquid chromatograph-quadrupole timeof-flight mass spectrometer (UPLC-QTOF-MS) method. Besides, the above kinetic experiments were also performed in river water and municipal WWTP effluent.

Analytical Methods
The water quality parameters were characterized as follows. The pH, conductivity, dissolved oxygen (DO) and oxidation reduction potential (ORP) were determined using a YSI Proplus multiparameter meter (YSI, United States). The dissolved organic carbon (DOC) was measured using a Shimadzu TOC-V CSH/CSN and TNM-1 analyzer (SHIMADZU, Japan). Total alkalinity was determined through acidic-titration using a pH 4.2 titration end-point (Yang et al., 2018). The H 2 O 2 solution was determined by Jinghua UV1800 UV-Vis spectrophotometry (Shanghai, China) based on ε 40.0 M −1 cm −1 at λ 240 nm.
The 17 benzodiazepines and pCBA were analyzed on a Waters Xevo TQ-S triple quadrupole mass spectrometry equipped with a Waters Acquity ultrahigh-performance liquid chromatograph. A BEH C18 column (2.1 × 50 mm, 1.7 μm) was used for the separation. The (A) buffer solution (5 mM CH 3 COONH 3 + 0.05% HCOOH) and (B) acetonitrile were used for the mobile phase with the gradient elution program. The gradient was programmed as follows: kept in 20% B at initial, increased to 95% at 5 min, decreased to 20% B to 5.5 min and post time was 1 min. The 17 benzodiazepines and pCBA were analyzed by multiple reaction monitoring (MRM) in positive and negative electrospray ionization mode, respectively. Details MRM transitions and collision were shown in Supplementary Table  S1. The method quantitation limits (MQLs), method detection limits (MDLs), quality assurance (QA) and quality control (QC) have been reported in our previous study (Lei et al., 2021).
The degradation products of alprazolam during the UV/H 2 O 2 treatment were analyzed by Agilent 6545 quadrupole time-offlight mass spectrometer equipped with an Agilent 1290 Ⅱ ultrahigh-performance liquid chromatograph. An Agilent Zorbax Eclipse Plus C18 column (3.0 × 150 mm, 1.7 μm) was used for the separation. For positive electrospray ionization (ESI+) analysis, the mobile phase consisted of (A) Milli-Q water with 0.1% formic acid and (B) acetonitrile with 0.1% formic acid at a flow rate of 0.3 ml min −1 . For negative electrospray ionization (ESI-) analysis, the mobile phase consisted of (A) buffer solution (1 mM CH 3 COONH 3 + 0.1% CH 3 COOH) and (B) acetonitrile with 0.1% acetic acid at a flow rate of 0.3 ml min −1 . The gradient was programmed as follows: kept in 10% B at first 2 min, increased to 98% at 21 min, kept in 98% B to 24 min, decreased to 10% B to 24.5 min and post time was 5.5 min. The column temperature was set at 30°C. The Agilent 6545 quadrupole time-of-flight mass spectrometer was performed in ESI + or ESI-mode with Dual AJS source. Details spectrometry conditions were shown in Supplementary Table  S3. The acquired MS and MS/MS data were processed by Agilent

RESULTS AND DISCUSSION
Removal of 17 Benzodiazepines by UV Photolysis and UV/H 2 O 2 Treatment Figure 1 depicts the removal of 17 benzodiazepines by UV irradiation and UV/H 2 O 2 treatment in phosphate buffer solution at pH 7.0. Under direct UV photolysis, 97.8% of chlordiazepoxide was rapidly removed within 8 min. Clobazam and temazepam were degraded 42.6 and 47.5% within 20 min, respectively. For the other 14 benzodiazepines, the removal rates only ranged from 6.8 to 28.6% after 20 min of UV irradiation. Thus, the UV photolysis has a low potential to remove these benzodiazepines in water. Kosjek et al. (2012) also reported that oxazepam cannot be degraded in 90 min UV irradiation and only 8% of diazepam was degraded. The 17 benzodiazepines cannot be degraded by 100 μM H 2 O 2 treatment within 20 min (data not shown). However, the removal of 17 benzodiazepines was dramatically enhanced by UV/H 2 O 2 treatment. The removal rates of 17 benzodiazepines achieved 73.8-100% in 20 min UV/H 2 O 2 treatment. This should be attributed to the high reactivity of hydroxyl radicals (·OH) towards benzodiazepines.

Kinetics of Benzodiazepines Removal
The degradation of 17 benzodiazepines followed pseudo-firstorder kinetics. The observed rate constants were calculated and are presented in Table 1 and Supplementary Figure S2. Table 1 shows the φ UV,BZDs 254nm and k ·OH BZDs for the reaction of 17 benzodiazepines degradations by UV photolysis and UV/H 2 O 2 treatment. Compared to the other 16 benzodiazepines, chlordiazepoxide has a higher quantum yield of 4.67 × 10 −2 mol E −1 , which is rapidly removed under direct UV photolysis. The reason might be attributed to the N 4 located in the nitrogencontaining heterocycles of the chlordiazepoxide, which would rapidly form N 4 -oxide function under UV photolysis (Ouedraogo et al., 2009). The φ UV,BZDs 254nm for the remaining 16 benzodiazepines ranged from 0.51 × 10 −3 to 4.28 × 10 −3 mol E −1 . The ·OH reacted with benzodiazepines in phosphate buffer solutions non-selectively, with the k ·OH BZDs from 3.48 × 10 9 to 2.44 × 10 10 M −1 s −1 . The secondorder rate constant for diazepam reacting with ·OH was determined to be 6.67 × 10 9 M −1 s −1 , which was close to that reported in the previous study (7.20 × 10 9 M −1 s −1 ) (Lopez et al., 2003), and comparable to the QASR predicted value (1.14 × 10 10 M −1 s −1 ) (Minakata et al., 2009;Minakata et al., 2014). Slight differences in the second-order rate constants between benzodiazepines might be due to the different groups on the benzodiazepine ring. For example, alprazolam, estazolam, and midazolam had an imidazole or triazole ring group, which had higher reactivity towards ·OH than other benzodiazepine. Compared to chlorine and ozone, the ·OH showed a higher reactivity to benzodiazepines. The second-order rate constant of diazepam, oxazepam and nordiazepam reacting with HOCl was only 1.2, 8.73, and 0.19 M −1 s −1 , respectively, (Carpinteiro et al., 2017;Yang et al., 2018). The diazepam reacting with O 3 was only 0.75 M −1 Frontiers in Environmental Science | www.frontiersin.org October 2021 | Volume 9 | Article 764841 s −1 (Huber et al., 2003). Thus, the UV/H 2 O 2 treatment could effectively remove benzodiazepines from water. In order to compare the relative contributions of UV photolysis and ·OH reaction during UV/H 2 O 2 treatment, the 17 benzodiazepines were grouped according to the specific value for k UV BZDs and k ·OH BZDs . Supplementary Figure S2 shows the pseudo-firstorder rate constants (k obs ) for the reaction of 17 benzodiazepines by UV/H 2 O 2 treatment. The observed fractions of k UV BZDs and k ·OH BZDs were significantly different for 17 benzodiazepines. For chlordiazepoxide in the group I (k ·OH BZDs /k UV BZDs <1), the direct UV photolysis played a dominant role in UV/H 2 O 2 treatment. For temazepam, clobazam, clonazepam, prazepam, flunitrazepam, nitrazepam, bromazepam, diazepam, and oxazepam in the group II (1 < k ·OH BZDs /k UV BZDs <10), both UV photolysis and ·OH reaction contributed to the degradation.  For lorazepam, midazolam, nordiazepam, flurazepam, alprazolam, clozapine, and estazolam in the group III (k ·OH BZDs /k UV BZDs >10), the ·OH reaction played a dominant role. The findings suggest that the treatment approach can be selected according to the occurrence of benzodiazepines in real water.

Effects of Operational Parameters
Experiments were performed to evaluate the effects of operational parameters and water quality on the degradation of a typical benzodiazepine, alprazolam, during UV/H 2 O 2 treatment.

Effect of H 2 O 2 Dosage
The degradation of alprazolam by UV irradiation, H 2 O 2 oxidation and UV/H 2 O 2 treatment in phosphate buffer at pH 7.0 is shown in Figure 2A. Alprazolam was not degraded by 100 μM H 2 O 2 treatment. In the presence of direct UV irradiation, only 18.3% of alprazolam was removed within 20 min. The removal efficiency of alprazolam by UV/H 2 O 2 treatment at a H 2 O 2 concentration of 25 μM was up to 70.2% within 20 min. The apparent rate constant (k obs ) of alprazolam increased from 1.0 × 10 −3 to 5.5 × 10 −3 s −1 when the H 2 O 2 concentration increased from 25 to 200 μΜ. A linear relationship between k obs and H 2 O 2 concentration (R 2 0.99) was observed, which was due to the enhanced formation of ·OH at higher H 2 O 2 concentrations. Similar results have been reported in degradation of oxytetracycline and tetracycline by UV/H 2 O 2 treatment (Lopez-Penalver et al., 2010;Liu et al., 2016). Figure 2B depicts the effect of solution pH on the alprazolam degradation by UV/H 2 O 2 treatment in phosphate buffer at pH 7.0. Results showed that the degradation of alprazolam was inhibited at both acidic and alkaline conditions, and the highest k obs (3.27 × 10 −3 s −1 ) was observed at pH 7.0. The inhibition at pH 3.0 and 5.0 might be attributed to the strong scavenging effect of H + , H 2 PO 4 2− , and H 2 PO 4 − at acidic conditions (Kong et al., 2016). At alkaline conditions, the selfdecomposition rate of H 2 O 2 increased rapidly (Lee et al., 2020). In addition, hydroperoxide anion (HO 2 − ) could consume ·OH and decreased the steady-state concentrations of ·OH Liu et al., 2016).

Effects of Coexisting Constituents
Effects of anions (Cl − , HCO 3 − , NO 3 − ) and humic acid (HA) that are ubiquitous in water and wastewater on the alprazolam degradation by UV/H 2 O 2 treatment were evaluated. As shown in Figure 2C, adding Cl − (0-5 mM) has a negative impact on UV/ H 2 O 2 treatment. When 5 mM NaCl was added to aqueous solutions, the removal rate of alprazolam decreased by 16.7%. A possible reason was that ·OH could react with Cl − to form reactive chlorine species (Cl 2 − , Cl − , ClHO − ), whose redox potentials are much lower than that of ·OH (Truong et al., 2004;Deng et al., 2013). The addition of HCO 3 − also affected the removal rate of alprazolam. With the increasing concentration of HCO 3 − (0-5 mM), the removal rate dropped gradually. HCO 3 − is a well-known scavenger of ·OH in solutions. HCO 3 − reacted with ·OH to produce carbonate radicals, which are more selective and less reactive towards organic contaminants than ·OH (Xiang et al., 2016;Yang et al., 2019). As exhibited, NO 3 − showed a negligible impact on the removal rate of Frontiers in Environmental Science | www.frontiersin.org October 2021 | Volume 9 | Article 764841 5 alprazolam. With addition of 5 mM NO 3 − in aqueous solutions, the removal rate only decreased 8.7%. This result may be due to the low concentration level of NO 3 − was used in this study. Besides, NO 3 − may have a weak scavenging effect on ·OH. Previous studies reported that low concentration of NO 3 − had a slight impact on 1,4-dioxane degradation by UV/H 2 O 2 treatment (Lee et al., 2020). The addition of humic acid (HA) had an obviously inhibitory effect on the removal rate of alprazolam. The removal rate reduced from 98.7 to 73.8% when HA was added from 0 to 5 mg/L in aqueous solutions. Generally, the HA is the main component of dissolved organic matter (DOM). It has a fast-quenching effect on ·OH and acts as a scavenger in UV/H 2 O 2 treatment. HA can significantly impact the oxidative degradation of micropollutants through electrostatic interactions , photosensitization (Latifoglu and Gurol, 2003) and binding effects . Thus, the possible reasons for our findings were as follows: Firstly, HA can compete with alprazolam for ·OH reaction; Secondly, the high concentration of HA could impede the UV irradiation and reduce the ·OH generation (Xie et al., 2019;Yang et al., 2021).

Products of Alprazolam Degradation
Six degradation products of alprazolam during UV/H 2 O 2 treatment were detected by quadrupole time-of-flight mass spectrometer. The experimental mass, proposed formula, MS/ MS fragments are listed in Table 2, and their chromatographic and mass spectrogram are provided in Supplementary Figures  S3-S6. The TP340-1 (m/z 341.0800) and TP340-2 (m/z 341.0800) were firstly reported as degradation products of alprazolam for UV/H 2 O 2 treatment. Two degradation products with m/z 341.0800 were detected at different retention times 9.375 and 12.006 min, indicating the addition of ·OH group on the benzene ring. But the exact substitution position could not be confirmed by MS/MS fragments. The TP248, TP324-1, TP324-2, and TP324-3 have been reported in the removal of alprazolam by heterogeneous photocatalysis and photo-Fenton oxidation processes (Finčur et al., 2017;Jimenez et al., 2017;Mitsika et al., 2021). According to two characteristic fragment ions m/z 220.0271 and 179.0003 amu, TP248 (m/z 249.0534) might be generated from the removal of phenyl group in alprazolam and the simultaneous addition of a hydroxyl group. The TP324 (m/z 325.0864) was observed in three different retention times 10.215, 11.052, and 13.842 min, respectively. Thus, TP324 (m/z 325.0864) should be ·OH adding in different positions of alprazolam.
Based on the identification of the degradation products, the hydroxylation substitution reaction was the major pathway for alprazolam degradation during UV/H 2 O 2 oxidation process. The ·OH is known as a reactive electrophilic reagent that reacts rapidly and non-selectively with most electron-rich sites of organic pollutants, mainly via H-atom abstraction, addition to olefins, and addition to aromatic compounds. As shown in Figure 3, the degradation of alprazolam could be divided into route A and route B. In route A, hydroxylation occurred at the initial of alprazolam degradation, producing hydroxylated alprazolam, TP324-1, TP324-2, and TP324-3. Further hydroxylation can form multi-hydroxylated byproducts TP340 (m/z 341.0800). In route B, the alprazolam (m/z 309.0919) might cleavage between seven number nitrogen-containing heterocycles and the benzene ring, corresponding to detach the phenyl group. Followed by hydroxylation, hydroxylated byproducts TP-248 (m/z 249.0534) was generated, according to the previous studies (Mitsika et al., 2021).

Degradation of 17 Benzodiazepines in Real Water Samples
The degradation of 17 benzodiazepines by UV/H 2 O 2 treatment was evaluated in the phosphate buffer solutions, river water (RW) and municipal WWTP effluent (WW), as shown in Figure 4. The k obs for the 17 benzodiazepines decreased from 26.9% (bromazepam) to 65.0% (prazepam) in RW and from 46.1% (chlordiazepoxide) to 84.0% (clobazam) in WW, compared to those in the phosphate buffer. The obvious inhibitory effect can be explained by the anions and dissolved organic matter in the RW and WW, which had been demonstrated according to Figure 2 and previous studies Wang et al., 2017;Ribeiro et al., 2019). The WW had the higher DOC concentration (7.2 mg L −1 ), Cl − concentration (78.1 mg L −1 ), NO 3 − concentration (32.1 mg L −1 ), and alkalinity (2.03 mg L −1 ) than that of RW (4.2 mg L −1 DOC, 40.3 mg L −1 Cl − , 22.9 mg L −1 NO 3 − , and 1.28 mg L −1 alkalinity). Consequently, the benzodiazepine psychoactive drugs can be effectively removed by the UV/H 2 O 2 treatment in real water.

CONCLUSIONS
The UV/H 2 O 2 treatment dramatically enhanced the removal of benzodiazepines in aqueous solutions, compared to UV irradiation  Frontiers in Environmental Science | www.frontiersin.org October 2021 | Volume 9 | Article 764841 7 or H 2 O 2 oxidation. This can be attributed to the high reactivity of ·OH reactivity towards benzodiazepines. The determined k ·OH BZDs in phosphate buffer solutions at pH 7.0 ranged from 3.48 × 10 9 M −1 s −1 to 2.44 × 10 10 M −1 s −1 . The water quality parameters, such as H 2 O 2 dosage, solution pH and co-existing species, had a significant effect on the removal of alprazolam by UV/H 2 O 2 treatment. Six degradation products of alprazolam were detected. The hydroxylation was proposed to be the main reaction mechanism in UV/H 2 O 2 treatment. This study demonstrated that UV/H 2 O 2 treatment is capable of degrading benzodiazepines in natural waters.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.