Indirect photodegradation of ofloxacin in simulated seawater: Important roles of DOM and environmental factors

Dissolved organic matter (DOM) plays a non-negligible role in the indirect photodegradation of organic contaminants. This research investigated the roles of DOM and the environmental factors (salinity, pH, NO3− , and HCO3− ) in the indirect photodegradation of ofloxacin (OFX) in simulated seawater. Results showed that DOM can significantly accelerate the indirect photodegradation of OFX, and 1O2 and 3DOM* were the main reactive intermediates (RIs) that could promote the indirect photodegradation of OFX. Fluorescence excitation–emission matrix spectroscopy–parallel factor analysis (EEMs-PARAFAC) was used to divide DOM into four fluorescence components. The indirect photodegradation rate of OFX was affected by DOM structure, and terrigenous DOM usually produced more RIs to promote the indirect photodegradation of OFX. Increased salinity significantly promotes the indirect photodegradation of OFX, while increased NO3− concentration had no effect on the OFX indirect photodegradation. pH affected the formation of RIs and the structure of OFX, thereby affecting the indirect photodegradation of OFX. The indirect photodegradation rate of OFX increased in the HCO3− solution, which is due to the formation of carbonate radical (CO3− ). This study is essential in understanding the role of DOM in OFX indirect photodegradation and providing a novel insight into the fate, removal, and transformation of OFX.


Introduction
Pharmaceuticals and personal care products (PPCPs) are considered new organic pollutants, including various antibiotics, biopharmaceuticals, and fungicides, which have raised scientific and public concern over the past decades. PPCPs are usually detected in drinking water, sewage, and wastewater treatment plants (WWTPs) due to their universal consumption, diversity, and unique chemical properties (Yang et al., 2017). Generally, most antibiotics and their metabolites ingested by humans and animals are excreted with urine and feces and subsequently spread to the environment. The concentrations of sulfonamides, fluoroquinolones (FQs), and macrolide compounds in wastewater from animal production and pharmaceutical companies are as high as 200, 98, and 5.6 mg L −1 , respectively . It has been demonstrated that antibiotic residues in the environment can lead to the evolution of new drug-resistant bacteria and increase bacterial resistance, which ultimately pose threats to aquatic ecosystems and human health. FQs are a family of synthetic antimicrobial agents for both humans and animals, which have been detected at high concentrations in aquatic environments in Europe, America, Australia, and Asia (Al Aukidy et al., 2012). Ofloxacin (OFX), one of the FQs, can be detected in a variety of natural water and even in drinking water and groundwater. It has been detected at a maximum of tens of mg L −1 in WWTPs and at hundreds of ng L −1 in surface water (Danner et al., 2019;Ding et al., 2020). Due to animal husbandry and aquaculture having produced a large amount of industrial and aquacultural waste, notable antibiotic residues were transported to coastal areas through riverine inputs and precipitation (Fu et al., 2003;Zhang et al., 2013). The ocean has become an important sink of many antibiotics. Hence, it is important to understand the behavior of OFX in seawater for its environmental risk assessment.
The removal of OFX from aquatic environments has been extensively researched, such as physical, chemical, and biological methods (Tian et al., 2021). Generally, photodegradation is considered the main degradation pathway of OFX in natural water. It is well known that direct photodegradation involves energy and electron transfer after dissolved organic matter (DOM) absorbs photons (Du et al., 2014). Indirect photodegradation refers to the process in which some photosensitizers generate reactive intermediates (RIs) to degrade organic contaminants (Yang et al., 2013). DOM is a diverse and complex mixture that acts as a photosensitizer in the natural environment. Generally, the photodegradation of DOM results in the formation of RIs, such as hydroxyl radical (·OH), singlet oxygen ( 1 O 2 ), and the reactive triplet state(s) of DOM ( 3 DOM*), which can facilitate the indirect photodegradation of OFX (Burns et al., 2012). However, DOM can also restrain the photodegradation process through light attenuation, quenching RIs, and back-reduction mechanisms (Wu et al., 2021). Numerous studies have demonstrated that the ability of DOM to generate RIs was dependent on the source of DOM and the content of functional groups, resulting in different effects on the photodegradation of organic pollutants (Guerard et al., 2009;Liu et al., 2021). For example, DOM from soil usually has high aromaticity and degree of humification (Xu et al., 2018), which can produce high concentrations of RIs to promote the photodegradation of organic pollutants. In contrast, DOM extracted from seawater contains lower aromatic C═C and C═O functional groups and exhibits lower steady-state concentrations of RIs compared with freshwater DOM, thus having less effect on the photodegradation of organic pollutants (Wang et al., 2019). Therefore, it is necessary to explore the effect of DOM components on the indirect photodegradation of OFX. Excitation-emission matrix spectroscopy (EEMs) can distinguish different classes of materials of terrestrial, autogenous, and anthropogenic origin by concatenating continuous emission spectra of a range of excitation wavelengths (Coble et al., 1998). The technique of EEMs combined with parallel factor analysis (PARAFAC) has been used for evaluating the composition, origin, and fate of DOM (Gueǵuen et al., 2011). Moreover, ultraviolet-visible spectroscopy (UV-Vis) technology can be used to reflect the degree of humification and aromaticity of DOM. Therefore, we used EEMs-PARAFAC and UV-Vis to study the relationship between DOM components and OFX indirect photodegradation.
Other environmental factors in seawater also can affect the generation of RIs, thus having an impact on the photodegradation of organic pollutants. It has been demonstrated that NO 3 − , HCO 3 − , salinity, and pH are crucial factors in seawater to influence the indirect photodegradation of organic pollutants (Bai et al., 2018;Wang et al., 2021). NO 3 − can produce ·OH through direct photolysis under solar radiation to affect the photodegradation of organic pollutants, and the quantum yield of ·OH was affected by pH (Bodrato and Vione, 2014). In natural water, HCO 3 − can react with ·OH to influence the formation of CO 3 − , which is a selective oxidant and plays an important role in aquatic oxidation reactions. The reaction between CO 3 − and organic pollutants mainly occurs through electron transfer or hydrogen extraction mechanism (Liu et al., 2018). Additionally, pH and salinity can affect the indirect photodegradation of organic pollutants by affecting the generation of RIs or the structure of organic pollutants.
Previous studies have reported that the indirect photodegradation of OFX in freshwater accounted for more than 50% of the observed photodegradation rate (Cheng et al., 2021). Lam and Mabury (2005) reported that the photosensitive reaction or photooxidant produced by DOM photolysis played a great role in the photodegradation of OFX in synthetic field water. However, the effect of DOM components on the indirect photodegradation of OFX is still unknown, and there are no data available on the indirect photodegradation of OFX in seawater. It limited the research about the mechanism of photodegradation of organic pollutants. In this study, we investigated the indirect photodegradation of OFX and environmental factors in simulated seawater, especially the roles of DOM in simulated seawater on the indirect photodegradation of OFX. This study aims to 1) explore the roles of DOM and RIs ( 1 O 2 , ·OH, and 3 DOM*) in the indirect photodegradation of OFX, 2) investigate the effects of DOM components on the indirect photodegradation of OFX, and 3) study the effect of other environmental factors (pH, salinity, NO 3 − , and HCO 3 − ) in DOM solutions on the indirect photodegradation of OFX. It provides an important theoretical basis for the scientific assessment of the fate and environmental risk of OFX.

Photochemical experiments
The XPA-7 merry-go-round photoreactor (Xujiang Electromechanical Plant, Nanjing, China) is equipped with a 1,000-W xenon arc lamp to perform photochemical experiments. The filter film was wrapped around the lamp to decrease the irradiation below 320 nm to reduce the direct photodegradation of OFX. To eliminate the influence of thermal and hydrolysis processes, the degradation rate of OFX was determined under dark conditions. The OFX initial concentration was 10 mM, and the pH of the OFX stock solution was controlled to 8.0 ± 0.1. The light intensity was maintained at 200 W m −2 from 320 nm to 800 nm, and the temperature was maintained at 25°C ± 1°C. The configured DOM solutions were irradiated in a photochemical reactor for 24 h, and then the concentration of OFX was determined.
FFA (20 mM) and IPA (10 mM) were added to OFX solutions containing JKHA to scavenge 1 O 2 and ·OH in ultrapure water to explore the self-sensitization of OFX. IPA (10 mM), indole (5 mM), THF (10 mM), and SA (20 mM) were used to scavenge ·OH, 1 O 2 , CO ·− 3 , and 3 DOM*, respectively. The JKHA solutions with different salinity were obtained by diluting IAPSO standard seawater to explore the effect of salinity on the indirect photodegradation of OFX. The dilute solutions of NaOH and HCl were used to adjust the pH to experimental values (5.0, 7.0, 9.0, 11.0), and different concentrations of nitrate (0-40 mM) and bicarbonate (2 mM) were prepared in JKHA solutions to explore the impacts of some common irons on the indirect photodegradation of OFX. During irradiation, a 2-ml sample solution was taken regularly and transferred to an amber liquid phase vial.
The light screening factor (S l ) and total light screening coefficient (f) were used to correct the light screening effects (Eqs 1, 2). S l = 1 − (10 −(a l +e l ½OFXl) ) 2:303(a l + e l ½OFX)l , (1) where a l is the light attenuation coefficient (cm −1 ), a l = 2.303A/l, A is the absorbance, e l is the molar absorptivity (L mol −1 cm −1 ), l is the light path length (cm), and [OFX] is the concentration of OFX (mol L −1 ). The indirect photodegradation rate of OFX is calculated as follows: where k obs(DOM) represents the observed photodegradation rate of OFX with the addition of DOM, k obs(non-DOM) is the observed photodegradation rate in ultrapure water, and k ind is the corrected indirect photodegradation rate of OFX.

Determination of RI content
RIs are commonly analyzed with probe components (PCs) in aquatic environments. FFA was used as a PC to detect the concertation of 1 O 2 . SA was used as a quantitative probe to determine the steady-state concentration of 3 DOM*. Terephthalic acid (TA) was used to determine the concentration of ·OH.
The concentration of 3 DOM* was measured by determining the concentration of SA. The removal rate of 3 DOM* by SA was obtained by dividing the rate of SA cis-trans isomer formation by 0.18, which is the yield of SA. The concentration of PC is divided by the rate constant of sorbic acid, and the probe concentration was plotted according to Eq. (4).
where R 3 DOM * represents the formation rate of 3 DOM*, k′ s is the reaction rate constants of quenchers with the triplets, and k SA,3 DOM * is the reaction rate between SA and 3 DOM*, which Timko et al. (2014) have already calculated as (3.35 ± 1.0) × 10 9 M −1 S −1 . Formation rates of triplets were calculated as the inverse of the slope, while k′ s and steady-state concentrations of 3 DOM* ([ 3 DOM*] ss ) (in the absence of probe) were determined as follows: The steady-state concentration of ·OH in DOM solutions was determined through the addition of TA (600 mM). The concentration of ·OH can be obtained by measuring the concentration of HOTA (Eqs 7, 8). Samples were taken at different time periods, and the concentration of HOTA was determined by HPLC.
where k ·OH,TA = 3.3 × 10 9 M −1 S −1 and g is the yield of HOTA, which is 0.35.
FFA was routinely used as the PC of 1 O 2 in determining the concentration of 1 O 2 . The DOM solution containing FFA (20 mM) was irradiated, and the steady-state concentration of 1 O 2 was determined by measuring the concentration of FFA (Eqs 9, 10).

Analytical method
The concentration of OFX was determined by an HPLC instrument (1100 series, Agilent Technologies, Santa Clara, CA, USA) equipped with UV-visible and fluorescence detectors and Agilent PLRP-S C18 column (4.6 × 150 mm, 4 mM). The flow rate, detection wavelength, mobile phase, and injection volume are shown in Table S1.
The EEMs of DOM were measured using a fluorescence spectrophotometer (Fluorolog3-11) equipped with a 450-W xenon arc lamp. The excitation wavelength of EEMs was from 240 to 480 nm, the emission wavelength was 250-580 nm, the interval of excitation and emission wavelengths was 5 nm, the integration time was 0.05 s, and the scanning speed was 1,200 nm min −1 . The slit width of excitation and emission was 5 nm. Before detection, the samples were all filtered through a 0.22-mM polyether sulfone membrane. The 3D Delaunay interpolation method was used to eliminate the scattering peaks of the fluorescence spectrum. The fluorescence spectrum was calibrated against the quinine sulfate standard, and the data were normalized to the unit of quinine sulfate (QSU).
The humification index (HIX) was obtained by dividing the emission intensity at 435-480 nm by the integral intensity at 300-345 nm at an excitation wavelength of 255 nm, and it was used to characterize the humification degree and source of DOM (Ohno and Technology, 2002). Specific UV absorbance (SUVA 254 ) was defined as the UV absorbance at 254 nm divided by the DOC concentration, which strongly correlated with DOM aromaticity (Weishaar et al., 2003). All data were measured by fluorescence spectrophotometer and analyzed by PARAFAC in MATLAB R2012a software.
SPSS 26.0 and Canoco 5.0 were used to analyze the correlation of the k ind , DOM components, HIX, SUVA 254 , and the concentrations of RIs. The indirect photodegradation experiments of OFX were performed in parallel three times, and the error was controlled within 5%.

Direct photodegradation of OFX
We found no loss of OFX concentration under dark conditions for 24 h, indicating that hydrolysis and biodegradation played negligible roles in the degradation of OFX. The photodegradation rate of OFX in ultrapure water followed pseudo-first-order kinetics, and OFX was degraded by approximately 27.7% after 24 h under simulated sunlight irradiation at pH 8.0 ( Figure S1A). We found that there was an absorption of OFX between 320 and 400 nm ( Figure S1B), indicating that the direct photodegradation of OFX was not completely deducted. Therefore, we used f to correct the indirect photodegradation of OFX. Niu et al. (2016) presented that the photodegradation process mediated by triplet states of pollutants and their induced reactive oxygen species was considered self-sensitizing photodegradation. Previous studies demonstrated that ·OH and 1 O 2 might play important roles in the self-sensitized photolysis of acebutolol (Ge et al., 2010;Ye et al., 2022). Zhu et al. (2014) reported that phenylarsonic compounds were mainly photodegraded by selfsensitization, starting with the formation of the excited tripletstate molecules. In our study, quenching experiments were performed to explore whether ·OH and 1 O 2 affect the selfsensitized photodegradation of OFX, for which the results are shown in Figure 1. The addition of IPA/FFA significantly inhibited the photodegradation of OFX. These results meant that the self-sensitized photodegradation of OFX might be conducted through ·OH and 1 O 2 .

Roles of DOM
DOM was the main form of organic matter in the aquatic environment, and it can combine with metals and organic pollutants, affecting the distribution and fate of organic pollutants. Meanwhile, photolysis of DOM can produce RIs to promote the indirect photodegradation of organic pollutants (Eqs 11-17) (Burns et al., 2012).
DOM· ! ·OH : In this study, the indirect photodegradation of OFX was investigated in different DOM solutions. The proportion of indirect photodegradation in OFX photodegradation is c a l c u l a t e d i n T a b l e S 2 . I n c o n t r a s t , t h e i n d i r e c t photodegradation rate of OFX increased irregularly with the increasing concentration (Figure 2A). It may be due to DOM having a dual effect on the indirect photodegradation of OFX. DOM can promote the indirect photodegradation of OFX by producing more RIs with increasing DOM concentration (Bai et al., 2018). Additionally, DOM can also inhibit indirect photodegradation by scavenging or reducing RIs. The combination of the stimulatory and inhibitory effects led to an irregular change in the rate of indirect photodegradation of OFX.

Roles of reactive intermediates
Many studies proved that indirect photodegradation was mediated by RIs derived from DOM. Wang et al. (2018) reported that three sulfonamide antibiotics were degraded by 3 DOM*induced indirect photodegradation. ·OH-mediated oxidation made more contributions to the indirect photodegradation of atrazine and S-metolachlor (Drouin et al., 2021). To further probe the effect of RIs on the indirect photodegradation of OFX, different scavengers were added to the OFX solutions containing JKHA. IPA was used as a scavenger to eliminate ·OH (Ye et al., 2022), and the roles of 3 DOM* and 1 O 2 were explored by the addition of SA and THF, respectively (Mateus et al., 2000;Han et al., 2017). As can be seen from Figure 2B, IPA did not significantly inhibit the indirect photodegradation of OFX compared to other scavengers, suggesting that the effect of ·OH was minor on the indirect photodegradation of OFX. In contrast, the indirect photodegradation rate was significantly reduced with the addition of SA and THF, indicating that 3 DOM* and 1 O 2 might dominate OFX indirect photodegradation. According to the indirect photodegradation rates of OFX with the addition of different quenchers, the contribution of RIs is shown in Table S3.

Roles of DOM fluorescent components
There were prominent different structural and compositional variability among DOM from different sources. Compared with DOM extracted from seawater, DOM from mariculture areas contained higher humic-like substances and aromatic and carbonyl structures, which exhibited a higher concentration of 3 DOM* to promote the photodegradation of sulfonamide (Wang et al., 2018). Therefore, EEMs-PARAFAC analysis was used to divide DOM into four different fluorescent components to investigate the roles of DOM components in OFX indirect photodegradation ( Figure 3; Table S4). The maximum Ex/Em peak of component 1 (C1) was 420/500 nm. The spectra of C1 were the same as those reported in earlier studies (Yamashita et al., 2015). This component can be characterized as aromatic and high-  Observed photodegradation rate of OFX. The data were presented as mean; error bars denote standard deviations (p< 0.05, n = 3). OFX, ofloxacin.
molecular-weight terrestrial humic-like materials (Kowalczuk et al., 2010). Component 2 (C2) exhibited broad excitation and emission maxima at 370 and 460 nm, which were typically observed as humic-like in the terrestrial environment (Quang et al., 2016). Component 3 (C3) had a maximum fluorescence peak at Ex/Em = 460/520 nm. The fluorescence peak position of C3 was similar to the reported soil fulvic acid known as the previous commercially available "Contech" FA (emission maxima occur at 492-521 nm, corresponding to excitation at 350-455 nm) (Lochmueller and Saavedra, 1986). Component 4 (C4) showed the Ex/Em wavelength of the fluorescence peak at 320/440 nm. Coble et al. (1998) found a fluorescent component that had an excitation maximum of 320 nm and an emission maximum of 420 nm in the Arabian Sea. The peak was considered to be present in visible humic-like. DOM with higher molecular weight may contain higher aromaticity, and the EEM peak occurred as a red-shift as molecular weight increased (Wei et al., 2016). Therefore, the molecular weights of the four components might be C3>C1>C2>C4. The maximum Ex/Em wavelengths of C3 might be due to increasing molecular weight and aromaticity (Baker and Inverarity, 2004). Additionally, we determined the fluorescence intensities of the four DOM components at 0 and 24 h, respectively (Table S5). It showed that the JKHA solution had the max total fluorescence intensity at the same DOC concentration, followed by SRFA and SRHA. After irradiation, the fluorescence intensity of C1, C2, and C3 decreased, else than C4. It may be because C4 cannot be excited effectively on account of the irradiation below 320 nm being the cutoff.
The proportion of indirect photodegradation in OFX photodegradation is calculated in Table S2. The indirect photodegradation of OFX accounted for 52.5%-95.4% of the JKHA solution. However, in SRFA and SRHA solutions, the indirect photodegradation of OFX only accounted for 35.6%-62.7% and 30.5%-80.1%, respectively. The results indicated that the indirect photodegradation was affected by the source and characteristics of DOM. To further research the mechanism of OFX indirect photodegradation in the presence of DOM solutions, we used PC to determine the steady-state concentrations of RIs in DOM solutions. The concentration of RIs increased with the DOM concentration (Table S6). Furthermore, the concentration of RIs in the JKHA solution was significantly higher than in other DOM solutions. It indicated that the JKHA solution can generate more RIs than other DOM solutions to promote the indirect photodegradation of OFX.
Additionally, Pearson's correlation analysis was carried out to explore the relationship between the DOM fluorescence intensity and the concentration of RIs (Table 1). The four fluorescence components had significant correlations with the concentration of RIs, suggesting that the photolysis of the four DOM components can generate RIs. It might contribute to the indirect photodegradation of OFX. The ability of the four fluorescent components to produce RIs can be evaluated as C1>C3>C4>C2. Meanwhile, C3 and C1 were the two main components in the JKHA solution (Table S5). Figure 2B pointed out that 1 O 2 and 3 DOM* were the main RIs to promote the indirect photodegradation of FIGURE 3 Fluorescence signatures of the four identified EEMs-PARAFAC components. EEMs-PARAFAC, fluorescence excitation-emission matrix spectroscopy-parallel factor analysis.
OFX. Therefore, the indirect photodegradation of OFX may be the fastest in the JKHA solution because there were high contents of C1 and C3, which can produce more 1 O 2 and 3 DOM*. The correlations among the OFX indirect photodegradation rate and four DOM components are shown in Table 1. The correlation order of the four DOM components was C1>C3>C2>C4, which indicated that C1 and C3 made more contributions than C2 and C4 to the indirect photodegradation of OFX. It confirmed that the large-molecularweight DOM contributes the most to the indirect photodegradation of OFX.
HIX and SUVA 254 were used to further explore the influence of DOM structure on OFX indirect photodegradation. SUVA 254 was significantly correlated with the concentrations of 1 O 2 and 3 DOM* (Table 1), indicating that highly aromatic DOM can produce more RIs. The aromatic structure was the main chromophore of DOM, which can absorb solar radiation, produce 1 DOM* in natural environmental water, and act on the formation of 3 DOM* (Boyle et al., 2009). The correlation between the indirect photodegradation rate of OFX and HIX is shown in Table 1, which indicated that terrestrial DOM with high humification contributed more to the indirect photodegradation of OFX. In Figure 2A, the OFX indirect photodegradation rate in the JKHA solution was prominently higher than that in other DOM solutions. This may be due to the higher SUVA 254 and HIX values of JKHA, which had a higher aromatic structure when compared to SRHA and SRFA. Therefore, higher concentrations of RIs can be generated in the JKHA solution to promote the indirect photodegradation of OFX.
Moreover, redundancy analysis (RDA) was conducted between DOM fluorescence components, HIX, and SUVA 254 with RIs, k ind . The RDA results revealed that the composition and properties of DOM can explain the variation in the database (96.53%) ( Figure  S2). HIX, SUVA 254 , 3 DOM*, and 1 O 2 were positively correlated with the indirect photodegradation rate of OFX. The finding was in accordance with the results of Pearson's correlation analysis.

Effects of environmental factors on the OFX indirect photodegradation 3.3.1 Effect of salinity
Environmental factors in seawater were important factors affecting the indirect photodegradation of organic contaminants. Trovo et al. (2009) studied the photodegradation phenomenon in freshwater and seawater, discovering that the photodegradation rate of sulfamethoxazole decreased after 7 h in simulated seawater. Sirtori et al. (2010) found that trimethoprim had a faster photodegradation rate in freshwater while showing a longer photodegradation time in seawater. de Bruyn et al. (2012) reported that it made no difference in the photodegradation rate of phenanthrene in seawater and freshwater. In contrast, the photodegradation rate of acetaminophen increased with the increase in salinity (Bai et al., 2018). In this study, the indirect photodegradation of OFX at different salinities in the JKHA solution was explored. Figure 4A shows that when the salinity increased to 35‰, the OFX indirect photodegradation rate was significantly enhanced, and the indirect photodegradation rate at S = 35‰ was approximately two times that at S = 5‰. Previous studies pointed out that the photodegradation of organic pollutants was affected by ionic strength, metal ions, and halogen ions in seawater . Firstly, Fe 2+ was contained in seawater, and Fe 2+ can react with H 2 O 2 to convert many organic compounds into inorganic compounds, which can effectively remove some refractory contaminants. Secondly, higher ionic strength can also reduce the attenuation rate of 3 DOM*, resulting in the promotion of the indirect photodegradation of OFX initiated by 3 DOM*, ·OH, and 1 O 2 . Thirdly, some halogen ions can produce halogen active radicals to react with electron-rich chromophores to accelerate the removal of antibiotics (Eqs 18, 19) (Lin et al., 2013). Finally, high salinity can increase the number of photolysis electrons in the hydrated DOM, thus increasing the  (Anastasio and Newberg, 2007).

Effect of pH
pH can affect the formation of RIs and the structure of pollutants to influence the photodegradation of organic pollutants. Previous studies found that the degradation of OFX by bimetallic doped catalysts exhibited high pH dependence (Tian et al., 2021), while the indirect photodegradation of OFX at different pH in simulated seawater was rarely studied. To explore the influence of pH on the indirect photodegradation of OFX in simulated seawater, we determined the indirect photodegradation rate of OFX from pH 5.0 to 11.0 in the JKHA solution. The indirect photodegradation rate of OFX was complex at different pH values ( Figure 4B). When the pH increased to 7.0, the indirect photodegradation rate was nearly three times that at pH 5.0. The pK a1 and pK a2 values of OFX have been reported to be 6.1 and 8.3, respectively (Luo et al., 2019). The existing forms of OFX changed from cation to zwitterion when pH was from 5.0 to 7.0. Fluoroquinolone makes it difficult for the defluorination process to occur under acidic conditions (Albini and Monti, 2003). Moreover, the main ionizable groups of OFX were the piperazine ring and the carboxyl. The electron cloud density was relatively high in the presence of zwitterion, and the piperazine ring of OFX was susceptible to electrophilic attack, resulting in a higher indirect photodegradation rate. However, when pH was in the range of 7.0 to 9.0, the decrease in electron cloud density led to a decrease in electrophilic attack, resulting in a decrease in the indirect photodegradation rate (Zhu et al., 2020). The indirect photodegradation rate increased continually at pH 11.0, and it was likely due to the high pH enhancing the role of 1 O 2 and ·OH and facilitating the indirect photodegradation of OFX (Bai et al., 2018).

Effects of nitrate and bicarbonate
Many studies have demonstrated that nitrate can produce ·OH to facilitate the photodegradation of organic pollutants in natural water (Zepp et al., 1987;Warneck and Wurzinger, 1988). The concentration of nitrate was approximately from 10 −5 to 10 −3 M in natural water, which was affected by the geographical environment and human activities. In this study, nitrate-induced indirect photodegradation of OFX was explored in a solution containing JKHA under simulated irradiation. The indirect photodegradation rate fluctuated by approximately 0.033 h −1 with the increase in the concentration of nitrate ( Figure 4C). It demonstrated that increased nitrate concentration had little effect on the indirect photodegradation of OFX. Other studies also discovered the same phenomenon of sulfadiazine and sulfamerazine Tang et al., 2021). John Mack (1999) reported that the direct photolysis of nitrate was related to the formation of ·OH and nitrite, but nitrite was always considered an excellent scavenger for ·OH. The sunlight absorption rate constant of nitrate increased sharply at 300-320 nm (Zepp et al., 1987), and the formation rate of ·OH was 9.1% during irradiation at approximately 320 nm at pH 8.0 (Warneck and Wurzinger, 1988). However, we used a filter to remove irradiation below 320 nm, which limited the production of ·OH. Meanwhile, ·OH contributed less to the indirect photodegradation of OFX according to the quenching experiments described in Section 3.3. Thus, the indirect photodegradation rate of OFX was not affected by the increasing nitrate concentration.
Bicarbonate was the ubiquitous inorganic salt present in natural water, which can produce CO 3 − to promote the photodegradation of organic contaminants (Eqs 25-28) (Yan et al., 2019). It is well known that ·OH can react very rapidly with almost any organic compound, but CO 3 − has high selectivity, and its concentration in the surface water exposed to sunlight was approximately hundreds of times higher than that of ·OH (Canonica et al., 2005). The role of CO 3 − may thus be important in controlling the persistence of various organic contaminants. It found that the indirect photodegradation rate of OFX increased significantly at pH 8.0 after the addition of HCO 3 − ( Figure 4D). In this study, indole, IPA, and SA were added to the HCO 3 − solution containing JKHA to explore the formation of RIs (S = 30‰, pH = 8.0). After the addition of indole, the OFX indirect photodegradation rate was close to that without HCO 3 − addition. It suggested that CO 3 − was the main factor promoting the OFX indirect photodegradation. Then, experiments were performed in the presence of SA and IPA in the HCO 3 − solution to explore the generation mechanism of CO 3 − . The indirect photodegradation rate decreased significantly after adding SA but was lower than that of k ind,OFX+SA ( Figure 4D), indicating that the 3 DOM* was responsible for the formation of CO 3 − . After the addition of IPA, the indirect photodegradation rate of OFX was decreased, suggesting that ·OH was responsible for the formation of HCO 3 − .

Conclusion
In this study, we investigated the effects of DOM and environmental factors on OFX indirect photodegradation. DOM was believed to enhance the indirect photodegradation of OFX due to the generation of RIs. Quenching experiments confirmed the roles of 1 O 2 and 3 DOM* in the indirect photodegradation of OFX in the JKHA solution. DOM was divided into four components by using EEMs-PARAFAC, and terrestrial humic-like components made more contributions to the indirect photodegradation of OFX by producing more RIs. The difference between indirect photodegradation of OFX in different DOM solutions was due to the structure of DOM. The indirect photodegradation of OFX was enhanced with salinity, which may be related to the halogen active radicals and the production of RIs. The indirect photodegradation of OFX by pH was complex because pH can affect both the structure of OFX and the production of RIs. NO 3 − had no significant effect on the indirect photodegradation of OFX, while HCO 3 − promoted the indirect photodegradation of OFX, and CO 3 − may be the main RIs responsible for OFX indirect photodegradation in HCO 3 − solution. This study was important for assessing the ecological risk and environmental fate of OFX.

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 authors.

Author contributions
SL: conceptualization, methodology, formal analysis, data analysis, and writing of the original draft. ZC: project administration and funding acquisition. YB: project administration, funding acquisition, and writing-review and editing. DD: investigation and visualization. JY: conceptualization and methodology. RS: project administration and writing-review and editing. KQ: project administration and funding acquisition. All authors contributed to the article and approved the submitted version.