Effects of nitrate and glucose on the formation of chloronitromethane (CNM) under UV/chlorine treatment

In recent years, disinfection by-products (DBPs) which were generated by reactions among inorganic nitrates, natural organic matter and chlorine in water have been paid more attention to in water treatment. Up to now, more than 600 kinds of DBPs, such as trihalomethanes (THMs), haloacetic acids (HAAs) and nitrogen-containing DBPs, had been found in water treatment (Hu et al. 2010). Although the concentration of nitrogen-containing DBPs (N-DBPs) in water was usually lower than the common DBPs of THMs and HAAs, the potential threat of N-DBPs to human health was much greater than THMs and HAAs (Hu et al. 2010). It was reported that N-DBPs could enter the human body through breathing, drinking and skin absorption, resulting in serious risk to human health (Yu & Reckhow 2017). Therefore, research into N-DBPs has become a hot topic in water treatment. Halonitromethanes (HNMs), a representative substance in N-DBPs, have received extensive attention.

As the world's population and food demand increase, the increase in the discharge of wastewater from petroleum, mining, chemical industry and other industries has resulted in the serious pollution of nitrate in water (Babiker 2004; Tyagi et al. 2018). In 2014, 39 groundwater monitoring results in northwest Mexico showed that the highest concentration of nitrate was 46.7 mg/L (Pasten-Zapata et al. 2014). In India, where agricultural and industrial activities were intensive, the nitrate concentration in groundwater reached as high as 630.7 mg/L (Kumari et al. 2013). Teo et al. (2015) studied 101 swimming pools and found that the average concentration of nitrate in the swimming pool was 8.6 mg/L. The situation in China is not optimistic. Liu et al. (2015) showed that the highest concentration of nitrate was up to 150.02 mg/L in 50 groundwater monitoring wells in Jilin Province. Besides, some scholars (Yu & Li 2014) found that the excess ratio of nitrate in more than 200 water samples in Tianjin Province reached 12.06%, in which the highest concentration of nitrate was 118.2 mg/L. It was reported that nitrate could react with chlorine to generate HNMs (Wang et al. 2006); therefore, nitrate as a precursor of HNMs should be widely concerned. In some wastewater, both nitrate and ammonium existed. The concentrations of ammonium and nitrate were 51 and 207 mg/L in mine and mill effluents, respectively, and 325 and 201 mg/L in a fertilizer factory, respectively (Deng et al. 2019). It was noteworthy that nitrogen from ammonia could be transformed to trichloronitromethane (TCNM) under UV/chlorine treatment (Zhou et al. 2020), so ammonia may be also transformed to CNM. Besides, Ca²⁺ in water significantly decreased the formation of DBPs such as trichlormethane (TCM) and dichloroacetonitrile (DCAN) under chlorination treatment (Zhang et al. 2019a). Therefore, NH₄NO₃, NaNO₃ and Ca(NO₃)₂ were selected as representative of nitrates.

UV irradiation as a new disinfection method has been widely used in the disinfection of drinking water and sewage treatment plants due to its characteristic of strong disinfection and safety (Liberti et al. 2003; Li & Blatchley 2009). Clancy et al. (2000) reported that UV irradiation could be effective to inactivate the cryptosporidium and the pathogen resistant to chlorine or chloramine. Moreover, the study reported by Huang et al. (2018) showed that the number of microorganisms was significantly lower in water under UV/chlorine treatment. It should be noted that UV irradiation does not possess residual disinfection capacity (Hijnen et al. 2006); therefore, UV irradiation needs to be combined with chlorine/chloramine.

There are some findings of HNMs formation from nitrate under only chlorine or UV irradiation followed by post-chlorination treatment. Guo et al. (2016a) reported that when water samples from Minhang No. 2 DWTP (MDWTP, Shanghai, China) mixed with nitrate and bromide were treated under UV irradiation followed by post-chlorination, TCNM, bromodichloronitromethane (BDCNM) and tribromonitromethane (TBNM) were detected and there was no significant increase of HNM formation with increasing nitrate concentration. Hong et al. (2015) found that trihalogenated-HNMs were the main HNM species when lake/river water samples collected from Qiangtang River and Tai Lake were treated under chlorine conditions. The study of Lyon et al. (2012) showed that water samples from Utility C, mixed with nitrate (1 and 5 mg N/L), could generate TCNM and TBNM under UV treatment followed by post-chlorination. However, there are few studies aimed at investigating the effects of nitrate on the formation of CNM under UV/chlorine treatment. Ultraviolet irradiation was able to stimulate chlorine or chloramine disinfectant in water with the generation of ·OH and halogen radicals (e.g. ·Cl) (Zhao et al. 2011; Yin et al. 2018; Cheng et al. 2019), and then the natural organic matter in the water reacted with ·OH and halogen radicals (e.g. ·Cl) to form DBPs. Besides, inorganic nitrogen such as ⁠, and in natural water may produce different nitrogen-containing radicals under UV/chlorine treatment, such as ·NO₂, ·NH₂, ·NH₂OO, ·NO and ·NHCl (Li & Blatchley 2009; Shah et al. 2011; Zhang et al. 2015; Guo et al. 2016a); therefore, it is possible that CNM will be generated.

The purpose of this study was to investigate the effects of nitrate on the formation of CNM during UV/chlorine treatment when glucose was used as the carbon source. The effects of nitrate type, initial glucose solution concentration, UV intensity, initial free chlorine concentration and pH on the formation of CNM were explored. To reveal the laws about the formation of CNM better, the changes of nitrogen in different forms from NaNO₃ and NH₄NO₃ were also investigated. The possible formation pathways of CNM formation were speculated. Meantime, water samples collected from a water supply plant and a sewage treatment plant were treated with chlorine or UV/chlorine to verify the laws found in the laboratory, which may provide references for controlling the concentration of CNM under UV/chlorine treatment.

All reactive experiments of CNM were carried out in the self-made double-layer reaction reactor (see Figure 1). The reactor contained a 500 mL quartz glass reactor (inner layer: 20.0 cm length, 9.0 cm diameter; outer layer: 1.0 cm wall thickness) with a condensing water circulation for controlling the temperature of reaction solution (22 ± 2 °C). Low pressure (LP) UV lamps (λ = 254 nm, Hangzhou yaguang lighting Co. Ltd, China) were fixed in the center of the reaction reactor. A magnetic stirrer was arranged at the bottom of the reactor to ensure a uniform mixture of the reaction solution.

In the experiments, all chemicals were analytical grade and used without further purification. All solutions were prepared from ultrapure water (>18 MΩ cm). The solution was controlled for pH 6–8 by 2 mM phosphate buffer. CNM was purchased from Sigma-Aldrich with the purity of 97%. Ammonium nitrate, sodium nitrate, calcium nitrate, sodium chloride, methyl tertiarybutyl ether (MTBE), concentrated hydrochloric acid, disodium hydrogen phosphate and sodium dihydrogen phosphate were purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Sodium thiosulfate was purchased from Comemi Chemical Reagent Co., Ltd. A stock of free chlorine solution was prepared by 5% sodium hypochlorite (NaOCl) purchased from Comemi Chemical Reagent Co., Ltd. Potassium persulfate was purchased from Sigma. Sodium potassium tartrate was purchased from Kelong Chemical Industry. Sodium sulfamate was purchased from Yuanye Biotechnology Co., Ltd. Nessler's reagent was purchased from Aladdin Chemical Trade Co., Ltd (Shanghai, China). CNM was dissolved in MTBE and stored at low temperature and in dark conditions.

A USB2000 UV-VIS-ES Spectroradiometer from Ocean Optics Company was employed to measure the actual radiation intensity of the UV lamp. The actual radiation intensity of the UV lamp (254 nm) was 25.81 μW/cm² (6 W), 40.43 μW/cm² (10 W) and 63.22 μW/cm² (16 W), respectively. A brown volumetric flask (500 mL) was used to prepare the reactive solution for the experiment. The reactive solution was then poured into the reactor in different experimental conditions and a magnetic stirrer rapidly mixed it. The UV lamp was kept stable for approximately 30 min before the experiment. At different time intervals, 5.0 mL samples were withdrawn from the reactor into a 10 mL brown glass. Then, the samples were extracted with 2.0 mL MTBE. 1 mL of the MTBE layer was transferred to a GC vial for analyzing CNM by gas chromatography (GC) system. All the experiments were done in triplicate. The experimental data were the mean values with an overall error of less than 5%.

For the analysis of CNM, 1 mL MTBE was taken to a GC vial from the extracted solution and analyzed using a GC equipped with an electron capture detection (Agilent GC). 1 μL of MTBE sample was injected into a DB-1 column (30 m × 0.32 mm × 5 μm) via an on-column injector with an AS2000 liquid auto-sampler. The GC temperature program consisted of an initial temperature of 50 °C for 5 min, followed by ramping up to 140 °C in 10 °C/min, then ramping up to 280 °C in 20 °C/min. The temperature of the injection port was 235 °C and the temperature of the detector was 280 °C. The injection volume was carried by a high purity nitrogen gas with 1.0 mL/min. The detection limit of CNM was around 0.1 μg/L, and the standard curve of CNM was Area = 925.45C_CNM − 490.83, R² = 0.9996.

The concentration of total nitrogen (TN) was determined by UV spectrophotometry referring to the Water quality-Determination of total nitrogen-Alkaline potassium persulfate digestion UV spectrophotometric method (HJ 636-2012). The concentration of was determined referring to Water quality-Determination of ammonia nitrogen-Nessler's reagent spectrophotometry (HJ 535-2009). The concentration of was determined referring to Water quality-Determination of nitrate-nitrogen-Ultraviolet spectrophotometry (HJ/T 346-2007). The concentration of DON was obtained by subtracting the concentrations of ⁠, and from the concentration of TN (Li et al. 2006). Since the reaction solution in this study always had a large concentration of free chlorine, which made the reaction solution under the condition of strong oxidation, the concentration of could be ignored in the solution.

In the experiments, the formation of CNM under only chlorine and UV/chlorine conditions was measured in two different actual water samples, which were filtered without disinfection from a water supply plant and a sewage treatment plant in Nanjing, China. The water quality index is shown in Table 1. All water samples were first filtered with 0.45 μm microfiltration membrane, and then were stored at 4 °C until use.

If −t_(0.025,2) < t < t_(0.025,2), it indicates that there is no significant difference in the formation of CNM. If |t| > t_(0.025,2), then there is a significant difference in the formation of CNM (t_(0.025,2) = 4.303).

To study the effects of nitrate species on the formation of CNM, NH₄NO₃, NaNO₃ and Ca(NO₃)₂ were selected in the experiments. The results are shown in Figure 2. Under only chlorine treatment, the concentration of CNM produced by three nitrate solutions increased slowly with contact time. The maximum concentration of CNM produced by NH₄NO₃ was only a little higher than that produced by NaNO₃ and Ca(NO₃)₂. At the same contact time, there was no significant difference in the formation of CNM among three nitrate solutions under only chlorine treatment (p > 0.05), except for 3 min between NH₄NO₃ and Ca(NO₃)₂ (p < 0.05). Under UV/chlorine treatment, the maximum concentrations of CNM produced by three kinds of nitrates were higher than those produced under only chlorine treatment. It should be noted that there was no significant difference in the formation of CNM between NaNO₃ and Ca(NO₃)₂ at the same contact time under UV/chlorine treatment (p > 0.05). The maximum concentrations of CNM produced by NH₄NO₃, NaNO₃ and Ca(NO₃)₂ were 23.4, 14.9 and 15.86 μg/L, respectively, which increased by 293.9, 192.2 and 322.9%, respectively, compared with the maximum concentrations of CNM under only chlorine treatment. In general, CNM was rapidly generated and reached its maximum concentration at 6 min, and then began to decrease with contact time under UV/chlorine treatment. It may be due to the photolysis of HOCl/OCl⁻, nitrate and monochloramine (produced by ammonium reacting with hypochlorite) with producing radicals such as ·OH, reactive chlorine species (·Cl, ·ClO, etc.), ·NO₂ and ·NH₂ under UV irradiation treatment (Lyon et al. 2012; Fang et al. 2014; Zhang et al. 2015), resulting in more CNM formation than chlorine treatment. At the same time, HNMs could be degraded by UV irradiation (Fang et al. 2013). After 6 min under UV/chlorine treatment, the degradation rate of CNM exceeded the formation rate of CNM, which resulted in the gradual decline of CNM. It was noted that the gap between the maximum concentration of CNM produced by NH₄NO₃ and that produced by the other two kinds of nitrates was widened under UV/chlorine treatment compared with only chlorine treatment, while the gap between the concentration of CNM produced by NaNO₃ and that produced by Ca(NO₃)₂ was still small. It could be concluded that under UV/chlorine treatment, NH₄NO₃ generated more CNM than NaNO₃ and Ca(NO₃)₂ at the same nitrogen content.

The effects of different concentrations of glucose on the formation of CNM are shown in Figure 3. The trends of the three nitrates were similar at different glucose concentrations under UV/chlorine treatment. Within the initial 6 min, the concentration of CNM rapidly increased to its maximum, and then gradually decreased with increasing contact time. The maximum values of CNM produced by NH₄NO₃ at 3, 6 and 12 mg/L glucose concentrations were 11.63, 18.24 and 23.25 μg/L, respectively, and decreased to 3.38, 5.47 and 8.30 μg/L at 25 min, respectively. The maximum values of CNM from NaNO₃ at 3, 6 and 12 mg/L glucose were 10.50, 16.96 and 18.43 μg/L, respectively, which were lower than those from NH₄NO₃. The maximum values of CNM from Ca(NO₃)₂ at 3, 6 and 12 mg/L glucose concentrations were 9.84, 16.09 and 20.14 μg/L, respectively, which were similar to those from NaNO₃. In terms of the maximum concentrations of CNM, there was a significant difference between NH₄NO₃ and NaNO₃ (p < 0.05), while there was an insignificant difference between NaNO₃ and Ca(NO₃)₂ (p > 0.05). The results indicated that the maximum concentrations of CNM from three nitrate solutions increased with increasing the concentration of glucose (p < 0.05, except for NaNO₃ at 6 and 12 mg/L glucose (p > 0.05)), and NH₄NO₃ was easier to form CNM than NaNO₃ and Ca(NO₃)₂ at the same nitrogen content under UV/chlorine treatment.

The effects of different concentrations of free chlorine on the formation of CNM are summarized in Figure 4. The maximum concentrations of CNM produced by NH₄NO₃ at 40, 60 and 80 mg/L of free chlorine were 24.73, 22.42 and 18.30 μg/L at 6 min, respectively. The concentrations of CNM decreased to 11.86, 11.04 and 7.02 μg/L at 25 min, respectively, which were 52.06, 50.75 and 61.66% lower than the maximum values, respectively. The maximum concentrations of CNM produced by NaNO₃ at 40, 60 and 80 mg/L of free chlorine were 21.10, 15.90 and 12.5 μg/L at 6 min, respectively, which were lower than those from NH₄NO₃. The maximum concentrations of CNM produced by Ca(NO₃)₂ at 40, 60 and 80 mg/L of free chlorine were 18.90, 15.20 and 11.33 μg/L at 6 min, respectively, which were similar to those from NaNO₃. It was worth noting that there was no significant difference in the formation of CNM between NaNO₃ and Ca(NO₃)₂ at the same concentration of free chlorine and the same contact time (p > 0.05), except for 40 mg/L of free chlorine at 1 min (p < 0.05). There was a significant difference in the maximum concentrations of CNM from 40 mg/L of free chlorine and 80 mg/L of free chlorine (p < 0.05). The results indicated that the higher the concentration of free chlorine in water, the less CNM was generated. This was due to the fact that CNM reacted with excess chlorine to produce DCNM and TCNM (the maximum concentrations of DCNM and TCNM increased with increasing the free chlorine concentration, but the data were not given) when the concentration of free chlorine increased, which resulted in the decrease of CNM concentration. Moreover, with increasing the free chlorine concentration, the total concentration of Cl-HNMs increased. The phenomenon might be interpreted by the fact that elevated chlorine resulted in a higher concentration of ·ClO due to ·OH and ·Cl reacting with chlorine species (Bulman et al. 2019), and ·ClO might promote the formation of Cl-HNMs.

Figure 5 displays the effects of different UV intensity on the formation of CNM. The maximum concentrations of CNM produced by NH₄NO₃ under 6, 10 and 16 W UV irradiation at 6 min were 16.51, 18.7 and 23.4 μg/L, respectively, and then decreased to 8.98, 10.12 and 14.67 μg/L at 25 min, respectively, which were 54.65, 43.35 and 44.47% lower than the maximum values, respectively. CNM could be degraded by UV irradiation (Fang et al. 2013) and radicals such as ·OH (Mincher et al. 2010); therefore, the formation and degradation of CNM existed at the same time. The maximum concentrations of CNM produced by NaNO₃ under 6, 10 and 16 W UV irradiation were 7.21, 13.91 and 18.37 μg/L, respectively, which were lower than those from NH₄NO₃. The maximum concentrations of CNM produced by Ca(NO₃)₂ under 6, 10 and 16 W UV irradiation were 7.81, 13.56 and 17.53 μg/L, respectively, which were similar to those from NaNO₃. It should be noted that there was no significant difference in the formation of CNM between NaNO₃ and Ca(NO₃)₂ at the same UV intensity and the same contact time (p > 0.05), except for 16 W UV irradiation at 1, 10 and 15 min (p < 0.05). The results indicated that the maximum CNM concentrations of three nitrate solutions increased with increasing the UV intensity (p < 0.05, except for NH₄NO₃ under 6 and 10 W UV irradiation and NaNO₃ under 10 and 16 W UV irradiation (p > 0.05)), which might be due to more active substances generated including ·NO₂, ·Cl and ·OH with enhancing UV irradiation. As a result, these active groups with strong oxidization could promote the subsequent reactions and generate more CNM.

The actual water from a water supply plant and a sewage treatment plant were selected as the target samples to verify the CNM formation laws under only chlorine or UV/chlorine treatment in the laboratory. Five groups of experiments under only chlorine or UV/chlorine treatment were designed as follows: (1) 500 mL of actual water containing 60 mg/L chlorine; (2) 500 mL of actual water containing 60 mg/L chlorine and 3 mg N/L NaNO₃; (3) 500 mL of actual water containing 60 mg/L chlorine and 6 mg/L C₆H₁₂O₆; (4) 500 mL of actual water containing 60 mg/L free chlorine, 3 mg N/L NaNO₃ and 6 mg/L C₆H₁₂O₆; (5) 500 mL ultrapure water containing 60 mg/L chlorine, 3 mg N/L NaNO₃ and 6 mg/L C₆H₁₂O₆. All the above experiments were carried out at pH = 7 and temperature of 22 °C.

Figure 7 depicts the formation of CNM in the actual water of a water supply plant under (a) chlorine and (b) UV/chlorine treatment. It could be seen from Figure 7 that the maximum concentration of CNM produced by the experiments (4) and (5) was almost the same regardless of actual water under chlorine (see Figure 7(a)) or UV/chlorine treatment (see Figure 7(b)). It indicated that there were few impurities in the actual water of the water supply plant. The maximum concentration of CNM produced by the experiments (2), (3) and (4) was higher than that of the experiment (1) under only chlorine (see Figure 7(a)) or UV/chlorine treatment (see Figure 7(b)). The results showed that the addition of NaNO₃ and C₆H₁₂O₆ increased the yield of CNM. Compared with only chlorine treatment, the peak value of CNM increased under UV/chlorine treatment. This was due to the fact that the oxidation and substitution ability of free chlorine were limited under chlorine treatment, while the radicals (i.e. ·OH, ·Cl, ·ClO and ·NO₂) produced by the photolysis of nitrate and HOCl/OCl⁻ (Lyon et al. 2012; Fang et al. 2014) might promote the formation of CNM. Under UV/chlorine treatment, the concentration of CNM increased first and then decreased, which was consistent with the laws of simulated water samples in the laboratory.

Figure 8 shows the formation of CNM in actual water of the sewage treatment plant under (a) chlorine and (b) UV/chlorine treatment. It could be seen from Figure 8 that the maximum concentration of CNM produced by the experiment (4) was higher than that produced by the experiment (5) regardless of actual water under chlorine (see Figure 8(a)) or UV/chlorine treatment (see Figure 8(b)). This was due to the fact that water samples from the sewage treatment plant contained DOM and nitrogen-containing substances. Therefore, the concentration of CNM produced by the experiment (1) was also higher than that produced by the experiment (5) regardless of actual water under chlorination (see Figure 8(a)) or UV/chlorine conditions (see Figure 8(b)). It was noted that the time to reach the peak concentration of CNM under UV/chlorine conditions was different. It may be related to the complex water quality index of actual water samples from the sewage treatment plant.

Generally, the laws found in simulated water samples were consistent with those found in actual water samples, which could provide references for controlling the concentration of CNM under UV/chlorine treatment in practical production.

The effects of UV/chlorine on organic compounds could be divided into three kinds: UV, HOCl/OCl⁻ and free radicals (Fang et al. 2014; Guo et al. 2016b; Hu et al. 2019; Pan et al. 2019). Previous studies proposed that chlorine could react with NOM in water by substitution, addition and oxidation reactions of C–C, C–O and C = O bonds in organic matter with the generation of DBPs (Watts & Linden 2007). Chlorine could produce ·OH and reactive chlorine species (such as ·Cl) under UV irradiation, and these radicals could react with different kinds of organic substances by means of electron transfer, hydrogenation and hydrogen extraction, which decomposed complex organic substances into simple ones. The study reported by Shah et al. (2011) showed that nitrate might be stimulated by UV irradiation (wavelength less than 250 nm and wavelength between 280 and 320 nm) to produce ·NO₂, and then organic substances in water were further decomposed into a simpler carbon chain structure. As shown in Figure 9, possible reaction pathways of CNM were speculated. Homolytic cleavage of glucose might occur under UV treatment, which could produce carbon-containing radicals. Meanwhile, carbon-containing radicals might be attacked by ·OH through additional reactions to generate substances containing the alcohol hydroxyl group. Due to two alcohol hydroxyl groups attached to a carbon being unstable, the aldehyde group was generated by dehydration. Through C–H bond cleavage, ·O might be transferred to the carbonyl group (Kong et al. 2018). Finally, the carbonyl group could react with nitrogenous radicals to generate CNM.

In order to understand mechanisms of CNM formation under UV/chlorine conditions better, changes of nitrogen in different forms from NaNO₃ and NH₄NO₃ were explored under only chlorine and UV/chlorine treatment. As shown in Figure 10(a), under chlorine treatment, the concentration of in NaNO₃ decreased slightly (<0.1 mg/L), an equal yield of DON was produced and there was no produced. As shown in Figure 10(b), under UV/chlorine treatment, the decline of the concentration of in NaNO₃ accounted for 27% of total nitrogen (TN) and 0.85 mg/L of DON was formed. As shown in Figure 10(c), the concentration of and in NH₄NO₃ decreased slightly under chlorine treatment (<0.1 mg/L), which corresponded to the fact that and could be the source of nitrogen in CNM (see Figure 2). The reaction ability of with chlorine was almost the same as that of with chlorine under chlorine treatment, which corresponded to the fact that the maximum concentrations of CNM produced by NH₄NO₃, NaNO₃ and Ca(NO₃)₂ under chlorine treatment were almost the same (see Figure 2). As shown in Figure 10(d), under UV/chlorine treatment, the concentration of and in NH₄NO₃ decreased by 28 and 35%, respectively, and 1.08 mg/L of DON was produced at the same time. The yield of DON from NH₄NO₃ was more significant than that from NaNO₃ under UV/chlorine treatment, which illustrated that was more likely to produce more CNM than ⁠. Moreover, the decline of and under UV/chlorine treatment was higher than that under chlorine treatment, which corresponded to the fact that more CNM was generated under UV/chlorine treatment compared with chlorine treatment.

Figure 10(c) shows that the concentration of decreased slightly under chlorine treatment for 200 s. According to a previous study (Zhang et al. 2015), the free chlorine, which reacted with ammonia, to ammonia molar ratio was less than 0.8 under our experimental conditions. Combined with the fact that the substance in the solution was mainly monochloramine when the free chlorine to ammonia molar ratio was less than or equal to 1 in the pH range from 6.5 to 8.5 (Li & Blatchley 2009), dichloramine and trichloramine could be neglected in our experiments. Therefore, some reactions containing monochloramine under UV/chlorine treatment should be considered in our experiments. According to the reaction (12) (Zhang et al. 2015), it was speculated that there was a reaction (17) in the experiment and ·NHCl and ·NH₂ produced by the reaction (18, 19) might have similar reaction pathways (Li & Blatchley 2009). Under UV/NH₂Cl treatment at pH 7.5, removal was mainly attributed to ·Cl among ·Cl, ·Cl₂ and ·OH, whose contribution accounted for 95.2% (Zhang et al. 2019b). It should be noted that the concentration of TN was a little lower with increasing contact time under UV/chlorine treatment, which was consistent with the fact that photolysis degradation of monochloramine could generate N₂O (23.6%) by reaction (16) at pH = 7.5. Moreover, the molar extinction coefficients of hypochlorite and monochloramine were 60 and 354 M⁻¹ cm⁻¹, respectively, at pH = 7.2 under a UV wavelength of 254 nm. The molar extinction coefficients of ammonia could be neglected. What mentioned above meant that photolysis of monochloramine and ammonia reacting with free radicals resulted in the decrease of under UV/chlorine treatment (Li & Blatchley 2009; Zhang et al. 2015, 2019b), and various nitrogen radicals were produced under UV/chlorine treatment resulting in more CNM formation than chlorine treatment. It could also be inferred that the free radicals of nitrogen formed by ammonium reacted more easily to produce CNM than those formed by nitrate.

The effects of NH₄NO₃, NaNO₃ and Ca(NO₃)₂ on the formation of CNM under UV/chlorine treatment was investigated when glucose was used as the carbon source. The results were as follows.

• (1)

  Under UV/chlorine treatment, the maximum concentrations of CNM produced by three kinds of nitrates (NH₄NO₃, NaNO₃ and Ca(NO₃)₂) were higher than that produced under only chlorine treatment.

• (2)

  Under UV/chlorine treatment, the gap between the maximum concentration of CNM produced by NH₄NO₃ and that produced by the other two kinds of nitrates (NaNO₃ and Ca(NO₃)₂) was widened.

• (3)

  The yield of CNM increased with increasing glucose concentration and UV intensity, while the yield of CNM decreased with increasing pH.

• (4)

  The yield of CNM formed by NH₄NO₃ was higher than that formed by NaNO₃ and Ca(NO₃)₂ at the same nitrogen content under UV/chlorine treatment.

• (5)

  The formation laws of CNM found in real water were consistent with those found in simulated water samples.

The findings can provide references for controlling the generation of CNM under UV/chlorine treatment in practical production. More studies about the role of nitrogen radicals were expected.

This work was financially supported by the National Natural Science Foundation of China (Nos. 22076023 and 21677032), the Natural Key R&D Program of China (Grant No. 2017YFC0504505) and the Fundamental Research Funds for the Central Universities. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

All relevant data are included in the paper or its Supplementary Information.