Performance and Mechanism Study of Simultaneous Removal of Carbamazepine and Ammonia from Water Using UV/Peroxymonosulfate Process

Abstract

Wastewater involving nitrogen-containing emerging contaminants is always accompanied by ammonia nitrogen. In this study, the 254 nm UV light activating peroxymonosulfate (PMS) process was investigated based on its performance and mechanisms for the simultaneous removal of carbamazepine (CBZ) and ammonia nitrogen. The results showed that both CBZ and ammonia could be simultaneously removed from water by the UV/PMS process, which was mainly attributed to the oxidation of SO₄^•− and •OH, respectively. Solution pH did not significantly affect CBZ degradation, but was a crucial factor for the removal of ammonia, and only the alkaline condition was effective for ammonia removal. The steady-state concentration of SO₄^•− (4.37 × 10⁻¹¹ M) at pH 10.5 was determined as 32 times that of •OH (1.35 × 10⁻¹² M), which made CBZ more competitive than ammonia in competing for radicals and more adaptable to coexisting anions. An appropriate increase in PMS concentration and light intensity could improve the removal of ammonia more significantly than that of CBZ, but an over-intense reaction could accelerate the decrease in solution pH, resulting in a plateau in ammonia removal. Moreover, the formation of nitrate and nitrogen gas was the primary transformation route of ammonia in the UV/PMS process. With the optimum PMS concentration of 2 mM, about 50% of the total nitrogen could be removed. The results of this study may provide some insights into applying the UV/PMS process for the simultaneous removal of emerging contaminants and ammonia nitrogen.

1. Introduction

The increased consumption of various emerging contaminants leads to their widespread distribution in various aquatic environments. Carbamazepine (CBZ), as a frequently used antiepileptic drug, stands out among these contaminants due to its persistence and resistance to biodegradation in conventional wastewater treatment [1]. The maximum concentration of CBZ detected in urban streams of different countries can reach as high as 67 μg/L [2], and significant ecological concerns can be aroused upon long-term exposure. The freshwater diatom was reported to be sensitive to CBZ with a 72 h EC₅₀ of 0.179 mg/L [3]. The inefficient treatment of urban sewage containing pharmaceuticals is one of the most important reasons for its migration into natural water [4]. Therefore, the application of an effective process to completely degrade CBZ is necessary. Moreover, due to the nitrogen-containing structure of CBZ, effective degradation of CBZ may also induce ionic by-products such as ammonium and nitrate [5], which may add to the nitrogen pollution burden in effluent water. It was also found that ammonium nitrogen significantly inhibited the biodegradation of CBZ in algal cultures [3]. Thus, an investigation of effective methods that can simultaneously remove CBZ and nitrogen from water is of great significance.

Due to the biorefractory property of CBZ, advanced oxidation processes (AOPs) are assumed to be promising methods for the degradation of CBZ. The activation of peroxymonosulfate (PMS) has been widely investigated for the effective degradation of various organic compounds mainly because of the efficient generation of highly reactive oxygen species (ROS), including radical species (•OH, SO₄^•−, O₂^•−, etc.) and non-radical species (e.g., ¹O₂) [6]. Among diverse activation approaches, using UV light to activate PMS has the advantages of cleanliness, being free of secondary pollution by metals, etc. In addition, the UV/PMS process is also a convenient alternative as a tertiary treatment for disinfection purposes [7]. The efficient degradation and even mineralization of CBZ by the UVC/PMS process have been reported [7], but the influence on nitrogen transformation was not mentioned.

Jiang et al. (2024) found that UV-activated persulfate showed high selectivity (>79%) toward oxidizing ammonia nitrogen to nitrogen gas with the aid of CaO to create an alkaline condition, in which •OH was identified as the critical radical for the oxidation of ammonia [8]. Using vacuum UV (VUV) to activate the mixture of PMS and H₂O₂ could remove 82.7% of NH₄⁺ with N₂ selectivity over 99%, and both •OH and SO₄^•− contributed to ammonia oxidation in the VUV/H₂O₂/PMS process [9]. The selective oxidation of ammonia nitrogen to N₂ as the final product is preferable for complete nitrogen removal from wastewater. Most of the current studies applying AOPs for ammonia treatment consistently proposed the alkaline condition as the preferable condition due to the easy reaction between molecular ammonia and radical species [10]. However, whether the preferable condition is also efficient for the degradation of organic compounds and the possible influence between ammonia removal and organic compounds degradation still need further investigation.

In this study, PMS was activated by 254 nm UVC light for an investigation of the performance and mechanisms of the simultaneous removal of CBZ and ammonia. The influence of different parameters and the coexistence of both CBZ and ammonia on their removal performance were investigated in detail. Moreover, the mechanisms of CBZ and ammonia were discussed. The results of this study may shed some light on the application of UV/PMS as an advanced treatment technology for the simultaneous removal of organic micropollutants and ammonia nitrogen.

2. Results and Discussion

2.1. The Removal Performance of CBZ and Ammonia in UV/PMS Process

The possible removal of CBZ and ammonia in the UV/PMS process and the corresponding constituent processes (UV, PMS) was investigated by adding CBZ and ammonia separately or simultaneously under an alkaline condition (pH₀ = 10.5). As shown in Figure 1a,b, PMS could weakly oxidize ammonia (ca. 10%) but hardly reacted with CBZ. Unlike ammonia, CBZ showed UV absorbance under 254 nm (Figure S1), which resulted in the partial degradation (ca. 15%) of CBZ under the UV process alone. The slight decrease in ammonia concentration (ca. 8%) under UV irradiation was likely due to the escape of NH₃ gas under pH 10.5 (pKa of NH₄⁺ = 9.3 [11]). It was clearly seen that both ammonia and CBZ could be rapidly removed in the UV/PMS process within 10 min. When CBZ and ammonia were in coexistence (Figure 1c,d), it was observed that both pollutants could still be completely removed within 10 min, but with decreased initial rates, which was ascribed to the competition for active species in the UV/PMS process.

2.2. Influence of Different Parameters

The influence of different parameters, including solution pH, PMS concentration, light intensity, pollutant concentration, and background anions, on the removal of CBZ and ammonia was examined systematically.

2.2.1. Influence of Solution pH

Solution pH could influence the existing state of ammonia, CBZ (pKa = 13.9 [12]), and PMS, which may further influence the removal performance of CBZ and ammonia in the UV/PMS process. Control tests were conducted with the UV alone process with varied pH (Figure S2), and indistinguishable effects were obtained when using different pH ranges. Similarly, pH also played a weak role in the direct reaction between PMS and both pollutants, with a slight enhancement in ammonia removal under alkaline conditions (Figure S3).

However, the removal of ammonia in the UV/PMS process with or without the presence of CBZ was found to be significantly affected by solution pH (Figure 2). The observable removal could only be detected when the initial solution pH was higher than 9, and the initial removal rate increased remarkably with the elevation of pH. Since the difference in the PMS activation rate under different pH ranges (Figure S4) was not as significant as that of ammonia removal, the above results could well imply that the active substances in the UV/PMS process may only react with ammonia rather than ammonium salt. Similar results were also reported in other AOPs [13,14].

In contrast, the degradation of CBZ was not influenced much by solution pH, with complete degradation achieved within 30 min under all pH conditions (Figure S5). The favorable alkaline condition was also advantageous for CBZ degradation. Moreover, it was also observed that the removal of CBZ generally followed pseudo-first-order kinetics under all pH conditions, but the removal of ammonia showed two-stage kinetics. This was likely due to the change in the existing form of ammonia nitrogen during the reaction, since solution pH could rapidly decrease to neutral or acidic conditions as the reaction went on without the addition of any buffer. At the initial pH of 10.5, almost all of the ammonia was in molecular form, which led to the fast oxidation of ammonia and complete removal within 10 min. Therefore, the following investigations were conducted at pH 10.5.

2.2.2. Influence of PMS Concentration

The influence of PMS concentration was also explored. As shown in Figure 3, the variation in PMS concentration was more influential in ammonia removal than CBZ. This may be partially because there were more moles of ammonia nitrogen (0.143 mM) than of CBZ (0.01 mM). The PMS concentration could directly determine the amount of active species in the UV/PMS process, which was more critical for ammonia at high concentrations. It was also noticed that the presence of CBZ (Figure 3a,c) could slow down the removal of ammonia when doped with the same amount of PMS. The initial removal rates of both ammonia and CBZ were further analyzed (Figure S6). It was demonstrated that the presence of CBZ significantly decreased the initial removal rate of ammonia, while the coexistence of ammonia only had a little impact on CBZ degradation. This may be explained by the differences in molecular dimensions. The large molecular size of CBZ was advantageous for the collision with active species to increase the reaction probability. Moreover, more reactive sites could be attacked in CBZ by radicals than in ammonia.

Moreover, as shown in Figure S6, an optimum PMS concentration could be obtained in the UV/PMS process for both ammonia and CBZ removal, which suggests the dual roles of PMS in the photochemical reaction. The variation in PMS concentration was monitored during the reaction (Figure S7). It was revealed that the initial consumption rate of 6 mM PMS was much faster than that of 4 mM PMS without improving the removal rate of both pollutants. Therefore, a too high dosage of PMS may intensify the self-quenching effect (Equations (1) and (2)) and PMS quenching effect (Equations (3) and (4)) due to the fast burst of reactive oxygen species (ROS) [15].

$$\mathrm{S}{\mathrm{O}}_4^{•-}+\mathrm{S}{\mathrm{O}}_4^{•-}\to {\mathrm{S}}_2{\mathrm{O}}_8^{2-}$$

(1)

$$•\mathrm{O}\mathrm{H}+•\mathrm{O}\mathrm{H}\to {\mathrm{H}}_2{\mathrm{O}}_2$$

(2)

$$\mathrm{S}{\mathrm{O}}_4^{•-}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{S}{\mathrm{O}}_5^{•-}+\mathrm{H}\mathrm{S}{\mathrm{O}}_4^{-}$$

(3)

$$•\mathrm{O}\mathrm{H}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{S}{\mathrm{O}}_5^{•-}+{\mathrm{H}}_2\mathrm{O}$$

(4)

2.2.3. Influence of Light Intensity

The effect of light intensity in the UV/PMS process was also investigated by changing the number of lamps (Figure 4). Control tests conducted in the sole-UV process disclosed that increasing the light intensity only weakly accelerated CBZ photolysis and showed a negligible effect on ammonia removal (Figure S8). Therefore, the light intensity was assumed to mainly affect the activating rate of PMS to change the rate and kinetics of pollutant removal in the UV/PMS process. As shown in Figure 4, when doped with enough PMS (4 mM), reducing light intensity could only slow down the initial removal rate of ammonia and CBZ, while complete removal could still be achieved even with two lamps. Moreover, the removal of ammonia also followed pseudo-first-order kinetics without showing lag phases when only two lamps were applied, which suggested that moderating the progression of the photochemical reaction in the UV/PMS process could alleviate the drop in solution pH, and the proper pH condition for ammonia removal could be maintained well. It was also worth mentioning that, when the light intensity increased by five times (from 2 lamps to 10 lamps), the initial removal rate of ammonia, as shown in Figure 4a, only increased by four times (from 0.17 mg/L/min to 0.69 mg/L/min). This was also because of the above-mentioned quenching effects (Equations (1)–(4)) when the photochemical reaction was too fast. Hence, properly controlling the reaction intensity was important to improve the cost-effectiveness.

2.2.4. Influence of Pollutant Concentration

The removal performance of ammonia and CBZ was also studied with varied initial target concentrations (Figure 5). The degradation kinetics of CBZ generally followed pseudo-first-order kinetics when changing CBZ concentration (Figure 5b), but different removal kinetics were observed for ammonia (Figure 5a). When the initial concentration of ammonia was low (≤2 mg/L), complete removal was achieved without showing a clear lag phase. However, a lag phase appeared for the removal of ammonia nitrogen with concentrations higher than 2 mg/L, which was also likely due to the decrease in solution pH during the late period of the reaction. It was interesting to note that the lag phase of ammonia removal appeared later with a higher initial concentration of ammonia. This should be related to the buffering effect caused by ammonia at high concentrations, which could ease the rapid pH change. The initial removal rate of ammonia was shown in Figure S9. It was concluded that, although the lower percentage of ammonia (C/C₀) was removed at higher initial concentrations, the initial removal rate was enhanced by increasing the concentration of ammonia. Therefore, increasing the initial concentration of ammonia could improve the utilization efficiency of PMS.

2.2.5. Influence of Background Anions

The influences of background anions (NO₃⁻, SO₄²⁻, CO₃²⁻, and Cl⁻) on the removal of ammonia and CBZ were compared (Figure 6). Adding 4 mM anions showed different effects on ammonia and CBZ removal. Only a weak inhibition of CBZ removal was found with the addition of these anions (Figure 6b), and complete degradation of CBZ could be achieved within 2 min. However, a more obvious influence was seen in ammonia removal. Generally, the addition of NO₃⁻ and SO₄²⁻ anions showed a weak inhibition of ammonia removal. The weak inhibition of the removal of both pollutants by SO₄²⁻ created a good foundation for applying the UV/PMS process. It was also noted that the presence of Cl⁻ and CO₃²⁻ could significantly reduce the removal rate of ammonia due to the quenching effect of Cl- and CO₃²⁻ (Equations (5)–(7)), which could transform partial SO₄^•− and •OH to less oxidative species (e.g., Cl^• and CO₃^•−) [16].

$$\mathrm{C}{\mathrm{l}}^{-}+\mathrm{S}{\mathrm{O}}_4^{•-}\to \mathrm{S}{\mathrm{O}}_4^{2-}+\mathrm{C}{\mathrm{l}}^{•}$$

(5)

$$\mathrm{C}{\mathrm{O}}_3^{2-}+\mathrm{S}{\mathrm{O}}_4^{•-}\to \mathrm{S}{\mathrm{O}}_4^{2-}+\mathrm{C}{\mathrm{O}}_3^{•-}$$

(6)

$$\mathrm{C}{\mathrm{O}}_3^{2-}+•\mathrm{O}\mathrm{H}\to \mathrm{O}{\mathrm{H}}^{-}+\mathrm{C}{\mathrm{O}}_3^{•-}$$

(7)

2.3. The Removal Mechanisms of Ammonia and CBZ

Quenching Experiments and ESR Analysis. The removal mechanisms of CBZ and ammonia were discussed in detail. The degradation mechanism of CBZ was first analyzed by adding EtOH or TBA as the radical quencher (Figure 7a). EtOH is capable of scavenging both •OH (1.2~2.8 × 10⁹ M⁻¹ s⁻¹) and SO₄^•− (1.6~7.7 × 10⁷ M⁻¹ s⁻¹) at high rates [17]. TBA is an effective scavenger for •OH (3.8~7.6 × 10⁸ M⁻¹ s⁻¹) and much less sensitive to SO₄^•− (4.0~9.1 × 10⁵ M⁻¹ s⁻¹) [18]. As shown in Figure 7a, the addition of 500 mM TBA could slow down the initial degradation rate of CBZ by 44%, but complete degradation of CBZ could still be achieved in 10 min. By adding 500 mM EtOH, however, the process was almost fully quenched. The difference in quenching effects between EtOH and TBA suggested that both •OH and SO₄^•− existed and contributed to CBZ degradation, while SO₄^•− played a more important role than •OH.

Either EtOH or TBA (10 mM) was also added during the oxidation of ammonia in the UV/PMS process. Since the presence of highly concentrated organic compounds disturbed the quantification of ammonia using Nessler’s reagent, the formation of nitrate was analyzed instead to investigate the influence of radical scavengers (Figure 7e). The results showed that both EtOH and TBA could significantly inhibit the formation of nitrate, and EtOH showed a slightly more obvious inhibition of nitrate formation compared to TBA. When 50 mM of EtOH or TBA was employed, however, hardly any nitrate was detected. Based on the quenching experiments, •OH was assumed to play a dominant role in ammonia oxidation, and the stronger quenching performance of EtOH was due to the blocking of the conversion from SO₄^•− to •OH under alkaline conditions (Equation (9)).

The radicals of SO₄^•− and •OH can be produced by the photolysis of HSO₅⁻ under 254 nm irradiation (Equation (8)). Under alkaline conditions, the SO₄^•− radicals were easily converted to •OH (Equation (9)). The electron spin resonance (ESR) technique was employed to determine the generated reactive oxygen species using DMPO as the trapping agent. Figure 7d showed four characteristic peaks of DMPO-•OH with an intensity ratio of 1:2:2:1 and a quartet pattern. The failure to detect the SO₄^•− signal was probably due to the instability of DMPO- SO₄^•−. In view of the possible photoactive property of nitrate ions [19], the ESR was also applied in the UV/NO₃⁻ process under alkaline conditions in an attempt to analyze the formation of radical species (Figure 7f). However, there were no clear signals detected, probably due to the relatively low concentration of nitrate (<2 mg/L-N) and the disadvantage of nitrate ions with low concentration in competing with PMS for 254 nm UV light (ε_PMS = 141 M⁻¹ cm⁻¹, ε_nitrate = 3.33 M⁻¹ cm⁻¹ at pH 10.5). Therefore, it can be concluded that the clear signal detected in UV/PMS was entirely attributed to PMS activation.

HSO₅⁻ + hν → SO₄^•− + •OH

(8)

SO₄^•− + OH⁻ → SO₄²⁻ + •OH (6.5 × 10⁷ M⁻¹ s⁻¹)

(9)

Quantification of ROS in UV/PMS process. Due to the difficulty of detecting SO₄^•− using ESR, the steady-state concentrations of •OH and SO₄^•− were quantified using 4-nitrobenzoic acid (4-NBA) and benzoic acid (BA) as the probe compounds. BA can react with both SO₄^•− (1.2 × 10⁹ M⁻¹ s⁻¹) and •OH (4.3 × 10⁹ M⁻¹ s⁻¹), while 4-NBA can react readily with •OH (2.6 × 10⁹ M⁻¹ s⁻¹) while being inert to SO₄^•− (<10⁶ M⁻¹ s⁻¹) [20].

The initial degradation of BA and 4-NBA in UV/PMS processes can be expressed as follows (Equations (10)–(13)):

$${\displaystyle \frac{\mathrm{d}\left[4\text{-}\mathrm{NBA}\right]}{\mathrm{dt}}}=-{ k}_{4\text{-}\mathrm{NBA},•\mathrm{OH}}\times {\left[•\mathrm{OH}\right]}_{\mathrm{ss}}\times [4\text{-}\mathrm{NBA}]-k_{\mathrm{UV},4\text{-}\mathrm{NBA}}\times [4\text{-}\mathrm{NBA}]$$

(10)

$$\ln {\displaystyle \frac{\left[4\text{-}\mathrm{NBA}\right]}{{\left[4\text{-}\mathrm{NBA}\right]}_0}}=-{\mathrm{k}}_{4\text{-}\mathrm{NBA},•\mathrm{OH}}{\times \left[•\mathrm{OH}\right]}_{\mathrm{ss} }\times \mathrm{t}-{ k}_{\mathrm{UV},4\text{-}\mathrm{NBA}}\times \mathrm{t}=-{\mathrm{k}}_{\mathrm{obs},4\text{-}\mathrm{NBA}}\times \mathrm{t}$$

(11)

$${\displaystyle \frac{\mathrm{d}\left[\mathrm{BA}\right]}{\mathrm{dt}}}=-({\mathrm{k}}_{\mathrm{BA},•\mathrm{OH}}{\times \left[•\mathrm{OH}\right]}_{\mathrm{ss}}+{ \mathrm{k}}_{\mathrm{BA},{\mathrm{SO}}_4^{•-}}\times {\left[{\mathrm{SO}}_4^{•-}\right]}_{\mathrm{ss}})\times \left[\mathrm{BA}\right]-{ k}_{\mathrm{UV},\mathrm{BA}}\times [\mathrm{BA}]$$

(12)

$$\ln {\displaystyle \frac{\left[\mathrm{BA}\right]}{{\left[\mathrm{BA}\right]}_0}}=-({\mathrm{k}}_{\mathrm{BA},•\mathrm{OH}}\times {\left[•\mathrm{OH}\right]}_{\mathrm{ss}}+{\mathrm{k}}_{\mathrm{BA},{\mathrm{SO}}_4^{•-}}{\times \left[{\mathrm{SO}}_4^{•-}\right]}_{\mathrm{ss}})\times \mathrm{t}-k_{\mathrm{UV},\mathrm{BA}}\times \mathrm{t}=-{\mathrm{k}}_{\mathrm{obs},\mathrm{BA}}\times \mathrm{t}$$

(13)

where [4-NBA] and [BA] are the concentrations of 4-NBA and BA (M) at moment t, respectively; k_4-NBA,•OH is the second-order rate constant of the reactions between 4-NBA and •OH; k_BA,•OH, and k_{BA, SO4•−} are the second-order rate constants of the reactions between BA and •OH, BA, and SO₄^•−, respectively; and [•OH]_ss and [SO₄^•−]_ss refer to the steady-state concentrations of •OH and SO₄^•− at the beginning of the reactions in UV/PMS process.

The observed first-order rate constants of 4-NBA (k_obs,4-NBA) and BA (k_obs,BA) degradation in the UV/PMS process were determined as 0.0066 s⁻¹ and 0.0788 s⁻¹, respectively (Figure S10b,d). The direct photolysis rate constants of 4-NBA and BA were determined as 0.0031 s⁻¹ and 0.0205 s⁻¹, respectively (Figure S10a,c). By substituting the kinetic values into Equations (11) and (13), the steady-state concentrations of [•OH]_ss and [SO₄^•−]_ss were calculated as 1.35 × 10⁻¹² M and 4.37 × 10⁻¹¹ M, respectively.

It was found that, although the SO₄^•− produced in the UV/PMS process has the chance to be transformed into •OH under alkaline conditions, the steady-state concentration of SO₄^•− was still about 32 times that of •OH at the beginning of the reaction. According to the quenching experiment results, CBZ degradation was more preferably induced by SO₄^•− due to the electron-rich property of its structure, but the removal of ammonia was dominantly attributed to •OH, although the concentration of •OH was low. This was likely due to the low reaction rate between SO₄^•− and NH₃ (3 × 10⁵ M⁻¹s⁻¹) compared to •OH and NH₃ (1 × 10⁸ M⁻¹s⁻¹) [21]. This can also explain the much easier disturbance of ammonia oxidation by different parameters (e.g., PMS concentration, existing anions, ammonia concentration) than that of CBZ degradation in view of the relatively low amount of •OH.

Discussion of Nitrogen Transformation. Specifically, NH₃ could react with •OH to form •NH₂ (Equation (14)) [9], which could be oxidized into NH₂OH by •OH (Equation (15)) [22]. NH₂OH was easily transformed to NO₂- and NO₃- with the help of •OH (Equations (16) and (17)) [23]. Under some conditions, ammonia might also be directly transformed to N₂ via the reaction with a high concentration of •OH (Equation (18)) [23]. Figure 7b shows the variation in ammonia concentration and the corresponding profile of nitrate concentrations. Hardly any nitrite could be detected in this study, which was possibly due to the fast transformation of nitrite to nitrate in the presence of PMS and ROS. As shown in Figure 7b, almost 75% of ammonia was transformed to nitrate with the addition of 4 mM PMS in the UV/PMS process. The measurement of total nitrogen (TN) also showed about 0.42 mg/L TN removal, which agreed well with the nitrogen balance between ammonia and nitrate nitrogen.

$$\mathrm{N}{\mathrm{H}}_3+•\mathrm{O}\mathrm{H}\to •\mathrm{N}{\mathrm{H}}_2+{\mathrm{H}}_2\mathrm{O}$$

(14)

$$•\mathrm{N}{\mathrm{H}}_2+•\mathrm{O}\mathrm{H}\to \mathrm{N}{\mathrm{H}}_2\mathrm{O}\mathrm{H}$$

(15)

$$\mathrm{N}{\mathrm{H}}_2\mathrm{O}\mathrm{H}+4•\mathrm{O}\mathrm{H}\to \mathrm{N}{\mathrm{O}}_2^{-}+3{\mathrm{H}}_2\mathrm{O}+{\mathrm{H}}^{+}$$

(16)

$$\mathrm{N}{\mathrm{O}}_2^{-}+2•\mathrm{O}\mathrm{H}\to \mathrm{N}{\mathrm{O}}_3^{-}+{\mathrm{H}}_2\mathrm{O}$$

(17)

$$\mathrm{N}{\mathrm{H}}_4^{+}+3•\mathrm{O}\mathrm{H}\to {\displaystyle \frac{1}{2}}{\mathrm{N}}_2+{\mathrm{H}}^{+}+{\mathrm{H}}_2\mathrm{O}$$

(18)

It was noticed from Equations (14)–(18) that the relative abundance between ammonia and radicals may influence the transformation route of ammonia. Therefore, the variation profiles of ammonia, nitrate, and TN concentrations in UV/PMS processes in the presence of different PMS concentrations were depicted in Figure S11. It was found that the concentration of PMS not only influenced the removal rate of ammonia but also resulted in different nitrate yields and TN removal. The removal amount of TN in the UV/PMS process supplied with different PMS concentrations was compared in Figure 7c. The results disclosed the promoting effect of increasing PMS concentration on TN removal, but further increasing the PMS concentration above the optimum level (2 mM) could gradually decrease TN removal. With the optimum PMS concentration of 2 mM, about 50% of TN could be removed.

3. Chemicals and Methods

3.1. Chemicals

Carbamazepine (CBZ, 99.5%), PMS (Oxone^®: KHSO₅, 0.5KHSO₄, 0.5K₂SO₄), and 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO, ≥97.0%) were purchased from Sigma Aldrich, Louis, MO, USA. Ammonium sulfate (≥99.0%), sodium thiosulfate pentahydrate (≥99.0%), sodium nitrate (≥99.0%) sodium carbonate (≥ 99.8%), sodium chloride (≥99.5%), potassium iodide (≥99%), mercuric iodide (≥99.5%), sulfuric acid (95~98%), 4-nitrobenzoic acid, benzoic acid (≥99.0%), tert-butanol (TBA, ≥99.0%), and ethanol (EtOH, ≥99.9%) were purchased from Sinopharm Chemical Reagent Co. Ltd., Shanghai, China. Seignette salt (99.5%), sodium hydroxide (≥96%, TN ≤ 0.0005%), and potassium persulfate (≥99.0%, TN ≤ 0.0005%) were obtained from Macklin, Shanghai, China. Ultrapure water was used exclusively. The solution pH was adjusted at the beginning using 1.0 M H₂SO₄ or 1.0 M NaOH without adding any buffer.

3.2. The Photochemical Reaction Apparatus and Experimental Conditions

The photochemical reactions were all carried out under air-equilibrated conditions without artificial aeration or temperature control. Experiments were conducted in a tailor-made photo-reactor with low-pressure mercury lamps installed on top of the ceiling. UV mercury lamps with a wavelength of 254 nm were purchased from Rayonet^®.org, Branford, CT, USA. Ten lamps were applied unless otherwise stated, and the incident photon intensities were about 10⁻⁵ Einstein/L/s, determined by chemical actinometer [19]. Quartz beakers with magnetic stirring were used throughout. All experiments were carried out in 250 mL beakers at room temperature. Unless otherwise stated, the initial concentrations of CBZ, ammonia, and PMS were fixed at 0.01 mM, 2.0 mg/L, and 4.0 mM, respectively. At certain time intervals, 1.0 mL of reaction solution was withdrawn, which was immediately added to vials containing 0.5 mL of concentrated Na₂S₂O₃ to quench any possible reactions between PMS and pollutants. Similarly, for the quantification of NH₄⁺-N, 1.0 mL of reaction solution was taken out of a vial containing Nessler reagent, and 15 min was allowed for color development. The required amount of acid or alkaline was predetermined by several trials in preliminary experiments, and the pH₀ values denoted the condition at the beginning of the photochemical reaction. The batch experiments were conducted at least twice until the errors were below 5%, and the average values obtained were used for plotting.

3.3. Analytical Methods

The CBZ concentration was analyzed by High-Performance Liquid Chromatography (HPLC; Dionex Ultimate 3000, Thermo Fisher Scientific, Waltham, MA, USA) equipped with a reverse-phase C18 column (250 mm × 4.6 mm × 5.0 μm) and an ultraviolet and visible (UV-Vis) spectrophotometry detector with the detection wavelength at 283 nm. The mobile phase was a mixture of 70/30% (v/v) methanol and 0.1% phosphoric acid water solution at a flow rate of 1.0 mL/min. The column temperatures were set at 35 °C. The concentration of NH₄⁺-N was measured using Nessler reagent on a UV-vis spectrophotometer at 420 nm [24]. Total nitrogen was measured using a standard method of alkaline potassium persulfate oxidation, and the absorbance was detected at 220 and 275 nm [22]. The concentrations of NO₃⁻ and NO₂⁻ were analyzed using an ion chromatograph system (Dionex ICS-600, Thermo Fisher Scientific, Waltham, MA, USA). For the quantification of HSO₅⁻, KI was used to react with HSO₅⁻ to form I₃⁻ (Equations (19) and (20)). The concentration of I₃⁻ could be quantified at λ_max = 352 nm [25]. The electron spin resonance (ESR) spectra were obtained by a Bruker EMX-10/12 (Bruker, Billerica, MA, USA) device with X-band field scanning.

2I⁻ + HSO₅⁻ + 2H⁺ → HSO₄⁻ + I₂ + H₂O

(19)

I₂ + I⁻ → I₃⁻

(20)

4. Conclusions

In this study, the performance and mechanisms of the simultaneous removal of CBZ and ammonia in the UV (254 nm)/PMS process were investigated in detail. The results showed that both CBZ and ammonia could be simultaneously removed from water via the UV/PMS process. Solution pH was a critical factor for the removal of ammonia, while not significantly influential to CBZ degradation, and efficient removal of ammonia was only achieved under alkaline conditions. Moreover, the removal of ammonia was more easily affected by PMS concentration, background anions, and pollutant concentration than CBZ. ESR measurement and quenching experiments showed the dominant roles of SO₄^•− and •OH in CBZ degradation and ammonia oxidation, respectively. The steady-state concentration of SO₄^•− (4.37 × 10⁻¹¹ M) at pH 10.5 was determined as 32 times that of •OH (1.35 × 10⁻¹² M), which was more beneficial to CBZ degradation. In addition, an appropriate increase in PMS concentration and light intensity could improve the removal of ammonia more significantly than that of CBZ, but an over-intense reaction could accelerate the decrease in solution pH, which could result in a plateau in ammonia removal. Increasing the concentration of ammonia could enhance the utilization efficiency of PMS. Moreover, the formation of nitrate and nitrogen gas was the primary transformation route of ammonia in the UV/PMS process. The dosage of PMS played dual roles in decreasing the amount of TN, and with the optimum PMS concentration of 2 mM, about 50% of the total nitrogen could be removed.