Introduction

Antihistamines have been widely used to treat a variety of diseases since they entered the market in the 1940s (Kristofco and Brooks 2017). Given the variety of uses, antihistamines have become the largest amount of drugs used to treat allergic diseases. According to the World Allergy Organization (WAO), the incidences of allergic asthma, rhinitis, urticarial, and pruritus have increased threefold, with a prevalence of 22% in the past 30 years. Antihistamines were included in the recommended drugs for first-line diagnosis and treatment (Zhang et al. 2021). These medicines have been developed through three generations. The use of the first-generation antihistamines, such as diphenhydramine and decloxizine, has decreased because of severe adverse effects. Second-generation antihistamines include loratadine (LOR), cetirizine (CER), and azelastine (Wu 2012); and third-generation antihistamines include fexofenadine (FEX) and levocetirizine. In China, the average annual production of LOR, FEX, and CER during 2015–2017 was 1.9, 0.79, and 40.6 tons, respectively (Editorial Department of Annual Report 2015, 2016, 2017).

All these drugs were discharged into wastewater treatment plants (WWTPs) in the form of the parent compounds or their metabolites after being taken by humans (Zhou et al. 2016). FEX and CER were excreted as the parent compounds; as a metabolite of LOR, desloratadine also had pharmacological activity (Kosonen and Kronberg 2009; Xu and Lin 2006; Zhuang et al. 2010). To date, the existing work on antihistamines mainly focused on adsorption, biodegradation, and advanced oxidation. Golovko et al. (2020) found that granular activated carbon had a high adsorption capacity for CER and FEX; Faria et al. (2020) indicated that the higher removal efficiency of LOR in EGSB-MBR was attributed to the adsorption and membrane retention, but not biodegradation; Gadipelly et al. (2016) applied the UV/S2O82− system to degrade CER and found that greater than 95% degradation was achieved within 90 min. Since the typical physicochemical and biological processes in traditional WWTPs are designed to remove major pollutants, pharmaceuticals may not be completely removed and are then discharged into the aquatic environment in the effluent. For example, influent and effluent concentrations of CER in a WWTP in Germany were detected at 0.49 μg/L and 0.51 μg/L, and biopersistence was found (Bahlmann et al. 2012). The concentrations of FEX in the influent and effluent of a WWTP in the Czech Republic were 0.18 μg/L and 0.17 μg/L, respectively (Golovko et al. 2014). Antihistamines were also found in the marine environment. CER was found in the Loimi River at approximately 7 ng/L in Finland (Kosonen and Kronberg 2009). The maximum concentration of CER detected in San Francisco Bay was 6.3 ng/L, and the maximum concentration of LOR detected in the Pacific Ocean was 57 ng/L (Nödler et al. 2014). Although concentrations typically in the ng/L to μg/L range have been detected in aquatic environments, the residue of drugs and their pseudo-persistence may cause serious ecological impacts and potential threats to human health (Kostich et al. 2014). LOR in concentrations of hundreds of μg/L could induce 50% mortality of Ceriodaphnia dubia (Iesce et al. 2019); FEX at 1.5 μg/L reduced the activity of the damselfly, and the escape time needed from potential predators was extended from 2 to 4 s (Jonsson et al. 2014).

Chlorine oxidation and ultraviolet oxidation are very common tertiary treatment processes in WWTPs; however, their degradation effect and the mechanism thereof on antihistamines are still rarely reported. As an emerging advanced oxidation technology, UV-NaClO combines the advantages of solely using UV or NaClO. In this process, direct photolysis, free radical degradation, and chlorine oxidation were complementary to each other. This enhances the degradation effect of organic compounds to a certain extent and brings great potential applications. UV-NaClO reduced the use of chemical oxidants, thereby reducing the generation of disinfection byproducts and maximizing the safety of the effluent from sewage plants. It could meet the demands of safe effluent under actual conditions and frontier cognition (Aghdam et al. 2017; Wu et al. 2019; Lv 2019).

In this study, three oxidation methods, NaClO oxidation, UV irradiation, and UV-NaClO oxidation, were applied to explore the degradation of three typical structural and functional drugs (LOR, CER, and FEX) and the effects of pH on their removal. We also explored the contribution of different active species to the photolysis process and the effects of different aqueous matrices on UV irradiation and NaClO oxidation. Acute toxicity tests and ultraviolet scans were conducted to estimate the drug degradation reaction system. Based on the UPLC-Q-TOF/MS results, we also proposed the analysis of the oxidation pathways of NaClO on the three drugs and compared the NaClO oxidation and UV photolysis products of CER. We hope that the research results can provide experimental data and a basis for the theoretical analysis of evaluating the removal of drugs in the water environment management and the impact of drugs on the ecosystem.

Materials and Methods

Chemicals

LOR (purity > 98%), CER (purity > 99%), FEX (purity > 98%), methanol, acetonitrile, glacial acetic acid, ammonium acetate, and ammonium dihydrogen phosphate were all chromatographically pure and purchased from Anpel Laboratory Technologies (Shanghai) Inc. (China). H3PO4, Na2HPO4, KH2PO4, NaCl, glycerol, yeast extract, and tryptone were all analytically pure and were purchased from Sinopharm Chemical Reagent Co., Ltd. (China). Sodium thiosulfate was analytically pure and was purchased from Damao (Tianjin) Chemical Reagent Factory (China). Sodium hypochlorite (available chlorine > 10%) was analytically pure and was purchased from Best (Tianjin) Chemical Co., Ltd. (China). KHCO3, KNO3, humic acid (HA), isopropyl alcohol (IPA), sodium azide (NaN3), and potassium sorbate (PS) were all of analytical pure and purchased from Aladdin Reagent Co., Ltd. (Shanghai, China).

Experimental methods

NaClO oxidation

To facilitate detection by HPLC, the initial concentrations of the drugs were set at 0.5 mg/L, which is higher than the actual concentration in the aqueous environment. The initial concentration of free chlorine in the NaClO system was 40 mg/L in a 100 ml reaction volume at 25℃. The pH values were adjusted to 3, 5, 7, and 10. Samples were collected at 0, 8, 15, 30, and 60 min and filtered into vials using a syringe filter with an excess of sodium thiosulfate solution to stop the reaction.

UV irradiation

A schematic diagram of the ultraviolet light device is shown in Fig. S1. The initial concentration of each drug was 0.5 mg/L, and the reaction volume was 100 ml with the temperature set at 25℃. The light (125 W medium-pressure mercury lamp, Shanghai Yaming Lighting Company) was turned on and stabilized for 5 min before the experiment. The distance between the liquid surface and light source was 35 cm, and the optical power density was 24 mW/cm2 (measured by a CEL-NP2000, Beijing Zhongjiaojinyuan Technologies Co. Ltd., Beijing, China). The pH values were adjusted to 3, 5, 7, and 10. Samples were collected at 0, 30, 60, 90, 120, and 150 s and filtered into vials using a syringe filter.

Free radicals quenching

According to the relevant literature (Bodhipaksha et al. 2017), we selected isopropyl alcohol (IPA), sodium azide (NaN3), and potassium sorbate (PS) to quench the active species produced in the photolysis process: IPA with a concentration (v/v) of 2% was used to quench hydroxyl radicals (·OH), 5 mM NaN3 was a scavenger for quenching ·OH and singlet oxygen (1O2), 10 mM PS was used to quench 1O2 and the triplet excited state of the drugs (3ANT*). Other experimental procedures are described in the Section of UV irradiation.

UV-NaClO oxidation

The reaction conditions were the same as in Section of NaClO oxidationand Section of UV irradation. Samples were collected at the set times (the sampling time of LOR and CER were 0, 30, 60, 90, 120 s; the sampling times of FEX were 0, 10, 20, 30, 40, 50, 60 s) and filtered into a vial using a syringe filter with an excess of sodium thiosulfate solution to stop the reaction.

Influence of different aqueous matrices

We examined the influence of different aqueous matrices on NaClO oxidation and UV irradiation. HCO3− (2, 5, 10 mM), NO3− (0.2, 0.5, 1 mM), and HA (2, 5, 10 mg/L) were added to the system. Other experimental procedures were described in Section of NaClO oxidation and Section of UV irradiation.

All experiments were conducted in triplicate, and the mean values were used to calculate the degradation efficiencies. The filtered samples were frozen at -20℃ until analysis.

Acute toxicity test

The acute toxicity of the degradation reaction system was tested by Photobacterium phosphoreum (purchased from Nanjing Cas Kuake Technology Co., Ltd. China). After activation and cultivation of the freeze-dried powder, the bacterial suspension was used to measure the luminescence intensity of the samples collected during degradation with a multifunctional microplate detection system (BoiTek Cytation 5, USA) according to the method of Li et al. (2021).

Analytical methods

Analysis methods of drugs concentration

The samples were all analyzed by high-performance liquid chromatography (HPLC, LC-20AT, Shimadzu, Japan) equipped with an Eclipse XDB chromatographic column (4.6 mm × 150 mm, 5 μm). The analytical conditions are shown in Table 1.

Table 1 HPLC analytical conditions of the three drugs
Full size table

Ultraviolet scan of degradation products

To determine the cleavage of oxidation products, 5 mL reaction solution was taken at different intervals in the reaction process, and an ultraviolet spectrophotometer (UV7250, Shanghai Longnik Instrument Co., Ltd.) was used to conduct measurements in the range of 200–400 nm. The absorbance of the drug solution treated by NaClO oxidation or UV irradiation was scanned. To increase the detection of drugs, the initial concentrations of the drugs were set at 20 mg/L, and the initial concentration of free chlorine was 100 mg/L.

Methods for identification of intermediate products

A UPLC 1290–6540B Q-TOF liquid chromatography–mass spectrometer coupled with an Eclipse plus C18 column (1.8 μm, 2.1 × 100 mm) was used to identify the degradation products of the drugs. The specific conditions were as follows:

Chromatographic conditions: mobile phase: A is methanol, B is water containing 0.2% formic acid; the elution started with 20% methanol, then the methanol concentration increased to 90% within 7 min and was sustained for 1 min, and finally was decreased to 10% within 0.5 min and run for 3.5 min to reach equilibrium before the next injection; injection volume: 10 μL; flow rate: 0.3 mL/min; column temperature: 40℃.

Mass spectrometry conditions: Mass spectrometry was performed in both the positive and negative ionization modes for analysis (ESI ±), and the parameter settings are shown in Table 2.

Table 2 Mass spectrometry parameter settings
Full size table

Calculation methods

Degradation efficiencies

The degradation efficiencies and the first-order reaction kinetics of the three oxidation methods are calculated by Eq. (1) and (2), respectively:

$$Degradation\;efficiency\:(\%)=\frac{(C_0-C_t)}{C_0}\times100\%$$
(1)
$$ln\frac{{C_{t} }}{{C_{0} }} = - kt$$
(2)

where Ct is the drug concentration at time t, mg/L; C0 is the initial concentration of the drug, mg/L; k is the first-order kinetic rate constant; and t is the reaction time.

Contribution rates of active species

The contribution rates of active species in the photolysis process can be calculated by Eq. (3), (4), and (5), respectively (Lin et al. 2020):

$$R_{ \bullet OH} = \frac{{k_{w} - k_{IPA} }}{{k_{w} }}$$
(3)
$$R_{{{}^{1}O_{2} }} = \frac{{k_{IPA} - k_{{NaN_{3} }} }}{{k_{w} }}$$
(4)
$$R_{{^{3} ANT^{*} }} = \frac{{k_{w} - k_{PS} }}{{k_{w} }} - R_{{{}^{1}O_{2} }}$$
(5)

where R·OH, \({R}_{{1}_{{O}_{2}}}\), and \({{\text{R}}_{ 3}}_{{ANT}^{*}}\) are the contribution rates of ·OH, 1O2, and 3ANT* to the drugs in the photolysis process; kIPA, \({k}_{Na{N}_{3}},\) and kPS are the photolysis rate constants after the addition of the quencher IPA, NaN3, and PS, s−1; and kw is the actual photolysis rate constant of the drugs in ultrapure water.

Light-shielding coefficients

The light-shielding coefficients of the aqueous matrices are calculated by Eq. (6) and (7). The smaller the total light-shielding coefficients are, the stronger the light shielding effects; when aqueous matrices exist, the theoretical photolysis rate constants of the drugs can be calculated by Eq. (8) (Li et al. 2021):

$$S_{\lambda } = \frac{{1 - \left( {10^{{ - \left( {\alpha_{\lambda } + \varepsilon_{\lambda } \left[ {ANT} \right]} \right)l}} } \right)}}{{2.303\left( {\alpha_{\lambda } + \varepsilon_{\lambda } \left[ {ANT} \right]} \right)l}}$$
(6)
$$\sum\limits_{200}^{400} {S_{\lambda } } \frac{{\sum {L_{\lambda } S_{\lambda } \varepsilon_{\lambda } } }}{{\sum {L_{\lambda } \varepsilon_{\lambda } } }}$$
(7)
$$k^{\prime} = k_{w} \bullet \sum\limits_{200}^{400} {S_{\lambda } }$$
(8)

where \({\text{S}}_{\lambda }\) represents the light-shielding coefficient of the aqueous matrices (HCO3−, NO3−, and HA) at wavelength \(\lambda\); \(\sum_{200}^{400}{{\text{S}}}_{\lambda }\) represents the total light-shielding coefficient with a wavelength range of 200–400 nm; [ANT] represents the concentration of the drugs; aλ represents the absorbance of the aqueous matrices at wavelength λ; \({\epsilon }_{\lambda }\) represents the molar absorption coefficient of the drugs at wavelength λ; l represents the light path; Lλ represents the percentage of light intensity of the mercury lamp; and k’ is the theoretical photolysis rate constant of the drugs.

Acute toxicity

Acute toxicity is assessed by the relative luminescence intensity (RLI), which can be calculated by Eq. (9).

$$RLI = \frac{{L_{t} }}{{L_{0} }} \times 100\%$$
(9)

where Lt is the luminous intensity at time t and L0 is the initial luminous intensity.

Results and discussions

Degradation effects of NaClO on the three drugs

The pH value is an important water quality parameter in the water treatment process. The existence of drugs and free chlorine may change at different pH levels, resulting in differences in degradation efficiencies. The acidity coefficients (pKa) of HOCl and the three drugs were as follows: HOCl (pKa = 7.54), LOR (pKa = 4.3), FEX (pKa1 = 4.2, pKa2 = 9.52), and CER (pKa1 = 2.1, pKa2 = 8). The degradation of the three drugs by NaClO at different pH values is shown in Fig. 1.

Fig. 1
figure 1

Degradation efficiencies of drugs with NaClO at different pH values

Full size image
  1. (1)

    LOR

As shown in Fig. 1, slight degradation of LOR by NaClO was observed after 1 h. The degradation efficiencies ranged from 3%–11%, which decreased with increase in pH. When pH < 4.3, LOR was mainly in the molecular form, while LOR− was mainly in the ionic form when pH > 4.3. Because HOCl is a weak acid, the dominant form is HOCl at pH < 7.54 and in the form of OCl− at pH > 7.54 (Ji et al. 2016). When the pH was in the range of 7–10, the solution was dominated by LOR− and OCl−. More energy was needed for the collision of charge-repelled anions, and a less effective impact reduced the degradation efficiency of LOR (Yan et al. 2017). When the pH was 3 and 5, the dominant forms in the solution were HOCl and LOR and HOCl and LOR−, respectively. HOCl is more electrophilic than OCl−. Therefore, the degradation efficiencies of LOR were higher under acidic conditions. In addition, we speculated that HOCl had a stronger oxidation effect on LOR, so when pH = 3, the degradation effect of LOR was more obvious than that at pH = 5.

  1. (2)

    FEX

For FEX, the reactivity of FEX under alkaline conditions was much higher than that under acidic conditions. The degradation efficiencies increased significantly from 11 to 92% in 1 h with increase in pH, especially from pH = 5 to pH = 10. Because FEX is a zwitterionic compound, the carboxyl group on FEX was gradually ionized and lost protons to form -COO− as the pH increased. The ionized form of the carboxyl group may increase the sensitivity of FEX (Gumieniczek et al. 2019), and it was also reported that deprotonated ions were more susceptible to the attack of electrophiles HOCl and ClO− (Gallard and von Gunten 2002). The reasons given above explain why the ability of FEX to react with chlorine was enhanced.

  1. (3)

    CER

For CER, which is also a zwitterionic compound, the reasons for its high degradation efficiencies under alkaline conditions were the same as that of FEX. In particular, the degradation efficiency of CER reached 85% in 1 h at pH = 3. HOCl was dominant in the solution at that time, and the piperazine ring, with a higher electron density on the CER molecular structure, was the best position for the electrophilic reagent to attack (Dodd et al. 2005; Pan et al. 2021), leading good degradation effect under acidic conditions.

Degradation effects of UV irradiation on the three drugs

At different pH values, the degradation effects of the three drugs oxidized by UV irradiation are shown in Fig. 2.

Fig. 2
figure 2

Degradation efficiencies of drugs with UV light at different pH values

Full size image

As shown in Fig. 2, UV irradiation had more efficient degradation for these three drugs than NaClO oxidation. When pH = 7, the degradation efficiencies at 150 s were all greater than 75%. The three drugs are photosensitive organics with poor stability under UV irradiation (Abounassif et al. 2005; Bergheim et al. 2014; Breier et al. 2008). In addition, other antihistamines, such as ebastine, terfenadine, and levocetirizine, all had good photosensitivity according to the literature, and these drugs also contain nitrogen-containing heterocycles (Gumieniczek et al. 2019; Tarozzi et al. 2003; Gunjal et al. 2011). Boreen et al. (2003) found that many macrolide antibiotics without chromogenic functional groups limited the photochemical reactions. Therefore, it was speculated that chromophoric groups such as the nitrogen-containing heterocycles contained in these three drugs promoted their photoreactivity. The degradation efficiencies of LOR were significantly increased from 9 to 93% in 150 s when the pH increased from 3 to 7. This was because the anionic form of LOR increased with increase in pH, thus promoting the attack by ·OH with strong electrophilicity and degradation effects (Ji et al. 2012a). When the pH was 10, the degradation efficiency of LOR decreased slightly because singlet oxygen generation was inhibited due to the increase in hydroxide concentration in the solution (Bilski et al. 1996). Different from LOR, pH had less effect on the photolysis of FEX and CER, but the removal efficiencies were above 75%, indicating that UV irradiation had a good degradation effect on the different existing forms of FEX and CER.

Degradation effects of UV-NaClO on the three drugs

The degradation effects of the three drugs oxidized by UV-NaClO at different pH values are shown in Fig. 3. As shown in Fig. 3, under the UV-NaClO system, the degradation efficiencies of LOR increased from 33 to 85% with the increase in pH from 3 to 10. When the pH values were 3 and 5, the degradation efficiencies of FEX at 60 s were 62% and 93%, respectively. When pH = 7 and pH = 10, FEX reached a complete degradation within 20 s. The degradation efficiencies of the two drugs under alkaline conditions were greater than those under acidic conditions. However, CER showed the best degradation effect under acidic conditions. UV-NaClO oxidation had a great degradation effect on drugs because of its synergistic effect on the degradation of organic compounds (Fang et al. 2014). Under UV irradiation, free chlorine produced active species (Formulas S1–S5), such as chlorine free radicals (Cl·), superoxide anions (O·−), and hydroxyl free radicals (HO·) (Dong et al. 2017), thus enhancing the degradation effect of organic compounds. In addition, we found that the change in the drugs in the UV-NaClO system was similar to that in the UV system under different pH conditions. UV irradiation played a leading role in the combination of UV and NaClO.

Fig. 3
figure 3

Degradation efficiencies of drugs with UV-NaClO at different pH values

Full size image

Comparison of degradation kinetics of the three oxidation methods

Under different pH conditions, drugs were oxidized by the three oxidation methods, all following the first-order kinetic equation (Fig. S2–S4). The rate constants are shown in Table 3. By comparing the data in Table 3, we found that UV-NaClO oxidation had the best degradation effect on drugs under most conditions. In addition, a few special considerations are discussed below.

Table 3 Oxidation rate constants of the three oxidation methods
Full size table

The rate constants of LOR in different treatments at pH 5 and 7 were ranked as follows: UV > UV-NaClO > NaClO. The removal effect of UV-NaClO was not as good as that of UV, which may indicate that the degradation efficiencies of LOR by NaClO were too small and could be ignored during the reaction time of 120 s. In contrast, the free chlorine in the solution produced by NaClO competed with LOR to absorb photons. A previous report showed that the ability of free chlorine to absorb photons was much greater than that of organic compounds (Jiang et al. 2018). That is why the removal efficiencies of UV-NaClO on LOR at pH 5 and 7 were not as high as those of UV irradiation alone.

Another point worth noting was that CER achieved the maximum degradation efficiencies in both UV-alone and NaClO-alone systems at pH 10. However, in the UV-NaClO system, the degradation at pH 10 was not the best as expected, which eventually led to rate constants ranked as UV > UV-NaClO > NaClO. Different from the above analysis of LOR, the effect of NaClO on CER could not be ignored. Therefore, in the UV-NaClO system, the reasons for the poor effect of CER when pH = 10 were as follows: 1) The quantum yield of HOCl was higher than that of OCl−. When the pH was increased to 10, the main component of free chlorine changed from HOCl to OCl−, and the reduction in the quantum yield reduced the generation of ·OH and Cl. 2) The degradation rate constants of HOCl and OCl− reacting with ·OH were 2.0 × 109 L/(mol·sec) and 8.8 × 109 L/(mol·sec), respectively. The scavenging effect of OCl− on ·OH was greater than that of HOCl. Therefore, lower degradation of CER was observed at pH 10 (Wu et al. 2016).

Contribution rates of the active species in photolysis

Comparing UV-NaClO oxidation and UV irradiation, there was no significant difference in drug removal between the two treatments. We investigated the free radicals generated in the UV system to explore their contribution rates during the degradation process.

Photolysis was divided into two processes: direct photolysis and indirect photolysis. In the former, the compound directly absorbed photons; the latter was a reaction between the compound and the various active species produced by photosensitization (Ozaki et al. 2021). As seen from Table 4, the contribution rates of 1O2 for FEX and CER reached 70%, indicating that both drugs mainly underwent indirect photolysis guided by 1O2. In addition, we found that the contribution rates of active oxygen species to these two drugs were relatively similar. This is because CER and FEX are zwitterionic compounds containing carboxyl groups and have relatively similar molecular structures. For LOR, the contribution rates of 1O2 and 3ANT* were relatively close, which meant that the direct photolysis and indirect photolysis of LOR were on the same level. According to the above analysis, the three active species were involved in the photolysis process, and the mechanism was speculated as follows: ANT absorbed the photons to obtain energy and then was transformed into 3ANT* to undergo direct photolysis; 3ANT* transferred part of the energy to 3O2, the ground state form of dissolved oxygen in water, to excite active 1O2. Then, 3O2 accepted the electrons transferred from the organic compounds to produce H2O2, which absorbed photons to produce ·OH (Derbalah et al. 2020; Ping et al. 2021).

Table 4 Contribution rates of the active species in photolysis
Full size table

The influence of aqueous matrices on NaOCl oxidation and UV irradiation

In natural water bodies and wastewater, many different aqueous matrices will affect oxidation and degradation efficiency due to stimulation, inhibition, or shielding effects (Lin et al. 2020). HCO3−, NO3−, and HA were selected to estimate their effect in the process of drug degradation, and the results of NaClO oxidation are shown in Fig. 4.

Fig. 4
figure 4

The influence of aqueous matrices on NaOCl oxidation

Full size image

HCO3− promoted drug degradation because it could generate CO3·−, which had stronger oxidizing properties. With the addition of NO3− and HA, they had a promoting effect on the degradation of FEX but had an inhibitory effect on the degradation of CER. For HA, the two different results may be because HA containing a large number of active groups could compete with the drugs for the available chlorine in the system, resulting in a decrease in the degradation efficiencies of the drugs. On the other hand, HA had good adsorption properties that could lead to great removal efficiencies of the drugs.

Figure 5 shows the effects of different aqueous matrices on degradation efficiencies, and Table 5 lists the kinetics parameters of the photolysis process.

Fig. 5
figure 5

The influence of aqueous matrices on UV irradiation

Full size image
Table 5 Total light-shielding effects of different aqueous matrices
Full size table

From Table 5, we can find that the total light-shielding coefficients of different aqueous matrices were all less than 1, indicating that the addition of aqueous matrices had light-shielding effects on the photolysis of drugs. However, k' and k were not equal, indicating that the aqueous matrices not only affected the photolysis process through the light-shielding effect but also through the removal and excitation of active species (Ping et al. 2021). Combining Fig. 5 and Table 5, the degradation efficiencies showed slight fluctuations with the addition of HCO3−. For LOR and FEX, HCO3− was a scavenger of ·OH, which reduced the amount of ·OH and had inhibitory effects on their degradation. (Formula S6) (Lin et al. 2020). In addition, we found that the contribution rates of ·OH were relatively low and that the addition of HCO3− did not have significant inhibitory effects on the drugs. For CER, the addition of HCO3− actually promoted the reaction because HCO3− produced CO3·− under ultraviolet irradiation, which had a better degradation effect on CER (Ji et al. 2012b).

We clearly observed that NO3− and HA had significant inhibitory effects on photolysis, and the degradation efficiencies decreased as the concentrations of NO3− and HA increased. With the decrease in the total light-shielding coefficient, the shielding effects of NO3− and HA on the photolysis of the drugs will be greater. The addition of NO3− and HA could compete with the drugs for the absorption of photons, and the decrease in absorbable photons resulted in lower degradation efficiencies. The inhibitory mechanisms of HA can react with ANT·+ produced by the reaction of ANT and free radicals, and ANT·+ is then transformed to the parent form (ANT) (Formulas S7-S8) (Wang et al. 2017).

Acute toxicity changes of the drugs

The samples were collected at intervals to test the acute toxicities of the three drugs, and the results are shown in Fig. 6.

Fig. 6
figure 6

Acute toxicity of drugs in different oxidation systems (When pH = 7, FEX reacted too fast and was completely degraded within 20 s, so we did not measure the acute toxicity of FEX in the UV-NaClO system.)

Full size image

Figure 6 shows that the luminous intensity showed a weakening trend in most cases, indicating that the toxicity of the drug degradation intermediates was greater than that of their parent drugs. For a certain point in time, the increase in relative luminous intensity was due to the generation of electrophilic groups such as carbonyl groups during the degradation process, which led to reduced toxicity (Cronin et al. 2001). Some research has reported the overall toxicity of antihistamines: Gadipelly et al. (2016) studied the luminescence intensity of Photobacterium phosphoreum in samples taken at different time intervals during the degradation of CER by UV/S2O82− and found that the toxicities of the samples first increased and then decreased; Borowska et al. (2016) found that the ozone-treated FEX sample (containing parent drugs and the products) inhibited the luminosity of Aliivibrio fischeri by 10%. The previous results and this study indicated that the toxicity of advanced oxidation needs more consideration. In the two systems, NaClO and NaClO-UV, the degree of toxicity of the samples was greater than that in the UV system. It was speculated that the addition of NaClO resulted in the generation of disinfection byproducts, which were highly toxic to luminescent bacteria. Based on our research, UV had the advantages of high disinfection efficiency and high safety. Therefore, we recommend the use of the UV method in actual projects.

Drugs degradation intermediates and pathways

Figure 7 shows the UV spectrum of the drug solution over time by NaClO oxidation and UV irradiation. The characteristic absorption peaks of LOR, FEX, and CER were approximately at 247 nm, 220 nm, and 230 nm, respectively. As the reaction proceeded, the absorbance of the three drugs at the characteristic absorption peaks gradually weakened, but new absorption peaks appeared. This indicated that the drugs were transformed into their degradation products but not completely mineralized during this process. Comparing NaClO oxidation and UV irradiation, it was found that not only did UV irradiation have stronger effects on drug degradation but also the wavelengths of the new absorption peaks generated by the oxidation of NaClO (at approximately 300 nm) were greater than those produced by UV irradiation (at approximately 250–270 nm). This was because the strongly oxidizable electrophiles such as hypochlorous acid and hypochlorite present in the NaClO solution system could react with the drugs to form chlorinated products. When chlorine atoms containing n electrons were connected with the ring, they could form a conjugated system that enhanced the color generation ability of the ring and caused the ultraviolet absorption to move to 300 nm in the redshift direction (Cai et al. 2013; Zhang et al. 2014).

Fig. 7
figure 7

Evolution of the UV spectrum of the drug solution over time during NaClO oxidation and UV irradiation

Full size image

Since UV scanning only gave macroscopic signals of drugs with functional groups such as benzene rings, their specific degradation pathways need tobe inferred by identifying their degradation products. The photolysis products of LOR and FEX were reported in previous articles (Iesce et al. 2019; Breier et al. 2008). Therefore, this study only identified the photolysis products of CER and the NaClO oxidation products of these three drugs.

Both primary and secondary mass spectrometry are widely used to analyze the molecular structure of intermediates (Soufan et al. 2013; Vione et al. 2009; Yassine et al. 2017). According to the results of MS1 and MS2 given by Q-TOF, some potential structures of the intermediates are presented (Table. S1-S4) and the degradation pathways are proposed.

  1. (1)

    NaClO oxidative degradation pathways of LOR (as shown in Fig. 8)

Fig. 8
figure 8

Possible degradation pathways of LOR oxidized by NaClO

Full size image

Pathways I and II

As a functional group rich in electron groups, the carbon–carbon double bond (C = C) is vulnerable to attack by electrophiles. Therefore, the C = C in the LOR molecule was oxidized and cracked to form a ketone (Yu and Zhao 2019), forming cpd-1. cpd-2 was found to be an isomer of the LOR molecule, and the unsaturated double bond could react with HOCl to form cpd-3.

Pathway III

It was speculated that cpd-4 was produced from the rupture of the bond between C1 and C2 of LOR. cpd-5 was formed by the further breaking of the C–Cl bond and the O3-C2 bond in the cpd-4 molecule. Armaković et al. (2016) calculated that the C–Cl bond dissociation energy (BDE) for LOR was 91.26 kcal/mol. Their study also showed that the aliphatic part of LOR was prone to bond rupture, and the BDEs between C1 and C2 and between O3 and C2 were relatively low, among which the weakest bond existed between O3 and C2. Therefore, we speculated that pathway III was the dominant pathway. In addition, the cpd-5 molecule was further decarboxylated to form cpd-6 due to the instability of the carbamate. Desloratadine, the human metabolite of LOR, was not found in the product identification. This may be due to the C–Cl bond fracture of desloratadine, which promoted the formation of cpd-6 in another way.

Pathway IV

In the process of degrading organic compounds with NaClO, chlorine atoms may be replaced by hydroxyl groups to form cpd-7, and further fracture of O3-C2 occurred to form cpd-8. In addition, cpd-5 may be formed due to the occurrence of side-chain fracture of cpd-8 during the reaction, which was another possible formation process of cpd-5. As an electron donor group, phenolic hydroxyl groups enriched with electrons facilitated substitution reactions between phenolic compounds and electrophiles such as HOCl (Liu et al. 2012). It was speculated that the chlorine in hypochlorous acid molecules attacked the positively charged carbon on the benzene ring and generated a variety of chlorine-substituted phenols (Hu et al. 2002). Therefore, a chlorine substitution reaction could occur in the benzene ring to generate compound cpd-9.

  1. (2)

    NaClO oxidative degradation pathways of FEX (as shown in Fig. 9)

    Fig. 9
    figure 9

    Possible degradation pathways of FEX oxidized by NaClO

    Full size image

Pathway I

There is a tertiary amine in the FEX molecule, and pathway I may be the dominant. Nitrogen in organic compounds was the main site for possible attack by oxidizing agents (Dyakonov et al. 2010). The oxidant attacked the tertiary amine nitrogen to form nitrogen oxide during the reaction, namely cpd-1. Then, cpd-1 was oxidized to a ketone to form cpd-2 with 2 Da less in the molecule.

Pathway II

cpd-3 was formed by the removal of two methyl groups from the FEX molecule. It was reported that certain types of substrates (various monohydroxybenzoic acid species) underwent halogenated decarboxylation during chlorination. This reaction was attributed to the direct action of the hydroxyl groups in the substrate structure on the carbon atoms of the carboxyl groups (Larson and Rockwell 1979). Therefore, decarboxylation and demethylation reactions occurred in cpd-3, and the part that is removed may be replaced by chlorine to form cpd-4.

Pathway III

Product cpd-5 was obtained by oxidizing the secondary hydroxyl group in the FEX molecule to form a benzoyl group (Henning et al. 2019). It was further speculated that cpd-5 underwent two processes: (1) demethylation to produce cpd-6, and (2) dealkylation of the tertiary amine to form cpd-7 and cpd-8. The hydroxylation reaction may occur during the chlorination process (Yang et al. 2019), leading to the formation of cpd-9, or chlorine attacking the amine functional group to form cpd-10. The reaction efficiency of a secondary amine and HOCl is much higher than that of a tertiary amine and HOCl (Boreen et al. 2003), so the formation of cpd-10 more easily occurred.

  1. (3)

    NaClO oxidative degradation and photolysis pathways of CER (shown in Fig. 10 and Fig. 11, respectively)

    Fig. 10
    figure 10

    Possible degradation pathways of CER oxidized by NaClO

    Full size image
    Fig. 11
    figure 11

    Possible degradation pathways of CER photolysis

    Full size image

Due to its strong oxidation capacity and structural characteristics, NaClO mainly reacts with organic compounds by oxidation, addition to unsaturated bonds, and electrophilic substitution (Ji et al. 2016). However, the degradation mechanisms of photolysis were completely different from those of NaClO oxidation. After absorbing photons, the substance was excited into a triple excited state, which was not stable. The products were then formed through reactions such as cracking and hydrogen extraction. In addition, light could produce active species such as 1O2 and HO·, which could form products through hydroxylation, ketonization, and other reactions (Gligorovski et al. 2015).

By comparing the products in Fig. 10 and Fig. 11, we found that the NaClO oxidative degradation and photolysis process of CER produced three groups of the same degradation products: cpd-1 and UV-1, cpd-2 and UV-2, and cpd-5 and UV-3.

For the first and second groups of the same products, cpd-1 and UV-1, cpd-2 and UV-2, their generation processes were different: in the chlorine oxidation process (Pathway Cl-I), C-N bond fracture occurred on the tertiary amine of CER, and the dealkylation of N1-amine led to the formation of carbonyl compounds and secondary amines, forming cpd-1 and cpd-2. In the photolysis process (UV-I), 1O2 generated in the aqueous solution attacked the tertiary amine in the CER molecule, resulting in the production of UV-1 and UV-2. From the previous analysis, 1O2 played the more obvious role in the photolysis of CER than 3ANT* and ·OH, so it could be inferred that UV-I was the main degradation reaction in the process of photolysis.

For cpd-5 and UV-3, the third group of the same products, the para-position of the chlorine substituent was active and easily oxidized (Pathway Cl-II), which led to the breaking of the carbon–carbon single bond and the removal of the chlorine-containing benzene ring to form cpd-5 (Li et al. 2019). In the photolysis process (Pathway UV-II), CER was transformed into a triple excited state (3CER*) after absorbing photons. 3CER* formed UV-3 through the fracture of chlorine-containing benzene rings and the hydrogen extraction reaction process.

The remaining products of chlorine oxidation and photolysis were different. For the chlorine oxidation pathway Cl-I, cpd-2, containing a secondary amine and a tertiary amine group, could be further oxidized on the nitrogen atom, which formed the corresponding N-oxide on the tertiary amine or the corresponding hydroxylamine on the secondary amine, namely cpd-3 and cpd-4.

Chlorine oxidation pathway Cl-III: The N2-amine in the CER molecule was dealkylated, and then the chlorine atom was replaced by the hydroxyl group to form cpd-6. The secondary amine was oxidized to form the corresponding hydroxylamine and subsequently to form cpd-7.

Chlorine oxidation pathway Cl-IV: The composition of NaClO in aqueous solution is complex, and the form of hypochlorous acid could decompose and generate new atomic oxygen [O], which has strong oxidizing properties (Sun et al. 2019). In addition, many studies have shown that tertiary amines can react with oxidants to form nitrogen oxides, so the tertiary amines in the CER molecule may be attacked by [O]. Similar to FEX, nitrogen may be the main attack site of oxidizing agents, and we speculated that Cl-IV was the main pathway. The CER molecule contained two tertiary amines, and the oxidant could attack the tertiary amine to form nitrogen oxides. However, it was not clear which tertiary amine was attacked first based on the mass spectrum information. Borowska et al. (2016) reported that N2-amine was more active than N1-amine, so the N2-oxide of CER was easier to form than the N1-oxide, forming cpd-8. It is worth noting that chlorine oxidation of amine-containing compounds had a unique reaction. Chlorine attacks the N in the compounds to form chlorinated amines. CER contained tertiary amines, but in the process of product identification, the expected product was not found; because of reaction with thiosulfate, chlorinated amines could be converted back into the parent compound (Pinkston and Sedlak 2004).

Photolysis pathways UV-III and UV-IV: 3CER* cracked and removed carboxyl groups to form UV-4, and subsequent dealkylation resulted in the formation of UV-5. Under ultraviolet light irradiation, dechlorination may be the first reaction to occur (Zheng and Ye 2001). CER absorbed photons and then dechlorinated to form UV-6. Chlorine was an electron-withdrawing group; after being removed, the electron cloud density on the benzene ring increased, and it was easily attacked by HO· radicals (Li and Ma 2006). Hydroxylation occurred to form UV-7, and UV-8 was formed by decarboxylation of UV-7.

Limitations and future works

As an emerging advanced oxidation technology, UV-NaClO requires further research on its reaction mechanism and influencing factors, and research on the degradation pathway of drugs also needs to be improved. To accurately assess the main degradation pathways of drugs in the oxidation process of UV alone and NaClO alone, we need to use more analytical methods to verify the degradation intermediates and quantify them for further analysis. We found that the acute toxicity of the intermediate product has increased, which will limit the application of disinfection methods to a certain extent. The choice of indicator organisms also has a certain influence on the effect of toxicity. In addition, we will explore the effects of the complicated components in the actual water body on the oxidative degradation of drugs to provide more valid and reliable information for the water treatment process.

Conclusions

This study comprehensively compared the degradation efficiencies of antihistamine drugs (loratadine, cetirizine, fexofenadine) by three different oxidation methods. The results showed:

  1. (1)

    In the NaClO system, the degradation efficiencies were affected by the existing forms of the drugs and free chlorine, which changed with solution pH. As the pH increased, the degradation efficiencies of LOR decreased from 11% to 3%, while FEX increased from 11% to 92%. CER had better degradation efficiencies under strong acid and strong alkaline conditions, which were greater than 85%. Due to the photosensitivity of the three drugs, their degradation efficiencies were more than 75% with UV irradiation within 150 s at a pH of 7, which were far more than those of NaClO oxidation. Among the three oxidation methods, UV-NaClO had the best effect on drug degradation under most conditions.

  2. (2)

    In the process of photolysis, 1O2 played an important role. For FEX and CER, the contribution rates of 1O2 reached 70%; for LOR, the contributions of 3ANT* and 1O2 were very close, and the rates were 49.7% and 46.2%, respectively.

  3. (3)

    Aqueous matrices had different influence on the oxidation efficienciesof the drugs. The influence of aqueous matrices on the UV system was more significant than that on the NaClO system. The aqueous matrices affected not only the photolysis process through the light-shielding effect but also the removal and excitation of active species.

  4. (4)

    The drugs could not be completely mineralized, and the toxicities of the degradation intermediates were even higher than those of the parent drugs. Several degradation pathways were first proposed by the identification of degradation intermediates.