Identification of novel disinfection byproducts in pool water: Chlorination of the algaecide benzalkonium chloride

Chemosphere 239 (2020) 124801

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Identification of novel disinfection byproducts in pool water: Chlorination of the algaecide benzalkonium chloride Nikolay V. Ul'yanovskii a, Dmitry S. Kosyakov a, *, Ilya S. Varsegov a, Mark S. Popov a, Albert T. Lebedev a, b, ** a b

Core Facility Center “Arktika”, Northern (Arctic) Federal University, Arkhangelsk, 163002, Russia Department of Organic Chemistry, Lomonosov Moscow State University, Moscow, 119991, Russia

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


Benzalkonium (BAC) algaecide gives a number of DBPs during chlorination.
Mono- and polychlorinated, oxo and hydroxy BAC derivatives were found in pool water.
Chlorination proceeds via radical substitution in the long aliphatic chains of BAC.
MS/MS allowed reliable elucidation of the structures of novel DBPs.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 July 2019 Received in revised form 26 August 2019 Accepted 6 September 2019 Available online 7 September 2019

The maintenance of public swimming pools requires numerous technological steps. One of the most important issues involves microbiological safety. Benzalkonium chloride (BAC) encompasses homologous alkylbenzyldimethylammonium chlorides with various alkyl chains, in particular C12 and C14, and is known as a popular algaecide for keeping water clean. In addition to BAC, NaOCl and UV-irradiation are also used to treat pool water as additional technological steps. Therefore, BAC itself can become a precursor of disinfection byproducts (DBPs). High-performance liquid chromatography e tandem mass spectrometry (HPLC-MS/MS), with accurate mass measurements, has allowed the discovery of several groups of DBPs that are related to BAC in public pool water in Arkhangelsk (Russia). These DBPs include numerous isomeric monochlorinated derivatives ([C21H37ClN]þ and [C23H41ClN]þ), hydroxyl derivatives ([C21H38NO]þ and [C23H42NO]), carbonyl ([C21H36NO]þ and [C23H40NO]þ), and dicarbonyl derivatives ([C21H34NO2]þ and [C23H38NO2]þ). In addition, chlorinated alcohols, ketones and ketoalcohols of BAC were also detected, including [C21H35ClNO]þ, [C21H37ClNO]þ and [C21H35ClNO2]þ for BAC-12; and [C23H39ClNO]þ, [C23H41ClNO]þ and [C23H39ClNO2]þ for BAC-14. MS/MS allowed reliable elucidation of the structures of novel DBPs, proving that chlorination starts via radical substitution in the long aliphatic chains of BAC. UV-irradiation dramatically accelerates the reaction completely destroying the original compounds in less than an hour, while the array of the intermediate products remains the same as in the dark. The formation of other DBPs proceeds due to further reactions of these primary products. The concentrations of novel DBPs in pool water reach mg L1 levels. These conclusions were proved by conducting model reactions of BAC with NaOCl. © 2019 Published by Elsevier Ltd.

Handling Editor: Xiangru Zhang Keywords: Benzalkonium chloride Algaecide LC-MS Disinfection byproducts Chlorination Pool water

* Corresponding author. Core Facility Center “Arktika”, Northern (Arctic) Federal University, Arkhangelsk, 163002, Russia. ** Corresponding author. Department of Organic Chemistry, Lomonosov Moscow State University, Moscow, 119991, Russia. E-mail addresses: [email protected] (D.S. Kosyakov), [email protected] (A.T. Lebedev). https://doi.org/10.1016/j.chemosphere.2019.124801 0045-6535/© 2019 Published by Elsevier Ltd.

2

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

1. Introduction The maintenance of modern public swimming pools involves a number of compulsory steps. The microbiological safety of the water of public pools requires treatment with chemical disinfecting agents and, in some cases, UV-irradiation (Tang et al., 2015; Cimetiere and Laat, 2014; Yang et al., 2016). The most popular agents for this purpose are compounds that contain “active” chlorine, such as sodium and calcium hypochlorites, chloramine and trichloroisocyanuric acid (Chowdhury et al., 2014; Carter and Joll, 2017). The interactions of “active” chlorine with dissolved organic matter from natural or man-made sources (natural organic matter, biological fluids, pharmaceuticals, personal care products, and reagents used in water treatment) unavoidably result in the formation of a wide range of DBPs (Kim et al., 2002; Weisel et al., 2009; Yeh et al., 2014). Many of these DBPs may possess carcinogenic, teratogenic, and mutagenic properties, and may irritate the skin, the respiratory system, and mucous membranes (Manasfi et al., 2017; Teo et al., 2015; Kelsall and Sim, 2001; Bessonneau et al., 2011; Richardson, 2003, 2012; Richardson et al., 2002; Richardson and Postigo, 2016). Trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs), chloramines and N-nitrosamines, as well as simple halogenated aldehydes and ketones, are the most important DBPs and are regulated worldwide. Their typical levels in disinfected pool water range from units to hundreds of mg L1 (Richardson et al., 2010). Nevertheless, more than 700 DBPs have been reported in the literature (Richardson and Postigo, 2016), and their number is increasing constantly due to the application of advanced analytical tools including comprehensive GC-HRMS, GC  GCHRMS, and HPLC-HRMS (hereinafter, the abbreviations of analytical methods see in the Supporting information) with atmospheric pressure ionization (Richardson et al., 2010; Li et al., 2017; Richardson and Postigo, 2018) that enable working with nonvolatile and thermolabile DBPs. Special attention has been given recently to nitrogen-containing disinfection byproducts (N-DBPs) due to their higher toxicity and reactivity compared to other DBPs (Yin et al., 2018; Plewa et al., 2008; Carter et al., 2018; Liu and Zhang, 2013; Yang and Zhang, 2013). Recently, using high-resolution mass spectrometry, new NDBP classes have been discovered in pool, spa, and drinking (tap) waters, including bromoimidazoles (Daiber et al., 2016), halogenated fatty amides (Kosyakov et al., 2017a) and N-chlorinated dipeptides (Huang et al., 2017a, b). In addition to biological liquids and fluids (Richardson et al., 2010), dissolved humic substances act as an important natural DBP precursor leading to numerous products (Gonsior et al., 2014) including chlorinated pyridine and pyrrole derivatives (Kosyakov et al., 2017b). Quaternary ammonium salts possess strong algaecide properties and are widely used to maintain pool water quality (Ash and Ash, 2009) and may also be precursors of N-DBPs. Among them, the most widely used for the treatment of swimming pool water is

benzalkonium chloride (BAC), which represents a group of homologous n-alkylbenzyldimethylammonium chlorides with 8e18 carbon atoms in the alkyl chain (Fig. 1) (Ford et al., 2002). In the extensive list of BAC formulations that are commercially available, the most frequently used mixture consists of two homologues with 12 and 14 carbon atoms in the aliphatic chain (BAC-12 and BAC-14). Occasionally, BAC-16 is used as well. The selection of a particular chain length depends on the selective influence of that chain on various microbial species (Ash and Ash, 2009; Ford et al., 2002). Taking into account that the scheme of pool water treatment involves, along with the use of algaecide, the periodic addition of new portions of chlorinating agents and UV-irradiation (Fig. S1) BAC itself becomes a potential precursor of DBPs. Nevertheless, despite the wide use of BAC together with agents containing “active” chlorine in the treatment of pool water, there is little information regarding its aqueous chlorination. Chang et al. (Chang et al., 2011; Chang and Wang, 2011) showed using GC-MS that BAC may transform into the highly toxic N-nitrosodimethylamine. Furthermore, high levels of THMs and HAAs were reported as well. Recently, Huang et al. (2017a, b), studying the advanced oxidation process of BAC involving simultaneous UV-irradiation and “active” chlorine treatment to process waste water, proposed a scheme for its transformation. In addition to several intermediate products that form due to benzyl-nitrogen bond cleavage, the authors demonstrated that the benzalkonium ion may be a precursor to known DBPs, including trichloromethane, chloral hydrate, di- and trichloropropanones, and dichloroacetonitrile. The authors thereby attracted attention to the need to control those DBPs. The relatively intense conditions of the applied UV/chlorine treatment probably did not allow studying initial processes of BAC transformation under actual pool water conditions. The present study was inspired by the discovery of novel chlorinated N-DBPs that are related to BAC in public swimming pool water in Arkhangelsk (Russia), where the BAC algaecide is used. The current research aims to identify these products and study the mechanism of BAC transformation under conditions of aqueous chlorination. 2. Materials and methods 2.1. Chemicals and reagents Benzalkonium chloride (>95%), consisting of BAC-12 (73.5%) and BAC-14 (24.6%), was purchased from Sigma-Aldrich (Steinheim, Germany) and was used as a standard for quantification as well as a model substrate for chlorination. Potassium permanganate (chem. pur.), hydrochloric acid (chem. pur., 38%), sulfuric acid (chem. pur.) and sodium hydroxide (chem. pur.) were purchased from Neva-Reaktiv (St. Petersburg, Russia). Sodium phosphate dibasic, potassium phosphate monobasic (ACS reagent, >99.0%, Sigma-Aldrich, Steinheim, Germany), and orthophosphoric acid (chem. pur., Komponent-Reaktiv, Moscow, Russia)

Fig. 1. Structural formula of the benzalkonium cation and some physico-chemical properties of BAC.

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

were used for the preparation of buffer solutions with predetermined pH values. Supergradient grade acetonitrile (PanReac, Barcelona, Spain), formic acid (ACS reagent, puriss p.a., Seelze, Germany) and ultrapure Milli-Q water were used for preparing the mobile phases. Gradient grade methanol (J.T. Baker, Phillipsburg, USA) was used as an elution solvent in solid phase extraction (SPE) procedures. Gaseous chlorine was obtained by reacting hydrochloric acid with potassium permanganate and was purified from trace HCl by passing through water (Brauer, 1963). A solution of sodium hypochlorite was obtained by bubbling gaseous chlorine through a 20% solution of sodium hydroxide at 0  C. The active chlorine content in the solution obtained was 22.5 g L1 as measured by iodometric titration (White et al., 1999). 2.2. Water samples and solid phase extraction The water samples were taken in November 2018 from an indoor swimming pool at the Northern (Arctic) Federal University (Arkhangelsk, Russia), that was treated with sodium hypochlorite and Algitin BAC-based algaecide (Marcopool Chemicals, Moscow, Russia) contained ~17% of BAC-12 and BAC-14 mixture using automatic dosage and pH adjustment systems, as well as a UVtreatment step (Fig. S1). According to manufacturer's recommendations, the algaecide dosage was 3 g m3 (0.5 mg L1 calculated for BAC) with repetitive treatments every 3 days. Sample 1 was collected from the pool water recycle line prior to the UV treatment and introduction of a disinfecting reagent and algaecide. Sample 2 was collected directly from the pool near the inlet of purified and disinfected water. The active chlorine concentrations determined by iodometric titration in the Samples 1 and 2 were 0.41 and 0.45 mg L1, respectively. For the solid phase extraction (SPE) and concentration of DBPs, 3 mL cartridges containing 100 mg of functionalized styrenedivinylbenzene polymeric sorbent Bond Elut PPL (Agilent Technologies, Santa Clara, USA) were used that provided good recovery rates for the wide range of moderately polar and nonpolar compounds (Gonsior et al., 2014; Phungsai et al., 2016). The recovery rates of BAC in the extraction procedure used in our experiments were 69 ± 2% and 72 ± 2% for BAC-12 and BAC-14, respectively. These values were determined using 200 mg L1 standard solution. Extraction was carried out using an automatic SPE system Sepaths UP (LabTech, Sorisole, Italy). Before extraction, SPE cartridges were conditioned consecutively with 10 mL of methanol and 10 mL of ultrapure water. One liter of pool water was acidified to pH 2e2.5 with sulfuric acid, passed through an SPE cartridge at a flowrate of 5 mL min1 and then washed with 10 mL of ultrapure water. After drying in nitrogen (99.99%) for 60 min the adsorbed analytes were eluted with 10 mL of methanol into glass vials. The obtained eluate was evaporated at 40  C to a volume of 1 mL under a stream of nitrogen (99.999%) in glass conical vials using a ReactiTherm/Reacti-Vap sample concentrator (Thermo Scientific, Bellefonte, USA). Simultaneously, a blank sample was prepared from 1 L of ultrapure water using the same procedure. Prior to analysis, the extract was centrifuged and the supernatant was transferred to a new glass vial and then injected into an HPLC-MS system. Each sample and its corresponding blank were processed and analyzed in triplicate. 2.3. Benzalkonium chlorination procedure A stock solution of BAC was prepared in ultrapure water (total concentration of homologues: 1 g L1) on a weight basis. Working solutions (20 mL) of BAC containing 10 mg L1 of BAC homologues at pH 7.20, 6.13 and 5.10 were prepared in 60 mL screw-cap glass

3

vials with PTFE septa by diluting the stock solution with 100 mM phosphate buffer. Sodium hypochlorite was added to each working solution to obtain an active chlorine concentration of 100 mg L1. The reaction was conducted in the dark at an ambient temperature (20 ± 1  C) for 18 days with continuous shaking. Similar solutions without chlorinating agents were maintained under the same conditions and served as reference samples. Periodically, 200 mL portions of reaction mixtures and reference samples were withdrawn, centrifuged and then injected into the HPLC-HRMS system. To study the effect of UV irradiation and conduct kinetic measurements the BAC working solution at pH 5.1 prepared by the procedure described above after addition of sodium hypochlorite was placed in a 15-mL borosilicate glass micro photochemical reactor with jacketed quartz immersion well (Ace glass, Vineland, USA), equipped with a quartz Pen-Ray UV (254 nm) 5.5 W low pressure mercury lamp (Analytik Jena, Upland, USA), and thermostated at 20  C. At specified intervals, 1-mL samples were taken. After quenching the reaction by adding 20 mL of 0.01 M sodium thiosulfate solution, they were subjected to HPLC-HRMS analysis. GC-MS was used for identification of volatile and semivolatile products of BAC chlorination after liquid-liquid extraction of sample aliquots. Five milliliter sample was taken 5 days after the start of the reaction and extracted sequentially at pH 2 and pH 11 (addition of sulfuric acid and sodium hydroxide solutions) with two portions (1 mL each) of dichloromethane, which is similar to the extraction procedure used in the EPA 8270 method (U.S. EPA 8270, 2007). The extracts were combined and then evaporated to 500 mL, as described in a previous section, and then injected into the GC. 2.4. Liquid chromatography e mass spectrometry analyses Analyses of the model samples and pool water extracts were carried out using a TripleTOF 5600þ high-resolution quadrupole time-of-flight (Q-TOF) mass spectrometer (AB Sciex, Concord, Canada) with a Duospray ion source combined with an LC-30 Nexera HPLC system (Shimadzu, Kyoto, Japan) consisting of a DGU-5A vacuum degasser, two LC-30AD chromatographic pumps, an SIL-30AC autosampler, and an STO-20A column thermostat. Chromatographic separations were carried out at 40  C on a Nucleodur PFP column (Macherey-Nagel, Duren, Germany) using a pentafluorophenyl-propyl stationary phase, 150  2 mm, particle size 1.8 mm. A mixture of water (A) and acetonitrile (B) containing 0.1% formic acid was used as a mobile phase. The gradient program was as follows: 0e1 min: 10% B, 1e25 min: linear gradient of B to 100%, 25e40 min: 100% B. The mobile phase flow rate was 0.25 mL min1 and the injection volume was 2 mL. Nontargeted screening of DBPs was performed in an information dependent acquisition (IDA) mode using positive electrospray ionization (ESIþ). The following ion source parameters were used: nebulizing (GS1) and drying (GS2) gas pressure 40 psi, curtain (CUR) gas 30 psi, capillary voltage (ISVF) 5500 V, and source temperature (TEM) 300  C. The parameters used for recording the mass spectra in a TOF MS mode were as follows: declusterization potential (DP) 80 V, m/z range 100e1000, and acquisition time 150 ms. Tandem (MS/MS) mass spectra were recorded using collisioninduced dissociation (CID) for precursor ions with signal intensities above a threshold of 100 cps. Nitrogen was used as the collision gas, collision energy (CE) 40 eV with a spread (CES) 20 eV. The maximum number of precursor ions simultaneously subjected to CID 15, m/z range 20e1000. Mass scale calibrations in the MS and MS/MS modes were performed prior to every run in an automatic regime using a sodium formate solution as a standard. Data processing was performed using MasterView and Formula Finder (AB Sciex, Concord, Canada) software packages. Elemental

4

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

compositions of the detected compounds were determined on the basis of the accurate masses of ions, their isotopic distributions, and product ions m/z in the MS/MS spectra. The following constraints were applied: maximal number of atoms: C e 100, H e 300, O e 20, N e 10, Cl e 10, Br e 5; mass error <5 ppm (MS) and <10 ppm (MS/ MS); signal-to-noise ratio (S/N) > 10. For the quantification of BAC-12 and BAC-14 in the pool water, extracted ion chromatograms (XIC) of m/z 304.300 and m/z 332.331 with 2 mDa mass window were used. To construct calibration curves, aqueous standard solutions of the mixture of BAC-12 and BAC-14 were prepared immediately prior to analysis and were used in a concentration range of 1e5000 mg L1. 2.5. Gas chromatography e mass spectrometry analyses The identification of the volatile and semivolatile products that formed in BAC chlorination experiments was performed using an Orbitrap Exactive GC high-resolution GC-MS system (Thermo, Waltham, USA) with an orbital ion trap mass spectrometer and a Trace 1310 gas chromatograph equipped with a TriPlus RSH autosampler and split/splitless injector. The separations were performed on a TG-5SILMS capillary column (Thermo, Waltham, USA), 30 m  0.25 mm (I.D.) and a film thickness of 0.25 mm. The carrier gas (He, 99.9999%) flow rate was 1.2 mL min1. The thermostat program was as follows: initial temperature 50 С for 3 min, ramp at 7 С min1 to 300 С, and hold at 300 С for 10 min. The injection volume was 1 mL with a 5:1 split. The injector and transfer line temperatures were 280 and 300  C, respectively. MS detection was performed in a m/z range of 35e550 with a spectral resolution of 60,000 (at m/z 200) and a C-trap filling target value (AGC target) of 5$105. Electron ionization (70 eV) with an ion source temperature of 200 С was used. For confirmation of molecular formulae, the GC-MS analyses were repeated using positive chemical ionization (PCI) with methane reagent gas (1.5 mL min1) at 150 С and a 30 eV ionization energy. 3. Results and discussion 3.1. Pool water targeted analysis The results of the analyses showed the identity of water samples taken at two chosen sample points within the typical error of the analytical method. Two major peaks (24% of the total area) dominate the HPLC-ESI(þ)-Q-TOF HRMS total ion current (TIC) chromatograms (Fig. S2) of the swimming pool water extracts. Their retention times (tR) are 14.74 and 16.14 min. These peaks are attributed to two individual compounds with molecular ions of m/z

304.3006 ([C21H38N]þ) and 332.3323 ([C23H42N]þ) and are identified as cations of benzalkonium BAC-12 and BAC-14, respectively (see XIC in Fig. 2). The concentrations of these compounds in the samples (taking into account determined recoveries with SPE) were 232 ± 6 and 94 ± 3 mg L1 respectively, making these compounds potential precursors for the formation of DBP and N-DBP. A targeted analysis of possible chlorinated products of BAC using the exact masses of [C21H37ClN]þ and [C23H41ClN]þ ions demonstrates the corresponding peaks in XIC, while their intensity reaches 0.5e1.0% of that of the original BAC (Fig. 2). Taking into consideration that benzalkonium ions as well as their chlorinated derivatives exist in solution as free cations with very similar properties, their yields in ESI (þ) may be treated as identical. Therefore, concentrations of monochlorinated species of BAC-12 and BAC-14 may be tentatively estimated using the BAC standard. Calculations show that their levels are 1.3 and 0.56 mg L1, respectively, and are comparable with the levels of various known N-DBPs in tap water and pool water worldwide (Carter and Joll, 2017; Kosyakov et al., 2017a; Manasfi et al., 2016). The shape of the monochlorinated BAC chromatographic peaks (Fig. 2b) indicates the presence of possible isomers, which cannot be completely separated due to their close retention times. The decrease in the tR values for these compounds using reversedphase chromatography, in comparison with the original BAC, indicates an increase in polarity after the introduction of a chlorine atom into the molecule. The calculations of the LogP values using an ACD/Labs software package (Advanced chemistry development Inc., Toronto, Canada) suggests the introduction of chlorine into the BAC aliphatic chain rather than into the aromatic ring. CID spectra of the detected chlorinated compounds BAC-12 and BAC-14 (Fig. 3) prove the above proposal, thereby reliably identifying their structures (except for the position of chlorine atoms in the aliphatic chain). A full fragmentation scheme for the monochlorinated BAC precursor ion is presented in the supporting information (Fig. S3). The [C21H37ClN]þ precursor ion, by losing an HCl molecule, gives rise to the primary ion [C21H36N]þ (m/z 302). An alternative fragmentation pathway involves the rupture of the BzeN bond. Direct cleavage of that bond results in the formation of a stable tropylium cation [C7H7]þ (m/z 91), while an accompanying hydrogen migration from the aliphatic substituents to the benzyl moiety with the loss of a C7H8 molecule forms a chloro-N,Ndimethyldodecan-1-iminium ion [C14H29ClN]þ (m/z 246). The latter ion also readily loses an HCl molecule, leading to the secondary product-ion [C14H28N]þ (m/z 210). Alternatively, the ion fragments with a split of the a-b CeC bond, forming a rather stable N,Ndimethylmethaniminium ion [C3H8N]þ (m/z 58). A similar array of fragment ions characterizes the mass spectrum of a chlorinated

Fig. 2. HPLC-ESI (þ)-Q-TOF HRMS XIC chromatograms of pool water BAC-12 and BAC-14 (A) and their monochlorinated derivatives (B).

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

5

Fig. 3. Tandem mass spectra of the monochlorinated derivatives of BAC-12 (A) and BAC-14 (B).

derivative of BAC-14, which differs from BAC-12 only by an increment of two additional methylene groups in the aliphatic chain. It is important to mention that neither chlorinated dimethylmethaniminium nor chlorinated benzylchloride (chlorotoluene) are present in either spectrum. Thus, the long alkyl chain represents the only moiety in the original molecules amenable to chlorination. That fact is unexpected, as alkanes (non-activated saturated alkyl chains) are the most stable moieties under the conditions of aquatic chlorination (Lebedev, 2007). The only possible reaction that could produce the observed result involves radical substitution of aliphatic hydrogens for chlorine. The mentioned process should produce an array of chlorinated products with chlorine atoms at various positions in the chain and with quite equal probability. Minor differences in their polarities may explain the peculiarities of the corresponding chromatographic peaks (Fig. 2) mentioned above. Competition of reactive centers in the reaction of aquatic chlorination is very important. It is difficult to predict the site of the primary attack (Lebedev, 2007; Trebse et al., 2016). It is the absence of other reactive centers in BAC molecules that makes radical substitution competitive. A benzene ring is not sufficiently activated to undergo aromatic electrophilic substitution (Tretyakova et al., 1994; Grbovic et al., 2013; Lebedev et al., 2004). Although benzyl methylene group is usually reactive (Lebedev et al., 2004), in the case of BAC it is attached to the positively charged N atom and is thus passivated. Where does the chlorine radical originate? Chlorination of interior swimming pool water usually occurs without light. However, the penetration of photons into the system is unavoidable.

One also cannot neglect the presence of dissolved oxygen, which may trigger radical reactions as well. Moreover, technical schemes for water treatment for the pool being studied (as well as for the majority of pools worldwide) involve additional UV-irradiation. Therefore, photons interact with Cl2 or HOCl molecules, which results in the formation of reactive radical chlorine species that can react with the CeH bonds of the aliphatic chain resulting in the formation of a hydrocarbon radical. Being unstable, the hydrocarbon radical immediately reacts with Cl2 or HOCl molecules causing the formation of the mentioned chlorination products or of hydroxyl derivatives (see below). The absence of BAC chlorinated in the a-position to the N-atom may involve steric effects, or their yields may simply be lower due to small number of exchangeable hydrogen atoms. Furthermore, the absence of the [C3H7ClN]þ ion in the CID spectra proves that the methylene group in the a-position in the aliphatic chain is also unavailable for a substitution reaction. Taking into account that the efficiency of chlorine radical formation increases under UV-irradiation, the formation of chlorinated BAC should benefit from the introduction of a UV-irradiation step in the pool water preparation scheme. Similarly, higher levels of these DBPs may be found in outdoor swimming pools, since they are subjected to intensive sun light. Laboratory experiments (see below) proved that conclusion. A targeted search for chlorinated BAC containing two or more chlorine atoms in pool water samples using XIC and accurate mass values was not successful. Therefore, these compounds are not formed, at least at levels greater than 10 ng L1 (estimated detection limit). The latter may be due to low concentrations of the

6

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

monochlorinated species, and consequently, low reaction rates for the chlorination reaction. Furthermore, primary organochlorines transform themselves via a nucleophilic substitution mechanism into the corresponding alcohols by reacting with water.

3.2. Non-targeted identification of BAC related DBPs Non-targeted screening of swimming pool water using HPLCESI(þ)-Q-TOF HRMS enabled the discovery of another class of DBPs that are related to BAC transformation. These DBPs have the following compositions: [C21H38NO]þ, [C21H36NO]þ and [C21H34NO2]þ for BAC-12, and [C23H42NO]þ, [C23H40NO]þ and [C23H38NO2]þ for BAC-14. The CID spectra of the DBPs (Table S1) are similar to the above discussed chlorinated derivatives and demonstrate fragment ions of N,N-dimethylmethaniminium, tropylium, and a corresponding oxygenated ion that arises from the loss of C7H8 from the precursor molecular ion. Therefore, these are benzalkonium cations with one hydroxyl group in the aliphatic chain ([C21H38NO]þ and [C23H42NO]þ), or similar structures containing one or two carbonyl groups. The variability of their position in the alkyl chain of dodecyl or tetradecyl groups significantly influences the molecules polarity, leading to a notable array of chromatographic peaks that are due to isomeric products (up to 11 for diketones) having the same mass (Fig. 4). An estimation of the areas of the chromatographic peaks demonstrates that the major components are keto derivatives (~30% of the original BAC content). Alcohols and diketones are formed to a lesser extent (approximately 12 and 9% of the original BAC content, respectively). One additional group of related derivatives was discovered during the study. These are hydroxyl and oxo BAC derivatives that contain one chlorine atom. For this group, it was possible to reliably assign three molecular formulas for BAC-12: [C21H35ClNO]þ, [C21H37ClNO]þ and [C21H35ClNO2]þ, and three for BAC-14: [C23H39ClNO]þ, [C23H41ClNO]þ and [C23H39ClNO2]þ. According to the degree of unsaturation, benzalkonium derivatives containing one oxygen atom are alcohols and ketones, while those containing two oxygen atoms are ketoalcohols. As with the oxygenated derivatives of BAC mentioned above, these species are present in pool water as an array of isomers containing corresponding groups (see XIC in Fig. S4). The concentration of these isomers in pool water (tentatively estimated level 0.15e0.30 mg L1) is approximately 10fold lower than that of the primary chlorinated products. Hydroxylated and monochlorinated benzalkonium ions are not reported in the literature as products of algaecide transformation during water treatment. In contrast, oxo derivatives were discovered earlier among the products of the oxidation of BAC with ozone

Fig. 4. XIC chromatograms of the oxygen-containing BAC derivatives from pool water.

in the presence of a catalyst, based on nickel oxide nano particles (Carbajo et al., 2016), as well as intermediate products (monooxo compounds) during BAC destruction by oxidation and with the simultaneous action of chlorine and UV-irradiation (Huang et al., 2017a, b). The results obtained in this work suggest a mechanism for the formation of the identified oxygenated BAC derivatives that involve the interaction of activated HOCl with hydrocarbon radical species or nucleophilic substitution of a chlorine atom for the hydroxyl group in the primary chlorinated BAC products (Fig. 5). Further oxidation of hydroxyl groups to carbonyl groups should be expected under the oxidative conditions of aquatic chlorination. Further transformation of the primary and secondary products of BAC aquatic chlorination involves numerous compounds with the following elemental compositions: [CnH2n-5ClNO]þ and [CnH2nþ 7ClNO2] , where n ¼ 13e20 (Table S2). According to their RDB (ring and double bond equivalent), it is quite possible to connect these products with BAC after the cleavage of CeC bonds in the aliphatic chain that is subjected to the radical initiated reactions. 3.3. Model chlorination of BAC To confirm the conclusions on the origin of novel DBPs in swimming pool water, model experiments on the chlorination of BAC-12 and BAC-14 mixtures were carried out under conditions (temperature, pH, and active chlorine/BAC ratio) close to that used in the pool water treatment. The higher absolute concentrations of BAC and active chlorine (10 and 100 mg L1, respectively) were used to avoid the pre-concentration step prior to analysis and, thus, to ensure the detection of widest range of the reaction products. Reaction of the studied compounds with active chlorine proceeds even in the absence of UV irradiation, although in that case the rate is rather low. Under those conditions, dissolved oxygen may be the most likely trigger for initiating the chain of transformation reactions. The initial concentrations of BAC decrease by an order of magnitude within 18 days and follow pseudo second-order kinetics (Fig. S5). At the same time, a decrease in the pH leads to a notable acceleration of the reaction due to an increase in the concentration of free chlorine in the solution (Table 1). The reaction products identified by HPLC-ESI(þ)-Q-TOF HRMS (Table 2) include monochlorinated BAC; hydroxy, oxo and dioxo derivatives; monochlorinated oxo-, hydroxy- and oxo-hydroxy derivatives as well as polychlorinated BAC containing 2e4 chlorine atoms. Except for the latter group, all of the indicated N-DBPs are identical to those found in pool water, i.e., their retention times, accurate masses and CID spectra were identical. The total number of peaks for isomeric oxygen-containing products in the XICs also coincides with those for the pool water extracts and model mixtures, despite a certain redistribution of intensities between them. Di-, tri- and tetrachlorinated BACs were found only among the products of model chlorination. It is likely that the absence of such compounds in pool water is due to the low probability of their formation in detectable amounts at low concentrations of active chlorine and primary monochlorides. However, that fact suggests the possibility of the formation of polychlorinated algaecide derivatives in great variety under certain conditions. As expected, the concentrations of polychlorinated derivatives decrease sharply with the introduction of each subsequent chlorine atom. Thus, tetrachlorinated BAC is formed in quantities that are 2e3 orders of magnitude lower than monochlorinated BACs (Table 2). The concentrations of major chlorination products of various classes in model solutions kept in dark reaches a maximum value after 4 days from the start of the reaction (Fig. S6). Afterwards, a decrease (most typical for lower pH) or stable levels are observed. Considering the continuous decrease in the concentration of the original BAC, the detected products may serve as intermediates and

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

7

Fig. 5. Tentative scheme for the transformation of BAC during aquatic chlorination.

Table 1 Observed second order rate constants for reaction of BAC chlorination in dark at 20  C. Compound

kobs, L mol1 h1 pH 7.2

pH 6.1

pH 5.1

BAC-12 BAC-14

340 1500

540 2300

1300 5400

undergo further transformation with degradation of the benzalkonium backbone. Lowering the pH in all cases leads to an increase in the concentrations of the products listed in Table 2. At the minimum pH value and reaction time of 4 days, the levels of formed mono- and dichloro-derivatives are approximately 10% and 5% of the corresponding initial BAC, respectively, while the oxygencontaining compounds are observed to be minor BAC transformation products in the range of 0.1e1.3%. This finding is in contrast to the results of the pool water analysis and can be explained by the different chlorination conditions in the pool water treatment system and model experiments. The conclusion on the radical mechanism of the BAC chlorination is confirmed by the results of a model experiment under the action of hard UV radiation. In this case, the composition of the resulting products turned out to be completely identical to that obtained in the dark (Table 2). Nevertheless, it is noteworthy, that the highest concentrations of chlorination products in the photochemical reactor at pH 5.1 are reached already 30 s after the start of

the reaction (Fig. S9), and their values are close to the maximal values achieved without UV-treatment (4 d after initiation of the reaction). This indicates an increase in the reaction rate of four orders of magnitude in the presence of light. It is worth mentioning that the extensive UV-irradiation leads to complete destruction of the original BAC and its derivatives in less than 1 h, which additionally proves the importance of photons as initiators of the radical reactions. In general, the identified products constitute less than half of the BAC consumed in the reaction. That observation suggests that the products will degrade further under aquatic chlorination conditions resulting in the formation of nonionic compounds. The effectiveness of their ESI ionization is orders of magnitude lower compared to benzalkonium cations. For the characterization of these products, we extracted the reaction mixtures 5 days after the start of the chlorination reaction in dark with dichloromethane and then performed GC-HRMS analysis. The obtained chromatograms comprise ~30 intense peaks (Fig. S7). The mass spectra of three major products that were identified in EI and PCI modes are presented in Fig. S8. All other chromatographic peaks demonstrate similar spectra. It is evident that molecular ions are absent, even when soft ionization is used. It is quite natural to chlorinated aliphatic oxygen containing compounds. Although the detailed structures of these products are not obtained and the possibility of their formation during the thermal degradation of the primary BAC chlorination products in the GC inlet cannot be completely ruled out, their detection proves the chlorination of the alkyl chain with any CeC bond cleavage at some

8

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801

Table 2 Products of BAC model chlorination in the dark (4 d after initiation of the reaction) and under UV irradiation (30 s after initiation of the reaction) identified by HPLC-ESI(þ)-QTOF HRMS. N

m/z measured (monoisotopic)

Elemental composition

m/z error, ppm

Assumed compound

Relative peak area, % of the original algaecide in dark

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

338.2618 366.2932 372.2228 400.2548 406.1843 434.2156 440.1454 468.1767 320.2955 348.3271 318.2800 346.3113 332.2595 360.2904 352.2411 380.2724 354.2569 382.2878 368.2362 396.2680

[C21H37ClN]þ [C23H41ClN]þ [C21H36Cl2N]þ [C23H40Cl2N]þ [C21H35Cl3N]þ [C23H39Cl3N]þ [C21H34Cl4N]þ [C23H38Cl4N]þ [C21H38NO]þ [C23H42NO]þ [C21H36NO]þ [C23H40NO]þ [C21H34NO2]þ [C23H38NO2]þ [C21H35ClNO]þ [C23H39ClNO]þ [C21H37ClNO]þ [C23H41ClNO]þ [C21H35ClNO2]þ [C23H39ClNO2]þ

2.6 2.7 2.3 3.9 3.3 3.2 3.2 2.9 2.3 2.8 2.8 2.3 3.3 1.9 2.6 2.4 3.1 1.8 3.1 4.1

stage. In fact, these semi-volatiles are complementary products to the [CnH2n-5ClNO]þ and [CnH2n-7ClNO2]þ charged products (Table S2).

Monochlorinated BAC-12 Monochlorinated BAC-14 Dichlorinated BAC-12 Dichlorinated BAC-14 Trichlorinated BAC-12 Trichlorinated BAC-14 Tetrachlorinated BAC-12 Tetrachlorinated BAC-14 HydroxyBAC-12 HydroxyBAC-14 OxoBAC-12 OxoBAC-14 DioxoBAC-12 DioxoBAC-14 Monochlorinated oxoBAC-12 Monochlorinated oxoBAC-14 Monochlorinated hydroxyBAC-12 Monochlorinated hydroxyBAC-14 Monochlorinated oxohydroxyBAC-12 Monochlorinated oxohydroxyBAC-14

under UV

pH 7.2

pH 6.1

pH 5.1

pH 5.1

3.89 4.87 1.20 1.10 0.22 0.23 0.02 0.01 0.49 0.69 0.35 1.22 0.02 0.05 0.11 0.21 0.28 0.38 0.02 0.03

8.21 8.53 3.07 3.49 0.69 0.99 0.08 0.17 0.87 1.16 0.50 1.05 0.06 0.10 0.22 0.38 0.64 0.97 0.07 0.12

10.3 11.0 4.69 5.44 1.25 1.84 0.15 0.42 1.05 1.23 0.70 1.03 0.11 0.18 0.38 0.61 0.90 1.19 0.13 0.23

7.1 5.2 3.2 2.4 1.1 0.97 0.26 0.33 1.1 1.3 1.3 1.2 0.06 0.02 1.2 0.83 0.91 0.82 0.15 0.14

Engineer V. Ageev for assistance with sampling and advice on the water treatment technology that is used. Appendix A. Supplementary data

4. Conclusions A brand new class of DBPs - halogenated and oxygenated derivatives of a popular all over the world algaecide benzalkonium chloride was discovered in an indoor swimming pool, where it is applied together with active chlorine species and UV-irradiation according to the sanitary and hygiene requirements. The reaction triggers by radical mechanism with substitution of hydrogen atoms in the long aliphatic chain for chlorine atoms. UV-irradiation accelerates the process dramatically. Interesting that only long aliphatic chains of the molecule are involved in the primary reaction, while benzylic moiety remains intact. The primary chlorinated species reacts further by the same mechanism, as well as by consecutive reactions of nucleophilic substitution and oxidation. The levels of the novel DBPs are in the range of mg L1. The structures of the novel DBPs were confirmed by MS/MS experiments and by carrying out an independent chlorination reaction of the standard benzalkonium chloride solution in various conditions. Toxicities of the novel DBPs remain unknown and require additional studies. Conflicts of interest The authors declare no conflicts of interest in relation to this research. Acknowledgments This research was performed using instrumentation at the Core Facility Center “Arktika” of Northern (Arctic) Federal University and was supported by the Russian Science Foundation (grant No. 17-1301112). The authors are grateful to the staff of the swimming pool at the Northern (Arctic) Federal University and personally to Chief

Supplementary data to this article can be found online at https://doi.org/10.1016/j.chemosphere.2019.124801. References Ash, M., Ash, I., 2009. Handbook of Preservatives. Synapse Information Resources Inc: Endicott, NY. ment, M., Thomas, O., 2011. Determinants of chloBessonneau, V., Derbez, M., Cle rination by-products in indoor swimming pools. Int. J. Hyg Environ. Health 215, 76e85. https://doi.org/10.1016/j.ijheh.2011.07.009. Brauer, G., 1963. Handbook of Preparative Inorganic Chemistry. Academic Press Inc, New York. , A., Leto  n, P., García-Calvo, E., Perdigo nCarbajo, J.B., Petre, A.L., Rosal, R., Berna  n, J.A., 2016. Ozonation as pre-treatment of activated sludge process of a Melo wastewater containing benzalkonium chloride and NiO nanoparticles. Chem. Eng. J. 283, 740e749. https://doi.org/10.1016/j.cej.2015.08.001. Carter, R.A.A., Joll, C.A., 2017. Occurrence and formation of disinfection by-products in the swimming pool environment: a critical review. J. Environ. Sci. 58, 19e50. https://doi.org/10.1016/j.jes.2017.06.013. Carter, R.A.A., Liew, D.S., West, N., Heitz, A., Joll, C.A., 2018. Simultaneous analysis of haloacetonitriles, haloacetamides and halonitromethanes in chlorinated waters by gas chromatography-mass spectrometry. Chemosphere 220, 314e323. https://doi.org/10.1016/j.chemosphere.2018.12.069. Chang, H., Chen, C., Wang, G., 2011. Identification of potential nitrogenous organic precursors for C-, N-DBPs and characterization of their DBPs formation. Water Res. 45, 3753e3764. https://doi.org/10.1016/j.watres.2011.04.027. Chang, H.H., Wang, G.S., 2011. Correlations between surrogate nitrogenous organic precursors and C-, N-DBP formation. Water Sci. Technol. 64, 2395e2403. https://doi.org/10.2166/wst.2011.823. Chowdhury, S., Al-Hooshani, K., Karanfil, T., 2014. Disinfection byproducts in swimming pool: occurrences, implications and future needs. Water Res. 53, 68e109. https://doi.org/10.1016/j.watres.2014.01.017. Cimetiere, N., Laat, J.D., 2014. Effects of UV-dechloramination of swimming pool water on the formation of disinfection by-products: a lab-scale study. Microchem. J. 112, 34e41. https://doi.org/10.1016/j.microc.2013.09.014. Daiber, E.J., DeMarini, D.M., Ravuri, S.A., Liberatore, H.K., Cuthbertson, A.A., Thompson-Klemish, A., Byer, J.D., Schmid, J.E., Afifi, M.Z., Blatchley, E.R., Richardson, S.D., 2016. Progressive increase in disinfection byproducts and mutagenicity from source to tap to swimming pool and spa water: impact of human inputs. Environ. Sci. Technol. 50, 6652e6662. https://doi.org/10.1021/ acs.est.6b00808. Ford, M.J., Tetler, L.W., White, J., Rimmer, D., 2002. Determination of alkyl benzyl

N.V. Ul'yanovskii et al. / Chemosphere 239 (2020) 124801 and dialkyl dimethyl quaternary ammonium biocides in occupational hygiene and environmental media by liquid chromatography with electrospray ionisation mass spectrometry and tandem mass spectrometry. J. Chromatogr., A 952, 165e172. https://doi.org/10.1016/S0021-9673(02)00082-1. Gonsior, M., Schmitt-Kopplin, P., Stavklint, H., Richardson, S.D., Hertkorn, N., Bastviken, D., 2014. Changes in dissolved organic matter during the treatment processes of a drinking water plant in Sweden and formation of previously unknown disinfection byproducts. Environ. Sci. Technol. 48, 12714e12722. Grbovi c, G., Trebse, P., Dolenc, D., Lebedev, A.T., Sarakha, M., 2013. LC/MS study of the UV filter hexyl 2-[4-(diethylamino)-2-hydroxybenzoyl]-benzoate (DHHB) aquatic chlorination with sodium hypochlorite. J. Mass Spectrom. 48, 1232e1240. https://doi.org/10.1002/jms.3286. Huang, G., Jiang, P., Li, X.-F., 2017a. Mass spectrometry identification of N-Chlorinated dipeptides in drinking water. Anal. Chem. 89, 4204e4209. https://doi. org/10.1021/acs.analchem.7b00228. Huang, N., Wang, T., Wang, W.L., Wu, Q.Y., Li, A., Hu, H.Y., 2017b. UV/chlorine as an advanced oxidation process for the degradation of benzalkonium chloride: synergistic effect, transformation products and toxicity evaluation. Water Res. 114, 246e253. https://doi.org/10.1016/j.watres.2017.02.015. Kelsall, H.L., Sim, M.R., 2001. Skin irritation in users of brominated pools. Int. J. Environ. Health Res. 11, 29e40. https://doi.org/10.1080/09603120020019629. Kim, H., Shim, J., Lee, S., 2002. Formation of disinfection by-products in chlorinated swimming pool water. Chemosphere 46, 123e130. https://doi.org/10.1016/ S0045-6535(00)00581-6. Kosyakov, D.S., Ul'yanovskii, N.V., Popov, M.S., Latkin, T.B., Lebedev, A.T., 2017a. Halogenated fatty amides e a brand new class of disinfection by-products. Water Res. 127, 183e190. https://doi.org/10.1016/j.watres.2017.10.008. Kosyakov, D.S., Ul'yanovskii, N.V., Popov, M.S., Latkin, T.B., Lebedev, A.T., 2017b. Characterization of disinfection by-products in Arkhangelsk tap water by liquid chromatography/high-resolution mass spectrometry. J. Anal. Chem. 73, 1260e1268 (Original Russian version in Mass-spektrometria, 2017, 14, 233-241. https://doi.org/10.1134/S1061934818130099. Lebedev, A.T., 2007. Mass spectrometry in the study of mechanisms of aquatic chlorination of organic substrates. Eur. J. Mass Spectrom. 13, 51e56. https://doi. org/10.1255/ejms.852. Lebedev, A.T., Shaidullina, G.M., Sinikova, N.A., Kharchevnikova, N.V., 2004. GC-MS comparison of the behavior of chlorine and sodium hypochlorite towards organic compounds dissolved in water. Water Res. 38, 3713e3718. https://doi. org/10.1016/j.watres.2004.06.007. Li, C., Wang, D., Xu, X., Wang, Z., 2017. Formation of known and unknown disinfection by-products from natural organic matter fractions during chlorination, chloramination, and ozonation. Sci. Total Environ. 587e588, 177e184. https:// doi.org/10.1016/j.scitotenv.2017.02.108. Liu, J., Zhang, X., 2013. Comparative toxicity of new halophenolic DBPs in chlorinated saline wastewater effluents against a marine alga: halophenolic DBPs are generally more toxic than haloaliphatic ones. Water Res. 65, 64e72. https://doi. org/10.1016/j.watres.2014.07.024. Manasfi, T., Coulomb, B., Boudenne, J.L., 2017. Occurrence, origin, and toxicity of disinfection byproducts in chlorinated swimming pools: an overview. Int. J. Hyg Environ. Health 220, 591e603. https://doi.org/10.1016/j.ijheh.2017.01.005. o, M., Coulomb, B., Di Giorgio, C., Boudenne, J.L., 2016. IdentifiManasfi, T., De Me cation of disinfection by-products in freshwater and seawater swimming pools and evaluation of genotoxicity. Environ. Int. 88, 94e102. https://doi.org/10.1016/ j.envint.2015.12.028. Phungsai, P., Kurisu, F., Kasuga, I., Furumai, H., 2016. Molecular characterization of low molecular weight dissolved organic matter in water reclamation processes using Orbitrap mass spectrometry. Water Res. 100, 526e536. https://doi.org/10. 1016/j.watres.2016.05.047. Plewa, M.J., Wagner, E.D., Muellner, M.G., Hsu, K.M., Richardson, S.D., 2008. Comparative mammalian cell toxicity of N-DBPs and C-DBPs. In: Disinfection By-Products in Drinking Water. American Chemical Society, Washington, DC, pp. 36e50. Richardson, S.D., 2003. Disinfection by-products and other emerging contaminants

9

in drinking water. Trends Anal. Chem. 22, 666e684. https://doi.org/10.1016/ S0165-9936(03)01003-3. Richardson, S.D., 2012. Mass spectrometry identification and quantification of toxicologically important drinking water disinfection by-products. In: Lebedev, A.T. (Ed.), Comprehensive Environmental Mass Spectrometry. ILM Publ, London, pp. 263e285, p. 510. Richardson, S.D., DeMarini, D.M., Kogevinas, M., Fernandez, P., Marco, E., , C., Heederik, D., Meliefste, K., McKague, A.B., Marcos, R., Lourencetti, C., Balleste Font-Ribera, L., Grimalt, J.O., Villanueva, C.M., 2010. What's in the pool? A comprehensive identification of disinfection by-products and assessment of mutagenicity of chlorinated and brominated swimming pool water. Environ. Health Perspect. 118, 1523e1530. https://doi.org/10.1289/ehp.1001965. Richardson, S.D., Postigo, C., 2016. Discovery of new emerging DBPs by highresolution mass spectrometry. Compr. Anal. Chem. 71, 335e356. https://doi. org/10.1016/bs.coac.2016.01.008. Richardson, S.D., Postigo, C., 2018. Liquid chromatographyemass spectrometry of emerging disinfection by-products. Compr. Anal. Chem. 79, 267e295. https:// doi.org/10.1016/bs.coac.2017.07.002. Richardson, S.D., Simmons, J.E., Rice, G., 2002. Disinfection byproducts: the next generation. Environ. Sci. Technol. 36, 198Ae205A. https://doi.org/10.1021/ es022308r. Tang, H.L., Ristau II, R.J., Xie, Y.F., 2015. Disinfection by-products in swimming pool water: formation, modeling, and control. In: Recent Advances in Disinfection By-Products. American Chemical Society, Washington, DC, pp. 381e403. Teo, T.L.L., Coleman, H.M., Khan, S.J., 2015. Chemical contaminants in swimming pools: occurrence, implications and control. Environ. Int. 76, 16e31. https://doi. org/10.1016/j.envint.2014.11.012. Trebse, P., Polyakova, O.V., Baranova, M., Kralj, M.B., Dolenc, D., Sarakha, M., Kutin, A., Lebedev, A.T., 2016. Transformation of avobenzone in conditions of aquatic chlorination and UV-irradiation. Water Res. 101, 95e102. https://doi. org/10.1016/j.watres.2016.05.067. Tretyakova, N.Yu, Lebedev, A.T., Petrosyan, V.S., 1994. Degradative pathways for aqueous chlorination of orcinol. Environ. Sci. Technol. 28, 606e613. https://doi. org/10.1021/es00053a012. U.S. EPA. 8270, 2007. Semivolatile Organic Compounds by Gas Chromatography/ mass Spectrometry (GC/MS). US Environmental Protection Agency. Weisel, C.P., Richardson, S.D., Nemery, B., Aggazzotti, G., Baraldi, E., Blatchley, E.R., Blount, B.C., Carlsen, K.-H., Eggleston, P.A., Frimmel, F.H., Goodman, M., Gordon, G., Grinshpun, S.A., Heederik, D., Kogevinas, M., LaKind, J.S., Nieuwenhuijsen, M.J., Piper, F.C., Sattar, S.A., 2009. Childhood asthma and environmental exposures at swimming pools: state of the science and research recommendations. Environ. Health Perspect. 117, 500e507. https://doi.org/10. 1289/ehp.11513. White, G.C., 1999. Handbook of Chlorination and Alternative Disinfectants. John Wiley & Sons, Inc., p. 1569. New York, Chichester, Weinheim, Brisbane, Singapore, and Toronto. Yang, L., Schmalz, C., Zhou, J., Zwiener, C., Chang, V.W., Ge, L., Wan, M.P., 2016. An insight of disinfection by-product (DBP) formation by alternative disinfectants for swimming pool disinfection under tropical conditions. Water Res. 101, 535e546. https://doi.org/10.1016/j.watres.2016.05.088. Yang, M., Zhang, X., 2013. Comparative developmental toxicity of new aromatic halogenated DBPs in a chlorinated saline sewage effluent to the marine polychaete Platynereis dumerilii. Environ. Sci. Technol. 47, 10868e10876. https://doi. org/10.1021/es401841t. , M.J., Stalter, D., Tang, J.Y.M., Molendijk, J., Escher, B.I., 2014. BioYeh, R.Y.L., Farre analytical and chemical evaluation of disinfection by-products in swimming pool water. Water Res. 59, 172e184. https://doi.org/10.1016/j.watres.2014.04. 002. Yin, J., Wu, B., Liu, S., Hu, S., Gong, T., Cherr, G.N., Zhang, X.X., Ren, H., Xian, Q., 2018. Rapid and complete dehalogenation of halonitromethanes in simulated gastrointestinal tract and its influence on toxicity. Chemosphere 211, 1147e1155. https://doi.org/10.1016/j.chemosphere.2018.08.039.