Precursors for brominated haloacetic acids during chlorination and a new useful indicator for bromine substitution factor

Science of the Total Environment 698 (2020) 134250

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Precursors for brominated haloacetic acids during chlorination and a new useful indicator for bromine substitution factor Lili Zheng 1, Hongjie Sun 1, Chouye Wu, Yibo Wang, Yuanyuan Zhang, Guangcai Ma, Hongjun Lin, Jianrong Chen, Huachang Hong ⁎ College of Geography and Environmental Sciences, Zhejiang Normal University, Jinhua 321004, PR China

H I G H L I G H T S

G R A P H I C

A B S T R A C T

• Water samples from city contained higher OM levels than those from country side. • SUVA of water was negatively related with Chl-a level. • Hydrophobicity of precursor was ranked as Br-HAAsbBrCl-HAAsbClHAAs. • Br/UVA is a good indicator for BSF of diHAAs and tri-HAAs. • Br/UVA is also a good index for BSF from other water samples and other DBPs.

a r t i c l e

i n f o

Article history: Received 2 August 2019 Received in revised form 1 September 2019 Accepted 1 September 2019 Available online 06 September 2019 Editor: Damia Barcelo Keywords: Brominated haloacetic acids (HAAs) Precursor Bromine substitution factor (BSF) Chlorination1

⁎ Corresponding author. E-mail address: [email protected] (H. Hong). 1 Means equal contribution to the paper.

https://doi.org/10.1016/j.scitotenv.2019.134250 0048-9697/© 2019 Elsevier B.V. All rights reserved.

a b s t r a c t Brominated haloacetic acids (HAAs) are much more cytotoxic and genotoxic than chlorinated one, yet little information is available for their organic precursors. In the present study, 8 water samples were collected in East China: 2 from lakes, 2 from rivers, 2 from reservoirs, a well and a mountain spring. Dissolved organic carbon (DOC), UV absorbance at 254 (UVA), specific UVA (SUVA) and chlorophyll a (Chl-a) were determined in raw water samples; formation of 9 HAA species as well as bromine substitution factor (BSF) were measured in chlorinated water samples. Results showed that water samples located in city generally contained higher levels of DOC (6.4–12.2 mg/L) and UVA (0.124–0.194/cm), while those in the country side, low DOC (2.4–5.9 mg/L) and UVA (0.061–0.109/cm) levels were observed. Negative relationship (p b 0.01) was found between SUVA values and Chl-a levels. Among 9 HAA species, 4 brominated HAA were detected. As for BAA and DBAA (i.e. Br-HAAs), their yields (μg/L) were significant related (p b 0.05) with DOC; In terms of BCAA and BDCAA (i.e. ClBr-HAAs), they were not only related with DOC, but also with UVA. These two results were quite different from DCAA and TCAA (Cl-HAAs), whose yields (μg/mg C) were only correlated with SUVA values, suggesting that precursors of Cl-HAA, Br-HAA and ClBr-HAA were different from each other, and their aromaticity/hydrophobicity may be in the order of Br-HAA b ClBr-HAA b Cl-HAA. Interrelationship between Br/DOC, SUVA and BSF revealed that BSF can be influenced by SUVA and Br/DOC, but in comparison, Br/UVA was the best indicator to describe BSF. This

2

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250

pattern is not only true in di-HAAs and tri-HAAs in this study, but also valid in other water samples and other species of disinfection by-product (e.g. trihalomethanes, dihaloacetonitriles, trihalonitromethanes). © 2019 Elsevier B.V. All rights reserved.

1. Introduction The chlorine is most widely used disinfectant in drinking water due to its high efficiency in killing microbes, low cost and easy operation. However, besides killing the hazardous microbial contaminants, it can result in formation of disinfection byproducts (DBPs) due to the reaction of chlorine with organic matter (OM) in source water (Sun et al., 2018; Ersan et al., 2019). To date, N600 species of DBPs have been detected. Haloacetic acid (HAA) is one of the DBP classes that detected in most frequency and existed in most abundant concentration (Krasner et al., 2006; Richardson et al., 2007; Park et al., 2019). Overall, nine HAA species commonly occur: three Cl-HAAs (i.e. chloro-(CAA), dichloro-(DCAA), trichloro-(TCAA)), three Br-HAAs (i.e. bromo-(BAA), dibromo-(DBAA), tribromo-(TBAA)), three ClBr-HAAs (i.e. bromochloro-(BCAA), bromodichloro-(BDCAA) and dibromochloroacetic acid (DBCAA)). Many studies have shown that high level of HAAs in drinking water may cause an increased risk of cancer and birth defects (Richardson et al., 2007; Pan et al., 2014; Kaufman et al., 2018). Now, HAAs have been regulated by many countries and districts. For example, U.S. Environmental Protection Agency (USEPA) set a maximum contaminant level of 60 μg/L for 5 HAA species (sum of CAA, DCAA, TCAA, BAA and DBAA) (EPA, 1998). In China, two HAA species, i.e. DCAA and TCAA are regulated, and the maximum level should not exceed 50 and 100 μg/L, respectively (MOH, 2006). Formation of HAAs is greatly influenced by treatment conditions and water quality parameters (Sohn et al., 2004; Hua and Reckhow, 2012; Hong et al., 2017). During chlorination, formation of HAAs generally increases with the increase of chlorine dose and reaction time. High bromide level could not only increase the total amount of HAAs yields, but also shift HAA formation to more brominated ones. Though HAA yields were usually positively related with dissolved organic carbon (DOC) or UV absorbance at 254 nm (UVA), precursor of trihaloacetic acids (tri-HAAs) is considered to have higher aromaticity and be more hydrophobic in property as compared to dihaloacetic acids (di-HAAs) and monohaloacetic acids (mono-HAAs) (Liang and Singer, 2003; Hua et al., 2015; Sun et al., 2018). Overall, current studies provided detailed information on HAAs formation. However, available studies regarding organic precursors of HAAs are primarily on chlorinated HAAs (e.g. DCAA, TCAA), di-HAAs, tri-HAAs and total HAAs (T-HAAs), very few on brominated HAAs (e.g. Br-HAAs, ClBr-HAAs) (Liang and Singer, 2003; Hua et al., 2015; Sun et al., 2018; Ersan et al., 2019). Considering the much higher cytotoxicity and genotoxicity of brominated HAAs (i.e. Br-HAAs and BrCl-HAAs) than chlorinated analogs (Cl-HAAs), clarifying the organic precursors for brominated HAAs is important in view of health risk control (Plewa et al., 2010; Wagner and Plewa, 2017). Bromine substitution factor (BSF), which is defined as the ratio of the molar concentration of bromine incorporated into a given class of DBPs divided by the total molar concentration of chlorine and bromine in that class, was usually used to evaluate the formation of brominated DBPs (including HAAs) (Obolensky and Singer, 2005; Hua and Reckhow, 2012; Hong et al., 2017). Many studies have shown that for a given water sample, BSF of DBPs (including HAAs) increased with bromide level (Hua et al., 2006; Hua and Reckhow, 2012; Ersan et al., 2019). But for a series of water samples from different sources, they probably not only differ in bromide level, but also differ in OM level and the property. So the single parameter of bromide is not enough to evaluate BSF for different water samples. Limited studies showed that Br/DOC may be a good indicator to evaluate BSF, the high the Br/DOC values, the high the BSF (Chow et al., 2007). Yet some other studies also revealed that bromine incorporation into DBPs (including HAAs) was more

reactive with hydrophilic OM (low Specific UVA (SUVA)) as compared to the hydrophobic one (high SUVA) (Liang and Singer, 2003; Ersan et al., 2019). Therefore, bromide level, DOC and SUVA should be considered together in order to comprehensively evaluate BSF of DBPs (including HAAs). However, no study was available currently. Based on above information, 8 water samples were collected from East China, which differed greatly in bromide level, DOC and SUVA values. The differences of organic precursor for Cl-HAAs, Br-HAAs and ClBr-HAAs were investigated and the combined effects of Br, DOC, SUVA on BSF for di-HAAs and tri-HAAs were evaluated. Finally the pattern related to the interrelationship between Br, DOC, SUVA and BSF were further tested using other water samples and DBPs. The result in the present study will add new information on organic precursors for brominated DBPs, and provide a new and useful indicator for BSF of DBP formation. 2. Material and methods 2.1. Water samples 8 source water samples were collected from East China: 2 from lake, 2 from river, 2 from reservoirs, a well, a mountain spring. Sampling site and corresponding locations (city, the longitude and latitude) are presented in Table 1. Field sampling was carried out in July 2018. The glassware used for sampling was washed with detergent and tap water, presoaked in 10% sulfuric acid for 24 h, and rinsed with Milli-Q water. Before water sampling, all of the glassware was flushed with source water 3 times. Water samples were collected at approximately 0.1–0.5 m below the surface water using an open-mouth glass bottle (5 L). For each sampling site, 3 water samples were collected and mixed in an even bigger glass bottle (15 L). Then the composite water samples were shipped to the laboratory immediately with an ice cooler box and then prepared for further analysis. 2.2. Analysis of water quality The source water samples were filtered through GC/F filters (pore size = 0.45 μm). The filtrate was then used to determine dissolved organic carbon (DOC, using ELEMENTAR Liqui TOCII TOC analyzer), UVA (Shimadzu UV-1601 UV–visible spectrophotometer), SUVA (SUVA = UVA × 100/DOC), bromide (Dionex ICS-2100 ion chromatography) and NH+ 4 (using salicylic acid-hypochlorous acid colorimetric method). All the above parameters were determined according to standard methods (APHA, 1998). Water samples for chlorophyll a (Chl-a) measurements were stored in 1 L bottles, wrapped in black plastic bags and preserved by the addition of 3–5 mL of 1% saturated magnesium carbonate. Each water sample was filtered through a GC/F filter (pore size = 0.45 μm). The retentate on the filter was then extracted with 95% ethanol. After that, the ethanol samples were centrifuged at 4000 rpm and the supernatant was determined with an UV–visible spectrophotometer according to the following formula (Wintermans and De Mots, 1965). Chl a ¼ ½13:7ðA665−A750Þ−5:76ðA649−A750Þ  E=F  l where E = extraction volume in mL; F = filtration volume in L; l = cuvette path length in cm; and Ax = absorbance at wavelength of x nm.

* TL1: near the shore, located at N31o27’5.48″, E120o19’29.72″ in Wuxi City); TL2: water intake of Gonghu Water Works, located at N31o22’14.99″, E120o13’42.15″ in Wuxi City; QTR: located at N30o10’52.34″, E120o07’27.43”in Hangzhou city; WR located at N29 o06’42.67, E119o40’33.93″ in Jinhua City; R1 (N29o12’49.88″, E119o14’17.21″), R2 (N29o07’28.32″, E119o29’39.66″), W (N29o09’44.72″, E119o38’16.08″) and MS (N29o09’49.35″, E119o38’8.21″) located in the country side in Jinhua. “ND” means not detectable.

ND ND ND ND ND ND ND ND 9.4 ± 0.1 8.1 ± 1.5 9.9 ± 0.1 5.1 ± 0.1 ND ND 1.4 ± 0.1 1.9 ± 0.6 10.3 ± 0.1 13.7 ± 1.5 1.3 ± 0.0 0.8 ± 0.0 ND ND 0.2 ± 0.0 0.3 ± 0.1 287.0 ± 4.2 191.0 ± 15.3 477.2 ± 5.9 236.0 ± 9.4 23.0 ± 0.7 151.8 ± 22.2 192.5 ± 0.5 312.6 ± 32.7 ND ND ND ND ND ND ND ND 216.6 ± 18.3 35.9 ± 2.3 4.4 ± 1.8 11.3 ± 0.5 32.5 ± 4.1 14.6 ± 1.0 0.8 ± 0.1 0.6 ± 0.1 0.151 ± 0.003 0.158 ± 0.015 0.194 ± 0.001 0.124 ± 0.005 0.061 ± 0.001 0.061 ± 0.000 0.109 ± 0.005 0.084 ± 0.009 12.2 ± 0.7 11.2 ± 1.1 9.5 ± 0.5 6.4 ± 0.6 5.9 ± 0.1 5.3 ± 0.0 4.1 ± 0.5 2.4 ± 0.0 Tai Lake 1*(TL1) Tai Lake 2*(TL2) Wu River (WR) Qiantang River (QTR) Reservoir 1 (R1) Reservoir 2 (R2) Well (W) Mountain Spring (MS)

UV254

/cm

DOC

(mg/L)

1.24 1.41 2.04 1.93 1.04 1.14 2.64 3.46

119.0 ± 4.6 141.7 ± 6.1 54.1 ± 1.7 26.2 ± 0.3 9.0 ± 1.1 6.3 ± 1.3 12.9 ± 0.9 11.9 ± 1.0

3.40 ± 0.01 3.46 ± 0.43 0.95 ± 0.02 2.11 ± 0.03 0.11 ± 0.00 0.42 ± 0.2 0.44 ± 0.05 0.48 ± 0.05

57.5 ± 0.1 48.8 ± 6.6 34.9 ± 0.2 19.0 ± 0.4 0.8 ± 0.0 4.1 ± 0.3 10.8 ± 0.8 12.7 ± 1.5

101.2 ± 1.6 76.5 ± 8.7 299.1 ± 0.6 168.1 ± 4.1 24.8 ± 0.0 130.4 ± 19 171.4 ± 0.6 261.1 ± 13

ND ND ND ND ND ND ND ND

TBAA DBCAA

(μg/L)

BDCAA

(μg/L)

TCAA

(μg/L)

DBAA

(μg/L)

BCAA DCAA

(μg/L) (μg/L) (μg/L) L/(mg/m)

BAA CAA

(μg/L)

Chl-a Br SUVA254

(μg/L)

(μg/L)

tri-HAAs di-HAAs mono-HAAs

−

Water quality parameters Sampling sites (short name)

Table 1 Water quality of source water and HAAs formation during chlorination.

(μg/L)

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250

3

2.3. Chlorination test The stock chlorine solution (NaClO) was obtained from J&k (reagent grade, 6%–10%), the chlorine was calibrated by the N,N diethylphenylenediamine (DPD) titrimetric method before disinfection (APHA, 1998). Water samples filtered through 0.45 μm were subjected to chlorination. All disinfection experiments were performed in a series of 100 mL glass tubes, and each water sample had two replicates. Prior to chlorination, phosphate buffer (final concentration = 5 mM) was added and adjusted pH = 7.5. Chlorine dose was performed according to Eq. (1) (Chu et al., 2013). Then, the water sample was cultured in the dark at 20 °C without headspace for 24 h. Cl2 dose ðmg=LÞ ¼ 3  DOCðmg=LÞ þ 7:6  NH4 þ −Nðmg N=LÞ þ 10ðmg=LÞ

ð1Þ

2.4. HAAs analysis HAA measurements were carried out by MTBE liquid-liquid extraction- acidification methanol derivatization-GC/ECD analysis (USEPA, 2003). The recovery rate of CAA, BAA, DCAA, TCAA, BCAA, DBAA, BDCAA, DBCAA and TBAA was 62%, 72%, 93%, 91%, 88%, 94%, 79%, 84% and 107%, respectively; the detection limit of CAA, BAA, DCAA, TCAA, BCAA, DBAA, BDCAA, DBCAA and TBAA was 0.192, 0.054, 0.128, 0.066, 0.046, 0.049, 0.077, 0.098 and 0.013 μg/L, respectively. Both of them have been reported in our previous study (Song et al., 2017). 2.5. Calculation of BSF Bromine substitution factor (BSF) was calculated according to Eq. (2): the ratio of the molar concentration of bromine incorporated into a given class of DBPs divided by the total molar concentration of chlorine and bromine in that class (Obolensky and Singer, 2005). For example, BSF of di-HAAs was calculated as shown in Eq. (3). BSF ¼

DBP−Br DBP−ðCl þ BrÞ

BSF−di−HAAs ¼

ð2Þ

BCAA þ 2  DBAA 2  ðDCAA þ BCAA þ BCAAÞ

ð3Þ

2.6. Statistical analysis DOC, UVA, SUVA, Chl-a and HAA yields were expressed as an average of 3 or 2 measurements (see Table 1). Linear regression models were used to study the relationships between Br/DOC, Br/SUVA, Br/UVA and BSF (Figs. 2–3), which were performed using SigmaPlot software (Version 9.0); Pearson correlations were used to examine the relationships between water quality parameters and HAA formation (Table 2), which were obtained by SPSS software (Version 16.0, SPSS, Inc.). Table 2 Relationship between water quality and HAAs formation Parameters

Br- 1 DOC1 UVA2541 SUVA2 Chl-a1

Cl-HAAs

Br-HAAs

DCAA

TCAA

BAA

0.230 0.26 0.737⁎ 0.917⁎⁎ 0.045

−0.257 −0.269 0.361 0.920⁎⁎ −0.383

0.917⁎⁎ 0.821⁎ 0.618 0.081 0.638

ClBr-HAAs DBAA 0.977⁎⁎ 0.818⁎ 0.537 −0.309 0.607

“⁎” and “⁎⁎” means significant level at 0.05 and 0.01, respectively. 1 HAAs in μg/L was used to perform regression analysis. 2 HAAs in μg/mg C was used to perform regression analysis.

BCAA

BDCAA

0.945⁎⁎ 0.893⁎⁎ 0.823⁎ 0.529 0.671

0.807⁎ 0.854⁎⁎ 0.953⁎⁎ 0.398 0.474

4

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250

3. Results and discussion 3.1. Water quality parameters Water quality parameters of the sampling sites are presented in Table 1. OM level in lakes and rivers (DOC: 6.4–12.2 mg/L; UVA: 0.124–0.194 L/mg/m) were dominantly higher than those in reservoirs, the well and the mountain spring (DOC: 2.4–5.9 mg/L; UVA: 0.061–0.109 L/mg/m). The different OM levels in these sampling sites may be related to the ambient environment and human activity. Both the lakes and rivers were located in the cities and have a dense population or developed secondary & third industry; while others were located in the country side with less human activity. SUVA values, an indicator for aromaticity/hydrophobicity of organic carbon, showed a trend of well water and mountain spring (2.64–3.46 L/mg/m) N river water (1.93–2.04 L/mg/m) N lakes and reservoirs (1.04–1.41 L/mg/m). Meanwhile, significant (p b 0.01) linear relationship was also observed between SUVA values and log chl-a levels (Fig. 1). These results suggested that autochthonous algae may be the main contributor to the organic matter in low SUVA waters (lakes and reservoirs); while for the high SUVA waters (well and mountain spring), the allochthonous organic matter (usually containing higher aromatic carbon) may be the major contributor. For the water with middle level of SUVA (river water), both the autochthonous and allochthonous organic matter may play important role. Bromide level, an inorganic precursor for DBPs (including HAAs) formation, also showed regional difference. Tai Lake water ranked highest (119.0–141.7 μg/L), followed by Wu River water (54.1 μg/L), while Qiantang River, reservoirs, well and the mountain spring had the lowest levels (6.3–26.2 μg/L). For Tai Lake, the high bromide level may be derived from Yangtze River water, which was drawn into Tai Lake to alleviate the low water level in winter and the algal bloom in summer (Hong et al., 2016). But the bromide level in Tai Lake in this study (2018) was lower as compared to that occurred in 2011 (248 μg/L) (Hong et al., 2015). Wu River is located in inner area (Jinhua), and is one of tributaries of Qiantang River, the relative high level of bromide may be derived from the waste water pollution. Qiantang River is located nearer the estuary as compared to Wu River, its low bromide level probably because the effective measurement taken by local government to protect the drinking water source (sampling site was in the water conservation area).

3.2. HAAs formation potential and its association with water quality parameters

SUVA (L/mg/m)

HAAs formation from chlorination of different source water is also presented in Table 1. DCAA and TCAA were dominant HAA species, accounting for 47.2–61.2% and 21.6–50.9% of the total yields, respectively. Among brominated HAAs, BCAA yields ranked highest for all water samples (Table 1, S-Fig. 1). Take lake water as an example, percentage yields for BCAA, BAA, DBAA and BDCAA were 12.3–14.3%, 0.7–1.0%, 2.2–4.0%

and 2.0–2.4%, respectively. CAA, DBCAA and TBAA were not detectable for all water samples. If the HAAs formation in this study was classified into mono-HAAs, di-HAAs and tri-HAAs, mono-HAAs was the minimum levels detected (Table 1). This is in agreement with other studies (Hong et al., 2013b; Sun et al., 2018). The underlying mechanisms may be that during the formation of mono-HAAs, the activated carbon atom in the β-diketone moiety (R'-CO − CH2 − CO − R, co-precursor for HAAs) is only substituted with one chlorine/bromine (R'-CO − CHX − CO − R, precursor for mono-HAA), but it is much more difficult for this structure to hydrolyze to CH2X − CO − R than it is for R'-CO − CX2 − CO − R (precursor for di-HAA/tri-HAA) to hydrolyze to CHX2 − CO − R, thus giving a very low yield of mono-HAA (when R = OH) (Liang and Singer, 2003). Relationships between HAAs formation and water quality parameters are shown in Table 2. Formation of BAA, DBAA, BCAA and BDCAA was positively and significantly (p b 0.01 or p b 0.05) related with bromide. This result is predictable, because bromide ion can be oxidized by free chlorine to produce hypobromous acid (HOBr, Eq. (9)), which can also react with OM in water to form DBPs (including HAAs). With increasing bromide levels, more bromide could be incorporated into HAAs, so the yields of brominated HAAs increased accordingly (Hong et al., 2017; Ersan et al., 2019). Moreover, the yields (μg/L) of BAA and DBAA (i.e. Br-HAAs) were significantly related (p b 0.05) with DOC, but not significantly (p N 0.05) related with UVA (yields in μg/L) or SUVA (yields in μg/mg C), indicating that aromatic carbon contributed little to their formation and DOC was a better index for their precursors. In terms of BCAA and BDCAA (i.e. ClBr-HAAs), they were not only significantly (p b 0.05) related with DOC, but also with UVA (Table 2), suggesting that both aromatic and non-aromatic carbon played important roles during their formation. Regarding DCAA and TCAA (Cl-HAAs), their yields are only significantly (p b 0.01) related with UVA (yields in μg/L) and/or SUVA (yields in μg/mg C) (Table 2), indicating that OM with UV absorbance (e.g. activated aromatic structures and other conjugated double bond) was main precursor for their formation during chlorination. All of these results suggested that precursors of Cl-HAA, Br-HAA and ClBrHAA may be different from each other. The aromaticity/hydrophobicity may be ranked as the following: precursor for Cl-HAA N precursor for ClBr-HAA N precursor for Br-HAA. The underlying mechanism may be that bromine is more reactive with aliphatic precursors than with aromatic precursors (in Eqs. (7)–(8), k3 N k4), but the opposite is true for chlorine (in Eqs. (4)–(5), k2 N k1) (Heller-Grossman et al., 1993; Liang and Singer, 2003). Therefore aromatic precursors (indicator by SUVA or UVA) contributed more to chlorinated HAAs; while aliphatic precursors contributed more to brominated HAAs.

ð4Þ

ð5Þ 3.0

y = -0.8653x + 2.7216 R² = 0.7819

2.0 1.0

ð6Þ 0.0

-1.0

0.0

1.0

2.0

3.0

lg Chl-a (µg/L) Fig. 1. Relationship between Chl-a and SUVA values.

ð7Þ

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250

BSF ¼ k 

Br=DOC þb SUVA

BSF ¼ k 

Br=DOC þb UVA=DOC

ð10Þ

BSF ¼ k 

Br þb UVA

ð11Þ

ð9Þ

Based on the above information, the relationships between BSF and Br/DOC, BSF and SUVA, BSF and Br/UVA were performed in this study (Fig. 2). Results showed that single SUVA is not related (p N 0.05) with BSF both for di-HAAs and tri-HAAs (Fig. 2a–b). This may be because the relationship between BSF and SUVA can be significant only when other conditions (e.g. chlorination conditions, Br/DOC level, etc) were the same or similar. For example in Fig. 2a, significant (p b 0.01) negative relationships were found between SUVA and BSF (di-HAAs) for 4 water samples in dotted circle, whose Br/DOC were in the middle Table 3 BSF for mono-HAAs, di-HAAs, tri-HAAs and some related parameters. Water samples

mono-HAAs (%)

di-HAA (%)

tri-HAA (%)

SUVA (L/mg/m)

Br/DOC (μg/mg)

Br/UVA (μg/L/cm)

TL1 TL2 WR QTR R1 R2 W MS

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

8.2 11.1 2.7 3.0 1.3 1.0 2.0 1.5

2.3 2.6 0.8 0.8 0.0 0.0 0.2 0.2

1.2 1.4 2.0 1.9 1.0 1.1 2.6 3.5

9.8 12.7 5.7 4.1 1.5 1.2 3.1 4.9

789.8 899.0 279.0 212.1 146.7 103.4 118.8 142.3

BSF-tri-HAAs (%)

4

8 4

y=1.5717-0.3838x (R2=0.1017, p=0.4413)

3 2 1 0

0 c) 12

1 2 3 SUVA (L/mg/m)

0

4

0

3

6 9 12 Br/DOC ( g/mg)

y=-0.0041+0.0115x (R2=0.9724, p<0.0001)

10 8 6 4 2

4

y=-0.4603+0.2451x (R2=0.9257, p=0.0001

3 2 1 0

15

e) 12

1 2 3 SUVA (L/mg/m)

4

d)

y=-0.9382+0.8935x

10 (R2=0.9157, p=0.0002) 8 6 4 2 0

BSF-tri-HAAs (%)

BSF of HAAs for all source waters share the same trend: BSF for mono-HAAs (100%) N BSF for di-HAAs (1.0%–11.1%) N BSF for tri-HAAs (0–2.6%) (Table 3). This may be because the precursor for mono-HAAs is extremely hydrophilic in nature as compared to di-HAAs and triHAAs (Sun et al., 2018). For a given water sample, BSF of DBPs usually increase with bromide level. However, for many water samples such as in the present study, they not only differ in bromide level, but also differ in OM level and their property, which may also influence bromine incorporation. Therefore, in order to comprehensively evaluate the relationship between BSF and water quality, factors such as Br, DOC and SUVA should be considered together. Chow et al. (2007)’s study showed that BSF of DBPs increased with the increase of Br/DOC ratio; Two other studies reported that bromine incorporation into DBPs (including HAAs) was more reactive in the OM with low SUVA than that with high SUVA (Liang and Singer, 2003; Ersan et al., 2019). From these two phenomena, it can be inferred that under certain conditions, BSF can be expressed as Eq. (9). Because SUVA equal to UVA divided by DOC, Eq. (9) can be transformed to Eq. (10). After eliminating DOC, Eq. (11) was finally obtained. Therefore, Br/UVA may be a comprehensive indicator to better evaluate BSF during DBP formation.

12

0

3

6 9 12 Br/DOC ( g/mg)

15

f) 3.0

BSF-tri-HAAs (%)

3.3. Bromine substitution factor and its association with water quality

b)

y=6.5673-1.4518x (R2=0.1083, p=0.4260)

0

BSF-di-HAAs (%)

Table 2 also shows that Chl-a was not related to any HAAs species. This result was not surprising. Algae really could contribute to HAA precursors in some waters (e.g. lakes and reservoirs, Table 1), but for other waters, HAA precursors were mainly derived from allochthonous OM (Section 3.1).

BSF-di-HAAs (%)

ð8Þ

BSF-di-HAAs (%)

a) 16

5

y=-0.19691+0.0031x (R2=09698, p<0.0001)

2.5 2.0 1.5 1.0 0.5 0.0

0 0

200 400 600 800 1000 Br/UVA ( g/L/cm)

0

200 400 600 800 1000 Br/UVA ( g/L/cm)

Fig. 2. Relationships between BSF and SUVA (a-b), BSF and Br/DOC (c-d), BSF and Br/UVA (e-f).

level (3.1–5.7 μg/mg) as compared to the whole data set (Br/DOC: 1.2–12.7 μg/mg, Table 3); Similarly in Fig. 2b, BSF for tri-HAAs was also significant (p b 0.05) related with SUVA if remove two samples (outside of the dotted circle) which have extreme low Br/DOC (1.2–1.5 μg/mg, Table 3). Different from SUVA, Br/DOC was significantly (p b 0.01) related with BSF (di-HAA: R2 = 0.9157; tri-HAA: R2 = 0.9257; Fig. 2c–d) without removing any water samples, suggesting Br/DOC is a better indicator for BSF; However by comparison, Br/UVA is an even better index to evaluate BSF (Fig. 2d–e), whose the R2 values reached 0.9724 (diHAA) and 0.9698 (tri-HAA). All these results suggested that Br/UVA may be the best indicator to describe BSF as compared to SUVA and Br/DOC. In order to test whether Br/UVA is also a good indicator for BSF in other chlorinated waters and DBPs, further study was carried out by using the data in our previous studies (Hong et al., 2013a, 2013b; Hong et al., 2015; Hong et al., 2016; Song et al., 2017; Lin et al., 2018), which mainly focused on the factors influencing DBPs formation or regression models. The data used here included 4 water samples (DOC: 1.13–10.34 mg/L; UVA: 0.017–0.054/cm; SUVA: 0.522–2.655 L/mg/m; background (Bg) Br− level = 9-248 μg/L). Each water sample was spiked with bromide and the final Br− levels reached Bg, Bg + 100, Bg + 200, Bg + 400 μg/L or Bg, Bg + 100, Bg + 300 μg/L. Chlorinated DBPs including trihalomethanes (THMs), di-HAAs, tri-HAAs, dihaloacetonitriles (di-HANs), trihalonitromethanes (tri-HNMs) were measured and the corresponding BSF values were calculated here. Then the relationships between BSF & SUVA (Fig. 3a1-a5), BSF & Br/ DOC (Fig. 3b1-b5), BSF & Br/UVA (Fig. 3c1-c5) were conducted. Obviously, Br/UVA was the best indicator for BSF as compared to SUVA and Br/DOC for all DBP classes. With the increase of Br/UVA, BSF values increased linearly at first. This is consistent with the results of Fig. 2. However, the increasing rate became slow after BSF values reached 60%. The overall curve fit the equation of “exponential rise to maximum”, i.e. y = a(1-exp(−bx)), which were presented on the bottom of the Fig. 3.

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250

60 40 20

0.0 0.5 1.0 1.5 2.0 2.5 3.0 SUVA (L/mg/m)

b3: tri-HAAs

20

20

0

0

70 140 210 280 350 Br/DOC ( g/mg)

c1: THMs

60 40

60 40

40

60 40

60 40 20

0

0

0

0

0

10000 20000 30000 Br/UVA ( g/L/cm)

y=78.3112(1-e-0.000078x) (R2=0.9526, p<0.0001)

0

10000 20000 30000

0

0

10000 20000 30000

Br/UVA ( g/L/cm)

Br/UVA ( g/L/cm)

y=71.4146(1-e-0.000088x) (R2=0.9517, p<0.0001)

y=74.9454(1-e-0.0001x) (R2=0.9438, p<0.0001)

70 140 210 280 350 Br/DOC ( g/mg)

120

80

20

40

70 140 210 280 350 Br/DOC ( g/mg) c4: di-HANs

100

20

60

0 0

20

80

20

120 c3: tri-HAAs

80

BSF (%)

BSF (%)

80

60

70 140 210 280 350 Br/DOC ( g/mg)

100 c2: di-HAAs

80

b5: tri-HNMs

100

0 0

100

100

b4: di-HANs

20

0 0

70 140 210 280 350 Br/DOC ( g/mg)

40

BSF (%)

20

SUVA (L/mg/m)

BSF (%)

40

60

40

120

100

BSF (%)

60

60

0.0 0.5 1.0 1.5 2.0 2.5 3.0

120

80

80

0

SUVA (L/mg/m)

100

a5: tri-HNMs

20

0.0 0.5 1.0 1.5 2.0 2.5 3.0

BSF (%)

BSF (%)

BSF (%)

40

20 0

b2: di-HAAs

60

40

SUVA (L/mg/m)

80

80

20

60

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0

b1: THMs 80

0

40

0

100

100

100

c5: tri-HNMs

100

BSF (%)

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 SUVA (L/mg/m)

BSF (%)

40

BSF (%)

BSF (%)

BSF (%)

60

120

a4: di-HANs

80

60

80

20

BSF (%)

a3: tri-HAAs

a2: di-HAAs

a1: THMs

80

100

80

100

100

BSF (%)

6

80 60 40 20 0

0

10000 20000 30000 Br/UVA ( g/L/cm)

y=84.9403(1-e-0.0002x) (R2=0.9688, p<0.0001)

0

10000 20000 30000 Br/UVA ( g/L/cm)

y=91.922(1-e-0.0003x) (R2=0.9517, p<0.0001)

Fig. 3. Relationship between BSF and SUVA (a1-a5), BSF and Br/DOC (b1-b5), BSF and Br/UVA (c1-c5) using the data in previous studies.

These results further demonstrated that Br/UVA is a good comprehensive indicator to describe BSF for chlorinated DBPs. 4. Conclusions Water samples derived from cities had higher DOC and UVA levels than those from the country side. SUVA of water was negatively related with Chl-a level. Among 9 HAA species, 4 brominated HAAs were detected. Precursors for Br-HAA (BAA, DBAA), ClBr-HAA (BCAA and BDCAA) and Cl-HAA (DCAA, TCAA) were different from each other, and the aromaticity/hydrophobicity followed the trend of Br-HAA b ClBr-HAA b Cl-HAA. Br/UVA was the best indicator to describe BSF as compared to Br/DOC and SUVA. This pattern is not only true for diHAAs and tri-HAAs in this study, but also valid in other water samples and other DBP species (e.g. THMs, di-HANs, tri-HNMs). Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2019.134250. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgement This study was financially supported by Basic Public Welfare Research Project in Zhejiang Province (LGF18H260005).

References APHA, 1998. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC, USA. Chow, A.T., Dahlgren, R.A., Harrison, J.A., 2007. Watershed sources of disinfection byproduct precursors in the Sacramento and San Joaquin Rivers, California. Environ. Sci. Technol. 41 (22), 7645–7652. Chu, W., Gao, N., Yin, D., Krasner, S.W., 2013. Formation and speciation of nine haloacetamides, an emerging class of nitrogenous DBPs, during chlorination or chloramination. J. Hazard. Mater. 260, 806–812. EPA., U.S., 1998. National Primary Drinking Water Regulations: disinfectants and disinfection byproducts, final rule. Fed. Regist. 63 (241), 69389–69476. Ersan, M.S., Liu, C., Amy, G., Karanfil, T., 2019. The interplay between natural organic matter and bromide on bromine substitution. Sci. Total Environ. 646, 1172–1181. Heller-Grossman, L., Manka, J., Limoni-Relis, B., Rebhun, M., 1993. Formation and distribution of haloacetic acids, THM and TOX in chlorination of bromide-rich lake water. Water Res. 27 (8), 1323–1331. Hong, H., Huang, F., Lin, H., Yu, H., Wang, F., Ding, L., 2013a. Factors affecting formation of haloacetonitriles and haloketones during chlorination/monochloramination of Jinlan Reservoir water. Water Sci. Technol. Water Supply 13 (4), 1123–1129. Hong, H., Xiong, Y., Ruan, M., Liao, F., Lin, H., Liang, Y., 2013b. Factors affecting THMs, HAAs and HNMs formation of Jin Lan Reservoir water exposed to chlorine and monochloramine. Sci. Total Environ. 444, 196–204. Hong, H., Qian, L., Xiong, Y., Xiao, Z., Lin, H., Yu, H., 2015. Use of multiple regression models to evaluate the formation of halonitromethane via chlorination/ chloramination of water from Tai Lake and the Qiantang River, China. Chemosphere 119, 540–546. Hong, H., Song, Q., Mazumder, A., Luo, Q., Chen, J., Lin, H., Yu, H., Shen, L., Liang, Y., 2016. Using regression models to evaluate the formation of trihalomethanes and haloacetonitriles via chlorination of source water with low SUVA values in the Yangtze River Delta region, China. Environ. Geochem. Health 38, 1303–1312. Hong, H., Yan, X., Song, X., Qin, Y., Sun, H., Lin, H., Chen, J., Liang, Y., 2017. Bromine incorporation into five DBP classes upon chlorination of water with extremely low SUVA values. Sci. Total Environ. 590, 720–728. Hua, G., Reckhow, D.A., 2012. Evaluation of bromine substitution factors of DBPs during chlorination and chloramination. Water Res. 46 (13), 4208–4216.

L. Zheng et al. / Science of the Total Environment 698 (2020) 134250 Hua, G., Reckhow, D.A., Kim, J., 2006. Effect of bromide and iodide ions on the formation and speciation of disinfection byproducts during chlorination. Environ. Sci. Technol. 40 (9), 3050–3056. Hua, G.H., Reckhow, D.A., Abusallout, I., 2015. Correlation between SUVA and DBP formation during chlorination and chloramination of NOM fractions from different sources. Chemosphere 130, 82–89. Kaufman, J.A., Wright, J.M., Evans, A., Rivera-Nunez, Z., Meyer, A., Narotsky, M.G., 2018. Associations between disinfection by-product exposures and craniofacial birth defects. J. Occup. Environ. Med. 60 (2), 109–119. Krasner, S.W., Weinberg, H.S., Richardson, S.D., Pastor, S.J., Chinn, R., Sclimenti, M.J., Onstad, G.D., Thruston, A.D., Jr., 2006. Occurrence of a new generation of disinfection byproducts. Environ. Sci. Technol.40(23), 7175–7185. Liang, L., Singer, P.C., 2003. Factors influencing the formation and relative distribution of haloacetic acids and trihalomethanes in drinking water. Environ. Sci. Technol. 37 (13), 2920–2928. Lin, J., Li, Z., Yao, Q., Zhu, A., Hong, H., 2018. Formation of disinfection by-products upon chlorination/chloramination of Qianjiang source water. Acta Sci. Circumst. 38 (8), 3027–3035 (in Chinese). MOH, 2006. Hygienic Standard for Drinking Water in People's Republic of China (GB5749–2006). Obolensky, A., Singer, P.C., 2005. Halogen substitution patterns among disinfection byproducts in the information collection rule database. Environ. Sci. Technol. 39 (8), 2719–2730. Pan, S., An, W., Li, H., Su, M., Zhang, J., Yang, M., 2014. Cancer risk assessment on trihalomethanes and haloacetic acids in drinking water of China using disability-adjusted life years. J. Hazard. Mater. 280, 288–294.

7

Park, K.-Y., Yu, Y.-J., Yun, S.-J., Kweon, J.-H., 2019. Natural organic matter removal from algal-rich water and disinfection by-products formation potential reduction by powdered activated carbon adsorption. J. Environ. Manag. 235, 310–318. Plewa, M.J., Simmons, J.E., Richardson, S.D., Wagner, E.D., 2010. Mammalian cell cytotoxicity and genotoxicity of the haloacetic acids, a major class of drinking water disinfection by-products. Environ. Mol. Mutagen. 51 (8–9), 871–878. Richardson, S.D., Plewa, M.J., Wagner, E.D., Schoeny, R., Demarini, D.M., 2007. Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: a review and roadmap for research. Mutat. Res. 636 (1–3), 178–242. Sohn, J., Amy, G., Cho, J., Lee, Y., Yoon, Y., 2004. Disinfectant decay and disinfection byproducts formation model development: chlorination and ozonation by-products. Water Res. 38 (10), 2461–2478. Song, Q., Ning, P., Sun, H., Lin, H., Chen, J., Hong, H., 2017. Regression models of HAAs formation upon chlorination of source water collected from Yangtze River Delta. J. Environ. Sci. 37 (6), 2048–2054 (in Chinese). Sun, H., Song, X., Ye, T., Hu, J., Hong, H., Chen, J., Lin, H., Yu, H., 2018. Formation of disinfection by-products during chlorination of organic matter from phoenix tree leaves and Chlorella vulgaris. Environ. Pollut. 243 (Pt B), 1887–1893. USEPA, 2003. Method 552.3: Determination of Haloacetic Acids and Dalapon in Drinking Water by Liquid-Liquid Microextraction, Derivatization, and Gas Chromatography with Electron Capture Detection. EPA 815-B-03-002. Revision 1.0. Wagner, E.D., Plewa, M.J., 2017. CHO cell cytotoxicity and genotoxicity analyses of disinfection by-products: An updated review. J. Environ. Sci. 58, 64–76. Wintermans, J.F.G.M., De Mots, A., 1965. Spectrophotometric characteristics of chlorophyll a and b and their pheophytins in ethanol. Biochim. Biophys. Acta 109, 448–453.