Introduction

Chlorine or chlorinated compounds are the most common disinfectants in drinking water treatment throughout the world for protecting the water against microbial contamination and preventing microorganism regrowth in the water distribution system (Sérodes et al. 2003). However, when chlorine reacts with natural organic matter—such as humic and fulvic acids, which is widely found in water bodies, a wide range of disinfections by-products (DBPs) were produced (Johnson et al. 1982; Peters et al. 1980). Special attention has been paid to the occurrence of DBPs due to their toxic and potential carcinogenic effect (Richardson et al. 2007). Trihalomethanes (THMs), the most commonly observed classes of DBPs in drinking water, were the first DBPs identified and occurred at higher concentrations than other DBPs (Bellar et al. 1974; Chang et al. 2010; Goslan et al. 2014; Loyola-Sepulveda et al. 2013; Ristoiu et al. 2009; Sérodes et al. 2003; Toroz and Uyak 2005; Wei et al. 2010). Most of the studies found that BDCM, DBCM, and TBM were generally not genotoxic in the standard test systems (Cancer 1999). Nonetheless, brominated THMs are activated to be mutagenic by glutathione S-transferase-theta in a transgenic strain of Salmonella (Pegram et al. 1997). Some published results link THMs to cancer, kidney, and liver damage, fetal growth retardation, and birth defects (Wright et al. 2004). When administered by gavage, TCM and DBCM induced live tumors in the mouse. BDCM produced renal and liver tumors in the mouse, and renal tumor in the rat. TBM induced intestinal tumors in the mouse and rat (Jorgenson et al. 1985; Program 1985, 2006). In order to minimize the risk of human health effect, regulations or guidelines have been promulgated to control THMs in many countries or international organizations (Richardson 2003).

In addition to THMs, other DBPs have been reported in the many research, including haloacetic acids, chlorophenols, chloral hydrate, haloacetonitriles, haloketones, haloacetonitriles, and haloacetamides, more than 600 DBPs (Ristoiu et al. 2009; Wang et al. 2015). It is almost impossible to investigate all the DBPs in a certain region, so adsorbable organic halides (AOX) were chosen as investigation index to reflect the overall level of DBPs. In addition, few literatures about the occurrence of AOX are available.

In summary, the most prevalent DBPs THMs and AOX, which reflect the total level of DBPs, were selected as the focus of the survey. The current survey aims at documenting the occurrence, speciation and temporal, and different water sources variability of THMs and AOX in 16 urban drinking water work using chlorine dioxide or chlorine as disinfectant. It was conducted in Shandong Province which is located on the eastern coast of China, on the lower Yellow River, with a population of 98.47 million and an area of about 157,900 km2 (from the 2015 census), and has a typical temperate and monsoonal climate with four clearly distinct seasons. It is also an important water-receiving region in Eastern-Line South-to-North Water Transfer Project. Accordingly, the results of the investigation are aimed at enriching database about the occurrence of DBPs in drinking water in China as a technical literature for future water quality management purpose and setting future DBPs regulations. Meanwhile, the current investigation provides a vital reference data for other countries or regions with temperate and monsoonal climate around the world to set relevant to DBPs occurrence in the drinking water systems.

The chemistry of chlorinated disinfectants and the toxicology of THM and AOX in water

Chemical behavior of chlorinated disinfectants

Upon dissolution in water, chlorine gas hydrolyses rapidly to yield hypochlorous acid (HClO), hypohalate anion, and ClO−, and the composition of the resulting solutions is pH dependent (Lopez et al. 2001).

When a substitute for chlorine, sodium hypochlorite (NaOCl), is used as disinfectant, HClO is also formed. Hypobromous acid (HBrO) was generated from the action of HClO on bromide ion and dissolved in most natural waters (Gunten and Oliveras 1998). HClO (pKa = 7.42) and HBrO (pKa = 8.70) primarily existed in acidic and neutral solution, both of them are strong electroaffinity, which participate in addition and substitution reactions with various organic matter (Boyce and Hornig 1983). HClO and ClO− can react with organic compounds by addition, substitution, and oxidation, and chlorine substitution can lead to the formation of halogenated compounds, such as TCM (Emmanuel et al. 2004). Similarly, the corresponding Br-substituted products, BDCM, DBCM, and TBM, are generally thought to result from the reaction of HBrO and BrO− with organic compounds (Heeb et al. 2014).

Toxicology of THMs and AOX

All four of the regulated THMs are carcinogenic in rodents. When administered by drinking water, TCM increased the yield of renal tubular adenomas and adenocarcinomas in male rats (Jorgenson et al. 1985), and BDCM was carcinogenic in the male F344/N rat based on an increased hepatocellular neoplasia (George et al. 2002). When administered by gavage, DBCM is toxic to the liver and kidneys and produces hepatocellular tumors in mice (Dunnick et al. 1985). When administered by injection, TBM induces a pulmonary adenoma response in mice (Theiss et al. 1977). Based on population-based case-control study, excess risks of rectal and bladder cancers for women, the overall association of bladder cancer risk with duration of THMs, and colon cancer risk for males associated with cumulative exposure to THMs were found (Cantor et al. 1998; King et al. 2000; Koivusalo et al. 1997). Research on the pharmacokinetics of bromodichloromethane in humans indicated that water uses involving dermal contact can result in much greater systemic BDCM doses than water ingestion (Leavens et al. 2007), emphasizing the significant of exposure route in risk assessment of the brominated THMs (Christian et al. 2007; Ross and Pegram 2003). Wang et al. also found that the cancer risks resulting from intakes of THMs are variable not only by the type of THM, but also by the route of exposure (Wang et al. 2007). Jin Lee et al. evaluated the lifetime cancer risks related with different exposure pathways by THMs in swimming pools and found that swimmers can be at the greater risk from inhalation exposure (Jin et al. 2009). THMs exposure risk to human health during non-potable reuse was estimated, and the lifetime cancer risks estimation presented that inhalation exposure to chloroform in the using peak THMFP values showed the highest cancer risks of 1.28 × 10−6 and 1.12 × 10−6 to residential adult and child receptors, respectively (Aina and Ahmad 2013). However, DBCM, BDCM, and TBM have not shown genotoxic in vivo and are most impossible to have any significant genotoxicity in mammals (Stocker et al. 1997). In addition, AOX, which generated from NaClO reacted with hospital wastewater, its concentrations were strongly associated with EC50 (TU) on daphnia. But there is very few information available about toxicity of AOX resulted from disinfection treatment in drinking water. To regulate or make policy decision, it is necessary for collecting more information about toxicity, and pharmacokinetics of THMs and AOX. It is more meaningful to study the toxicity of total amount of DBPs (represented as AOX) than to study individual DBPs, since THMs or haloacetic acid or nitrogen-containing DBPs are never singular in drinking water, they are usually mixed together.

Materials and methods

Reagent

The THMs standards containing the four trihalomethane (trichloromethane [TCM], bromodichloromethane [BDCM], dibromochloromethane [DBCM], and tribromomethane [TBM]) obtained from Supelco (USA). Methanol was purchased from Fisher Chemicals (New Jersey, USA) with HPLC grade. Distilled water was prepared by a Milli-Q Synthesis water purifier (Millipore, Bedford, MA, USA).

Sample collection

The sampling sites are presented in Fig. 1, and the treatment process of each plant is described in Table 1. Sample collection lasted for 3 years, and a total of 538 samples (raw water and treated water) were collected for this study. In this study, water samples from 16 water works were distributed in 13 cities of Shandong province. These water works are also scientifically interesting because they use different types of sources as raw water (Yellow River reservoir, local reservoir and ground water).

Fig. 1
figure 1

Map of sampling locations

Full size image
Table 1 Summary of source water, treatment, and disinfection at 16 water plants
Full size table

Sample vials contained a preservative and quenching agent in keeping with US Environmental Protection Agency (EPA) Methods 551.1 and standard method 5320 B. Samples were collected in amber glass vials packed in a cooling box (4 °C) and delivered to the laboratory within 48 h. Samples were preserved in the dark at 4 °C until used as soon as they arrived in the lab. All samples were analyzed within 7 days of collection.

Analytical methods

THMs were extracted using a modified form of USEPA Method 551.1. The analytes were detected by gas chromatography (GC) (Agilent 6890) with electron capture detection (ECD). The method detection limits (MDLs) of TCM, BDCM, DBCM, and TBM were 4, 1.3, 2, and 1.4 μg/L, respectively. AOX was determined using AOX analyzer (Multi X 2500 Halogen analyzer, Analytik Jena AG). The procedure was based on standard methods 5320 B (APHA et al. 1998). The method detection limits (MDLs) of AOX were 10 μg organic Cl–/L. A TOC-VCPH analyzer was used for DOC detection (Shimadzu, Japan). The UV254 was measured by a UV-1800 UV/Visible spectrophotometer (Shimadzu, Japan).

Results and discussions

Occurrence of DBPs

Figure 2 presents the concentration profiles of total THMs and AOX in all the samples collected from 16 drinking water works in Shandong. Nine water works are disinfected with chlorine, and others are chlorine dioxide. The THMs concentration of the samples disinfected with chlorine ranged from 2.1–105 μg/L and N.D.–47.6 μg/L with chlorine dioxide. The THMs concentration of 93% of the samples disinfected with chlorine and all of the samples disinfected with was below 80 μg/L. That is the permissible levels of THMs in the USA (USEPA 1998). Nevertheless, according to Chinese water quality standards, the maximum contaminant level (MCL) of total THMs was not regulated, but the sum of the ratio of each THMs concentration to its respective MCL must be less than or equal to one (GB5749-2006 2006).

$$ \frac{C_{CH{ Cl}_3}}{60}+\frac{C_{CH{ Cl}_2 Br}}{60}+\frac{C_{CH Cl{ Br}_2}}{100}+\frac{C_{CH{ Br}_3}}{100}\le 1.0 $$
Fig. 2
figure 2

Concentration of DBPs in Shandong drinking water

Full size image

The calculated ratios were summed and 92% of the samples disinfected with chlorine and all of the samples disinfected with chlorine dioxide were below 1.0, indicating that the vast majority of the samples met Chinese total THMs regulations. Seventy-five percent of the THMs concentrations of the samples disinfected with chlorine were below 51.5 μg/L, and the median concentration of total THMs was 35 μg/L. That was similar to previous investigations on a national scale in the USA with THMs concentration 39 μg/L (Wang et al. 2015), and slightly lower than some European cities (Barcelona 85 μg/L, Scotland 74 μg/L) (Goslan et al. 2009; Goslan et al. 2014) and Canadian city (Quebec ,62 μg/L) (Sérodes et al. 2003), while was much higher than that reported from observations in other cities in China, Shenzhen (19.9 μg/L), Guangzhou (17.7 μg/L), Beijing (14.1 μg/L), and Taiwan (4.0–24.4 μg/L) (Chang et al. 2010; Gan et al. 2013; Huang et al. 2017; Wei et al. 2010).

The AOX concentrations of the samples disinfected with chlorine ranged from 11–238 μg/L that was similar to the USA (21–237 μg/L) (Weinberg et al. 2002), and the median concentration was 61 μg/L that was much higher than European countries (22 μg/L) (Palacios et al. 2000) The AOX concentrations of the samples disinfected with chlorine dioxide ranged from N.D.–102 μg/L, and the median concentration was 36 μg/L.

The median concentration of total THMs and AOX of the samples disinfected with chlorine dioxide was much lower than chlorine. Volk et al. found that the substitution of chlorine by chlorine dioxide corresponded to an 81% reduction in THMs concentrations (Volk et al. 2002). The results of several studies indicate that under the same reaction conditions, the AOX generated with chlorine dioxide is 1–25% of the AOX generated with chlorine (Aieta and Berg 1986). It has been well established that the samples treated with chlorine-free chlorine dioxide do not form THMs (Hua and Reckhow 2012; Richardson et al. 2000). However, when chlorine was used as the secondary disinfectant after the application of chlorine dioxide, THMs were found (Richardson et al. 2000). Appreciable amounts of THMs (5–25 μg/L over 50%) which were found in water works used chlorine dioxide as disinfectant, which indicates that chlorine dioxide used in these water works is not pure, and by-products chlorine may be produced during on-site preparation of chlorine dioxide. It is advised that the purity of chlorine dioxide must be increased until no chlorine is formed during the preparation of chlorine dioxide, if they want to further reduce the production of THMs.

Seasonal DBPs variations

There are apparent temporal variations in the concentration levels of THMs and AOX, and the distribution of concentration of DBPs in four seasons is summed up in Fig. 3. THMs and AOX levels were high in spring and low in summer. This may be partly related to the fact that the spring water temperature is higher, thus accelerating the production reaction of THMs and AOX (Guilherme and Rodriguez 2014; Huang et al. 2017; Mercier Shanks et al. 2013; Wei et al. 2010). Even though the highest average temperature appears in summer, rainfall is greatly heavy in summer in Shandong, which causes a decrease in DOC level in drinking water. Besides, the average doses of disinfectant used (chlorine and chlorine dioxide) were lower in summer than in spring (Table 2). However, although the average temperatures observed in autumn and winter were much lower than in summer (Table 2), the concentrations of DBPs in autumn and winter were generally higher than in summer. This is likely explained by comprehensive factors, relating to the fact that the average usage doses of disinfectant (chlorine and chlorine dioxide) and the average DOC concentrations were higher in autumn and winter than in summer (Table 3).

Fig. 3
figure 3

Seasonal variation in DBPs concentrations in Shandong drinking water

Full size image
Table 2 Average values of water quality and operational parameters at the treatment plant in the different seasons
Full size table
Table 3 Average values of water quality and operational parameters at the treatment plant of the different water sources
Full size table

Distribution of DBPs concentration in different water sources

As presented in Fig. 4, the change of DBPs concentration of different water sources was in the order local reservoir > Yellow River reservoir > ground water. There are three main sources of drinking water in Shandong: local reservoir water, Yellow River reservoir water, and ground water. The average concentration for THMs was 31.9 μg/L in local reservoir water, 21.1 μg/L in Yellow River reservoir water, and 2.5 μg/L in ground water and for AOX in local reservoir water, Yellow River reservoir water, and ground water were 70 μg/L, 53 μg/L, and 25 μg/L, respectively. The reason why the DBPs level of ground water was the lowest may be that the DOC concentration of ground water sources was lower than the two other water sources. Moreover, most of the water works that use ground water as water source are disinfected with chlorine dioxide. The operational parameters at the treatment plants and the DOC level of water quality of local reservoir water were similar to Yellow River reservoir water, but DBPs level of local reservoir water was slightly higher than Yellow River reservoir water (Table 3). This was possibly related to that fraction distributions of DBPs precursors were different between local reservoir water and Yellow River reservoir water, while DBPs formation highly rests with the nature of dissolved organic matter (Niu et al. 2015; Wei et al. 2008; Zheng et al. 2016). It should be noted that the fraction distributions of DBPs precursors of the two types of water sources need to be studied.

Fig. 4
figure 4

The DBPs of different water sources in Shandong drinking water

Full size image

Speciation of the THMs

The distribution of individual species of THMs is summarized in Fig. 5. The mean concentrations of TCM, BDCM, DBCM, and TBM of samples disinfected with chlorine were 5.8, 9.0, 10.1, and 5.8 μg/L, respectively, higher than chlorine dioxide (impurity, mixed with chlorine) 0.6, 2.9, 2.5, and 4.4 μg/L. TCM, BDCM, DBCM, and TBM of samples disinfected with chlorine account for 19%, 29%, 33%, and 19% of total THMs, and 6%, 28%, 24%, and 42% for chlorine dioxide, respectively. Apparently, as chlorine replaced by chlorine dioxide as disinfectant, the ratio of TCM, DBCM to total THMs decreased, but TBM increased significantly. It is significantly noted that the amount of THMs formation was suppressed when water was treated with ClO2 and C12, but the percentage of TBM may prominently improve. Jun Wen Li et al. had drawn a similar conclusion that the four species THMs formation potential decreased as various usage amount of ClO2 and C12, and the ratio of TCM, DBCM to total THMs decreased by 9% and 2% respectively, TBM increased by 13% when ClO2/C12 was from 0 to 0.2 (Li et al. 1996).

Fig. 5
figure 5

The ratio of individual speciation to total THMs with different disinfectants

Full size image

Bromine incorporation

According to some toxicological studies, brominated DBPs may present higher health risks than chlorinated DBPs, so the formation of bromine-containing DBPs is of particular interest (Richardson 2003). The bromine substitution factor (BSF) was introduced in some studies, which represent the ratio of the molar concentration of bromine to the total halogen of some class of DBPs and applied to measure the bromine substitution among different DBPs classes (Hua et al. 2006). For THMs, the BSF can be

$$ \mathrm{BSF}\left(\mathrm{THMs}\right)=\frac{0\times \left[\mathrm{CH}{\mathrm{Cl}}_3\right]+1\times \left[\mathrm{CH}\mathrm{Br}{\mathrm{Cl}}_2\right]+2\times \left[\mathrm{CH}\mathrm{Cl}{\mathrm{Br}}_2\right]+3\times \left[\mathrm{CH}{\mathrm{Br}}_3\right]}{3\times \left(\left[\mathrm{CH}{\mathrm{Cl}}_3\right]+\left[\mathrm{CH}\mathrm{Br}{\mathrm{Cl}}_2\right]+\left[\mathrm{CH}\mathrm{Cl}{\mathrm{Br}}_2\right]+\left[\mathrm{CH}{\mathrm{Br}}_3\right]\right)} $$

calculated by the expanded equation above. The BSF of THMs ranged from 0.19 to 0.8, and the mean value was 0.50, much higher than Shenzhen. The reason is possibly that much higher level of bromide (109 μg/L on average) has been obtained in reservoir water in Shandong than Shenzhen (28 μg/L on average). This phenomenon was consistent with the report that a positive correlation was observed between BSF and initial concentration of bromide for THMs formation studied (Hua et al. 2006). Figure 6 shows the relationship between BSF of THMs and initial concentration of bromide. Here we can see that all chlorination data (Cl2) has a Spearmanρ (ρ) of 0.38, and was lower than the result 0.59, that was found in Scotland (Goslan et al. 2009). The difference was attributed to the ratio of bromine consumption to chlorine consumption which has been recognized as a significant role in bromine substitution during THMs formation (Wu and Chadik 1998).

Fig. 6
figure 6

Bromide concentration against trihalomethane (THMs) bromine substitution factor in reservoir water

Full size image

Conclusion

This study determined the occurrence of THMs and AOX in 16 drinking water works using chlorine or chlorine dioxide as disinfectant. These water works distribute in Shandong province of China. The resulting data depict portraits of THMs and AOX in drinking water in the area, and there is currently very little data available on the subject. The portrait shows that the median concentration of total THMs and AOX of the samples disinfected with chlorine was 35 μg/L and 61 μg/L, respectively, much higher than chlorine dioxide. It is possibly inferred that if water works desire to further inhibit the formation of THMs and AOX, chlorine dioxide may be an alternative disinfectant to substitute chlorine. The seasonal variation for THMs and AOX followed the order spring > autumn-winter > summer. The variation of THMs and AOX concentration in different water sources was as follows: local reservoir > Yellow River reservoir > ground water. The ratio of TBM to THMs of samples disinfected with chlorine dioxide was higher than chlorine. The BSF of THMs and initial concentration of bromide shows a weak positive correlation. This survey succeeded in enriching the knowledge of the presence, the temporal, and different water resource variabilities of the concentrations of THMs and AOX in drinking water works. In this work, no observations were made on the concentration of THMs, AOX, DCA, TCA, DOC, and chlorine residues in the water available at consumer faucets. In addition, the TOC has not been studied on water from the following ZZ, JN, and LC points where chlorine disinfection is the only treatment given to these raw waters; in order to complete and validate these results in the study areas, it seems necessary to carry out further studies on THMs and AOX in drinking water.