Serving as the last barrier to secure drinking water safety, household water treatment and safe storage (HWTS) is perceived as an interim measure for removing pathogens from drinking water and reducing disease risk. In recent years, the application of HWTS has shown a growing trend, and its performance in controlling chemicals has also received much attention. Disinfection by-products (DBPs) are formed by the reaction of chemical disinfectants and precursors, and are present at sub-μg·L−1 or low-to-mid-μg·L−1 levels in drinking water. Although precursor control and disinfection operation modification could contribute to DBP mitigation to some degree, DBP removal after their formation emerges as an important strategy due to the ubiquitous existence of DBPs in distribution systems and tap water. In order to figure out how DBP concentrations vary during the residence time of drinking water in households, this review summarizes the effectiveness and mechanism of HWTS and combination technologies for DBP control in municipal tap water, and makes a comparison with regard to technologies implementing different removal mechanisms as well as DBPs possessing different natures. Based on these results, this article provides an insight into DBP risk assessment and human health protection.
The performances and mechanisms of household treatment for DBP control were reviewed.
Point-of-use filtration, boiling and relevant combination technologies could remove DBPs significantly.
Different technologies showed varying removal rates regarding different DBPs species.
DBPs formation, transformation and removal could occur simultaneously during household treatment.
adsorbable organic bromine
adsorbable organic chlorine
adsorbable organic iodine
adsorbable organic halogen
carbonaceous disinfection by-products
drinking water distribution system
drinking water treatment plants
sum of dihalogenated HAAs
sum of MCAA, DCAA, TCAA, MBAA and DBAA
household water treatment
household water treatment and safe storage
- log Kow
logarithmic octanol-water partitioning coefficient
limit of quantification
natural organic matter
nitrogenous disinfection by-products
sum of TCM, BDCM, DBCM and TBM
total organic chlorine
total organic bromine
total organic halogen
sum of trihalogenated HAAs
World Health Organization
United Nations Children's Fund
US Environmental Protection Agency
According to recent reports published by the World Health Organization (WHO) and United Nations Children's Fund (UNICEF), although 90% of the global population uses a basic drinking water source, over two billion people worldwide lack access to safely managed drinking water services, and approximately 485,000 diarrheal deaths in low- and middle-income countries are attributable to unsafe drinking water annually, accounting for over half of the diarrheal disease burden all over the world (WHO 2019; WHO/UNICEF 2019). Drinking water safety risks associated with the disease could be minimized from catchment to consumer, and within this framework, household water treatment and safe storage (HWTS) serves as an important interim measure used for waterborne pathogen removal and disease risk reduction (WHO 2012). The main types of household water treatment (HWT) methods include disinfection with chemicals, disinfection with heat, disinfection with solar light or ultraviolet light (UV), filtration as well as flocculation-disinfection. Rosa & Clasen (2010) extracted data on reported HWT practices from 67 national surveys and supposed that the actual number of people practicing HWT in low- and middle-income countries might exceed 1.5 billion. On the one hand, HWT is an effective intervention in households where continuous access to safe piped-in water is not available, on the other hand, the employment of HWTS shows an increasing trend in institutional settings such as schools and health care facilities (WHO 2019). The microbiological performance of HWTS technologies has been investigated in different countries or regions (Brown & Sobsey 2012; Lantagne & Clasen 2012) and evaluated globally based on WHO criteria (WHO 2016, 2019). The effectiveness of HWTS in removing chemical contaminants (e.g., trace organic compounds, metals, nitrates, residual chlorine, and other inorganics) also raises concern (Anumol et al. 2015; Brown et al. 2017). However, there is a current paucity of studies reviewing the performance of HWTS in controlling disinfection by-products (DBPs), which are regarded as a series of important emerging contaminants in drinking water (Richardson & Ternes 2011).
Disinfection of drinking water supplies commenced early in the 20th century with the purpose of deactivation of waterborne pathogens and has been rightly hailed as one of the greatest achievements in public health (Li & Mitch 2018). Nevertheless, disinfection could result in an unintended hazard concerning human health: DBP formation caused by the reaction of chemical disinfectants (chlorine, chloramines, ozone, chlorine dioxide, etc.) with natural organic matter (NOM), halide and anthropogenic contaminants (Richardson & Ternes 2011). It has been demonstrated that DBPs exhibit toxicological properties (e.g., cytotoxicity and genotoxicity) and are potentially responsible for an increase in the risk of cancer and other adverse outcomes, including reproductive and developmental effects (Cantor et al. 1998; Richardson et al. 2007; Nieuwenhuijsen et al. 2009). The occurrence (Krasner et al. 2006; Templeton & Chen 2010; Bei et al. 2016), analytical methods (Yang & Zhang 2016; Ding & Chu 2017), precursors and formation mechanisms (Shah & Mitch 2011; Bond et al. 2012; Ding et al. 2019a), health effects (Hrudey 2009; Wagner & Plewa 2017), as well as control technologies (Singer 1994; Chu et al. 2015; Hu et al. 2018) of DBPs in drinking water have gained considerable attention since the first discovery of DBPs in the mid-1970s (Bellar et al. 1974; Rook 1974).
The strategies for DBP mitigation include precursor control (source control and precursor removal), disinfection operation modification as well as DBP removal after their generation (Singer 1994; Sun et al. 2019). However, DBP precursors are present in drinking water inevitably and residual disinfectants is required to maintain microbial safety, thus formation of DBPs will further occur in drinking water distribution systems (DWDS) and plumbing systems (Dion-Fortier et al. 2009; Chowdhury et al. 2011; Li et al. 2019), making it significant to consider removal of DBPs just before human ingestion. Serving as the last barrier to ensure drinking water safety, HWT methods (boiling, filtering, UV treatment, etc.) combined with storage have been reported to result in DBP variations (Weinberg et al. 2006; Chowdhury et al. 2010), and the related processes include: (1) DBP formation due to thermal cleavage of larger halogenated species and additional reactions between precursors and residual disinfectants; (2) DBP transformation mainly through hydrolysis and decarboxylation; and (3) DBP removal by means of membrane rejection, physical adsorption, UV photolysis, volatilization, and so on.
With the purpose of understanding the effectiveness of HWTS in managing drinking water quality risks from a DBP perspective and improving the assessment accuracy of human exposure to DBPs by ingestion, this paper attempts to provide a review of the DBP concentration variations during HWTS processes. First, the microbial performance and worldwide application status of HWT methods as well as the occurrence level and regulation (or guideline) of DBPs are briefly introduced. Then, the effects of HWTS on DBP concentrations and the mechanisms relevant to DBP variations are presented. The final part gives a summary of HWTS performance for DBP control, provides insight about DBP exposure assessment and human health protection, and makes a conclusion.
HOUSEHOLD WATER TREATMENT
Another important aspect about HWT technologies is their application across the globe. Figure 1 shows the application status of several typical HWT methods (including the proportion of individual methods used in households and the percentage of population treating drinking water at home) in six selected regions with low- and middle-income families surveyed by Rosa & Clasen (2010). In general, 33% of households were reported to treat water at home, with the practice being widespread in the Western Pacific region (66.8%) while less common in the Eastern Mediterranean (13.6%) and Africa (18.2%). Boiling was reported to be the most dominant HWT method and was described by approximately one-fifth of households in these countries (Rosa & Clasen 2010). Moreover, in early 2015, the WHO commissioned a rapid market assessment of commercial HWT products using literature review, interview, and field visits. A strong growth in filter markets in parts of Asia (e.g., Vietnam, China, South Korea) was found, which might be associated with consumers' growing awareness of drinking water quality and human health, and the wide availability and affordability of HWT products (WHO 2016). Different from people who use HWT mainly for pathogen control, North Americans are more familiar with point-of-use (PoU) treatment (i.e., HWT) for taste and odor control, or less commonly for removal of metals or organics (Berg 2015). It was estimated that ∼40% of the population in the United States used water purifiers at home (Anumol et al. 2015).
DBPS IN DRINKING WATER
Chemical constituents of drinking water may come from source waters, treatment process, distribution system, and consumers' plumbing (Ministry of Health 2018). Since DBPs are formed after the addition of disinfectants and their concentrations might change during delivery (Fang et al. 2019a; Li et al. 2019), research relevant to DBP occurrence mostly collected water samples from drinking water treatment plants (DWTPs) outlet, DWDS, and water faucets. Standards or guidelines for drinking water quality proposed by different countries, regions, and organizations generally provide regulatory compliance for DBPs, including their regulatory limits or guideline values, as well as corresponding monitoring requirements (sampling sites, monitoring frequencies, etc.).
Organic halogenated DBPs commonly occurring in drinking water include carbonaceous DBPs (C-DBPs), such as trihalomethanes (THMs), haloacetic acids (HAAs), haloacetaldehydes (HALs) and nitrogenous DBPs (N-DBPs), such as halonitromethanes (HNMs), haloacetonitriles (HANs) and haloacetamides (HAMs). These DBPs possess a similar CX3R structure (X = H, Cl, Br or I, R = functional group) and exhibit significantly higher concentration or higher toxicity, thus are classified as CX3R-type DBPs (Ding et al. 2019b; Fang et al. 2019b). CX3R-type DBPs are generally present in drinking water at sub-μg·L−1 or low-to-mid-μg·L−1 levels, and THMs and HAAs are two major classes on a weight basis. The levels of representative CX3R-type DBPs in finished water of DWTPs, DWDS water, and tap water are summarized in Table 2. The concentrations and speciation of DBPs in drinking water are dependent on several factors, including nature and concentration of DBP precursors (Westerhoff & Mash 2002; Chow et al. 2005; Chow 2006), disinfection scenarios (e.g., disinfectant type, disinfectant dose, disinfectant contact time) (Barrott 2004; Hua & Reckhow 2007), and other characteristics of water quality such as pH and water temperature (Yang et al. 2007; Hua & Reckhow 2008).
Physiochemical properties are of significance when assessing the behavior of DBPs in natural waters, DWTPs, DWDS, as well as households. For instance, logarithmic octanol-water partitioning coefficient (log Kow) is acknowledged to be associated with polarity and water solubility, and the substances with high log Kow values tend to be more readily adsorbed on organic matter owing to their low affinity to water. Furthermore, boiling point (BP) and Henry's law constant are related to DBP volatilization, which is a predominant process for volatile DBP reduction during boiling treatment and can also occur in DWTPs (Wu et al. 2001; Qiu et al. 2019). As suggested by Krasner & Wright (2005), although Henry's law constants are not determined for boiling water at 100 °C, DBPs which are not volatile (Henry's law constant at 25 °C <10−7 atm-m3/mol) would still be considered relatively non-volatile in boiling water. Table 3 lists some physicochemical properties of commonly occurring CX3R-type DBPs.
Regulations and guidelines
Numerous countries in Asia, Europe, America, Oceania, as well as Africa have included DBPs in their standards or guidelines for drinking water quality (European Union 1998; Ministry of Health of the People's Republic of China 2006; WHO 2006; USEPA 2009; National Health & Medical Research Council & National Resource Management Ministerial Council 2011; Ministry of Health Labour & Welfare 2015; South African National Standard 2015; England & Wales 2016; WHO 2017; Ministry of Health 2018; Health Canada 2019), and the regulatory limits or guideline values of different organic DBPs in selected standards or guidelines are listed in Table 4. THMs and HAAs are regulated in the majority of cases, while only a few standards or guidelines consider N-DBPs such as HANs and N-nitrosodimethylamine (NDMA), which are of greater perceived health risk than regulated C-DBPs (Muellner et al. 2007; Bond et al. 2011).
With respect to drinking water quality monitoring, the WHO (2017) proposed that sampling locations should include points near the extremities of DWDS and taps connected directly to the mains in houses when the concentrations of target constituents can change during distribution, while sampling at DWTPs or at the head of the DWDS may be sufficient for constituents whose concentrations do not change during delivery. In mainland China, samples aimed at monitoring regulated DBPs should be collected from the outlet of DWTPs and extremities of DWDS (Ministry of Construction of the People's Republic of China 2005; Ministry of Health of the People's Republic of China 2006). In the case of the USA, USEPA (2006) provided different monitoring locations for different DBPs and the requirements are as follows: THM4/HAA5 (several points in DWDS), bromate (one point at entry point to DWDS), and chlorite (one point at entry point to DWDS and three points in DWDS). As for the UK, the points of compliance set for THMs and bromate are consumers' taps (England & Wales 2016).
DBP VARIATIONS IN HOUSEHOLD TREATMENT PROCESS
Water purification with PoU filters (filtration method)
A PoU filter attached directly to the consumer's tap or pitchers is perceived as an effective solution to DBP issues, and its performance for DBP removal may vary depending on the features of the filter (e.g., filter technique, filter manufacturer, aging condition) and the characteristics of water to be treated (Egorov et al. 2003; Leuesque et al. 2006; Wright et al. 2006). To our knowledge, much work on DBP removal through membrane rejection (size exclusion, charge repulsion, and hydrophobic interaction effects) or physical adsorption has been carried out (Tung et al. 2006; Fujioka et al. 2013; Wang et al. 2018). Most of these studies focused on removal mechanisms and influencing factors, thus solutions merely containing one or several DBP compounds with extremely high concentration were employed in these experiments to avoid interference caused by other matters and amplify the removal effect. However, since drinking water is a complex mixture comprising various substances, this section mainly takes into account the results obtained in experiments involving commercially available PoU filters and real tap water, which seems to be more practical and representative. Filtering systems consisting of activated carbon (AC) and ion exchange resin, reverse osmosis (RO) filters, microfiltration (MF) filters, as well as non-membrane pressure filters have been used to investigate the impact of domestic filtering devices on DBPs in municipal tap water (Gibbons & Laha 1999; Egorov et al. 2003; Ahmedna et al. 2004; Leuesque et al. 2006; Weinberg et al. 2006; Chowdhury et al. 2010; Rahman et al. 2011; Carrasco-Turigas et al. 2013; Stalter et al. 2016). The tested filters, target DBPs, and key findings are shown in Table 5.
Gibbons & Laha (1999) conducted an investigation into water purification systems employed for treating municipal water, and reported that the home water filtration system involving AC and ion exchange resin performed an almost 50% reduction in total THMs while commercial filter units did not contribute to a water quality improvement, which was likely owing to the improper operation. The removal efficiency of THMs, HAAs, and 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) with three commercially available home filters was assessed by Egorov et al. (2003). In this study, the removal rate of MX (up to >99%) and TCM by tested filters was consistently higher than that of HAAs, and TCM removal efficiency of 90% by Brita filters and 95% by Aquaphor filters was reported after 150 L of use, which was the manufacturer-recommended capacity of both filters. As Aquaphor filter use increased from 1 to 150 L, reductions of 97 and 59% for DCAA as well as 95 and 63% for TCAA were observed. However, a negative TCAA removal (−19%) by a Brita filter after 150 L of use demonstrated that longer use of a cartridge in filters might result in the elution of previously accumulated HAAs, which was probably attributed to HAA replacement by compounds with higher affinity to binding sites (Egorov et al. 2003).
The claim that effectiveness of DBP removal by pitcher-PoU devices depended on DBP physicochemical properties was suggested by Weinberg et al. (2006) according to the observation that pitcher-PoU devices had poorer average removal efficiency of TCM, DCAA, and TCAA than their bromine-containing counterparts; this might be because the latter possessed lower aqueous solubility and polarity, making them more amenable to carbon adsorption. In this study, reduction in THMs (93–99%) and HAAs (68–95%) by four different AC PoU devices were reported. In addition, 48% reduction in THM and 100% removal of HAAs was achieved in this research with the use of RO/deionized water system; the considerably high removal rate of HAAs could be explained by the fact that HAAs are in an anionic state at the pH of tap water and thus could be amenable to removal on the deionization resins through ion exchange (Weinberg et al. 2006). With regard to filters' artificial aging, it has been found that there was no real differences in THM reduction when using new filters (91%) or used filters (89%) for Brita pitchers, while the artificial aging of the filter had some effect on HAA removal, and HAA reduction was approximately 72 and 54% for new filters and used filters, respectively (Leuesque et al. 2006). Average reduction of HAAs by filter involved in this study was lower than that obtained for THMs, which might be because HAAs have a higher affinity to water than THMs, and thus THMs could be more easily adsorbed on the carbon particles contained in the filter.
Smith & El Komos (2009) evaluated the treatment performance of several PoU units with respect to various water quality parameters and reported that both properly functioning granular AC (GAC) units and RO units could remove about 70% of THMs. With the intention of developing models to predict the effects of indoor handling strategies on DBPs, Chowdhury et al. (2010) collected water samples and found that using commercial filters could reduce THMs and HAAs by 84–89% and 58–71%, respectively. Another study examining the effects of tap water processing on DBPs revealed that jug filter containing AC and ion exchange resin could produce substantial and statistically significant reductions in THMs (93%) and HAAs (82%) (Rahman et al. 2011).
In addition to the aforementioned organic DBPs, inorganic DBP such as bromate was also considered and bromate showed a less consistent pattern (Carrasco-Turigas et al. 2013). In the same study, RO appeared to be a good method for brominated THM elimination (99%) while TCM had higher removal rate using pitcher filter (91% for new filter) compared to a RO system (56% reduction), and MX concentration decreased below the limit of quantification (LOQ) after filtration (Carrasco-Turigas et al. 2013). In contrast with previous researchers who mainly focused on specific DBP species, Stalter et al. (2016) selected adsorbable organic halogen (AOX), which is also known as TOX, as a sum parameter for all halogenated DBPs, and examined the effect of 11 filters on AOX level. Of the filters examined, RO and a gravity filter containing AC removed >94% of AOX, and the removal efficacy of adsorbable organic bromine (AOBr) was generally higher than that of adsorbable organic chlorine (AOCl) and adsorbable organic iodine (AOI) for most filters. Also, the results of toxicity assays showed that 7 out of the 11 filters tested reduced cytotoxicity, oxidative stress response, and genotoxicity by >60% (Stalter et al. 2016).
Water treatment with boiling devices (thermal method)
DBP variations during water boiling might be associated with several processes, including DBP removal through volatilization, hydrolysis or destruction, DBP further formation as a result of cleavage of larger halogenated intermediate species or accelerated reactions between residual disinfectants and corresponding precursors, as well as DBP transformation from one species to another, including base-catalyzed hydrolysis (e.g., DCAN to dichloroacetamide) and thermal decarboxylation (e.g., TCAA to TCM) (Li & Sun 2001; Reckhow et al. 2001; Wu et al. 2001; Zhang & Minear 2002; Krasner & Wright 2005; Pan et al. 2013; Zhang et al. 2015; Ma et al. 2017; Shi et al. 2017). Several documented variation mechanisms of CX3R-type DBPs during boiling water are shown in Figure 2. Therefore, DBPs possessing different volatility or thermo-stability may behave in different ways under the same conditions, making it critical to consider individual species when estimating thermal effects on DBPs.
In order to provide a more realistic and intensive understanding about the effect of boiling water on DBPs in drinking water, and distinguish boiling water discussed in this section from heating water since water temperature has been proven as a significant factor influencing DBP volatilization (Batterman et al. 2000), what follows will focus on DBP variations in real tap water during the process of boiling, and the experimental conditions and a series of key findings are detailed in Table 6. Although volatilization occurring during tap water boiling could potentially increase airborne DBP concentrations and thus inhalation exposures, small impacts were expected given the large amount of dilution in most indoor environments (Batterman et al. 2000).
The effect of boiling water on concentrations of 21 target DBPs and TOX in chlorinated tap water was explored by Wu et al. (2001) using glass beakers covered with watch glasses. As boil time increased from 1 to 5 min, reductions of 68 and 83% for TCM as well as 74 and 95% for BDCM were observed, along with a decrease in TCAA (31–46%) and an increase in MCAA (27–30%). Moreover, a two-fold increase of DCAA levels upon boiling was worth noting, and other identifiable DBPs declined dramatically. With regard to TOX, boiling of tap water reduced TOX by 28 and 44% after 1 and 5 min, respectively. THM variations upon boiling was dominated by volatilization reactions in an open system, where THMs were almost completely removed, while the volatilization of HAAs could be ignored during boiling. It has also been suggested that HAA destruction reactions took over at higher temperatures and for more highly halogenated species, while HAA formation was more important for less halogenated species (Wu et al. 2001).
Since different residual disinfectants would contribute to DBPs formation to varying degrees and overall DBP variations would depend on the competition of formation rate and removal rate, Krasner & Wright (2005) employed a tea kettle on a stovetop to determine the impact of boiling on DBP concentrations in both chloraminated and chlorinated tap waters with a boiling time of 1, 2, and 5 min. As for HAAs in chloraminated water, no significant change was detected in the concentrations of the DXAAs (sum of dihalogenated HAAs) upon boiling, whereas the levels of the TXAAs (sum of trihalogenated HAAs) decreased over time (e.g., 9–37% for TCAA, 32–61% for TXAAs). However, increased DXAA concentrations (58–68%) were detected in the boiled chlorinated sample while TCAA concentration was unchanged after boiling chlorinated water for 1 min, but a 30% reduction was observed after 5 min. Furthermore, THM concentrations were reduced in both chloraminated (74–98%) and chlorinated (64–98%) waters upon boiling, and the reduction of 75% for TCM in chloraminated water as well as the reduction of 34% for TCM in chlorinated water after a 1 min boil indicated that simultaneous formation and volatilization of TCM was occurring. Most of the remaining DBPs (e.g., TCAL, HAN, HNM) were removed by at least 90% after 1 min of boiling in both samples.
Tests have also been performed with the utilization of an electric kettle that automatically cut off when boiling began (Weinberg et al. 2006). This study found that boiling of cold tap water with free chlorine residual removed more than 98% of the THM4 concentration. Leuesque et al. (2006) evaluated the extent to which boiling water influenced DBP concentration under a relatively different condition, where chlorinated tap water was boiled using a plastic kettle which was turned off 30 s after the onset of boiling. THM reduction alone by boiling water for 30 s averaged 83% while average total concentrations of HAAs apparently did not change during boiling. Regarding individual species of HAAs, boiling resulted in a statistically significant increase in DCAA levels (on average 35%, p value <0.001) and a statistically obvious decrease in TCAA levels (on average 42%, p value <0.001) (Leuesque et al. 2006). A reduction of 82–83% for THMs and a decrease of 34–62% in TCAA upon boiling was found in another study, as well as an increase of 22–49% in DCAA (Chowdhury et al. 2010).
Even though the aforementioned studies provided a detailed description about the effect of boiling water on DBP concentrations in tap water, other research attempted to make a comparison among the performances of various household boiling devices. A domestic electric kettle keeping water at a rolling boil for 10–15 s and an instant boiling water unit were compared for treating DBPs in chloraminated water (Rahman et al. 2011). In general, the mean concentrations of THMs were consistently and substantially reduced by boiling (86–94%) while HAAs showed a different behavior. Kettle boiling did not reduce total HAAs but instant boiling produced an average reduction of approximately 28%. Furthermore, boiling appeared to reduce mean TCAA concentrations (25% for kettle and 94% for instant unit) but increase mean DCAA concentrations (12% for kettle and 17% for instant unit), and bromochloroacetic acid (BCAA) concentrations were substantially reduced in instant boiled water (averaging 53%) but not in kettle-boiled water, suggesting that a longer time at a high temperature was required for BCAA breakdown (Rahman et al. 2011).
Interestingly, Carrasco-Turigas et al. (2013) took lidded electric kettle, saucepan, and microwave oven into consideration and made a systematical comparison. Among the three boiling devices, the highest THM reduction was observed for the microwave oven (97%) with a very high decrease of the brominated forms, and the kettle experienced the lowest reduction in THM4 levels (48%). In addition, TCM presented a higher percentage of removal than brominated analog in a saucepan, where bromate concentration increased after boiling water. As for MX, the concentration of MX decreased after all boiling experiments. Possible explanations for these results included: (1) the presence of a lid created a semi-sealed environment for kettle experiments; (2) in a microwave, big bubbles were created when water started boiling and ‘hot spots’ could occur due to faster temperature rise and homogeneously dissipated energy; and (3) when water was boiled using a saucepan, the heat source was not very powerful and it took longer for water to boil, leading to the formation of very small bubbles as well as little agitation (Carrasco-Turigas et al. 2013).
Since the adverse health effects of halogenated DBPs in humans have been proven to be associated with tap water ingestion (Li & Mitch 2018), the effect of boiling water on DBP concentration as well as developmental toxicity of water has been explored by Liu et al. (2015). The concentrations of total organic chlorine (TOCl) and total organic bromine (TOBr) in the two tap waters tested substantially reduced after 5 min boiling in open glass beakers. TOCl and TOBr in tap water 1 were reduced by 39 and 44%, respectively, and TOCl and TOBr in tap water 2 were reduced by 52 and 38%, respectively. Another important result was that 5 min boiling could significantly decrease the toxicity of tap water based on the data of developmental toxicity against the polychaete Platynereis dumerilii (Liu et al. 2015).
Water disinfection with UV light (photolytic method)
Most household water treatment technologies employ low-pressure lamps that emit UV radiation at 254 nm to inactivate microorganisms, and the effectiveness of UV disinfection is dependent on the delivered UV fluence, which is associated with intensity and exposure time (WHO 2019). Compared with other typical HWT methods, the home use of UV light for water disinfection might be limited due to its requirement of power sources, professional maintenance, and relatively higher cost. Very few studies have explored the performance of household UV treatment in removing DBPs in municipal drinking water; however, UV treatment is regarded as an effective method for DBP elimination at reuse facilities as well as swimming pools (Mitch et al. 2003; Hansen et al. 2013).
Chuang et al. (2016) investigated the UV direct photolysis of 26 halogenated DBPs and presented their corresponding photolysis rate constants as well as quantum yields under low pressure UV irradiation at 254 nm. Considering the lack of research relevant to UV photolysis of iodinated N-DBPs, Zhang et al. (2019) reported the direct UV photolysis degradation rate constants of 40 halogenated DBPs with the use of a low-pressure mercury vapor lamp emitting irradiation at 254 nm and established a quantitative structure–activity relationship model based on the experimental results. Both of these studies focused on the fluence-based UV photolysis rate constants of DBPs (unit: cm2 mJ−1), which depend on DBP molar extinction coefficients and their quantum yields. It has been found that the fluence-based photodegradation rate constants of DBPs by UV254 alone are affected by DBP halogen substitution patterns including halogen substituents species and halogenation degrees, and the rate constants follow the trend of iodo- > bromo- > chloro-DBPs and tri- > di- > monohalogenated DBPs; besides, DBP classes with more functional groups and larger molecular volume are more susceptible towards UV photolysis (Chuang et al. 2016; Wang et al. 2017; Zhang et al. 2019).
In addition, water storage was also considered as a handling scenario to explore DBPs variations during the residence time of drinking water in households. In a study where tap water was refrigerated at 4 °C in a high density polypropylene bottle overnight for 12 h, Weinberg et al. (2006) reported that refrigeration of cold tap water in an open container appeared to have negligible impact on the THM levels, with a removal rate of only ∼8%. Considering that DBP removal by volatilization during storage might be influenced by the seal condition of containers, experiments storing tap water in covered or uncovered polypropylene pitcher in refrigerators at about 4 °C for 4 and 48 h was conducted by Leuesque et al. (2006). Average THM reduction of 17% by the storage of water in a covered pitcher was observed in this study. In addition, a reduction of 61% for THMs was found in the uncovered container after 48 h from the moment the water was stored in the refrigerator. As for HAAs which are not volatile compounds, average HAA concentrations remained practically the same during the time of storage regardless of the presence of a lid, suggesting that there was no further formation of HAAs even if the water contained appreciable concentrations of residual chlorine and precursors at the time of storage in refrigeration (Leuesque et al. 2006).
However, an increase in THMs between 4 and 48 h storage of tap water was observed when the storage treatment was performed in a closed pitcher in the refrigerator at about 4 °C, which might be attributed to the extended reactions between disinfectant and THM precursors (Chowdhury et al. 2010). Consistent with the previously obtained results, reduction of THMs continued from 27% to 61% after 4 to 48 h of storage in the case of open pitcher, and no significant changes in HAAs were reported during storage in the refrigerator regardless of the seal condition (Chowdhury et al. 2010). Similarly, Rahman et al. (2011) proposed that refrigerating tap water at 4 °C in closed containers with about 5% air space for 5 h resulted in no or little change in DBP concentrations, although all the levels were decreased.
Besides previously noted results, Leuesque et al. (2006) also assessed the effect of combined household treatment technologies on DBP variations in tap water, including boiling water followed by storage as well as filtering water followed by storage (in the refrigerator at 4 °C). Results showed that the reduction of THMs by boiling and subsequently storing water was considerable, with an average total removal rate of 83 and 90% observed for further storage in a covered and uncovered pitcher, respectively, and a subsequent THM reduction during storage emerged in the case of the uncovered pitcher (85% reduction after 48 h storage). Filtration of tap water followed by storage resulted in slightly higher reductions of THMs compared to filtration alone, and the average total removal rate of 92 and 91% was found for the employment of a new filter and a used filter, respectively, with a significant additional reduction associated with water storage after filtering (34–39%). Moreover, average HAA reduction was about 72 and 60% in the case of new and used filters followed by storage, respectively (Leuesque et al. 2006).
The result that further variations of THMs during storage after boiling depended on the seal condition of containers was also reported by Chowdhury et al. (2010). In this study, boiling followed by storage decreased THMs by 82 and 97% in the case of storage for 48 h in a closed pitcher and an open pitcher, respectively, while HAAs showed varying results under the same conditions. With regard to the combination of filtration and storage, THMs were reduced by 90 and 93% when tap waters were filtered with new and used filters followed by storage for 48 h in a refrigerator, respectively. In addition, reductions of HAAs were observed using new and used filters followed by storage under identical conditions (71 and 59%, respectively) (Chowdhury et al. 2010).
Rahman et al. (2011) treated tap water with an instant boiling water unit followed by a jug fitted with ion exchange and AC filtration, and found that there were relatively large reductions for total THMs (88%) and HAAs (81%) in instant boiled-filtered water. A facility consisting of a polypropylene cotton filter, two GAC filter cartridges, and a boiler was used to determine DBP removal by PoU facility (Wang et al. 2019). Based on the data obtained in the one-year-long survey, the new PoU facility removed 40–60% of THMs and its efficiency gradually decreased with time, and the treated water occasionally contained higher THMs than untreated water in the summer. Furthermore, this study demonstrated that HALs and residual chlorine were completely removed regardless of old or new PoU facilities, and the removal efficiency of HAAs (84% in summer and 58% in autumn) was higher than that of THMs (Wang et al. 2019).
SUMMARY AND INSIGHT
Table 7 and Figure 3 summarize the performance of HWTS and combination technologies in controlling DBPs in municipal tap water. Generally, PoU filters could substantially reduce regulated DBP levels, while diverse DBP classes have different fates. Compared with HAAs, the filters which contain AC could perform higher THM reduction, which could be explained by the fact that THMs possess higher log Kow values (see Table 3) and exhibit lower aqueous solubility and polarity, making them more amenable to carbon adsorption. Similarly, filters containing AC showed poorer average removal efficiency of TCM, DCAA, and TCAA than their bromine-containing counterparts, since the latter have a lower affinity to water and tend to be more readily adsorbed on the carbon particles. As for filters in which ion exchange happens, DBPs existing in an anionic state could be easily eliminated in these filters. Furthermore, most filters showed higher removal rate of AOBr than AOCl and AOI. Different from PoU filters, boiling treatment might increase the concentration of HAA species, especially MCAA and DCAA in chlorinated tap water, leading to an unpredictable change of total HAA levels. However, it has been reported that boiling could eliminate a series of DBPs through volatilization, hydrolysis, or destruction, including THMs, HALs, HANs, and HNMs, while the removal rate of TOX by boiling was lower than the case of PoU filtration. Storage did contribute to DBP removal to some extent, while the effect was not clear when the target DBPs were not volatile (see Table 3, DBPs with Henry's law constant at 25 °C <10−7 atm-m3/mol are considered not volatile) or the storage was performed in a closed system. Also, combination technologies including filtration followed by storage, boiling followed by storage, as well as boiling coupled with filtration all showed unnegligible removal effect on DBPs in municipal tap water.
Although domestic water purifiers fitted with UV lamp are rarely studied for their effect on DBP control, the processes and rate constants of UV direct photolysis for DBPs have been explored. As for household chemical disinfection, there is no doubt that chlorination as well as flocculation followed by chlorination would cause DBP formation, and DBP yields might depend on characteristics of water to be treated as well as the dosage of disinfectant. Smith et al. (2010) compared DBP formation from iodine-based disinfectants at their recommended dosages for point of use to chlorination, and reported that iodoform with high toxicity was the predominant THM formed during iodination and total organic iodine concentrations during iodination with lower oxidant dosage exceeded TOCl concentrations during chlorination on a molar basis.
As mentioned above, HWTS has a significant impact on DBP levels before drinking water ingestion, which could have important implications for DBP exposure assessment and public health protection. In terms of DBP exposure assessment, on the one hand, studies conducted for occurrence survey or regulatory purposes mainly obtained monitoring data of DBPs in DWTP effluent, DWDS or tap water, which may not adequately represent DBP exposure and risks to human health since the population is not necessarily exposed directly to these waters (Dion-Fortier et al. 2009; Liu & Reckhow 2013; Chowdhury 2016). On the other hand, in epidemiological studies, researchers mostly used average DBP concentrations in water systems to estimate individual-level exposures, or used tap water consumption as a surrogate for DBP exposures to understand the dose–response relationships, which may lead to inaccuracy and misclassification in exposure assessment (Waller et al. 1998; Weinberg et al. 2006; Wright et al. 2006; Hamidin et al. 2008; Evans et al. 2013). Owing to the important role played by HWTS in DBP control, the variations of DBPs during the residence time of drinking water in households should be taken into account when studies of exposure characterization are conducted for DBPs in drinking water.
Also, HWTS is related to drinking water quality improvement, and consequently, human health protection. For emergencies in which tap water contains DBPs with a high level, HWTS seems to be an interim and useful solution to DBP elimination. Household treatment including filtration (especially in the case of AC filter), boiling water, storage in an open system as well as relevant combination technologies could help vulnerable populations who are exposed to THMs with high level. As for involatile HAAs, indoor treatment that favors DBP volatilization, including boiling and storage, are not recommended for HAA removal, while the filters could contribute to HAA control, and notably the filters contain ion exchange resins. From a perspective of TOX which is a surrogate for all halogenated DBPs and has been proven to be associated with toxicity (Yang et al. 2014), both filtration and boiling could reduce TOX levels in drinking water. However, what should be noted is that the efforts to reduce DBP concentrations must not compromise the effectiveness of pathogen removal by HWTS. For instance, storage in an open system could favor removal of some volatile DBPs, while the treated water may be susceptible to recontamination under such conditions. Microbial contamination in filters could occur when regular cleaning of filters and frequent replacement of spare parts are not carried out. On the other hand, it is well known that membrane filtration could remove some minerals and nutrients, resulting in another trade-off associated with pollutants and nutrients in drinking water.
In this study, the research in the field of HWTS in DBP control have been reviewed, and the involved technologies include filtration, boiling, photolytic method, storage in refrigerators, as well as combination technologies. Generally, HWTS could remove DBP to some extent, and the performances of different methods are dependent on the removal mechanisms performed by HWTS, the physicochemical properties of target DBPs as well as the characteristics of water to be treated. However, DBP concentration might show an increased trend, which could be attributable to accelerated DBP formation at high temperature as well as DBP elution from filter cartridges. Understanding the link between HWTS and DBPs will help improve the assessment accuracy of human exposure to DBPs and provide some interim advice for humans exposed to DBPs with a high level. Similarly, a trade-off between microbial and chemical risks also exists during household treatment, but pathogen removal should not be compromised in attempting to control DBPs as well. Future research relevant to DBP control by means of HWTS should take into account more DBP species possessing higher toxicity and other types of domestic devices used for drinking water purification.
The authors gratefully acknowledge the National Natural Science Foundation of China (No. 51822808; 51778445), the National Major Science and Technology Project of China (No. 2017ZX07201005), Shanghai City Youth Top Talent Project, Tongji University Youth 100 program and Social Development projects of Shanghai Science and Technology Commission (No. 19DZ1204400).
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.