Occurrence and spatio-temporal variability of halogenated acetaldehydes in full-scale drinking water systems

Highlights

• Halogenated acetaldehydes (HALs) were monitored in full-scale drinking water systems.

• Dibromoacetaldehyde was detected after ozonation but before post-chlorination.

• Water temperature and ozone dose were highly correlated to the formation of HALs.

• THM levels were more strongly correlated to the levels of tri-HALs than di-HALs.

Abstract

As the third largest group of identified disinfection by-products (DBPs) by weight, halogenated acetaldehydes (HALs), were monitored for one year at numerous locations in two full-scale drinking water systems applying an ozone-chlorine sequential disinfection strategy. The HALs that were targeted included four trihalogenated acetaldehydes (THALs): chloral hydrate (CH), bromodichloroacetaldehyde (BDCAL), dibromochloroacetaldehyde (DBCAL) and tribromoacetaldehyde (TBAL). Three dihalogenated acetaldehydes (DHALs) were also included: dichloroacetaldehyde (DCAL), bromochloroacetaldehyde (BCAL) and dibromoacetaldehyde(DBAL). In addition to various sampling points in two distribution networks, this study also investigated the formation of HALs during water treatment and for the first time, reports the formation of DBAL before chlorine is applied. Low bromide levels in source waters from both systems resulted in the rare detection of DBAL and TBAL. CH accounted for >50% of total HALs (HAL7) with DHALs accounting for as little as 10% of HAL7, presumably due to the use of ozone-chlorine instead of ozone-chloramine. In the presence of chlorine residuals and with increasing water residence times, most HALs continued to form, more readily in warm water than in cold water. However, the spatial and temporal patterns for each HAL differed depending on speciation (THAL vs. DHAL) and water temperature. Compared to the relatively stable bromine incorporation factor (BIF) of THMs in the distribution systems, the decreasing BIFs of HALs according to water residence time increases suggested that bromine-containing THMs are more stable than their corresponding HALs. Re-chlorination at the extremities of the distribution networks demonstrated a significant impact on the occurrence and speciation of DBPs. In both full-scale systems, water temperature was shown to be the biggest contributing factor to HAL formation. The strong correlations between THM levels and THAL levels make it possible to predict the occurrence of THALs based on THMs.

Introduction

Due to the regulations for trihalomethanes (THMs) and haloacetic acids (HAAs), many drinking water utilities have switched their disinfection processes from chlorination to using alternative disinfectants such as ozone, chloramines, and chlorine dioxide (Krasner et al., 2006; Richardson et al., 2007). However, other problems have emerged with the use of these chemicals. For example, ozonation may substantially minimize the formation of THMs and HAAs, but it may lead to the formation of bromate, which is problematic (Moore and Chen, 2006; Umemura and Kurokawa, 2006). Similarly, the formation of nitrosamines, chlorite and chlorate were reported with the use of chloramines or chlorine dioxide (Choi and Valentine, 2002; Padhi et al., 2019). Along with bromate, these latter disinfection by-products (DBPs) have been regulated due to their toxicity (Health Canada, 2014; Liteplo and Meek, 2001; NTP, 2005; Ueno et al., 2000; USEPA, 2010). To date, >600 DBPs have been identified in drinking water that has been disinfected using chlorine or alternative disinfectants (Richardson, 2011). Compared to regulated DBPs, the occurrence, formation and toxicology of most unregulated DBPs, such as halogenated acetaldehydes (HALs), have not been as thoroughly assessed through research (Richardson et al., 2007).

In a U.S. nationwide DBP occurrence study, HALs were reported to be the third most abundant DBP class (by weight) in drinking water, just after THMs and HAAs, which are regulated (Krasner et al., 2006; Weinberg et al., 2002). In early Canadian nationwide DBP occurrence surveys, the fully chlorinated form of HAL, chloral hydrate (CH, the hydrated form of trichloroacetaldehyde, TCAL), was found to be the third most prominent chlorinated DBP (Koudjonou et al., 2008). Recently, toxicologists have evaluated the toxicity of HALs by comparing their overall cytotoxicity and genotoxicity with five other DBP classes that include THMs and HAAs. These evaluations ranked HALs as the second most cytotoxic DBP class (Jeong et al., 2015; Postigo et al., 2015). Also, Chuang and Mitch (2017) have identified HALs as the drivers of overall toxicity in chloramine-disinfected drinking water. Although there have been no reports of HALs forming directly as a result of ozonation, the relationship between ozonation and HAL formation has been widely discussed in the literature. Early mechanistic research has shown that in water that has been ozonated, the subsequent chlorination will result in higher CH formation, via the formation of acetaldehyde, which may react with chlorine to form CH (McKnight and Reckhow, 1992): this observation was confirmed in subsequent research on CH and other trihalogenated acetaldehydes (THALs) (Dabrowska and Nawrocki, 2009; Koudjonou et al., 2008; Krasner et al., 2012; LeBel et al., 1997; Mitch et al., 2009; Zeng et al., 2016). However, ozone-chloramine treatment was reported to decrease the formation of CH and increase the formation of dihalogenated acetaldehydes (DHALs), indicating that the source of precursors for DHALs may be different than the source of THAL precursors (Jacangelo et al., 1989; Krasner et al., 2006; Mitch et al., 2009; Weinberg et al., 2002). Overall, HALs are formed from disinfection using chlorine or chloramine, but higher levels of HALs may be expected when combining ozone-chlorine/chloramine for disinfection (Richardson et al., 2000).

The current literature on DBP occurrence shows that in most studies, the only HAL that has been investigated is CH (Dabrowska and Nawrocki, 2009; Gan et al., 2013; Golfinopoulos and Nikolaou, 2005; Golfinopoulos et al., 2003; Jeong et al., 2012; Kawamoto and Makihata, 2004; Krasner et al., 1989; LeBel et al., 1997; Lee et al., 2001; McGuire et al., 2003; Nikolaou et al., 2004; Simpson and Hayes, 1998; Villanueva et al., 2012; Wei et al., 2010; Williams et al., 1995; Williams et al., 1997; Włodyka-Bergier et al., 2014). One study revealed that there were extremely high CH levels (i.e., 263 μg/L) in Canadian drinking water samples (Koudjonou et al., 2008). However, median values for CH levels in drinking water barely exceeded 10 μg/L in various DBP occurrence studies (Gan et al., 2013; Huang et al., 2017; Lee et al., 2001; McGuire et al., 2003; Wei et al., 2010). With the recent commercial availability of analytical standards of other HALs, the amount of literature on the occurrence of HALs other than CH is growing (Koudjonou et al., 2008; Koudjonou and LeBel, 2006). Generally, drinking water was sampled quarterly (4 times/one year) or at even lower-frequencies, which is not sufficient or statistically significant to evaluate the temporal (seasonal) variability of HALs (Aranda-Rodriguez et al., 2008; Charisiadis et al., 2015; Gan et al., 2013; Kawamoto and Makihata, 2004; Krasner et al., 1989; Lee et al., 2001; Simpson and Hayes, 1998; Weinberg et al., 2002; Williams et al., 1997). In addition, it is difficult to fully grasp the temporal variability of DBPs because it is dependent on the regional climate (e.g., temperature, precipitation) and the operational parameters used (e.g., disinfectant dose) (Delpla and Rodriguez, 2017; Delpla et al., 2015; Rodriguez and Sérodes, 2001). For example, the subtropical monsoon climate in Southern China (i.e., average water temperature of 29 °C in summer and up to 20 °C in winter) is not comparable to the northern temperate climate in Canada (i.e., 20 °C in summer and as low as 0 °C in winter) in terms of temporal variability (Gan et al., 2013; Mercier Shanks et al., 2013). Moreover, to our knowledge, there are no data available in the literature for the occurrence of HALs during drinking water treatment, and very few studies have investigated the effects of network re-chlorination on the occurrence and speciation of HALs (Legay et al., 2015). Because the analytical standards of most HALs have only recently been made commercially available, there is limited knowledge on the occurrence and particularly the spatio-temporal variability of most HALs (Koudjonou et al., 2008; Koudjonou and LeBel, 2006).

The objectives of this study were to (i) investigate the spatial and temporal variations of HALs in two Canadian full-scale drinking water systems that apply the ozone-chlorine sequential disinfection method in their drinking water treatment plants (DWTP) and apply re-chlorination in the distribution systems, (ii) identify factors (i.e., water quality, operational parameters) responsible for these variations, (iii) evaluate the relationship between the levels of regulated THMs and unregulated HALs in order to provide better estimations for HAL occurrence in future studies. This study is based on a robust dataset acquired from field sampling campaigns that involved numerous locations in the systems being studied.