N-nitrosamines, emerging disinfection by-products of health concern: an overview of occurrence, mechanisms of formation, control and analysis in water

Chlorination is the primary method used to disinfect water for human consumption. Its use is largely justified by its effectiveness in inactivating pathogens, its persistence, its ease of use, and reasonable cost. However, chlorination leads to the formation of disinfection by-products (DBPs) by reaction with organic matter naturally present in water. The DBPs are the subject of increasing concern because of the potential human health risk (Krasner 2009; Nieuwenhuijsen et al. 2009; Hebert et al. 2010). Several epidemiological studies have found associations between consumption of chlorinated drinking water and an increased risk of certain health outcomes, particularly bladder cancer (Nieuwenhuijsen et al. 2009). To date, a wide variety of DBPs (more than 600) were detected in disinfected water resulting from the use of chlorine and other disinfectants (chlorine, chloramine, and ozone). Among them, N-nitrosamines (NAms) are one class of non-halogenated compounds which have been identified in different matrices. NAms are alkylating agents characterized by the presence of the N-nitroso group and may be aliphatic or ring structures. Different studies reported that these molecules are significantly more toxic than the regulated DBPs (Oya et al. 2008). NAms have recently become the subject of great concern because of suspected effects on human health, especially of carcinogenic and genotoxic order, by long-term consumption of chlorinated water, even at low concentration levels (Nawrocki & Andrzejewski 2011).

Historically, NAms have been detected in many food products (Ventanas et al. 2006), beverages (beer) (Jurado-Sánchez et al. 2007; Pérez et al. 2008), consumer products (cosmetics) (Ma et al. 2011), tobacco smoke (Lee et al. 2007c), and environmental samples such as water from chlorinated swimming pools, wastewater, treated wastewater, groundwater, and drinking water (Cheng et al. 2006; Walse & Mitch 2008; Jurado-Sánchez et al. 2010; Nawrocki & Andrzejewski 2011; Krasner et al. 2013; Zhou et al. 2014).

It is known that drinking water disinfection with free chlorine or monochloramine leads to the formation of NAms, including N-nitrosodimethylamine (NDMA). In 1989, NDMA was first detected as a disinfection by-product in the Province of Ontario (Canada), and in 1999, this compound was found in drinking waters throughout California (USA) (Charrois & Hrudey 2007a). The natural organic matter (NOM) present in water may contain different precursors that lead to the formation of various NAms. Over the last decade, interest has been growing in the formation of some N-nitrosamines in the process of water treatment. Nowadays, nine NAms have been detected in waters. Their chemical structures are shown in Figure 1.

NAms are polar compounds, usually water soluble with low octanol/water (K_o/w) partition coefficients, therefore are difficult to extract with organic solvents (Table 1). Presently, there are no standard analytical methods for the determination of nine NAms in drinking water at the low nanogram per liters levels, which causes serious analytical problems. Future regulation will increase interest to develop reliable and sensitive analytical methods for the identification and quantification of these compounds.

This paper gives an overview of the current knowledge on the occurrence, precursors, and formation mechanisms of the N-nitrosamines in water. The existing regulations established by different organisations around the world are discussed. The sample preparation and the analysis of these compounds in water samples are also described.

One of the first occurrence studies was focused on NDMA detection in US and Canadian drinking water treatment plants (DWTPs) (Taguchi et al. 1994). Significant concentrations of NDMA were detected in both finished water from DWTPs (up to 100 ng L⁻¹) and distributed water (66 ng L⁻¹), which can be mainly attributed to the use of chlorine and chloramines for disinfection (Charrois et al. 2007b). Moreover, a nationwide survey of NDMA in raw water for DWTPs in Japan indicated that NDMA was detected in 15 of 31 raw water samples collected in the summer at levels up to 2.6 ng L⁻¹, and in nine of 28 raw water samples collected in winter at levels up to 4.3 ng L⁻¹ (Asami et al. 2009).

Other studies confirmed the presence of NDMA, NDEA, NMOR, NPYR, NPIP, and NDPHA in drinking water (Zhao et al. 2006; Goslan et al. 2009; Richardson 2009) and of NDMA, NMOR in surface water (Zhao et al. 2008; Kosaka et al. 2010). Furthermore, seven NAms, NDMA, NDEA, NMOR, NDPA, NPYR, NPIP, and NDBA, have been reported in wastewater with concentrations in the range 5–25 ng L⁻¹ (Sedlak et al. 2005; Krauss et al. 2009; Padhye et al. 2009).

More recently, N-nitrosamines data reported from the first rounds of samples collected under the second Unregulated Contaminants Monitoring Rule (UCMR2) and the Ontario Drinking Water Surveillance Program were reviewed to assess the frequency, magnitude of occurrence, and the effect of disinfectant type and other treatment factors on reported N-nitrosamine concentrations. Monitoring data reveal that NDMA was detected in drinking water at concentrations higher than the UCMR2 minimum reporting level (MRL) of 2 ng L⁻¹ in one of every 10 samples. Other N-nitrosamines such as NDEA, NDBA, NPYR and NMEA were rarely detected at levels above their MRLs (Russell et al. 2012). Other N-nitrosamines occurrence surveys conducted in the UK, Australia, and China from 2008–2012 indicated that NDMA was the N-nitrosamine most frequently detected (Zhao et al. 2008; Templeton & Chen 2010, Ma et al. 2012; Kristiana et al. 2013).

In recent years, the formation of NAms in drinking water has attracted a significant scientific and regulatory attention, especially since some of these molecules have been classified as probable human carcinogens (Walse & Mitch 2008). The chemical pathways of NAms formation in water are complex and uncertain. Many reactions can occur simultaneously; the products depend on the reaction conditions, including the concentrations of reactants, catalysts, and inhibitors, the occurrence of other reactions, and other competing reactions. NAms formation is a complicated process that also is influenced by many operating conditions including pH, content, and concentration of NOM in source water, as well as the type and concentration of the disinfectants used, and disinfectant contact time. Several studies have investigated the effects of different disinfection treatments on NAms formation. It was found that the water disinfection processes with chlorine, monochloramine, chlorine dioxide, and ozone generally lead to the formation of NAms in treated waters (Schreiber & Mitch 2005; Andrzejewski et al. 2008; Changha et al. 2007). Shah & Mitch (2012) published a critical article focused on the formation pathways of N-DBPs including N-nitrosamines associated with the type disinfection treatment.

The determination of a formation mechanism with a simple compound related to the N-nitrosamines structure seems obvious and more probable. That is why DMA is often used as a model precursor. Choi et al. (2002) has shown that NDMA formation was initially attributed to the reaction of DMA with monochloramine, followed by the oxidation of unsymmetrical dimethyl hydrazine (UDMH) as an intermediate, which is further oxidized to NDMA. This pathway established a reversible transfer mechanism between free chlorine and ammoniac, monochloramine, and DMA-Cl (Figure 2.I) (Choi et al. 2002).

Schreiber & Mitch (2006a) showed that dichloramine concentration and dissolved oxygen enhance the formation of NAms. Dichloramine reacts with DMA to form a chlorinated UDMH (UDMH-Cl). In the presence of dissolved oxygen, this product may cause the formation of the NAms (Figure 2.II). Selbes et al. (2013) investigated the yield of NDMA produced by different tertiary aliphatic and aromatic amines to establish the effect of functional groups attached to the basic DMA structure.

According to Keefer & Roller (1973), Choi & Valentine (2003) the formation of NDMA can also occur in the absence of ammoniac by the nitrosation pathway. The reaction of HOCl with nitrite forms as intermediates a dinitrogen tetroxide (N₂O₄), which is favored at neutral pH or a dinitrogen trioxide (N₂O₃) at low pH. The formation of NDMA is made possible by the reaction of these intermediates on DMA as represented in Figure 2.III.

Maximum NDMA formation rates are observed between pH 6 and pH 8 (Mitch & Sedlak 2002). Le Roux et al. (2012) suggested that bromine-containing oxidant species can enhance the formation of NDMA from some tertiary amines or DMA. UMDH-Cl favors the incorporation of dissolved O₂ over UDMH due to the weakness of the N − Cl bond.

NDMA precursors are distinct from the precursors of the regulated chlorinated DBPs, such as trihalomethanes and haloacetic acids. Currently, the precursors involved in the formation of the other N-nitrosamines are not well defined, although a possible list of NDMA precursors has been reported (Nawrocki & Andrzejewski 2011). Humic and fulvic acids are responsible for the formation of regulated DBPs (Krasner et al. 2009a, b). However, these compounds are not substantial precursors of NDMA (Mitch & Sedlak 2004). It is obvious that dimethylamine (DMA) is the most natural and most likely precursor of NDMA (Nawrocki & Andrzejewski 2011). Other compounds, such as tertiary and quaternary amines (Lee et al. 2007b; Kemper et al. 2010), undefined organic nitrogen as a part of NOM (Gerecke & Sedlak 2003; Chen & Valentine 2006), cationic flocculants (Krasner et al. 2013; Wilczak et al. 2003), and anion exchange resins that are used for ion-exchange in water treatment are also identified as NDMA precursors (Kimoto et al. 1980).

Lee et al. (2007a) and Lee et al. (2008) noted the formation of NDMA as the result of oxidation of several NDMA precursors (such as organic compounds with dimethylamine or trimethylamine groups) with high doses of chemical oxidants (ozone, chloramines, and chlorine dioxide) (Lee et al. 2007a; Oya et al. 2008; Planas et al. 2008; Schmidt & Brauch 2008). Some researchers have shown that oxidation of DMA may lead to the formation of NDMA, particularly for a pH value ranging from 7 to 8 (Andrzejewski et al. 2005; Andrzejewski & Nawrocki 2007, 2009).

Recent studies, have shown that some secondary amines, including DMA and diphenylamine (DPHA), are also important precursors for NDMA and NDPHA (Zhou et al. 2009). In addition to DMA and DPHA, there are other secondary amines, such as diethylamine (DEA), methylethylamine (MEA), di-n-propylamine (DPA), di-n-butylamine (DBA), morpholine (MOR), pyrrolidine (PYR), and piperidine (PIP), that are considered as principal precursors corresponding to the seven NAms (i.e. NDEA, NMEA, NDPA, NDBA, NMOR, NPYR, and NPIP, respectively) (Table 2). Wang et al. (2011) investigated the occurrence of the nine secondary amines in the 12 source water and finished water samples. DMA and DEA were detected in most of the samples, with concentrations in source water samples ranging from 0.2 to 3.9 μg L⁻¹ and 0.3 to 2.4 μg L⁻¹, respectively. The other amines (such as MOR, PYR, PIP, DBA and DPHA) existed at lower concentrations and detection frequency. DMA and DEA were also detected in most of the finished water samples, with concentrations ranging from 0.4 to 4.0 μg L⁻¹ and 0.1 to 1.8 μg L⁻¹, respectively (Table 2). Moreover, the detectable rates of the nine NAms in finished water samples were correlated positively with the detectable rates of the corresponding secondary amines in source water samples (r² = 0.87).

Shen & Andrews (2011) demonstrated that several tertiary amines including pharmaceutical ranitidines are nitrosamine precursors during chloramines disinfection. Le Roux et al. (2011) demonstrated that NDMA formation from chloramination of ranitidine exhibited a maximum yield of 59.6% at pH 7.9, which is similar to 62.9% reported by Schmidt & Brauch (2008) for the same conditions.

Since studies have confirmed that DMA is not the only NDMA precursor, researchers have turned their focus to dissolved organic nitrogen (DON). This parameter is contained in NOM and was found to constitute 10% of the total dissolved nitrogen in raw waters (Lee et al. 2007d). Dotson et al. (2009) investigated the positive correlation between DON concentrations and NDMA formation in natural waters. They found that waters with higher DON concentrations could lead to the formation of NDMA and other NAms via nitrosation of available amines (Richardson 2003). Mitch & Sedlak (2004), Schreiber & Mitch (2006a, b), also demonstrated the implication of DON to form NDMA in wastewater effluents. Other authors indicated that a subset of DON constituents in treated wastewater effluent organic matter (EfOM) is likely to account for the majority of NDMA formation in source waters (Shah & Mitch 2012; Krasner et al. 2013).

Researchers have also investigated the relationship between DON and common water quality parameters such as dissolved organic carbon (DOC), UV absorbance at a wavelength of 254 nm (UV254), and SUVA as well as the NDMA formation potential. Xu et al. (2011) studied the characteristics of DON in raw water from the Huangpu River, including the molecular weight (MW) distribution of dissolved organic matter distribution and NDMA formation potential. The results from linear regression analysis showed that DON is moderately correlated to DOC, UV254, specific ultraviolet absorbance (SUVA) (UV254/DOC), and NDMA formation potential. Chen & Valentine (2007) found a relationship between SUVA (272 nm) and NDMA formation of Iowa River water. They also found a linear relationship between oxidized NOM and NDMA formation potential (Chen & Valentine 2006). Zhao et al. (2008) showed no relationship between total organic carbon, absorbance, pH, and NDMA formation in surface waters. In addition, Chen & Valentine (2006) found that the amount of NDMA can be correlated linearly with a decrease of SUVA value at 274 nm.

However, little is known about the role of NOM and the effects of its nature and reactivity on the formation of NAms. Several studies have investigated the reactivity of hydrophobic and hydrophilic fractions of NOM for the formation of these molecules. Chen & Valentine (2007) reported the formation of NDMA from the chloramination of different polarity-based fractions of NOM (hydrophobic acid, hydrophobic base, hydrophobic neutral, hydrophilic acid, hydrophilic base, and hydrophilic neutral). They found that the chloramination of DMA at 0.5 μg L⁻¹ in river water accounts for 15% of NDMA formation. Moreover, the hydrophilic acid fraction of NOM tended to form more NDMA than its hydrophobic fraction, and basic fractions of NOM tended to form more NDMA than the acidic fractions. In addition, Gerecke & Sedlak (2003) confirmed again the role of hydrophilic fractions and basic fractions in the formation of NDMA.

Gerecke & Sedlak (2003), Chen & Valentine (2006; 2007) have indicated that during chloramination, some NOM fractions were found to produce significant amounts of NDMA in water. Kristiana et al. (2013) investigated the influence of NOM MW characteristics on the formation of eight compounds of NAms following chlorination or chloramination from diverse source waters. NDMA was the most frequently detected species, and was measured at the highest concentrations ranging from 0.02 to 0.08 nM and from 0.03 to 0.18 nM during chlorination and chloramination, respectively. Regardless of the MW characteristics, chloramination demonstrated a higher potential to form NAms than chlorination. Moreover, in chloramination, low to medium MW fractions of NOM were identified as containing significant amounts of NAms precursors, where fractions with apparent MW <2.5 kDa formed the highest concentrations of NAms.

Many cationic flocculants (e.g. polyamines, poly-diallyldimethyl ammonium chloride (poly-DADMAC) used in drinking water treatments have also been found to be a significant precursor of NDMA in drinking water. Park et al. (2009) and Wilczak et al. (2003) reported that the enhancement of the poly-DADMAC concentration increased the NDMA formation after the chloramination processes. Other studies showed that polyamines produced more NDMA than poly-DADMAC, and the highest levels of NDMA were found around pH 8 during chloramination (Park et al. 2009). Freshly used, or just regenerated, resin ion exchangers can pollute water with NDMA (Kemper et al. 2008). Anion exchange resins employed in drinking water treatment are strong-base quaternary amine ion- exchangers. Resins were found to release NDMA precursors, which can react with chlorine or chloramine to increase nitrosamine formation (Nawrocki & Andrzejewski 2011).

The formation of NDMA in drinking water treatment does not depend only on the NDMA precursors. It can be affected by many water quality parameters and operating conditions. The influence of several parameters such as pH, concentration of dissolved oxygen, reaction time, and disinfectant dose has been investigated in previous studies. According to Mitch & Sedlak (2002), the rate of NDMA formation varies with pH, with a maximum NDMA formation rate between pH 6 and pH 8. Le Roux et al. (2011) demonstrated that the pH and dissolved oxygen content of the solution were found to play a major role for the formation of NDMA from ranitidine. NDMA was formed in higher amounts at pH 8 and a lower concentration of dissolved oxygen dramatically decreased NDMA yields. NDMA formation was significantly inhibited for low oxygen concentration (0.2 mg O₂/L, molar yields of 4.01%) compared to ambient oxygen concentration (9 mg O₂/L, molar yields of 54%). Charrois et al. (2007b) found that NDMA concentrations in the distribution system vary and tend to increase with increasing distribution residence time. Moreover, bench-scale experiments with free-chlorine contact (2 h) before chloramination resulted in significant reductions in NDMA formation (up to 93%) compared to no free-chlorine contact time. Krasner (2009) have found that a lower Cl₂:N (<10) ratio generates more NDMA than chlorine for a ratio >10.

Water treatment processes can directly or indirectly affect the formation of NAms. Their impact on the formation of these molecules is complex and variable. Zhao et al. (2008) investigated the formation of nine NAms following eleven different disinfection treatments in water from seven locations in North America. NDMA, NMEA, NMOR, and NDPHA were detected above the limit of detection in some of the disinfected samples of the source waters from six locations. The NDMA was the most frequently detected nitrosamine with the highest concentrations. NDPHA, NMEA, and NMOR were also identified in some of the disinfected water samples. NDPHA was formed after disinfection with hypochlorite ion (OCl⁻), monochloramine (NH₂Cl), and ozone (O₃). NMEA was produced with OCl⁻, while NMOR formation was associated with O₃. Additionally, NDMA formation was more than one order of magnitude higher when NH₂Cl was applied compared to OCl⁻. Recently, Luo et al. (2012) investigated the occurrence of nine NAms by following different water treatment processes (Table 3). It was conducted using samples from seven drinking water treatment plants from three cities and tap waters from one city in China. The species and concentrations of nine NAms were simultaneously analyzed in waters and the performance of unit processes (coagulation, filtration and/or biological activated carbon adsorption) in DWTPs adopting chloramine, ozone, chlorine, and ultraviolet as disinfection process was investigated. Obtained results showed that, the species and concentrations of the NAms varied with the used disinfection method and with the source waters. NDMA, which is the N-nitrosamine of greatest concern, was identified in raw water, disinfecting water, finished water, and tap water samples, ranging from 0.8 to 21.6, 0.12 to 24.2, no detection to 8.8, and no detection to 13.3 ng L⁻¹, respectively (Table 3). Chloramination alone produced the most significant amounts of NDMA, while ozonation followed by chloramination led to moderately reduced levels. In addition, chlorination produced relatively less NDMA than chloramination.

Regulatory actions have been directed almost exclusively toward trihalomethanes and haloacetic acids due to the availability of analytical techniques to detect these molecules as well as their relatively high abundance in drinking water. Currently, there are no drinking water guidelines for nine NAms.

Regulatory guideline or assessment levels for environmental media and in particular for drinking water have been set by a number of international, national, and subnational authorities (Mhlongo et al. 2009).

Concerning the health risks, NAms are classified by a number of international organizations and regulatory authorities in function of their carcinogenicity. The degree of carcinogenicity among these compounds varies dramatically. NDEA is the most potent carcinogen among the NAms. The United States Environmental Protection Agency (US EPA) has classified these molecules into the group indicating probable carcinogenicity to humans (Table 1). In addition to NDMA, the US EPA has listed NAms, including NDEA, NMEA, NDPA, NDBA, and NPIP, in the Unregulated Contaminant Monitoring Rule 2 (UCMR 2) to be monitored from 2008 to 2010 (US EPA 2010). The European Union (EU) categorizes NDMA and NDEA as category 1B (presumed to have carcinogenic potential for humans; largely based on animal evidence). The Ontario Ministry of the Environment (OME) (2003) and the California Department of Public Health (CDPH) (2007) have recommended a maximum contaminant level MCL in drinking water for NDMA, NDPA, and NDEA. The World Health Organization (WHO) (2011), Health Canada (HC) (2010), and Australian National Health and Medical Research Council (ANHMRC) have also established a guideline value for NDMA in drinking water. In the EU, NAms are not specifically listed in the Drinking Water Directive (Council Directive 98/93/EC), but a few EU member states have regulated their presence in drinking water, as higher or more stringent standards are allowed under the directive. Moreover, provisional standard values were proposed in the Netherlands and in Germany for NDMA and NMOR (Planas et al. 2008; Reyes-Contreras et al. 2012). All of these regulations guidelines and the associated MCL values are summarized in Table 4. A recent work indicated that the US EPA will make a preliminary regulatory determination for NAms in 2013 (Krasner et al. 2013).

N-nitrosamines are typically detected in drinking water at low concentration levels. Consequently, the analytical methods to determine NAms in waters must be highly sensitive to be able to detect a few ng L⁻¹. In the literature, the most commonly used methods for sample preparation are liquid–liquid extraction (LLE) (Taguchi et al. 1994), solid-phase micro-extraction (SPME) (Grebel et al. 2006; Ventanas et al. 2006; Hung et al. 2010), and solid-phase extraction (SPE) (Zhao et al. 2006; Lee et al. 2007c; Boyd et al. 2011; McDonald et al. 2012; Wang et al. 2012). The micro-liquid–liquid extraction (Cheng et al. 2006) and pressurized liquid extraction (PLE) were also used for the determination of NAms (Ramírez et al. 2012). However, few LLE methods have been published in the past few years. Recently, SPE method with different stationary phases or adsorbents (i.e. Strata X, Coconut Charcoal, and Oasis HLB) has often been used for the preconcentration of NAms in water samples (Charrois et al. 2004; Munch & Bassett 2006; Jurado-Sánchez et al. 2009). Although automation is possible, the SPE method has been commonly used for the enrichment of these molecules because of its low cost, short processing time, and ease of use (Jurado-Sánchez et al. 2007; Planas et al. 2008).

Several selective analytical techniques have been developed for the quantification of NAms in waters. The analytical methods currently used are based on two steps, that is, analysis by gas chromatography (GC) or liquid chromatography (LC) and extraction/concentration steps. The NAms have been analyzed in water samples by using GC coupled to different types of detectors such as: nitrogen–phosphorus detection (Grebel et al. 2006; Jurado-Sánchez et al. 2010), thermal energy analyser (GC/TEA) (Byun et al. 2004; Drabik-Markiewicz et al. 2010), nitrogen chemiluminescence detection (Grebel & Suffet 2007; Ramírez et al. 2012), mass spectrometry (MS) (Ventanas & Ruiz 2006; Pozzi et al. 2011; Reyes-Contreras et al. 2012; Huang et al. 2013), and tandem mass spectrometry (MS/MS) (Llop et al. 2012; McDonald et al. 2012). The US EPA has developed an analytical method covering seven NAms (Table 5). However, some of NAms may be undetectable by GC/MS/MS, because they are thermally unstable or non-volatile (Munch & Bassett 2004). For example, NDPHA is contained in the US EPA Method 521 standard mix but is not included in the GC/MS/MS method, because it is thermally unstable and decomposes in the GC injector.

Moreover, the analysis of NAms has been performed by LC using a fluorescence detector (Cha et al. 2006) and high pressure liquid chromatography-tandem mass spectrometry (HPLC/MS/MS) (Zhao et al. 2006; Plumlee et al. 2008; Krauss et al. 2009). Ultra high pressure liquid chromatography-tandem mass spectrometry (UHPLC/MS/MS) was used for the analysis of NDMA in water samples from drinking water treatment plants (Asami et al. 2009; Wang et al. 2010). Kulshrestha et al. (2010) described a new strategy to measure the molar concentration of total N-nitrosamines (TONO) in disinfected pools adapted of the widely used assay for nitrite determination.

Ripollés et al. (2011) has developed a novel UHPLC/MS/MS method for detection and quantification of NDMA, NMOR, NMEA, NDEA, NPIP, NPYR, NDPA, and NDBA in drinking water, using an atmospheric pressure chemical ionization. The limits of detection were found to be in the range of 1–8 ng L⁻¹. Recently, a fast and sensitive method using UHPLC/MS/MS and electrospray ionization mode was applied for the analysis of NDMA, NMOR, NMEA, NDEA, NPIP, NPYR, NDPA, NDBA, and NDPHA in chlorinated and chloraminated drinking water (Wang et al. 2011; Kadmi et al. 2014). The obtained detection limits varied from 0.2 to 0.9 ng L⁻¹ and 0.1 to 0.7 ng L⁻¹ for source water and finished water samples, respectively. In addition, Lee et al. (2013) recently described a new analytical HPLC-PCUV procedure for the N-nitrosamines monitoring based on the Griess colorimetric determination of nitrite generated by UV-254 nm photolysis of nitro(so) compounds after separation by HPLC (HPLC-Post Column UV photolysis/Griess reaction (HPLC-PCUV)).

 Table 5 gives a detailed presentation of some of the analytical methods developed for the analysis of N-nitrosamines as well as the extraction recoveries and detection limits.

N-nitrosamines belong to a group of non-halogenated molecules recently identified as DBPs in chlorinated waters. Other disinfection water processes with monochloramine, chlorine dioxide, and ozone can lead to their formation. This class of compounds is considered as extremely toxic, mostly carcinogenic emerging organic pollutants, and may pose some threat to drinking water consumers. Currently, nine NAms have been identified and their occurence in drinking water may be caused by raw water pollution from nitrosamines or by their formation during water treatment processes. The formation mechanisms of NAms are complex and still not well understood; current results are limited. Different mechanisms have been proposed in the literature, especially for NDMA formation. The formation of these molecules during water treatment requires the presence of N-containing precursors; most of them are still unknown. Moreover, to the best of our knowledge, there are not many studies on the occurrence of these molecules in European drinking water, surface water or ground water. This research field has been only partially explored.

Several analytical methods have been developed and reported, but the accuracy of these methods is mainly tested for drinking water; less information is present for other matrices. Some of the proposed methods are labor-intensive and use large amounts of organic solvents while achieving low recovery. In the case of NAms, several factors complicate their analysis, such as detecting of low concentrations and detecting both stable and labile NAms. More work is required to develop cost-effective, fast, and accurate methods for monitoring the NAms in water. However, further studies are needed in order to understand their formation mechanisms and to identify and characterize the precursors responsible for their formation in treated waters.