Fate of bromine-containing disinfection by-products precursors during ozone and ultraviolet-based advanced oxidation processes

International Journal of Environmental Science and Technology https://doi.org/10.1007/s13762-018-1652-8 ORIGINAL PAPER Fate of bromine‑containing disinfection by‑products precursors during ozone and ultraviolet‑based advanced oxidation processes M. Petronijević1 · J. Agbaba1 · S. Ražić2 · J. Molnar Jazić1 · A. Tubić1 · M. Watson1 · B. Dalmacija1 Received: 17 May 2017 / Revised: 26 September 2017 / Accepted: 4 January 2018 © Islamic Azad University (IAU) 2018 Abstract This research evaluates the effect of ultraviolet photolysis, ozonation and ozonation/ultraviolet advanced oxidation processes on different disinfection by-product precursors, during the treatment of water with low organic matter and moderate bromide contents. After different combinations of ultraviolent fluence and ozone, the formation potentials of trihalomethanes and haloacetic acids were investigated. Bromine incorporation factors were used to give specific insight into the behaviour of brominated disinfection by-products, and inorganic bromate formation was also determined. The ozone/ultraviolet process was found to be more effective in reducing the total natural organic matter content than ozonation or ultraviolet photolysis alone. Ultraviolet photolysis was more successful removing the precursors of brominated trihalomethanes than chlorinated trihalomethanes, but slightly increased the precursors of both brominated and chlorinated haloacetic acids. During ozonation, reductions in the haloacetic acid formation potential were significantly better than those of the trihalomethanes formation potential (up to 54 and 27%, respectively). In the combined ozonation/ultraviolet process, increasing the ultraviolet fluence had a varying effect on trihalomethane and haloacetic acid precursor behaviour, depending on the ozone dose applied. Bromine incorporation after ozonation alone increased to up to 38% of the total bromide, largely as a result of bromate formation. The combined process curtailed all bromate formation, but increased the bromine incorporation up to 48% at higher ozone doses, with disinfection by-product formation shifting towards the more toxic brominated species. Keywords Advanced oxidation processes · Bromine incorporation · Disinfection by-product precursors · Natural organic matter · Ozonation · Ultraviolet photolysis Introduction The formation of disinfection by-products (DBPs) during water treatment presents a significant and complex problem for water utilities. More than 600 different DBPs have been identified, formed by reactions between the applied disinfectants and different water constituents such as inorganic compounds (bromide/iodide) and natural organic matter Editorial responsibility: M. Abbaspour * J. Agbaba jasmina.agbaba@dh.uns.ac.rs 1 Department of Chemistry, Biochemistry and Environmental Protection, Faculty of Sciences, University of Novi Sad, Trg Dositeja Obradovića 3, Novi Sad 21000, Republic of Serbia 2 Department of Analytical Chemistry, Faculty of Pharmacy, University of Belgrade, Vojvode Stepe 450, Belgrade 11221, Republic of Serbia (NOM) (Richardson et al. 2007; Mao et al. 2014). Many of the DBPs formed are classified as possible human carcinogens, including the trihalomethanes (THMs) and the haloacetic acids (HAAs), species which are both commonly detected in waters after chlorination. Their concentrations in drinking water are therefore regulated worldwide. In the Republic of Serbia (Official Gazette RS No. 42/98 and 44/99) and the European Community (Directive 98/83/EC), the sum of four THMs is limited to 100 µg/L, whereas the US Environmental Protection Agency (USEPA) regulates totals for both THMs and HAAs at 80 and 60 μg/L, respectively (USEPA 2006). The presence of bromide in waters has been shown to shift the distribution of DBPs to more brominated species, which exacerbates the threat posed by DBPs to public health, as in general, the brominated DBPs have been found to be more toxic than their chlorinated analogues (Mosteo et al. 2009). In order to better quantify the risk of brominated DBPs, Obolensky and Singer (2005) defined the bromine 13 Vol.:(0123456789) International Journal of Environmental Science and Technology incorporation factor (BIF) as the ratio of the molar concentration of bromine incorporated into a certain class of DBPs to the molar concentration of DBPs in that class. Depending on how many halogen atoms are available for substitution, the BIFs of THMs and HAAs may range from 0 to 3 (Goslan et al. 2009; Hua and Reckhow 2012). Greater bromide-tochlorine ratios in the source water result in the incorporation of more bromine (Hua and Reckhow 2012; Yang et al. 2012). Source water pH has also been shown to influence bromine incorporation, whereby increasing the pH (from 6.5 to 8.5) increased bromine incorporation for THMs and bromate, although bromine incorporation into HAAs was unchanged (Sohn et al. 2006). In order to control DBPs in treated drinking water, it is generally considered most efficient to prevent their formation in the first place (Serrano et al. 2015). Given the necessity for some form of disinfection, this implies removing the other DBP precursors such as NOM. Many different water treatment technologies have been applied to this end, such as coagulation and flocculation (Tubić et al. 2013; Matilainen and Sillanpää 2010), membrane filtration (GarcíaVaquero et al. 2014), adsorption (Bhatnagar and Sillanpää 2017) and electrochemical methods (Särkkä et al. 2015), but they are not always capable of fully degrading NOM and other non-biodegradable compounds. Advanced oxidation processes (AOPs) represent a more powerful solution, with the combination of O3/UV having been shown to be an efficient and environmentally friendly treatment technology in drinking water preparation, especially when applied prior to a final water polishing step (Toor and Mohseni 2007; Moncayo-Lasso et al. 2008). Combined O3/UV has even been applied in wastewater treatments for the degradation of NOM and other industrial waste products (Boczkaj and Fernandes 2017). AOPs generate highly reactive and nonselective hydroxyl radicals (HO·) which are capable, via the creation of lower molecular weight (MW) intermediates, of mineralising most organic compounds to non-toxic carbondioxide and water (Sanly et al. 2007). Among the AOPs used for drinking water treatment, ozone/ultraviolet (O3/UV) has been shown to result in significant NOM removal and reduction of DBPs formation potential, although the changes in NOM properties during chlorination may result in the formation of other DBPs instead (Chin and Bérubé 2005; Matilainen and Sillanpää 2010; Lamsal et al. 2011; Agbaba et al. 2016). UV irradiation was found to successfully decrease the ozone dose required during ozonation, which in turn reduces the formation of carcinogenic bromate (Meunier et al. 2006). Although many researchers have shown that O3/UV prior to chlorination is an effective method for removal of NOM and DBP precursors, data relating to the effect of O3/UV on the content of specific THM and HAA precursors in water are limited in the literature, especially in relation to the more toxic brominated DBPs, which have mainly been 13 investigated only in ozonated and chlorinated waters (Yang et al. 2012; Mao et al. 2014). The primary objective of this research was therefore to establish how the different DBP precursors are affected by O3/UV advanced processes during the treatment of water with low organic matter and moderate bromide contents. Different O3 and UV doses were applied during the AOPs, to investigate their influence on the species of THMs and HAAs formed, and BIFs were also calculated, in order to explore how these processes affect the conversion of bromine to DBPs during water treatment. All the research included in this work was carried out between October 2014 and November 2015, in Novi Sad, Republic of Serbia. Materials and methods The water used in this investigation was groundwater from Bačka (Republic of Serbia) drawn from a 159–170 m deep well. The groundwater was characterised (see Table 1) using the analytical methods detailed below. In order to assess the efficacy of the treatments investigated, before and after each oxidation treatment, the samples were analysed for their DOC contents, UV254 absorbance and disinfection byproduct formation potentials. Table 1 Characterisation of the natural groundwater investigated Parameter Units Mean value ± SD No. of measurements pH Total alkalinity Conductivity TOC DOC UV254 SUVA Br− THMFP CLFP BDCMFP DBCMFP BRFP HAAFP MCAAFP MBAAFP DCAAFP TCAAFP BCAAFP DBAAFP HANFP – mg CaCO3/L μS/cm mg/L mg/L cm−1 L/mg m µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L µg/L 7.9 ± 0.6 982 ± 35.8 797 ± 14.0 2.17 ± 0.22 2.06 ± 0.38 0.051 ± 0.001 2.43 ± 0.21 50.0 ± 10.0 180 ± 44.0 162 ± 40.3 16.6 ± 2.8 1.44 ± 0.03 < PQL 165 ± 12.4 < PQL < PQL 101 ± 20.0 60.8 ± 12.4 1.21 ± 0.45 2.25 ± 0.84 < PQL 10 10 10 11 22 22 22 6 22 22 22 22 22 22 22 22 22 22 22 22 6 International Journal of Environmental Science and Technology Standards for THMs [chloroform (CL), bromodichloromethane (BDCM), dibromochloromethane (DBCM) and bromoform (BR)], HAAs [monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA), bromochloroacetic acid (BCAA) and dibromochloroacetic acid (DBAA)], haloacetonitriles (HANs) [trichloroacetonitrile (TCAN), dichloroacetonitrile (DCAN), bromochloroacetonitrile (BCAN) and dibromoacetonitrile (DBAN)], bromide and bromate analysis were purchased from Supelco at a concentration of 2000 μg/mL. Methanol and methyl tert-butyl ether for organic residue analysis were purchased from J.T. Baker. All other chemicals were analytical grade and were used without further purification. All solutions were made up with ultrapure deionized water (prepared by LABCONCO, WaterPro RO/PS Station). Experimental procedures Ozonation of natural water was carried out without prior pH correction in a 6-L glass column (85 mm diameter). The electrochemical Argentox ozone generator used has a capacity of 1 g/h. Ozone was introduced to the water at a flow rate of 7 L/h via a diffuser at the bottom of the column. The contact times ranged from 3 to 10 min, in order to achieve a range of transferred ozone concentrations of 0.5–3.0 mg O3/mg DOC. Excess ozone in the gas phase was removed by nitrogen flow after ozonation. In order to facilitate the comparison of results, the ozone doses, which were chosen to investigate the influence of different doses on changes in NOM content and DBPs formation, were calculated in relation to the initial DOC concentration of the water. A commercially available stainless steel reactor was used in batch mode for the UV photolysis and O3/UV experiments. This 0.7-L vessel has a low-pressure mercury UV lamp (Philips TUV, 16 W) and is described in more detail in Molnar et al. (2015). The lamp emits UV light at a wavelength of 254 nm with an intensity of radiation of 10 mW/ cm2. The range of UV doses was from 600 to 6000 mJ/cm2 (either 1, 5 or 10 min). Analytical methods Samples were filtrated through a 0.45-μm membrane filter and analysed for DOC content by Elementar LiquiTOC II, using combustion at 850 °C to oxidise the carbon (SRPS 2007). UV254 absorbance measurements were performed in accordance with standard methods (APHA 2012) on a PG Instruments LtdT80 + UV/VIS spectrophotometer at a wavelength of 254 nm, with a 1 cm quartz cell, and the specific UV absorbance values (SUVA) were calculated. Disinfection by-products formation potential (DBPFP: THMFP, HAAFP and HANFP) was determined according to the standard method for measurement of THM formation potential (APHA 2012). At the end of the 7-day reaction period, the samples were dechlorinated and THMs, HAAs and HANs were analysed. THMs were directly analysed with a purge and trap system (Tekmar 3100) coupled to a gas chromatograph (Agilent Technologies 7890A), equipped with a DB-5 capillary column (30 m × 0.25 mm × 0.25 µm) and a mass selective detector (Agilent Technologies 5975C), based on USEPA Methods 5030B and 8260B (USEPA 1996a, b). HAAs were analysed after MTBE extraction–acid methanol methylation followed by analysis with GC/µECD instrument (Agilent 6890 N), equipped with a DB-XLB capillary column (4 × 250 mm), based on standard method USEPA 552.3 (USEPA 2003). HANs were determined by liquid–liquid extraction with MTBE followed by GC/µECD analysis (Agilent 6890N), according to USEPA Method 551.1 (USEPA 1995). The practical quantitation limits (PQL) for the analytical methods applied ranged from 0.34 to 0.74 μg/L for the THMs, 0.30–2.8 μg/L for the HAAs and 0.30–0.70 μg/L for the HANs. Bromide and bromate contents were measured by Dionex ICS-3000 Ion Chromatography System, with AG9-HC, AS9-HC anion column and a conductivity detector. Sodium carbonate solution (9 mM Na2CO3) was used for mobile phase, and a flow rate of 1 mL/min was applied. Additional operating conditions were: run-time 40 min and injection volume 75 µL. pH measurements were carried out on a WTW InoLab portable instrument. Turbidity was determined by nephelometry, against standard formazin polymer using a Hanna model HI 93703 instrument. Water conductivity was measured by conductometer Hanna model HI 933000. Water alkalinity (p- and m-alkalinity) was measured by standard volumetric method (APHA 2012). The concentration of ozone transferred to the water was calculated from the difference in the input and output ozone concentrations in the gas phase, which were measured by iodometric titration (APHA 2012). The equations for calculation of the BIFs for the THMs, HAAs and HANs are given in Eqs. (1–3) (Obolensky and Singer 2005) [ ] [ ] [ ] CHBrCl2 + 2 CHBr2 Cl + 3 CHBr3 BIF (THMs) = [ ] [ ] [ ] ] [ CHCl3 + CHBrCl2 + CHBr2 Cl + CHBr3 (1) [ ] [CHBrClCOOH] + 2 CHBr2 COOH BIF (HAAs) = [ ] [ ] CHCl2 COOH + [CHBrClCOOH] + CHBr2 COOH (2) 13 International Journal of Environmental Science and Technology [ ] [CHBrClCN] + 2 CHBr2 CN BIF (HANs) = [ ] [ ] CHCl2 CN + [CHBrClCN] + CHBr2 CN (3) The overall percentage of Br– incorporation in organic DBPs (THMs, HAAs and HANs) and bromate was also calculated as follows, whereby Br– is the bromide content in the investigated water (Sohn et al. 2006): Percentage of Br− incorporation in DBPs (%) (4) = [THM-Br, HAA-Br, HAN-Br] × 100∕Br− THM-Br (μg L−1 ) ([ ] [ ] [ ]) 2 CHBr2 Cl 3 CHBr3 CHBrCl2 = 79.9 × + + 163.80 208.25 252.70 (5) −1 HAA-Br (μg L ) = 79.9 × + [CHBrClCOOH] + 173.40 HAN-Br (μg L−1 ) = 79.9 × Influence of UV photolysis, O3 and O3/UV on the NOM content of the water ([ ] CH2 BrCOOH 138.90 [ ]) 2 CHBr2 COOH 217.80 ( [CHBrClCN] + 119.90 (6) [ ]) 2 CHBr2 CN 198.80 Percentage of Br− incorporation in bromate (%) [ [ ] ] BrO−3 BrO−3 as Br− × 100 = 0.625 × × 100. = Br− Br− (7) (8) Results and discussion Characteristics of the raw groundwater The natural groundwater samples used for this study were collected between October 2014 and July 2015. The results of the groundwater characterisation are given in Table 1. The natural water was characterised by low values of TOC, DOC and UV254 (Table 1), which is correlated with low levels of NOM. The DOC represents more than 95% of the TOC found in the groundwater. The mean SUVA value (2 < SUVA < 4 L/mg m) suggests the NOM present contains a mixture of hydrophilic and hydrophobic species with different MW (Uyak and Toroz 2007; Shan et al. 2012; Tian et al. 2013). The groundwater has a moderately high bromide content (Table 1), increasing the risk of bromate and brominated DBP formation during chemical oxidation. The high alkalinity of the water suggests that the carbonate and bicarbonate ions (known OH· scavengers) present in the water may inhibit bromate formation during treatment. The NOM content of the raw water is low, but showed a high reactivity towards chlorine, resulting in high values 13 of THM and HAA formation potential after chlorination (Table 1). The dominant THM compound was chloroform (90% of the total THMFP), with bromodichloro- and dibromochloromethane precursors also present (9 and 1%, respectively). At this bromide concentration, bromoform was not detected after chlorination of the raw water. From the group of HAAs, precursors of the chlorinated HAAs were present the most (di- and tri-chloro acetic acids 61 and 37%, respectively). The formation potential of brominated HAAs was 2%, whilst precursors of monohalogenated-HAA were not detected during the entire investigation. Because of the high reactivity of the NOM towards THM and HAA formation, UV photolysis and oxidation treatments were applied in order to investigate the mechanism of NOM oxidation and the subsequent effect on the formation and distribution of certain organic and inorganic DBPs. The influence of UV photolysis, ozonation and the O3/UV AOP on the character and content of the NOM in natural water was quantified by measuring the DOC, UV254 and SUVA values (Fig. 1). During UV photolysis, reduction of NOM was very low (up to 5% DOC, 18% UV254 and 13% SUVA). The UV treatment had a minor impact on the NOM structure, changing the UV254 absorbance and SUVA, but had almost no impact on DOC concentration. Other authors report that significant NOM removal can be achieved only at very high UV doses, such as 2.7–26.1 J/cm2 (Sindelar et al. 2014), 200 J/ cm2 (Parkinson et al. 2003), up to 1110 J/cm2 (Goslan et al. 2006), much higher than those used in this study. By applying ozone in the treatment, significantly better NOM removals were achieved (up to 11% DOC, 66% UV254 and 61% SUVA). Further increasing the ozone dose from 0.5 to 3.0 mg O3/mg DOC did not significantly improve the NOM removals. The greater reduction of UV254 absorbance in comparison with the DOC removal suggests conjugated double bonds were oxidised with minimal mineralisation. Thus, the oxidation of the hydrophobic NOM structure by ozone reduced the UV254 values and changed the NOM to be more hydrophilic in nature, resulting in lower SUVA values < 1 L/mg m. These by-products are less susceptible to attacks by OH· and therefore are not completely removed (Sanly et al. 2007; Matilainen and Sillanpää 2010). The combination of ozone with UV irradiation (O3/UV) reduced DOC, UV254 and SUVA up to 15, 80 and 75%, respectively. The NOM reduction slightly increased with increasing UV and ozone dose, with the maximum reduction achieved at the highest applied doses of ozone and UV irradiation (3.0 mg O3/mg DOC, 6000 mJ/cm2). The greater reduction in UV254 values indicates that the high International Journal of Environmental Science and Technology (a) mg O3/mg DOC 2.5 0 0.5 1 3 DOC (mg/L) 2.0 1.5 1.0 0.5 0.0 0 1000 2000 3000 4000 5000 6000 (Oh et al. 2003). This is in agreement with Lamsal et al. (2011), who observed 31% TOC removal after treatment by O3/UV treatment under similar conditions (~ 1.4 mg O3/mg DOC, 1140 mJ/cm2). The reductions in SUVA and UV254 achieved by the combined process (Fig. 1b, c) are significant as they indicate that the remaining organic carbon will be easier to remove by subsequent treatment steps. Bazri et al. (2012) demonstrated that SUVA and assimilable organic carbon are inversely correlated after oxidative treatment, implying that technologies such as biologically active carbon, which can be applied as a final treatment step for water polishing, will be increasingly effective at lower SUVA values (Toor and Mohseni 2007). UV dose (mJ/cm2) Influence of UV photolysis, O3 and O3/UV on DBP formation and speciation mg O3/mg DOC (b) 0.06 0 0.5 1 3 0.05 UV254 (cm-1) 0.04 0.03 0.02 0.01 0.00 0 1000 2000 3000 UV dose (c) 4000 5000 6000 (mJ/cm2) mg O3/mg DOC 3.0 0 0.5 1 3 SUVA (L/mg m) 2.5 2.0 1.5 1.0 0.5 0.0 0 1000 2000 3000 4000 5000 6000 UV dose (mJ/cm2) Fig. 1 Changes in a DOC, b UV254 absorbance, c SUVA values during the water treatments MW chromophores in the NOM were quickly broken down into low MW by-products with almost no UV absorbance, a finding in agreement with other authors (Matilainen and Sillanpää 2010). The increased degradation observed in the O3/UV process compared to ozonation and UV photolysis alone is probably a result of an increased formation in OH· The total THMFP and HAAFP of the natural and treated waters are shown in Fig. 2a, b. Note that the contents of THM and HAA precursors in the raw groundwater were similar, 180 ± 44 and 165 ± 12.4 µg/L, respectively. After UV photolysis alone, the dominant THM in each treatment was chloroform (around 90% of the total), with brominated THM precursors BDCM and DBCM contributing less than 10%. Increasing the UV fluence (600–6000 mJ/ cm2) resulted in greater THM precursor removals, with 6000 mJ/cm2 removing 7% of the chloroform and 16% of the brominated precursors. HAA formation potentials were again dominated by chlorinated precursors, with the brominated precursors yielding just 2% of the total HAAFP. However, in the case of HAA, UV photolysis generally increased the formation potentials, with the maximum UV fluence yielding 15% increases in both chlorinated and brominated HAA precursors. The slight changes in THM and HAA precursors are likely due to the lesser impact of UV irradiation on UV254 values and changes in the MW distribution of organic compounds (Lamsal et al. 2011). Other authors have also shown that by itself, UV photolysis is ineffective at removing the precursors of THMs and HAAs (Chin and Bérubé 2005; Sakai et al. 2013). Ozonation alone had a very straightforward positive effect, with higher ozone doses removing more than 27% of the total THM precursors at 3.0 mg O3/mg DOC. At lower ozone doses there was a shift from monobromo- into dibromo-substituted THMs. However, at the higher ozone dose, the available bromide was spent on the formation of other brominated DBP and there was no DBCM present after chlorination in this sample. Other authors have shown that ozonation of drinking water shifts the subsequent formation of DBPs to more brominated species, as a result of higher NOM hydrophilicity (as compounds which are more reactive with bromine than chlorine) in water (Wert and RosarioOrtiz 2011). Precursors of both chlorinated and mono- and 13 International Journal of Environmental Science and Technology (a) 200 BR DBCM BDCM CL 180 THMFP (µg/L) Fig. 2 Changes of a THMFP, b HAAFP in the water during UV photolysis alone and combined with ozone at doses of 0.5 mg O3/mg DOC, 1.0 mg O3/mg DOC and 3.0 mg O3/mg DOC 160 140 120 100 0 600 3000 6000 UV photolysis 0 600 3000 6000 0.5 mg O3/mg DOC 0 600 3000 6000 1.0 mg O3/mg DOC 0 600 3000 6000 3.0 mg O3/mg DOC Water treatment (b) 300 DBAA BCAA TCAA DCAA 0 3000 6000 250 HAAFP (µg/L) 200 150 100 50 0 0 600 3000 6000 UV photolysis dibromo-HAAs decreased during ozonation, although the highest dose was much less effective at removing HAA precursors than the other two doses, with removals falling from around 50 to 15%. Ozone is a strong electrophilic agent and at very high doses could change the NOM structure sufficiently to create more reactive precursors for some DBPs. In the range of the two lower ozone doses, Mao et al. (2014) reported similar results, with increasing ozone dose improving the removal of chlorinated DBP precursors and shifting the DBP distribution to more brominated DBPs species. The results of the combination of ozone and UV photolysis are considerably less straightforward, with strong evidence for an interaction effect between the ozone dose and UV fluence. For the THM precursors, the combined 13 0 600 3000 6000 600 1.0 mg O3/mg DOC 0.5 O3/mg DOC Water treatment 0 600 3000 6000 3.0 mg O3/mg DOC treatment only showed an improvement over ozonation alone at the highest ozone dose. At the low 0.5 mg O3/mg DOC dose, increasing the UV fluence has a significant negative effect on the THMFP, with removals falling from 18 to 4% at the highest UV fluence. In contrast, at the highest ozone dose, total THM precursor removals generally improved with increasing UV fluence, with removal maxima for the precursors of chlorinated and brominated THMs at 3000 and 600 mJ/cm2, respectively. Although UV had a negative impact on the THM precursor removals at 1.0 mg O3/mg DOC, increasing the UV fluence again improved the combined treatment performance. The THM precursor results at the two higher ozone doses are in agreement with Lamsal et al. (2011), who achieved 75% removal of THM precursors International Journal of Environmental Science and Technology under similar treatment conditions. These results are also comparable with our previous research, carried out in a different groundwater with greater NOM content, whereby a slight reduction (17%) in THMFP was observed after the O3/ UV process (0.5 mg O3/mg DOC; 3000 mJ/cm2) (Agbaba et al. 2016). The HAA precursors behaved somewhat differently, with the same general trend observed at each ozone dose: addition of UV to the process has a negative impact on treatment performance, with lesser removals by the combined process with increasing UV fluence. This negative effect is particularly pronounced for the precursors of the chlorinated HAA, which increased by almost 60% at 1.0 mg O3/mg DOC and 3000 mJ/cm2, whereas brominated HAA precursors only increased by 29% under the same conditions. Although Lamsal et al. (2011) reported significant improvements with O3/ UV in comparison with O3 alone, it should be noted that the groundwater from Bačka in this study differs considerably from the surface water Lamsal et al. investigated, with NOM of different hydrophobicity and an alkalinity which is greater by two orders of magnitude. It is possible the high alkalinity of the water (Table 1) may scavenge OH· radicals in the water (Bazri et al. 2012), therefore favouring the oxidation of NOM by molecular ozone, potentially explaining the lower HAAFP removals under these conditions. The contrasting results for the THMFP and HAAFP during the combined O3/UV treatment indicate that in this groundwater source, at least some of their precursors originate in different NOM fractions. This is not unusual, as THM and HAA concentrations in treated waters do not always correlate (Malliarou et al. 2005). The UV254 and SUVA results presented above show that the combined treatment favours the oxidation of aromatic NOM to more hydrophilic compounds. Together with the fact that increasing the UV and O3 doses more consistently improved the removals of THM precursors, this suggests that the THM precursor material is Fig. 3 Changes in BIFs during water treatment for a THM, b HAA, c HAN more hydrophobic than the HAA precursor material, and is therefore easier to oxidise (Molnar et al. 2013). Note that after all the treatments which applied 3.0 mg O3/ mg DOC ozone, precursors of DBAN were also detected in the water. DBAN precursors were the only HAN precursors detected and were not present in the raw water (Petronijević et al. 2015; Table 1). Previous work has confirmed that the residual bromide present after AOPs can react with residual NOM during disinfection to form brominated HANs (Molnar et al. 2015; Petronijević et al. 2015). Bromine incorporation during water treatment The bromine incorporation factor (BIF) proposed by Obolensky and Singer (2005) was calculated to show the degree of bromine substitution in DBPs after the different oxidation processes investigated. In general, the BIF values obtained for the THMs ranged from 0.078 to 0.11 and were about 2–7 times greater than the HAA values, which ranged from 0.016 to 0.040 (Fig. 3). In the UV treatment, increasing the UV fluence reduced the BIF, mostly as a result of BDCM precursor degradation (Fig. 2a). However, UV photolysis by itself did not have a significant impact on the BIF values for HAA precursors. Overall, the UV treatment therefore had a positive effect on the reduction of brominated DBPs. In contrast, ozonation alone up to 1.0 mg O3/mg DOC increased the BIF for THMs, although the HAA BIF was not affected. At ozone doses of 0.5–1.0 mg O 3/mg DOC, THM BIF values were higher than the raw water values. A further increase in the ozone dose to 3.0 mg O 3/mg DOC was sufficient to avoid significantly increasing in THM BIF, as at this dose, all the BDCM precursors were degraded. This may be explained by the inorganic bromate formation during ozonation, which increases with increasing ozone dose (up to 11.1 µg BrO3−/L). Using high ozone doses also THM HAA HAN 2.00 0.12 0.10 0.08 0.06 0.04 0.02 6000 3000 600 0 6000 3000 600 0 6000 3000 600 0 6000 3000 600 0.00 0 BIF (DBPs) 0.14 UV photolysis 0.5 mg O3/mg DOC 1.0 mg O3/mg DOC 3.0 mg O3/mg DOC Water treatment 13 International Journal of Environmental Science and Technology Influence of the investigated oxidation processes on bromide speciation In the final part of this work, the bromide mass balance in water during the different water treatments is presented (Fig. 4). The total percentage of brominated DBPs in natural water (22% bromide contribution) was calculated as the sum of brominated THMs and HAAs (18 and 4%, respectively). The UV treatment alone had almost no impact on the distribution of brominated DBPs, but ozonation greatly increased bromine incorporation (by up to 38%), mainly as a result of bromate formation. The percentage conversion of THM HAA HAN Bromate 80 60 40 20 0.5 mg O3/mg DOC 1.0 mg O3/mg DOC 3.0 mg O3/mg DOC Water treatment 6000 3000 600 0 6000 3000 600 6000 600 0 6000 600 3000 0 UV photolysis 13 Others 100 0 Percentage of bromine incorporaon in DBPs (%) Fig. 4 Bromide mass balance in water during a UV photolysis and AOPs, b 0.5 mg O3/mg DOC, c 1.0 mg O3/mg DOC, d 3.0 mg O3/mg DOC 0 bromine into bromate increased by up to 14% during ozonation, although the bromate content remained far below that of the brominated organic DBPs. Interestingly, the addition of UV irradiation to the ozonation process completely eliminated bromate formation, favouring the formation of brominated THMs at the lowest ozone dose, brominated HAAs at the middle dose, and brominated HAN at the highest dose. Brominated HANs are more toxic compounds than the brominated THMs and HAAs (Richardson et al. 2007), and under these conditions represented 22–30% of the bromine contribution, which is in agreement with Hu et al. (2010). In the experimental conditions applied in this work, the total conversion of bromine into brominated DBPs was very high, ranging from 22 to 48%. This may be explained by the presence of excess OH· in the weakly basic conditions of the water investigated, which promote the base-catalysed hydrolysis of certain functional groups during the formation of brominated THMs and bromate, as postulated by Sohn et al. (2006). Those authors reported similar bromine incorporation percentages for twelve surface and groundwaters, in the ranges 6–41% for THMs, 4–13% for HAAs and 4–22% for bromate (Sohn et al. 2006). The results in this work are also in agreement with von Gunten (2003), who suggests that in waters with bromide contents above 20 µg/L, the formation of excessive brominated organic DBP and bromate may be especially problematic under certain oxidation process conditions. Ultimately, when designing a drinking water treatment plant, the choice of which technologies to apply will be informed by the need to maximise the safety of the resulting water whilst minimising the capital and operation and maintenance costs. In this context, although a full cost analysis is beyond the scope of this present work, it should be noted that several studies have shown that for certain oxidation applications, the UV/O3 combined process need not be more expensive than ozonation alone (Crozes et al. 2003; Boczkaj and Fernandes 2017): the costs of UV technology are small 3000 increases the NOM hydrophilicity (SUVA value decreased to < 1 L/mg m, Fig. 1c), resulting in the formation of brominated HANs (Fig. 3). The higher BIF values obtained therefore confirm the shift towards the formation of more brominated DBPs due to ozonation. As explained above, DBAN was the only HAN detected during these experiments, hence the maximum BIF for HAN found at 3.0 mg O3/mg DOC. The addition of UV irradiation to the ozonation process increases the THM BIF values at each ozone dose applied. This again implies that the combined O 3/UV process has oxidised the THM precursors present in the water into more hydrophilic structures, which are more reactive with bromide than chloride during the disinfection process. Note that at the lower ozone dose of 0.5 O3/ mg DOC, the combined process significantly reduced the HAA BIF by almost 50%. This is due to the proportionately larger increase in chlorinated HAAs in comparison with the brominated HAA. At the higher ozone doses, the addition of UV irradiation had no significant effect on HAA BIF values compared to ozonation alone. From these results, it appears that under these conditions, there are THM precursors which are considerably more reactive towards bromide than the HAA precursors. International Journal of Environmental Science and Technology in comparison with ozonation and may well be offset by the increased process efficacy allowing for the application of smaller capacity O3 generators. As shown in this work, the specific characteristics of the water matrix, especially the concentration and nature of the NOM present, can have a major impact on the process efficacy and can therefore also impact on the final treatment cost. For this reason, treatment technologies must be carefully optimised for each particular water supply. Conclusion This work investigated the effect of three processes, UV irradiation, ozonation and O3/UV AOP, on the NOM content and disinfection by-product formation potentials of groundwater. The groundwater investigated has a moderate bromide concentration and low NOM content, but showed high values of THM and HAA formation potential after chlorination. Efficacy in removing NOM followed the following sequence: O3/UV > ozonation > UV irradiation. UV irradiation alone had a better effect at removing the precursors of brominated THMs than chlorinated THMs, but slightly increased the precursors of both brominated and chlorinated HAAs. During ozonation, reductions in the HAAFP were significantly larger than those of THMFP, compared to the natural water. In the combined O3/UV treatment, increasing the UV fluence had a contrasting effect on precursor behaviour, depending on the ozone dose applied. Calculations of the BIFs and the mass balance of brominated DBPs revealed how sensitive the structure of the NOM present in the raw water is to changes in the O3/UV process conditions. Ozonation alone resulted in the formation of increasing amounts of bromate at higher ozone doses. However, the combined O3/UV process completely eliminated the formation of bromate, with the formation of brominated THMs, HAAs or HANs instead being favoured depending on the ozone dose applied. The total percentage conversion of bromine into brominated DBPs reached maximum values at the highest applied ozone dose during both ozonation and the combined O3/UV process. Given the varying toxicities of the different by-products, and the increased toxicity of brominated disinfection byproducts in particular, further work is required to investigate the interaction observed between ozone dose and UV fluence, in order to ensure the combined treatment process can be optimised to fully minimise the risks posed by DBPs to human health. Acknowledgements The authors gratefully acknowledge the support of the Ministry of Education, Science and Technological Development of the Republic of Serbia (Projects III43005 and III43009) and Anahem d.o.o. laboratory in Belgrade for the analytical support. References Agbaba J, Molnar Jazic J, Tubic A, Watson M, Maletic S, Kragulj Isakovski M, Dalmacija B (2016) Oxidation of natural organic matter with processes involving O3, H2O2 and UV light: formation of oxidation and disinfection by-products. RSC Adv 6:86212–86219. https://doi.org/10.1039/c6ra18072h APHA (2012) American Public Health Association, American Water Works Association. Water pollution control federation, 20th ed. Standard methods for the examination of water and wastewater Bazri MM, Barbeau B, Mohseni M (2012) Impact of UV/H2O 2 advanced oxidation treatment on molecular weight distribution of NOM and biostability of water. Water Res 46(16):5297–5304. https://doi.org/10.1016/j.watres.2012.07.017 Bhatnagar A, Sillanpää M (2017) Removal of natural organic matter (NOM) and its constituents from water by adsorption: a review. Chemosphere 166:497–510. https://doi.org/10.1016/j.chemosph ere.2016.09.098 Boczkaj G, Fernandes A (2017) Wastewater treatment by means of advanced oxidation processes at basic pH conditions: a review. J Chem Eng 320:608–633. http s://doi.org/10.1016 /j.cej.2017 .03.084 Chin A, Bérubé PR (2005) Removal of disinfection by-product precursors with ozone-UV advanced oxidation process. Water Res 39:2136–2144. https://doi.org/10.1016/j.watres.2005.03.021 Crozes G, Cushing R, Rennecker J, Cleveland C, Wright H (2003) Synergies of UV disinfection and ozone in water treatment, International UV Association, Vienna, pp 1–5 Directive 98/83/EC (1998) Directive 98/83/EC of 16 February 1998 on the quality of water for human consumption. Official J. L 123.24.4, pp 26–40 García-Vaquero N, Lee E, Jiménez Castañeda R, Cho J, López-Ramírez JA (2014) Comparison of drinking water pollutant removal using a nanofiltration pilot plant powered by renewable energy and a conventional treatment facility. Desalination 347:94–102. https:// doi.org/10.1016/j.desal.2014.05.036 Goslan EH, Gurses F, Banks J, Parsons SA (2006) An investigation into reservoir NOM reduction by UV photolysis and advanced oxidation processes. Chemosphere 65:1113–1119. http s://doi. org/10.1016/j.chemosphere.2006.04.041 Goslan EH, Krasner SW, Bower M, Rocks SA, Holmes P, Levy LS, Parsons SA (2009) A comparison of disinfection by-products found in chlorinated and chloraminated drinking waters in Scotland. Water Res 43:4698–4706. http s://doi.org/10.1016 /j.watr es.2009.07.029 Hu J, Song H, Karanfil T (2010) Comparative analysis of halonitromethane and trihalomethane formation and speciation in drinking water: the effects of disinfectants, pH, bromide, and nitrite. Environ Sci Technol 44:794–799. https://doi.org/10.1021/es90 2630u Hua G, Reckhow DA (2012) Evaluation of bromine substitution factors of DBPs during chlorination and chloramination. Water Res 46(4):208–4216. https://doi.org/10.1016/j.watres.2012.05.031 Lamsal R, Walsh ME, Gagnon GA (2011) Comparison of advanced oxidation processes for the removal of natural organic matter. Water Res 45:3263–3269. https://doi.org/10.1016/j.watres.2011 .03.038 Malliarou E, Collins C, Graham N, Nieuwenhuijsen MJ (2005) Haloacetic acids in drinking water in the United Kingdom. Water Res 39:2722–2730. https://doi.org/10.1016/j.watres.2005.04.052 Mao Y, Wang X, Yang H, Wang H, Xie YF (2014) Effects of ozonation on disinfection by-product formation and speciation during subsequent chlorination. Chemosphere 117:515–520. https://doi. org/10.1016/j.cej.2013.02.08 13 International Journal of Environmental Science and Technology Matilainen A, Sillanpää M (2010) Removal of natural organic matter from drinking water by advanced oxidation processes. Chemosphere 80:351–365. https://doi.org/10.1016/j.chemosphere.2010 .04.067 Meunier L, Canonica S, von Gunten U (2006) Implications of sequential use of UV and ozone for drinking water quality. Water Res 40:1864–1876. https://doi.org/10.1016/j.watres.2006.02.030 Molnar J, Agbaba J, Dalmacija B, Tubić A, Krčmar D, Maletić S, Tomašević D (2013) The effects of matrices and ozone dose on changes in the characteristics of natural organic matter. J Chem Eng 222:435–443. https://doi.org/10.1016/j.cej.2013.02.087 Molnar J, Agbaba J, Tubić A, Watson M, Kragulj M, Rončević S, Dalmacija B (2015) The effects of ultraviolet/H2O2 advanced oxidation on the content and characteristics of groundwater natural organic matter. Water Sci Technol WS 15:34–41. https://doi. org/10.2166/ws.2014.081 Moncayo-Lasso A, Pulgarin C, Benítez N (2008) Degradation of DBPs’ precursors in river water before and after slow sand filtration by photo-Fenton process at pH 5 in solar CPC reactor. Water Res. https://doi.org/10.1016/j.watres.2008.07.014 Mosteo R, Miguel N, Martin-Muniesa N, Ormad MP, Ovelleiro JL (2009) Evaluation of trihalomethane formation potential in function of oxidation processes used during the drinking water production process. J Hazard Mater 172:661–666. https://doi.org/10.1016 /j.jhazmat.2009.07.048 Obolensky A, Singer PC (2005) Halogen substitution patterns among disinfection byproducts in the information collection rule database. Environ Sci Technol 39:2719–2730. https://doi.org/10.1021 /es0489339 Official Gazette RS No. 42/98 and 44/99 (1998) Regulation on the hygienic quality of drinking water Republic of Serbia (on Serbian) Oh HJ, Kim WJ, Choi JS, Gee CS, Hwang TM, Kang JG, Kang JW (2003) Optimization and control of ozonation plant using raw water characterization method. Ozone Sci Eng 25:383–392. http s://doi.org/10.1080/01919510390481702 Parkinson A, Roddick FA, Hobday MD (2003) UV photooxidation of NOM: issues related to drinking water treatment. J Water Supply Res Technol 52:577–586 Petronijević M, Agbaba J, Molnar Jazić J, Watson M, Dalmacija B (2015) The influence of an O3/UV advanced oxidation process on the content of certain disinfection byproduct precursors in water. In: XI Symposium of “Modern technology and economic development” October 22–23. Leskovac, Serbia 211–222. UDC 628.166.094.3-926.214:547-302 Richardson SD, Plewa MJ, Wagner ED, Schoeny R, DeMarini DM (2007) Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: a review and roadmap for research. Mutat Res 636:178–242. https://doi. org/10.1016/j.mrrev.2007.09.001 Sakai H, Autin O, Parsons S (2013) Change in haloacetic acid formation potential during UV and UV/H2O2 treatment of model organic compounds. Chemosphere 92:647–651. http s://doi. org/10.1016/j.chemosphere.2013.01.072 Sanly ML, Chiang K, Amal R, Fabris R, Chow C, Drikas M (2007) A study on the removal of humic acid using advanced oxidation processes. Sep Sci Technol 42:1391–1404. https://doi.org/10.1080 /01496390701289799 Särkkä H, Vepsäläinen M, Sillanpää M (2015) Natural organic matter (NOM) removal by electrochemical methods: a review. J Electroanal Chem 755:100–108. https://doi.org/10.1016/j.jelechem .2015.07.029 Serrano M, Montesinos I, Cardador MJ, Silva M, Gallego M (2015) Seasonal evaluation of the presence of 46 disinfection by-products throughout a drinking water treatment plant. Sci Total Environ 517:246–258. https://doi.org/10.1016/j.scitotenv.2015.02.070 13 Shan J, Hu J, Kaplan-Bekaroglu SS, Song H, Karanfil T (2012) The effects of pH, bromide and nitrite on halonitromethane and trihalomethane formation from amino acids and amino sugars. Chemosphere 86:323–328. https://doi.org/10.1016/j.chemosphere.2011 .09.004 Sindelar HR, Brown MT, Boyer TH (2014) Evaluating UV/H2O2, UV/ percarbonate, and UV/perborate for natural organic matter reduction from alternative water sources. Chemosphere 105:112–118. https://doi.org/10.1016/j.chemosphere.2013.12.040 Sohn J, Amy G, Yoon Y (2006) Bromide ion incorporation into brominated disinfection by-products. Water Air Soil Pollut 174:265– 277. https://doi.org/10.1007/s11270-006-9104-3 SRPS ISO 8245:2007 (2007) Water quality. Guidelines for determination of total organic carbon (TOC) and dissolved organic carbon (DOC) in water Tian C, Liu R, Guo T, Liu H, Luo Q, Qu J (2013) Chlorination and chloramination of high-bromide natural water: DBPs species transformation. Sep Purif Technol 102:86–93. https://doi. org/10.1016/j.seppur.2012.09.034 Toor R, Mohseni M (2007) UV-H2O2 based AOP and its integration with biological activated carbon treatment for DBP reduction in drinking water. Chemosphere 66:2087–2095. https://doi. org/10.1016/j.chemosphere.2006.09.043 Tubić A, Agbaba J, Dalmacija B, Molnar J, Maletić S, Watson M, Ugarčina Perović S (2013) Insight into changes during coagulation in NOM reactivity for trihalomethanes and haloacetic acids formation. J Environ Manag 118:153–160. https://doi.org/10.1016 /j.jenvman.2012.11.046 USEPA (1995) Method 551.1: determination of chlorination disinfection byproducts, chlorinated solvents, and halogenated pesticides/ herbicides in drinking water. National Exposure Research Laboratory, Office of Research and Development, Environmental Protection Agency, Cincinnati USEPA (1996a) Method 5030B: purge and trap for aqueous samples. U.S. Environmental Protection Agency, Cincinnati USEPA (1996b) Method 8260B: volatile organic compounds by gas chromatography/mass spectrometry (GC/MS). U.S. Environmental Protection Agency, Cincinnati USEPA (2003) Method 552.3: determination of haloacetic acids and dalapon in drinking water by liquid–liquid microextraction, derivatization, and gas chromatography with electron capture detection. Technical support center office of ground water and drinking water. U.S. Environmental Protection Agency, Cincinnati USEPA (2006) National primary drinking water regulations: stage 2 disinfectants and disinfection by products rule: final rule. Fed Regist 71(2):388–493 Uyak V, Toroz I (2007) Investigation of bromide ion effects on disinfection by-products formation and speciation in an Istanbul water supply. J Hazard Mater 149:445–451. http s://doi.org/10.1016 /j.jhazmat.2007.04.017 von Gunten U (2003) Ozonation of drinking water: part I. Oxidation kinetics and product formation. Water Res 37:1443–1467. https:// doi.org/10.1016/S0043-1354(02)00457-8 Wert EC, Rosario-Ortiz FL (2011) Effect of ozonation on trihalomethane and haloacetic acid formation and speciation in a full-scale distribution system. Ozone Sci Eng 33:14–22. http s://doi. org/10.1080/01919512.2011.536504 Yang X, Peng J, Chen B, Guo W, Liang Y, Liu W, Liu L (2012) Effects of ozone and ozone/peroxide pretreatments on disinfection byproduct formation during subsequent chlorination and chloramination. J Hazard Mater 239–240:348–354. https://doi. org/10.1016/j.jhazmat.2012.09.006