ARTICLE IN PRESS

WATER RESEARCH 41 (2007) 1481– 1490

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Formation of oxidation byproducts from ozonation of wastewater

Eric C. Werta,Ã, Fernando L. Rosario-Ortiza, Doug D. Druryb, Shane A. Snydera aSouthern Nevada Water Authority (SNWA), 243 Lakeshore Rd., Boulder City, NV 89005, USA bClark County Water Reclamation District, 5857 E. Flamingo Road, Las Vegas, NV 89122, USA article info abstract

Article history: Disinfection byproduct (DBP) formation in tertiary wastewater was examined after

Received 27 July 2006 ozonation (O3) and advanced oxidation with O3 and (O3/H2O2). O3 and

Received in revised form O3/H2O2 were applied at multiple dosages to investigate DBP formation during coliform

11 January 2007 disinfection and trace contaminant oxidation. Results showed O3 provided superior

Accepted 16 January 2007 disinfection of fecal and total coliforms compared to O3/H2O2. Color, UV absorbance, and

SUVA were reduced by O3 and O3/H2O2, offering wastewater utilities a few potential Keywords: surrogates to monitor disinfection or trace contaminant oxidation. At equivalent O3 dosages, O3/H2O2 produced greater concentrations of assimilable organic carbon (5–52%), Advanced oxidation process (AOP) aldehydes (31–47%), and carboxylic acids (12–43%) compared to O3 alone, indicating that Hydroxyl radicals organic DBP formation is largely dependent upon hydroxyl radical exposure. Bromate Wastewater formation occurred when O3 dosages exceeded the O3 demand of the wastewater. Bench- Water reuse scale tests with free chlorine showed O3 is capable of reducing total organic halide (TOX) Disinfection byproducts (DBP) formation potential by at least 20%. In summary, O3 provided superior disinfection Bromate compared to O3/H2O2 while minimizing DBP concentrations. These are important Aldehydes considerations for water reuse, aquifer storage and recovery, and advanced wastewater Carboxylic acids treatment applications. & 2007 Elsevier Ltd. All rights reserved.

1. Introduction trace contaminant oxidation make ozonation an attractive alternative for advanced . Ozonation has been shown to be highly effective for water During ozonation, the formation of organic disinfection. In drinking , ozonation is used to (e.g., assimilable organic carbon (AOC), aldehydes, carboxylic meet United States Environmental Protection Agency (USEPA) acids, and ketones) and inorganic (e.g., bromate) disinfection regulations for the inactivation of viruses, Cryptosporidium,and byproducts (DBPs) has been well documented (Richardson Giardia (USEPA, 1991, 2003). In wastewater treatment, ozona- et al., 1999; Huang et al., 2005). Organic DBPs from ozonation tion has been used to meet discharge requirements for have been linked to increased bacterial regrowth in drinking coliform and virus inactivation since the 1970s (Rice et al., water distribution systems (LeChevallier et al., 1992). Drinking 1981). In recent years, ozonation has gained attention for its water utilities often employ biological filtration to remove ability to oxidize endocrine disrupting chemicals (EDCs) and these byproducts creating a biologically stable water prior to pharmaceuticals in both drinking water and wastewater distribution (Huck et al., 1991; Krasner et al., 1993). In water (Zwiener and Frimmel, 2000; Huber et al., 2003, 2005; Snyder reuse applications, AOC has contributed to increased bacter- et al., 2006). The combination of microbial disinfection and ial regrowth and more rapid chlorine decay in distribution

ÃCorresponding author. Tel.:+1 702 567 2306; fax:+1 702 567 2085. E-mail address: [email protected] (E.C. Wert). 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.01.020 ARTICLE IN PRESS

1482 WATER RESEARCH 41 (2007) 1481– 1490

systems (Ryu et al., 2005). Currently, bromate is the only tion operated by the Clark County Water Reclamation District ozone (O3) DBP regulated in drinking water by the USEPA, in Las Vegas, Nevada, USA. The WWTP treats wastewater which established a maximum contaminant level (MCL) of by first adding 14–16 mg/L of ferric chloride followed by 10 mg/L (USEPA, 1998). primary clarification and sedimentation, allowing heavier Dissolved organic carbon (DOC) concentrations are typically particles to be removed. Then, aerobic and anaerobic greater in wastewater than in surface water, resulting in faster biological treatment occurs followed by secondary clarifica-

O3 decomposition rates and increased hydroxyl radical (dOH) tion. During the tertiary treatment stage, 4–13 mg/L of alum is exposures (Buffle et al., 2006a). As a result, higher O3 dosages added followed by flocculation and filtration. are required to meet wastewater treatment goals, potentially light is used to achieve disinfection goals. Samples for this leading to increased DBP formation. The organic composition study were collected after filtration, but prior to UV disinfec- of wastewater is also different from the natural organic matter tion. Seasonal variations were examined by collecting sam- (NOM) found in surface water supplies and is commonly ples in June 2005 and January 2006. Table 1 shows the tertiary known as effluent organic matter (EfOM). EfOM is composed of wastewater quality was very similar during both sampling recalcitrant NOM from drinking water, synthetic organic events. chemicals added during anthropogenic use (including disin- fection by-products), and soluble microbial products (Shon et 2.3. Ozonation at bench scale al., 2006). Sparse data exist regarding the influence of EfOM on DBP formation. Sirivedhin and Gray (2005) studied the impact Bench-scale tests were performed using a batch reactor to of EfOM on halogenated DBP formation during chlorination of obtain information about O3 decomposition, dOH exposure, South Platte River water. They found that locations with and bromate formation. A sample of NanopureTM water was greater EfOM influence were less reactive with chlorine on a placed inside a water-jacketed flask and cooled to 2 1C. Once per carbon basis than locations with minimal EfOM influence. cooled, 11% gaseous O3 was diffused into the water using an The primary objective of this study was to quantify the oxygen-fed generator (model CFS-1A, Ozonia North America formation of known organic and inorganic O3 DBPs at dosages Inc., Elmwood Park, NJ USA). O3 stock solution concentrations required for coliform disinfection and trace contaminant and dissolved O3 residuals were measured according to destruction in tertiary wastewater. Conventional O3 and Standard Methods 4500-O3 (Bader and Hoigne, 1982; APHA advanced oxidation processes (AOPs) with O3 and hydrogen et al., 1998). O3 dosages were administered by injecting an peroxide (O3/H2O2) were evaluated at bench-scale and pilot- aliquot of the stock solution into a 1-L amber glass container scale to examine DBP formation. Subsequent chlorination with a repeating pipette dispenser containing the tertiary examined the impact of O3 and O3/H2O2 on total organic treated wastewater at room temperature (20 1C). During O3/ halide (TOX) formation. These results will determine which H2O2 experiments, H2O2 was added 30 s prior to the addition ozonation technique is optimal to meet multiple wastewater of O3 stock solution. Duplicate experiments were performed treatment objectives while minimizing DBP formation. with the addition of 1.75 mM (274 mg/L) para-chlorobenzoic

acid (pCBA), an O3-resistant probe compound which reacts selectively with dOH (Elovitz and von Gunten, 1999). Dis-

2. Materials and methods solved O3 residual and pCBA samples were collected at 10 s intervals during the first minute of reaction and each minute

2.1. Experimental plan thereafter to investigate O3 decomposition and dOH exposure.

Dissolved O3 residual and pCBA samples were collected until

Bench-scale and pilot-scale experiments were conducted the O3 residual had decayed to less than 0.05 mg/L, or until a with tertiary wastewater to quantify the DBP formation contact time of 15 min was achieved. Bromate samples were during O3 and O3/H2O2 and subsequent TOX formation during collected after each test was completed. chlorination. The testing was conducted in three phases: (1) bench-scale O3 and O3/H2O2, (2) pilot-scale O3 and O3/H2O2, and (3) bench-scale chlorination. Bench-scale O3 and O3/H2O2 tests provided information regarding O3 demand, decay rate, Table 1 – Water quality summary of tertiary wastewater and dOH exposure. Due to the 1 L sample volume limitation during bench-scale testing, pilot-scale experiments were used Water quality Units June January to gather adequate sample volumes to measure coliform parameter 2005 2006 disinfection, DBP formation (AOC, carboxylic acids, alde- hydes, and bromate), and trace contaminant removal. Results Ammonia mg-N L1 o 0.08 o 0.08 of trace contaminant testing (i.e., EDCs and pharmaceuticals) Bromate mg L1 o 0.001 o 0.001 1 are presented elsewhere (Snyder et al., 2006). Bench-scale Bromide mg L 0.28 0.21 1 chlorination tests using free chlorine determined the impact Nitrate mg-N L 13.6 13.8 pH Units 6.92 7.02 of ozonation on TOX formation. Temperature 1C 27.2 20.4 Total alkalinity mg L1 130 110 2.2. Wastewater treatment facility Total organic carbon mg L1 7.10 7.20 UV254 cm1 0.128 0.119 1 1 Tertiary wastewater was collected from a 416,000 m3/d waste- SUVA L mg m 1.80 1.65 water treatment plant (WWTP) with nitrification/denitrifica- ARTICLE IN PRESS

WATER RESEARCH 41 (2007) 1481– 1490 1483

2.4. Ozonation at pilot scale H1-S, IN USA Inc., Needham, MA, USA) were used to calculate

and control the O3 dosage. Due to the column height of 0.83 m A 1 L/min bench-top pilot plant (BTPP) consisting of all inert and diameter of 0.055 m, the transfer efficiency varied materials including glass, stainless steel, and fluorocarbon between 40% and 70% depending on the desired dose and polymers was used to conduct the testing (Fig. 1). A peristaltic corresponding gas flow rate. The off-gas was collected from pump was used to transport the wastewater from a 208-L the top of each cell into a central manifold and destroyed by stainless steel drum into the contactor. During O3/H2O2 manganese dioxide catalyst. experiments, H2O2 (3% stock, Fisher Scientific, Pittsburgh, PA O3 dosages were selected based upon bench-scale demand USA) was injected into the wastewater flow stream followed tests to evaluate a range of exposures for coliform disinfec- by a static mixer prior to entering the contactor. The O3 tion and trace contaminant oxidation. O3 exposure was contactor consisted of 12 glass chambers each providing calculated by integrating the dissolved residual concentration

2 min of contact time for a total of 24 min. O3 feed gas was over time (CT) according to the extended integrated CT10 produced from oxygen gas using a laboratory-scale generator method (Rakness et al., 2005). During each test, the waste- (model LAB2B, Ozonia North America Inc., Elmwood Park, NJ water was maintained at room temperature (20 1C). Dissolved

USA). O3 was added in the first contactor chamber with O3 measurements were collected after 2, 6, 10, 14, and 18 min counter-current flow through a glass-fritted diffuser with to examine demand and decay rates. Water quality samples bubble size of 0.1 mm. A mass-flow controller (model were collected for AOC, carboxylic acids, aldehydes, bromate, AFC2600D, Aalborg Instruments and Controls, Inc., Orange- total coliforms and fecal coliforms. In order to measure dOH burg, NY, USA) and a feed gas concentration analyzer (model exposure, duplicate experiments were performed with 75.6 L

Fig. 1 – Schematic of bench-top pilot plant (BTPP). ARTICLE IN PRESS

1484 WATER RESEARCH 41 (2007) 1481– 1490

of tertiary wastewater spiked with 1 mM (156 mg/L) of pCBA. exceeded the IOD, H2O2 promoted O3 decomposition thereby Samples for pCBA analysis were collected after 2, 6, 10, 14, increasing initial dOH exposure. and 18 min and quenched with a small aliquot of sodium thiosulfate (Na2S2O3). 3.2. Hydroxyl radical exposure

2.5. Analytical methods Oxidation attributed to dOH exposure was measured by the depletion of pCBA as shown in Fig. 3. During the IOD phase Water samples were collected, preserved, and refrigerated at (timeo30 s), results showed pCBA depletion was 13–23% 4 1C until analyzed. Standard methods (SM) were used for the greater during O3/H2O2 than O3 when equivalent O3 dosages determination of total organic carbon (TOC) (SM 5310B), UV were applied. For example, 71% pCBA depletion occurred absorbance at 254 nm (SM 5910), total coliform (SM 9221B), fecal during O3/H2O2 versus 48% during O3 when both processes coliform (SM 9221E), and color (SM 2120C) (APHA et al., 1998). employed an O3 dose of approximately 5 mg/L. After 10 min of AOC samples were preserved through pasteurization and reaction, pCBA depletion was within 7–10% during O3 and O3/ measured through a bioassay using Pseudomonas fluorescens H2O2 when similar O3 dosages were applied. During first- d strain P-17 and Spirillum strain NOX according to SM 9217. Seven order O3 decay, OH formation continued at a slower rate, aldehydes (acetaldehyde, butanal, formaldehyde, glyoxal, m- resulting in comparable dOH exposures. These results agree glyoxal, pentanal and propanal) were analyzed by gas chroma- with Acero and von Gunten (2001), who demonstrated that d tography with electron capture detection (GC/ECD) according to the overall OH exposure was similar between O3 and O3/ SM 6252. Six carboxylic acids (acetate, formate, ketomalonate, H2O2 in surface waters (DOC ¼ 3.2 mg/L), and that H2O2 d oxalate, propionate and pyruvate) were measured by ion addition accelerates the rate of O3 decomposition into OH. chromatograph by previously established methods (Kuo, 1998; During pilot-scale experiments, pCBA depletion was similar Randtke, 2001). TOX concentrations were analyzed by EPA to bench-scale results as shown in Table 2. When comparing Method 9020B. Bromide and bromate concentrations were O3 and O3/H2O2 at similar dosages, pCBA was depleted 17–18% analyzed by ion-chromatography with inductively coupled more with O3/H2O2 than with O3 during the first 2 min. After plasma-mass spectroscopy detection (IC-ICP/MS)(Quinones et 6 min of contact time, pCBA depletion was again within 10%, al., 2006). Quantification of pCBA was achieved using high- coinciding with bench-scale results. These results confirmed pressure liquid chromatography (HPLC) equipped with an RP- bench-scale findings indicating that the initial dOH exposure

C18 column with a 45/55 mixture of 10 mM H3PO4 pH 2/MeOH as (timeo30 s) was greater during O3/H2O2 than O3; although, mobile phase and UV detection at 234 nm. The detection limit overall dOH exposure (time46 min) was similar for both was estimated at 0.02 mM(3.1mg/L). processes when applying similar O3 dosages.

3.3. Color and UV absorbance 3. Results and discussion

O3 and O3/H2O2 reduced color and UV absorbance at 254 nm

3.1. O3 decomposition (UV254) as shown in Table 3. Color was reduced from 24 to 5–8

units by O3 and to 7 units by O3/H2O2. Color reduction can be

Initial O3 demand and decay from bench-scale tests are important for public perception during water reuse or aquifer shown in Fig. 2. Results demonstrate minimal seasonal storage and recovery projects. UV254 and SUVA gradually changes in O3 decomposition, which was expected due to decreased as the O3 dose increased during O3 and O3/H2O2. similar TOC concentrations. The instantaneous O3 demand The decrease in SUVA demonstrates that the aromatic carbon

(IOD) was measured by the difference between the trans- content decreases as the O3 dose increases (Weishaar et al., ferred dose and the dissolved O3 residual after 30 s. Results 2003). These changes in color, UV254, and SUVA offer waste- showed the IOD varied between 2 and 4 mg/L depending on water utilities a few potential surrogates to monitor disinfec- the transferred dose. The IOD phase of ozonation has shown tion or trace contaminant oxidation with or without a similarities to O3-based AOPs, resulting in dOH exposure measurable O3 residual. (Buffle et al., 2006b).

After the IOD phase was complete, dissolved O3 residual 3.4. Coliform disinfection decayed according to first-order rate kinetics, as expected

(inset of Fig. 2). When targeting disinfection in drinking water Coliform disinfection was evaluated using both O3 and O3/ applications, a measurable dissolved O3 residual is required to H2O2 to determine the conditions required to achieve less meet CT guidelines established by the USEPA and assist with than 200 fecal coliforms per 100 mL to comply with discharge

O3 process control. Therefore, pilot-scale wastewater experi- regulations. Results shown in Table 3 indicate that an O3 ments targeted CT values up to 9.90 mg-min/L to evaluate dosage of 2.1 mg/L can achieve effective coliform disinfection wide range of O3 exposures (Table 2). in the absence of a stable ozone residual. These results

During bench-scale O3/H2O2 experiments, the O3 residual coincide with pilot-scale results reported by Janex et al. (2000), rapidly decayed within 1.5 min of reaction time. During pilot- who showed that 2-logs of fecal coliform inactivation could be scale O3/H2O2 experiments, detection of residual O3 after the achieved in wastewater before the IOD was met. Significant initial 2-min mass transfer period indicated that the O3/H2O2 oxidation of trace contaminants was also observed during the reaction was incomplete and a brief period of O3 oxidation IOD phase as shown in Snyder et al. (2006).However,O3 occurred. These results also show that when O3 dosages dosages exceeding the IOD are required to maintain proper ARTICLE IN PRESS

WATER RESEARCH 41 (2007) 1481– 1490 1485

8 0 O3=11.1 mg/L (Jan) O3=6.9 mg/L (June) 7 O3=4.3 mg/L (June) -1

O3=4.7 mg/L (Jan) ]o) 3 O3=5.1 mg/L, H2O2=2.5 mg/L (Jan) -2

6 O =11.2 mg/L, H O =5 mg/L (Jan) ]/[O

3 2 2 3 -3

5 ln ([O -4

-5 4 0246810121416 Time (min) 3

2 Dissolved Ozone Residual (mg/L)

1

0 0 2 4 6 8 10121416 Time (min)

Fig. 2 – Dissolved O3 residual decay curves during bench-scale experiments. The inset presents the data to show first-order rate kinetics.

Table 2 – O3 residual decay and pCBA reduction during pilot scale experiments

1 1 1 Date Transferred O3 dose (mg L )H2O2 dose (mg L ) Ozone residual Ozone CT (mg-min L ) pCBA reduction (mg/L) (%)

2min 6min 2min 6min

Jun-05 4.9 — 1.55 0.12 1.52 —— 7.3 — 2.71 0.54 4.20 —— 8.7 — 3.43 0.96 6.25 ——

Jan-06 2.1 — 0.09 o 0.05 — 15 — 2.1 1.0 o 0.05 o 0.05 ——— 3.6 — 0.83 0.05 0.74 48 55 3.6 2.0 0.36 o 0.05 — 65 — 7.0 — 3.75 1.39 9.90 80 92 7.1 3.5 1.72 o 0.05 — 98 499

disinfection levels during periodic demand fluctuations and above 3.1 mg/L, a linear relationship was observed between

flow rate changes (Rakness et al., 1993). bromate formation and dose. At O3 dosages above 4.5 mg/L,

The addition of H2O2 decreased disinfection efficiency. bromate formation exceeded the 10 mg/L drinking water MCL;

When O3/H2O2 was applied using O3 dosages near the IOD however, there are no US regulations on bromate in waste-

(2.1 mg/L O3 with 1.0 mg/L H2O2), there was no measurable water effluents. Additional research is needed regarding the reduction in total coliform. Less efficient disinfection by O3/ fate of bromate in receiving waters to determine if detection

H2O2 as compared to O3 can be expected due to a lack of O3 is expected in drinking water supplies influenced by waste- residual and the quantities of available dOH scavengers in water. wastewater, specifically DOC and alkalinity (Acero and von During O3/H2O2 experiments, bromate formation was not

Gunten, 2001; von Gunten, 2003). observed at lower O3/H2O2 conditions likely due to the lack of

a stable O3 residual and the reduction of hypobromous acid by

3.5. Bromate formation excess H2O2 (von Gunten and Oliveras, 1998). However,

0.013–0.016 mg/L of bromate was formed at the highest O3

Bromate formation results during bench-scale and pilot-scale dose of 7.1 mg/L with 3.5 mg/L of H2O2. Bromate formation testing are shown in Fig. 4. Bromate formation was observed during this AOP condition was likely due to the presence of a after the IOD of the wastewater was exceeded. At O3 dosages high initial O3 residual after the first 2 min of mass transfer. ARTICLE IN PRESS

1486 WATER RESEARCH 41 (2007) 1481– 1490

1.0

0.9 O3=2.3 mg/L O3=2.3 mg/L, H2O2=2 mg/L 0.8 O3=4.7 mg/L O3=5.1 mg/L, H2O2=2.5 mg/L 0.7 O3=11.1 mg/L O3=11.2 mg/L, H2O2=5 mg/L

o 0.6

0.5

0.4 pCBA/pCBA

0.3

0.2

0.1

0.0 0241 3 56 79108 Time (min)

Fig. 3 – pCBA reduction during bench-scale O3 and O3/H2O2 experiments.

Table 3 – Coliform disinfection using conventional O3 and O3/H2O2

Test O3 dosage H2O2 dose Reaction Color TOC UV254 SUVA Total Fecal date (mg L1) (mg L1) time (min) (units) (mg L1) (cm1) (L mg1 m1) coliform (#/ coliform (#/ 100 mL) 100 mL)

Jun- —— ——7.2 0.128 1.78 6000 2600 05 4.9 — 18 — 7.1 0.067 0.95 o2 o2 7.3 — 18 — 7.0 0.060 0.86 o2 o2 8.7 — 18 — 6.9 0.055 0.80 o2 o2

Jan- —— —24 7.2 0.119 1.66 41600 41600 06 2.1 — 6 — 7.2 0.087 1.20 30 8 2.1 1.0 10 — 7.3 0.089 1.22 41600 500 3.6 — 18 8 7.1 0.075 1.06 o2 o2 3.6 2.0 10 7 6.9 0.071 1.04 80 14 7.0 — 18 5 7.0 0.058 0.83 12 4 7.1 3.5 10 7 6.9 0.047 0.68 4 2

3.6. Organic DBP formation not detected during any of the testing conditions. During O3/

H2O2 conditions, propanal was detected at concentrations

Organic DBP formation during O3 and O3/H2O2 pilot-scale between 2.9 and 3.9 mg/L. Formation of the remaining four testing is shown in Table 4. Results showed that O3/H2O2 aldehydes is shown in Fig. 5. Formaldehyde was formed at formed 5–52% more AOC than O3. AOC comprised about greater concentrations than the other aldehydes. Since O3/

14–16% of the TOC after O3 and 15–25% of the TOC after O3/ H2O2 produced greater concentrations of formaldehyde than

H2O2. AOC was further investigated by identifying aldehyde O3, formation can be attributed mostly to dOH exposure. and carboxylic acid formation. Minimal data exist regarding However, the results in Fig. 3 indicated that the overall dOH the natural attenuation of these organic DBPs in receiving exposures were within 10% after O3 and O3/H2O2. Therefore, it waters. can be deduced that the majority of formaldehyde formation

Total aldehyde concentrations showed that O3/H2O2 formed occurred during the greater initial dOH exposure within the

31–47% more aldehydes than O3. Butanal and pentanal were IOD phase (timeo30 s) of O3 decomposition. ARTICLE IN PRESS

WATER RESEARCH 41 (2007) 1481– 1490 1487

70

60 BTPP (June 2005) BS (Jan 2006) 50 BTTP (Jan 2006)

40

30 y = 8.2785x - 25.918 R2 = 0.8959 Bromate (µg/L)

20

10

0 0 246810 12 Ozone Dose (mg/L)

Fig. 4 – Relationship between bromate formation and O3 dose during bench scale (BS) and bench-top pilot plant (BTPP) experiments.

Table 4 – Summary of biodegradable byproducts formed during pilot scale ozonation

O3 Dosage H2O2 dose Reaction time AOC AOC/TOC Total aldehydes Total carboxylic acids (mg L1) (mg L1) (min) (mgL1) (%) (mgL1) (mgL1)

———331 4.6 26 90 2.1 — 6 —— — 376 2.1 1.0 10 —— — 421 3.6 — 18 971 14 114 623 3.6 2.0 10 1019 15 149 737 7.0 — 18 1149 16 211 937 7.1 3.5 10 1752 25 310 1340

Carboxylic acids were formed at much greater concentra- 3.7. TOX tions than aldehydes. Total carboxylic acid concentrations showed that O3/H2O2 formed 12–43% more carboxylic acids Simulated distribution system (SDS) testing was con- than O3. Of the six carboxylic acids analyzed, ketomalonate ducted by adding free chlorine to ozonated and non-ozonated and propionate were not detected under any testing condi- samples and holding for 24 h as shown in Table 5. Each O3 tions. Formation of the four remaining carboxylic acids is condition reduced the 1-h chlorine demand by at least shown in Figs. 6 and 7. Formate, oxalate, and acetate 1 mg/L and reduced TOX formation by 20%. The chlorine formation appear to be related to the O3 dose. Formate residual information shows the chlorine exposure accounted for 55–60% of all carboxylic acid formation during (CT) was higher in the ozonated samples. As a result, both O3 and O3/H2O2, and may be the result of formaldehyde even greater TOX reduction can be expected when oxidation by O3 (Can and Gurol, 2003). Results also indicate lowering the chlorine dosages to target a constant chlorine that O3 may have oxidized pyruvate, while O3/H2O2 (dOH CT. A TOX reduction was expected due to the lower SUVA exposure) may have formed pyruvate. Due to the detection values after ozonation as shown in Table 3. SUVA has been limit of 20 mg/L, it is difficult to determine whether these shown to be a good indicator of TOX formation during the relationships are valid or due to analytical variability. These chlorination of surface waters (Korshin et al., 1997). TOX results coincide with aldehyde results indicating that car- results from O3/H2O2 testing are not shown due to variable boxylic acid formation occurs during the greater initial dOH chlorine demand observed, likely due to the presence of exposure within the IOD phase of ozonation. excess H2O2. ARTICLE IN PRESS

1488 WATER RESEARCH 41 (2007) 1481– 1490

160 146 Formaldehyde 140 Glyoxal M-Glyoxal 120 Acetaldehyde

100 95

80 71 65 59 60 Concentration (µg/L) 44 45 42 40 29 31 28 29 31 23 21 18 21 20 4 3 3 0 Tertiary O3=3.6 mg/L O3=3.6 mg/L O3=7.0 mg/L O3=7.1 mg/L Effluent H2O2=2.0 mg/L H2O2=3.5 mg/L

Fig. 5 – Aldehyde formation during O3 and O3/H2O2 pilot-scale experiments.

600

526 Formate 500 Oxalate Acetate Pyruvate 400 377

300 274

209

Concentration (µg/L) 200 136 116 100 100 84 53 45 41 40 30 30 <20 <20 0 Tertiary O3=2.1 mg/L O3=3.6 mg/L O3=7.0 mg/L Effluent

Fig. 6 – Carboxylic acid formation during O3 pilot-scale experiments.

did not enhance the overall dOH exposure during 4. Conclusions ozonation.

Coliform disinfection was more effective using O3 as

O3/H2O2 resulted in 13–23% greater dOH exposure than O3 compared to O3/H2O2 when similar O3 dosages are applied.

during the IOD phase (30 s). However, the overall dOH Sufficient disinfection was achieved with O3 when apply-

exposure was within 7–10% after the first-order O3 ing dosages at or above the IOD. O3 dosages above the IOD

decay was completed. These results show that H2O2 are recommended for process control during full-scale

accelerated the rate of O3 decomposition into dOH, but application. ARTICLE IN PRESS

WATER RESEARCH 41 (2007) 1481– 1490 1489

900

800 Formate 779 Oxalate 700 Acetate Pyruvate 600

500 417 400 311

Concentration (µg/L) 300 239 197 200 160 111 89 100 52 50 53 27 27 40 <20 <20 0 Tertiary O3=2.1 mg/L O3=3.6 mg/L O3=7.1 mg/L Effluent H2O2=1.0 mg/L H2O2=2.0 mg/L H2O2=3.5 mg/L

Fig. 7 – Carboxylic acid formation during O3/H2O2 pilot scale experiments.

similar after 10 min of reaction, the majority of organic Table 5 – Summary of free chlorine residual decay and DBP formation likely occurred during the initial dOH

subsequent TOX formation exposure within the IOD phase (timeo30 s) of O3 decom- position. O reduced TOX potential by 20% after chlorination. TOX O3 dose Free Free chlorine residual TOX 3 (mg L1) chlorine (mg/L) (mgL1) reduction would benefit water reuse systems by reducing dose trihalomethane and haloacetic acid formation. (mg L1) As a result of these findings, O3 may be preferred to O3/ 2 min 60 min 960 min H2O2 in wastewater applications. In addition, O3/H2O2 is

— 5 3.48 1.35 o 0.05 499 prone to leaving a H2O2 residual in water. If subsequent 2.1 5 4.32 2.55 0.53 420 chlorination or chloramination is required, the H2O2 will 3.6 5 4.60 2.66 0.54 384 quench these oxidants resulting in increased chlorine

7.0 5 4.40 2.82 0.47 424 dosages. O3 provided optimal coliform disinfection and trace contaminant removal (Snyder et al., 2006) while minimizing DBP formation.

Color was reduced from 24 units to 5–8 units by either O3

or O3/H2O2 at the dosages tested. Color reduction can be important for public perception for water reuse or aquifer Acknowledgements storage and recovery projects. UV254 and SUVA gradually decreased as the O dose increased. These changes offer 3 The authors thank Oscar Quin˜ ones and Janie Holady-Zeigler wastewater utilities a few potential surrogates to monitor at the Southern Nevada Water Authority’s Water Quality disinfection or trace contaminant oxidation with or with- Research and Development Division for their valuable analy- out a measurable O3 residual. tical support and assistance conducting bench-scale and Bromate formation occurred when applying O3 dosages pilot-scale experiments. The authors also thank Devon greater than the IOD. The lack of O3 residual coupled with Morgan at the Clark County Water Reclamation District for the presence of excess H2O2 inhibited bromate formation providing the wastewater and performing the coliform during the O3/H2O2 process. While not regulated in analysis. wastewater effluents, bromate could be a concern during

indirect potable reuse applications, or during the treat- REFERENCES ment of surface water heavily influenced by wastewater effluent. O /H O produced greater concentrations of AOC (5–52%), 3 2 2 Acero, J.L., von Gunten, U., 2001. Characterization of oxidation aldehydes (31–47%), and carboxylic acids (12–43%) than O3 processes: ozonation and the AOP O3/H2O2. J. Am. Water at similar dosages, indicating their formation is largely Works Assoc. 93 (10), 90–100. dependent upon dOH exposure. Since dOH exposure was APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater. American Public Health Associa- ARTICLE IN PRESS

1490 WATER RESEARCH 41 (2007) 1481– 1490

tion, American Water Works Association, and Water Environ- Rakness, K.L., Corsaro, K.M., Hale, G., Blank, B.D., 1993. Waste- ment Federation, Washington, DC. water disinfection with ozone—process control and operating Bader, H., Hoigne, J., 1982. Determination of ozone in water by the results. Ozone: Sci. Eng. 15 (6), 497–514. indigo method; a submitted standard method. Ozone: Sci. Eng. Rakness, K.L., Najm, I., Elovitz, M., Rexing, D., Via, S., 2005. 4 (4), 169–176. Cryptosporidium log-inactivation with ozone using effluent

Buffle, M.-O., Schumacher, J., Meylan, S., Jekel, M., von Gunten, U., CT10, geometric mean CT10, extended integrated CT10 and 2006a. Ozonation and advanced oxidation of wastewater: extended CSTR calculations. Ozone: Sci. Eng. 27 (5), 335–350.

effect of O3 Dose, pH, DOM and HO scavengers on ozone Randtke, S.J., 2001. An improved ion chromatographic method for decomposition and HO generation. Ozone: Sci. Eng. 28 (4), carboxylic acids in drinking water. In: Proceedings of the 247–259. American Water Works Association Water Quality Technology Buffle, M.-O., Schumacher, J., Salhi, E., Jekel, M., von Gunten, U., Conference, Nashville, TN. 2006b. Measurement of the initial phase of ozone decom- Rice, R.G., Evison, L.M., Robson, C.M., 1981. Ozone disinfection of position in water and wastewater by means of a continuous municipal wastewater—current state-of-the-art. Ozone: Sci. quench-flow system: application to disinfection and pharma- Eng. 3 (4), 239–272. ceutical oxidation. Water Res. 40 (2006), 1884–1894. Richardson, S.D., Thruston Jr., A.D., Caughran, T.V., Chen, P.H., Can, Z.S., Gurol, M., 2003. Formaldehyde formation during Collette, T.W., Floyd, T.L., Schenck, K.M., Lykins Jr., B.W., 1999. ozonation of drinking water. Ozone: Sci. Eng. 25 (1), 41–51. Identification of new ozone disinfection byproducts in drink- Elovitz, M.S., von Gunten, U., 1999. Hydroxyl radical/ozone ratios ing water. Environ. Sci. Technol. 33 (19), 3368–3377.

during ozonation processes. I. The RCT concept. Ozone: Sci. Ryu, H., Alum, A., Abbaszadegan, M., 2005. Microbial character- Eng. 21 (3), 239–260. ization and population changes in nonpotable reclaimed Huang, W.-J., Fang, G.-C., Wang, C.-C., 2005. The determination water distribution systems. Environ. Sci. Technol. 39 (22), and fate of disinfection by-products from ozonation of 8600–8605. polluted raw water. Sci. Total Environ. 345 (2005), 261–272. Shon, H.K., Vigneswaran, S., Snyder, S.A., 2006. Effluent organic Huber, M.M., Canonica, S., Park, G.-Y., von Gunten, U., 2003. matter (EfOM) in wastewater: constituents, effects, and Oxidation of pharmaceuticals during ozonation and advanced treatment. Crit. Rev. Environ. Sci. Technol. 36, 327–374. oxidation processes. Environ. Sci. Technol. 37 (5), 1016–1024. Sirivedhin, T., Gray, K.A., 2005. Comparison of the disinfection by- Huber, M.M., Gobel, A., Joss, A., Hermann, N., Loffler, D., Mcardell, product formation potentials between a wastewater effluent C.S., Ried, A., Siegrist, H., Ternes, T.A., von Gunten, U., 2005. and surface waters. Water Res. 39 (2005), 1025–1036. Oxidation of pharmaceuticals during ozonation of municipal Snyder, S.A., Wert, E.C., Rexing, D.J., Zegers, R.E., Drury, D.D., 2006. wastewater effluents: a pilot study. Environ. Sci. Technol. 39 Ozone oxidation of endocrine disruptors and pharmaceuticals (11), 4290–4299. in surface water and wastewater. Ozone: Sci. Eng. 28 (6), Huck, P.M., Fedorak, P.M., Anderson, W.B., 1991. Formation and 445–460. removal of assimilable organic carbon during biological USEPA, 1991. Guidance Manual for Compliance with the Filtration treatment. J. Am. Water Works Assoc. 83 (12), 69–80. and Disinfection Requirements for Public Water Systems Janex, M.L., Savoye, P., Roustan, M., Do-Quang, Z., Laine, J.M., Using Surface Water Sources. Lazarova, V., 2000. Wastewater disinfection by ozone: influ- USEPA, 1998. National primary drinking water regulations: disin- ence of water quality and kinetics modeling. Ozone: Sci. Eng. fectants and disinfection byproducts final rule. Fed. Reg. 63 22 (2), 113–121. (241), 69389. Korshin, G.V., Li, C.-W., Benjamin, M.M., 1997. The decrease of UV USEPA, 2003. Long Term 2 Enhanced Surface Water Treatment absorbance as an indicator of TOX formation. Water Res. 31 Rule Toolbox Guidance Manual (DRAFT). EPA 815-D-03-009. (4), 946–949. von Gunten, U., 2003. Ozonation of drinking water: part II. Krasner, S.W., Sclimenti, M.J., Coffey, B.M., 1993. Testing biologi- Disinfection and by-product formation in presence of bro- cally active filters from removing aldehydes formed during mide, iodide or chlorine. Water Res. 37 (2003), 1469–1487. ozonation. J. Am. Water Works Assoc. 85 (5), 62–71. von Gunten, U., Oliveras, Y., 1998. Advanced oxidation of Kuo, C.-Y., 1998. Improved application of ion chromatographic bromide-containing waters: bromate formation mechanisms. determination of carboxylic acids in ozonated drinking water. Environ. Sci. Technol. 32 (1), 63–70. J. Chromatogr. A 804 (1), 265–272. Weishaar, J.L., Aiken, G.R., Bergamschi, B.A., Fram, M.S., Fujii, R., LeChevallier, M.W., Becker, W.C., Schorr, P., Lee, R.G., 1992. Mopper, K., 2003. Evaluation of specific ultraviolet absorbance Evaluating the performance of biologically active rapid as an indicator of the chemical composition and reactivity of filters. J. Am. Water Works Assoc. 84 (4), 136–140. dissolved organic carbon. Environ. Sci. Technol. 37 (20), Quinones, O., Snyder, S.A., Cotruvo, J.A., Fisher, J.W., 2006. 4702–4708. Analysis of bromate and bromide in blood. Toxicology 221 Zwiener, C., Frimmel, F.H., 2000. Oxidative treatment of pharma- (2), 229–234. ceuticals in water. Water Res. 34 (6), 1881–1885.