- Email: [email protected]
Solar photolysis kinetics of disinfection byproducts
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
Solar photolysis kinetics of disinfection byproducts Baiyang Chen a,*, Wontae Lee b, Paul K. Westerhoff c, Stuart W. Krasner d, Pierre Herckes e a
Chinese Environmental Scholars and Professionals Network, 11900 Stonehollow Drive, Apartment 338, Austin, TX 78758, USA HDR Engineering, Inc., 3200 East Camelback Road, Suite 350, Phoenix, AZ 85018, USA c Arizona State University, School of Sustainable Engineering and The Built Environment, Engineering Center (G-Wing), Room ECG-252, Tempe, AZ 85287-5306, USA d Metropolitan Water District of Southern California, Water Quality Laboratory, 700 Moreno Avenue, La Verne, CA 91750, USA e Arizona State University, Department of Chemistry and Biochemistry, Tempe, AZ 85287-1604, USA b
article info
abstract
Article history:
Disinfection byproducts (DBPs) discharged from wastewater treatment plants may impair
Received 17 December 2009
aquatic ecosystems and downstream drinking-water quality. Sunlight photolysis, as one
Received in revised form
process by which DBPs may dissipate in the receiving surface water, was investigated.
8 March 2010
Outdoor natural sunlight experiments were conducted in water for a series of carbona-
Accepted 8 March 2010
ceous DBPs (trihalomethanes, haloacetic acids, halopropanones, and haloacetaldehydes)
Available online 17 March 2010
and nitrogenous DBPs (nitrosamines, halonitromethanes, and haloacetonitriles). Their pseudo-first-order rate constants for photolytic degradation were then used to calibrate
Keywords:
quantitative structureeactivity relationship (QSAR) parameters, which, in return, predicted
Photolysis
the photolysis potentials of other DBPs or related compounds. Nitrogenous DBPs were
QSAR
found to be more susceptible to solar irradiation than carbonaceous DBPs, with general
DBP
rankings for the functional groups as follows: N-nitroso (N-NO) > nitro (NO2) > nitrile
NDMA
(C^N) > carbonyl (C]O) > carboxyl (COOH). Compounds containing a high degree of
THM
halogenation (e.g., three halogens) were usually less stable than less halogenated species
HAA
(e.g., those with two halogens). Bromine- or iodine-substituted species were more photosensitive than chlorinated analogs. While most bromine- and chlorine-containing trihalomethanes and haloacetic acids persisted over the 6-h test, nearly complete removal (>99%) of nitrosamines occurred within 1 h of sunlight exposure. Indoor laboratory experiments using simulated sunlight demonstrated that the degradation of nitrosamines was w50% slower when organic matter was present, and w11% slower in non-filtered water than in filtered water. ª 2010 Elsevier Ltd. All rights reserved.
1.
Introduction
In addition to water recycling and reclamation programs, unintentional reuse of treated wastewater has occurred over the past few decades and will likely increase in the future
as upstream wastewater treatment plants (WWTPs) discharge effluents into rivers or lakes that serve as downstream drinking-water resources. Drought and competing in-stream demands may result in substantial contribution of treated wastewater in water bodies (Krasner et al., 2008). Attention
* Corresponding author. Tel.: þ1 480 840 4647. E-mail addresses: [email protected] (B. Chen), [email protected] (W. Lee), [email protected] (P.K. Westerhoff), [email protected] (S.W. Krasner), [email protected] (P. Herckes). 0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2010.03.014
3402
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
has been drawn to the impact of WWTP discharges as a source of multiple contaminants, such as pharmaceuticals, to aquatic life and downstream drinking-water sources. Due to the common disinfection practice at U.S. WWTPs of adding chlorine, disinfection byproducts (DBPs) are likely to be present in WWTP discharges. Wastewater-derived DBPs may pose risks to downstream aquatic organisms (Mizgireuv et al., 2004) and human health. In a U.S. survey, for WWTPs that achieved breakpoint chlorination (i.e., free chlorine residual), the median concentrations of trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs), and trihalogenated acetaldehydes were 57, 70, 16, and 16 mg/L, respectively (Krasner et al., 2009). For WWTPs without breakpoint chlorination (i.e., chloramine residual), a median of 11 ng/L (75th percentile ¼ 24 ng/L) of N-nitrosodimethylamine (NDMA) was detected (Krasner et al., 2009). Chlorination of effluent organic matter (EfOM) in an earlier study also resulted in the formation of chloral hydrate (i.e., trichloroacetaldehyde) and dichloroacetonitrile (Trehy et al., 1986). As WWTP effluents contain relatively high levels of bromide and iodide due to anthropogenic activities (Krasner et al., 2008), certain brominated and potentially iodinated DBPs may form too. Some nitrogenous DBPs (including certain HANs, halonitromethanes [HNMs]) are more cytotoxic and genotoxic than the regulated carbonaceous DBPs (i.e., THMs, HAAs), and their toxicity was impacted by halogen substitution type: iodine > bromine > chlorine (Plewa and Wagner, 2009). The U.S. Environmental Protection Agency’s (USEPA’s) Integrated Risk Information System (IRIS) database indicates for NDMA and seven other nitrosamines that concentrations of low ng/L level in water are associated with a 106 lifetime cancer risk. Therefore, the treatment, fate, and transport of wastewaterderived DBPs in water resources require study. In conjunction with sorption, biodegradation, volatilization, hydrolysis, and other biogeochemical processes, sunlight photolysis may reduce DBP concentrations to varying degrees in surface waters (Table 1). For example, dichloroacetonitrile spiked into drinking water was photodegradable (Lekkas and Nikolaou, 2004), and chloropicrin (i.e., trichloronitromethane [TCNM]) in the atmosphere was degraded under sunlight conditions (Allston et al., 1978). HAAs underwent varying levels of photolysis (Lifongo et al., 2004), and nitrosamines exhibited substantial photolytic losses under simulated sunlight conditions (Megan and Reinhard, 2007). However, the experimental conditions in these studies were different, which makes a comparison of the photolysis degradation of different groups of DBPs difficult. Quantitative structureeactivity relationship (QSAR) analysis is an empirical chemical property estimation method that can predict the behaviors of one chemical from other structurally similar chemicals. The method was developed by Hammett (1935), explored by Hansch et al. (1995), and widely applied in many areas, including the prediction of toxicity testing and pharmaceutical behavior (Hansch et al., 1995), the photolysis of polycyclic aromatic hydrocarbons (Chen et al., 2001) and haloaromatics (Peijnenburg et al., 1993), and the hydrolysis of trihaloacetic acids (Zhang and Minear, 2002). The QSAR application is based upon a key assumption that chemicals with similar structures behave similarly and their activity differences are attributed only to the type, number,
steric, and electronic effect of their functional group(s). Because DBPs in one group share similar chemical structures, the QSAR technique should be applicable to DBP photolysis. The purpose of this paper was to investigate the photolysis of several groups of DBPs and related compounds under uniform solar irradiation conditions. The pseudo-first-order degradation rate constants were obtained from experiments and subsequently used to calibrate the QSAR models. In turn, the calibrated QSAR models were used to predict the photolytic degradation of other DBPs or related compounds, such as iodinated DBPs, which have not been studied experimentally. Although the reactive mechanism and pathway of each DBP is uncertain and many other factors, such as receiving waterbody geometry (width, depth of the photic zone), flowrate, compound sensitivity, microbial activity, solar light spectra, and meteorological conditions are also critical in understanding the photolytic degradation of DBPs (Chen et al., 2008), the outcome of this study may serve as a basis for further and more thorough exploration of the role of photolysis in DBP removal.
2.
Materials and methods
2.1.
Natural sunlight (outdoor) experiment
Sunlight photolysis of DBPs was evaluated in 9-mL quartz test tubes (outside diameter ¼ 13 mm). Concentrated DBP solutions were spiked into 200 mL of pH-buffered organic-pure water (pH w7.2; phosphate buffer) in volumetric flasks to result in concentrations of w500e1000 mg/L of each halogenated DBP and w10,000 ng/L of each nitrosamine and were distributed evenly to multiple test tubes. Sample vials were capped to minimize volatilization or evaporation, and placed on an inclined platform on top of the laboratory roof in Tempe, Arizona, USA (33 250 12.3500 N, 111 550 55.1100 W). Samples were partially air cooled during the tests by a fan, as the surrounding air temperature was w30 C (Arizona Meteorological Network, http://ag.arizona.edu/azmet). Energy intensities were monitored on-site by a photometer (IL 1700, International Light Technologies, Inc., Massachusetts, USA). The solar spectrum was also extracted from a model named SPCTRAL2 (http://rredc.nrel.gov/solar/models/spectral/ SPCTRAL2). The irradiation intensities appeared similar and remained constant in the range of 1150e1300 W/m2 between 11 am and 3 pm (<10% difference). Halogenated DBP and nitrosamine samples were exposed for 0, 0.5, 1.5, 3, and 6 h from 11 am to 5 pm on May 20, 2005 for the first run of the experiment. On July 25, 2005, nitrosamine photolysis experiments were repeated but for a shorter exposure time (i.e., 0, 1, 3, 5, 10, 20, 30, and 60 min). On August 10, 2005, a second run was conducted for up to 8 h for the THMs and HANs. A 6- or 8-h time period was chosen in order to have a nearly constant solar irradiation for each experiment. The total solar radiations of the three dates were recorded as 696, 681, and 405.7 langleys for May 20, July 25, and August 10, 2005, respectively, by Arizona Meteorological Network Mesa Station (http://ag.arizona.edu/azmet). The average air temperatures of the three dates however were similar, being 31.7, 33.3, and 33.9 C, respectively.
3403
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Table 1 e Literature review of photolysis kinetics of halogenated chemicals. DBP or related chemical
Conditions
Pseudo-first-order rate constant (kP, h1) and/or quantum yield (F)
References
Halomethanes CH2I2 CH2IBr CH2ICl
pH ¼ 7.65; deionized water, salt water, seawater; sunlight; 55 N, 1 W
4.32, F ¼ 0.31 1.45 101, F ¼ 0.29 7.81 102, F ¼ 0.13
Jones and Carpenter, 2005
CHCl3
River/lake; 20 C
1.81 104
SCDM, 1997a
THMs
pH ¼ 2.5e10; UV ¼ 112e400 nm; 0.1% H2O2
CHBr3 > CHClBr2 > CHCl2Br > CHCl3, up to 92% removal
Rudra et al., 2005
CHBr3
Air; UV ¼ 266e324 nm; 31 C
F ¼ 0.76
Bayes et al., 2003
CHBr3
Air; UV ¼ 248 nm
F ¼ 0.2e0.3
CH3Br
Neutral pH; sunlight, UV at 254 nm
Nil under sunlight, 6.12 10 under UV at 254 nm
CHCl3 CHCl2Br CHClBr2 CHBr3
pH ¼ 7.5; <30 min; UV at 253.7 nm; 20 C
<5%
Haloacetic acids Br2CH(COOH) BrClCH(COOH) Cl3CH(COOH) Br3CH(COOH)
Zou et al., 2004 3
Castro and Belser, 1981
Nicole et al., 1991
F ¼ 0.43 0.1
Mercury lamps (l > 400 nm); 15 C
6.91 104 1.84 104 2.30 104 3.00 103
Lifongo et al., 2004
Halonitromethanes Cl3C(NO2)
Neutral pH; sunlight, UV at 254 nm; 20 C,
9.61 103 (sunlight) 5.00 101 (UV)
Castro and Belser, 1981
Cl3C(NO2)
Air; UV ¼ 190e400 nm; 295 K
2.05 101 to 2.88
Allston et al., 1978
Nitrosamines N-Nitrosodimethylamine
pH ¼ 3 and 7; 1 mM; UV ¼ 200e300 nm
156, F ¼ 0.3
Stefan and Bolton, 2002
N-Nitrosodimethylamine N-Nitrosodiethylamine N-Nitrosodipropylamine N-Nitrosodibutylamine N-Nitrosopiperidine
River/lake; 20 C
6.88 101 8.75 102 6.88 101 8.75 102 3.07 103
SCDM, 1997a
N-Nitrosodimethylamine
Neutral
F ¼ 0.11 0.03
Ho et al., 1996
2.40 0.12, F ¼ 0.41 2.94 0.18, F ¼ 0.61 2.70 0.24, F ¼ 0.43 3.00 0.24, F ¼ 0.46 2.88 0.06, F ¼ 0.52 3.30 0.36, F ¼ 0.55 3.42 0.12, F ¼ 0.51
Megan and Reinhard, 2007
N-Nitrosodimethylamine N-Nitrosomethylethylamine N-Nitrosodiethylamine N-Nitrosodipropylamine N-Nitrosodibutylamine N-Nitrosopyrrolidine N-Nitrosopiperidine
2
765 W/m in Milli-Q water (pH ¼ 6); 100 mg/L except for NPIP (1000 mg/L); simulated sunlight lamp
a SCDM: Superfund Chemical Data Matrix program, 1997 version, from USEPA.
After each time interval, samples were immediately preserved as appropriate (i.e., pH lowered to w3e4 for samples of THMs, HANs, halopropanones, HNMs, and haloacetaldehydes), refrigerated, and shipped to the laboratory for DBP analysis (Krasner et al., 2009). The photolysis samples (9-mL) were diluted to a 200- to 500-mL volume before the analyses, so that there would be sufficient volume for the analytical methods used, which diluted the halogenated DBPs (in the absence of photolytic degradation) to w20 mg/L and the nitrosamines to w180 ng/L. These concentration levels were selected to be able to observe losses of up to w90e99%. Briefly,
the THMs, HANs, halopropanones, and chloropicrin were analyzed by liquid/liquid extraction (LLE) and gas chromatography (GC) with an electron capture detector (ECD); other HNMs and haloacetaldehydes were analyzed by either LLEeGC/ECD or solid-phase extraction and GC/mass spectrometry (MS); the HAAs were analyzed by LLE, derivatization with acidic methanol, and GC/ECD; and the nitrosamines were analyzed by micro LLE and chemical ionization GC/MS. Throughout the photolysis study, two sets of control samples were employed, which included “control 1” wrapped in aluminum foil and stored at 4 C in a refrigerator (i.e., no light
3404
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
or heat exposure), and “control 2” samples wrapped in aluminum foil but placed in a box near the samples for the entire hold time (i.e., same outdoor conditions and handling as that of the testing samples, except for no light exposure). Control 1 samples had minimal effects from other loss mechanisms, as they were immediately preserved (e.g., pH lowered to w3e4 for samples of THMs, HANs, halopropanones, HNMs, and haloacetaldehydes), whereas control 2 samples could experience other loss mechanisms such as hydrolysis and/or volatilization during the outdoor holding period. The pseudo-first-order rate constant of photolysis (kP) of the DBPs was calculated, which was adjusted for any other loss mechanism observed in control 2. The degradation kinetics were assumed to be pseudo-first-order for all DBPs.
2.2.
Laboratory photolysis (indoor) experiment
In order to understand the effects of organic and particulate matter on the photolysis of nitrosamines, which were the most labile species studied, indoor laboratory experiments were conducted in different types of water. The sample was placed in a thermostated and jacketed reactor (100-mL capacity). The reactor was constantly stirred and was irradiated through a quartz window on top with simulated sunlight. The simulated light was provided by a 300-W xenon arc lamp (Spectraphysics Oriel, 91160A), and the output of the arc lamp was filtered through a standardized air mass 1.5 filter (AM 1.5 Global), which yielded a spectrum representative of the solar spectrum at ground level when the sun was at a zenith angle of 48.2 , and the lamp output power was 1325 W/m2. This experiment was different from the outdoor work in the following ways: (1) light intensity was higher and more consistent throughout the duration, (2) larger sample volumes were used in this setup (40 versus 9 mL), and (3) water temperature was low and maintained constant at room temperature (20 C). Three types of waters were spiked with eight nitrosamines: (1) organic-free water; (2) non-filtered grab samples from two WWTP effluents (effluent #1 had a turbidity of 1.5 NTU and effluent #2 had a turbidity of 8.8 NTU); and (3) two filtered (0.45-mm) WWTP effluents from (2). WWTP effluent #1 had a total organic carbon (TOC) concentration of 12.2 mg/L and ultraviolet absorbance at 254 nm (UVA254) of 0.18 cm1, whereas WWTP effluent #2 had a TOC of 16.5 mg/L and UVA254 of 0.16 cm1. UVevisible absorbance spectra were compared for organic-free water and the two types of effluents, both filtered and non-filtered. In the non-filtered samples, substantially more (>30%) light interception was observed at wavelengths greater than 350 nm, which means less photons can reach DBPs when particulate matter is present (Krasner et al., 2008).
2.3.
QSAR modeling
In addition to experimental approaches, a fundamental QSAR analysis equation (Eq. (1)) was employed to predict the pseudo-first-order rate constants of non-tested DBPs or related compounds with similar structure(s). The equation had been applied to predicting DBP hydrolysis in previous
research (Zhang and Minear, 2002), but had not been used for DBP photolysis analysis until now. logðkP Þ ¼ d$Es þ r$s þ c0
(1)
where kP is the pseudo-first-order observed photolysis (P) rate constant (k) of the target chemical of the group; Es is the steric effect constant of a substituent; s is the electronic effect constant of the substituent; d is the sensitivity coefficient of the steric effect of the substituent; r is the sensitivity coefficient of the electronic effect of the substituent; and c0 is an empirical constant (Hansch et al., 1995). Based upon previous studies, 0.50, 0.47, 0.42, and 0.38 were obtained as the s values (electronic constants) and 0.46, 0.96, 1.16, and 1.40 were obtained as the Es values (steric constants) for fluoride, chloride, bromide, and iodide, respectively (Hansch et al., 1995). The effects of multiple halides were assumed to be the sum of the effects of single halides, which follows an early study (Zhang and Minear, 2002). Accordingly, Eq. (1) became a triplex equation with three unknowns (d, r, and c0 ). Solving the triplex equations required three sets of inputs (i.e., kP, s, and Es) and once the constants d, r, and c0 values were obtained, Eq. (1) was, in turn, used to estimate the kinetics of other chemicals. During our experiments the oxidation mechanisms probably involved both direct photolysis and indirect radical reactions. Direct photolysis is a function of quantum yield. Rates of indirect reactions involve both the DBP chemical structure and radical concentrations. As such, the intent of this paper is not to relate fundamental chemical properties (e.g., chemical bond strength, thermodynamic properties) to the observed pseudofirst-order rate loss constants because multiple oxidation mechanisms may be involved. Efforts along these lines would require greater focus on a smaller number of DBPs. Instead, we hypothesize that within a specific DBP class the effects of halogenated substituents can be represented by Es, s, d, and r, collectively, and correlated with the observed kP. Note that DBPs can be chlorine-, bromine-, and/or iodine-substituted. Fluorinated compounds were also evaluated for completeness in the QSAR analysis, but are not DBPs. Moreover, not all of the potential iodinated analogs of the DBPs have been identified in water.
3.
Results and discussion
This paper presents both experimentally obtained and QSAR predictive data. The experimental conditions of this study are representative of a near-surface water exposed to intensive solar irradiation, such that the results of this study may reflect the photolysis degradation potentials of the DBPs and related compounds. The photolysis rate constants were estimated for a range of halogenated compounds (Table 2). The following discussion focuses primarily on the regulated and emerging DBPs of concern.
3.1.
Halomethanes (HMs)
Over the 6- or 8-h intensive solar exposure periods, individual THM species in control 2 samples exhibited losses of 5e11%
3405
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Table 2 e QSAR predictions of log kP of halogenated compoundsa under sunlight conditions in May 2005 in Tempe (AZ). Halomethane, log kP (h1)
Haloacetic acid, log kP (h1)
Haloacetaldehyde, log kP (h1)
Haloacetonitrile, log kP (h1)
Halonitromethane, log kP (h1)
Mono-halogenated compounds F Cl Br I
2.74 2.10 1.84 1.45
4.47 2.68 1.95 0.69
3.39 2.42 2.03 1.38
2.61 1.55 1.12 0.45
2.49 1.44 1.02 0.36
Dihalogenated compounds F2 FCl FBr FI Cl2 ClBr ClI Br2 BrI I2
2.95 2.30 2.05 1.66 1.65 1.40 1.01 1.14 0.75 0.36
6.73 4.94 4.21 2.95 3.15 2.42 1.16 1.68 0.43 0.82
4.35 3.38 2.99 2.34 2.41 2.02 1.37 1.63 0.98 0.33
3.27 2.21 1.79 1.11 1.15 0.72 0.05 0.30 0.37 1.04
3.04 1.99 1.57 0.91 0.93 0.52 0.14 0.10 0.56 1.22
Trihalogenated compounds F3 F2Cl F2Br F 2I FCl2 FClBr FClI FBr2 FBrI FI2 Cl3 Cl2Br Cl2I ClBr2 ClBrI ClI2 Br3 Br2I BrI2 I3
3.16 2.51 2.26 1.87 1.86 1.61 1.22 1.35 0.96 1.66 1.21 0.96 0.57 0.70 0.31 0.08 0.45 0.06 0.33 0.72
8.99 7.20 6.47 5.21 5.41 4.68 3.42 3.94 2.69 2.95 3.62 2.88 1.63 2.15 0.90 0.36 1.42 0.17 1.09 2.34
5.31 4.34 3.95 3.30 3.37 2.98 2.33 2.59 1.94 2.34 2.40 2.01 1.36 1.62 0.97 0.32 1.23 0.58 0.07 0.73
3.93 2.87 2.45 1.78 1.81 1.39 0.71 0.96 0.29 1.11 0.75 0.32 0.35 0.10 0.77 1.44 0.52 1.19 1.86 2.53
3.60 2.54 2.12 1.47 1.49 1.07 0.41 0.65 0.01 0.91 0.43 0.01 0.64 0.41 1.06 1.72 0.82 1.48 2.14 2.79
Halogen type & number
a Many of these compounds are not DBPs, but are related chemicals.
in May and 10e28% in August 2005. These losses were probably the result of volatilization, as THMs are volatile species (Henry’s constants ¼ 4.35 104e7.83 103 atm m3/mol). Hydrolysis, which is a common loss mechanism for other DBPs, is not an issue for THMs (Mabey and Mill, 1978). After 6 h of exposure in May, the three brominated THMs were substantially reduced in concentration (>50%), and the degree of photolysis increased with increasing bromine substitution (Fig. 1), whereas the value of trichloromethane was unchanged. Pseudo-first-order rate constants (kP) were calculated: bromoform (CHBr3) (0.32 h1) > dibromochloromethane (CHClBr2) (0.21 h1) > bromodichloromethane (CHCl2Br) (0.11 h1). The degree of THM degradation was lower in the August experiments, but the observed trend remained the same: CHBr3 > CHClBr2 > CHCl2Br. Dibromochloromethane and bromoform exhibited relatively minor losses in August, with kP values of 0.026 and 0.040 h1, respectively. The difference was likely due to less sunlight radiation (405 langleys in
August versus 681 langleys in May) or other experimental factors (e.g., ability to maintain constant temperature, role of clouds, etc). Fig. 2 shows the best-fitting calibrated coefficients of QSAR analysis, and compares the experimental results from May and the QSAR data for three bromine-substituted THMs, along with three iodinated methanes obtained from the literature (Jones and Carpenter, 2005), which reported the kP of diiodomethane, bromoiodomethane, and chloroiodomethane to be 4.32, 0.145, and 0.078 h1, respectively, under an experimental condition similar to this study (pH ¼ 7.65; 54 N, 1 W; outdoor solar photolysis). Only data from May were used, because similar (albeit lower) reactivity trends were observed in August. The differences between observed data and QSARpredicted data for the halomethanes were usually less than 0.1-log relative standard deviation (RSD), except for diiodomethane (1.0-log RSD). QSAR was used to predict the kP of six other chlorine- and bromine-substituted methanes using the calibrated parameters (Fig. 2). In general, the predicted
3406
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
CHCl2Br
CHClBr2
CHBr3
CHCl2Br control
CHClBr2 control
CHBr3 control
30.0
THM (μg /L)
25.0 20.0 15.0 10.0 5.0 0.0 0
2
4 Time (hr)
6
8
Fig. 1 e Solar photolysis of bromine-containing THMs in organic-free water (May 2005) (CHCl2Br [ bromodichloromethane, CHClBr2 [ dibromochloromethane, CHBr3 [ bromoform) (RSDs between replicates analyzed at time zero, 1.5 h, and 6 h were 1.1e6.4%).
kP values of halomethanes were related to the number of halogens (tri- > di- > mono-halogenated) and the size of the halogen substituted, ranked as I > Br > Cl > F (Table 2). The model predicted the kP of trichloromethane to be 6.17 102 h1, which corresponds to a half-life of 11 h (or 30% loss over 6 h), whereas there was no observable loss during the 6-h test (<10%). The gap indicates that solar photolysis of trichloromethane is a complex process controlled by many factors, such that the QSAR method cannot replace laboratory efforts for precise kinetic evaluations. Triiodomethane was predicted to be the most vulnerable halomethane, with a log kP of 0.72 h1 or a half-life of 8 min.
3.2.
Haloacetic acids (HAAs)
Overall, HAAs were relatively stable in water. HAA concentrations for all nine species in the control 2 sample did not change significantly (<20%) throughout the experiment, Log(kp ) = - 1.18Es - 1.51 σ - 2.54, R2 =0.733, n=6 1.0 Observed Predicted
Log (kp ), 1/hour
0.0
-1.0
-2.0
which is within the coefficient of variation (CV) of the analytical method. Due to their protonated form in water, with low acidebase dissociation constants (pKa <3), HAAs are neither volatile nor hydrolysable. However, low levels of degradation of three brominated species (e.g., 21% for tribromoacetic acid) were observed in May 2005. The pseudofirst-order rate constants for monobromo-, dibromochloro-, and tribromoacetic acid were calculated to be 0.011, 0.007 and 0.038 h1, respectively. QSAR analysis of the photolysis results for the HAAs is presented for the three species with experimental results; the kP values matched the QSAR-predicted data. In an earlier study, kP values were reported for bromochloro- (1.84 104 h1), dibromo- (6.91 104 h1), trichloro- (2.3 104 h1), and tribromoacetic acid (3.0 103 h1) under simulated sunlight conditions, whereas monochloro- and monobromoacetic acid photolysis was not observed (Lifongo et al., 2004). Although there were differences in the absolute values between the two studies, the trends were similar (i.e., brominated HAAs were more susceptible than their chlorinated analogs). The half-life of triiodoacetic acid, which is probably the most susceptible HAA, was predicted by QSAR to be 11 s (Table 2). Zhang and Minear (2002) estimated the half-life of triiodoacetic acid due to hydrolysis to be 0.014 day (i.e., 20 min). Either way, this HAA is expected to be highly labile.
3.3.
Haloacetaldehydes (HAs)
Photolysis of dihalogenated acetaldehydes containing chlorine was not observed under the conditions examined (data not shown). However, dibromoacetaldehyde appeared to undergo some photolysis (kP ¼ 0.0045 h1). The brominated trihalogenated acetaldehydes (in particular, tribromoacetaldehyde) exhibited low-to-moderate levels (5e38%) of degradation relative to the control 2 sample. Based on a CV of 20%, only the degradation of the tribromoacetaldehyde was greater than the analytical variability. After accounting for loss in the control 2 samples, the pseudo-first-order rate constants for bromodichloro-, dibromochloro-, and tribromoacetaldehyde were estimated to be 0.011, 0.017, and 0.070 h1, respectively. The degradation of haloacetaldehydes under photolysis increased with increasing degree of bromine incorporation. QSAR analysis for the haloacetaldehydes was based only on the experimental data from the three trihalogenated species. Each of the chlorinated and brominated haloacetaldehydes was predicted to undergo photodegradation, with first-order rate constants ranging from 0.005 to <0.1 h1. The QSARpredicted value for dibromoacetaldehyde agreed with the experimental result with an RSD of 28%. Triiodoacetaldehyde was estimated to be more sensitive than the other haloacetaldehydes, with a half-life of w8 min under the sunlight conditions of this study (Table 2).
3.4.
Halopropanones (HKs)
CH2 I 2
C H 2 I Br
C H 2 I Cl
CHClBr 2
CHBr 3
CHCl 2 Br
CHCl 3
CH 2 Br 2
CH2 ClBr
CH2 Cl2
C H 3 Br
C H 3 Cl
-3.0
Fig. 2 e Comparison of experimental results (-) and QSAR predictions (>) of kP for halomethanes.
Two haloketones, 1,1-dichloro- and 1,1,1-trichloropropanone, were investigated in this study. Both exhibited low levels of degradation (e.g., 14% for the dichlorinated species, where control 2 only differed from control 1 by 3%). Their kP values were estimated to be 0.022 and 0.008 h1 for 1,1-dichloro- and
3407
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Halonitromethanes (HNMs)
Fig. 3 shows the photolysis of three HNM species examined in this study. Degradation typically appeared to occur more rapidly for this class of nitrogenous DBPs than for carbonaceous DBPs in this study. The pseudo-first-order rate constants ranged from 0.09 to w0.8 h1 for five species. A comparison of QSAR data and experimental data is shown in Fig. 4, with an excellent correlation between predicted and measured values (R2 ¼ 0.98). Increasing the number of halogens and/or the degree of bromination tended to increase the photolysis kinetics. Iodinated species are predicted to be more unstable than brominated HNMs (Table 2), and triiodonitromethane had the highest predicted value of kP of 621 h1, which is equivalent to a half-life of around 4 s.
10.0 5.0 0.0 0
2
4 Time (hr)
6
8
Fig. 3 e Solar photolysis of selected HNMs in organic-free water (May 2005) (BNM [ bromonitromethane, DCNM [ dichloronitromethane, TCNM [ trichloronitromethane) (RSDs between replicates analyzed at the five sample times were 0.7e6.3% for BNM and DCNM and were 6.5e33% for TCNM).
irradiation intensity (765 W/cm2) light source (Megan and Reinhard, 2007). These results demonstrate that sunlight intensity (actinic flux) plays an important role in the degradation kinetics of these compounds. The three nitrosamines with cyclic side groups (i.e., N-nitrosomorpholine, N-nitrosopyrrolidine, N-nitrosopiperidine) had some of the highest kP values among the nitrosamines, whereas the three nitrosamines with only methyl and/or ethyl side groups (i.e., NDMA, N-nitrosomethylethylamine, N-nitrosodiethylamine) had the lowest kP values among the nitrosamines. These results suggest that the nature of the side groups may impact the photolytic loss of each nitrosamine. QSAR analysis was not conducted for the nitrosamines due to the concurrent changes in steric structure and carbon numbers, as opposed to halogen substitution on single backbone structures for the halogenated DBPs. Log(kp ) = - 1.90Es - 2.85σ - 1.94, R2 =0.979, n=5 1.0 Observed Predicted
0.0
-1.0
Br3 C(NO2 )
ClBr2 C(NO2 )
Cl2 BrC(NO2 )
Cl3 C(NO2 )
Br2 CH(NO2 )
-2.0 Cl2 CH(NO2 )
With the solar intensity of 1150e1300 W/cm2, all eight nitrosamines were rapidly photolyzed (w99%) within 1 h. The halflives were approximately 8e10 min, with kP around 4.9 h1. The indoor tests, using simulated sunlight (1325 W/m2), exhibited more rapid photolysis, with rate constants of 5e10 h1 (Fig. 5), which corresponded to half-lives of 3e9 min. Both sets of experimental results were faster than an earlier study (half-lives ¼ 12e18 min), which employed a lower
ClBrCH(NO2 )
Nitrosamines (NAs)
BrCH2 (NO2 )
3.7.
15.0
ClCH2 (NO2 )
3.6.
TCNM TCNM control
20.0
Haloacetonitriles (HANs)
Over the 6-h test period in May, the HANs in control 2 degraded to varying levels: dichloroacetonitrile was reduced in concentration by 46%, bromochloroacetonitrile by 26%, and dibromoacetonitrile by 12%. This was probably due to hydrolysis (Croue´ and Reckhow, 1989). In addition, photolysis of the HANs was clearly observed. The kP values of bromochloro- and dibromoacetonitrile were calculated to be 0.329 and 0.324 h1, respectively, whereas dichloroacetonitrile’s photolytic loss rate was low (kP ¼ 0.063 h1) after accounting for the hydrolysis effect. QSAR analyses for experimental data and predictions of other HANs are presented in Table 2. Increasing the number of halogen atoms and/or replacing chlorine with bromine increased the photolysis decay of the HANs. The trend is consistent with previous findings that dichloroacetonitrile in water was photodegradable, whereas monochloroacetonitrile was resistant to photolysis (Lekkas and Nikolaou, 2004). Triiodoacetonitrile is probably the HAN that can be photolyzed most rapidly, with an estimated half-life of less than 8 s (Table 2).
DCNM DCNM control
25.0
Log (kp ), 1/hour
3.5.
BNM BNM control
HNM ( μg /L)
1,1,1-trichloropropanone, respectively. Although the trihalogenated species had a somewhat lower (not higher) rate of photolysis than the dihalogenated propanone, both values were low. Moreover, 1,1,1-trichloropropanone underwent a loss (12%) in the control 2 sample, which may have been due to hydrolysis, so determination of the rate of decay due to photolysis required separating out the other loss mechanism. QSAR was not conducted for this group because photolysis rate constants were only determined for two chlorinated species.
Fig. 4 e Comparison of experimental results (-) and QSAR predictions (>) of kP for halonitromethanes.
3408
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Nanopure Water Effluent #1 Filtered Effluent #2 Filtered
Effluent #1 Unfiltered Effluent #2 Unfiltered 100 9.06/hr
5.16/hr
5.80/hr
7.14/hr
10.38/hr
NDEA
NMOR
10.20/hr
7.74/hr
NPIP
NPYR
kp/knanopure (%)
7.43/hr
50
0 NDBA
NDPA
NDMA
NMEA
Nitrosamine Species
Fig. 5 e Comparison of kP of nitrosamines in non-filtered and filtered EfOM to samples prepared in organic-free water under simulated sunlight (the numbers beside the bars denote the kP values in organic-free water) (NDMA: N-nitrosodimethylamine, NMEA: N-nitrosomethylethylamine, NDEA: N-nitrosodiethylamine, NDPA: N-nitrosodipropylamine, NDBA: N-nitrosodibutylamine, NMOR: N-nitrosomorpholine, NPYR: N-nitrosopyrrolidine, NPIP: N-nitrosopiperidine).
0.0
HAN HNM
-1.0
-2.0
-3.0
3Br
1Cl,2Br
2Cl,1Br
3Cl
2Br
1Cl,1Br
-4.0 2Cl
This study represents among the first investigations of the solar photolysis effect on a wide variety of carbonaceous and nitrogenous DBPs and related compounds under a set of common exposure conditions. All rates of DBP transformation are intended to be apparent rates rather than mechanistic rate constants. As such, they should be used in relationship to each other to identify more or less photo-reactive DBP classes and influences of halogen substitution. The QSAR-predicted values agreed with the experimental results with R2 values of 0.73 or higher. Although only brominated and chlorinated DBPs were commonly detected during wastewater treatment
HM HAA HA
1Br
Conclusion
1.0
1Cl
4.
(Krasner et al., 2009), including data on iodinated related chemicals from the literature with our observed brominated and chlorinated DBP data allowed us to demonstrate the importance of halogen substitution. Moreover, QSAR was used to predict kP values for fluorinated, chlorinated, brominated, and iodinated species not tested (some of which are not DBPs). Fig. 6 compares the QSAR estimates for five groups of halogenated DBPs. The photolysis kinetics appeared to be a function of functional group, halogen type(s) and number(s).
Log (kp), 1/hour
In other studies, it was found that the advanced oxidation of nitrosamines with free radicals could be mechanistically understood (in part) based on structureeactivity relationships (Mezyk et al., 2008). In the absence of sunlight, the average decrease in concentration of the nitrosamines in the control 2 sample was 9%, which was less than the analytical CV, indicating that no other mechanism(s) contributed to nitrosamine loss during these experiments. For nitrosamines spiked into EfOM, their first-order rate constants were around 50% below those spiked into organic-free water (Fig. 5). Typically, nitrosamines in effluent #2, which had higher TOC (16.5 mg/L), exhibited slower photolysis kinetics than those in effluent #1 (TOC ¼ 12.2 mg/L). In addition, there was (on average) an 11% higher kP (i.e., faster kinetics) for filtered water as compared to non-filtered samples. Thus, organic or particulate matter can impact the rate of photolysis. For example, increased turbidity can decrease the actual actinic flow in the solution.
Fig. 6 e Comparison of kP of five groups of halogenated DBPs under sunlight conditions in May 2005 in Tempe (AZ) as predicted by QSAR analysis (the x-axis denotes the number(s) and type(s) of halogen(s) in the compounds) (HM [ halomethane, HAA [ haloacetic acid, HA [ haloaldehyde, HAN [ haloacetonitrile, HNM [ halonitromethane).
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Including non-halogenated compounds, the nitrogenous DBPs (i.e., nitrosamines, HNMs, HANs) had higher photolytic rates than carbonaceous DBPs (i.e., THMs, HAAs, haloacetaldehydes, halopropanones), generally ranked as follows: nitroso (N-NO) > nitro (NO2) > nitrile (C^N) > carbonyl (C]O) > carboxyl (COOH). Although most HAAs and haloacetaldehydes did not degrade over the 6-h photolysis test period, near complete removals (99%) of nitrosamines were observed within 1 h of sunlight exposure. Within one group of DBPs, the photolysis kinetics usually increased with increasing number of halogens. Brominated DBPs were more susceptible than chlorinated DBPs, and iodinated species (some of which may be DBPs) are probably even more unstable than the brominated analogs.
Acknowledgements The authors thank the American Water Works Association Research Foundation (AwwaRF) and the USEPA for its financial, technical, and administrative assistance in funding and managing the project. The comments and views detailed herein may not necessarily reflect the views of AwwaRF, its officers, directors, affiliates or agents, or the views of the U.S. federal government. The project manager was Alice Fulmer. The photolysis setup was funded by NSF ATM 0530718. Thanks are also extended to Metropolitan staff who conducted various analyses to support this study, including Michael J. Sclimenti, Salvador J. Pastor, Russell Chinn, Cordelia J. Hwang, Yingbo C. Guo, Eduardo Garcia-Zayas, Donnie Morquecho, Wing Chow, Rachel M. Kelley, Mary Ann Viernes, Sikha Kundu, Chih Fen Tiffany Lee, and Jacob Nikonchuk. In addition, the authors thank the participating utilities for supplying treated wastewater samples for testing.
references
Allston, T.D., Fedyk, M.L., et al., 1978. Photoabsorption spectra of some halo-substituted methyl nitrogen oxides. J. Photochem. 9 (2e3), 116e118. Bayes, K.D., Friedl, R.R., Sander, S.P., Toohey, D.W., 2003. Measurements of quantum yields of bromine atoms in the photolysis of bromoform from 266 to 324 nm. J. Geophys. Res. 108 (D3), 4095. doi:10.1029/2002JD002877. Castro, C.E., Belser, N.O., 1981. Photohydrolysis of methyl bromide and chloropicrin. J. Agric. Food Chem. 29, 1005e1008. Chen, B., Westerhoff, P., Krasner, S.W., 2008. Fate and transport of wastewater-derived disinfection by products in surface waters. In: Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y. (Eds.), Disinfection By-products in Drinking Water: Occurrence, Formation, Health Effects, and Control. American Chemical Society, Washington, DC, pp. 257e273. Chen, J.W., Peijnenburg, W., et al., 2001. Is it possible to develop a QSPR model for direct photolysis half-lives of PAHs under irradiation of sunlight? Environ. Pollut. 114 (1), 137e143. Croue´, J.-P., Reckhow, D.A., 1989. Destruction of chlorination byproducts with sulfite. Environ. Sci. Technol. 23 (11), 1412e1419. Hammett, L.P., 1935. Some relations between reaction rates and equilibrium constants. Chem. Rev. 17 (1), 125e136.
3409
Hansch, C., Leo, A., Hoekman, D., 1995. Exploring QSAR: Hydrophobic, Electronic, and Steric Constants. In: Heller, S.R. (Ed.), American Chemical Society Professional Reference Book Washington, DC. Ho, T.-F.L., Bolton, J.R., Lipczynska-Kochany, E., 1996. Quantum yields for the photodegradation of pollutants in dilute aqueous solution: phenol, 4-chlorophenol and N-nitrosodimethylamine. J. Adv. Oxid. Technol. 1 (2), 170e178. Jones, C., Carpenter, L.J., 2005. Solar photolysis of CH2I2, CH2ICl, and CH2IBr in water, saltwater, and seawater. Environ. Sci. Technol. 39 (16), 6130e6137. Krasner, S.W., Westerhoff, P., Chen, B., Amy, G., Nam, S.-N., Chowdhury, Z.K., Sinha, S., Rittmann, B.E., 2008. Contribution of Wastewater to DBP Formation. Awwa Research Foundation, Denver, USA. Krasner, S.W., Westerhoff, P., Chen, B., Rittmann, B.E., Amy, G., 2009. Occurrence of disinfection byproducts in United States wastewater treatment plant effluents. Environ. Sci. Technol. 43 (21), 8320e8325. Lifongo, L.L., Bowden, D.J., et al., 2004. Photodegradation of haloacetic acids in water. Chemosphere 55 (3), 467e476. Lekkas, T.D., Nikolaou, A.D., 2004. Degradation of disinfection byproducts in drinking water. Environ. Eng. Sci. 21 (4), 493e506. Mabey, W., Mill, T., 1978. Critical review of hydrolysis of organic compounds in water under environmental conditions. J. Phys. Chem. Ref. Data 7 (2), 384e415. Megan, H.P., Reinhard, M., 2007. Photochemical attenuation of N-nitrosodimethylamine (NDMA) and other nitrosamines in surface water. Environ. Sci. Technol. 41 (17), 6170e6176. Mezyk, S.P., Landsman, N.A., Swancutt, K.L., Bradford, C.N., Cox, C.R., Kiddle, J.J., Clore, T.J., 2008. Free radical chemistry of advanced oxidation process removal of nitrosamines in waters. In: Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y. (Eds.), Disinfection By-products in Drinking Water: Occurrence, Formation, Health Effects, and Control. American Chemical Society, Washington, DC, pp. 319e333. Mizgireuv, I.V., Majorova, I.G., Gorodinskaya, V.M., Khudoley, V.V., Revskoy, S.Y., 2004. Carcinogenic effect of N-nitrosodimethylamine on diploid and triploid zebrafish (Danio rerio). Toxicol. Pathol. 32 (5), 514e518. Nicole, I., De Laat, J., et al., 1991. Degradation of trihalomethanes in dilute aqueous solution by UV irradiation e determination of the quantum yield of photolysis at 253.7 nm. Environ. Technol. 12 (1), 21e31. Peijnenburg, W.J.G.M., de Beer, K.G.M., den Hollander, H.A., Stegeman, M.H.L., Verboom, H., 1993. The development of quantitative structure activity relationships for the direct photolysis of substituted haloaromatics in aqueous environments. Sci. Total Environ. 134 (Suppl. 2), 1397e1408. Plewa, M.J., Wagner, E.D., 2009. Quantitative Comparative Mammalian Cell Cytotoxicity and Genotoxicity of Selected Classes of Drinking Water Disinfection By-products. Awwa Research Foundation, Denver, USA. Rudra, A., Thacker, N.P., Pande, S.P., 2005. Hydrogen peroxide and ultraviolet irradiations in water treatment. Environ. Monit. Assess. 109, 189e197. Stefan, M.I., Bolton, J.R., 2002. UV direct photolysis of N-nitrosodimethylamine (NDMA): kinetic and product study. Helv. Chim. Acta 85, 1416e1426. Trehy, M.L., Yost, R.A., Miles, C.J., 1986. Chlorination byproducts of amino acids in natural waters. Environ. Sci. Technol. 20 (11), 1117e1122. Zhang, X., Minear, R.A., 2002. Decomposition of trihaloacetic acids and formation of the corresponding trihalomethanes in drinking water. Water Res. 36, 3665e3673. Zou, P., Shu, J., Sears, T.J., Hall, G.E., North, S.W., 2004. Photodissociation of bromoform at 248 nm: single and multiphoton processes. J. Phys. Chem. A108, 1482e1488.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
Solar photolysis kinetics of disinfection byproducts Baiyang Chen a,*, Wontae Lee b, Paul K. Westerhoff c, Stuart W. Krasner d, Pierre Herckes e a
Chinese Environmental Scholars and Professionals Network, 11900 Stonehollow Drive, Apartment 338, Austin, TX 78758, USA HDR Engineering, Inc., 3200 East Camelback Road, Suite 350, Phoenix, AZ 85018, USA c Arizona State University, School of Sustainable Engineering and The Built Environment, Engineering Center (G-Wing), Room ECG-252, Tempe, AZ 85287-5306, USA d Metropolitan Water District of Southern California, Water Quality Laboratory, 700 Moreno Avenue, La Verne, CA 91750, USA e Arizona State University, Department of Chemistry and Biochemistry, Tempe, AZ 85287-1604, USA b
article info
abstract
Article history:
Disinfection byproducts (DBPs) discharged from wastewater treatment plants may impair
Received 17 December 2009
aquatic ecosystems and downstream drinking-water quality. Sunlight photolysis, as one
Received in revised form
process by which DBPs may dissipate in the receiving surface water, was investigated.
8 March 2010
Outdoor natural sunlight experiments were conducted in water for a series of carbona-
Accepted 8 March 2010
ceous DBPs (trihalomethanes, haloacetic acids, halopropanones, and haloacetaldehydes)
Available online 17 March 2010
and nitrogenous DBPs (nitrosamines, halonitromethanes, and haloacetonitriles). Their pseudo-first-order rate constants for photolytic degradation were then used to calibrate
Keywords:
quantitative structureeactivity relationship (QSAR) parameters, which, in return, predicted
Photolysis
the photolysis potentials of other DBPs or related compounds. Nitrogenous DBPs were
QSAR
found to be more susceptible to solar irradiation than carbonaceous DBPs, with general
DBP
rankings for the functional groups as follows: N-nitroso (N-NO) > nitro (NO2) > nitrile
NDMA
(C^N) > carbonyl (C]O) > carboxyl (COOH). Compounds containing a high degree of
THM
halogenation (e.g., three halogens) were usually less stable than less halogenated species
HAA
(e.g., those with two halogens). Bromine- or iodine-substituted species were more photosensitive than chlorinated analogs. While most bromine- and chlorine-containing trihalomethanes and haloacetic acids persisted over the 6-h test, nearly complete removal (>99%) of nitrosamines occurred within 1 h of sunlight exposure. Indoor laboratory experiments using simulated sunlight demonstrated that the degradation of nitrosamines was w50% slower when organic matter was present, and w11% slower in non-filtered water than in filtered water. ª 2010 Elsevier Ltd. All rights reserved.
1.
Introduction
In addition to water recycling and reclamation programs, unintentional reuse of treated wastewater has occurred over the past few decades and will likely increase in the future
as upstream wastewater treatment plants (WWTPs) discharge effluents into rivers or lakes that serve as downstream drinking-water resources. Drought and competing in-stream demands may result in substantial contribution of treated wastewater in water bodies (Krasner et al., 2008). Attention
* Corresponding author. Tel.: þ1 480 840 4647. E-mail addresses: [email protected] (B. Chen), [email protected] (W. Lee), [email protected] (P.K. Westerhoff), [email protected] (S.W. Krasner), [email protected] (P. Herckes). 0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2010.03.014
3402
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
has been drawn to the impact of WWTP discharges as a source of multiple contaminants, such as pharmaceuticals, to aquatic life and downstream drinking-water sources. Due to the common disinfection practice at U.S. WWTPs of adding chlorine, disinfection byproducts (DBPs) are likely to be present in WWTP discharges. Wastewater-derived DBPs may pose risks to downstream aquatic organisms (Mizgireuv et al., 2004) and human health. In a U.S. survey, for WWTPs that achieved breakpoint chlorination (i.e., free chlorine residual), the median concentrations of trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs), and trihalogenated acetaldehydes were 57, 70, 16, and 16 mg/L, respectively (Krasner et al., 2009). For WWTPs without breakpoint chlorination (i.e., chloramine residual), a median of 11 ng/L (75th percentile ¼ 24 ng/L) of N-nitrosodimethylamine (NDMA) was detected (Krasner et al., 2009). Chlorination of effluent organic matter (EfOM) in an earlier study also resulted in the formation of chloral hydrate (i.e., trichloroacetaldehyde) and dichloroacetonitrile (Trehy et al., 1986). As WWTP effluents contain relatively high levels of bromide and iodide due to anthropogenic activities (Krasner et al., 2008), certain brominated and potentially iodinated DBPs may form too. Some nitrogenous DBPs (including certain HANs, halonitromethanes [HNMs]) are more cytotoxic and genotoxic than the regulated carbonaceous DBPs (i.e., THMs, HAAs), and their toxicity was impacted by halogen substitution type: iodine > bromine > chlorine (Plewa and Wagner, 2009). The U.S. Environmental Protection Agency’s (USEPA’s) Integrated Risk Information System (IRIS) database indicates for NDMA and seven other nitrosamines that concentrations of low ng/L level in water are associated with a 106 lifetime cancer risk. Therefore, the treatment, fate, and transport of wastewaterderived DBPs in water resources require study. In conjunction with sorption, biodegradation, volatilization, hydrolysis, and other biogeochemical processes, sunlight photolysis may reduce DBP concentrations to varying degrees in surface waters (Table 1). For example, dichloroacetonitrile spiked into drinking water was photodegradable (Lekkas and Nikolaou, 2004), and chloropicrin (i.e., trichloronitromethane [TCNM]) in the atmosphere was degraded under sunlight conditions (Allston et al., 1978). HAAs underwent varying levels of photolysis (Lifongo et al., 2004), and nitrosamines exhibited substantial photolytic losses under simulated sunlight conditions (Megan and Reinhard, 2007). However, the experimental conditions in these studies were different, which makes a comparison of the photolysis degradation of different groups of DBPs difficult. Quantitative structureeactivity relationship (QSAR) analysis is an empirical chemical property estimation method that can predict the behaviors of one chemical from other structurally similar chemicals. The method was developed by Hammett (1935), explored by Hansch et al. (1995), and widely applied in many areas, including the prediction of toxicity testing and pharmaceutical behavior (Hansch et al., 1995), the photolysis of polycyclic aromatic hydrocarbons (Chen et al., 2001) and haloaromatics (Peijnenburg et al., 1993), and the hydrolysis of trihaloacetic acids (Zhang and Minear, 2002). The QSAR application is based upon a key assumption that chemicals with similar structures behave similarly and their activity differences are attributed only to the type, number,
steric, and electronic effect of their functional group(s). Because DBPs in one group share similar chemical structures, the QSAR technique should be applicable to DBP photolysis. The purpose of this paper was to investigate the photolysis of several groups of DBPs and related compounds under uniform solar irradiation conditions. The pseudo-first-order degradation rate constants were obtained from experiments and subsequently used to calibrate the QSAR models. In turn, the calibrated QSAR models were used to predict the photolytic degradation of other DBPs or related compounds, such as iodinated DBPs, which have not been studied experimentally. Although the reactive mechanism and pathway of each DBP is uncertain and many other factors, such as receiving waterbody geometry (width, depth of the photic zone), flowrate, compound sensitivity, microbial activity, solar light spectra, and meteorological conditions are also critical in understanding the photolytic degradation of DBPs (Chen et al., 2008), the outcome of this study may serve as a basis for further and more thorough exploration of the role of photolysis in DBP removal.
2.
Materials and methods
2.1.
Natural sunlight (outdoor) experiment
Sunlight photolysis of DBPs was evaluated in 9-mL quartz test tubes (outside diameter ¼ 13 mm). Concentrated DBP solutions were spiked into 200 mL of pH-buffered organic-pure water (pH w7.2; phosphate buffer) in volumetric flasks to result in concentrations of w500e1000 mg/L of each halogenated DBP and w10,000 ng/L of each nitrosamine and were distributed evenly to multiple test tubes. Sample vials were capped to minimize volatilization or evaporation, and placed on an inclined platform on top of the laboratory roof in Tempe, Arizona, USA (33 250 12.3500 N, 111 550 55.1100 W). Samples were partially air cooled during the tests by a fan, as the surrounding air temperature was w30 C (Arizona Meteorological Network, http://ag.arizona.edu/azmet). Energy intensities were monitored on-site by a photometer (IL 1700, International Light Technologies, Inc., Massachusetts, USA). The solar spectrum was also extracted from a model named SPCTRAL2 (http://rredc.nrel.gov/solar/models/spectral/ SPCTRAL2). The irradiation intensities appeared similar and remained constant in the range of 1150e1300 W/m2 between 11 am and 3 pm (<10% difference). Halogenated DBP and nitrosamine samples were exposed for 0, 0.5, 1.5, 3, and 6 h from 11 am to 5 pm on May 20, 2005 for the first run of the experiment. On July 25, 2005, nitrosamine photolysis experiments were repeated but for a shorter exposure time (i.e., 0, 1, 3, 5, 10, 20, 30, and 60 min). On August 10, 2005, a second run was conducted for up to 8 h for the THMs and HANs. A 6- or 8-h time period was chosen in order to have a nearly constant solar irradiation for each experiment. The total solar radiations of the three dates were recorded as 696, 681, and 405.7 langleys for May 20, July 25, and August 10, 2005, respectively, by Arizona Meteorological Network Mesa Station (http://ag.arizona.edu/azmet). The average air temperatures of the three dates however were similar, being 31.7, 33.3, and 33.9 C, respectively.
3403
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Table 1 e Literature review of photolysis kinetics of halogenated chemicals. DBP or related chemical
Conditions
Pseudo-first-order rate constant (kP, h1) and/or quantum yield (F)
References
Halomethanes CH2I2 CH2IBr CH2ICl
pH ¼ 7.65; deionized water, salt water, seawater; sunlight; 55 N, 1 W
4.32, F ¼ 0.31 1.45 101, F ¼ 0.29 7.81 102, F ¼ 0.13
Jones and Carpenter, 2005
CHCl3
River/lake; 20 C
1.81 104
SCDM, 1997a
THMs
pH ¼ 2.5e10; UV ¼ 112e400 nm; 0.1% H2O2
CHBr3 > CHClBr2 > CHCl2Br > CHCl3, up to 92% removal
Rudra et al., 2005
CHBr3
Air; UV ¼ 266e324 nm; 31 C
F ¼ 0.76
Bayes et al., 2003
CHBr3
Air; UV ¼ 248 nm
F ¼ 0.2e0.3
CH3Br
Neutral pH; sunlight, UV at 254 nm
Nil under sunlight, 6.12 10 under UV at 254 nm
CHCl3 CHCl2Br CHClBr2 CHBr3
pH ¼ 7.5; <30 min; UV at 253.7 nm; 20 C
<5%
Haloacetic acids Br2CH(COOH) BrClCH(COOH) Cl3CH(COOH) Br3CH(COOH)
Zou et al., 2004 3
Castro and Belser, 1981
Nicole et al., 1991
F ¼ 0.43 0.1
Mercury lamps (l > 400 nm); 15 C
6.91 104 1.84 104 2.30 104 3.00 103
Lifongo et al., 2004
Halonitromethanes Cl3C(NO2)
Neutral pH; sunlight, UV at 254 nm; 20 C,
9.61 103 (sunlight) 5.00 101 (UV)
Castro and Belser, 1981
Cl3C(NO2)
Air; UV ¼ 190e400 nm; 295 K
2.05 101 to 2.88
Allston et al., 1978
Nitrosamines N-Nitrosodimethylamine
pH ¼ 3 and 7; 1 mM; UV ¼ 200e300 nm
156, F ¼ 0.3
Stefan and Bolton, 2002
N-Nitrosodimethylamine N-Nitrosodiethylamine N-Nitrosodipropylamine N-Nitrosodibutylamine N-Nitrosopiperidine
River/lake; 20 C
6.88 101 8.75 102 6.88 101 8.75 102 3.07 103
SCDM, 1997a
N-Nitrosodimethylamine
Neutral
F ¼ 0.11 0.03
Ho et al., 1996
2.40 0.12, F ¼ 0.41 2.94 0.18, F ¼ 0.61 2.70 0.24, F ¼ 0.43 3.00 0.24, F ¼ 0.46 2.88 0.06, F ¼ 0.52 3.30 0.36, F ¼ 0.55 3.42 0.12, F ¼ 0.51
Megan and Reinhard, 2007
N-Nitrosodimethylamine N-Nitrosomethylethylamine N-Nitrosodiethylamine N-Nitrosodipropylamine N-Nitrosodibutylamine N-Nitrosopyrrolidine N-Nitrosopiperidine
2
765 W/m in Milli-Q water (pH ¼ 6); 100 mg/L except for NPIP (1000 mg/L); simulated sunlight lamp
a SCDM: Superfund Chemical Data Matrix program, 1997 version, from USEPA.
After each time interval, samples were immediately preserved as appropriate (i.e., pH lowered to w3e4 for samples of THMs, HANs, halopropanones, HNMs, and haloacetaldehydes), refrigerated, and shipped to the laboratory for DBP analysis (Krasner et al., 2009). The photolysis samples (9-mL) were diluted to a 200- to 500-mL volume before the analyses, so that there would be sufficient volume for the analytical methods used, which diluted the halogenated DBPs (in the absence of photolytic degradation) to w20 mg/L and the nitrosamines to w180 ng/L. These concentration levels were selected to be able to observe losses of up to w90e99%. Briefly,
the THMs, HANs, halopropanones, and chloropicrin were analyzed by liquid/liquid extraction (LLE) and gas chromatography (GC) with an electron capture detector (ECD); other HNMs and haloacetaldehydes were analyzed by either LLEeGC/ECD or solid-phase extraction and GC/mass spectrometry (MS); the HAAs were analyzed by LLE, derivatization with acidic methanol, and GC/ECD; and the nitrosamines were analyzed by micro LLE and chemical ionization GC/MS. Throughout the photolysis study, two sets of control samples were employed, which included “control 1” wrapped in aluminum foil and stored at 4 C in a refrigerator (i.e., no light
3404
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
or heat exposure), and “control 2” samples wrapped in aluminum foil but placed in a box near the samples for the entire hold time (i.e., same outdoor conditions and handling as that of the testing samples, except for no light exposure). Control 1 samples had minimal effects from other loss mechanisms, as they were immediately preserved (e.g., pH lowered to w3e4 for samples of THMs, HANs, halopropanones, HNMs, and haloacetaldehydes), whereas control 2 samples could experience other loss mechanisms such as hydrolysis and/or volatilization during the outdoor holding period. The pseudo-first-order rate constant of photolysis (kP) of the DBPs was calculated, which was adjusted for any other loss mechanism observed in control 2. The degradation kinetics were assumed to be pseudo-first-order for all DBPs.
2.2.
Laboratory photolysis (indoor) experiment
In order to understand the effects of organic and particulate matter on the photolysis of nitrosamines, which were the most labile species studied, indoor laboratory experiments were conducted in different types of water. The sample was placed in a thermostated and jacketed reactor (100-mL capacity). The reactor was constantly stirred and was irradiated through a quartz window on top with simulated sunlight. The simulated light was provided by a 300-W xenon arc lamp (Spectraphysics Oriel, 91160A), and the output of the arc lamp was filtered through a standardized air mass 1.5 filter (AM 1.5 Global), which yielded a spectrum representative of the solar spectrum at ground level when the sun was at a zenith angle of 48.2 , and the lamp output power was 1325 W/m2. This experiment was different from the outdoor work in the following ways: (1) light intensity was higher and more consistent throughout the duration, (2) larger sample volumes were used in this setup (40 versus 9 mL), and (3) water temperature was low and maintained constant at room temperature (20 C). Three types of waters were spiked with eight nitrosamines: (1) organic-free water; (2) non-filtered grab samples from two WWTP effluents (effluent #1 had a turbidity of 1.5 NTU and effluent #2 had a turbidity of 8.8 NTU); and (3) two filtered (0.45-mm) WWTP effluents from (2). WWTP effluent #1 had a total organic carbon (TOC) concentration of 12.2 mg/L and ultraviolet absorbance at 254 nm (UVA254) of 0.18 cm1, whereas WWTP effluent #2 had a TOC of 16.5 mg/L and UVA254 of 0.16 cm1. UVevisible absorbance spectra were compared for organic-free water and the two types of effluents, both filtered and non-filtered. In the non-filtered samples, substantially more (>30%) light interception was observed at wavelengths greater than 350 nm, which means less photons can reach DBPs when particulate matter is present (Krasner et al., 2008).
2.3.
QSAR modeling
In addition to experimental approaches, a fundamental QSAR analysis equation (Eq. (1)) was employed to predict the pseudo-first-order rate constants of non-tested DBPs or related compounds with similar structure(s). The equation had been applied to predicting DBP hydrolysis in previous
research (Zhang and Minear, 2002), but had not been used for DBP photolysis analysis until now. logðkP Þ ¼ d$Es þ r$s þ c0
(1)
where kP is the pseudo-first-order observed photolysis (P) rate constant (k) of the target chemical of the group; Es is the steric effect constant of a substituent; s is the electronic effect constant of the substituent; d is the sensitivity coefficient of the steric effect of the substituent; r is the sensitivity coefficient of the electronic effect of the substituent; and c0 is an empirical constant (Hansch et al., 1995). Based upon previous studies, 0.50, 0.47, 0.42, and 0.38 were obtained as the s values (electronic constants) and 0.46, 0.96, 1.16, and 1.40 were obtained as the Es values (steric constants) for fluoride, chloride, bromide, and iodide, respectively (Hansch et al., 1995). The effects of multiple halides were assumed to be the sum of the effects of single halides, which follows an early study (Zhang and Minear, 2002). Accordingly, Eq. (1) became a triplex equation with three unknowns (d, r, and c0 ). Solving the triplex equations required three sets of inputs (i.e., kP, s, and Es) and once the constants d, r, and c0 values were obtained, Eq. (1) was, in turn, used to estimate the kinetics of other chemicals. During our experiments the oxidation mechanisms probably involved both direct photolysis and indirect radical reactions. Direct photolysis is a function of quantum yield. Rates of indirect reactions involve both the DBP chemical structure and radical concentrations. As such, the intent of this paper is not to relate fundamental chemical properties (e.g., chemical bond strength, thermodynamic properties) to the observed pseudofirst-order rate loss constants because multiple oxidation mechanisms may be involved. Efforts along these lines would require greater focus on a smaller number of DBPs. Instead, we hypothesize that within a specific DBP class the effects of halogenated substituents can be represented by Es, s, d, and r, collectively, and correlated with the observed kP. Note that DBPs can be chlorine-, bromine-, and/or iodine-substituted. Fluorinated compounds were also evaluated for completeness in the QSAR analysis, but are not DBPs. Moreover, not all of the potential iodinated analogs of the DBPs have been identified in water.
3.
Results and discussion
This paper presents both experimentally obtained and QSAR predictive data. The experimental conditions of this study are representative of a near-surface water exposed to intensive solar irradiation, such that the results of this study may reflect the photolysis degradation potentials of the DBPs and related compounds. The photolysis rate constants were estimated for a range of halogenated compounds (Table 2). The following discussion focuses primarily on the regulated and emerging DBPs of concern.
3.1.
Halomethanes (HMs)
Over the 6- or 8-h intensive solar exposure periods, individual THM species in control 2 samples exhibited losses of 5e11%
3405
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Table 2 e QSAR predictions of log kP of halogenated compoundsa under sunlight conditions in May 2005 in Tempe (AZ). Halomethane, log kP (h1)
Haloacetic acid, log kP (h1)
Haloacetaldehyde, log kP (h1)
Haloacetonitrile, log kP (h1)
Halonitromethane, log kP (h1)
Mono-halogenated compounds F Cl Br I
2.74 2.10 1.84 1.45
4.47 2.68 1.95 0.69
3.39 2.42 2.03 1.38
2.61 1.55 1.12 0.45
2.49 1.44 1.02 0.36
Dihalogenated compounds F2 FCl FBr FI Cl2 ClBr ClI Br2 BrI I2
2.95 2.30 2.05 1.66 1.65 1.40 1.01 1.14 0.75 0.36
6.73 4.94 4.21 2.95 3.15 2.42 1.16 1.68 0.43 0.82
4.35 3.38 2.99 2.34 2.41 2.02 1.37 1.63 0.98 0.33
3.27 2.21 1.79 1.11 1.15 0.72 0.05 0.30 0.37 1.04
3.04 1.99 1.57 0.91 0.93 0.52 0.14 0.10 0.56 1.22
Trihalogenated compounds F3 F2Cl F2Br F 2I FCl2 FClBr FClI FBr2 FBrI FI2 Cl3 Cl2Br Cl2I ClBr2 ClBrI ClI2 Br3 Br2I BrI2 I3
3.16 2.51 2.26 1.87 1.86 1.61 1.22 1.35 0.96 1.66 1.21 0.96 0.57 0.70 0.31 0.08 0.45 0.06 0.33 0.72
8.99 7.20 6.47 5.21 5.41 4.68 3.42 3.94 2.69 2.95 3.62 2.88 1.63 2.15 0.90 0.36 1.42 0.17 1.09 2.34
5.31 4.34 3.95 3.30 3.37 2.98 2.33 2.59 1.94 2.34 2.40 2.01 1.36 1.62 0.97 0.32 1.23 0.58 0.07 0.73
3.93 2.87 2.45 1.78 1.81 1.39 0.71 0.96 0.29 1.11 0.75 0.32 0.35 0.10 0.77 1.44 0.52 1.19 1.86 2.53
3.60 2.54 2.12 1.47 1.49 1.07 0.41 0.65 0.01 0.91 0.43 0.01 0.64 0.41 1.06 1.72 0.82 1.48 2.14 2.79
Halogen type & number
a Many of these compounds are not DBPs, but are related chemicals.
in May and 10e28% in August 2005. These losses were probably the result of volatilization, as THMs are volatile species (Henry’s constants ¼ 4.35 104e7.83 103 atm m3/mol). Hydrolysis, which is a common loss mechanism for other DBPs, is not an issue for THMs (Mabey and Mill, 1978). After 6 h of exposure in May, the three brominated THMs were substantially reduced in concentration (>50%), and the degree of photolysis increased with increasing bromine substitution (Fig. 1), whereas the value of trichloromethane was unchanged. Pseudo-first-order rate constants (kP) were calculated: bromoform (CHBr3) (0.32 h1) > dibromochloromethane (CHClBr2) (0.21 h1) > bromodichloromethane (CHCl2Br) (0.11 h1). The degree of THM degradation was lower in the August experiments, but the observed trend remained the same: CHBr3 > CHClBr2 > CHCl2Br. Dibromochloromethane and bromoform exhibited relatively minor losses in August, with kP values of 0.026 and 0.040 h1, respectively. The difference was likely due to less sunlight radiation (405 langleys in
August versus 681 langleys in May) or other experimental factors (e.g., ability to maintain constant temperature, role of clouds, etc). Fig. 2 shows the best-fitting calibrated coefficients of QSAR analysis, and compares the experimental results from May and the QSAR data for three bromine-substituted THMs, along with three iodinated methanes obtained from the literature (Jones and Carpenter, 2005), which reported the kP of diiodomethane, bromoiodomethane, and chloroiodomethane to be 4.32, 0.145, and 0.078 h1, respectively, under an experimental condition similar to this study (pH ¼ 7.65; 54 N, 1 W; outdoor solar photolysis). Only data from May were used, because similar (albeit lower) reactivity trends were observed in August. The differences between observed data and QSARpredicted data for the halomethanes were usually less than 0.1-log relative standard deviation (RSD), except for diiodomethane (1.0-log RSD). QSAR was used to predict the kP of six other chlorine- and bromine-substituted methanes using the calibrated parameters (Fig. 2). In general, the predicted
3406
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
CHCl2Br
CHClBr2
CHBr3
CHCl2Br control
CHClBr2 control
CHBr3 control
30.0
THM (μg /L)
25.0 20.0 15.0 10.0 5.0 0.0 0
2
4 Time (hr)
6
8
Fig. 1 e Solar photolysis of bromine-containing THMs in organic-free water (May 2005) (CHCl2Br [ bromodichloromethane, CHClBr2 [ dibromochloromethane, CHBr3 [ bromoform) (RSDs between replicates analyzed at time zero, 1.5 h, and 6 h were 1.1e6.4%).
kP values of halomethanes were related to the number of halogens (tri- > di- > mono-halogenated) and the size of the halogen substituted, ranked as I > Br > Cl > F (Table 2). The model predicted the kP of trichloromethane to be 6.17 102 h1, which corresponds to a half-life of 11 h (or 30% loss over 6 h), whereas there was no observable loss during the 6-h test (<10%). The gap indicates that solar photolysis of trichloromethane is a complex process controlled by many factors, such that the QSAR method cannot replace laboratory efforts for precise kinetic evaluations. Triiodomethane was predicted to be the most vulnerable halomethane, with a log kP of 0.72 h1 or a half-life of 8 min.
3.2.
Haloacetic acids (HAAs)
Overall, HAAs were relatively stable in water. HAA concentrations for all nine species in the control 2 sample did not change significantly (<20%) throughout the experiment, Log(kp ) = - 1.18Es - 1.51 σ - 2.54, R2 =0.733, n=6 1.0 Observed Predicted
Log (kp ), 1/hour
0.0
-1.0
-2.0
which is within the coefficient of variation (CV) of the analytical method. Due to their protonated form in water, with low acidebase dissociation constants (pKa <3), HAAs are neither volatile nor hydrolysable. However, low levels of degradation of three brominated species (e.g., 21% for tribromoacetic acid) were observed in May 2005. The pseudofirst-order rate constants for monobromo-, dibromochloro-, and tribromoacetic acid were calculated to be 0.011, 0.007 and 0.038 h1, respectively. QSAR analysis of the photolysis results for the HAAs is presented for the three species with experimental results; the kP values matched the QSAR-predicted data. In an earlier study, kP values were reported for bromochloro- (1.84 104 h1), dibromo- (6.91 104 h1), trichloro- (2.3 104 h1), and tribromoacetic acid (3.0 103 h1) under simulated sunlight conditions, whereas monochloro- and monobromoacetic acid photolysis was not observed (Lifongo et al., 2004). Although there were differences in the absolute values between the two studies, the trends were similar (i.e., brominated HAAs were more susceptible than their chlorinated analogs). The half-life of triiodoacetic acid, which is probably the most susceptible HAA, was predicted by QSAR to be 11 s (Table 2). Zhang and Minear (2002) estimated the half-life of triiodoacetic acid due to hydrolysis to be 0.014 day (i.e., 20 min). Either way, this HAA is expected to be highly labile.
3.3.
Haloacetaldehydes (HAs)
Photolysis of dihalogenated acetaldehydes containing chlorine was not observed under the conditions examined (data not shown). However, dibromoacetaldehyde appeared to undergo some photolysis (kP ¼ 0.0045 h1). The brominated trihalogenated acetaldehydes (in particular, tribromoacetaldehyde) exhibited low-to-moderate levels (5e38%) of degradation relative to the control 2 sample. Based on a CV of 20%, only the degradation of the tribromoacetaldehyde was greater than the analytical variability. After accounting for loss in the control 2 samples, the pseudo-first-order rate constants for bromodichloro-, dibromochloro-, and tribromoacetaldehyde were estimated to be 0.011, 0.017, and 0.070 h1, respectively. The degradation of haloacetaldehydes under photolysis increased with increasing degree of bromine incorporation. QSAR analysis for the haloacetaldehydes was based only on the experimental data from the three trihalogenated species. Each of the chlorinated and brominated haloacetaldehydes was predicted to undergo photodegradation, with first-order rate constants ranging from 0.005 to <0.1 h1. The QSARpredicted value for dibromoacetaldehyde agreed with the experimental result with an RSD of 28%. Triiodoacetaldehyde was estimated to be more sensitive than the other haloacetaldehydes, with a half-life of w8 min under the sunlight conditions of this study (Table 2).
3.4.
Halopropanones (HKs)
CH2 I 2
C H 2 I Br
C H 2 I Cl
CHClBr 2
CHBr 3
CHCl 2 Br
CHCl 3
CH 2 Br 2
CH2 ClBr
CH2 Cl2
C H 3 Br
C H 3 Cl
-3.0
Fig. 2 e Comparison of experimental results (-) and QSAR predictions (>) of kP for halomethanes.
Two haloketones, 1,1-dichloro- and 1,1,1-trichloropropanone, were investigated in this study. Both exhibited low levels of degradation (e.g., 14% for the dichlorinated species, where control 2 only differed from control 1 by 3%). Their kP values were estimated to be 0.022 and 0.008 h1 for 1,1-dichloro- and
3407
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Halonitromethanes (HNMs)
Fig. 3 shows the photolysis of three HNM species examined in this study. Degradation typically appeared to occur more rapidly for this class of nitrogenous DBPs than for carbonaceous DBPs in this study. The pseudo-first-order rate constants ranged from 0.09 to w0.8 h1 for five species. A comparison of QSAR data and experimental data is shown in Fig. 4, with an excellent correlation between predicted and measured values (R2 ¼ 0.98). Increasing the number of halogens and/or the degree of bromination tended to increase the photolysis kinetics. Iodinated species are predicted to be more unstable than brominated HNMs (Table 2), and triiodonitromethane had the highest predicted value of kP of 621 h1, which is equivalent to a half-life of around 4 s.
10.0 5.0 0.0 0
2
4 Time (hr)
6
8
Fig. 3 e Solar photolysis of selected HNMs in organic-free water (May 2005) (BNM [ bromonitromethane, DCNM [ dichloronitromethane, TCNM [ trichloronitromethane) (RSDs between replicates analyzed at the five sample times were 0.7e6.3% for BNM and DCNM and were 6.5e33% for TCNM).
irradiation intensity (765 W/cm2) light source (Megan and Reinhard, 2007). These results demonstrate that sunlight intensity (actinic flux) plays an important role in the degradation kinetics of these compounds. The three nitrosamines with cyclic side groups (i.e., N-nitrosomorpholine, N-nitrosopyrrolidine, N-nitrosopiperidine) had some of the highest kP values among the nitrosamines, whereas the three nitrosamines with only methyl and/or ethyl side groups (i.e., NDMA, N-nitrosomethylethylamine, N-nitrosodiethylamine) had the lowest kP values among the nitrosamines. These results suggest that the nature of the side groups may impact the photolytic loss of each nitrosamine. QSAR analysis was not conducted for the nitrosamines due to the concurrent changes in steric structure and carbon numbers, as opposed to halogen substitution on single backbone structures for the halogenated DBPs. Log(kp ) = - 1.90Es - 2.85σ - 1.94, R2 =0.979, n=5 1.0 Observed Predicted
0.0
-1.0
Br3 C(NO2 )
ClBr2 C(NO2 )
Cl2 BrC(NO2 )
Cl3 C(NO2 )
Br2 CH(NO2 )
-2.0 Cl2 CH(NO2 )
With the solar intensity of 1150e1300 W/cm2, all eight nitrosamines were rapidly photolyzed (w99%) within 1 h. The halflives were approximately 8e10 min, with kP around 4.9 h1. The indoor tests, using simulated sunlight (1325 W/m2), exhibited more rapid photolysis, with rate constants of 5e10 h1 (Fig. 5), which corresponded to half-lives of 3e9 min. Both sets of experimental results were faster than an earlier study (half-lives ¼ 12e18 min), which employed a lower
ClBrCH(NO2 )
Nitrosamines (NAs)
BrCH2 (NO2 )
3.7.
15.0
ClCH2 (NO2 )
3.6.
TCNM TCNM control
20.0
Haloacetonitriles (HANs)
Over the 6-h test period in May, the HANs in control 2 degraded to varying levels: dichloroacetonitrile was reduced in concentration by 46%, bromochloroacetonitrile by 26%, and dibromoacetonitrile by 12%. This was probably due to hydrolysis (Croue´ and Reckhow, 1989). In addition, photolysis of the HANs was clearly observed. The kP values of bromochloro- and dibromoacetonitrile were calculated to be 0.329 and 0.324 h1, respectively, whereas dichloroacetonitrile’s photolytic loss rate was low (kP ¼ 0.063 h1) after accounting for the hydrolysis effect. QSAR analyses for experimental data and predictions of other HANs are presented in Table 2. Increasing the number of halogen atoms and/or replacing chlorine with bromine increased the photolysis decay of the HANs. The trend is consistent with previous findings that dichloroacetonitrile in water was photodegradable, whereas monochloroacetonitrile was resistant to photolysis (Lekkas and Nikolaou, 2004). Triiodoacetonitrile is probably the HAN that can be photolyzed most rapidly, with an estimated half-life of less than 8 s (Table 2).
DCNM DCNM control
25.0
Log (kp ), 1/hour
3.5.
BNM BNM control
HNM ( μg /L)
1,1,1-trichloropropanone, respectively. Although the trihalogenated species had a somewhat lower (not higher) rate of photolysis than the dihalogenated propanone, both values were low. Moreover, 1,1,1-trichloropropanone underwent a loss (12%) in the control 2 sample, which may have been due to hydrolysis, so determination of the rate of decay due to photolysis required separating out the other loss mechanism. QSAR was not conducted for this group because photolysis rate constants were only determined for two chlorinated species.
Fig. 4 e Comparison of experimental results (-) and QSAR predictions (>) of kP for halonitromethanes.
3408
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Nanopure Water Effluent #1 Filtered Effluent #2 Filtered
Effluent #1 Unfiltered Effluent #2 Unfiltered 100 9.06/hr
5.16/hr
5.80/hr
7.14/hr
10.38/hr
NDEA
NMOR
10.20/hr
7.74/hr
NPIP
NPYR
kp/knanopure (%)
7.43/hr
50
0 NDBA
NDPA
NDMA
NMEA
Nitrosamine Species
Fig. 5 e Comparison of kP of nitrosamines in non-filtered and filtered EfOM to samples prepared in organic-free water under simulated sunlight (the numbers beside the bars denote the kP values in organic-free water) (NDMA: N-nitrosodimethylamine, NMEA: N-nitrosomethylethylamine, NDEA: N-nitrosodiethylamine, NDPA: N-nitrosodipropylamine, NDBA: N-nitrosodibutylamine, NMOR: N-nitrosomorpholine, NPYR: N-nitrosopyrrolidine, NPIP: N-nitrosopiperidine).
0.0
HAN HNM
-1.0
-2.0
-3.0
3Br
1Cl,2Br
2Cl,1Br
3Cl
2Br
1Cl,1Br
-4.0 2Cl
This study represents among the first investigations of the solar photolysis effect on a wide variety of carbonaceous and nitrogenous DBPs and related compounds under a set of common exposure conditions. All rates of DBP transformation are intended to be apparent rates rather than mechanistic rate constants. As such, they should be used in relationship to each other to identify more or less photo-reactive DBP classes and influences of halogen substitution. The QSAR-predicted values agreed with the experimental results with R2 values of 0.73 or higher. Although only brominated and chlorinated DBPs were commonly detected during wastewater treatment
HM HAA HA
1Br
Conclusion
1.0
1Cl
4.
(Krasner et al., 2009), including data on iodinated related chemicals from the literature with our observed brominated and chlorinated DBP data allowed us to demonstrate the importance of halogen substitution. Moreover, QSAR was used to predict kP values for fluorinated, chlorinated, brominated, and iodinated species not tested (some of which are not DBPs). Fig. 6 compares the QSAR estimates for five groups of halogenated DBPs. The photolysis kinetics appeared to be a function of functional group, halogen type(s) and number(s).
Log (kp), 1/hour
In other studies, it was found that the advanced oxidation of nitrosamines with free radicals could be mechanistically understood (in part) based on structureeactivity relationships (Mezyk et al., 2008). In the absence of sunlight, the average decrease in concentration of the nitrosamines in the control 2 sample was 9%, which was less than the analytical CV, indicating that no other mechanism(s) contributed to nitrosamine loss during these experiments. For nitrosamines spiked into EfOM, their first-order rate constants were around 50% below those spiked into organic-free water (Fig. 5). Typically, nitrosamines in effluent #2, which had higher TOC (16.5 mg/L), exhibited slower photolysis kinetics than those in effluent #1 (TOC ¼ 12.2 mg/L). In addition, there was (on average) an 11% higher kP (i.e., faster kinetics) for filtered water as compared to non-filtered samples. Thus, organic or particulate matter can impact the rate of photolysis. For example, increased turbidity can decrease the actual actinic flow in the solution.
Fig. 6 e Comparison of kP of five groups of halogenated DBPs under sunlight conditions in May 2005 in Tempe (AZ) as predicted by QSAR analysis (the x-axis denotes the number(s) and type(s) of halogen(s) in the compounds) (HM [ halomethane, HAA [ haloacetic acid, HA [ haloaldehyde, HAN [ haloacetonitrile, HNM [ halonitromethane).
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 4 0 1 e3 4 0 9
Including non-halogenated compounds, the nitrogenous DBPs (i.e., nitrosamines, HNMs, HANs) had higher photolytic rates than carbonaceous DBPs (i.e., THMs, HAAs, haloacetaldehydes, halopropanones), generally ranked as follows: nitroso (N-NO) > nitro (NO2) > nitrile (C^N) > carbonyl (C]O) > carboxyl (COOH). Although most HAAs and haloacetaldehydes did not degrade over the 6-h photolysis test period, near complete removals (99%) of nitrosamines were observed within 1 h of sunlight exposure. Within one group of DBPs, the photolysis kinetics usually increased with increasing number of halogens. Brominated DBPs were more susceptible than chlorinated DBPs, and iodinated species (some of which may be DBPs) are probably even more unstable than the brominated analogs.
Acknowledgements The authors thank the American Water Works Association Research Foundation (AwwaRF) and the USEPA for its financial, technical, and administrative assistance in funding and managing the project. The comments and views detailed herein may not necessarily reflect the views of AwwaRF, its officers, directors, affiliates or agents, or the views of the U.S. federal government. The project manager was Alice Fulmer. The photolysis setup was funded by NSF ATM 0530718. Thanks are also extended to Metropolitan staff who conducted various analyses to support this study, including Michael J. Sclimenti, Salvador J. Pastor, Russell Chinn, Cordelia J. Hwang, Yingbo C. Guo, Eduardo Garcia-Zayas, Donnie Morquecho, Wing Chow, Rachel M. Kelley, Mary Ann Viernes, Sikha Kundu, Chih Fen Tiffany Lee, and Jacob Nikonchuk. In addition, the authors thank the participating utilities for supplying treated wastewater samples for testing.
references
Allston, T.D., Fedyk, M.L., et al., 1978. Photoabsorption spectra of some halo-substituted methyl nitrogen oxides. J. Photochem. 9 (2e3), 116e118. Bayes, K.D., Friedl, R.R., Sander, S.P., Toohey, D.W., 2003. Measurements of quantum yields of bromine atoms in the photolysis of bromoform from 266 to 324 nm. J. Geophys. Res. 108 (D3), 4095. doi:10.1029/2002JD002877. Castro, C.E., Belser, N.O., 1981. Photohydrolysis of methyl bromide and chloropicrin. J. Agric. Food Chem. 29, 1005e1008. Chen, B., Westerhoff, P., Krasner, S.W., 2008. Fate and transport of wastewater-derived disinfection by products in surface waters. In: Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y. (Eds.), Disinfection By-products in Drinking Water: Occurrence, Formation, Health Effects, and Control. American Chemical Society, Washington, DC, pp. 257e273. Chen, J.W., Peijnenburg, W., et al., 2001. Is it possible to develop a QSPR model for direct photolysis half-lives of PAHs under irradiation of sunlight? Environ. Pollut. 114 (1), 137e143. Croue´, J.-P., Reckhow, D.A., 1989. Destruction of chlorination byproducts with sulfite. Environ. Sci. Technol. 23 (11), 1412e1419. Hammett, L.P., 1935. Some relations between reaction rates and equilibrium constants. Chem. Rev. 17 (1), 125e136.
3409
Hansch, C., Leo, A., Hoekman, D., 1995. Exploring QSAR: Hydrophobic, Electronic, and Steric Constants. In: Heller, S.R. (Ed.), American Chemical Society Professional Reference Book Washington, DC. Ho, T.-F.L., Bolton, J.R., Lipczynska-Kochany, E., 1996. Quantum yields for the photodegradation of pollutants in dilute aqueous solution: phenol, 4-chlorophenol and N-nitrosodimethylamine. J. Adv. Oxid. Technol. 1 (2), 170e178. Jones, C., Carpenter, L.J., 2005. Solar photolysis of CH2I2, CH2ICl, and CH2IBr in water, saltwater, and seawater. Environ. Sci. Technol. 39 (16), 6130e6137. Krasner, S.W., Westerhoff, P., Chen, B., Amy, G., Nam, S.-N., Chowdhury, Z.K., Sinha, S., Rittmann, B.E., 2008. Contribution of Wastewater to DBP Formation. Awwa Research Foundation, Denver, USA. Krasner, S.W., Westerhoff, P., Chen, B., Rittmann, B.E., Amy, G., 2009. Occurrence of disinfection byproducts in United States wastewater treatment plant effluents. Environ. Sci. Technol. 43 (21), 8320e8325. Lifongo, L.L., Bowden, D.J., et al., 2004. Photodegradation of haloacetic acids in water. Chemosphere 55 (3), 467e476. Lekkas, T.D., Nikolaou, A.D., 2004. Degradation of disinfection byproducts in drinking water. Environ. Eng. Sci. 21 (4), 493e506. Mabey, W., Mill, T., 1978. Critical review of hydrolysis of organic compounds in water under environmental conditions. J. Phys. Chem. Ref. Data 7 (2), 384e415. Megan, H.P., Reinhard, M., 2007. Photochemical attenuation of N-nitrosodimethylamine (NDMA) and other nitrosamines in surface water. Environ. Sci. Technol. 41 (17), 6170e6176. Mezyk, S.P., Landsman, N.A., Swancutt, K.L., Bradford, C.N., Cox, C.R., Kiddle, J.J., Clore, T.J., 2008. Free radical chemistry of advanced oxidation process removal of nitrosamines in waters. In: Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y. (Eds.), Disinfection By-products in Drinking Water: Occurrence, Formation, Health Effects, and Control. American Chemical Society, Washington, DC, pp. 319e333. Mizgireuv, I.V., Majorova, I.G., Gorodinskaya, V.M., Khudoley, V.V., Revskoy, S.Y., 2004. Carcinogenic effect of N-nitrosodimethylamine on diploid and triploid zebrafish (Danio rerio). Toxicol. Pathol. 32 (5), 514e518. Nicole, I., De Laat, J., et al., 1991. Degradation of trihalomethanes in dilute aqueous solution by UV irradiation e determination of the quantum yield of photolysis at 253.7 nm. Environ. Technol. 12 (1), 21e31. Peijnenburg, W.J.G.M., de Beer, K.G.M., den Hollander, H.A., Stegeman, M.H.L., Verboom, H., 1993. The development of quantitative structure activity relationships for the direct photolysis of substituted haloaromatics in aqueous environments. Sci. Total Environ. 134 (Suppl. 2), 1397e1408. Plewa, M.J., Wagner, E.D., 2009. Quantitative Comparative Mammalian Cell Cytotoxicity and Genotoxicity of Selected Classes of Drinking Water Disinfection By-products. Awwa Research Foundation, Denver, USA. Rudra, A., Thacker, N.P., Pande, S.P., 2005. Hydrogen peroxide and ultraviolet irradiations in water treatment. Environ. Monit. Assess. 109, 189e197. Stefan, M.I., Bolton, J.R., 2002. UV direct photolysis of N-nitrosodimethylamine (NDMA): kinetic and product study. Helv. Chim. Acta 85, 1416e1426. Trehy, M.L., Yost, R.A., Miles, C.J., 1986. Chlorination byproducts of amino acids in natural waters. Environ. Sci. Technol. 20 (11), 1117e1122. Zhang, X., Minear, R.A., 2002. Decomposition of trihaloacetic acids and formation of the corresponding trihalomethanes in drinking water. Water Res. 36, 3665e3673. Zou, P., Shu, J., Sears, T.J., Hall, G.E., North, S.W., 2004. Photodissociation of bromoform at 248 nm: single and multiphoton processes. J. Phys. Chem. A108, 1482e1488.