Modeling the Formation of N-Nitrosodimethylamine (NDMA) from the

Oct 13, 2006 - This paper presents mechanistic studies on the formation of NDMA, a newly identified chloramination disinfection byproduct, from reacti...
7 downloads 9 Views 184KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 7290-7297

Modeling the Formation of N-Nitrosodimethylamine (NDMA) from the Reaction of Natural Organic Matter (NOM) with Monochloramine† ZHUO CHEN AND RICHARD L. VALENTINE* Civil and Environmental Engineering, 4105 Seamans Center for the Engineering Arts and Sciences, The University of Iowa, Iowa City, Iowa 52242-1527

This paper presents mechanistic studies on the formation of NDMA, a newly identified chloramination disinfection byproduct, from reactions of monochloramine with natural organic matter. A kinetic model was developed to validate proposed reactions and to predict NDMA formation in chloraminated water during the time frame of 1-5 days. This involved incorporating NDMA formation reactions into an established comprehensive model describing the oxidation of humic-type natural organic matter by monochloramine. A rate-limiting step involving the oxidation of NOM is theorized to control the rate of NDMA formation which is assumed to be proportional to the rate of NOM oxidized by monochloramine. The applicability of the model to describe NDMA formation in the presence of three NOM sources over a wide range in water quality (i.e., pH, DOC, and ammonia concentrations) was evaluated. Results show that with accurate measurement of monochloramine demand for a specific supply, NDMA formation could be modeled over an extended range of experimental conditions by considering a single NOM sourcespecific value of θNDMA, a stoichiometric coefficient relating the amount of NDMA produced to the amount of NOM oxidized, and several kinetic parameters describing NOM oxidation. Furthermore, the oxidation of NOM is the rate-limiting step governing NDMA formation. This suggests that NDMA formation over a 1-5 day time frame may be estimated from information on the chloramine or free chlorine demand of the NOM and the source-specific linear relationship between this demand and NDMA formation. Although the proposed model has not yet been validated for shorter time periods that may better characterize the residence time in some distribution systems, the improved understanding of the important reactions governing NDMA formation and the resulting model should benefit the water treatment industry as a tool in developing strategies that minimize NDMA formation.

Introduction Many utilities have implemented the use of chloramines as a secondary disinfectant in drinking water distribution †

This article is part of the Emerging Contaminants Special Issue. * Corresponding author phone: (319) 335-5653; fax: (319) 3355660; e-mail: [email protected]. 7290

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006

systems because this generally results in lower concentrations of most regulated disinfection byproducts (DBPs), which are frequently produced in unacceptable amounts when free chorine is utilized (1-5). Recently however, N-nitrosodimethylamine (NDMA), a potent carcinogen, has been identified as a new disinfection byproduct in drinking water that can be formed by several reactions including those involving chloramines and free chlorine (6, 7). Mechanistic studies have shown that monochloramine, the dominant form of chloramines normally occurring (8), can react with dimethylamine (DMA), a relatively ubiquitous substance in water, to produce NDMA (9, 10). This mechanism involves the oxidation of DMA to unsymmetrical dimethylhydrazine (UDMH) as an intermediate, which in turn, is further oxidized to NDMA, and explains why NDMA formation is generally slow and associated with chloramination. A reaction of free chlorine with dimethylamine has also been observed, which is approximately 1 order of magnitude slower than that with monochloramine (10). Choi and Valentine (11) showed that the oxidation of nitrite by free chlorine in the presence of DMA could also result in NDMA formation. However, the importance of this reaction in the formation of NDMA in drinking water is probably very small because of generally low nitrite concentrations. While DMA has been implicated as a potentially important precursor in natural waters, studies have indicated that naturally occurring levels of DMA may be insufficient to account for the amounts of NDMA typically produced in chloraminated drinking water (12). It has been suggested that natural polymeric organic matter may be an important source of NDMA precursors in drinking water (13, 14). The reactivity of natural organic matter (NOM) with free chlorine and monochloramine to form halogenated DBPs is widely recognized (15-18). More precisely, it is generally considered that it is the humic fractions of NOM that account for DBP formation (19, 20). It is logical to hypothesize that the structurally complex humic substances could also serve as a source of NDMA precursors. The C/N ratio of humic substances generally varies from 10:1 to 100:1 (21, 22). Carbon content varies in the range of 50∼80%, while nitrogen usually contributes 1∼5%. Given a typical dissolved organic carbon content of several milligrams/L, there is ample organic nitrogen content to account for the nanograms/L levels of NDMA that have been observed in chloraminated drinking water (6). The nature of the reactions however, that might lead to NDMA formation from these potential precursors have not been studied. An improved understanding is critical to the development of effective strategies to minimize NDMA formation as a consequence of chloramination. The objective of this research was to elucidate reaction pathways and develop a kinetic model describing NDMA formation in the presence of monochloramine and natural organic matter (NOM). Such a model provides a valuable tool in validating proposed reactions including the premise that humic type substances are important precursors. Critical to the approach taken is the simple hypothesis that the rate of NDMA formation is limited by the rate of NOM oxidation. Therefore, understanding the reaction mechanism and kinetics of NOM oxidation becomes paramount in the subsequent description of how NDMA forms. As a starting point in the elucidation of reaction pathways and development of a model describing NDMA formation, the reaction model describing the oxidation of humic-type NOM by monochloramine developed by Duirk et al. (23-25) was adopted and then modified to account for NDMA formation. The proposed reaction model 10.1021/es0605319 CCC: $33.50

 2006 American Chemical Society Published on Web 10/13/2006

was validated using three different sources of NOM over a wide range of reaction conditions. Model Development. The formation of NDMA is a relatively slow process generally occurring on a time scale of days (12, 26), which is comparable to the time scale of monochloramine loss from autodecomposition (27) and from oxidation of NOM (28). The monochloramine-NOM reaction model proposed by Duirk (23-25) was used to describe these parallel processes to predict both the concentration of monochloramine and the amount of NOM oxidized with time. A schematic of the proposed model is shown in Figure S.1. Table S.1 tabulates the essential equations and constants included in the monochloramine-NOM-NDMA reaction model finally adopted. Duirk et al. (23) modified a description of monochloramine autodecomposition (29) (reactions 1-10 in table S.1) by incorporating simple second-order kinetic description of NOM oxidation involving two types of reactive sites (reactions 11-12 in table S.1) expressed as a fraction in terms of the molar amount of monochloramine that is reduced by the NOM normalized to the total molar dissolved organic carbon content. Site type S1 represents a comparatively small fraction of reactive dissolved organic carbon (DOC) that reacts very rapidly producing an “initial chloramine demand”. Site type S2, comprising more than 90% of the reactive DOC, reacts quite slowly (over days) accounting for most of the monochloramine loss typically observed over a 5-day period. While the rate constants k1 and k2 are associated with these two types of reactive sites respectively, constant k1 is simply set conveniently high so that the demand due to site type S1 is accounted for rapidly, within an hour. The value of k2, on the other hand, is critical to predicting the slow loss of monochloramine with time. NOM oxidation is primarily attributed to reaction 12 in table S.1 involving a direct reaction of DOC with HOCl that exists in minute amounts in chloramine solutions, produced from the hydrolysis of monochloramine (reaction 2 in table S.1). Duirk et al. (23) validated the model using several whole water sources, NOM concentrates obtained by reverse osmosis, and fractionated humic isolates. The model could account for such dependencies as pH, ammonia, DOC, and monochloramine concentrations. Results were consistent with the assumption that humic-type substances were the dominant reactive source of NOM. We modified this model to account for NDMA formation by assuming that the rate of NDMA formation was proportional to the rate of NOM oxidation. This is similar to the approach recently taken to describe the formation of dichloroacetic acid from the oxidation of NOM (30). We incorporated a simple stoichiometric coefficient θNDMA to linearly correlate the formation of NDMA with the oxidation of NOM, similar to what previous researches have done to estimate the formation of trihalomethanes (THMs) and haloacetic acids (HAAs) from NOM (31, 32). This is equivalent to the assumption that the ratio of NDMA formed to NOM oxidized as measured by the molar quantity of monochloramine reduced by the NOM, is a constant θNDMA over the entire course of the reaction (eq 1):

d[DOC]OX d[NDMA] ) θNDMA dt dt

(1)

According to reactions 11 and 12 in table S.1, the rate of NDMA formation can, therefore, be expressed as eq 2 in terms of model parameters and variables:

d[NDMA] ) θNDMA{ k1[NH2Cl][DOC] × S1 + dt k2[HOCl][DOC] × S2} (2)

S1 ) DOC short-term reactive site fraction, S2 ) DOC longterm reactive site fraction, k1, k2 ) reaction rate constants, and θNDMA ) stoichiometric coefficient All constants characterizing monochloramine autodecomposition were obtained from literature (sources shown in table S.1). Only k1, S1, k2, S2 and θNDMA were determined for this model as NOM source specific parameters. The model was expressed as a system of ordinary differential equations (ODEs), the whole set of which was solved using Scientist (33). Scientist uses a modified Powell algorithm to minimize the unweighted sum of the squares of the residual error between the predicted and experimentally determined values to estimate the model parameters. The experimental monochloramine decay data in the presence of tested NOM at pH 7.0 and C/N molar ratio of 0.7 were used to estimate the parameters including k1, S1, k2, and S2, which were further verified at different pH values and various DOC or ammonia concentrations. The shortterm reaction parameters k1 and S1 and the long-term reaction parameters k2 and S2 were estimated independently. The short-term reaction parameters were adjusted to account for the 5-10% rapid monochloramine loss (the initial demand) during the initial phase of the reactions typically occurring within 1 h. The remaining data set was used to estimate the parameters characterizing the long-term slow reaction. As an alternative to estimating S1 and S2 for each NOM source, the total reactive site fraction ST, which is the sum of S1 and S2, was also estimated by titration with free chlorine (23). The value of NDMA formation coefficient θNDMA was estimated after the parameters characterizing NOM oxidation were determined. This was accomplished by dividing the amount of NDMA formed by the amount of NOM oxidized as calculated by the model, and not by nonlinear regression. Final estimation of parameters and standard deviations were determined by simply averaging all parameters determined for each source specific data set.

Experimental Section This study utilized extracted and concentrated natural organic matter (NOM) as well as whole natural water samples from three sources. NOM was concentrated from the Iowa River (IRW) located in Iowa and from Valentine Pond in the Keweenaw Peninsula of the Upper Michigan Peninsula (UPW) using a RealSoft PROS/2S reverse osmosis (RO) unit (Stone Mountain, GA). While the Iowa River is highly impacted by agricultural practices that result in significant additions of inorganic and organic nitrogen, the latter source is located in a largely uninhabited and pristine area. Therefore, comparisons are expected to yield valuable information on possible human/animal sources of NDMA precursors. Whole water samples from these sources as well as from the Cedar River (CRW) were also used in this study. The Cedar River is also located in Iowa and is significantly impacted by agricultural practices (34, 35). All experiments were conducted in batch reactors (oneliter capacity clear Pyrex bottles with PTFE screw caps) utilizing additions of preformed monochloramine stock solutions. This was done to avoid artifacts caused by reactions between NOM and free chlorine which may exist if monochloramine was formed in-situ. Two types of NDMA formation experiments were conducted. Studies aimed at elucidation of reaction mechanisms used monochloramine dosages of 0-0.05 mM, which span a range typical of chloramination practices. The buffer capacity was augmented by addition of 4 mM sodium bicarbonate and the pH values were adjusted by concentrated HCl and NaOH to the range from 6.0 to 9.0. Variable free ammonia concentrations were obtained by adding excess amount of NH4Cl so that Cl/N molar ratio ranged from 0.1 to 0.7. Duirk (24) demonstrated that ionic strength does not affect the monochloramine loss VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7291

TABLE 1. Model Estimated Parameters with Standard Deviation for Tested NOM Sources Including IRW, UPW. and CRW NOM sources

IRWa

UPWb

CRWc

k1(M-1h-1) S1e k2(M-1h-1) S2e S1+S2e STd,e θNDMAf

1.21 × 104 ( 6.94 × 103 0.011 ( 0.005 5.92 × 105 ( 3.02 × 104 0.53 ( 0.06 0.54 ( 0.06 0.59 2.85 × 10-5 ( 8.34 × 10-6

1.12 × 104 ( 8.90 × 103 0.009 ( 0.004 4.88 × 105 ( 2.32 × 104 0.48 ( 0.05 0.57 ( 0.05 0.45 1.67 × 10-5 ( 5.53 × 10-6

1.43 × 104 ( 3.13 × 103 0.015 ( 0.006 6.01 × 105 ( 6.43 × 104 0.62 ( 0.05 0.63 ( 0.06 0.65 3.23 × 10-5 ( 2.24 × 10-6

a IRW: Iowa River water, n ) 10. b UPW: Valentine Pond water in Upper Peninsular of Michigan, n ) 7. c CRW: Cedar River water, n ) 4. d Total reactive sites determined by free chlorine titration. e Reactive organic carbon/total organic carbon, dimensionless. f Mole of NDMA formed/ equivalent mole of NOM oxidized as measured by monochloramine reduced, dimensionless.

in the presence of NOM. However, ionic strength was adjusted at 8 mM by sodium perchlorate because its influence on NDMA formation was not known. NDMA and monochloramine concentrations were monitored up to 5 days. Additional experiments were conducted to determine the NDMA formation potential using a high dosage of monochloramine (1 mM) and reaction conditions (pH 7.0, reaction time of 7 days) presumed to exhaust all NDMA precursors. This approach was based upon a similar logic behind the THM formation potential test (36-38) and was first proposed by Mitch et al. (12) for the determination of the NDMA formation potential in water. NOM total reactive site fraction ST was determined using the free chlorine titration method proposed by Duirk (23). In this approach, it is assumed that the ultimate free chlorine demand is due to reactions with NOM and is the same as would be exerted by excess monochloramine if sufficient time were allowed for the reaction with NOM to go to completion in the absence of losses due to auto-oxidation (constraints that cannot be met experimentally). Free chlorine was initially dosed at 0.2 mM in the presence of 2.0 mgDOC/L of IRW, UPW, and CRW NOM at pH 9.0, a value shown to exceed the NOM demand and minimize the third-order free chlorine decomposition involving Cl2O‚H2O and ClO2(39). The remaining free chlorine concentration was monitored for up to 20 days, and demand attributed to reaction with NOM was assumed to be complete when three consecutive samples over a period of 5-10 days did not change. The total reactive site concentration was calculated as the difference between the loss of free chlorine in the control and the final free chlorine concentration in solutions containing NOM. In general, the free chlorine loss in the control sample was less than 5% of the initial amount added. Free chlorine residual and monochloramine were determined using the Standard Method 4500-Cl F DPD-FAS titrimetric method (40). Ammonia was measured using an Orion Scientific 960 meter with an Orion 95-12 gas sensing membrane electrode (Orion Research, Inc., Beverly, MA.). DOC was determined using Shimadzu TOC 5000 (Shimadzu Scientific, Columbia, MD). UV absorbance and spectral characteristics of the natural organic mater were obtained with a Shimadzu UV1601 dual beam spectrophotometer. NDMA analysis was accomplished using solid-phase extraction, isotope dilution gas chromatography/mass spectrometer (GC/MS) methodology similar to that originally proposed by Taguchi et al. (41) and used by others (42, 43). Briefly, 1 liter of water sample was extracted with 200 mg of carbonaceous polymeric beads (Ambersorb 572, Aldrich) by shaking for 1 h at 200 RPM. The Ambersorb beads were then filtered onto a glass fiber filter. After air-dried for roughly 30 min, they were transferred to a 2 mL amber vial and 0.4 mL methylene chloride was added to re-extract the adsorbate. The methylene chloride phase was then used for gas chromatography injection. An 8 µL aliquot of methylene chloride extract was injected into Varian CP-3800 gas 7292

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006

chromatography coupled with Saturn 2200 MS/MS. A DB1701 capillary column was used for NDMA determination. Identifications were based on the comparison of the experimental full scan mass spectra with standard NDMA mass spectrum. d6-NDMA was used as the internal standard, and quantifications were accomplished using selective ion monitoring (SIM) based on the characteristic mass peaks of NDMA in CI mode, specifically m/z 75 for NDMA and m/z 81 for d6-NDMA. The MDL was determined to be at 0.67 ng/L.

Results and Discussion Precursor Recovery Efficiency. Recovery of NDMA precursors in the RO concentrate is an issue since we wished to study NDMA formation as a function of DOC concentration, which was varied by adding different amounts of the concentrate. A significant loss as a consequence of the concentrating process could result in remaining precursor material that is not representative of the whole water samples. Kitis et al. (16) showed that reverse osmosis isolation appears to maintain the integrity and reactivity of the NOM with respect to chlorine demand and DBP formation suggesting that this should also be true for NDMA formation especially if the same precursor material is involved. We determined the recovery of NOM and NDMA precursors as measured by DOC, SUVA254, SUVA272, and NDMA formation potential as previously described. Figure S.2 shows that the NDMA formation potential of the original whole Iowa River (IRW) sample steadily increased with time attaining a value of 112 ng/L at the end of the 7-day test period. In comparison, the concentration-time profile was similar in water sample prepared using the IRW NOM concentrate at the same DOC concentration. Approximately 100 ng/L of NDMA was formed in the laboratory-prepared water at the end of the 7-day test period indicating a 90% recovery of NDMA precursors. This is consistent with the high recovery of DOC and comparable values of SUVA254 and SUVA272 listed in table S.2. We conclude that the RO concentration method produced concentrated NOM from natural water with excellent preservation of the NOM characteristics and reactivity with respect to NDMA formation. Estimated Model Parameters. The comprehensive reaction model resolves monochloramine loss into two pathways: auto-decomposition and oxidation of NOM. The initial modeling activity required determining the parameters that characterize monochloramine loss independently of the value of θNDMA, which is only used to predict NDMA formation. The final estimated model parameters for each particular NOM source, obtained by averaging parameter estimates for all data sets for the same NOM source, are listed in Table 1. The model estimation of NOM reactive sites are consistent with the total reactive sites determined by the free chlorine titration methodology. This shows that titrating the NOM with free chlorine serves as a good independent method to determine reactive site fractions.

FIGURE 1. Monochloramine residual loss in the presence of NOM ([NH2Cl]0 ) 0.05 mM, I ) 8 mM, line represents model prediction). (A) IRW concentrate, pH ) 7.0. (B) IRW concentrate, DOC ) 3.4 mg/L. (C) IRW concentrate, pH ) 7.0, DOC ) 3.4 mg/L, [NH3] ) 0.07∼0.5 mM. All the rate constants as well as the site fraction values were remarkably similar, within a factor of 25% of each other (Table 1), indicating very similar NOM oxidation characteristics among tested NOM sources. The value of θNDMA, however, varied by a factor of 2, ranging from a high of 3.23 × 10-5 for source CRW NOM to a low value of 1.67 × 10-5 for water contain UPW NOM. While the values for CRW and IRW derived NOM were statistically similar, the value of θNDMA for UPW NOM was significantly smaller. This difference presumably reflects differences in the nature of the NDMA

precursors. Perhaps this difference is indicative of differences in NOM characteristics, or the presence of other types of precursors possibly associated with agricultural discharges into the two rivers. Comparison of Predicted and Measured Results. The ability of the model to accurately predict NOM oxidation is measured by the fit between the measured and predicted monochloramine concentrations during the investigated time frame of 1-5 days in the presence of sufficient DOC to significantly increase its loss relative to that due to the VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7293

FIGURE 2. Linear relationship between NDMA formation and NOM oxidation ([NH2Cl]0 ) 0.05 mM, I ) 8 mM, NOM oxidized is equivalent to NH2Cl reduced by NOM, θNDMA is mole of NDMA formed per mole of NOM oxidized). (A) IRW concentrate DOC ) 3.4 mg/L. (B) UPW concentrate, DOC ) 3.4 mg/L. (C) CRW concentrate, DOC ) 4.5 mg/L. autodecomposition pathway. Using Aldrich humic acids as an NOM source, Vikesland et al. (28) demonstrated that NOM predominantly reacts with monochloramine as a reductant which exerts additional monochloramine demand. The predicted monochloramine concentrations were found to be very close to the measured values determined in the presence of IRW and UPW NOM as a function of DOC at pH 7.0 (Figures 1. A and figure S.3) over a range of DOC concentrations from 0 to 6.8 mg/L. A pronounced influence of pH on the monochloramine loss in the presence of NOM was observed experimentally and was also captured by the model. Figures 1. B, figure S.4, and figure S.5 show the loss in monochloramine in the presence of IRW, UPW, and CRW NOM respectively at pH 7294

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006

values from 6 to 9. The acceleration of monochloramine loss with decreasing pH is attributed to acid-catalyzed monochloramine autodecomposition (reaction 5 in table S.1) (44) and to an increased rate of NOM oxidation caused by the increase in the concentration of HOCl with decreasing pH. Consistent with the role of HOCl in NOM oxidation, the rate of monochloramine reduction by NOM decreased with increasing the concentrations of ammonia (Figure 1. C), since excessive ammonia reduces the formation of HOCl which is derived from the hydrolysis of monochloramine. With the establishment of the model’s capability to describe the loss of monochloramine and NOM oxidation using the three NOM sources, NDMA formation was studied next. This required quantifying the amount of NOM oxidized

FIGURE 3. Model prediction of NDMA formation at various pH and ammonia concentration from the reactions between monochloramine and NOM ([NH2Cl]0 ) 0.05 mM, I ) 8 mM, line represents model prediction). (A) Effect of pH, IRW concentrate DOC ) 3.4 mg/L. (B) Effect of pH, UPW concentrate DOC ) 3.4 mg/L. (C) Effect of pH, CRW DOC ) 4.5 mg/L, (D) Effect of ammonia, IRW DOC ) 3.4 mg/L, pH ) 7.0. (as equivalent amount of monochloramine reacted with NOM) from the difference between the measured monochloramine loss and the calculated amount attributed to autodecomposition. Figure 2 shows a linear relationship between NDMA formed and NOM oxidized for all three NOM sources at three pH values of 6, 7, and 8. The three sourcedependent θNDMA values shown in Table 1 resulted from the best fit linear regression for the pooled pH-dependent data sets. This linear relationship is consistent with the assumption that the rate of NDMA formation is given by eq 1 and that the value of θNDMA is a source-specific constant that is independent of pH and hence the rate of NOM oxidation. Figure S.6 and S.7 compare measured and predicted NDMA formation from 1 to 5 days in IRW and UPW NOM at DOC concentrations of 1.7-6.8 mg/L, which are in a DOC range representative of many water supplies (45). A higher concentration of DOC resulted in higher NDMA formation. During the early stage of the reaction course, the model slightly underestimated NDMA production at low DOC levels and overestimated it at high DOC levels. But overall, the model predicted NDMA formation quite well. Figure 3A-C shows the influence of pH from 6.0 to 9.0. A good match between experimental data and model prediction is again indicated during the examined time frame. The effect of free ammonia on NDMA formation was also investigated using a fixed initial monochloramine concentration and chlorine to ammonia molar ratio ranging from 0.1 to 0.7 (Figure 3D). Results show that increasing ammonia reduces the rate of NDMA formation, consistent with the expected influence of ammonia on the rate of NOM oxidation. Although ionic strength was found to have no effect on monochloramine loss in the presence of aquatic NOM (24), it may potentially have an impact on NDMA formation. Experiments conducted using 0.05 mM NH2Cl at pH 7.0 and an IRW NOM DOC concentration of 3.4 mg/L, however,

showed no dependence of NDMA yield on ionic strength, which was varied from 8 mM to 100 mM with sodium perchlorate (Figure S.8). The success of the model in predicting NDMA formation over a wide range of reaction conditions, especially the influence of pH and ammonia, shows that NDMA formation is limited by the oxidation of NOM by the protonated chorine species, HOCl. Additionally, there is no evidence for a pH dependent pathway leading to NDMA formation that involves NOM oxidation products. Otherwise, the yield of NDMA would likely be pH dependent in a manner that cannot be explained simply by the proposed reactions and kinetics of NOM oxidation. The model described NDMA formation over a 1-5 day period in the presence of DOC from three water sources including two heavily impacted by agricultural practices and one from a pristine source. The reader is cautioned that applicability over shorter time periods that may be important has not been determined. While key estimated parameters did differ among the source waters, all constants were remarkably similar except for θNDMA values. The value of this coefficient for NDMA formation in NOM obtained from a pristine (UPW) source was approximately one-half the value of those determined for NOM from the agriculturally impacted waters. This may be attributable to differences in the relative contribution of various NDMA precursors formed by the oxidation of NOM or perhaps to factors associated with elevated concentrations of other specific NDMA precursor compounds associated with elevated organic nitrogen content expected in the non-pristine water sources. The results of this study provide important insight into the process and reactions that control NDMA formation in drinking water. The proposed model should allow estimation of NDMA formation from the reactions between NOM and monochloramine without detailed probing into the exact VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7295

reaction mechanisms but only measurement of monochloramine/chlorine demand once the source specific relationship between this demand and NDMA formation is established.

Acknowledgments This research was funded by USGS project no. 2002IA16G, award no. 02HQGRO136.

Supporting Information Available Table S.1 lists essential reactions in the model, while table S.2 lists the DOC, SUVA, and NDMA formation potential recoveries after RO concentration. Figure S.1 shows the model schematics, and figure S.2 compares the NDMA formation potential profile in both IRW whole water and reconstituted RO concentrated water. Additional figures (S.3-S.8) show the model prediction of NH2Cl residual and NDMA concentration from the three tested NOM sources under different conditions (i.e., pH, DOC, ionic strength). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Wallis-Lage, C.; et al. Disinfection strategies: not as simple as they used to be. In WEFTEC.04, Conference Proceedings, Annual Technical Exhibition & Conference, 77th, New Orleans, LA, United States, Oct. 2-6, 2004; Water Environment Federation: Alexandria, VA, 2004; pp 892-903. (2) Richardson, S. D.; Thruston, A. D., Jr.; Collette, T. W.; Caughran, T. V.; Patterson, K. S.; Lykins, B. W., Jr. 1998. Alternative Disinfectants for Drinking Water. In Chlorine and Chlorine Compounds in the Paper Industry, [an International Symposium], Chicago, August, 1995; Turoski., V. , Ed.; Ann Arbor Press: Chelsea, MI, 1998; pp 183-192. (3) EPA. Alternative Disinfectants and Oxidants Guidance Manual; EPA 815-R-99-014; United States Environmental Protection Agency, Office of Water: Washington, DC, 1999. (4) Carlson, M.; Hardy, D. Controlling DBPs with monochloramine. J. Am. Water Works Assoc. 1998, 90 (2), 95-106. (5) Diehl, A. C.; et al. DBP formation during chloramination. J. Am. Water Works Assoc. 2000, 92 (6), 76-90. (6) Valentine, R. L.; Choi, J.; Chen, Z.; Barrett, S. E.; Hwang, C.; Yingbo Guo, Y.; Wehner, M.; Fitsimmons, S.; Andrews, S. A.; Werker, A. G.; Brubacher, C.; Kohut, K. Factors Affecting the Formation of NDMA in Water and Occurrence; AWWA Research Foundation: Denver, CO, 2005. (7) Mitch, W. A.; Sharp, J. O.; Trussel, R. R.; Valentine, R. L.; AlvarezCohen, L.; Sedlak, D. L. N-Nitrosodimethylamine (NDMA) as a drinking water contaminant: A review. Environ. Eng. Sci. 2003, 20(5), 389-404. (8) Karimi, A. A.; et al. A chloramine conversion evaluation of the nation’s most complex water system. In ProceedingssAnnual Conference, American Water Works Association; AWWA: Denver, CO, 2002; pp 590-599. (9) Choi, J.; Valentine, R. L. Studies on the formation of Nnitrosodimethylamine (NDMA) in drinking water: a new chloramination disinfection by-product. In ProceedingssAnnual Conference, American Water Works Association; AWWA: Denver, CO, 2001; pp 47-55. (10) Mitch, W. A.; Sedlak, D. L. Formation of N-nitrosodimethylamine (NDMA) from dimethylamine during chlorination. Environ. Sci. Technol. 2002, 36 (4), 588-595. (11) Choi, J.; Valentine, R. L. N-Nitrosodimethylamine formation by free-chlorine-enhanced nitrosation of dimethylamine. Environ. Sci. Technol. 2003, 37 (21), 4871-4876. (12) Mitch, W. A.; Gerecke, A. C.; Sedlak, D. L. A N-nitrosodimethylamine (NDMA) precursor analysis for chlorination of water and wastewater. Water Res. 2003, 37 (15), 3733-3741. (13) Gerecke, A. C.; Sedlak, D. L. Precursors of N-nitrosodimethylamine in natural waters. Environ. Sci. Technol. 2003, 37 (7), 1331-1336. (14) Luo, X.; Clevenger, T. E.; Deng, B. Role of NOM in the formation of N-nitrosodimethylamine (NDMA) in surface waters. In Abstracts of Papers, 229th ACS National Meeting, San Diego, CA, United States, March 13-17, 2005; American Chemical Society: Washington, DC, 2005; p ENVR-135. (15) Gallard, H.; von Gunten, U. Chlorination of natural organic matter: kinetics of chlorination and of THM formation. Water Res. 2001, 36 (1), 65-74. 7296

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006

(16) Kitis, M.; Karanfil, T.; Kilduff, J. E. The reactivity of natural organic matter to disinfection by-products formation and its relation to specific ultraviolet absorbance. Water Sci. Technol. 2001, 43(2, 1st World Water Congress, Part 2: Industrial Wastewater and Environmental Contaminants, 2000), 9-16. (17) Wei, Y.; et al. How organic nitrogen compounds in surface water relate to DBP formation in the drinking water treatment process, ProceedingssWater Quality Technology Conference; American Water Works Association: Denver, CO, 2002; pp 107-120. (18) Wu, W. W.; Chadik, P. A.; Delfino, J. J. The relationship between disinfection by-product formation and structural characteristics of humic substances in chloramination. Environ. Toxicol. Chem. 2003, 22 (12), 2845-2852. (19) Chae, S. Evaluation of drinking water treatment processes focusing on natural organic matter removal and on disinfection by-product formation. Water Sci. Technol.: Water Supply 2002, 2 (5-6), 459-464. (20) Huang, W. J.; Chen, L. Y.; Peng, H. S. Effect of NOM characteristics on brominated organics formation by ozonation. Environ. Int. 2004, 29 (8), 1049-1055. (21) Croue, J. P.; Korshin, G. V.; Benjamin, M. Characterization of Natural Organic Matter in Drinking Water; AWWA Research Foundation: Denver, CO, 1999. (22) Guo, L.; Lehner, J. K.; White, D. M.; Garland, D. S. Heterogeneity of natural organic matter from the Chena River, Alaska. Water Res. 2003, 37 (5), 1015-1022. (23) Duirk, S. E.; Gombert B.; Croue J.P.; Valentine R. L. Modeling monochloramine loss in the presence of natural organic matter. Water Res. 2005, 39 (14), 3418-3431. (24) Duirk, S. E.; Gombert, B.; Choi, J.; Valentine R. L. Monochloramine loss in the presence of humic acid. J. Environ. Monit. 2002, 4 (1), 85-89. (25) Duirk, S. E.; et al. Modeling monochloramine loss in the presence of natural organic matter and relationship to haloacetic acid formation, Preprints of Extended Abstract, ACS National Meeting. Am. Chem. Soc., Div. Environ. Chem. 2001, 41 (2), 799803. (26) Choi, J.; Valentine, R. L. Formation of N-nitrosodimethylamine (NDMA) in chloraminated water: New disinfection by-product. Preprints of Extended Abstract, ACS National Meeting. Am. Chem. Soc., Div. Environ. Chem. 2001, 41 (1), 8-11. (27) Vikesland, P. J.; Ozekin, K.; Valentine, R. L. Monochloramine decay in model and distribution system waters. Water Res. 2001, 35 (7), 1766-1776. (28) Vikesland, P. J.; Ozekin, K.; Valentine, R. L. Effect of natural organic matter on monochloramine decomposition: Pathway elucidation through the use of mass and redox balances. Environ. Sci. Technol. 1998, 32 (10), 1409-1416. (29) Jafvert, C. T.; Valentine, R. L. Reaction scheme for the chlorination of ammoniacal water. Environ. Sci. Technol. 1992, 26 (3), 577-586. (30) Duirk, S. E.; Valentine, R. L. Modeling dichloroacetic acid formation from the reaction of monochloramine with natural organic matter. Water Res. 2006, 40 (14), 2667-2674. (31) Clark, R. M.; Thumau; R. C.; Sivaganesan, M.; Ringhand, P. Predicting the formation of chlorinated and brominated byproducts. J. Environ. Eng. (Reston, Virginia) 2001, 127 (6), 493501. (32) Gang, D. D.; Segar, R. L.; Clevenger, T. E.; Banerji, S. K. Using chlorine demand to predict TTHM and HAA9 formation. J. Am. Water Works Assoc. 2002, 94 (10), 76-86. (33) Scientist, Version 2.02; Software Package. Micromath; Salt Lake City, UT, 1995. (34) Becher, K. D.; Kalkhoff, S. J.; Schnoebelen, D. J.; Barnes, K. K.; Miller, V. E. Water-Quality Assessment of the Eastern Iowa Basins-Nitrogen, Phosphorus, Suspended Sediment, and Organic Carbon in Surface Water, 1996-98; U.S. Geological Survey: Iowa city, IA, 2001. (35) Schnoebelen, D. J.; Kalkhoff, S. J.; Becher, K. D.; Thurman, E. M. Water-Quality Assessment of the Eastern Iowa Basins: Selected Pesticides and Pesticide Degradates in Streams, 199698. In Water-Resources Investigations Report 03-4075; U.S. Geological Survey: Iowa City, IA, 2003. (36) Bruchet, A.; et al. THM formation potential and organic content: a new analytical approach. Aqua (Oxford) 1987, 2, 102-109. (37) Chu, H. P.; Wong, J. H. C.; Li, X. Y. Trihalomethane formation potentials of organic pollutants in wastewater discharge. Water Sci. Technol. 2002, 46 (11-12, Water Quality and Environmental Management in Asia), 401-406. (38) Mega, J. Water quality assessment and trihalomethane formation potential. Vodni Hospodarstvi 2005, 55 (2), 33-35.

(39) Adam, L. C.; Fabian, I.; Suzuki, K.; Gordon, G. Hypochlorous acid decomposition in the pH 5-8 region. Inorg. Chem. 1992, 31 (17), 3534-3541. (40) APHA, AWWA, and WEF Standard Methods for the Examination of Water and Wastewater, 18th ed.; American Public Health Association: Washington DC, 1992. (41) Taguchi, V. Y.; et al. Determination of N-nitrosodimethylamine by isotope dilution, high-resolution mass spectrometry. Can. J. Appl. Spectrosc. 1994, 39 (3), 87-93. (42) Luo, X.; Clevenger, T. E.; Gang, D. NDMA analytical methods comparisons and its occurrence in Missouri. ProceedingssWater Quality Technology Conference; American Water Works Association: Washington, DC, 2003; pp 131-142. (43) Choi, J.; Valentine, R. L. Formation of N-nitrosodimethylamine (NDMA) from reaction of monochloramine: a new disinfection by-product. Water Res. 2002, 36 (4), 817-824.

(44) Valentine, R. L.; Jafvert, C. T. General acid catalysis of monochloramine disproportionation. Environ. Sci. Technol. 1988, 22 (6), 691-696. (45) Roth, J. G.; Ozment, C. L. How effective is enhanced coagulation and TOC removal in controlling DBPs? Results from two Illinois treatment plants. ProceedingssAnnual Conference, American Water Works Association; AWWA: Washington, DC, 1998; Vol. D (Water Quality), pp 347-363.

Received for review March 7, 2006. Revised manuscript received August 17, 2006. Accepted August 28, 2006. ES0605319

VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7297