Relative Importance of Nitrite Oxidation by Hypochlorous Acid under

Publication Date (Web): May 9, 2012 ... Chlorine disinfection remains quite popular in the United States,(1-4) but because ... (7, 8) During nitrifica...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Relative Importance of Nitrite Oxidation by Hypochlorous Acid under Chloramination Conditions David G. Wahman*,† and Gerald E. Speitel, Jr.‡ †

United States Environmental Protection Agency, Office of Research and Development, Cincinnati, Ohio 45268, United States University of Texas at Austin, Department of Civil, Architectural and Environmental Engineering, Austin, Texas 78712, United States

‡

S Supporting Information *

ABSTRACT: Nitrification can occur in water distribution systems where chloramines are used as the disinfectant. The resulting product, nitrite, can be oxidized by monochloramine and hypochlorous acid (HOCl), potentially leading to rapid monochloramine loss. This research characterizes the importance of the HOCl reaction, which has typically been ignored because of HOCl’s low concentration. Also, the general acid-assisted rate constants for carbonic acid and bicarbonate ion were estimated for the monochloramine reaction. The nitrite oxidation reactions were incorporated into a widely accepted chloramine autodecomposition model, providing a comprehensive model that was implemented in AQUASIM. Batch kinetic experiments were conducted to evaluate the significance of the HOCl reaction and to estimate carbonate buffer rate constants for the monochloramine reaction. The experimental data and model simulations indicated that HOCl may be responsible for up to 60% of the nitrite oxidation, and that the relative importance of the HOCl reaction for typical chloramination conditions peaks between pH 7.5 and 8.5, generally increasing with (1) decreasing nitrite concentration, (2) increasing chlorine to nitrogen mass ratio, and (3) decreasing monochloramine concentration. Therefore, nitrite’s reaction with HOCl may be important during chloramination and should be included in water quality models to simulate nitrite and monochloramine’s fate.

■

reactive with NO2− than NH2Cl.10 As HOCl’s ability to react with NO2− may be limited by HOCl production from NH2Cl hydrolysis, previous research has either dismissed this reaction pathway,11 operated under conditions where it would be minimally important compared to NO 2− oxidation by NH2Cl,10,12,13 or ignored the kinetically controlled nature of chloramine decomposition in their analysis.14 To our knowledge, Vikesland et al.11 represents the only peer-reviewed study under drinking water relevant conditions to evaluate incorporation of the NH2Cl/NO2− reaction scheme of Margerum et al.10 into the widely accepted chloramine autodecomposition model of Jafvert and Valentine.15 This work, however, must be considered preliminary for several reasons. First, Vikesland et al.11 reported only a single experiment focusing on NO2− and NH2Cl. Second, they excluded the HOCl/NO2− reaction from their kinetic analysis, citing previous research10,13 as evidence that the HOCl/NO2− reaction was unimportant. The cited work, however, was conducted under conditions that would mask the contribution of the HOCl/NO2− reaction. Third, the kinetic modeling of the NH2Cl/NO2− reaction directly incorporated the rate constants from Margerum et al.10 specific to a 50 mM phosphate buffer. Because the NH2Cl/NO2− reaction is general acid-assisted,10 refined estimates of the rate constants are needed to represent the lower-concentration carbonate buffer systems typical of

INTRODUCTION Chlorine disinfection remains quite popular in the United States,1−4 but because of the Stage 1 and Stage 2 Disinfectants and Disinfection Byproducts Rules, many United States utilities now use combinations of chlorine and chloramines to avoid excessive trihalomethane (THM) and haloacetic acid (HAA) formation. A recent survey reported that 30% of the respondents currently chloraminate to maintain distribution system residual,3 and other recent surveys suggest that between 8 and 12% of drinking water utilities are contemplating a future switch to chloramination.3,5 Upon implementation of the Stage 2 Disinfectants and Disinfection Byproducts Rule (i.e., ∼2013− 2015), chloramination for secondary disinfection in the United States is predicted to increase to 57% of all surface and 7% of all groundwater treatment systems.6 Chloramination may promote the growth of nitrifying bacteria [i.e., ammonia-oxidizing bacteria (AOB) and nitriteoxidizing bacteria (NOB)] because of naturally occurring ammonia; residual ammonia remaining from initial chloramine formation; and ammonia released from chloramine decay, oxidation of natural organic matter (NOM), corrosion, pipe surface reactions, and nitrite (NO2−) oxidation under various conditions in chloraminated water systems.7,8 During nitrification, AOB biologically oxidize free ammonia to NO2−. The NO2− produced can subsequently react with monochloramine (NH2Cl), and this mechanism is cited as a possible cause for rapid NH2Cl loss.9 Another possible pathway for NO2− is direct reaction with extremely low hypochlorous acid (HOCl) concentrations (e.g., 10−3−10−6 mg Cl2 L−1) in chloraminated systems because HOCl is approximately 1.8 × 105-times more © 2012 American Chemical Society

Received: Revised: Accepted: Published: 6056

March 8, 2012 May 7, 2012 May 9, 2012 May 9, 2012 dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

Table 1. Chloramine Model Implementation Reactions and Associated Reaction Rate and Equilibrium Constants at 25 °C no.

reaction stoichiometry

rate expression

rate or equilibrium constant (25 °C)

ref(s)

1 2 3 4 5 6 7 8 9 10 11 12

HOCl+ NH3→NH2Cl + H2O NH2Cl + H2O→HOCl + NH3 HOCl + NH2Cl→NHCl2 + H2O NHCl2 + H2O→HOCl + NH2Cl NH2Cl + NH2Cl→NHCl2 + NH3 NHCl2 + NH3→NH2Cl + NH2Cl NHCl2 + H2O→Ib I + NHCl2→HOCl + N2 + 3H+ + 3Cl− I + NH2Cl→N2 + 3H+ + 3Cl− NH2Cl + NHCl2→N2 + 3H+ + 3Cl− HOCl ⇆ H+ + OCl− NH4+ ⇆ NH3 + H+

k1M[HOCl][NH3] k2M[NH2Cl] k3M[HOCl][NH2Cl] k4M[NHCl2] k5M[NH2Cl]2 k6M[NHCl2][NH3][H+] k7M[NHCl2][OH−] k8M[I][NHCl2] k9M[I][NH2Cl] k10M[NH2Cl][NHCl2]

k1M = 3.07 × 106 M−1 s−1 k2M = 2.1 × 10−5 s−1 k3M = 2.8 × 102 M−1 s−1 k4M = 6.5 × 10−7 s−1 k5Ma k6M = 6.0 × 104 M−2 s−1 k7M = 1.1 × 102 M−1 s−1 k8M = 2.8 × 104 M−1 s−1 k9M = 8.3 × 103 M−1 s−1 k10M = 1.5 × 10−2 M−1 s−1 KHOCl = 10−7.5 KNH4+ = 10−9.3

31 31 32,33 32 11,34 35 36,37 38 38 38 39 39

13

H2CO3 ⇆ HCO3− + H+

KH2CO3 = 10−6.3

39

14

HCO3−

KHCO3 = 10

39

15

H2O ⇆ OH− + H+

Kw = 10−14

39

⇆ CO3 + H 2‑

+

−

−10.3

k5M = k5MH[H+] + k5MHCO3[HCO3−] + k5MH2CO3[H2CO3] where k5MH = 6.9 × 103 M−2 s−1, k5MHCO3 = 2.2 × 10−1 M−2 s−1, k5MH2CO3 = 1.1 × 101 M−2 s−1. bI = unidentified monochloramine autodecomposition intermediate. a

natural waters. Thus, a deeper understanding of NO 2− oxidation is needed now in light of the growing use of chloramination and ongoing concerns about distribution system nitrification. A further motivation for this research is that both the NH2Cl/NO2− and HOCl/NO2− reactions result in nitryl chloride (NO2Cl) formation.10,16 NO2Cl has been implicated in N-nitrosodimethylamine (NDMA) formation;17 therefore, proper accounting for NO2Cl formation in models may provide insight into NDMA formation. In the present study, the importance of the HOCl/NO2− reaction under chloramination conditions was evaluated using batch kinetic experiments and a chloramine model implemented into the computer program AQUASIM to determine if inclusion of the HOCl/NO2− reaction was required to accurately simulate the experimental data. Because drinking water systems are carbonate buffered and general acid-assisted rate constants only exist for phosphate,10 an additional objective was determining carbonate buffer general acid-assisted rate constants for the NH2Cl/NO2− reaction. Based on these results and with the revised model, two practical and pHdependent implications for chloraminated drinking water distribution systems are discussed: (1) the maximum NO2− concentration that will accumulate during a nitrification event, and (2) interpretation of chemical concentrations [e.g., total free ammonia (NH3+NH4+; TOTNH3) and NO2−] used for detection of nitrification occurrence.

solution was allowed to mix for 15 min before use, and scans of NH2Cl stock solutions were conducted on a Nicolet Evolution 300 UV−visible spectrophotometer (Thermo Electron Scientific Instruments) to verify NH2Cl formation. Batch Kinetic Experiments. Batch kinetic experiments were conducted at room temperature (22 ± 1 °C) in 4 or 10 mM sodium bicarbonate buffered ultrapure water spiked with the appropriate aliquot of NH2Cl stock solution to achieve the desired NH2Cl concentration. This prepared NH2Cl solution was then placed in headspace-free 500-mL, glass, gastight syringes (VICI Precision Sampling). Before each experiment, syringes were made chlorine-demand-free by soaking in a 5000 mg Cl2 L−1 free chlorine solution for 24 h, rinsed with distilled water, and air-dried. Syringes contained small Teflon-coated stir bars for mixing and were wrapped in aluminum foil. Before NO 2 − addition, initial samples were taken to verify experimental conditions. Next, the appropriate amount of NO2− stock solution was injected through the syringe nose to start an experiment. Subsequently, samples for NH2Cl, TOTNH3, NO2−, nitrate (NO3−), temperature, and pH were collected over time. To evaluate the implemented model over conditions applicable to drinking water treatment, 21 experiments were conducted (Supporting Information (SI), Table S1), spanning the range that may be seen during drinking water chloramination and associated nitrification events. These experiments sought to validate model implementation and estimate carbonate buffer general acid-assisted rate constants (i.e., kH2CO3 and kHCO3) relevant to drinking water. The range of conditions used in these experiments represents those typically found in drinking water; care should be taken in using the resulting model at conditions (e.g., pH, Cl2:N mass ratio) beyond those used in the current research without further validation. Analytical Methods. During batch kinetic experiments, pH and TOTNH3 were measured on a model 250 pH/ISE/ conductivity meter with a pH and ammonia electrode (Denver Instrument), respectively; and NH2Cl and NO2− were measured on a Nicolet Evolution 300 UV−visible spectrophotometer at 655 nm using HACH Method 10171 and at 507 nm using HACH Method 8507, respectively. At experiment end, a

■

EXPERIMENTAL SECTION Reagent Preparation. Solutions were prepared in ultrapure water (Barnstead NANOpure Diamond). Stock NO2− solutions were prepared from reagent-grade sodium nitrite. Stock chlorine solutions were prepared by diluting 4−6% sodium hypochlorite and were standardized periodically with sodium thiosulfate in accordance with Standard Methods 4500B. 18 Stock TOTNH3 solutions were prepared by dissolving ammonium sulfate in ultrapure water and adjusting to pH 8.3. Stock NH2Cl solutions were prepared by additions of stock TOTNH3 solutions to ultrapure water and then adding an aliquot of the stock chlorine solution to this well-stirred TOTNH3 solution (pH > 8.3) to achieve a 4:1 or 1:1 chlorine to nitrogen (Cl2:N) mass ratio as required. The NH2Cl stock 6057

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

Table 2. Nitrite Reaction Implementation for Hypochlorous Acid after Johnson and Margerum16 and Monochloramine after Margerum et al.10 monochloramine reaction 1

+

H + NH 2Cl +

k1 ′ NO−2 XooooY k−1 ′

hypochlorous acid k1

HOCl + NO−2 XoooY NO2 Cl + OH−

NH3 + NO2 Cl

reaction 2

k−1

k2

NO2 Cl + NO−2 XoooY N2O4 + Cl− k−2

reaction 3

N2O4 + OH− → NO−3 + NO−2 + H+(fast)

reaction 4

NO2 Cl XoooY NO+2 + Cl−

k3

k4

k−4

reaction 5

k5

NO+2 + OH− → NO−3 + H+ (fast)

rate expression

HOCl + NO−2 → Cl− + NO−3 + H+

NH 2Cl + NO−2 + H 2O → NH3 + NO−3 + Cl −+ H+

overall a

(

k1 ′[H+][NH 2Cl][NO−2 ] 1 + k−1 ′ [NH3] k4

(

+ 1+

)

(

k2 [NO−2 ] k4

k1[HOCl][NO−2 ] 1 +

)

k2 [NO−2 ] k4

k−1 [OH−] k4

(

+ 1+

)

k2 [NO−2 ] k4

)

k2 [NO−2 ] k4

where

k1 ′ = kH + k H2CO3 K=

[H 2CO3] [HCO−3 ] + k HCO3 + [H ] [H+]

k −1 ′/k4 k1 ′

⎛ k H CO [H CO ] k HCO3 [HCO−3 ] ⎞ k −1 ′ ⎟ = KkH⎜1 + 2 3 2 + 3 + k4 [H ] [H+] ⎠ kH kH ⎝ a

Rate expression derivations assume a constant NO2Cl concentration.

final sample for NO2− and NO3− was analyzed by ion chromatography using EPA method 300.0; and TOTNH3 was analyzed on the WESTCO SmartChem 200 autoanalyzer (Mandel Scientific Company) using EPA method 350.1. Chloramine Model Implementation. Reaction Rate Expressions and Stoichiometry. A chloramine model15 was implemented (Table 1) into the computer program AQUASIM19 along with the HOCl16 and NH2Cl10 NO2− reactions (Table 2). For implementation, acid−base chemistry was modeled as extremely fast forward and reverse reactions (SI, Table S2) governed by their respective equilibrium constants.20 Temperature dependency (SI, Table S3) of important rate and equilibrium constants was also implemented.11,12,21 Ionic strength corrections were implemented using the Davies equation.22 In addition to HOCl, it is worth noting that chlorine monoxide (Cl2O) and molecular chlorine (Cl2) have been recently highlighted as kinetically important free available chlorine species.23,24 For the experimental conditions of this research and conservatively assuming that both Cl2O and Cl2 are reacting at diffusion controlled rates of 1010 M−1 s−1,22 their combined maximum reaction rate is less than 1% of that attributed to HOCl (data not shown). Model Parameter Estimation. Because experimental conditions were specifically chosen to represent a range covering relevant drinking water conditions and the NO2− oxidation rate equations are complex, pseudo-first-order assumptions were not valid. To estimate parameters in this nonlinear system, all experiments were simultaneously fit using measured NH2Cl concentrations (297 data points) and the parameter estimation function in AQUASIM, which was configured to minimize the weighted residual sum of squares (WRSS) between measurements and calculated model results (eq 1):

⎛ ymeas, i − WRSS = ∑ ⎜ ⎝ W i=1 n

yi ⎞2 ⎟ = ⎠

⎛y − ∑ ⎜⎜ meas, i ymeas, i i=1 ⎝ n

yi ⎞ ⎟ ⎟ ⎠

2

(1)

In eq 1, ymeas,i is the i-th NH2Cl measurement, W is the weighting factor, and yi is the model calculated NH2Cl concentration corresponding to the i-th measurement. Because NH2Cl measurements were changing over an order of magnitude, ymeas,i was implemented for W to prevent higher concentrations from biasing the fitting procedure, resulting in a dimensionless WRSS.25,26 The secant algorithm in AQUASIM was used for nonlinear parameter estimation for three parameters (Table 2) associated with NO2− oxidation by NH2Cl: two proposed acid-assisted rate constants for the carbonate system [carbonic acid (kH2CO3) and bicarbonate ion (kHCO3)] and K. The values for the carbonate system (kH2CO3 and kHCO3) represent the first estimates of these parameters while K represents a re-estimation of this parameter from Margerum et al.10 Because kinetic parameters were simultaneously estimated, experiments focused on varying the initial experimental conditions rather than conducting all experiments in duplicate. To confirm that the experiments could be duplicated, one experimental condition was duplicated and validated the experimental procedure used (SI, Figure S1).

■

RESULTS AND DISCUSSION Model Evaluation and Parameter Determination. Evaluation and Adjustment of Published Reaction Rate Constants. During model development, the work of Johnson and Margerum16 and Margerum et al.10 was critically reviewed. For the NH2Cl/NO2− reaction, it was determined that the 6058

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

originally estimated acid-assisted rate constants10 for hydrogen ion [kH = (7.6 ± 1.2) × 106 M−2 s−1] and phosphate (kH2PO4 = 46 ± 8 M−2 s−1) should be reanalyzed. First, the reanalysis used the complete set of data (22 data points) presented by Margerum et al.10 in their Table 3, allowing inclusion of data where both phosphate concentration (0.020−0.140 M) and pH (6.09−7.20) varied versus the original analysis that included a subset (16 data points corresponding to rows 7−22 in Table 3 of Margerum et al.10) of these data at a single pH (6.57 ± 0.03). Second, the reanalysis accounted for phosphate being a multiprotic acid, using the phosphate pKas (pKa1 = 1.75, pKa2 = 6.8, pKa3 = 11.30) presented by Kumar et al.27 Using the pseudo-first-order rate constants (kobs) presented by Margerum et al.,10 a multiple linear regression was conducted in Excel where kobs/[NO2−] = kH[H+] + kH2PO4[H2PO4−], resulting in significantly different parameter estimates [kH = (1.2 ± 0.08) × 107 M−2 s−1; kH2PO4 = 13 ± 6 M−2 s−1]. A comparison of the calculated kobs (Calculated kobs) to the experimentally determined kobs (Experimental kobs) for both the original Margerum et al.10 analysis and the current reanalysis is presented in SI, Figure S2 where the six additional data points are noted as filled symbols, demonstrating that the reanalysis provides a better representation of the experimental data. The rate equation derived for the HOCl/NO2− reaction by Johnson and Margerum16 was based on the assumption that k−1[OH−] ≫ (k4 + k2[NO2−]). We removed this assumption and used the complete rate equation (Table 2) in the current model, which requires knowing k1. The value of k1 was not initially determined by Johnson and Margerum,16 but k1 (4.4 × 104 M−1 s−1) was later estimated by Margerum et al.10 based on a nucleophilicity correlation using kH. This methodology was followed (SI, Figure S3 and Table S4) to produce a new estimate of k1 (1.8 × 105 M−1 s−1) based on the revised estimate of kH previously described. Comparison of Various Model Simulations in Current Research. Using experimental data collected in the current research, an evaluation of whether (1) the HOCl/NO2− reaction and (2) carbonate buffer general acid-assisted rate constants were required to simulate the experimental data was conducted. Evidence supporting these two model inclusions is presented in Figure 1 where experiments were conducted with the same initial conditions except pH (SI, Table S1, Experiments 1, 2, and 4). Figure 1 shows the impact of different model simulations with varying assumptions on how NO2− oxidation was incorporated into the chloramine model previously described: (1) “No Nitrite Reactions (Published)” represents the chloramine model without inclusion of any NO2− oxidation reactions and provides a baseline for NH2Cl autodecomposition and the relative importance of the NO2− reactions, (2) “Monochloramine Reaction (Published)” represents direct incorporation of the NH2Cl/NO2− reaction from Margerum et al.10 without modification, (3) “Both Reactions (Published)” represents direct incorporation of the HOCl/ NO2− reaction from Johnson and Margerum16 and NH2Cl/ NO2− reaction from Margerum et al.10 without modification, (4) “Both Reactions (Estimated)” represents inclusion of the HOCl/NO2− reaction from Johnson and Margerum16 and NH2Cl/NO2− reaction from Margerum et al.10 with the modifications described previously and the best-fit model parameters estimated from the current research and described in the following section.

Figure 1. Comparison of four model simulations for Experiments 4 (A, pH 7.6), 1 (B, pH 8.4), and 2 (C, pH 9.0). Initial conditions: 4 mM carbonate buffer, 1.5 mg Cl2 L−1 monochloramine, 4:1 Cl2:N mass ratio, and 2 mg N L−1 nitrite. See text (Comparison of Various Model Simulations in Current Research) for definitions.

Figure 1 highlights that inclusion of the HOCl/NO2− reaction and carbonate buffer rate constants for these experimental conditions improved model simulations. The general-acid assisted nature of the carbonate buffer and impact of free ammonia (NH3) concentration contribute to the pH effects on model simulations as the model with published parameters underpredicts NH2Cl loss at pH 7.6 (Figure 1, 6059

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

Table 3. Summary (Value ± 95% Confidence Interval) of Published Nitrite Reaction Parameters with Hypochlorous Acid (HOCl) from Johnson and Margerum16 and Monochloramine (NH2Cl) from Margerum et al.,10 Parameters Reanalyzed in the Current Research, and Model Estimated Parameters in the Current Research nitrite reactant

a

parameter

units −1

published value

reanalysis of published value

−2e

initial model estimation

final model estimation

s−1 s−1

(2.3 ± 0.6) × 10 5.0 ± 0.2e 217 ± 54e 4.4 × 104bf (1.9 ± 0.3) × 106ef (7.6 ± 1.2) × 106f 46 ± 8f

ND ND ND 1.8 × 105b (7.7 ± 1.1) × 106c (1.2 ± 0.08) × 107d 13 ± 6d

ND ND ND ND ND ND ND

ND ND ND ND ND ND ND

K kH2CO3

Ms M−2 s−1

(4.0 ± 1.0) × 10−2f ND

ND ND

(7.6 ± 0.4) × 10−2d 920 ± 140d

(7.6 ± 0.3) × 10−2d 940 ± 130d

kHCO3

M−2 s−1

ND

ND

1.2 ± 5.2d

ND

HOCl HOCl HOCl and NH2Cl HOCl HOCl NH2Cl NH2Cl

(k1k4)/(k−1) (k1k2)/(k−1) (k2)/(k4) k1 (k−1)/(k4) kH kH2PO4

s M−1 M−1 M−1 M−1 M−2 M−2

NH2Cl NH2Cl NH2Cl

b

s−1 s−1

a

c

d

ND = Not determined. Calculated from nucleophilicity. Calculated from estimated parameters. Estimated in current research. eRef 16. fRef 10.

ableness of the carbonic acid general-acid assisted rate constant determined in the current research. Simulation Summary. As an example to show that the model was accurately simulating NH2Cl and the other nitrogencontaining species during the experiments, Figure 3 summarizes experimental and model simulated profiles of NH2Cl, NO2−,

Panel A) and overpredicts NH2Cl loss at pH 9.0 (Figure 1, Panel C). Model Parameter Estimation Summary. Table 3 summarizes the published parameters for the oxidation of NO2− by NH2Cl and HOCl, the reanalysis of certain parameters, and parameters estimated in the current research by a simultaneous fit of the experimental data as previously described. As mentioned previously, the NH2Cl/NO2− reaction is generalacid assisted. Initially, estimation of both the carbonic acid and bicarbonate ion terms was conducted, resulting in a bicarbonate ion term that was not significantly different from zero (Table 3); therefore, a final estimation was conducted without inclusion of this term. Both estimations resulted in the same WRSS (3.9), resulting in selection of the model including only the carbonic acid term for further evaluation. The results from both model parameter estimations are included in Table 3. To evaluate the estimated carbonic acid (kH2CO3) term, a Brønsted plot (Figure 2) was constructed using the parameters

Figure 2. Brønsted plot for general acid-assisted rate constants for the reaction of monochloramine and nitrite.

presented in SI, Table S5. The correlations were generated using estimated values for kH, kH2PO4, and kH2CO3. Figure 2 shows both the original correlation determined by Margerum et al.10 and results from the current study where both plots result in the same slope (0.62), providing evidence for the reason-

Figure 3. Comparison of monochloramine, nitrite, total free ammonia, and nitrate model simulations and experimental data for Experiments 1 (A, initial 2.0 mg N L−1 nitrite) and 7 (B, initial 0.5 mg N L−1 nitrite). Initial conditions: pH 8.4, 4 mM carbonate buffer, 1.5 mg Cl2 L−1 monochloramine, and 4:1 Cl2:N mass ratio. 6060

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

Figure 4. Summary of a subset of model simulations performed to evaluate the significance of nitrite oxidation by hypochlorous acid under conditions relevant to chloraminated drinking water distribution systems [pH, nitrite concentration, monochloramine concentration, and chlorine to nitrogen (Cl2:N) mass ratio].

TOTNH3, and the final NO3− concentrations for Experiments 1 and 7, which represent both a low (0.5 mg N L−1) and high (2.0 mg N L−1) NO2− concentration. In addition to NH2Cl, the model accurately simulated the nitrogen-containing species throughout the experiment. Only the final nitrate value and relevant model simulation period is shown in Figure 3 because a nitrate sample was only taken at the end of each experiment. A summary of all the experiments conducted and the associated simultaneous model fits is presented in SI, Figure S4. Considering that a simultaneous fit was conducted for the complete data set, the model represented the experimental data very well for most experimental conditions evaluated in this research. In two cases (i.e., Experiments 19 and 21), the model indicated a somewhat more rapid monochloramine decay than was observed experimentally. These two experiments were conducted at the highest monochloramine concentration (3 mg Cl2 L−1) and lowest Cl2:N ratio (1:1), which are conditions at the outer limits of what might be encountered in practice. Overall, incorporation of the HOCl/NO2− reaction better represented batch kinetic experimental data for the NH2Cl loss in the presence of NO2−. Also, the carbonate buffer general acid assisted reaction rate constants for the NH2Cl/NO2− reaction are essential to the accurate modeling of typical distribution systems. Relative Importance of Nitrite Oxidation by Hypochlorous Acid. Using the model developed in this research, an evaluation of the relative significance of NO2− oxidation by HOCl and NH2Cl was conducted under conditions spanning the range of those that may be encountered in chloraminated drinking water systems, considering the possibility of nitrification: pHs 6, 7, 8, and 9; chlorine to nitrogen mass ratios (Cl2:N) of 1:1, 2:1, 3:1, 4:1, and 5:1; 1, 2, and 4 mg Cl2 L−1 NH2Cl; 0.5 and 2 mg N L−1 NO2−, 4 mM carbonate buffer, and 25 °C temperature. A subset of these simulations (Figure 4) showed that HOCl may be responsible for up to 60% of the NO2− oxidation in chloraminated systems. Based on the trends

in Figure 4, the relative importance of NO2− oxidation by HOCl peaks around pH 7.5 to 8.5 and generally increases with (1) decreasing NO2− concentration, (2) increasing Cl2:N mass ratios, and (3) decreasing NH2Cl concentration. These trends represent a balance between the HOCl/NO2− and NH2Cl/ NO2− reaction rates (Table 2) and the availability of HOCl through its release from NH2Cl hydrolysis (Table 1, Reaction 2). Figure 5 provides NH2Cl concentrations when including and not including the HOCl/NO2− reaction in model simulations at three pHs (7, 8, and 9) for a typical chloramination condition (2 mg Cl2 L−1 NH2Cl, 5:1 Cl2:N mass ratio) with 0.5 mg N L−1 NO2−. The time to reach the first half-life (t1/2) of NH2Cl (i.e., 1 mg Cl2 L−1) under the conditions simulated in Figure 5 increased with pH. In general, NH2Cl stability is affected by pH because NH2Cl disproportionation is general acid-catalyzed (Table 1, Reaction 5), but there is a further pH impact on NH2Cl stability when NO2− is present because pH dependent terms are present in both the numerator (i.e., H+ and HOCl) and denominator (i.e., NH3 and OH−) of the NH2Cl/NO2− and HOCl/NO2− rate equations (Table 2). When only the NH2Cl/NO2− reaction is considered, t1/2 increased by similar factors from pH 7 to 8 (factor of 12) and pH 8 to 9 (factor of 16), but when the HOCl/NO2− reaction is included, a greater variation occurs as the increase from pH 7 to 8 is only a factor of 7, but the subsequent increase to pH 9 is a factor of 24. Based on these model simulations, the HOCl/NO2− reaction may be a relevant reaction in chloraminated systems and should be included in water quality models to model NH2Cl and NO2− concentrations. Practical Implications of Nitrite Oxidation by Hypochlorous Acid in Chloraminated Systems. The HOCl/ NO2− reaction has at least two practical implications for chloraminated drinking water distribution systems that are system pH dependent: (1) the simulated maximum NO2− concentration that will accumulate during a nitrification event, 6061

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

distribution system. Interest by the USEPA29 in changing the sampling protocol for NO2− and NO3− measurements in distribution systems indicates the need for a more comprehensive understanding of all mechanisms of NH2Cl loss and NO2− and NO3− formation. Depending on the system operating conditions (e.g., influent NO2−, NH2Cl, and TOTNH3 concentrations) and extent of nitrification, it is theoretically possible to exceed the 1 mg N L−1 NO2− concentration in the distribution system while meeting the current MCL requirements leaving the treatment plant. Second, because direct AOB monitoring in distribution systems is currently not practical, NO2− and other chemical parameters (e.g., TOTNH3, NH2Cl) are commonly used as surrogates for AOB presence and indicate the onset of nitrification events9 with a concentration of 0.010 mg N L−1 NO2− representing an alert level for the beginning of nitrification. NO2− accumulation will depend on both the activity of the microorganisms producing (i.e., AOB) and using (i.e., NOB) NO2−, chemical (i.e., HOCl and NH2Cl) NO2− oxidation rates, and competing reactions for NH2Cl and HOCl (e.g., bromide and natural organic matter). Depending on the operating conditions (e.g., system pH), a single NO2− level may not be representative of the extent of nitrification that is occurring in the distribution system. To simulate how the inclusion of the HOCl/NO2− reaction may impact simulated NO2− concentrations, biological activity was incorporated into the chloramine model. In AQUASIM, a batch reactor was simulated with a 2 mg TSS L−1 initial starting biomass concentration and the Monod kinetic parameters of the Lake Austin Mixed Culture (kTOTNH3 = 2.39 mg TOTNH3 mg TSS−1 d−1; KsNH3−N = 0.069 mg NH3−N L−1).30 Because the hydrodynamics were modeled as a batch reactor, this simulation (Figure 6) would be an approximation of a stagnant flow condition (e.g., dead-end mains, premise plumbing, poorly mixed storage tanks) and shows that NO2− concentration is pH dependent and is consistently lower if the oxidation by HOCl is included. If an alert level is based on a 0.010 mg N L−1 NO2− concentration,9 the indication of whether active nitrification is occurring is different at pH 7 versus pHs 8 and 9. Because TOTNH3 represents the AOB growth substrate, complementing monitoring of NH2Cl and NO2− with TOTNH3 warns of impending nitrification and accumulation indicates chloramine instability.9 As with NO2−, TOTNH3 concentrations (Figure 6) vary with pH and change if the HOCl/NO2− reaction is included, making interpretation of TOTNH3 concentrations also pH dependent. Overall, for simulations that include the HOCl/NO2− reaction, the following observations can be made: • pH 7 (Figure 6, Panel A): NO2− alert level not exceeded; TOTNH3 accumulates • pH 8 (Figure 6, Panel B): NO2− alert level exceeded; TOTNH3 accumulates, and • pH 9 (Figure 6, Panel C): NO2− alert level exceeded; TOTNH3 does not accumulate. Based on the simulation results, a nitrification monitoring program should be customized based on system operating pH as typical monitoring parameters (e.g., NO2− and TOTNH3) do not follow the same trends with pH changes and optimal parameters to monitor for nitrification events in distribution systems may vary with pH. These simulation trends indicate that TOTNH3 is more sensitive at pH 7 and NO2− becomes more sensitive as pH increases to 9.

Figure 5. Model simulations including and not including nitrite oxidation by hypochlorous acid at pH 7 (A), pH 8 (B), and pH 9 (C). Time to the first half-life (t1/2) of monochloramine is indicated. Initial conditions: 4 mM carbonate buffer, 2.0 mg Cl2 L−1 monochloramine, 5:1 Cl2:N, 0.5 mg N L−1 nitrite, and 25 °C.

and (2) interpretation of chemical concentrations (e.g., TOTNH3 and NO2−) used for detection of nitrification occurrence. First, the maximum NO2− concentration that will occur is of importance with regard to acute NO2− exposure. NO2− is regulated in the United States entering a distribution system with a maximum contaminant level (MCL) of 1 mg N L−1,28 but no current requirement exists for monitoring NO2− in the 6062

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

■

Article

AUTHOR INFORMATION

Corresponding Author

*Phone: (513) 569-7733; fax: (513) 487-2543; e-mail: [email protected]; mail: USEPA, 26 W. Martin Luther King Dr., Cincinnati, OH 45268. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Keith Kelty and David Griffith for analytical support. The USEPA collaborated in the research described herein. It has been subjected to the Agency’s peer and administrative review and has been approved for external publication. Any opinions expressed are those of the authors and do not necessarily reflect the views of the Agency; therefore, no official endorsement should be inferred. Any mention of trade names or commercial products does not constitute endorsement or recommendation for use.

■

Figure 6. Model simulations showing the impact of considering nitrite oxidation by hypochlorous acid when biological activity is occurring at pH 7 (A), pH 8 (B), and pH 9 (C). Initial conditions: 4 mM carbonate buffer, 4.0 mg Cl2 L−1 monochloramine, 5:1 Cl2:N, 25 °C.

■

REFERENCES

(1) AWWA Water Quality and Technology Division Disinfection Systems Committee. Committee report: Disinfection at small systems. J. Am. Water Works Assoc. 2000, 92 (5), 24−31. (2) AWWA Water Quality and Technology Division Disinfection Systems Committee. Committee report: Disinfection at large and medium-size systems. J. Am. Water Works Assoc. 2000, 92 (5), 32−43. (3) AWWA Water Quality and Technology Division Disinfection Systems Committee. Committee report: Disinfection survey, part 2 alternatives, experiences, and future plans. J. Am. Water Works Assoc. 2008, 100 (11), 110−124. (4) AWWA Water Quality and Technology Division Disinfection Systems Committee. Committee report: Disinfection survey, part 1 recent changes, current practices, and water quality. J. Am. Water Works Assoc. 2008, 100 (10), 76−90. (5) Seidel, C. J.; McGuire, M. J.; Summers, R. S.; Via, S. Have utilities switched to chloramines? J. Am. Water Works Assoc. 2005, 97 (10), 87−97. (6) USEPA. Economic Analysis for the Final Stage 2 Disinfectants and Disinfection Byproducts Rule; 815-R-05-010; Washington, DC, 2005. (7) Kirmeyer, G.; Martel, K.; Thompson, G.; Radder, L.; Klement, W.; LeChevallier, M.; Baribeau, H.; Flores, A., Optimizing Chloramine Treatment, 2nd ed.; AWWA Research Foundation: Denver, CO, 2004. (8) Wilczak, A.; Jacangelo, J. G.; Marcinko, J. P.; Odell, L. H.; Kirmeyer, G. J.; Wolfe, R. L. Occurrence of nitrification in chloraminated distribution systems. J. Am. Water Works Assoc. 1996, 88 (7), 74−85. (9) American Water Works Association. Fundamentals and Control of Nitrification in Chloraminated Drinking Water Distribution Systems (AWWA Manual M56), 1st ed.; American Water Works Association: Denver, CO, 2006. (10) Margerum, D. W.; Schurter, L. M.; Hobson, J.; Moore, E. E. Water chlorination chemistry: Nonmetal redox kinetics of chloramine and nitrite ion. Environ. Sci. Technol. 1994, 28 (2), 331−337. (11) Vikesland, J., P.; Ozekin, K.; Valentine, R. L. Monochloramine decay in model and distribution system waters. Water Res. 2001, 35 (7), 1766−1776. (12) Hao, O. J.; Chien, C. M.; Valentine, R. L. Kinetics of Monochloramine Reactions with Nitrite. J. Environ. Eng. 1994, 120 (4), 859−874. (13) Valentine, R. L., Disappearance of monochloramine in the presence of nitrite. In Water Chlorination: Environmental Impact and Health Effects; Jolley, R. L., Ed.; Lewis Publishers: Chelsea, MI, 1985; Vol. 5, pp 975−984. (14) Yang, H.; Cheng, H. Controlling nitrite level in drinking water by chlorination and chloramination. Sep. Purif. Technol. 2007, 56 (3), 392−396.

ASSOCIATED CONTENT

S Supporting Information *

Five tables and four figures providing additional information. This information is available free of charge via the Internet at http://pubs.acs.org/. 6063

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064

Environmental Science & Technology

Article

(38) Leao, S. F. Kinetics of combined chlorine: reactions of substitution and redox. Dissertation, University of California, Berkeley, CA, 1981. (39) Snoeyink, V. L.; Jenkins, D. Water Chemistry; John Wiley & Sons, Inc.: New York, 1980.

(15) Jafvert, C. T.; Valentine, R. L. Reaction scheme for the chlorination of ammoniacal water. Environ. Sci. Technol. 1992, 26 (3), 577. (16) Johnson, D. W.; Margerum, D. W. Non-metal redox kinetics: A reexamination of the mechanisms of the reaction between hypochlorite and nitrite ions. Inorg. Chem. 1991, 30, 4848−4851. (17) Choi, J.; Valentine, R. L. N-Nitrosodimethylamine Formation by Free-Chlorine-Enhanced Nitrosation of Dimethylamine. Environ. Sci. Technol. 2003, 37 (21), 4871−4876. (18) American Public Health Association. Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association: Washington, DC, 1998. (19) Reichert, P. AQUASIM - a tool for simulation and data analysis of aquatic systems. Water Sci. Technol. 1994, 30 (2), 21−30. (20) Musvoto, E. V.; Wentzel, M. C.; Loewenthal, R. E.; Ekama, G. A. Integrated chemical-physical processes modelling-I. Development of a kinetic-based model for mixed weak acid/base. Water Res. 2000, 34 (6), 1857−1867. (21) Cachaza, J. M.; Casado, J.; Castro, A.; Quintela, M. A. L. Kinetics of oxidation of nitrite by hypochlorite ions in aqueous basic solution. Can. J. Chem. 1976, 54 (21), 3401−3406. (22) Benjamin, M. M. Water Chemistry, 1st ed.; McGraw-Hill: New York, 2002. (23) Sivey, J. D.; McCullough, C. E.; Roberts, A. L. Chlorine Monoxide (Cl2 O) and Molecular Chlorine (Cl 2 ) as Active Chlorinating Agents in Reaction of Dimethenamid with Aqueous Free Chlorine. Environ. Sci. Technol. 2010, 44 (9), 3357−3362. (24) Sivey, J. D.; Roberts, A. L. Assessing the Reactivity of Free Chlorine Constituents Cl2, Cl2O, and HOCl Toward Aromatic Ethers. Environ. Sci. Technol. 2012, 46 (4), 2141−2147. (25) Robinson, J. A. Determining microbial kinetic parameters using nonlinear regression analysis: Advantages and limitations in microbial ecology. Adv. Microbial Ecol. 1985, 8, 61−114. (26) Draper, N. R.; Smith, H. Applied Regression Analysis; John Wiley & Sons, Inc.: New York, 1998. (27) Kumar, K.; Day, R. A.; Margerum, D. W. Atom-transfer redox kinetics - General-acid-assisted oxidation of iodide by chloramines and hypochlorite. Inorg. Chem. 1986, 25 (24), 4344−4350. (28) USEPA. National Primary Drinking Water Regulations; 40CFR141.62; 2010; pp 448−449. (29) Fed. Reg. 2010, 75, (59), 15499−15572. (30) Wahman, D. G.; Katz, L. E.; Speitel, G. E., Jr. Trihalomethane cometabolism by a mixed-culture nitrifying biofilter. J. Am. Water Works Assoc. 2006, 98 (12), 2−20. (31) Morris, J. C.; Isaac, R. A., A critical review of kinetic and thermodynamic constants for the aqueous chlorine-ammonia system. In Water Chlorination: Environmental Impact and Health Effects; Jolley, R. L., Brungs, W. A., Cotruvo, J. A., Cumming, R. B., Mattice, J. S., Jacobs, V. A., Eds.; Ann Arbor Science: Ann Arbor, MI, 1983; Vol. 4, pp 49−62. (32) Margerum, D. W.; Gray, E.; Huffman, R. P., Chlorination and the formation of N-chloro compounds in water treatment. In Organometals and Organometalloids: Occurrence and Fate in the Environment; Brinckman, F. E., Bellama, J. M., Eds.; American Chemical Society: Washington, DC, 1978; pp 278−291. (33) Wei, I. W. Chlorine-ammonia breakpoint reactions. Dissertation, Harvard University, Cambridge, MA, 1972. (34) Granstrom, M. L. The disproportionation of monochloramine. Dissertation, Harvard University, Cambridge, MA, 1955. (35) Hand, V. C.; Margerum, D. W. Kinetics and mechanisms of the decomposition of dichloramine in aqueous solution. Inorg. Chem. 1983, 22 (10), 1449−1456. (36) Jafvert, C. T. A unified chlorine-ammonia speciation and fate model (breakpoint, monochloramine, dichloramine, hypochlorous acid). Dissertation, The University of Iowa, Iowa City, IA, 1985. (37) Jafvert, C. T.; Valentine, R. L. Dichloramine decomposition in the presence of excess ammonia. Water Res. 1987, 21 (8), 967−973. 6064

dx.doi.org/10.1021/es300934x | Environ. Sci. Technol. 2012, 46, 6056−6064