Environ. Sci. Technol. 2002, 36, 588-595
Formation of N-Nitrosodimethylamine (NDMA) from Dimethylamine during Chlorination WILLIAM A. MITCH AND DAVID L. SEDLAK* Department of Civil and Environmental Engineering, 609 Davis Hall, University of California, Berkeley, Berkeley, California 94720
Chlorine disinfection of secondary wastewater effluent and drinking water can result in the production of the potent carcinogen N-nitrosodimethylamine (NDMA) at concentrations of approximately 100 and 10 parts per trillion (ng/L), respectively. Laboratory experiments with potential NDMA precursors indicate that NDMA formation can form during the chlorination of dimethylamine and other secondary amines. The formation of NDMA during chlorination may involve the slow formation of 1,1-dimethylhydrazine by the reaction of monochloramine and dimethylamine followed by its rapid oxidation to NDMA and other products including dimethylcyanamide and dimethylformamide. Other pathways also lead to NDMA formation during chlorination such as the reaction of sodium hypochlorite with dimethylamine. However, the rate of NDMA formation is approximately an order of magnitude slower than that observed when monochloramine reacts with dimethylamine. The reaction exhibits a strong pH dependence due to competing reactions. It may be possible to reduce NDMA formation during chlorination by removing ammonia prior to chlorination, by breakpoint chlorination, or by avoidance of the use of monochloramine for drinking water disinfection.
Introduction Recently, several water agencies in California have observed the formation of N-nitrosodimethylamine (NDMA) after chlorine disinfection. The observation of NDMA is due to improvements in analytical techniques that have enabled detection of concentrations as low as 1 ng/L (1), rather than changes in treatment techniques. Under conditions employed during chlorine disinfection, NDMA formation can exceed 100 ng/L during chlorination of secondary wastewater effluent (2). Chlorination of surface waters typically results in the formation of less than 10 ng/L NDMA (2). The US EPA integrated risk information service (IRIS) database classifies NDMA as a probable human carcinogen and lists a drinking water concentration resulting in a 10-6 risk of contracting cancer of 0.7 ng/L for NDMA (3). In contrast, for the same level of risk, the IRIS database provides a 6 µg/L drinking water concentration for chloroform, a trihalomethane. In response to the recent detection of NDMA in chlorinated waters, the California Department of Health * Corresponding author: telephone (510) 643-0256; fax (510) 6427483; e-mail:
[email protected]. 588
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002
Services has set an action level of 20 ng/L for NDMA. In locations where chlorinated wastewater effluent is used for indirect potable reuse, the presence of NDMA is a significant concern. For example, two drinking water production wells, under the influence of recharge water from the advanced wastewater treatment system of the Orange County Water District’s Water Factory 21, recently suspended operations due to their inability to meet the 20 ng/L action level (4). NDMA and related nitrosamines can be formed by nitrosation of secondary amines by nitrite. Toxicologists concerned with nitrosamine formation in the stomach have proposed the following sequence of reactions to explain the kinetics of NDMA formation (5, 6):
HNO2 T NO2- + H+
K1) 10-3.37 (5)
(1)
2 HNO2 T N2O3 + H2O
K2 ) 10-6.70 (7)
(2)
(CH3)2NH2+ T (CH3)2NH + H+
K3 ) 10-10.72 (5)
(3)
N2O3 + (CH3)2NH f (CH3)2NNO + HNO2
(4)
As a result of the dissociation of protons from nitrous acid and dimethylamine (reactions 1 and 3), the formation of NDMA exhibits a strong dependence on pH with a maximum rate near pH 3.4:
d[NDMA] ) k[(CH3)2NH][HNO2]2 dt k ) 1.5 × 10-5 M-2 s-1 (5) Some researchers have hypothesized that NDMA formation during water and wastewater treatment is attributable to the presence of nitrite (8, 9). However, these kinetic expressions indicate that the rate of formation of NDMA by this mechanism is extremely slow under the conditions encountered in wastewater. For example, 100 µM nitrite and 100 µM dimethylamine (DMA) would form less than 10-21 g/L of NDMA in 24 h at pH ) 7. In addition, at near-neutral pH, hypochlorite oxidizes nitrite to nitrate via a two-electron transfer with a half-life less than 1 s under conditions typically encountered during water and wastewater chlorination (10). We propose an alternative explanation for the formation of NDMA during water and wastewater treatment involving chlorination reactions resulting in the formation and oxidation of 1,1-dimethylhydrazine, which is also known as unsymmetrical dimethylhydrazine (UDMH). UDMH, a component of liquid rocket fuel, was manufactured during Germany’s V-2 rocket program by the reaction of monochloramine and dimethylamine at pH values greater than 10 via the Raschig Process (11) (Figure 1). In the presence of an oxidant at moderate pH levels, UDMH is transformed into a variety of products (Figure 2). Experiments performed by researchers interested in destruction of UDMH in waste rocket fuel indicate the formation of dimethylcyanamide (DMC), dimethylformamide (DMF), formaldehyde dimethylhydrazone (FDMH), formaldehyde monomethylhydrazone (FMMH), and NDMA (12-15). In all experiments, NDMA only accounted for a small percentage of the overall product yield (i.e., less than 5%). Neither of the two mechanisms for NDMA formation described above (i.e., nitrosation by nitrite or oxidation of UDMH) has been studied extensively under the conditions encountered during water and wastewater treatment. To identify the factors controlling the formation of NDMA during 10.1021/es010684q CCC: $22.00
2002 American Chemical Society Published on Web 12/28/2001
FIGURE 1. Proposed reaction scheme for NDMA formation via the UDMH pathway: UDMH formation.
FIGURE 2. Proposed reaction scheme for NDMA formation via the UDMH pathway: UDMH oxidation.
water and wastewater chlorination, we have conducted a series of experiments using dimethylamine as a model precursor for NDMA formation. The formation of NDMA and other nitrogen-containing products was quantified following addition of hypochlorite to dimethylamine alone, and in the presence of ammonia. While detailed mechanistic studies remain to be done, our results suggest that formation of NDMA by reaction between monochloramine and organic nitrogen species such as dimethylamine via the UDMH pathway could explain observed NDMA production in fullscale treatment systems. Although these experiments do not fully replicate the conditions encountered during chlorination of water and wastewater, they provide a new understanding of NDMA formation that can be used to identify NDMA precursors and to assess control strategies.
Materials and Methods Materials. All experiments were conducted using deionized water produced from a Barnstead Nanopure II water purifying system. Acros dimethylamine hydrochloride (99%), trimethylamine hydrochloride (98%), unsymmetrical dimethylhydrazine (99%), and sodium nitrite (97+%), Fisher Scientific ammonium chloride (99.9%) and sodium hypochlorite
(4-6% purified grade), and commercial grade hydrogen peroxide (3%) reagents were used without further purification. Acros N-nitrosodimethylamine (99+%), N,N-dimethylformamide (99+%), and dimethylcyanamide (97+%), and Cambridge Isotopes Laboratories N-nitrosodimethylamined6 (98%) were used as standards without further purification. Buffers used for various pH ranges included: potassium acetate (Mallinckrodt, Inc.) for pH < 6.0, a mixture of sodium phosphate monobasic and sodium phosphate dibasic (Fisher Scientific) for 6.0 < pH < 8.0, and boric acid (Mallinckrodt, Inc.) for pH > 8.0. Solution pH values were further adjusted by addition of sodium hydroxide or hydrochloric acid. Fisher Scientific L-ascorbic acid and GC Resolve methylene chloride were used for quenching chlorine and for sample liquidliquid extractions, respectively. Burdick-Jackson high purity methanol was used for analysis by gas chromatography with chemical ionization/tandem mass spectrometry. Preparation of Chlorinated Species. Monochloramine solutions were prepared fresh daily by dissolving ammonium chloride in deionized water adjusted to pH ) 8 with sodium hydroxide and chilled to 5 °C. Sodium hypochlorite was added slowly to a rapidly stirred solution at a molar ratio of at least 1:1.2 for hypochlorite to ammonia (16). The adjustment of the pH minimizes the disproportionation of monochloramine to dichloramine. Maintaining a slight excess of ammonia and slow addition of hypochlorite reduces the potential for breakpoint chlorination resulting from localized excesses of hypochlorite due to poor mixing. Chlorinated dimethylamine stock solutions were prepared by addition of sodium hypochlorite to dimethylamine in equimolar concentrations. All chlorinated compound stock concentrations were standardized iodometrically (17). Chlorination of NDMA. NDMA (3 mM) was mixed at various pH values with varying concentrations of sodium hypochlorite, monochloramine, and chlorinated dimethylamine (CDMA) taken from standardized stock solutions to evaluate the transformation of NDMA. All reaction vessels were maintained at 25 °C using a temperature controlled water bath. Samples were periodically withdrawn and analyzed by UV/visible spectrophotometry using a PerkinElmer Lambda 14 UV-Visible spectrophotometer and a quartz cuvette with a 1-cm path length. Wavelengths used for quantification were chosen to provide the greatest difference in molar absorptivities between NDMA and the chlorinating agents. Because the spectra of NDMA and the chlorinating agents overlapped, the concentrations were calculated using matrix algebra. Calculated concentrations were confirmed by using two pairs of wavelengths for each experiment and averaging the results [(262 nm, 292 nm) and (292 nm, 333 nm) for NaOCl at pH ) 6.8, (333 nm, 265 nm) and (333 nm, 350 nm) for NaOCl at pH 8 and 9.2, (331 nm, 350 nm) and (331 nm, 360 nm) for chlorinated dimethylamine, and (350 nm, 331 nm) and (300 nm, 300 nm) for monochloramine and dichloramine]. NDMA Formation. All glassware used during these experiments was rinsed with acetone and baked at 400 °C for at least 4 h prior to use. Reactions were conducted at 25 °C in volumetric flasks placed in a temperature-controlled water bath shielded from light. Reactions were conducted at buffer concentrations ranging between 10 and 50 mM. Total chlorine was determined iodometrically (17) at the end of each reaction and at other relevant times. Reactions were halted by the addition of excess ascorbic acid (10 to 40 mM) to quench the chlorinating agent. While most experiments were conducted in deionized water, some experiments were conducted using secondary wastewater effluent obtained from the Dublin/San Ramon Services District’s municipal wastewater plant located in Pleasanton, California. The sample was collected prior to chlorine disinfection and was filtered within 2 h of collection VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
589
TABLE 1. Experimental Results reagent 1
reagent 2
reagent 3
expt
reaction
identity
mM
identity
mM
1
A B C D E F G H A Be Cf A B C A B C A B B C D
NaOCl NH2Cl CDMAc CDMA NaOCl NH2Cl NH2Cl NH3 DMA DMA DMA DMA DMA DMA NH3 NH3 NH3 NH2Cl NH2Cl UDMH UDMH UDMH
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1.9 1.9 1.9 1.1 1.1 1 1 1
NH3 DMAb NH3 NH2Cl DMA TMAd DMA DMA NH3 NH3 NH3 NH2Cl NH2Cl NH2Cl NaOCl NaOCl NaOCl DMA DMA
1 1 1 1 1 1 1 1 1 1 1 1 1 1 0.7 1.4 4.5 1 1
HOOH NaOCl
1 1
2 3 4 5 6
identity
NDMA formation mM
NaNO2 HOOH NaOCl NaOCl NaOCl
1 1 2 2 2
WWh WW WW HCO3HCO3-
10 50
µΜ/h
