Environ. Sci. Technol. 2006, 40, 3203-3210
Occurrence and Fate of Nitrosamines and Nitrosamine Precursors in Wastewater-Impacted Surface Waters Using Boron As a Conservative Tracer I. MARIE SCHREIBER AND WILLIAM A. MITCH* Department of Chemical Engineering, Yale University, Mason Lab 313b, 9 Hillhouse Avenue, New Haven, Connecticut 06520
Using boron as a conservative tracer of municipal wastewater effluents, a mass balance was developed to determine river flowrates that requires only wastewater discharge flowrates and boron concentrations in wastewater effluents and in the river upstream and downstream of these discharges. Furthermore, this method permits calculation of the percentage of the river deriving from wastewater. This method could be useful within river sections featuring no independent data regarding river discharge. We assessed the decay of nitrate and N-nitrosodimethylamine (NDMA) precursors within an engineered treatment wetland and, using our boron analysis technique to account for dilution, within the Quinnipiac River (CT). Although both decayed with several day half-lives, their slow decay indicates they can persist to impact downstream drinking water supplies. Concentrations of NDMA and N-nitrosomorpholine (NMOR) were measurable within the river, but concentrations of four other nitrosamines, their precursors, and NMOR precursors were not detectable.
Introduction Certain hydrophilic xenobiotics have been detected at high frequencies in wastewater-impacted streams (1, 2). To assess their transformations within surface waters, the hydrologic fate of water parcels is required to distinguish transformations from dilutions associated with baseflows from tributaries, groundwater inflows, or agricultural runoff. Using boron as a conservative tracer of municipal wastewater effluents (3), we developed a method to assess the hydrologic fate of discharges that requires only discharge rates and boron concentrations in effluents and upstream and downstream of the discharges. Using the Quinnipiac River (CT) as a case study, we demonstrate that this method can determine river flowrates in sections of the river not serviced by USGS gauging stations and the fraction of the river volume consisting of municipal wastewater discharges. The persistence of nitrogen in streams impacted by secondary wastewater effluents is a concern when these streams discharge to eutrophic waters. In addition to * Corresponding author phone: (203) 432-4386; fax: (203) 4324387; e-mail:
[email protected]. 10.1021/es052078r CCC: $33.50 Published on Web 04/06/2006
2006 American Chemical Society
inorganic nitrogen, wastewater-impacted surface waters contain significantly higher dissolved organic nitrogen (DON) concentrations than non-impacted surface waters (4). In regions where the removal of inorganic nitrogen via nitrification/denitrification processes at treatment plants does not prevent eutrophication (e.g., the Truckee River watershed (CA/NV) (5)), the persistence of DON is a concern. Limited research indicates that DON in wastewater effluents is more bioavailable (6) than humic substances in natural waters (7). Further assessment of bioavailability is complicated by the lack of DON characterization. Nitrosamines are a potent family of carcinogens that can form through a reaction between chloramines and organic nitrogen precursors (8-10); this DON fraction constituted a focus of this study. Nitrosamine formation is mainly associated with non-nitrified chlorinated wastewater effluents because these waters exhibit higher concentrations of DON precursors than natural waters (11, 12) and their high ammonia concentrations promote chloramine formation. The most probable DON precursors for nitrosamines include corresponding secondary amines (e.g., morpholine for NMOR) and tertiary amines with corresponding secondary amines as functional groups (11). Previous research on the organic nitrogen precursors of NDMA indicated that biological treatment effectively removed dimethylamine to concentrations accounting for 1: fdownstream_n ) fupstream_n
(
)
Qwwtp_n Qupstream_n + Qdownstream_n Qdownstream_n
Equation 6 is comprised of two components: the fraction of the river which is (1) derived from wastewater upstream of plant n and (2) originated as wastewater from treatment plant n. The fraction of the river upstream of plant n, which is attributable to wastewater discharges from plants 1 to n 1, can be expressed as
fupstream_n ) xn-1(fdownstream_n-1)
(4)
where
xn )
The fraction of the river downstream of plant n, which originated as wastewater, can be expressed as
If n ) 1:
After preliminary dilution, wastewater discharges are subject to two hydrologic fates between treatment plants: (1) mass removal at withdrawal locations and (2) concentration dilution due to inflow from tributaries, groundwater, and discharges from industrial facilities containing low boron concentrations. Withdrawals do not alter the tracer concentration, only the total tracer mass, in the river. Therefore, total dilution can be assessed by comparing boron concentrations downstream of a discharge site n and upstream of the subsequent discharge site n + 1. The degree of dilution can be evaluated using the following equation:
xnCdownstream_n + (1 - xn)Cbackground ) Cupstream_n+1
FIGURE 2. Concentration of NDMA precursors (2) and nitrogen species (0 ) nitrate, ] ) DON) within an engineered treatment wetland receiving effluent from the Mt. View Sanitary District’s wastewater treatment plant (Martinez, CA) over 7 d. NDMA precursor samples were collected at each sampling location for four consecutive days (April 22-25, 2003). Nitrate and DON samples were collected on the third and fourth day. Error bars represent 1 standard deviation. Note y-axis break and scale change.
(7)
which accounts for the degree of dilution between two consecutive treatment plants. The background boron concentration (Cbackground) for the Quinnipiac River, and tributaries and groundwater involved in dilution, was assumed to be the upstream boron concentration at Southington (SU, Figure 1).
Results Engineered Treatment Wetland: Nitrate concentrations declined by 40%, while DON concentrations doubled from 0.6 to 1.2 mg/L in the engineered treatment wetland (Figure 2) over ∼7 d of treatment. The increase in DON concentrations is likely attributable to exudates from bacteria and algae following nitrate uptake. NDMA concentrations in effluent from the treatment plant, which practices ultraviolet disinfection, and within the marsh were negligible. Dimethylamine concentrations declined from 1.2 µg/L ((0.2 µg/L standard deviation) at the first sampling location (∼0.1 d) to 1: Cupstream_n )
(
)
fupstream_n (C - Cbackground) fdownstream_n-1 downstream_n- 1 + Cbackground
Cdownstream_n ) Cupstream_n(1 - Xn) + Cwwtp_nXn Xn )
(
)
fdownstream_n - fupstream_n 1 - fupstream_n
where
Cbackground ) background concentration (mg/L) Cdownstream_n ) concentration downstream (mg/L) Cupstream_n ) concentration upstream (mg/L) Cwwtp_n ) plant effluent concentration (mg/L) fdownstream_n ) fraction of river downstream originating from wastewater fupstream_n ) fraction of river upstream originating from wastewater Xn ) fraction of the river downstream originating only from effluent n Since the wastewater fractions (f) correct the predicted concentrations for dilution effects, any significant decrease in measured concentrations compared to predictions would indicate degradation within the river. Although NDMA was detected within all wastewater treatment plant effluents (Table 2), only the concentrations in the Meriden (ME) and Wallingford (WE) effluents were above the 12 ng/L practical quantitation limit (i.e., three times the detection limit). NDMA concentrations within the river did not significantly exceed the detection limit except downstream of the Wallingford (WD) treatment plant, where NDMA concentrations increased by a factor of 3-4 during both seasons (Figure 4A and B). Wallingford effluent (WE) exhibited NDMA concentrations over an order of magnitude
TABLE 2. Nitrosamine, Nitrosamine Precursor, and Nitrogen Species Concentrationsa in Municipal Wastewater Effluents municipal wastewater treatment plant
NDMA (ng/L)
NMOR (ng/L)
Spring
NDMA-FPb (ng/L)
Southington Cheshire Meriden Wallingford
7.6 ( 2.1c 108 ( 49 617 ( 9 8.2 ( 1.8 56 ( 10 694 ( 37 52 ( 8 1390 ( 260 850 ( 95 253 ( 31 78 ( 31 1310 ( 105
Southington Cheshire Meriden Wallingford
10.3 ( 5.9 11.1 ( 5.6 13.4 ( 6.7 400 ( 96
Summer
59 ( 17 23.5 ( 7.8 94 ( 23 25.8 ( 7.6
nitrate (mg/L) 20.9 ( 0.3 30.3 ( 1.1 10.4 ( 1.2 17.8 ( 3.3
511 ( 73 13.8 ( 0.7 247 ( 51 17.7 ( 0.7 611 ( 165 6.9 ( 0.6 760 ( 145 12.0 ( 0.5
a Concentrations of NMEA, NDEA, NPYR, NPIP, and all measured secondary amines were below detection limits (Table SI-2). Concentrations of ammonia, nitrite, and DON were not statistically significant. b NDMA Formation Potential (NDMA-FP) concentrations corrected for background NDMA concentrations. c (standard error (n ) 4).
higher than those from the other three plants (Table 2). NDMA in the Wallingford effluent (WE) likely originated from an industrial source as Wallingford, a large industrial town, does not practice chlorine disinfection. NMOR was detected in all effluents; however, concentrations only exceeded the practical quantitation limit in effluents from Southington (SE), Meriden (ME), and Wallingford (WE, spring only) (Table 2). During the spring, NMOR concentrations within the river did not significantly exceed the detection limit except downstream of the Meriden (MD) plant, where concentrations were ∼100 ng/L (Figure 4C). Other nitrosamines did not exceed their detection limits (Table SI-2) within plant effluents or the river. Because NDMA and NMOR concentrations were not quantifiable in most river sections, an assessment of their persistence was not feasible. NDMA (Figure 4A and B) and NMOR (Figure 4C) concentrations decreased between Wallingford (WD) and North Haven (NHU). However, these decreases could not be definitively attributed to degradation, because boron loadings (Figure 3) not associated with municipal effluents in this section prevented the use of the boron tracer analysis to account for dilution from tributaries such as Wharton Brook. NDMA precursor and nitrate effluent concentrations (Table 2) generally were at least 3 times greater than concentrations measured in the Quinnipiac River. Between the Southington (SD) and Cheshire (CU), and Cheshire (CD) and Meriden (MU) treatment plants, NDMA precursors (Figure 5A and B) decreased by ∼10% and 30% in the spring and summer, respectively. The decline in NDMA precursor concentrations between subsequent treatment plants (Southington to Meriden) can be attributed to degradation. NDMA precursor concentrations were ∼35% and ∼60% lower than predicted concentrations upstream of the North Haven (NHU) plant during the spring and summer, respectively; however, we could not use the boron tracer analysis to control for dilution in this section. NDMA precursor concentrations increased by more than 50% between the Meriden (MD) and Wallingford (WU) plants during both seasons. Since there are no major tributaries between these plants, the source of these additional precursors is unclear. A wastewater odor was detected upstream of the Wallingford (WU) plant. The additional precursors may have arisen from inputs of raw sewage resulting from broken sewer lines. Other nitrosamine precursors, including those of NMOR, and secondary amines were not detectable (Table SI-2) either in the river or in wastewater effluents. Variability in daily nitrate concentrations in the spring precluded the assessment of degradation. In the summer, VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3207
FIGURE 4. Nitrosamine concentrations measured (]) and predicted (gray bars) along the Quinnipiac River (CT). NDMA concentrations during the spring (A) and the summer (B). NMOR concentrations during the spring (C). Sampling sites: Southington (S), Cheshire (C), Meriden (M), Wallingford (W), and North Haven (NH) wastewater treatment plants either upstream (U) or downstream (D) of wastewater discharge. Error bars represent 1 standard error (n ) 4). 3208
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
FIGURE 5. Organic and inorganic nitrogen concentrations measured (]) and predicted (gray bars) along the Quinnipiac River (CT). NDMA precursor concentrations during the spring (A) and the summer (B). Nitrate concentrations during the summer (C). Sampling sites: Southington (S), Cheshire (C), Meriden (M), Wallingford (W), and North Haven (NH) wastewater treatment plants either upstream (U) or downstream (D) of a wastewater discharge. Error bars represent 1 standard error (n ) 4). nitrate (Figure 5C) concentrations increased downstream of each plant but declined by ∼15% between consecutive plants upstream of Wallingford. Overall nitrate degradation upstream of Wallingford (WU) was ∼40% compared with
predicted concentrations. During both seasons, DON concentrations were not statistically significant due to high nitrate relative to total nitrogen concentrations.
Discussion Due to concerns regarding the fate in surface waters of hydrophilic contaminants within municipal wastewater discharges, researchers have sought conservative tracers for wastewater, including caffeine (20) and boron (3, 21). Previous studies have used grab samples of these tracers to identify wastewater-impacted waters (3, 21, 22). Two recent studies using gadolinium (which was shown to correlate sufficiently with boron (22)) as a wastewater tracer in the Teltow canal, Germany (22), and the Vltava River, Czech Republic (23), demonstrated significant increases in gadolinium concentrations after wastewater discharges to the rivers. The Czech study combined available hydrologic information regarding river flowrates, wastewater discharges and drinking water withdrawals in mass balance models to predict gadolinium concentrations in the river. These values were successfully compared to measured concentrations. To our knowledge, the use of conservative tracers to assess the hydrologic fates of wastewater discharges within mass balance models has required independent information regarding river flowrates, discharges, and withdrawals. In the United States, such hydrologic information generally is not available in river sections not serviced by USGS gauging stations. These gauging stations are relatively rare (e.g., only one station was present among the five municipal treatment plants along the Quinnipiac River). The determination of river discharge via stations that do not feature a weir can be labor intensive, because frequent measurements of changing river cross-sectional areas are required to convert measured velocities into flowrates. Our tracer technique uses only municipal wastewater effluent flowrates and boron concentrations in effluents and at river locations upstream and downstream of each discharge site. Although no independent measurements of flowrates were needed, our method successfully determined river flowrates within 10% of those recorded by a gauging station. This could be a valuable technique for river sections not serviced by gauging stations. In addition, this technique might be less labor-intensive than the current methods for discharge measurements used by gauging stations not featuring a weir. Four limitations were indicated for our method. First, boron mass loadings from municipal treatment plants with low discharge flowrates (e.g., the Cheshire plant, 0.1 m3/s) may be insufficient to cause the significant increases in boron concentrations required for flowrate determinations. Second, even relatively large treatment plants will not generate measurable increases in boron concentrations downstream when a significant fraction of the river is already comprised of wastewater effluent (e.g., Wallingford). Third, high background boron concentrations might prevent use of the method. Last, industrial discharges of boron will interfere. The last two limitations might be avoided if a tracer that features low background concentrations and has minimal industrial uses, such as gadolinium, were used. The ability to assess the fraction of a river consisting of wastewater would be useful in evaluating the impacts of wastewaters on water quality. Using our tracer analysis, we determined that the percentage of the Quinnipiac River deriving from wastewater increased along the river to 16% at Wallingford during the spring high baseflow season and 41% during the summer low baseflow season. These values were consistent with those determined by dividing the total measured discharges of wastewater plant effluents by the USGS measured river flowrate at Wallingford. Besides requiring USGS data, the latter calculation would be much less accurate had significant withdrawals occurred upstream.
Actual withdrawal volumes generally are unknown. In 2000, the Connecticut Department of Environmental Protection derived a water balance using maximum volumes for the 103 registered or permitted withdrawals along the Quinnipiac River which predicted that withdrawals exceeded discharges by 3.3 m3/s. For comparison, the mean annual discharge of the Quinnipiac River at Wallingford was only 5.9 m3/s that year. Because our results are consistent with the simple calculations based on discharge volumes, this indicates actual withdrawals are much lower than permitted withdrawals. Although useful, wastewater fraction calculations are cumulative, requiring boron measurements and calculations at all upstream municipal wastewater plants. While manageable for a small river, such as the Quinnipiac, this technique may not be feasible for rivers, such as the Mississippi, with many wastewater discharges. Eutrophication resulting from anthropogenic inputs of nitrogen to estuaries, such as Long Island Sound, is a current concern. Using our boron tracer analysis to correct for dilution within the Quinnipiac River, we determined that ∼40% of the nitrate was removed during 2-5 d of travel time. Similarly, nitrate was removed by 40% over the ∼7 d of treatment in the engineered wetland. Observed denitrification rates (i.e., ∼900 µmol N m-2 h-1) within the Quinnipiac River were comparable to those observed in wastewater-impacted rivers in France (24). Although NDMA and NMOR concentrations were too low within the river to assess their persistence, we measured concentrations greater than 15 ng/L and 100 ng/L, respectively. Fortunately, the Quinnipiac River is not used as a drinking water supply, because although no federal drinking water maximum contaminant level (MCL) for NDMA has been established, 10 ng/L has been established as an action level in California (25) and as a guideline in Massachusetts (26). No contaminant levels have been established for NMOR, even though it was recently detected in drinking water at a concentration of 1 ng/L (27). NMEA, NDEA, NPYR, and NPIP were not detectable either in wastewater effluents or within the river. The use of wastewater-impacted surface waters as drinking water supplies is escalating. A 1980 study estimated water supplies, which consisted of nearly 100% municipal wastewater effluent during low baseflow conditions, supplied approximately 4 million of the 62 million people served by surface water supplies (28). Communities exploiting such waters could be negatively impacted by the persistence of NDMA-forming precursors, particularly when their drinking water plants practice chloramination. NDMA precursors were removed by ∼10% (spring) and 30% (summer) between treatment plants in the upper section of the Quinnipiac and by approximately 50% after 7 d of wetland treatment. Assessment of significant differences in observed decay rates of NDMA precursors within the river (0.2 d-1 (spring), 0.3 d-1 (summer)) and the engineered treatment wetland (0.1 d-1) are hampered by variations in water temperature (e.g., 8 °C (spring), 22 °C (summer) for the river, (15)) and flowrate (e.g., ∼0.5 d (spring), ∼1.0 d (summer) between treatment plants). Nevertheless, the low decay rates observed in all systems indicate that NDMA precursors could persist and negatively impact downstream communities. For example, the half-life of NDMA precursors at an observed decay rate of 0.2 d-1 would be 3.5 d. Over this time period, a discharge into the Mississippi River from St. Louis (MO) would travel ∼200 miles (15), or nearly the distance to Memphis (TN).
Acknowledgments Acknowledgment is made to the Donors of the American Chemical Society Petroleum Research Fund for partial support of this research. I.M.S was supported by a National Science Foundation graduate fellowship. We would like to VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3209
thank the personnel of the Mt. View, Southington, Cheshire, Meriden, and Wallingford wastewater treatment plants, and Kathryn Johnson and Laura U for help with sample collection.
Supporting Information Available Water flow diagrams and residence time calculations for the Mt. View Sanitary District Wastewater Treatment plant, and quantitation ions and detection limits for nitrosamines and secondary amines. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Golet, E. M.; Alder, A. C.; Giger, W. Environmental exposure and risk assessment of fluoroquinolone antibacterial agents in wastewater and river water of the Glatt Valley Watershed, Switzerland. Environ. Sci. Technol. 2002, 36, 3645-3651. (2) Daughton, C. G.; Ternes, T. A. Pharmaceuticals and personal care products in the environment: Agents of subtle change? Environ. Health Perspect. 1999, 107, 907-938. (3) Vengosh, A.; Heumann, K. G.; Juraske, S.; Kasher, R. Boron isotope application for tracing sources of contamination in groundwater. Environ. Sci. Technol. 1994, 28, 1968-1974. (4) Westerhoff, P.; Mash, H. Dissolved organic nitrogen in drinking water supplies: a review. J. Water Supply: Res. Technol.sAQUA 2002, 51, 415-448. (5) Pehlivanoglu, E.; Sedlak, D. L. Bioavailability of wastewaterderived organic nitrogen to the alga Selenastrum Capricornutum. Water Res. 2004, 38, 3189-3196. (6) Seitzinger, S. P.; Sanders, R. W. Contribution of dissolved organic nitrogen from rivers to estuarine eutrophication. Mar. Ecol.: Prog. Ser. 1997, 159, 1-12. (7) Stepanauskas, R.; Leonardson, L.; Tranvik, L. J. Bioavailability of wetland-derived DON to freshwater and marine bacterioplankton. Limnol. Oceanogr. 1999, 44, 1477-1485. (8) Schreiber, I. M.; Mitch, W. A. Influence of the order of reagent addition on NDMA formation during chloramination. Environ. Sci. Technol. 2005, 39, 3811-3818. (9) Mitch, W. A.; Sedlak, D. L. Formation of N-nitrosodimethylamine (NDMA) from dimethylamine during chlorination. Environ. Sci. Technol. 2002, 36, 588-595. (10) Choi, J. H.; Valentine, R. L. Formation of N-nitrosodimethylamine (NDMA) from reaction of monochloramine: a new disinfection by-product. Water Res. 2002, 36, 817-824. (11) Mitch, W. A.; Sedlak, D. L. Characterization and fate of N-nitrosodimethylamine precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 2004, 38, 1445-1454. (12) Gerecke, A. C.; Sedlak, D. L. Precursors of N-nitrosodimethylamine in natural waters. Environ. Sci. Technol. 2003, 37, 13311336. (13) “National Primary Drinking Water Regulations: Disinfectants and Disinfection Byproducts; Final Rule,” United States Environmental Protection Agency. Fed. Reg., 63:241:69390-69476, 1998.
3210
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
(14) Bogaert, D. Personal Communication from Mt. View Sanitary District, Martinez, CA, April 2003. (15) US Geological Survey; USGS - Water Resources of the United States; 2004; http://water.usgs.gov/. (16) “Preliminary assessment of water withdrawals and streamflows in the Quinnipiac River watershed,” Quinnipiac River Watershed Partnership Work Group, July 2000. (17) Fischer, H. B.; List, J. E.; Koh, C. R.; Imberger, J.; Brooks, N. H. Mixing in Inland and Costal Waters; Academic Press: NY, 1979. (18) 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, 3733-3741. (19) Standard Methods for the Examination of Water and Wastewater, 20th ed.; Eaton, A. D., Clesceri, L. S., Greenberg, A. E., Eds.; American Public Health Association, American Water Works Association, Environment Federation Publishers: Washington D.C., 1998. (20) Buerge, I. J.; Poiger, T.; Muller, M. D.; Buser, H. R. Caffeine, an anthropogenic marker for wastewater contamination of surface waters. Environ. Sci. Technol. 2003, 37, 691-700. (21) Chetelat, B.; Gaillardet, J. Boron isotopes in the Seine River, France: A probe of anthropogenic contamination. Environ. Sci. Technol. 2005, 39, 2486-2493. (22) Knappe, A.; Moller, P.; Dulski, P.; Pekdeger, A. Positive gadolinium anomaly in surface water and ground water of the urban area Berlin, Germany. Chem Erde-Geochem. 2005, 65, 167189. (23) Moller, P.; Paces, T.; Dulski, P.; Morteani, G. Anthropogenic Gd in surface water, drainage system, and the water supply of the city of Prague, Czech Republic. Environ. Sci. Technol. 2002, 36, 2387-2394. (24) Chesterikoff, A.; Garban, B.; Billen, G.; Poulin, M. Inorganic nitrogen dynamics in the River Seine downstream from Paris (France). Biogeochemistry 1992, 17, 147-164. (25) California Department of Health Services; NDMA in California Drinking Water; March 15, 2002; http://www.dhs.ca.gov/ps/ ddwem/chemicals/NDMA/history.htm. (26) Massachusetts Department of Environmental Protection; Drinking Water Standards, Guidelines, and Policies; http://www. mass.gov/dep/brp/dws/standard.htm. (27) Charrois, J. W. A.; Arend, M. W.; Froese, K. L.; Hrudey, S. E. Detecting N-nitrosamines in drinking water at nanogram per liter levels using ammonia positive chemical ionization. Environ. Sci. Technol. 2004, 38, 4835-4841. (28) Swayne, M. D.; Boone, C. H.; Bauer, D.; Lee, J. S. “Wastewater in Receiving Waters at Water Supply Abstraction Points,” United States Environmental Protection Agency. EPA-600/2-80-044, 1980.
Received for review October 19, 2005. Revised manuscript received March 7, 2006. Accepted March 13, 2006. ES052078R
