Environ. Sci. Technol. 2010, 44, 4678–4684
Metolachlor and Atrazine in the Great Lakes PERIHAN B. KURT-KARAKUS,† DEREK C.G. MUIR,† T E R R Y F . B I D L E M A N , * ,‡ J E F F S M A L L , † SEAN BACKUS,† AND ALICE DOVE† Science and Technology Branch, Environment Canada, 867 Lakeshore Road, Burlington, L7R 4A6, ON, Canada, and Center for Atmospheric Research Experiments, Environment Canada, 6248 Eighth Line, Egbert, L0L 1N0, ON, Canada
Received February 17, 2010. Revised manuscript received April 27, 2010. Accepted April 30, 2010.
Concentrations of atrazine and metolachlor and stereoisomer fractions (SF ) herbicidally active/total stereoisomers) of metolachlor were determined in 101 surface water samples collected from the five Laurentian Great Lakes in 2005-2006. Geometric mean (GM) concentrations of atrazine ranged from 5.5 to 61 ng L-1, decreasing from lakes Ontario ∼ Michigan ∼ Erie > Huron > Superior, while metolachlor concentrations ranged from 0.28 to 14 ng L-1 and showed similar trends among the lakes. Median SFs ranged from 0.527 (Superior) to 0.844 (Erie) with an overall value of 0.708, and were significantly different among the Great Lakes (p < 0.05), except for Michigan vs Huron and Michigan vs Ontario. The SF in Erie was closest to that of the dominant product in use, S-metolachlor (SF ) 0.880), while Superior showed an SF similar to that of racemic metolachlor (SF ) 0.500). The median SFs in lakes Ontario, Huron and Erie were significantly lower than the median SF in Ontario stream samples collected in 2006-2007. The lower SFs in lakes suggest in-lake stereoselective processing of metolachlor or hold-up of older racemic metolachlor residues.
Introduction The Great Lakes basin drains nearly 790 000 km2 and surveys in the 1980s indicated that one-third of the land use in the Great Lakes basin was based on agricultural activities, varying among the five watersheds from 3% Lake Superior, 27% Lake Huron, 44% Lake Michigan, 39% Lake Ontario, and 67% Lake Erie (1). Between 1982-2001, farmland in the U.S. basin states experienced a 12% decline due to development (2). Corn and soybeans account for approximately 70% of reported crop acreage in the basin (3) and the preemergent herbicides metolachlor [(aRS,1RS)-2-chloro-6′-ethyl-N-(2methoxy-1-methylethyl)acet-o-toluidide] and atrazine (6chloro-N-ethyl-N′-(1-methylethyl)-1,3,5-triazine-2,4-diamine) (Supporting Information (SI) Figure S1) are heavily used. Applications in Great Lakes watersheds of the U.S. during 2002 amounted to 1112 tonnes metolachlor and 2466 tonnes atrazine (4). Use in Ontario during 2003 was 543 tonnes metolachlor and 514 tonnes atrazine (5). In the U.S., 75% of metolachlor and 84% of atrazine applications nationally were to corn while soybeans accounted for 14% of metolachlor * Corresponding author phone: +1 705 458 3322; fax: +1 705 4583301; e-mail:
[email protected]. † Science and Technology Branch. ‡ Center for Atmospheric Research Experiments. 4678 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 12, 2010
use (6). Metolachlor in Ontario was used mainly on corn (57%) and soybeans (34%), and to a lesser extent on other crops, while 97% of atrazine was applied to corn (5). Although concentration data are available for selected organochlorine insecticides in the Great Lakes (7, 8), limited information is available for herbicides in open lake waters. Schottler and Eisenreich (9) measured atrazine and metolachlor in surface and deep water of the lakes between 1990-1993, while Tierney et al. (10) reported predicted atrazine concentrations for 1991-1995. Struger et al. (11) measured a larger group of insecticides and herbicides, including atrazine and metolachlor, in Great Lakes surface water between 1994-2000. Other recent studies have focused on streams and tributaries within the basin (12-15), or have emphasized atmospheric transport and deposition as a source to the lakes (16-20). Metolachlor consists of four stable stereoisomers due to two chiral centers, one on a carbon atom and the other from hindered rotation around the C-N axis (atropisomerism) (21, 22) (SI Figure S1). These are designated as aR,1′R; aR,1′S; aS,1′R; aS,1′S, where a ) axial chirality and 1′ ) carbon chirality, two pairs of enantiomers (aR,1′R; aS,1′S and aR,1′S; aS,1′R) and two pairs of diastereomers (aR,1′R; aS,1′R and aR,1′S; aS,1′S) (22, 23). Metolachlor was introduced into the market in 1976 as the racemic compound in which the enantiomers in each pair are in a 1:1 relationship, although the ratios of diastereomers need not be 1:1 (23). Most (95%) of its herbicidal activity resides in the aR,1′S and aS,1′S stereoisomers (23). The product enriched in these isomers is known as S-metolachlor, with a content of approximately 90% (1′S)-diastereomers (21), which is reported to achieve the same herbicidal effect at about 60% of the racemic product use rate (24). S-metolachlor has been used in Canada since 1998 and in the U.S. since 1996 (25). Registrations for racemic metolachlor in Canada expired at the end of 2006 and only formulations containing S-metolachlor have been sold since (26). Racemic (generic) metolachlor is still registered in the U.S. (27). In previous work, concentrations and stereoisomer proportions of metolachlor were determined in Ontario streams between 2003-2004 (13) and 2006-2007 (28). These studies showed that stereoisomer proportions in high-concentration samples were similar to those in S-metolachlor, whereas low concentration samples were characterized by stereoisomer profiles between racemic and S-metolachlor. This suggested carry over of older (racemic) metolachlor products, or stereoselective degradation in the watersheds. In this study, we report the concentrations of parent compounds and stereoisomer profiles of metolachlor in surface waters of the Great Lakes. Atrazine is included because of its similar usage pattern to metolachlor and to update concentrations in the lakes since the last measurements nearly a decade ago.
Materials and Methods Locations and Sample Collection. Details on sampling procedures for surface and depth samples are given in SI Table S1. Briefly, Great Lakes surface samples were collected as part of Environment Canada’s Great Lakes Surveillance Program (Figure 1) in 2005-2006. The ship-based surveys were conducted in the spring, when the lakes are isothermal. Samples were collected from approximately 4 m below the water surface (i.e., below the draft of the ship) using a clean technique sampler (29). In the upper lakes (Superior, 2005, n ) 14; Michigan, 2006, n ) 7; and Huron (including Georgian Bay), 2007, n ) 20), 24 L of filtered water was collected. In the lower lakes (Erie, 2006, n ) 10 and Ontario, 2005 and 10.1021/es100549v
Published 2010 by the American Chemical Society
Published on Web 05/26/2010
FIGURE 1. Sampling locations for Great Lakes surface water and depth profile samples. 2006, n ) 50), 16 L was collected. Samples were filtered into precleaned 4 L glass bottles, preserved in the field with dichloromethane (DCM; 50 mL per 4 L sample), and transported to the Canada Centre for Inland Waters in Burlington for extraction and analysis. Additionally, surface and deep water samples were collected for metolachlor from two locations in Lake Ontario in July 2007 from Site 1 and Site 2 (Figure 1). Surface water (70-100 L) was sampled using method described above, while depth profiles (∼38 L) were obtained with GO-FLO bottles (General Oceanics Inc., Miami, FL) Analytical Methods. Materials. Atrazine-d5, metolachlord6 and unlabeled analytical standard of atrazine were from C/D/N Isotopes (Pointe-Claire, Quebec, Canada), racemic metolachlor was from AccuStandard (New Haven, CT). Two metolachlor mixtures containing: a) aS,1’S, aR,1’S and b) aR,1’R, aS,1’R isomers were gifts from Syngenta Crop Protection (Guelph, ON, Canada). Dual MagnumR, a formulation of S-metolachlor, was used to characterize the isomer composition of a commercial metolachlor product that is currently used in Canada and was gifted by Muck Crops Research Station of University of Guelph. All solvents were chromatographic grade. Details regarding suppliers of solvents are given in SI Table S2. Extraction and Cleanup. Great Lakes surface water samples from spring campaigns were extracted using a Goulden large volume continuous-flow extractor (30, 31) with DCM while XAD-2 resin columns from the July 2007 expedition on Lake Ontario were eluted with MeOH. Cleanup of extracts in both cases was done on silica gel columns. Details on extraction and cleanup methods can be found in SI Table S3. Quantitative Analysis. Analysis was conducted by gas chromatography with mass selective detection (GC-MSD) using a 6890 GC - 5973 MSD (Agilent Technologies, U.S.). Ions monitored (target/qualifier) were metolachlor (162/238), atrazine (200/215). Samples were quantified using external standards and peaks were integrated using MSD Productivity Chemstation Software (G1701DA Rev. D.03.00SP1, Agilent). Details on instrumentation and operating conditions are given in SI Table S4. Chiral Analysis of Metolachlor Stereoisomers. Analysis of metolachlor stereoisomers was carried out by high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). The ion source was APCI (atmospheric pressure
chemical ionization) operated at 250 °C in positive ion mode. The ion transition monitored was 284.2-251.1. SI Table S4 gives more detailed information on the stereospecific separation procedure for metolachlor. Racemic metolachlor, two isomer mixtures from Syngenta and commercial S-metolachlor (Dual MagnumR) were used as standards in the chiral metolachlor analysis. The isomer elution order aR,1’R; aR,1’S; aS,1’R; aS,1’S was identified according to Polcaro et al. (32), who used the same chiral LC column with polarimetry and proton nuclear magnetic resonance (HNMR) spectroscopy for identification. Peak areas were integrated using Analyst software (version 1.4.1). Details regarding these procedures and example chromatograms of compound separations are given in the main paper and the Supporting Information of ref 13. Data are reported as the stereoisomer fraction (SF) of herbicidally active/total components (eq 1) (28). SF )
[aR, 1′S + aS, 1′S] [aR, 1′R + aR, 1′S + aS, 1′R + aS, 1′S]
(1)
Quality Control. A batch approach was used to process the Goulden extracted samples, with an analytical run consisting of 12-15 samples, a method blank and two spikes for recovery checks. Blank samples were treated in the same manner as lake water samples by extracting and cleaning up 20 L of Milli-Q water. Method spikes were prepared by adding a known quantity of unlabeled target analyte (50 ng atrazine and 125 ng metolachlor) into the Milli-Q water. Recoveries of spiked compounds were atrazine 97 ( 21%, ranging from 63 to 140% (n ) 24), metolachlor 140 ( 38%, ranging from 79 to 210% (n ) 24). Due to spike recovery problems for atrazine in one set of Lake Superior surface water samples, only 8 of 14 samples resulted with satisfactory atrazine concentrations. Therefore, results for atrazine are reported for eight samples while metolachlor results are reported for 14 samples. For field spiked samples processed through XAD, recoveries of metolachlor-d6 (1000 ng, n ) 19) and atrazined5 (1010 ng, n ) 19) were 108 ( 23% (range 58-139%) for metolachlor-d6 and 109 ( 17% (range 73-131%) for atrazined5. Results were not recovery corrected. The method detection limit (MDL) was calculated according to Gomez-Taylor et al. (33). The MDL was detected using ten portions of Milli-Q water which were spiked at concentrations of 0.25 ng L-1 (atrazine) and 0.63 ng L-1 (metolachlor) and subjected to the Goulden extraction and VOL. 44, NO. 12, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4679
4.7 33.6 1039 297 6630d 7 26 47 118 113
