Environ. Sci. Technol. 2007, 41, 6933-6939
Comparison of Fate and Transport of Isoxaflutole to Atrazine and Metolachlor in 10 Iowa Rivers MICHAEL T. MEYER,* ELISABETH A. SCRIBNER, AND STEPHEN J. KALKHOFF United States Geological Survey, 4821 Quail Crest Place, Lawrence, Kansas 66049
Isoxaflutole (IXF), a newer low application rate herbicide, was introduced for weed control in corn (Zea mays) to use as an alternative to widely applied herbicides such as atrazine. The transport of IXF in streamwater has not been well-studied. The fate and transport of IXF and two of its degradation products was studied in 10 Iowa rivers during 2004. IXF rapidly degrades to the herbicidally active diketonitrile (DKN), which degrades to a biologically inactive benzoic acid (BA) analogue. IXF was detected in only four, DKN in 56, and BA in 43 of 75 samples. The concentrations of DKN and BA were approximately 2 orders of magnitude less than those of the commonly detected triazine and acetamide herbicides and their degradation products. Concentrations of IXF, DKN, and BA were highest during the May through June postplanting period. The concentration ratio of BA/DKN was similar to the deethylatrazine/atrazine ratio with smaller ratios occurring during May and June. The relative temporal variation of DKN and BA was similar to that observed for atrazine and deethylatrazine. This study shows that low application rate herbicides can have similar temporal transport patterns in streamwater as compared to more widely applied herbicides but at lower concentrations.
Introduction Numerous surface water studies have shown the pulsed transport of herbicides applied in row-crop agriculture in response to rainfall following application (1-7). Although the transport of triazine and acetamide herbicides has been addressed previously (6, 8-9), much less is known about the transport of the newer low application rate herbicides. Because these newer herbicides are effective at much lower application rates than traditional herbicides, deleterious effects on water quality may occur at lower concentrations than those observed for the older and more extensively used herbicides. It is also important to understand the effect of changing herbicide usage on the mixtures of compounds that are transported into surface water and groundwater from agricultural sources throughout the year. Isoxaflutole (IXF; [5-cyclopropyl-4-(2-methylsulfonyl-4(trifluoromethylbenzoyl)isoxazole]; see Figure S1 in the Supporting Information), a member of the isoxazole class of herbicides, is a relatively new low application rate herbicide that was licensed for use in 1999 under the trade name Balance. It is a restricted-use preemergent herbicide for the * Corresponding author phone: (785)832-3564; fax: (785)832-3500; e-mail:
[email protected]. 10.1021/es070903t Not subject to U.S. Copyright. Publ. 2007 Am. Chem. Soc. Published on Web 09/22/2007
control of broadleaf weeds and grass (10) and currently is registered for use on corn (Zea mays) in 18 states. The U.S. Environmental Protection Agency (U.S. EPA) has classified IXF as a probable human carcinogen (11), based on statistically significant increases in liver tumors in both sexes of mice and rats (12) and because of the concern for leaching, has a label restriction that prohibits application to sandy loam or sand surface soils with less than 2% organic matter (OM, upper 30 cm of soil) where the groundwater table is less than 7.6 m below ground surface (11). The U.S. EPA has also reported that IXF is expected to persist and accumulate in surface and groundwater and that IXF and DKN may accumulate at concentrations that would be toxic to nontarget plants (11). Several studies have explored the efficacy and degradation of IXF (13-19) and estimated the potential transport of IXF in surface and groundwater (20-21). However, to date, no published studies have evaluated the seasonal occurrence, fate, and transport of IXF and its sequential degradation products, diketonitrile (DKN; [1-(2-methylsulfonyl-4-trifluoromethylphenyl)-2-cyano-3-cyclopropylpropan-1, 3-dione]) and benzoic acid analogue (BA; [2-methylsulfonyl-4-(trifluoromethyl)benzoic acid]), in surface water. IXF usage has fluctuated from 97 000 kg in 1999, peaking at 199 000 kg in 2001, then decreasing to 146 000 kg in 2003 (22). Usage in Iowa was about 28 000 kg in 1999 and 2000, 104 000 kg in 2001, 67 000 kg in 2002, and 61 000 kg in 2003 (22). A high potential for crop injury, rotational restrictions, and cost were cited as concerns in the application of IXF in a Wisconsin environmental impact statement (23). These general concerns may, in part, explain the decrease in IXF usage from 1999 to 2003. Application rates for IXF range from 75 to 140 g of active ingredient per hectare per year as opposed to approximately 1-2 kg of active ingredient per hectare per year for herbicides such as atrazine and metolachlor (22). IXF is designed to rapidly undergo hydrolysis to form the herbicidally active DKN degradation product (24-26) through the opening of the isoxazole ring (27-28); DKN, in turn, is biotically degraded into its BA analogue, which is biologically inactive (16). In this sense, IXF is a precursor rather than a parent compound (29). The hydrolysis of IXF to DKN in soil is affected primarily by abiotic factors (e.g., soil moisture, temperature, and pH (30)), whereas the conversion of DKN to BA occurs primarily through biotic processes (26-28). A summary of physiochemical characteristics with references for IXF, atrazine, and metolachlor and selected degradates is shown in Table 1. The aqueous solubility of IXF is approximately 6.2 mg/L, and its reported half-life ranges from 0.3 to 7 days. Theoretically, IXF sorbs to the soil and under wet conditions is degraded to the more soluble (326 mg/L), and more stable with a half-life of about 8-61 days, and biologically active DKN (16). Degradation studies of IXF also have shown that DKN is degraded to BA (solubility, 8460 mg/L; half-life about 20-977 days. A 14C-IXF study (17) found two additional but minor degradation products of DKN presumably formed through biotic processes. Thus, IXF is a rapidly transformed precursor of the active herbicide, DKN, and BA is the primary degradation product of the herbicide DKN. The purpose of this study was to determine the seasonal occurrence and fate of IXF and its two sequential degradation products DKN and BA in surface water and to compare their transport to that of atrazine and metolachlor. VOL. 41, NO. 20, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6933
TABLE 1. Summary of Physiochemical Parameters for IXF, Atrazine, Metolachlor, and Selected Degradation Products
a
compounds
solubility (mg/L)
IXF DKN BA atrazine DEA DIA metolachlor metolachlor ESAa metolachlor OXAb
6.2 326 8460 33.8 3200 670 484
ESA ) ethane sulfonic acid
8500 b
The selection of sampling sites, sample collection and processing methods, and materials used for the analysis of IXF and its degradation products is provided in the Supporting Information. The locations, site names, and drainage areas of the sampling sites are provided in the Supporting Information as well (see Figure S2 and Table S1). Method for Analysis of IXF. IXF, DKN, and BA were analyzed using the method of Meyer et al. (42). Samples were extracted using Oasis hydrophilic-liophilic balanced (HLB) solid-phase extraction (SPE) cartridges (6 mL, 500 mg) on a vacuum manifold. The cartridges were rinsed with 5 mL of 50:50 methanol/acetonitrile followed by 5 mL of reagent water. A vacuum was applied to load the samples onto the SPE cartridges in 15-20 min. Cartridges then were rinsed with 3 mL of 1% formic acid (aqueous) and then eluted with four 4 mL aliquots of 50:50 methanol/acetonitrile. One hundred microliters of 0.615 ng/µL ISTD solution then was added to the sample eluates. All sample eluates then were evaporated at 45 °C under a gentle stream of nitrogen using a turbovap (Zymark, Inc., Hopkinton, MA) to a final volume of 200 µL, transferred to vials with 200 µL glass inserts, and stored in a freezer at -10 °C until analysis. Sample eluates (40 µL) were injected, and the compounds were separated and analyzed on an Agilent 1100 model D series liquid chromatograph (LC; Wilmington, DE) and a Waters Quattro Micro tandem mass spectrometer (MS/MS) with electrospray ionization (ESI) in negative-ion mode using multiple-reaction monitoring (MRM). The specific LC/MS/ MS conditions are described in Meyer et al. (42). The method detection levels (MDL) were established at 0.002 µg/L for IXF, DKN, and BA. Analysis of Triazine and Acetamide Herbicides and Degradation Products. The two methods used to analyze for triazine and acetanilide herbicides and degradation products are briefly described in the Supporting Information. Quality Assurance. The general description of the field and laboratory quality assurance samples and measurements and general quality assurance/quality control (QA/QC) results for the IXF method are provided in the Supporting Information. The MS performance was evaluated prior to each sample run by injecting 20 µL of the 0.123 ng/µL working standard mix and assessing the product-ion abundances and ratios. If the abundances were low, the general tune parameters and compound specific tune parameters were reoptimized.
Results and Discussion Comparison of Detections and Concentrations. The most striking feature of this study is the concentration distribution (Figure 1) and the frequency of detections (see Table S2 in the Supporting Information) of the herbicidally active IXF degradation product, DKN, relative to the much more extensively used triazine and acetamide herbicides. To understand the fate and transport of IXF, DKN, and BA, it is useful to compare their concentration distributions. Distri9
54-258 35 51 3.0-114 2.3-5 7
t1/2 (days)
refs
0.3-7 8-61 20-977 12-162 100 100 9-16
13, 21, 28 13, 21, 28 13, 21 31-39 38, 39 38, 39 40, 41 40, 41 40, 41
OXA ) oxalinic acid
Materials and Methods
6934
Koc
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007
butions of the detectable concentrations show that the concentrations of IXF, DKN, and BA varied from 0.002 to 0.077, 0.002 to 0.55, and 0.001 to 0.017 µg/L, respectively (Figure 1). The median concentrations for DKN and BA were 0.010 and 0.007 µg/L (Figure 1), respectively. The median concentrations of atrazine, metolachlor, and selected triazine and acetamide herbicide degradation products were approximately 2 orders of magnitude greater than the median concentration distribution of IXF and its degradation products (Figure 1). The relative distribution of the concentrations between DKN and BA was similar to the pattern observed between atrazine and its major dealkylated degradation product, deethylatrazine (DEA). For example, DKN and atrazine had a wider distribution and slightly higher median concentrations than their degradation products. However, a nonparametric Mann-Whitney U test showed no significant difference (p g 0.05) between the concentrations of DKN and the concentrations of BA and only a marginally significant difference (p < 0.05) between atrazine and DEA. Metolachlor also showed a wider distribution in concentrations than its degradation products, but the median concentrations of metolachlor ethanesulfonic acid (ESA) and metolachlor oxanilic acid (OXA) were higher than for metolachlor. A nonparametric (Mann-Whitney U) test showed a significant difference (p < 0.001) in the distribution of the concentration between metolachlor and metolachlor ESA. These tests indicate that there is probably a different relationship between the stability and the solubility of DKN and atrazine to their degradates than there is to metolachlor and its ESA degradate. The detection frequencies for the 75 samples collected were 5, 75, and 57% for IXF, DKN, and BA, respectively (see Table S2 in the Supporting Information). The low detection frequency and low concentration of IXF support the findings of previous studies (25-26), which showed that IXF degrades rapidly. The patterns shown by the study data described herein also were consistent with the greater solubility and stability of DKN relative to IXF (refs 27-28 and Table 1) and support the findings that DKN and BA can be readily transported to surface water (16, 20). The difference in the concentration distributions between IXF and its sequential degradation products as compared to the triazine and acetamide herbicides and their degradation products also is consistent with (i) the low application rate of IXF and (ii) the lower usage of IXF in Iowa (17% crop acreage for IXF, in comparison to 70% for atrazine and 20% for metolachlor (22)). The difference in atrazine and IXF usage also may be indicated by their concentration distribution during the May through June postplanting period (Figure 2). Atrazine is applied to approximately 4.1 times more acres in Iowa than IXF and is applied at about 13 times the rate of IXF. Thus, about 50 times more atrazine than IXF is applied in Iowa. During the May through June postplanting period, the median concentration of the IXF degradate, DKN, was approximately
FIGURE 1. Semilog box-plot distributions of detected concentrations of selected herbicides and degradation products in samples collected from 10 Iowa rivers during March through September 2004. 50 times less than the median concentration of atrazine. These data indicate that the transport of atrazine and IXF as DKN in surface water during this period reflects the difference in the overall usage and application rate of atrazine and IXF in Iowa. In addition, a nonparametric (Mann-Whitney U) test showed a significant difference (p < 0.001) in the distribution of the concentration between DKN and atrazine for the postplanting period. Temporal Variation of Detections and Concentrations. Previous studies also have shown temporal differences or patterns in the transport of herbicides and their degradation products in surface water (e.g., refs 6, 7, and 9). To assess the seasonal variation in detections of IXF and its sequential degradation products, DKN and BA, the samples were divided into three collection periods, March through April (preplanting), May through June (postplanting), and July through September (late-summer; see Table S3 in the Supporting Information). IXF was detected in only four samples and only during the postplanting period, whereas DKN was detected in more than 60% of the preplanting samples, 100% of the postplanting samples, and 59% of the late-summer samples. BA was detected in more than 30% of the preplanting and late-summer samples and in 85% of the postplanting
samples. The infrequent detection of IXF and its detection only in the postplanting samples were consistent with its low solubility and rapid degradation (Table 1). The relatively frequent detection of DKN and BA in the preplanting samples suggests that some DKN and BA persists in the soil from year to year and is subsequently available for transport into streamwater by overland flow runoff and/or discharge of contaminated groundwater to rivers during base-flow. Potential year-to-year carryover of DKN in the soil is also suggested by Bresnahan et al. (43), and groundwater contributions of herbicides to streamwater during base flow have also been shown (5, 48). The large percentage of detections of DKN and BA in the postplanting samples indicates that DKN is rapidly formed and that it degrades to BA but at a slower rate than it is formed. The frequent detection rates of DKN (59%) and BA (48%) in the late-summer samples also indicate that DKN degrades slowly to BA and that BA also degrades slowly. Reported soil half-lives of DKN range from 8 to 61 days and for BA from 20 to 977 days (Table 1). The higher solubility of BA relative to DKN and the longer half-life also may have an effect on the seasonal distributions of concentrations between BA and DKN. VOL. 41, NO. 20, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6935
FIGURE 2. Semilog box-plot distributions of detected concentrations of selected herbicides and degradation products in samples collected during preplanting (March through April), postplanting (May through June), and late summer (July through September) from 10 Iowa rivers in 2004. Table S3 in the Supporting Information shows the total number of samples and the detection frequencies for atrazine and metolachlor and selected degradation products of these herbicides. The data for the other triazine and acetamide herbicides and their measured degradation products are available in Scribner et al. (44). Atrazine and several of its degradation products were detected in 63-100% of the samples from all sampling periods with the exception of deisopropylatrazine (DIA), which was detected in only 32% of the preplanting samples. During the postplanting period, all the selected triazine and acetamide compounds were detected in more than 90% of the samples. Metolachlor, metolachlor ESA, metolachlor OXA, and acetochlor/metolachlor ESA secondary amide degradation products were 6936
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 20, 2007
detected in 93-100% of the samples from all three sampling periods. These data are consistent with detection frequencies from several other studies of triazine and acetamide herbicides (45-47). The detection frequency of DKN in the preplanting samples is approximately 30% less than that of atrazine and metolachlor. The detection frequency for DKN is similar for atrazine and metolachlor in postplanting samples (100 vs 100 and 97%, respectively); however, it is about 40% less in the late-summer samples (59 vs 100 and 100%, respectively; see Table S3 in the Supporting Information). The lower detection frequency of DKN and BA in the preplanting samples relative to atrazine and metolachlor and their degradation products most likely is related to the low
application rate of IXF and the continued degradation of the low levels of DKN and BA in the soil throughout the year after application. Figure 2 shows the concentration distribution for the preplanting, postplanting, and late-summer samples of DKN, BA, atrazine, DEA, metolachlor, and metolachlor ESA. The most obvious observations are (i) the concentrations of DKN and BA are generally less than 0.20 µg/L, which is the method reporting level for the triazine and acetamide analytical methods, (ii) the median concentrations of DKN and BA were 1-2 orders of magnitude less than the median concentrations of the triazine and acetamide compounds for the same sampling periods, and (iii) the highest median concentrations and widest range in concentrations for each of the compounds except metolachlor ESA were in the postplanting period. The data in Figure 2 also show that concentrations of BA and DEA are larger than those of their respective parent compounds, DKN and atrazine, in the preplanting samples and are much less than their respective parents in the postplanting samples. The data also show that there is substantial overlap of BA and DEA with their respective parent compounds in the late-summer samples. In contrast, the relation between metolachlor and its degradation product, metolachlor ESA, is much different with the concentrations of metolachlor ESA far exceeding those of metolachlor in the preplanting and late-summer samples but with metolachlor ESA concentrations very similar to metolachlor concentrations in the postplanting samples. The relative variation in concentrations with respect to sampling period for DKN and BA is more similar to that of atrazine and DEA than metolachlor and metolachlor ESA. These data suggest that the degradation processes that affect the transport of DKN and atrazine to surface water are similar. Because IXF degrades rapidly to DKN and DKN is the preemergent herbicide that controls the weeds, DKN can be considered a parent herbicide as is atrazine. Variation of Degradation Product/Parent Ratios. Studies have shown that the concentration ratios of deethylatrazine/ atrazine (DAR) in surface water are typically low during the spring flush because herbicides are applied primarily in the form of parent compounds with very little degradation products (48-50). As the parent compound degrades, the concentrations of the degradation products increase relative to the amount of the parent compound resulting in increasing DAR values in soil (48-49, 51). As atrazine is transported through the unsaturated zone, it degrades to DEA, which is preferentially transported, creating the high DAR values (4849, 51). Additionally, DAR is higher in surface water during low-flow periods when the relative contribution of groundwater to streamflow is substantial (5, 48). DAR values varied in this study from approximately 0.8 to 2 in the preplanting samples, from 0.1 to 0.8 in the postplanting samples, and increased to approximately 0.50.9 in late-summer samples (Figure 3). This variation in DAR has been observed in several studies (5, 48, 50). A DAR of 1 is indicative of longer soil residence times, either through slow unsaturated zone transport or transport of residual atrazine that was sorbed to the soil over the winter and transported into surface water with the first spring rains prior to application. A low DAR (
