Introduction

Chloroacetamide herbicides for pre-emergence weed control in various crops were first introduced in the early 1950s and 1960s. Although some of the first compounds were eventually replaced, the addition of safeners has extended the use of chloroacetamide herbicides for broadening the weed control spectrum when used in tank mixes.¹ Acetanilide-derived herbicides are used for pre-emergence weed control in corn and soybean, as well as sorghum (Sorghum bicolor), sunflower (Helianthus annuus), peanut (Arachis hypogaea), and cotton (Gossypium spp.).²

Chloroacetamide herbicides are often used in tank-mixes with other herbicides to provide broad-spectrum weed control. These herbicides play a vital role in managing herbicide-resistant weeds, including those resistant to glyphosate, but their limited activity on many broadleaf weeds has increased the need for additional herbicide chemistries to achieve effective weed control. Evaluations of new herbicide compounds are crucial for the development of new types of chemistry for weed control and for incorporation of these compounds into cropping systems to give farmers more diverse tools to manage weeds that are inherently difficult to control.

Pyroxasulfone is a new, pyrazole-based herbicide with the same mechanism of action as chloroacetamides, eg, the inhibition of very long chain fatty acid biosynthesis.³ This compound is used pre-emergent to control a broad spectrum of grasses and small-seeded broadleaves and is selective in corn, soybean, wheat, and sunflower.⁴ The advantage of pyroxasulfone over the chloroacetamide herbicides is its low use rate and activity on important broadleaf weeds such as Amaranthus spp.⁵ Pyroxasulfone recently received an Environmental Protection Agency (EPA) federal registration to be used in combination with flumioxazin in corn in the United States (Valent U.S.A. press release, 2011).

Herbicide sorption to soil influences a compound's environmental fate, persistence in the soil, and biological activity.⁶ It is important to examine the affinity of pyroxasulfone to soil to provide insight into its interaction with soil when applied under field conditions. Soil texture and chemical properties influence herbicide binding and subsequent weed control. Understanding the influences of these properties on the binding of pyroxasulfone will help us predict the behavior of the compound in different soils. Pyroxasulfone has been shown to provide comparable weed control to chloroacetamide herbicides when applied at much lower use rates,⁷ and investigating soil adsorption rates could provide insight into this difference in weed control efficacy in terms of amounts of available herbicide for plant uptake.

There have been numerous studies on how chloroacetamide herbicides bind to soil components. For example, soil organic matter is the predominant adsorbent for s-metolachlor.^6,8 This type of information is limited for pyroxasulfone, so we cannot compare soil interactions to commonly used chloroacetamide herbicides. Measuring the soil-sorption coefficients for pyroxasulfone and comparing these values to commonly used chloroacetamide herbicides will give us a predictor of how this compound will act in the field and how it can be incorporated into current agricultural practices.

The objectives of this study were to (a) compare the relative soil binding between pyroxasulfone, s-metolachlor, and dimethenamid-p across 25 different soil types and (b) evaluate the influence of different soil texture and chemical properties on soil binding of these three herbicides.

Materials and Methods

Herbicide Soil Adsorption

A mixed herbicide stock solution was prepared by combining 1 mg mL⁻¹ of pyroxasulfone, s-metolachlor, and dimethenamid-p together in a 0.02 M CaCl₂ solution. Given that herbicide adsorption is a competitive process, pilot studies were conducted to confirm that adsorption rates at 1 mg mL⁻¹ were similar for combined herbicide solutions compared to individually applied herbicides. Batch equilibrium studies were conducted by combining 10 g of dry soil with 10 mL of herbicide stock solution in capped, 50-mL glass centrifuge tubes, which were shaken horizontally for 24 hours on a table shaker. Control herbicide solutions without soil were also included and used for initial total herbicide concentrations in order to account for herbicide binding to glassware. After shaking, the samples were centrifuged at 1000 rcf (relative centrifugal force) for 10 minutes to separate the soil and herbicide solution. Three milliliters of the supernatant was combined with 3 mL of toluene and shaken for two hours on a horizontal shaker. After being centrifuged at 1000 rcf for 10 minutes, 2 mL of the toluene supernatant was transferred to a volumetric flask and spiked with 500 ng L⁻¹ butylate as an internal standard and then injected in a gas chromatography/mass spectrometry (GC/MS) column to quantify herbicide concentrations in the solution. The herbicide concentrations in the toluene phase were analyzed using a gas chromatograph equipped with a mass spectrometer (Shimadzu GC-17A and GCMS QO 5050A, Shimadzu Scientific Instruments, Inc., Columbia, MD, USA), which monitored the masses for butylate (m/z 146), s-metolachlor (m/z 162.15), dimethenamid-p (m/z 229.10), and pyroxasulfone (m/z 179.10). An RTX-5 30-m × 0.25-mm column (Restek, Bellefonte, PA) was used with a flow of helium at 1 mL min⁻¹.

Herbicide concentrations in the liquid solution were subtracted from initial total concentrations without soil to calculate the amount of herbicide bound to the soil. Ratios were then calculated by dividing the concentration of herbicide bound to the soil by the concentration in the soil solution as mathematically represented by equation (1).

K_oc for each herbicide was calculated as

where f_oc is the soil organic C mass-fraction 100 g soil⁻¹ that was measured for each soil f_oc was calculated as

where som is the soil organic matter and 1.72 is a conversion factor to estimate the fraction of organic carbon from the amount of soil organic matter.

Results

Herbicide Sorption

Soil textural and chemical properties for all soils are listed in Table 1. The K_d and K_oc values for all three herbicides in different soils are displayed in Table 2. The greatest amount of herbicide adsorption for all three compounds occurred in the Dodge County, Wisconsin, silt loam soil. For s-metolachlor and dimethenamid-p, the lowest adsorption occurred in the Pasco, Washington, sand soil. For pyroxasulfone, the lowest amount of adsorption occurred in the 3-River Montana sandy clay loam soil. Sorption coefficients ranged from 0.76 to 16.67 L kg⁻¹ for s-metolachlor, 0.32 to 9.57 L kg⁻¹ for dimethenamid-p, and 0.49 to 5.91 L kg⁻¹ for pyroxasulfone. When combined across all soil types, the mean and median K_d values were 4.0 and 2.8 L kg⁻¹ for s-metolachlor, 2.3 and 1.7 L kg⁻¹ for dimethenamid-p, and 1.7 and 1.5 L kg⁻¹ for pyroxasulfone, respectively. Pearson correlations were calculated comparing K_d values to soil textural and chemical properties, which are displayed in Table 3.

Table 1.

Soil physical and chemical properties for the 25 soils that were evaluated in order of increasing organic matter.

Table 2.

Sorption coefficients (K_d and K_oc) for s-metolachlor, dimethenamid-p, and pyroxasulfone in order of increasing organic matter.

Table 3.

Pearson correlations were evaluated for s-metolachlor, dimethenamid-p, and pyroxasulfone K_d and K_oc values versus soil chemical and physical properties.

When comparing the K_d values for all three herbicides, we observed a significant relationship with organic matter, silt, and sand at r² = 0.89, 0.63, and −0.52 for s-metolachlor; r² = 0.92, 0.65, −0.55 for dimethenamid-p; and r² = 0.94, 0.67, −0.66 for pyroxasulfone (Table 3). Soil pH, cation exchange capacity, and clay content were not significantly correlated to K_d for all three herbicides (Table 3). Correlations between organic matter, sand, and silt content resulted in significant linear relationships of −0.676 and 0.688, respectively (Table 4). High correlations among soil properties across all soils resulted in correlations between these soil properties and herbicide binding. Given that organic matter was highly correlated to herbicide binding (>0.89 for all herbicides), and also highly correlated to sand and silt content, this would explain the significant correlation with both sand and silt content and herbicide binding.

Table 4.

Pearson correlations between soil textural and chemical properties.

Fisher's LSD for all three herbicides was run across all soil types. Fisher's LSD = 0.6179, which resulted in statistical differences between s-metolachlor and both pyroxasulfone and dimethenamid-p, but there was no statistical difference between binding for pyroxasulfone and dimethenamid-p (Table 5). K_oc (organic carbon–water partitioning coefficient) is the distribution coefficient (K_d) normalized to total organic carbon content and is calculated from equation (2). K_oc values were calculated for all three herbicides across all soils. The average K_oc values ± standard error for s-metolachlor was 273 ± 23 L kg⁻¹, dimethenamid-p 151 ± 10 L kg⁻¹, and pyroxasulfone 117 ± 7 L kg⁻¹.

Table 5.

Fisher's LSD grouping for all three herbicides across all soil types (α = 0.05) LSD = 0.6179.

Discussion

Typically compounds with lower water solubility will have higher sorption coefficient values and will result in higher amounts of herbicide bound to the soil.⁹ Pyroxasulfone is a unique compound in that it has very low water solubility (3.49 mg L⁻¹ at 20°C), and yet has reduced soil binding compared to s-metolachlor and dimethenamid-p, which have higher water solubility values of 530 mg L⁻¹ and 1450 mg L⁻¹ at 20°C, respectively. Based solely on water solubility values, we hypothesized that pyroxasulfone would have the greatest sorption coefficient (K_d) value followed by s-metolachlor and then dimethenamid-p. However, when sorption coefficient studies were measured across 25 different soil types, we observe the trend of pyroxasulfone = dimethenamid-p < s-metolachlor in order of increasing soil binding. Average K_oc values listed in the WSSA Herbicide Handbook (9th Edition, 2007) for s-metolachlor (200 L kg⁻¹) and dimethenamid-p (55-125 L kg⁻¹) were similar to our average K_oc values of 268 and 149 L kg⁻¹ for s-metolachlor and dimethenamid-p, respectively. Although there is little sorption coefficient data for dimethenamid-p in the literature, our data show that relative differences in adsorption between s-metolachlor and dimethenamid-p were similar to the listed K_oc values for these two compounds.

With all three herbicides, organic matter resulted in a significant correlation with herbicide soil adsorption. Generally, compounds with lower water solubility are adsorbed to a greater extent by organic matter.^10,11 This would explain why pyroxasulfone binding was highly correlated to organic matter (OM); however, results for all three herbicides indicate that herbicide adsorption was highly correlated to organic matter (>0.89 for all three herbicides) regardless of differences in their water solubility (Table 6). Generally, adsorption of a herbicide is positively correlated with its octanol–water partition coefficient and negatively correlated with the compounds' water solubility.¹² These herbicide characteristics appeared to predict the relative differences in soil binding between s-metolachlor and dimethenamid-p, but failed to predict the amount of soil binding for pyroxasulfone based on its water solubility and log K_ow (Table 6).

Table 6.

Chemical structure, chemical formula, water solubility, log K_ow, and vapor pressure values for pyroxasulfone, s-metolachlor, and dimethenamid-p.

In general within a group of structurally related compounds, the phytotoxicity of the herbicides of higher water solubility was less influenced by organic matter than that for those materials of lower water solubility.^10,11,13 Bailey et al⁹ also concluded that within a chemically homologous series, the extent of adsorption was directly related to or governed by the compounds' water solubility. This would explain relative differences in adsorption between dimethenamid-p and s-metolachlor. It also shows the impact of differences in chemical structure and the influence of pyrazole- and chloroacetamide-based molecules on soil adsorption.

Weber et al¹⁴ found that nonionizable organic herbicides generally bind to OM more readily than to other soil colloids such as clay minerals or metallic hydrous oxides. In their study, they showed that only two out of eight nonionizable herbicides K_d were correlated to clay minerals. Whereas all eight of the nonionizable herbicides tested were correlated to organic matter content.¹⁴ Previous work with s-metolachlor has shown that soil retention was correlated to OM.^8,15,16 Others have shown that herbicide retention is correlated to both OM and clay content.¹⁷¹⁸¹⁹ Weber et al²⁰ showed that K_d values were highly correlated with soil OM (r = 0.97), clay content (r ≥ 0.79), and CEC (r ≥ 0.94), although they did show a high correlation between percent clay and CEC (r = 0.93), which would explain their high correlation between K_d values and CEC. However, they did not list correlations between clay and OM for the soils tested, which could explain the correlation between sorption and clay content. Clay content and OM are typically correlated to each other since soils with higher clay content are typically more productive and tend to return more carbon into the system annually, which contributes to increased OM content over time.²¹ In our study we observed a low correlation between OM and clay content (r = 0.208) and OM and CEC (r = 0.25) across all 25 different soil types (Table 6).

Although we did not have significant correlations to clay content or CEC, it is hard to say if correlations from previous literature were due to correlation between soil properties or if binding was actually correlated to these soil properties. For the 25 soil types used in our experiment, organic matter appeared to be the only soil property that was highly correlated to binding. Other soil properties and their correlation to herbicide binding could be explained by that soil properties correlation to organic matter. Our results are in agreement with previous authors, who found that OM was the main constituent for predicting binding. Results also differ from other previous reports that both OM and clay content were highly related to binding. Since we did not have a strong correlation between OM and either CEC or clay content, we can only conclude that OM was the dominant soil characteristic in terms of predicting herbicide binding.

Conclusion

Across all 25 soils evaluated, sorption coefficients indicate that pyroxasulfone had a lower degree of soil binding than dimethenamid-p and s-metolachlor, although pyroxasulfone behaved more similar to dimethenamid-p. Averaged across all soils evaluated, pyroxasulfone had the lowest average K_d value of 1.725 mg mL⁻¹ while dimethenamid-p and s-metolachlor had higher average values of 2.278 mg mL⁻¹ and 4.009 mg mL⁻¹, respectively. Results suggest that pyroxasulfone should be more available in the soil water solution than dimethenamid-p and s-metolachlor. Using a broad range of soils with diverse physical and chemical properties, statistical analysis suggests that organic matter correlates the best with herbicide binding for all three herbicides. Pearson correlations between sorption coefficient values and soil characteristics would support previous claims that pyroxasulfone activity under field conditions is inversely related to organic matter content.²²

Author Contributions

Conceived and designed the experiments: EPW and DLS. Analyzed the data: EPW. Wrote the first draft of the manuscript: EPW. Contributed to the writing of the manuscript: EPW, DLS, KB, RK. Agree with manuscript results and conclusions: EPW, DS, KB, RK. Jointly developed the structure and arguments for the paper: EPW. Made critical revisions and approved final version: DLS, KB, RK. All authors reviewed and approved of the final manuscript.