The propiconazole standard was obtained from the U.S. Environmental Protection Agency National Pesticide Standard Repository (Ft. Meade, MD, USA); its purity was 99.6%. Solutions of the fungicide were prepared at 5 μg/mL in methyl tert-butyl ether (MTBE) and analyzed by GG-MS to check for purity, which nominally matched the given percentage purity. Methanol, MTBE, other organic solvents, sodium hydroxide and inorganic buffer salts were of analytical grade from Fisher Chemicals (Fair Lawn, NJ, USA). CE reagents, including 2-hydroxypropyl γ-cyclodextrin (the chiral selector) and sodium dodecyl sulfate (for micelle formation) were obtained from Sigma-Aldrich. Reagent water for all experiments was produced by a Barnstead Nanopure Infinity water purification system (Thermo Fisher Scientific, Waltham, MA, USA).

Soil 1 (USDA): J. Phil Campbell Natural Resource Conservation Center, Watkinsville, GA, USA. This is a composite agricultural soil collected from the A horizon of a cropped field on this USDA property. After collection, the soil was air-dried, sieved to <2 mm and riffled for homogenization.

Soil 2 (UGA): Univ. of Georgia Horticulture farm, Watkinsville, GA. This is a composited soil collected from the top 10 cm of the peach tree area of the horticulture farm. It was air-dried, sieved to <1 mm and homogenized (this soil was collected at a different time from soil 1 and inadvertently sieved to a smaller particle size.)

Physicochemical characteristics of soils [CEC = cation exchange capacity (meq/100g) and OM = organic matter]:

Soil 1: pH, 6.2; CEC, 8.9; sand, 70%; soil, 19%; clay, 11%; OM, 3.0%.

Soil pH was measured in a slurry of 1:2.5 soil to 0.01 M CaCl₂ solution (wt:vol). The other soil parameters were measured by a contractor using standard methods.

The experiment required preparation of 50 μg/mL of propiconazole in 1:5 (wt:vol) soil-water slurries of each soil in mineral/nutrient water. For this, 25 mg of propiconazole was dispensed in methanol solution onto the bottom of an autoclaved (121 °C for 20 min) Erlenmeyer flask; the solvent was allowed to evaporate under ambient conditions. Simultaneously, 500 mL of an aqueous solution of minerals and nutrients used to support growth of standard microbial cultures [21] was autoclaved, covered with foil, and cooled. This solution was added to the flask containing the propiconazole and the contents were stirred overnight with a magnetic stirrer at room temperature to produce 50 μg/mL of the fungicide in the mineral/nutrient water, as verified by the GC/MS analytical method described below. Aliquots of 20 mL of this aqueous solution were transferred to each of several 50 mL amber serum vials which had been autoclaved. Four grams of soil 1 (not autoclaved) were then added to each of these vials. This process was repeated with soil 2. These vials were then stoppered and shaken overnight, and constituted the active (spiked) treatments.

There were also produced, by autoclaving the soils, a corresponding number of controls. For this, 4 g of soil 1 was wetted with 2.0 mL of water and transferred to each of several amber serum vials; this was repeated for soil 2. These soils were autoclaved (121 °C for 20 min) once a day for 3 days; then 20 mL of the above autoclaved mineral/nutrient solution containing 50 ug/mL of propiconazole was added to each vial and the vials were shaken overnight. All vials, of both active and control treatments, for these kinetic experiments were shaken continuously in the dark in a thermostated shaker at 20 °C for the course of the experiment.

At t₀ and each successive time point, aliquots of the water phase were taken from all vials, active and control, for CE and, in some cases, GC-MS analysis. First, each vial was manually shaken to mix the slurry, then the stopper was immediately removed and about 1mL of the slurry was collected with a wide-mouth pipet and dispensed into a 2 mL centrifuge tube. The tube and contents were centrifuged at 14,000 rpm for 20 min (Eppendorf centrifuge 5418). As much as possible of the aqueous phase was removed with an automatic pipet. For CE analysis, 40 μL of this was fitered through a 0.45 μm syringe filter into a CE analysis vial. For GC-MS, 0.5 mL of this phase was extracted with 0.5 mL of methyl tert-butyl ether (MTBE) by vibration with a vortex mixer; 100 μL of this extract was diluted with 900 μL of MTBE for final GC-MS analysis.

Recovery from deionized water spiked at 50 μg/mL; average of duplicates: 97%.

Recovery from spiked soil-water slurries: a slurry of 5 mL of water and 0.5 g of soil 1 containing 50 μg/mL propiconazole was prepared as described above (section 2.3) and shaken overnight at 24 °C. The slurry was vortexed, a 1 mL sample of the slurry was collected, and the water was extracted as described above (see Extraction and Sample Preparation Methods) for subsequent analysis by CE. The solid (soil phase) remaining in the centrifuge tube was washed with 1.0 mL water, which was discarded. (This step was included to remove most of the soil pore water, which may have contained dissolved propiconazole, from the soil phase to allow for more accurate soil analysis.) The wet soil was extracted by vortexing with 1.0 mL of MeOH 2 times; extracts were then combined and evaporated to 0.5 mL. MeOH was selected as the extraction solvent because of its water miscibility. This extract was diluted with 1.5 mL of water to achieve an appropriate 25% MeOH aqueous solution for CE analysis; 40 μL of this was filtered through a 0.45 μm syringe filter. Both water and soil phases were analyzed by CE (see section 2.6). This procedure was repeated for soil 2, in duplicate.

Average % recovery (duplicates) of propiconazole from water phase of spiked soil:water slurry: Soil 1, 34.7%; Soil 2, 30.9%.

Average % recovery (duplicates) of propiconazole from soil phase of spiked soil:water slurry: Soil 1, 40.6%; Soil 2, 34.1%.

Total recoveries (from water + soil): Soil 1, 75.3%; Soil 2, 65.0%. Apparently, about 25% and 35%, respectively, of propiconazole was irreversibly sorbed to the soil, lost in the 1 mL of water used to wash the soil, or lost by sorption to glassware during the extraction process. The total recoveries are considered adequate for these experiments.

CE analysis was with a liquid cooled Beckman P/ACE System 5500 CE, with diode array UV detector, hydrodynamic injection, power supply up to 30 kV, and System Gold version 8.1 chromatography software (Beckman Instruments, Fullerton, CA, USA). The CE column was uncoated fused silica: 75 μm id, 300 μm od, 57 cm total length, and 50 cm effective length (MicroSolv Technology Corp., Long Branch, NJ, USA). Samples were contained in vial inserts for small volumes (40 μL or 400 μL, Beckman Instruments). The CE samples prepared during the sample extraction procedure described above were filtered through a 0.45 μm nylon syringe filter (or equivalent) just before transfer to the CE vials.

Before analysis of each sample, the column was washed with distilled water for 2 min, 0.10 M NaOH for 2 min, water again for 2 min, and electrolyte solution for 2 min. The column was maintained at 23 °C. Sample injection was hydrodynamic, usually for 6.5 sec; injections of 10 sec were used if more sensitivity was needed. Typical MEKC electrolyte composition was: 25 mM NaPO4 buffer, pH 7; 75 mM sodium dodecyl sulfate; 30 mM 2-hydroxypropyl-γ-cyclodextrin; and 10% methanol and 5% acetonitrile as organic modifiers. Instrumental conditions: temperature, 23 °C; detector wavelength, 190 nm; voltage, 30 kV; and run time, 30 min. The limit of quantitation for each of the 2 minor propiconazole stereoisomers in the water phase of a water-soil slurry was observed to be about 0.3 mg/L. However, for quantitation of all four stereoisomers (total propiconazole), the limit is about 2.5 mg/L.

GC-MS analysis of propiconazole standards and extracts was by use of a Hewlett-Packard 5973 mass spectrometer linked to a 6890 gas chromatograph equipped with a BGB 172 (BGB Analytik AG, Switzerland) chiral column. Column description: 30 m × 0.25 mm ID × 0.25 μm film thickness; chiral stationary phase, 20% tert-butyldimethylsilylated-β-cyclodextrin. GC conditions were: injection, splitless at temp. 275 °C; column temp. program, 150–220 °C at 4°/min, followed by temp. hold for 60 min; helium gas flow, 1.5 mL/min; MS inlet temp., 275°; MS source temp., 230°; and fragmentation voltage, 70 eV. Sample injection volume was 1 μL. Detection was by selected ion monitoring (SIM); SIM ions were m/z 173, 259 and 261 for propiconazole.

Quantitation for both CE and GC-MS was by comparison of stereoisomer peak areas to those of propiconazole standards of similar concentration analyzed the same day; quality control included analysis of at least one standard per day as well as analysis of a standard before and after each 10 samples. These standards were in turn referenced to a standard curve of propiconazole stereoisomer peak area vs. concentration. For example, a standard curve was generated by CE using a series of 3, 5, 8, 10, 50 and 100 mg/L standards of propiconazole in water plotted against CE peak area to give a linear plot with R² of 0.9922. Reproducibility of propiconazole quantitation by CE was measured using a 5 mg/L standard solution in water, with n = 6. The percent relative standard deviation for peak 1 stereoisomer was 5.6%, and for peak 2, 7.3%. The RSD for the 2 minor peaks were higher because of their low concentrations (about 0.5 mg/L each): peak 3 was 31.9% and peak 4 was 36.1%. The RSD for total propiconazole was 18.9%.

Propiconazole consists of four stereoisomers (Figure 1) [1,14,15]. The separate stereoisomers were not available to us, so we were not able to assign absolute configurations or signs of optical rotation to the individual peaks in the electrophorogram [Figure 2(A)]. It is assumed however, from other work [11], that the two sets of peaks with identical enantiomer concentrations within each set indicate the two diastereomers of propiconazole (actually, this term is used loosely—as shown in Figure 1, the pairs of diastereomers are divided between the two sets of enantiomers). It is also assumed in this work that diastereomer A is the larger of the two, since one report of the synthesis of propiconazole states that diastereomer A constitutes 60% of the product while diastereomer B is 40% [22]. The proportions of our standard propiconazole are different, however; diastereomer A is 81% while B is 19% [Figure 2(A)]. Regardless, it is assumed here that the differences in proportions of diastereomer A vs. B results from the synthesis process and not from any differences in UV absorbance. This is verified by the fact that the ratios of diastereomers as measured by CE with UV detection are very similar to those measured by GC-MS (see DF a/b, Table 1).

Figure 2(A) is the CE electropherogram of a sample prepared from the aqueous phase of the soil 1 soil-water slurry after exposure of propiconazole to the slurry for 2 hours. The slurry had been spiked at t₀ with 50 mg/L of standard propiconazole. This standard gave an identical electropherogram pattern to that of Figure 2(A). The concentration of total propiconazole at 2 hours in the water phase of soil 1 was 9.8 mg/L, as measured by CE. The 2 hour samples were the t₀ samples for subsequent kinetic runs; it would have been difficult to obtain and process samples for analysis in a shorter time. Apparently the majority of the fungicide had sorbed to the soil by 2 hours (see section 2.5). The kinetic data in Figure 3 show that the t_1/2 of propiconazole loss in both soils is on the order of 50 days under the conditions used here, so only a small amount would have been lost by chemical reaction in 2 hours. Figure 2(B) is the electropherogram of an aqueous-phase sample after exposure of the soil 1 slurry for 96 days; the concentration of total propiconazole is 3.8 mg/L.

The reproducibility of CE migration times is sometimes problematic. Although stable electropherograms and baseline separation of the enantiomers are seen in both Figure 2(A) and 2(B), it is also seen that the migration time of propiconazole in 2(B) relative to 2(A) has decreased from about 20.6 min to 19.1 min for diastereomer A and a corresponding amount for diastereomer B. Slight changes in buffer composition from day to day or in capillary column surface character, which can happen as the column ages, affects migration times by changing the degree of protonation of the silica column surface. Such changes occurred even with propiconazole standard analytes, and sometimes even during the same day. This is not critical for experiments such as described here; since the stereochemical pattern of the propiconazole is apparent (and would be even after some stereoselective biotransformation) the analyte is easily identified, even with some migration time change.

GC-MS was used occasionally as an adjunct or confirmatory analytical tool for chiral analysis of propiconazole. GC-MS chromatograms of a propiconazole standard, injected concentration of 5 μg/mL, were generated using a BGB-172 chiral column. The two diastereomers were separated by 4.7 min; enantiomer separation averaged 0.21 min (13% separation) for the first eluting pair and 1.16 min (baseline separation) for the second pair. Similar diastereomer concentrations were obtained by GC-MS as by CE: A constitutes 80%, while B is 20% of the propiconazole mixture (Table 1). The enantiomer peaks for each of the diastereomeric pairs were nominally of equal area, indicating the expected racemic mixtures; the EF (enantiomer fraction) of the first set of enantiomers was 0.48, while that of the second set was 0.50. The lack of a racemic value (0.50) for the first pair of enantiomers is because of poor chromatographic separation of the two peaks; the average length of peak separations at peak apexes relative to total peak heights was only 13%. No other columns were tried for better stereoisomer separation—there possibly are better columns for this purpose than the BGB-172. Contrary to these results, excellent separation of all four stereoisomers of triadimenol, a triazole fungicide similar in structure to propiconazole, were obtained on the same BGB-172 column in earlier work of the authors [11]. The most obvious reason for this discrepancy is that triadimenol and propiconazole have enough differences in structure—e.g., the former compound has a secondary OH and a tert-butyl group connected to a carbon chiral center at the same relative position where propiconazole has a substituted dioxolane ring with two carbon chiral centers—to affect their interactions with the tert-butyldimethylsilylated-β-cyclodextrin chiral phase of the BGB 172 column.

The main objective of this research was to demonstrate the capability of CE to separate and quantify stereoisomers of conazole fungicides and similar chiral compounds in environmental water and soil samples. Therefore, an application to spiked environmental matrices was attempted.

To measure loss of propiconazole from soil slurries, each of two soils was shaken in an aerobic slurry of 1:5 soil-water for about five months; the water had been spiked with 50 μg/mL of propiconazole. Although several products are reported to be formed by biotransformation or abiotic degradation of propiconazole in soils [1], no reaction products were detected in any of the soil or water samples analyzed here. However, our CE method relied only on UV detection and the particular CE conditions used for propiconazole stereoisomer separation and would possibly not have detected any products. The GC-MS method used the SIM mode of detection, which would not have detected products unless they contained the specific SIM ions used for propiconazole. Therefore, there could have been unobserved products of either an abiotic or biotic reaction.

Figure 3 shows plots for the first order rate equations for loss of propiconazole from the water phases of soils 1 and 2. Each time point represents only one CE analysis. The plots begin with day 5 data, assuming that most sorption had occurred by that time. It is seen that the rates of loss are very similar; k values are 0.0136 and 0.0154 d⁻¹ for soils 1 and 2 respectively, with half lives of 51 and 45 days. This compares well with the DT₅₀ ranges of 40–70 days [1] and 30–112 days [6] obtained by earlier investigators for loss of propiconazole under aerobic conditions. The close k values are not surprising because the properties of the two soils are quite similar (see section 2.2); the pH values are 6.2 and 6.3, while the percent organic matter is 3.0 and 5.4 and the percent clay is 11 and 10.

Analysis of the controls (data shown in Figure 3), which had been autoclaved to reduce or eliminate microbial activity and resulting biotransformation of propiconazole, showed that very little, if any, propiconazole was lost from the soil 1 and 2 water phases during the experiment. This indicates only a very small amount of microbial activity in the controls. Therefore it is concluded that the disappearance of propiconazole from the active samples was probably caused by aerobic biotransformation. However, it is known that autoclaving can cause changes in the character of soil organic matter, and such changes might influence soil sorbtivity. Because the controls were autoclaved and the active samples were not, there could have been some sorption by the active samples and not by the controls. So only abiotic transformation can be excluded from pathways of propiconazole loss in the soils studied here.

For the 2 hour and 96 day exposure samples of soils 1 [Figure 2(A) and 2(B], respectively] and 2, the enantiomer fractions (EF A and EF B) of each pair of propiconazole enantiomers and the diastereomer fractions (DF) of diastereomers A and B are shown in Table 1.

The EF and DF are calculated according to the following equations, where brackets indicate concentration: EF = [ 1 st   eluting   enantiomer ] / [ 1 st   eluting   enantiomer + 2 nd   eluting   enantiomer ] DF = [ 1 st   eluting   diastereomer ] / [ 1 st   eluting   diastereomer + 2 nd   eluting   diastereomer ]

It is seen from the table that the EF of both pairs of enantiomers of soil 1 changed very little during these 96 days, but the DF tended to decrease (statistics could not be calculated because there were not replicate data points). For soil 2, the EF of both pairs of enantiomers tended to decrease, as well as the DF. These trends indicate a low level of stereoselectivity between both pairs of enantiomers and the diastereomers during the biotransformation of propiconazole in soil 2, and a small degree of diastereomer selectivity for soil 1. Contrary to these low levels, the enantiomer fraction of the conazole fungicide triadimefon was shown to change from about 0.45 to 0.60 in its biotransformation to triadimenol, indicating a higher degree of stereoselectivity [11].

The reproducibility of EF measurements was tested by multiple runs of a 10 mg/L standard of propiconazole. For CE analysis of enantiomers 1 and 2, the EF (EF A in Table 1) was 0.50 ± 0.03 (n = 5), while the EF for enantiomers 3 and 4 (EF B) was 0.49 ± 0.03 (n = 5). EF and DF values for GC-MS analysis of propiconazole stereoisomers are also provided in Table 1 for comparison with the CE data, and are seen to be similar.

As evidenced by Figure 2, CE is capable of excellent enantiomer separation with minimum background noise and reasonable sensitivity. The level of quantitation for propiconazole under the conditions used here is about 0.3 mg/L for each of the minor stereoisomers in the injected sample and about 2.5 mg/L for total propiconazole, which includes quantitation of all four stereoisomers. One distinct advantage of CE for analysis of environmental water and soil samples, in addition to high resolution and the small volume of sample needed, is the minimum sample preparation required. For the type of sample involved here, the water sample can be injected directly after filtration through a syringe filter. To increase sensitivity in the water phase, one needs only to perform a simple liquid-liquid extraction of a larger amount of water sample, evaporate the organic extraction solvent to a small volume, say 250 μL, and dilute this to 1 mL with water (CE can accommodate such solutions of water and soluble extraction solvents). For solid phase samples such as soil or sediment, conventional extractions with methanol or acetonitrile, for example, can be adapted for CE (Figure 4). The organic extract is simply diluted with water and filtered.

Thus, CE, especially in the MEKC mode, is a useful analytical tool for measuring the kinetics of disappearance/biotransformation of stereoisomers of chiral pesticides and other pollutants from soil, sediment or aquatic microcosms spiked with 10 to 50 mg/L or so of the pollutant (this much may be necessary to follow the concentration as it decreases to low mg/L levels) [18,19], and could be used to simply detect propiconazole or similar pollutants in water or soil at about 1–2 mg/L. In fact, recently developed on-line preconcentration methods coupled with MEKC may allow increases in sensitivity for pesticide detection by up to 100-fold [20].

In addition to its inherent low sensitivity, which can be improved with equipment and technique changes, a disadvantage of CE for trace pollutant analysis is, in some cases, low migration time reproducibility. This can confuse electropherogram interpretation and affect peak identity. As discussed above, changes in buffer/electrolyte composition and/or CE column surface protonation can change migration times on a day-by-day basis. However, an intentional change of the fused silica column may create large migration time changes. The new column used for soil phase analysis (Figure 4) resulted in reduction of time of the propiconazole diastereomer A from 20.6 [Figure 2(A)] to 11.9 minutes (Figure 4). The column was likely not well conditioned, and the change from an aqueous (water phase) to a soil phase extract probably introduced critical salt content and other changes in the injected sample.

Kinetic data developed in this study using CE showed that soils 1 and 2, which have similar physiochemical properties, support reaction with propiconazole in soil-water slurries to reduce its concentration in the overlying water phase by about two-thirds over a period of three months (Figure 2). The very low loss of propiconazole from the water phase of the autoclaved control samples argues against loss from the active samples by abiotic processes; aerobic biotransformation, however, is a possible and perhaps the likely process for depletion of propiconazole in the environmental systems studied here. The low level of stereoselectivity observed during biotransformation is unusual but, as with the case of imazaquin in aerobic soil slurries [19] and metalaxyl in aerobic soils at pH 4–5 [23], is not unknown.

There remains the possibility of propiconazole loss by a long-term and log-linear sorption process, and the lack of pronounced stereoselectivity supports this because it is believed that sorption of chiral organics to soils and sediments is not stereoselective except in the cases of pure mineral sorbents or definite mineral surfaces [24]. However, diastereomers, as opposed to enantiomers, differ in their abiotic chemical and physical properties. Thus, the observation of slight decreases of DF values in both soils at 96 days exposure (Table 1) could be caused by diastereomer-selective sorption, as was observed by other workers with the higher sorption affinity of 17β-estradiol over that of 17α-estradiol [25].

As with other chiral pesticides, the stereoisomers of propiconazole are independent entities in respect to their biological properties. Stereoisomers may differ in their toxicities to a variety of species, and may be transformed by microbes at different rates, even with different products. These observations should be brought to bear when the exposure and toxicity of chiral pesticides are being measured or considered in environmental matrices [26]. The effective exposure and toxicity of a chiral pesticide depends upon its stereochemical composition at the point of exposure; furthermore, because of stereoselective biotransformation, this composition may well differ from that of standards to which comparison may be attempted [11].