Anal. Chem. 2000, 72, 3665-3670
Pressurized Fluid Extraction of Nonpolar Pesticides and Polar Herbicides Using In Situ Derivatization Michael D. David,*,† Sonia Campbell, and Qing X. Li*
University of Hawaii, Department of Molecular Biosciences and Biosystems Engineering, 1955 East-West Road, Honolulu, Hawaii 96822
Analysis of polar acidic herbicides has traditionally presented a challenge because of their strong adsorption to and ionic interactions with soil. One approach which has been successful for extraction of these polar compounds from soil is supercritical fluid extraction (SFE) coupled with in situ derivatization. This technique involves the addition of common derivatization reagents directly into the extraction chamber, where the acid herbicides are derivitized to extractable esters or ethers. This study describes the application of an in situ derivatization technique to pressurized fluid extraction (PFE) for the herbicides 2,4-D, 2,4,5-T, dicamba, silvex, trichlopyr, and bentazone. The efficiency of in situ derivatization PFE for these analytes is compared with a conventional basic extraction method followed by ex situ derivatization. The variables of temperature, pressure, static extraction time, and derivatization-reagent amount were optimized for recovery of these analytes from soil. Average recovery for these six analytes was 107% for in situ derivatization PFE from spiked sand, 93% for the same method from a highconcentration spiked soil (50 mg/kg), and 68% for the optimized in situ derivatization PFE method from lowconcentration soil (0.5 mg/kg). The in situ derivatization PFE method has substantial advantages of simplicity of methodology and reduction in extraction time compared with the conventional technique. A second in situ derivatization PFE strategy was investigated using sodium EDTA in the extraction chamber for the extraction of 2,4-D from soil. Preliminary results demonstrate improved recovery with the use of Na4EDTA. Extraction efficiency of PFE for nonpolar organochlorine insecticides and slightly polar triazine herbicides from soil is also presented and compared with that of Soxhlet extraction. Quantitative extraction of soil residues of chlorophenoxy acid herbicides has traditionally presented a challenge because of their strong binding to soil. The conventional method for extracting these compounds from soils uses alkaline aqueous extraction followed by protonation of the analyte and liquid/liquid extraction into an organic solvent. Because of the dissociable proton of the acid, gas chromatographic (GC) analysis is accomplished only 10.1021/ac000164y CCC: $19.00 Published on Web 07/07/2000
© 2000 American Chemical Society
after derivatization by one of several methods.1,2 Combining the aqueous extraction and derivatization methods for extraction and analysis of acidic herbicides in soil is time-consuming and laborious. High-pressure extraction techniques such as supercritical fluid extraction (SFE), typically using CO2,3,4 offer an alternative to traditional solvent extraction methods such as Soxhlet extraction with advantages of significantly reduced extraction time and solvent volume. Unmodified CO2 is effective in extracting nonpolar compounds, and moderately polar compounds can be extracted with the addition of polarity modifiers to the CO2.5-9 Application of SFE to very polar analytes such as those with a dissociable proton (e.g., acids) has been difficult. One strategy for improving recovery of polar acidic herbicides from environmental residues using SFE is in situ derivatization.10,11 The in situ derivatization strategy includes derivatization of polar compounds (such as phenoxy acetic acids or phenols) to less polar esters or ethers in the extraction chamber. The analytes are then collected in a solvent and analyzed by GC. Reactions applied to this method include methylation/alkylation,11-16 silylation,17,18 and pentaflurobenzyl derivatization.13,19 These methods have the advantages of SFE combined with the advantage of ease † American Cyanamid Company, P.O. Box 400, Princeton, NJ, 08543. (1) EPA method 8151A., revision 1; USEPA: Washington, DC, December, 1996. (2) Lee, H. B.; Stokker, Y. D.; Chau, A. S. Y. J.sAssoc. Off. Anal. Chem. 1986, 69, 557-560. (3) EPA method 3560, revision 1; USEPA: Washington, DC, December, 1996. (4) EPA method 3561, revision 1; USEPA: Washington, DC, December, 1996. (5) Fahmy, T. M.; Paulaitis, M. E.; Johnson, D. M.; McNally, M. E. P. Anal. Chem. 1993, 65, 1462-1469. (6) Kahn, S. U. J. Agric. Food Chem. 1995, 43, 1718-1723. (7) Locke, M. A. J. Agric. Food Chem. 1993, 41, 1081-1084. (8) Lee, H. B.; Peart, T. E.; Hong-You, R. L.; Gere, D. R.J. Chrom. A 1993, 653, 83-91. (9) Benner, B. A. Anal. Chem. 1998, 70, 4594-4601. (10) Field, J. A. J. Chromatogr., A 1997, 785, 239-249. (11) Hawthorne, S. B.; Miller, D. J.; Nivens, D. E.; White, D. C. Anal. Chem. 1992, 64, 405-412. (12) King, J. W.; France, J. E.; Snyder, J. M. Fresenius’ J. Anal. Chem. 1992, 344, 474-478. (13) Lopez-Avila, V.; Dodhiwala, N. S.; Beckert, W. F. J. Agric. Food Chem. 1993, 41, 2038-2044. (14) Lee, H.; Peart, T. E.; Hong-You, R. L. J. Chromatogr. 1992, 605, 109-113. (15) Croft, M. Y.; Murby, E. J.; Wells, R. J. Anal. Chem. 1994, 66, 4459-4465. (16) Alzaga, Y. R.; Bayona, J. M. Anal. Chem. 1994, 66, 1161-1167. (17) Hills, J. W.; Hill, H. H. Anal. Chem. 1991, 63, 2152-2155. (18) Hills, J. W.; Hill, H. H. J. Chromatogr. Sci. 1993, 31, 6-12. (19) Hillmann, R.; Bachmann, K. J. High Resolut. Chromatogr. 1994, 17, 350352.
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000 3665
of GC analysis of derivatized compounds. Methylation reactions have been demonstrated to be more efficient with in situ derivatization.13 Pentaflurobenzyl derivatization, however, has added advantage for the analysis of ultratrace residue levels in the environment because of the enhanced sensitivity to detection by GC equipped with either an electron capture detector (ECD) or negative chemical ionization mass spectrometer (NCI-MS)2. Ion pairing is a more recent strategy for improved SFE efficiency of herbicides. Quantitative recovery of chloroacetic herbicides from soil by adding water and ethylenediamine tetraacetic acid tetrasodium salt (Na4EDTA) has been reported.20,21 Other analytes, including aromatic alcohols and other polar compounds,22-25 have also demonstrated improved recovery by SFE in the presence of ion-pair reagents. Pressurized fluid extraction (PFE), also referred to by the trade name of accelerated solvent extraction (ASE), provides a more recent alternative to SFE for high-pressure extraction.26 PFE has the same advantages of reduced time and solvent requirements as SFE and has a greater flexibility of solvent mixtures.26-28 The variables of temperature and pressure are less important in PFE relative to SFE because maintaining supercritical conditions is not a factor in PFE.28 Quantitative analyte recoveries can also be superior to SFE.28,29 Application of PFE to the extraction of chloroacetic acid herbicides from soil has proved challenging, but the use of acidified extraction solvent has been reported effective for some matrixes.26,30 Application of in situ derivatizations or the use of ion-pair reagents have not been reported for PFE. The primary objective of this work is to investigate the feasibility of in situ pentaflurobenzyl derivatization for the quantitative recoveries of the acidic herbicides 2,4-D, 2,4,5-T, trichlopyr, silvex, dicamba, and bentazone. Experimental variables of amount of derivatization reagents, pressure, temperature, and extraction time have been optimized for recovery and repeatability for the six analytes as a group. Other methods for the improved recovery of chloroacetic herbicides from soil using PFE are also evaluated, including acid-modified extraction fluids, and the use of Na4EDTA. Recoveries of less polar agrochemicals, including organochlorines, triazines, and carbofuran, are also reported and compared with that of Soxhlet extraction in order to validate the PFE procedures. MATERIALS AND METHODS Chemicals. Standards of the validation chemicals included carbofuran, atrazine, simazine, heptachlor, chlordane, aldrin, DDE, (20) Guo, F.; Li, Q. X.; Alcantara-Licudine, J. P. Anal. Chem. 1999, 71, 13091315. (21) Alcantara-Licudine, J. P.; Kawate, M. K.; Li, Q. X. J. Agric. Food Chem. 1997, 45, 766-773. (22) Jimenez-Carmona, M. M.; Tena, M. T.; Luque de Castro, M. D. J. Chromatogr., A. 1995, 95, 269-276. (23) Jimenez-Carmona, M. M.; Manclus, J. J.; Montoya, A.; Luque de Castro, M. D. J. Chromatogr., A. 1997, 97, 329-336. (24) Tena, M. T.; Luque de Castro, M. D.; Valcarel, M. Chromatographia 1995, 40, 197-203. (25) Field, J. A.; Miller, D. J.; Field, T. M.; Hawthorne, S. B.; Giger, W. Anal. Chem. 1992, 64, 3161-3167. (26) Ezzell, J. L.; Richter, B. E.; Felix, W. D.; Black, S. R.; Meikle, J. E. LC-GC 1995, 13, 390-398. (27) Richter, B. E.; Jones, B. A.; Ezzel, J. L.; Porter, N. L. Anal. Chem. 1996, 68, 1033-1039. (28) David, M. D.; Seiber, J. N. Anal. Chem. 1996, 68, 3038-3044. (29) Heemken, O. P.; Theobald, N.; Wenclawiak, B. W. Anal. Chem. 1997, 69, 2171-2180. (30) Nemato, S.; Lehotay, S. J J. Agric. Food Chem. 1998, 46, 2190-2199.
3666
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
DDT (EPA reference standards, 99%), and nonachlor (Chem Service, 99%). Standards of herbicidal acid analytes included 2,4,5-T (Dow, 99%), silvex (AAPCO, 99%), dicamba (Velsico, 99.9%), bentazone, 2,4-D, and triclopyr (EPA reference standards, 99%). Extraction solvents were Optima grade (Fisher Scientific). Extraction and derivatization reagents used included pentaflurobenzyl bromide (PFBBr) (99%, Aldrich), potassium carbonate (K2CO3), and Na4EDTA (Matheson Coleman and Bell, 99%). Soil. Soil used in this study is described in detail elsewhere.20 It is referred to as ‘‘Wahiawa” soil and is considered variably charged with 4.4% organic carbon and 86% clay. Ottawa sand (2030 mesh, Fisher Scientific) was used for spike-recovery studies and as an inert solid matrix for PFE studies. Spiked soil preparations were made by diluting a stock mixture of standards in 100-150 mL of acetone and mixing with 200 g of soil. The solvent was evaporated in a rotary evaporator, and the soil was allowed to dry in a fume hood. Two separate preparations of spiked Wahiawa soil were used for the validation chemicalss one containing carbofuran, atrazine, and simazine and the other with heptachlor, nonachlor, aldrin, chlordane, DDT, and DDE. These two soils spiked with the validation chemicals were aged for 60 days at room temperature (25 ( 2 °C) before extraction. Several batches of soil spiked with the herbicidal acid mixture were prepared as needed, aged for 24 h at room temperature, and thereafter stored at 2-4 °C. Two spiking levels for the herbicides on Wahiawa soil were used: a high-concentration spike at 50 mg/kg, and a lower concentration of 0.5 mg/kg. Ottawa sand was spiked in the same manner with the herbicides at the higher spike levels. PFE Validation. A fully automated Dionex ASE200 (Dionex, Salt Lake City) was used for all extractions. Extraction cells (22 mL), with cell caps (Dionex) were prepared for PFE by securing the lower cap to the bottom of the cell, inserting two cellulose filters (Dionex) which were packed to the bottom of the cell, and adding 2-3 cm3 Ottawa sand as an inert matrix. The spiked soil or sand was then added above the sand, and the cell was filled the rest of the way with additional Ottawa sand and sealed with the top cell cap. This sample-loading procedure applies to validation chemical extractions and those prepared for the in situ derivatization method development. Validation of the PFE methods and apparatus was accomplished by triplicate extractions of the two soils spiked with the validation chemicals described above. Validation chemicals were extracted with two methods, one using acetone and one using methanol, under the same conditions: 1500 psi, 100 °C, 5-min static extraction, 100% flush volume, and 60-s purge time. Conventional Extractions. Soxhlet extraction for comparison of the PFE results of the validation chemicals was accomplished in cellulose thimbles containing approximately 5 g of spiked Wahiawa soil. Triplicate extractions were done using 175 mL of ethyl acetate for 10-12 h with an approximately 10-min reflux cycle time. The conventional method of extracting herbicidal acids for comparison to in situ derivatization PFE was based on a method previously described.11 Soil to be extracted with the conventional method was first subjected to alkaline conditions in order to hydrolyze any acid esters and deprotonate the analytes to partition them into the aqueous phase. This was accomplished by adding
80 mL of 0.5 N KOH in a 10% KCl aqueous solution to 20 g of spiked soil and rotating in a boiling water bath for 30 min. The soil/base suspension was then centrifuged in two 80-mL centrifuge vials. The soil pellet was resuspended, and the basic extract was repeated. The combined aqueous supernatants were washed twice with 50 mL of chloroform and acidified with H2SO4 to pH < 1. The reprotonated acids were recovered from the aqueous phase with two 70-mL portions of chloroform, which were dried over sodium sulfate. The chloroform was reduced and exchanged for acetone prior to derivatization of the acids. Derivatization. The ex situ derivatization procedure is based on the EPA standard method 8151A.1 PFBBr was prepared in acetone at 30% (v/v). Potassium carbonate was prepared at 10% (wt/v) in water. Herbicide standards or extracts were brought to 8-10 mL of acetone in a test tube or reaction flask with closure, to which was added 480 µL of the 30% pentaflurobenzyl bromide solution and 720 µL of the 10% K2CO3. The reaction tube was heated in a water bath to 60 °C for 3 h. Investigations into the kinetics of this reactions demonstrated complete derivatization after 2-2.5 h. After reaction, the acetone solution was reduced under nitrogen to less than 0.5 mL. Hexane (2 × 2 mL) was added to the remainder and reduced to near dryness. The residue was redissolved in a graduated volumetric test tube in 8 mL of a 9:1 toluene/ hexane mixture and brought to 10.0 mL with hexane. One milliliter of this solution was removed to another graduated volumetric test tube and dried under nitrogen. This final residue was redissolved in 4 mL of the 9:1 toluene/hexane and brought to 5.0 mL of hexane. This solution was then analyzed by GC/MS. Herbicide Extraction. Three strategies for PFE of herbicidal acids spiked onto Wahiawa soil were employed: (1) extraction with acidic solvent, followed by ex situ derivatization, (2) in situ derivatization/extraction, and (3) Na4EDTA assisted extraction. An extraction method using acid-modified solvent based on that of Ezzell26 was employed followed by the ex situ derivatization method. The extraction solvent was a 1:2 mixture of methylene chloride/acetone with 4% of a 1:1 solution of phosphoric acid and water. The extraction cell was pressurized to 1500 psi and heated to 100 °C for a static extraction time of 5 min. Extraction solutions were dried over acidified sodium sulfate and were treated with 5 mL of 37% KOH for 90 min to hydrolyze any esters of the acids. These solutions were brought to 10 mL and derivitized by the ex situ derivatization method previously described. An in situ derivatization was developed by optimizing for the variables of reagent amount, pressure, temperature, and extraction time. The lower concentration herbicide-spiked soil and acetone for the PFE extraction solvent was used in all of the in situ derivatization method development experiments. The typical PFE extraction conditions of 1500 psi, 100 °C, and 5 min of static extraction time were used as starting points to investigate the effect of concentration of derivatization reagent. Triplicate PFEs were done using 480 µL of either 3, 10, or 30% PFBBr in acetone added directly to the top of the extraction cell after it was loaded with the spiked soil as described previously. An aliquot of 720 µL of 10% aqueous K2CO3 solution was then added to the top of the cell, and it was sealed with the top cap and loaded onto the extractor.
The effects of pressure and temperature on the in situ derivatization were studied while using the 30% PFBBr solution and 5-min static extraction time. The effect of static extraction time was investigated using 30% PFBBr solution, 1500 psi, and 100 °C and triplicate extractions using 5-, 10-, 15-, and 30-min static extraction times. The optimized conditions of 1500 psi, 100 °C, 30-min static extraction time, and 30% PFBBr solution were then repeated for five additional extractions in order to study the optimum recovery and repeatability of the optimized in situ derivatization method. A preliminary investigation of the effect of Na4EDTA on in situ derivatization PFE was accomplished by the addition of 0.3 g (5% w/w) of Na4EDTA to 6 g of soil spiked with 2,4-D at 0.5 µg/g. Water was then added (1.0 mL, approximately 15% w/w), and the prepared sample was allowed to incubate at room temperature for 1 h. After the sample was loaded into the extraction chamber, the in situ derivatization reagents were added in the same amounts as previously described, and the sample was extracted using a 5-min static extraction time, 1500 psi, and 100 °C. Post-extraction sample preparation and analysis was accomplished by the same methods as the in situ derivatization PFE samples. Analysis. A Hewlett-Packard 5890II GC interfaced with a 5989A MS was employed for all analyses. Two columns were used: a DB-5ms (J&W scientific) GC column with dimensions of 30 m × 0.25 mm i.d. with a 0.25-µm film thickness and a DB-17ms with the same dimensions. An injection port temperature of 250 °C and a GC/MS transfer line temperature of 280 °C were used for all analyses. A recently silanized injection port liner was required for analysis of carbofuran to prevent degradation in the injection port. Analysis of the extracts of the validation chemicals was accomplished in two groups. Atrazine, simazine, and carbofuran were analyzed on the DB-17ms column using a temperature program of 110 °C, initial temperature, for 2 min; an initial ramp of 15°/min to 140 °C; a second ramp of 4 °C to 280 °C; and a final time of 5 min. The organochlorine chemicals were analyzed on the DB-5ms with the same temperature program. For both of these groups, the mass spectrometer was operated in electron impact (EI) mode at 70 eV, with a source temperature of 250 °C. The high-concentration herbicide extracts and the ex-situ derivatization method validation trials were analyzed in EI mode with a temperature profile of 110 °C for 1 min, a ramp of 5 °C/ min to 210 °C, and a final time of 5 min. All other herbicide extracts and derivatizations were analyzed with the same GC temperature profile with negative chemical ionization (NCI) MS. For all NCI analyses, the source temperature was kept at 200 °C and 1% ammonia in methane was used for a reagent gas with an ion source pressure of 1.3-1.5 Torr. Safety. Use of high-pressure extraction equipment may result in leaking of high-temperature organic solvent or vapors, which can pose human health and fire risks. Commercially available extraction equipment, such as the system used in this study, have safety features such as solvent leak detectors to minimize the risk. Similar extraction efficiencies can be achieved with laboratoryassembled devices,28 which can be equipped with safety shields. With either apparatus, valves and solvent flow lines should be checked for leaks regularly. Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3667
extraction and uses about 40 mL of solvent. Soxhlet extraction takes 8 h or more for each extraction and up to 200 mL of solvent. PFE is therefore faster and uses less solvent compared with Soxhlet extraction. These two significant advantages for the extraction of these low- and intermediate-polarity compounds from soil make PFE the preferred method, especially when processing numerous samples. Herbicide Extraction. The conventional basic solvent extraction technique proved efficient in the extraction of all six herbicide analytes from Wahiawa soil. The percent recoveries for these analytes using this technique averaged 102% and ranged from 91 to 114% recovery (Table 2). This technique was significantly more complex and time-consuming than in situ derivatization PFE. The phase separations of the basic water extracts and the solvent washes in the separatory funnel were particularly time-consuming, with each solvent wash taking up to several hours for separation. Total time for each extraction averaged 24-48 h. The primary advantages of this method are the fact that several extractions can be done simultaneously and no specialized laboratory equipment is required. Application of the solvent extraction method of Ezzell26 followed by ex situ PFBBr derivatization was unsuccessful for these spiked soils. We were unable to recover more than 5% of any of the six herbicides tested, failing to replicate the reported 86-98% recoveries for these analytes at similar concentrations from clay. The complex ionic nature of the Wahiawa soil, considered a variablecharge soil,20 may account for these results. In the development of the in situ derivatization PFE method, use of the initial conditions of 1500 psi, 100 °C, 30% PFBBr solution, and 5-min static extraction time provided quantitative recovery of the six herbicidal acids from spiked sand and from soil at high concentrations (Table 2). This method is therefore appropriate for quantitative analysis of these herbicides from an inert matrix such as sand or at higher concentrations from natural soils. Application of this method to the lower concentration spiked
Table 1. Comparison of Acetone and Methanol as PFE Solvent for Nine Validation Compounds acetone analyte
%
heptachlor chlordane nonachlor aldrin DDE DDT simazine atrazine carbofuran avg
recoverya 93 103 102 104 105 95 103 102 94 100
methanol %
RSDb
% recoverya
% RSDb
78 116 105 105 109 90 81 82 81 94
12 5 7 6 6 10 1 1 4 6
16 7 7 6 7 9 1 1 2 6
a Percent recovery compared with mean of triplicate Soxhlet extraction. b Percent relative standard deviation ) (standard deviation/mean) × 100.
Organic solvents and reagents used in these experiments should be handled and used with routine laboratory safety. The derivatization reagent PFBBr is a lacrymator and can pose a health risk and, therefore, should be prepared and used in a fume hood. RESULTS AND DISCUSSION Validation Chemicals. Mean recoveries of triplicate PFEs using acetone and methanol are compared with those of Soxhlet extractions for the nine validation chemicals from soils aged for 60 d (Table 1). Average recoveries for these nine analytes using the two solvents are similarly quantitative, at 100% for acetone and 94% for methanol. Methanol, however, demonstrated more variability, with recoveries ranging from 78 to 116%, compared with from 93 to 105% for acetone. Acetone, therefore, is the favored extraction solvent for these compounds. The quantitative recovery of these pesticides from soil by PFE is accomplished in less than 30 min for sample preparation and
Table 2. Comparison of Percent Recoveries of Spiked Herbicides Using Conventional and In Situ Derivatization Techniques
a
dicamba
2,4-D
triclopyr
silvex
2,4,5-T
bentazone
avg
conventional technique soil
114
91
103
105
95
102
102
in situ derivatization sand, 50 mg/kg
112
96
105
110
109
107
107
in situ derivatization soil, 50 mg/kg 5 min
92
60
98
104
68
138
93
in situ derivatization soil, 0.5 mg/kg 5 min
45
31
49
55
37
74
49
in situ derivatization soil, 0.5 mg/kg 30 min
61
58
77
76
54
78
67
literaturea
83
24
1-3 mg/kg of spike onto soil; 1 mL of 5% PFBBr solution; n ) 113.
3668 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
34
47
Figure 1. The effect of pressure on PFE/in situ derivatization recovery of six herbicides: B ) dicamba; 9 ) 2,4-D; 2 ) triclopyr; 1 ) silvex; ( ) 2,4,5-T; 0 ) bentazone.
Figure 2. The effect of temperature on PFE/in situ derivatization recovery of six herbicides: B ) dicamba; 9 ) 2,4-D; 2 ) triclopyr; 1 ) silvex; ( ) 2,4,5-T; 0 ) bentazone.
soil resulted in a mean recovery of 46% for the six analytes and a range of 31-74% (Table 2). Experimental optimization of the four primary extraction variables was therefore attempted using the low-spiked soil. Figures 1 and 2 show the effect of pressure and temperature, respectively, on recoveries of the six polar herbicides. These experiments were run using 5-min static extraction time and 30% PFBBr solution. The pressure study was conducted at 100 °C. The effect of pressure variations on herbicide recovery over the range of 500 to 2500 psi (Figure 1) was slight, with maximum recovery occurring in the 1000-2000 psi range. The temperature experiment was therefore conducted at 1500 psi. Recoveries for all six analytes were maximized in the temperature range of 100150 °C (Figure 2). Some analytes demonstrated better recovery at 150 °C, but the higher temperature increased the length of the
Figure 3. The effect of static extraction time on PFE/in situ derivatization recovery of six herbicides: B ) dicamba; 9 ) 2,4-D; 2 ) triclopyr; 1 ) silvex; ( ) 2,4,5-T; 0 ) bentazone.
extraction steps because of the greater time required for sample heating. The mean relative standard deviations for recovery of all six analytes were also lower at 100 °C than at 150 °C. Therefore, the initial values of 1500 psi and 100 °C were used for all subsequent extractions. These studies are consistent with a previous study demonstrating the effects of pressure and temperature conditions on recovery with PFE.27,28 The effect of static extraction time on herbicide recovery was studied using the established pressure and temperature conditions and using the 30% solution of PFBBr. Extraction time was optimized, and the effect on recovery is shown in Figure 3. Recoveries were variable, between 5 and 20 min, with no clear trend observed, but 30-min static extraction provided significantly better recovery than shorter times. Figure 4 shows the effect of the concentration of PFBBr solution used in the in situ derivatization. The 30% PFBBr solution used in all previous in situ derivatization extractions represents an excess of the reagent 5000-10 000 times greater than the molar amount required to derive all herbicides present in the soil. Experimental extractions using the reduced PFBBr concentrations of 3 and 10% were conducted in order to minimize the amount of this reagent used. Despite the reagent still being present in large excess, significantly reduced recoveries with lower levels of PFBBr were observed. This effect may be attributable to cross-reaction of the active derivatization reagent with active sites of the soil or quenching. If this is true, it follows that the extraction of larger amounts of soil than the six to seven grams used in the extractions reported here may require higher concentrations of derivatizing reagent in order to allow for complete derivatization of the herbicide analytes. Five additional extractions of the low-concentration spiked soil using the conditions of a 30-min static extraction time, 1500 psi, 100 °C, and 30% PFBBr solution were repeated in order to assess the repeatability of this optimized method. The mean recoveries of the six herbicides using these conditions was 67%, and ranged from 54 to 78% (Table 2). These results compare favorably with the only other reported study of in situ derivatization using PFBBr Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3669
Figure 4. The effect of PFBBr concentration on PFE/in situ derivatization recovery of six herbicides: B ) dicamba; 9 ) 2,4-D; 2 ) triclopyr; 1 ) silvex; ( ) 2,4,5-T; 0 ) bentazone.
for recovery of these herbicides13 (Table 2), which used SFE rather than PFE and 1-3 mg/kg spiking levels. The use of Na4EDTA with in situ derivatization PFE of the 2,4-D spiked soil improved the recovery dramatically. Figure 5 shows the percentage recovery of 2,4-D using in situ derivatization PFE with and without Na4EDTA. Using a 5-min static extraction time, the presence of Na4EDTA increased recovery of 2,4-D from 31 to 94%, which is comparable to the recovery using the conventional method. CONCLUSIONS The developed method using PFBBr for in situ derivatization with PFE has been shown to quantitatively extract residues of chloroacetic herbicides from soil at a 50 mg/kg spike level. At residue levels of 0.5 mg/kg this method will extract greater than 75% of residues of triclopyr, silvex, and bentazone. The recoveries of dicamba, 2,4-D, and 2,4,5-T under the same extraction conditions ranged from 54 to 61%. Preliminary experiments with 2,4-D show that recovery of these compounds at the lower spike levels may benefit from the addition of Na4EDTA to the extraction chamber. Other derivatization reagents have been demonstrated to be more effective for complete in situ derivatization using SFE,10,13
3670 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
Figure 5. Comparison of 2,4-D recovery from soil using a conventional method, optimized in situ derivatization/PFE, and in situ derivatization/PFE with Na4EDTA. Error bars represent one standard deviation (n ) 3).
but the use of PFBBr has greater application for analysis of trace residues of the herbicidal acids. The sensitivity required for detection of these compounds on soil in the part-per-billion range (ng/g) is facilitated by the addition of the pentaflurobenzyl moiety. With the addition of this group, the analytes are quantified in the low picogram range using either an electron-capture detector or quantified and confirmed simultaneously with negative chemical ionization mass spectrometry in selected ion monitoring mode. Analysis using this PFE/in situ derivatization method will vary based on differing analytes and matrixes and, therefore, the method variables may need to be optimized for a specific analyte of interest from varying soil types. These experiments demonstrate that the in situ derivatization strategy can be successfully applied to PFE, resulting in a method much less time-consuming and therefore more applicable to routine analysis than existing conventional techniques. ACKNOWLEDGMENT This work was funded in part by USDA Grant PSW-98-019RJVA, the State of Hawaii Department of Agriculture, the State of Hawaii Department of Health, and the Hawaiian Electric Co., Inc. Received for review February 8, 2000. Accepted April 3, 2000. AC000164Y
