Anal. Chem. 1997, 69, 3314-3320
Determination of the Metabolites of the Herbicide Dimethyl Tetrachloroterephthalate in Drinking Water by High-Performance Liquid Chromatography with Gas Chromatography/Mass Spectrometry Confirmation Robin A. Carpenter, Richard H. Hollowell,* and Kenneth M. Hill
Division of Medical-Legal Investigation and Forensic Sciences, Suffolk County Department of Health Services, Veterans Memorial Highway, Hauppauge, New York 11788
Two metabolites of the herbicide dimethyl tetrachloroterephthalate, (DCPA), monomethyl tetrachloroterephthalate (MM) and tetrachloroterephthalic acid (TCPA), are assayed via high-performance liquid chromatography with ion pairing. Samples are analyzed via direct injection, without preparation, and analyte detection is accomplished with an ultraviolet photodiode array detector. The metabolites are extracted from positive samples with a petroleum ether/diethyl ether mixture, derivatized with N,O-bis(trimethylsilyl) trifluoroacetamide, and confirmed by way of gas chromatography/mass spectrometry. The HPLC analysis of 20 spiked drinking water samples yielded a recovery range of 92-106% with a mean recovery of 101% for TCPA and a recovery range of 92101% with a mean recovery of 96% for MM. The minimum detection limits for TCPA and MM were 2.4 and 2.7 µg/L, respectively. In addition, the GC/MS analysis of spiked reagent water yielded mean recoveries of 91% for MM and 86% for TCPA. Twenty drinking water samples were split and analyzed by the HPLC and GC/MS methods and by USEPA Method 515.1. Comparable results were obtained. The HPLC method, which is amenable to automation, typically allows for the analysis of up to 40 samples overnight. Dimethyl tetrachloroterephthalate (DCPA) is an important preemergent herbicide used for crabgrass, green/yellow foxtails, broadleaf weeds, etc. on golf courses and fruit, vegetable, and sod farms. Depending on temperature and soil conditions present, DCPA exhibited a half-life (t1/2) of 45-90 days, and the resulting composition of the degradation products was 1% monomethyl tetrachloroterephthalate (MM) and 99% tetrachloroterephthalic acid (TCPA).1 The half-life value of MM in soil is reported to be 2.8 days, whereas TCPA showed virtually no loss at room temperature or at 38 °C for a period of over 290 days.2 The discovery of the breakdown products of DCPA, also known as Dacthal,3 in the groundwater aquifers that supply drinking water (1) Kileen, J.; Eilich, G., Diamond Shamrock, personal communication with Dr. L. Dal Cortivo, 18 March, 1982. (2) Wettasinghe, A.; Tinsley, I. J. Bull. Environ. Contam. Toxicol. 1993, 50, 226-231. (3) Sine, C., Ed. Farm Chemicals Handbook, 75th anniversary ed.; Meister Publishing Co.: Willoughby, OH, 1989.
3314 Analytical Chemistry, Vol. 69, No. 16, August 15, 1997
for the residents of local communities has created a need for a rapid and sensitive method of analysis. According to the U.S. Environmental Protection Agency (USEPA) National Pesticide Survey (NPS), TCPA was the most commonly found herbicide degradation product in the groundwater samples collected from across the United States. The highest value for TCPA in that report was 7.2 µg/L.4 In addition, the New York State Department of Health has established a drinking water standard of 50 µg/L for unregulated organic compounds (UOCs) such as DCPA, MM, and TCPA.5 USEPA Method 515.1, which employs gas chromatography with electron capture detection, is not amenable for screening large numbers of samples for the presence of the two metabolites. The sample preparation steps for this method are very labor intensive, requiring manual extractions with large volumes of organic solvents, evaporation/concentration, and a derivatization step. The derivatizing agent used, diazomethane, is carcinogenic and potentially explosive.6 As a result of the initial hydrolysis step at pH ) 12 for 1 h, DCPA and MM are hydrolyzed to TCPA. More importantly, methylation, via diazomethane, converts the two degradation products, TCPA and MM, back to the parent compound, DCPA. This precludes the determination of the individual degradation products in the original sample.7 Distinguishing the concentration of each product is significant for the correct assessment of the public health significance of the contamination. We describe a reverse phase high-performance liquid chromatography (HPLC) analysis for the two metabolites of DCPA that incorporates ion pairing with ultraviolet (UV) detection. Ion pair chromatography is employed in order to add a second ion, of opposite charge, to the eluent which will combine with the ionic TCPA molecule creating a neutral ion pair, which will then partion between the stationary and mobile phases.8 This procedure allows (4) National Pesticide Survey; Summary Results of EPA’s National Survey of Pesticides in Drinking Water Wells; United States Environmental Protection Agency, Office of Water, Office of Pesticides and Toxic Substances: Washington, DC, Fall 1990. (5) Maximum Contamination Levels. Public Health Law. Section 225. Public Water System Subpart 5-1, January 6, 1993; pp 1-103. (6) Lewis, R. J., Sr. Carcinogenically Active Chemicals, A Reference Guide; Van Nostrand Reinhold: New York, 1991. (7) U.S. EPA Environmental Monitoring and Support Laboratory. Method 515.1: Determination of Chlorinated Acids in Water by Gas Chromatography with an Electron Capture Detector; Revision 4.0; USEPA EMSL: Cincinnati, OH, 1989. S0003-2700(96)01212-7 CCC: $14.00
© 1997 American Chemical Society
on-column injection of large volumes of unprocessed well water to provide the needed sensitivity for potable water assays. This analysis is amenable to unattended, overnight analyses of multiple samples, which maximizes productivity and reduces the cost of each sample analyzed. Positive well water samples, calibration standards, and spiked recoveries are extracted, derivatized, and analyzed by gas chromatography/mass spectrometry (GC/MS) for analyte quantitation and mass spectral confirmation. The extraction procedure consists of selectively partitioning the chemical species from water into a small volume of organic solvent. Prior to derivatization, the extraction solvent is removed by means of a nitrogen blowdown technique. The derivatizing agent, N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA), reacts with the carboxyl group(s) of the degradation products, forming a trimethylsilyl (TMS) ether(s). The difference in chemical structure between the derivatized TMS ethers of MM and TCPA permits complete separation of the two compounds in the GC/MS analysis. In addition, and important for confirmation, the mass spectrum of each derivatized compound is different, as can be seen in Figures 4, 6, and 7. Furthermore, the derivatizing agent is less dangerous to use than diazomethane. EXPERIMENTAL SECTION Apparatus. The high-performance liquid chromatography system consists of the following: A Perkin-Elmer Series 400 solvent delivery system was interfaced with a Waters Associates 712 WISP autosampler. The standard 0.25 mL syringe on the sampler was replaced with one of 1.0 mL volume (Waters, No. 76457). The sample loop volume was 2.0 mL. A Supelco LCPAH-modified C18 analytical column, 15.0 cm × 4.6 mm × 5 µm (Catalog No. 5-8318), was used in conjunction with an Upchurch (Catalog No. C-130B) 2 mm i.d. × 2 cm guard column. The guard column was packed with 30-40 µm pellicular Perisorb RP8 from Upchurch Scientific (Part No. C602) and 2 µm frits (Part No. A100X). An Upchurch precolumn filter (Part No. A-318) with a 2 µm replaceable frit (Part No. A-101x) was also installed in front of the guard column. Column temperature was maintained at 34 °C by employing a Beckman Instruments oven (Part No. 235781). The detector was a Perkin-Elmer LC-480 autoscan diode array detector equipped with an 18 µL flow cell. Detector signal monitoring and data reduction were accomplished with a PE Nelson 900 series interface and a Perkin-Elmer Turbochrom 3.1 software package installed on a 486/33 personal computer system with 8 MB of RAM. Chromatographic conditions appear in Table 1. The GC/MS system used in the confirmation analyses consisted of a Hewlett-Packard (HP) 5890 Series II gas chromatograph, a HP 5971 Series mass selective detector, a HP 7673 autosampler injector interfaced with a HP chemstation (DOS Series), and a HP Laser Jet IIIp printer. The GC column was a HP-1, cross-linked methyl silicone, 12 m × 0.2 mm × 0.33 µm thickness, from Hewlett-Packard (Catalog No. 19091-60312). The autosampler used a 10 µL syringe from Hamilton Co. (75ASN 23S GA 1.7 in., Part No. m87987). The splitless injection port liner was a deactivated quartz glass open tube from Hewlett Packard (Part No. 5181-8818). Tetrahydrofuran was substituted for hexane in the solvent wash vials used for rinsing the syringe in the (8) Yost, R. W.; Ettre, L. S.; Conlon, R. D. Practical Liquid ChromatographysAn Introduction; The Perkin Elmer Corp.: Norwalk, CT, 1980; pp 116-117.
Table 1. HPLC Chromatographic Conditionsa
step
duration (min)
flow (mL/min)
0 1 2 3
0.1 8.0 5.0 10.0
0.1 1.0 1.0 1.0
amount of reagent (%) A B C 20 40 20 20
70 50 70 70
10 10 10 10
linear linear
a Reagent A, 15 mL of 0.5 M TBAP/1 L of CH CN; reagent B, 15 3 mL of 0.5 M TBAP + 50 mL of CH3CN/1 L of H2O; reagent C, 15 mL of 0.5 M TBAP + 25 mL of CH3CN/1 L of 0.01 M K2HPO4‚3H2O (TBAP ) tetrabutylammonium phosphate). Retention times: TCPA, 6.94 min; MM, 12.50 min. Wavelength of PDA scans, 190-430 nm. Quantitation wavelength, 238 nm.
Table 2. GC/MS Chromatographic Conditions injector temp injection volume injection mode He flow rate analytical column initial oven temp ramp rate 1 ramp rate 2 final temp MS scan range source retention times DCPA heptachlor epoxide MM TCPA
220 °C 3 µL splitless, 1 min 1 mL/min capillary, HP-1, 12 m, 0.20 mm i.d. 100 °C, 0.5 min 20 °C/min, to 260 °C 40 °C/min, to 300 °C 300 °C, 2 min 300-450 amu electron ionization, 70 eV 5.79 min 6.02 min 6.69 min 7.50 min
autosampler. This solvent significantly reduces the possibility of the autosampler plunger binding. GC/MS chromatographic conditions are listed in Table 2. Reagents. HPLC grade water used throughout was obtained by processing all glass-distilled water with a Barnstead Nanopure system (Model D4754). The hexane, suitable for HPLC and GC analysis (Baxter Diagnostics), was dried using molecular sieves, size 5 Å (Alltech Associates). HPLC grade acetonitrile (Burdick and Jackson Laboratories); Baker analyzed sulfuric acid (Fisher Scientific); Baker Resi-Analyzed anhydrous diethyl ether and petroleum ether (VWR Scientific); analytical grade potassium phosphate dibasic (Mallinckrodt); commercially prepared HPLC grade, 0.5 M, pH 7.5 ionate solution (tetrabutylammonium phosphate, Pierce); neat N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA, Pierce); certified ACS grade spectranalyzed isooctane (Fisher); and spectrophotometric grade tetrahydrofuran and reagent grade hydrochloric acid, 37% (Aldrich), were used. Analytical standards were purchased from Chem Service Inc. or obtained from SDS Biotech Inc. Preparation of Standards. Approximately 10 mg, accurately weighed, of TCPA or MM was dissolved in 10 mL of methanol to provide each primary stock standard solution. When stored at -10 °C, these solutions are stable for at least 6 months. A secondary stock solution, 25 µg/mL of each compound, was prepared by adding 250 µL of each stock standard solution into a 10 mL volumetric flask and diluting to volume with methanol. The secondary stock solution was prepared monthly and stored at -10 °C. Calibration standards were prepared weekly via appropiate dilutions of the primary or secondary stock solutions with HPLC grade water to provide standards of approximately 12.5, 25, 50, 125, and 200 µg/L concentrations of each compound. Quality Analytical Chemistry, Vol. 69, No. 16, August 15, 1997
3315
Table 3. Accuracy and Precision for DCPA Metabolites by HPLC
compound TCPA TCPA TCPA TCPA MM MM MM MM
true concn (µg/L) 10a 50b 25c 125c 10a 50b 25c 125c
tR (%RSD) (min)
mean (µg/L)
recovery (%)
6.23 (0.21) 6.15 (0.32)
9.3 50.8 23.5 128.0 10.3 52.3 24.1 124.0
92.6 102.0 93.8 102.0 103.0 104.0 96.4 99.3
12.25 (0.06) 12.20 (0.06)
%RSD
average deviation (µg/L)
standard deviation (µg/L)
bias (µg/L)
9.9 2.9 7.4 2.5 10.1 6.4 5.7 4.0
0.8 1.2 1.3 2.5 0.8 2.8 1.1 4.0
0.9 1.5 1.7 3.2 1.0 3.3 1.4 5.0
-0.7 0.8 -1.6 2.5 0.3 2.3 -0.9 -0.9
a Fifteen consecutive determinations. b Fifteen consecutive determinations, 1 week later. c Weekly prepared standards analyzed over a 7 month period, for a total of 23 determinations.
Figure 1. HPLC chromatogram of (a) a mixed standard containing (1) tetrachloroterephthalic acid (55.5 µg/L) eluting at 6.94 min and (2) monomethyl tetrachloroterephthalate (52.0 µg/L) eluting at 12.5 min; (b) a drinking water sample with (1) tetrachloroterephthalic acid (158 µg/L) eluting at 6.91 min; and (c) reagent water blank.
control standards were prepared from primary stock standard solutions from a secondary source. Derivatized calibration standards for the GC/MS analysis were prepared in glassware that was rinsed with 1 N HCl, heated at 105 °C for 1 h, cooled to room temperature, rinsed with diethyl ether/petroleum ether (50:50), and allowed to evaporate to dryness prior to use. The HCl reacts with alkaline substances on the surface of the glassware, minimizing losses of MM or TCPA. Calibration standards of 20, 40, 60, 80, and 100 µg/L were prepared by the addition of the appropiate amounts of primary stock solutions into centrifuge tubes containing 15 mL of the diethyl ether/petroleum ether mixture. The calibration standards were then derivatized in the same manner as the sample extracts. The primary stock solution was used in order to minimize the amount of methanol injected, as the methanol also reacts with the derivatizing agent to form a trimethylsilyl (TMS) ether. Heptachlor epoxide was chosen as the internal standard because it has never been detected in any of the over 20 000 drinking water samples from Suffolk County that were analyzed via USEPA Method 505 at a 0.2 µg/L detection limit. Ap3316 Analytical Chemistry, Vol. 69, No. 16, August 15, 1997
Figure 2. Background-corrected UV spectra of (a) tetrachloroterephthalic acid (55.5 µg/L) standard eluting at 6.94 min and (b) a drinking water sample that is contaminated with tetrachloroterephthalic acid (158 µg/L) eluting at 6.91 min.
proximately 10 mg of the pure standard was weighed out into a 10 mL volumetric flask and diluted to volume with hexane. One milliliter of this internal standard stock was transferred to a 100 mL volumetric flask and diluted to volume with hexane. The resulting concentration was 10 ng/µL heptachlor epoxide. HPLC/UV Sample Preparation and Analysis. Samples were collected in 4 oz, high-density polyethylene bottles. Unless analyzed immediately, samples were stored in a -10 °C freezer. Fresh or completely thawed samples were transferred into 4 mL disposable autosampler vials (Sun Brokers, No. 200510 or 2006381562A) for loading into the HPLC autosampler. A 500 µL sample volume was injected sequentially every 23 min. Prior to each automated HPLC run, a 50 µg/L calibration working standard was injected, and the areas observed for each peak in the run were then used by the computer to generate the day’s response factor, based on the ultraviolet (UV) absorbance at 238 nm. Although the UV absorbance of both MM and TCPA is greater at 214 nm, 238 nm was selected to minimize the background rise of the chromatogram and still obtain adequate sensitivity. The stability and linearity of the calibration curve were evaluated by analyzing a working standard after every fifth sample.
Table 4. Results of Stability Studies by HPLC Analysis no. of compound samples
sample type groundwater groundwater spiked reagent spiked reagent
TCPA TCPA TCPA MM
6 37 19 10
holding time at -10 °C
mean recovery (%)
80-90 days 80 days-17 months 14 weeks 14 weeks
98.3 96.9 97.2 96.6
Table 5. Accuracy, Precision, and Recoveries for GC/MS Analysis true average standard compound concn recovery deviation deviation bias MDLa (n ) 10) (µg/L) (%) (µg/L) (µg/L) (µg/L) (µg/L) DCPA MM TCPA a
86.4 63 70.2
95 91 86
2.6 5.3 6.9
3.2 5.9 8.4
-4.3 -5.7 -9.7
2.5 4.6 4.8
Eleven consecutive injections of our lowest standard (∼20 µg/L).
In addition, a 150 µg/L QC sample, prepared from a different primary stock solution, was also analyzed during each overnight run. As part of the quantitative analysis, the UV spectra of positive samples and standards were compared via a computer generated overlay. GC/MS Sample Preparation and Analysis. A 50 mL aliquot of a well water sample was measured out and transferred into a 125 mL separatory funnel. This separatory funnel had a Teflon stopcock and was rinsed with 50:50 diethyl ether/petroleum ether mixture prior to use. It is important that the pH of the sample be
