Environ. Sci. Technol. 2001, 35, 3190-3192
High Levels of Monoaromatic Compounds Limit the Use of Solid-Phase Microextraction of Methyl tert-Butyl Ether and tert-Butyl Alcohol LISA BLACK AND DENNIS FINE* ManTech Environmental Research Services Corporation, National Risk Management Research Laboratory, Subsurface Protection and Remediation Division, P.O. Box 1198, Ada, Oklahoma 74821-1198
Recently, two papers reported the use of solid-phase microextraction (SPME) with poly(dimethylsiloxane)(PDMS)/ Carboxen fibers to determine trace levels of methyl tertbutyl ether (MTBE) and tert-butyl alcohol (tBA) in water. Attempts were made to apply this technique to the analysis of water samples containing high levels of benzene, toluene, ethylbenzene, xylenes, and trimethylbenzenes (BTEXsTMBs) as would be expected at leaking underground storage tank sites. It was found that when the sample contained total aromatic compounds above 1 ppm, the response of the internal standards, deuterated MTBE and tBA, dropped by more than 65%. As this decrease in internal standard peak area was unacceptable, a static headspace method was used instead. This headspace method was used successfully to analyze groundwater from 670 monitoring wells at 74 service stations located in the northeast United States. In these monitoring wells, 30% of the samples contained total BTEXsTMBs above 1 ppm. If the SPME method was used to analyze these samples, dilution of more than 200 samples would be required to minimize the adverse matrix effect that high aromatic content had on the internal standard peak area.
Introduction In solid-phase microextraction (SPME) (1), the solid phase is coated on a fused silica fiber that is attached to the end of a wire plunger. This fiber is pushed through a syringe needle and is immersed in the aqueous liquid phase or exposed to the headspace above the liquid. Organic compounds present in the liquid or partitioned into vapor phase are absorbed into or adsorbed onto the solid phase. After a specified sampling time, the compounds are thermally desorbed from the solid phase in the heated injection port of a gas chromatograph. Recently, two papers were published that describe the use of SPME with gas chromatography/ mass spectrometry to determine methyl tert-butyl ether (MTBE) in surface water and MTBE and tert-butyl alcohol (tBA) in groundwater. These papers recommend aqueousphase extraction using poly(dimethylsiloxane)/Carboxen SPME fibers. Achten and Puttmann (2) found increased compound extraction when the manual fiber holder was * Corresponding author phone: (580)436-8669; fax (580)436-8501; e-mail:
[email protected]. 3190
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
cooled with ice and the aqueous sample temperature was maintained at 19 °C. Their method included the addition of sodium chloride at 25%, a 60-min extraction time, and a 1.5-mL sample volume. A method detection limit (MDL) of 10 ppt with a relative standard deviation of 12% was reported for MTBE. Deuterated methyl-d3 tert-butyl ether was used as an internal standard. Cassada et al. (3) did the SPME extraction at room temperature using a 3.8-mL sample volume containing 25% sodium chloride with a 25-min extraction time. They report a MDL of 8 ppt for MTBE and 1.8 ppb for tBA. An internal standard was not incorporated into their method. With MDLs at 8-10 ppt for MTBE and 1.8 ppb for tBA, these SPME methods could provide an advantage in sensitivity when trace analysis is needed. These MDLs are better than those reported by Halden et al. (4) for standard purgeand-trap GC methods using mass spectrometry, photoionization detection, and flame ionization detection with column switching. With a sample prepared in reagent water, they determined MDLs ranging from 0.2 to 1.1 ppb for MTBE and from 3 to 35 ppb for tBA. A significant part of their study dealt with how total petroleum hydrocarbons (TPH) at levels expected at leaking underground storage tank (LUST) sites would affect the performance of each of the standard methods. They found excellent results for EPA Method 8260B, which uses a mass spectrometer to analyze for MTBE and tBA in samples containing TPH up to 50 ppm. With EPA Method 8021B, which uses a photoionization detector, the high TPH matrix contributed gasoline components that coeluted with tBA and caused high false-positive concentrations. This paper provides strong support for purge-andtrap with GC/MS as a method for determining MTBE and tBA in LUST samples. In the two SPME methods, neither of the studies investigated possible matrix effects that could be caused by the presence of petroleum hydrocarbons including aromatic hydrocarbons. Initially, we attempted to utilize SPME with the poly(dimethylsiloxane)/Carboxen SPME fiber using similar parameters as reported, but significant matrix effects were found when high levels of aromatic compounds were present. The response of deuterated MTBE and tBA was examined as total concentration of aromatic compounds increased. These deuterated compounds were used as internal standards for the quantitation of MTBE and tBA. These experiments were repeated using a static headspace method.
Experimental Section Materials and Methods. An internal standard stock mixture containing deuterated methyl-d3 tert-butyl ether (MTBE-d3) (CDN Isotopes) and tert-butyl-d9 alcohol (tBA-d9) (CDN Isotopes) was prepared from neat compound at 50 ppm in purge-and-trap grade methanol (J. T. Baxter). A 10 000 ppm mix of MTBE; benzene; toluene; ethylbenzene; p-, m-, and o-xylene; and 1,3,5-, 1,2,4-, and 1,2,3-trimethylbenzene in methanol (Restek Corp.) was used for preparing standards. Solutions of MTBE and tBA (Aldrich Chem.) in methanol were prepared from neat materials. Standards and samples were prepared in 10 mL of water using 16-mL SPME sample vials or 22-mL headspace vials each containing 3.6 g of sodium chloride (ACS certified, Aldrich Chemical) and 10 µL of 50 ppm of MTBE-d3 and tBA-d9. Three SPME fibers (Supelco) were studied: 65 µm poly(dimethylsiloxane) (PDMS)/divinylbenzene (DVB), 75 µm PDMS/Carboxen, and 85 µm PDMS/Carboxen StableFlex. While keeping the internal standards at 50 ppb, the fuel components (BTEXsTMBs) were 10.1021/es010539c CCC: $20.00
2001 American Chemical Society Published on Web 06/22/2001
TABLE 1. Extraction and Desorption Parameters for SPME Analysisa
fiber
extraction phase
65 µm PDMS/DVB 85 µm PDMS/Carboxen 65 µm PDMS/DVB 75 µm PDMS/Carboxen
liquid liquid headspace headspace
extraction desorb desorb time time temp (min) (min) (°C) 25 25 1 5
2 2 3 2
240 250 200 320
a Parameters used for headspace SPME were obtained by optimization. Parameters used for liquid SPME were taken from refs 2 and 3.
increased at 10× increments ranging from 20 to 2000 ppb. This corresponds to a total aromatic content of BTEXsTMBs from 0.18 to 18 ppm. Method blanks containing internal standards were run at the beginning and at the end of the sample queue. Instrumentation and Methods. A Varian 8200 autosampler with SPME option was used for headspace and liquid SPME sampling. A Supelco SPME injector and fiber were placed in the autosampler syringe holder. Extraction times, desorption times, and desorption temperatures for headspace and liquid-phase extractions are provided in Table 1. Each fiber was heated at its recommended conditioning temperature for 1 h at the beginning of each sample queue. The automated fiber agitator was used for all SPME analyses. After sampling, the SPME fiber was desorbed in the split/ splitless injection port of a Hewlett-Packard 5890 series II gas chromatograph. The injection port was equipped with a deactivated 0.75 mm i.d. liner (Supelco). A split flow of 10 mL/min was used for all analyses. During desorption, compounds were cyrotrapped on 9 cm of 0.53 mm i.d. deactivated fused silica precolumn (J & W Scientific). This precolumn was maintained at -125 °C for 3 min by a MicroCyro Trap (Scientific Instruments Serv.) using liquid nitrogen. After 3 min, the MicroCyro trap was heated to 200 °C, and the compounds were eluted onto a 30-m Rtx-Wax Crossbond-PEG (Restek) capillary column with a 0.25 mm i.d. and 0.5-µm film thickness. At the beginning of the SPME injection, the GC oven was held isothermal at 35 °C for 3 min and then temperature programmed to 62 °C at 8 °C/min, held there for 3.5 min, ramped to 135 °C at 8 °C/min, and finally ramped to 250 °C at 40 °C/min. A Tekmar 7000 headspace autosampler and a Tekmar 7050 automated injection module were used for the static headspace sampling. For this analysis, the headspace vial was heated at 80 °C for 25 min. The sample vial was then pressurized with helium at 10 psi for 1 min and equilibrated for 0.25 min. The pressurized headspace filled a 1.0-mL sample loop, and after a loop fill equilibration time of 0.20 min, the sample was transferred to the injection port of the Hewlett-Packard 5890 series II GC mentioned earlier. The flow through the transfer line to the injector was 11 mL/min. The Tekmar sample loop valve and transfer line were maintained at 200 °C. From the injector, the split flow was set at 10 mL/min, and the column flow was calculated to be 1.1 mL/min at 35 °C. An injector/column pressure was maintained at 8 psi by the electronic pressure control of the GC. The septum purge line was plugged. For transfer of headspace sample to the capillary column, a 1 mm o.d. stainless steel tube, 60 mm long with 0.6 mm i.d., was attached to the transfer line using a zero dead volume fitting. This tube was extended 50 mm into the injection port septum. The capillary column was inserted 30 mm into this tube. The GC temperature program was the same as described above. A Finnigan GCQ ion trap mass spectrometer was used for compound detection and quantitation in the SPME and headspace sampling methods. The capillary transfer line and
FIGURE 1. Plot of changes in normalized peak area of 50 ppb MTBEd3 as a function of the total concentration of BTEXsTMBs. Comparisons are made using (O) a 65-µm PDMS/DVB SPME fiber with liquid sampling, (0) an 85-µm PDMS/Carboxen StableFlex SPME fiber with liquid sampling, (4) a 65-µm PDMS/DVB SPME fiber with headspace sampling, (3) an 85-µm PDMS/Carboxen SPME fiber with headspace sampling, and (]) a Tekmar headspace sampler. ion source temperatures were maintained at 225 and 200 °C, respectively. A mass spectrum range of m/z 46-250 was scanned at 0.5 s for all of the analyses. Deuterated MTBE and tBA were quantified from mass chromatograms of ions at m/z 76 and 65, respectively. Both internal standards were present at 50 ppb per 10 mL of prepared standard solution or samples. Aromatic compounds were quantified using the most intense ion for each compound.
Results and Discussion The matrix effect of aromatic compounds on the SPME of MTBE and tBA was examined by determining how the response of MTBE-d3 and tBA-d9, each at 50 ppb, changed as the total aromatic content was increased from 0 to 18 ppm. Three SPME fibers were used for the extractions: 65µm PDMS/DVB, 75-µm PDMS/Carboxen, and 85-µm PDMS/ Carboxen StableFlex. Headspace and liquid sampling were used for the SPME analysis. The static headspace analysis was done using the same concentrations of internal standards and aromatic compounds. In each case, the initial peak area of MTBE-d3 and tBA-d9 in the mass chromatogram of the sample blanks was used as a reference to normalize the responses. This allowed comparison between the SPME fibers, the SPME extraction methods, and the static headspace sampling method. Figures 1 and 2 show that, for each of the SPME fibers, the normalized peak area of the internal standards decreased by more than 65% when the total BTEXsTMBs concentration increased from 0.18 to 1.8 ppm. This occurred with both headspace and aqueous SPME sampling. At 18 ppm, the normalized peak areas decreased to less than 90% of the highest peak area. This decrease in response of polar compounds in the presence of less polar compounds has been characterized by Gorecki et al. (5). They indicated that the composition of the matrix can significantly affect the amount of analyte extracted by the fiber when compounds with higher affinity for the SPME stationary phase replace compounds with lower affinity. Using a PDMS/DVB fiber with aqueous-phase sampling, they showed significant decreases in extraction of methyl isobutyl ketone when benzene was added as the interfering compound and also decreased extraction of 2-propanol as the amount of methyl isobutyl ketone increased. They explained that extraction using the PDMS/DVB fiber is based on adsorption processes VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3191
FIGURE 2. Plot of changes in normalized peak area of 50 ppb tertbutyl alcohol-d9 as a function of the total concentration of BTEXsTMBs. Comparisons are made using (O) a 65-µm PDMS/DVB SPME fiber with liquid sampling, (0) an 85-µm PDMS/Carboxen StableFlex SPME fiber with liquid sampling, (4) a 65-µm PDMS/ DVB SPME fiber with headspace sampling, (3) an 85-µm PDMS/ Carboxen SPME fiber with headspace sampling, and (]) a Tekmar headspace sampler. where nonlinearity exists between the amount of a compound extracted by the fiber and its concentration in the sample. As such, linearity is only obtained when the concentration of the analyte is at low levels and interfering compounds are absent. With static headspace sampling, Figures 1 and 2 show that the normalized peak areas of MTBE-d3 and tBA-d9 did not drop below 80% as the total aromatic content increased. This response indicates that the partial vapor pressures of these compounds in the headspace are unaffected as the total aromatic content increased up to 18 ppm. Although not shown in the figures, this trend continued as the total BTEXsTMBs concentration increased to 45 and 90 ppm. The static headspace method described here was used to analyze groundwater samples from 670 monitoring wells at 74 different service stations located in the northeast United States. Thirty percent of these monitoring wells contained total BTEXsTMBs concentrations above 1 ppm. If the SPME method had been used, a significant matrix effect and unacceptable decrease in internal standard response would be expected in these samples. This decrease in internal standard peak area could be remedied by diluting the sample so that the matrix effect due to the aromatic content is minimized. However, with this dilution, the method sensi-
3192
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
tivity would decrease by the dilution factor applied to the sample. Dilution of a LUST sample with a high aromatic content of 140 ppm would require a dilution factor greater than 150 to bring the total aromatic content below 1 ppm. The sensitivity advantage of the SPME method would be lost with this sample as the MDL would be increased from 8 ppt to ∼1 ppb for MTBE and from 1.8 ppb to ∼250 ppb for tBA. The method detection limits (6) for the described static headspace method were determined in reagent water to be 0.2 ppb for MTBE and 5 ppb for tBA. These MDLs compare well with those reported by Halden et al. (4) for the EPA Method 8260B purge-and-trap GC/MS at 0.2 ppb for MTBE and at 3 ppb for tBA. Both static headspace GC/MS and purgeand-trap GC/MS methods are good alternatives to the SPME method when samples containing high TPH or high aromatic content are suspected. If SPME methods are used, internal standards should be considered a necessity along with monitoring of internal standard peak areas and total aromatic content.
Acknowledgments Dr. John Wilson (NRMRL, U.S. EPA, Ada, OK) is thanked for his encouragement and helpful comments during the preparation of this paper. The authors are grateful to Dr. Bruce Pivetz for his careful review of our manuscript. Although the research in this report has been funded wholly by the U.S. Environmental Protection Agency through EPA Contract 68-C-98-138 (Project Officer Roger Cosby, National Risk Management Research Laboratory, Subsurface Protection and Remediation Division, Ada, OK), it has not been subjected to Agency review and, therefore, does not necessarily reflect the views of the Agency, and no official endorsement should be inferred.
Literature Cited (1) Pawliszyn, J. Solid-Phase Microextraction. Theory and Practice; Wiley-VCH: New York, 1997. (2) Achten, C.; Puttmann, W. Environ. Sci. Technol. 2000, 34, 13591363. (3) Cassada, D.; Zhang, Y.; Snow, D.; Spalding, R. Anal. Chem. 2000, 72, 4654-4658. (4) Halden, R.; Happel, A.; Schoen, S. Environ. Sci. Technol. 2001, 35, 1469-1474. (5) Gorecki, T.; Yu, X.; Pawliszyn, J. Analyst 1999, 124, 643-649. (6) Glaser, J.; Foerst, D.; McKee, G.; Quave, S.; Budde, W. Environ. Sci. Technol. 1981, 15, 1426-1435.
Received for review January 18, 2001. Revised manuscript received May 4, 2001. Accepted May 4, 2001. ES010539C
