Anal. Chem. 2000, 72, 3102-3108
Performance-Enhanced “Tunable” Capillary Microwave-Induced Plasma Mass Spectrometer for Gas Chromatography Detection Angela M. Zapata and Albert Robbat, Jr.*
Department of Chemistry, Center for Field Analytical Studies and Technology, Tufts University, Medford, Massachusetts 02155
Improvements in the stability and performance of a capillary microwave-induced plasma-mass spectrometer (MIP-MS) were achieved by optimizing power transfer to the cavity using a tunable coaxial MIP. The MIP, operating at atmospheric pressure, was sustained with 30 mL/min He and 60 W of power. Measurement precision and sensitivity for the standard waveguide and coaxial systems were determined using 16 organochlorine pesticide solutions separated by gas chromatography (GC). The linear dynamic range obtained with the tunable MIP-MS extended over 3 orders of magnitude, a 10 time improvement with respect to the standard MIP. Detection limits were between 3 and 19 pg of Cl mol-1 s-1, 7 times lower than the detection limits obtained with the nontunable MIP-MS. Analysis of pesticides containing sulfur atoms was also possible, further demonstrating multielement MIP-MS detection. Excellent accuracy (10% recovery) and precision (5% RSD) were found for the detection of the pesticides in a petroleum-contaminated reference soil. By placing the GC column at the plasma expansion stage, molecular fragmentation of a mixture of volatile organic compounds was also demonstrated. With the MS operated in the selected ion monitoring mode, measurement sensitivity was ∼500 pg/s per compound. Plasma mass spectrometry (MS) is an excellent multielement technique capable of providing both elemental and isotopic analysis at ultratrace detection levels. Among the various plasma sources, the He microwave-induced plasma (MIP) provides unique advantages when coupled to gas chromatography (GC) systems.1-3 He plasmas are highly efficient at ionizing high ionization potential atoms, which can then be used as probes in the detection of organic compounds. The low plasma gas flow rates used to sustain the MIP best match the flow rate of high-resolution GC columns, minimizing analyte dilution, while 100% of the GC effluent is transported into the MIP, maximizing detector sensitivity. Both reduced and atmospheric pressure MIPs have been coupled with MS to detect heteroatoms in organotin,4 halogen,5-11 * Corresponding author: (phone) 617-627-3474; (fax) 617-627-3443; (e-mail)
[email protected]. (1) Uden, C. J. Chromatogr., A 1995, 703, 393-416. (2) Long, G. L.; Ducatte, G. R.; Lancaster, E. D. Spectrochim. Acta, Part B 1994, 49B, 75-87. (3) Olson, L. K.; Caruso, J. A. Spectrochim. Acta, Part B 1994, 49B, 7-30.
3102 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
phosphorus, and sulfur12,13 compounds. The atmospheric plasmas used in those investigations employed He flow rates typically between 3 and 12 L/min and 150-460 W of power. In most cases, the torch box of commercially available ICPMSs were replaced with similarly large MIP plasmas with little reduction in instrument size or consumables (e.g., power, gas, and water). Although the contribution of those studies to the advancement of MIP-MS detection for GC is significant, only one reference demonstrates practical applicability of GC/MIP-MS by analyzing real-world samples.9 Moreover, chemical speciation has become increasingly important in the environmental, biochemical, clinical, and nutritional fields. It is recognized that specific elements play distinct roles depending on their chemical environment. Although the development of plasma sources has kept pace with the quantitative challenges of detecting atoms in complex mixtures, little advances have been made to obtain molecular structure information. The ability to generate compound-specific fragmentation patterns and total elemental composition with the same instrument would result in structural, elemental, and quantitative analysis capabilities unmatched by other techniques. There have been few reports on the use of microwave plasma sources for molecular ionization. Pousell et al.14 used a low-pressure (4 × 10-2 Torr) surfatron MIP operated at 35 W and 1-2 mL/min gas flow. Fragmentation of aliphatic and aromatic compounds, introduced via an injection valve, was similar to that obtained by electron impact (EI) ionization. Olson et al.15 also used a low-pressure MIP to fragment (4) Satzger, R. D.; Fricke, F. L.; Brown, P. G.; Caruso, J. A. Spectrochim. Acta, Part B 1987, 42B, 705-712. (5) Brown, P. G.; Davidson, T. M.; Caruso, J. A. J. Anal. At. Spectrom. 1988, 3, 763-769. (6) Creed, J. T.; Mohamad, A. H.; Davidson, T. M.; Ataman, G.; Caruso, J. A. J. Anal. At. Spectrom. 1988, 3, 923-926. (7) Creed, J. T.; Davidson, T. M.; Shen, W. L.; Caruso, J. A. J. Anal. At. Spectrom. 1990, 5, 109-113. (8) Mohamad, A. H.; Creed, J. T.; Davidson, T. M.; Caruso, J. A. Appl. Spectrosc. 1989, 43, 1127-1131. (9) Read, P.; Beere, H.; Ebdon, L.; Leizers, M.; Hetheridge, M.; Rowland, S. Org. Geochem. 1997, 26, 11-17. (10) Pack, B.; Broekaert, J.; Guzowski, J.; Poehlman, J.; Hieftje, G. Anal. Chem. 1998, 70, 3957. (11) Story, W. C.; Olson, L. K.; Shen, W. L.; Creed, J. T.; Caruso, J. A. J. Anal. At. Spectrom. 1990, 5, 467-470. (12) Story, W. C.; Caruso, J. A. J. Anal. At. Spectrom. 1993, 8, 571-575. (13) Suyani, H.; Creed, J.; Caruso, J.; Satzger, R. D. J. Anal. At. Spectrom. 1089, 4, 777-782. (14) Pousell, E.; Mermet, J.; Deruaz, D.; Beaugrand, C. Anal. Chem, 1988, 60, 923-927. 10.1021/ac9914418 CCC: $19.00
© 2000 American Chemical Society Published on Web 06/17/2000
alkanes and aromatic compounds. The plasma was sustained with 20 W and 26 mL/min He, with an expansion stage pressure of 0.13 mbar. In their study, analytes were also introduced by syringe injection or by opening a metering valve. The fragmentation patterns obtained were similar to those by EI. In contrast, Shen and Satzger16 used an atmospheric MIP operated at 30 W and 5 L/min He. Molecular ions were obtained for both gaseous and liquid organic compounds using an injector probe without disturbing the plasma. In these studies, the MIP-MS served as a detector without the benefit of GC separation of mixture components. In an earlier publication,17 we described a GC/MIP-MS system where a capillary MIP source operated at atmospheric pressure and, with He flow rates of 30 mL/min and 50 W power, was coupled to a mass spectrometer. Using Cl-selective detection, plasma stability was demonstrated by the excellent measurement precision, accuracy, and selectivity obtained for organochlorine pesticides detected in an extremely high hydrocarbon concentration (diesel fuel/engine oil) mixture. However, extremely poor sensitivity was observed unless O2 was used as the reagent gas, which minimized carbon deposition on the discharge tube walls. Although better sensitivity was obtained for the pure standard mixture, ∼300 pg of Cl/compound, suggesting less than optimum power transfer to the cavity. In addition, detection of 32S, present in some of the pesticides, required H2 (0.1 mL/min) as a reagent gas. The estimated detection limit for 32S in the standard pesticide solution was 19 ng of S/compound injected. Occasional impedance mismatches also caused hot spots within the cavity, which resulted in the discharge tube and O-ring melting, with subsequent leaking of water into the vacuum system. For highly complex environmental samples, two reagent gases were needed to obtain reasonable instrument performance, e.g., measurement precision and accuracy, with sensitivity and selectivity compromised depending upon the level of hydrocarbons in the mixture. In this paper, we demonstrate that optimum power transfer to the cavity can be obtained by better matching the impedance of the cavity to that of the generator. Results are presented based on microwave transmission from the magnetron/waveguide through a 50 Ω coaxial cable to the cavity. The MIP source is tuned by measuring and maintaining low reflected power by adjusting a double stub tuner. Elemental (Cl probe) measurement precision and sensitivity are compared for the coaxial and waveguide MIP GC/MS systems for 16 organochlorine pesticides separated by GC. The accuracy obtained with the coaxial MIP is demonstrated using a certified reference material soil contaminated with pesticides, polycyclic aromatic hydrocarbons, and petroleum. Finally, the ability of the capillary MIP-MS to generate molecular fragments is shown by analyzing a mixture of volatile organic compounds, separated by GC, typically found as waste solvents or gasoline constituents at hazardous waste sites. EXPERIMENTAL SECTION Instrumentation. Tables 1 and 2 list the GC/MIP-MS operating conditions used in the elemental and molecular fragmentation (15) Olson, L.; Story, W.; Creed, J.; Shen, W.; Caruso, J. J. Anal. At. Spectrom, 1990, 5, 471-475. (16) Shen, W.; Satzger, R. Anal. Chem. 1991, 63, 1960-1964. (17) Zapata, A. M.; Bock, C. L.; Robbat, A., Jr. J. Anal. At. Spectrom., 1999, 14, 1187-1192.
Table 1. Operating Conditions Employed with GC/MIP-MS GC GC column PTE-5, 30 m × 0.32 mm × 0.25 µm injector/detector/transfer line T 300 °C oven program 50-150 °C at 30 °C/min 150-220 °C at 4 °C/min 220-300 °C at 50 °C/min carrier gas He, 1.2 bar at 150 °C, constant flow solvent vent time 3 min plasma gas flow rate chamber 1 pressure chamber 2 pressure chamber 3 pressure extraction lens voltages SIM ion monitored dwell time
MIP-MS 30 mL/min He 1.3 mbar 3.1 × 10-5 mbar 6.0 × 10-6 mbar -87.3, -89.1 V 35 amu 100 ms
Table 2. Operating Conditions Employed with GC/ MIP-MS To Generate Molecular Ions
GC column injector/detector/transfer line T oven program carrier gas solvent vent time plasma gas flow rate chamber 1 pressure chamber 2 pressure chamber 3 pressure extraction lens voltages
GC DB-624, 30 m × 0.53 mm × 3 µm 80 °C 40-80 °C at 4 °C min-1 He, 0.2 bar at 50 °C, constant flow 1 min
MIP-MS 2.5 mL min-1 He 0.1 Torr 6.7 × 10-6 Torr 5.0 × 10-6 Torr -95.5, -204.5 V
studies, respectively. An Agilent Technologies (Palo Alto, CA) model G1800A GC/MS was also used to illustrate the complexity of the certified reference material. The same chromatographic conditions as shown for the GC/MIP-MS experiments were used. The mass spectrometer was operated in the total ion current mode and was scanned from 45 to 425 amu every 1.25 s. A detailed description of the GC/MIP-MS system is found in ref 17. Briefly, an Agilent pressure programmable GC, with a heated four-port solvent vent valve housed within the oven, was used. The GC was connected to the waveguide MIP via a heated transfer line. The MIP was mated to the MS through a threestage differentially pumped vacuum interface. The MS analyzer, a conventional Agilent 5972 shifted 2 mm off axis, along with the ion extraction and focusing lenses were located in a modified manifold. A schematic of the coaxial MIP source is shown in Figure 1. The microwave power at 2450 MHz, was generated by a Panasonic 2M211A magnetron tube (1; AMI, Eagle Grove, IA) into a rectangular waveguide (2). An antenna-coaxial adaptor (3), made from a brass rod/cylinder on one end and a coaxial N-type connector on the other end, was used to transmit the microwaves from the waveguide to coaxial cable. All of the coaxial cables used in the instrument were 50 Ω (RG 8/U, Pasternack Enterprises, Irvine, CA). A variable power attenuator (4; E-103, Alfred Electronics, Palo Alto, CA) was connected to the waveguide. Its purpose was to control the power going to the cavity and to act as an isolator by helping to dissipate the reflected power going back to the magnetron tube. A dual directional coupler (7; 777D, Agilent Technologies) was connected to the attenuated output of the Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3103
Figure 1. Schematic of coaxial MIP: (1) magnetron tube, (2) rectangular waveguide, (3) antenna to coaxial adaptor, (4) variable power attenuator, (5, 6) termination loads, (7) dual directional coupler, (8) thermistor mounts, (9) power meters, (10) dual stub tuner, (11) coaxial tocoupling loop adaptor, (12) TM010 cavity, and (13) waveguide test port.
Figure 2. GC transfer line for generation of molecular fragments: (1) 5971 MSD transfer line, (2) 1/16-in. stainless steel tubing, (3) capillary column, (4) skimmer cone, and (5) plasma discharge tube.
variable attenuator, to continuously and simultaneously monitor the forward and reflected power via thermistor mounts (8; 478A, Agilent Technologies) and two independent power meters (9; 432A and 431B, Agilent Technologies). The cavity was tuned by manipulating the dual stub tuner (10; DS109L, Weinschel Engineering, Gaithersburg, MD) placed at the output end of the directional coupler. The original re-entrant TM010 cavity antenna was replaced by a coaxial adaptor (11), which consisted of an N-type connector on one end and a 2.6 cm, 25 Ω coaxial section on the other screwed into the coupling loop of the cavity (12). GC/MS and GC/MIP-MS data were collected with the HP ChemStation software. A customized GC transfer line was connected to the side port of chamber 1 (see Figure 2) for the generation of molecular fragments. Approximately 30 cm of 1/ -in. stainless steel tubing (2) was housed inside a 5971 MSD 16 GC transfer line (1). The stainless steel tubing was resistively heated at temperatures up to 250 °C. The GC capillary column (3) was threaded through the inside of the tubing and was aligned ∼3 mm in front of the skimmer cone (4) orifice. Monitoring and control of the heating components was achieved by employing available ports (injector B and auxiliary) on the GC heating circuit panel. Measurement precision, accuracy, and sensitivity were 3104 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
determined using the Ion Fingerprint Detection (IFD) mass spectrometry software postacquisition. IFD was obtained from Ion Signature Technology, Inc. (Cambridge, MA). Chemicals and Reference Materials. The organochlorine pesticide mixture (US-102BN, Ultra Scientific, N. Kingstown, RI) contained the following compounds at 2000 µg/mL each: aldrin, R-BHC, β-BHC, δ-BHC, 4,4′-DDD, 4,4′-DDE, 4,4′-DDT, dieldrin, endosulfan I, endosulfan II, endosulfan sulfate, endrin, endrin aldehyde, γ-BHC, heptachlor, and heptachlor epoxide, with 2,2′,4,6-tetrachlorobiphenyl and decachlorobiphenyl used as internal and surrogate standards, respectively. Optima grade hexane (Fischer Scientific, Pittsburgh, PA) was used to prepare standard solutions for the dynamic range and detection limit experiments and for extraction of soil samples. Percent recovery experiments were performed using a certified reference soil (IRM-105) obtained from Ultra Scientific. The contaminated material came from a Superfund site in the western United States and was found to contain 8 of the target analyte pesticides, 17 polycyclic aromatic hydrocarbons, pentachlorophenol, carbazole, and dibenzofuran as well as petroleum. He (UHP grade) was obtained from Northeast Air (Salem, NH); vinyl chloride (2000 ppm in He) was obtained from Airgas (Radnor, PA). The chlorinated volatile organic compound (VOC) mixture (Supelco Inc., Bellefonte, PA) used in the molecular ionization experiments contained the following compounds at 2000 µg/mL each: 1,1,1-trichloroethane, 1,2-dichloroethane, cis-1,3-dichloropropylene, trans-1,3-dichloropropylene, tetrachloroethylene, ethylbenze, and p-xylene. Standards were prepared by serial dilution in Optima grade methanol (Fischer Scientific). Preparation of Soil Samples. The surrogate, decachlorobiphenyl, was added to the reference soil, homogenized, and allowed to dry at room temperature. A 2 g sample was placed in a 7 mL glass vial with a Teflon-lined screw cap. A total of 2 mL of hexane was added to the vial; the resultant mixture was shaken for 3 min and then centrifuged for 5 min. One milliliter of extract was transferred to a GC amber vial for analysis. Determination of Linear Dynamic Range and Instrument Detection Limit. A standard chlorinated pesticide solution (2000 µg/mL) was serially diluted until the 35Cl signal (peak area) was less than 3σ of the blank. This calibration curve, which consisted of 11-13 points, was used to determine the concentration of pesticides in the reference material. Response factors were determined for each pesticide using 2,2′,4,6-tetrachlorobiphenyl as the internal standard. The response factor (RF) was calculated as the ratio of AxCis/AisCx, where Cx was the amount of target analyte injected and Ax the signal obtained. Cis and Ais were the corresponding values for the internal standard. For the instrument detection limit study, seven consecutive injections were made of standard mixtures whose pesticide signals were 3σ of the blank. Each pesticide’s concentration was determined from the calibration curve. The standard deviation was then calculated and the resulting value was multiplied by the Student’s t value at the 99% confidence level. Pesticide detection limits are reported based on a Cl/s basis, which was obtained by dividing the Cl detection limit by the width of the chromatographic peak at half-height. RESULTS AND DISCUSSION Our initial investigation employed a MIP cavity that was tuned by the manufacturer. Several disadvantages were observed. For
Figure 3.
35Cl
and
32S
chromatograms for a standard pesticide solution.
example, analysis of pesticides in the presence of more than 10% 2:1 diesel fuel-to-engine oil mixture resulted in carbon deposition on the walls of the discharge tube and a nondetectable Cl signal. The addition of O2 as a reagent gas prevented the carbon deposition and provided the means for quantifying pesticides in the presence of a 70% petroleum mixture. The estimated minimum quantitation level in this mixture was 15 ng of Cl/pesticide injected. In addition, for the standard pesticide solution or the diesel fuel/engine oil mixtures sulfur, present in three of the pesticides, was detectable only when H2 was used as a reagent gas. The estimated minimum quantitation level in the petroleum mixture was 19 ng of S/pesticide injected. To improve instrument performance, the MIP was modified to gain better control of the power transfer to the cavity. Experiments were performed to determine the optimum location for the waveguide reflector walls so that an incident wave maximum was positioned at the antenna-coaxial interface. This was done by measuring the forward and reflected power, respectively, at locations 3 and 13 in Figure 1. Using vinyl chloride as the tuning gas, the MIP-MS signal was optimized daily by adjusting the double stub tuner and MS ion optics until a compromise was reached between low reflected power and maximum 35Cl signal/noise ratio. Optimum performance was achieved when the forward power was ∼60 W and the reflected power was ∼15 W. In contrast to the early study, no reagent gas (O2) was needed with the tunable MIP to obtain the same analytical figures of merit, namely, 15% RSD precision and 10% accuracy, for the highest concentration petroleum mixture. These results are consistent with the fact that carbon deposition was visually nondetectable on the discharge tube walls. The 32S signal was also observed with the tunable source without the need of H2 as the reagent gas. Figure 3 shows the 35Cl (top) and 32S (bottom) chromatograms obtained from a 50 ng/compound injection of a standard pesticide solution. The three peaks in the 32S chromatogram were produced from endosulfan I, endosulfan II, and endosulfan sulfate and demonstrate the feasibility of using S as a simultaneous detection probe. 32S was detected in the 70% diesel fuel/engine oil mixture at concentration levels 10 times below what the
nontunable waveguide MIP could produce. Other studies have shown that the high background levels at m/z 32 found in He MIP sources required that either the less abundant isotope 34S be detected12 or that nitrogen be used at reduced pressure.11 As compared to other MIP instruments, the tunable MIP is easier to operate and provides the opportunity for multielement detection. Table 3 compares the linear dynamic range and instrument detection limit for the tunable and non tunable sources. The linear dynamic range for the coaxial MIP for all 16 pesticides extended over 3 orders of magnitude, nearly 1 order of magnitude greater than that of the nontunable waveguide. Both the correlation coefficient and the average response factor percent relative standard deviation (RF% RSD) calculated from ∼12 points over the calibration curve for each pesticide was much better for the coaxial versus waveguide systems. For example, the average RSD for the coaxial MIP was 9% as compared to 14% for the waveguide MIP. Increased sensitivity was also observed for the tunable MIP. Pesticide concentrations were between 3 and 19 pg/s, which is an improvement of about seven times the detection limit obtained with the waveguide. Pesticide detection limit studies obtained with plasma source instruments are listed in Table 4. For comparison purposes, the coaxial MIP-MS detection limits were converted to picograms by multiplying the values in Table 3 by the corresponding peak width. All values are corrected for the amount of Cl in each pesticide. An Agilent 5921A atomic emission detector (AED) was used with GC by Miyahara et al.18 for the analysis of eight organochlorine pesticides. They compared results of the MIP-AED with flame photometric (FPD), electron capture (ECD), and electron impact MS detectors. MIP-AED sensitivity was between 145 and 560 pg. Although they were able to obtain better sensitivity with the ECD (∼30 pg), AED was better than MS and FPD. The EPA will publish a GC/MIP-AED method for the analysis of chlorinated pesticides based on data produced by Cummings and co-workers.19 The detection limit was similar to that reported by Miyahara. Gurka (18) Miyahra, M.; Suzuki, T.; Saito, Y J. Agric. Food Chem. 1992, 40, 64-69. (19) Method 8085: Compound-Independent Elemental Quantitation of Pesticides by Gas Chromatography with Atomic Emission Detection (GC/AED); U.S. Environmental Protection Agency, Port Orchard, WA, January 1999.
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3105
Table 3. Comparison of Linear Dynamic Range and Detection Limit for 16 Pesticides Obtained with Waveguide and Coaxial MIP’s dynamic range (n ) 3) coaxial pesticide
decades
r2
R-BHC β-BHC γ-BHC δ-BHC heptachlor aldrin heptachlor epoxide endosulfan I dieldrin 4,4′-DDE endrin endosulfan II 4,4′-DDD endrin aldehyde endoslfan sulfate 4,4′-DDT
3.4 3.4 3.4 3.2 3.4 3.4 3.4 3.4 3.4 3.2 3.4 3.2 3.4 3.2 3.2 3.1
1.000 1.000 1.000 1.000 1.000 1.000 0.999 0.998 0.998 1.000 0.999 0.999 0.997 0.999 0.999 0.999
RF
% RSD
2.51 4.40 4.45 1.93 1.69 1.11 1.50 1.13 1.32 0.901 0.915 0.986 0.778 0.866 0.808 0.636
7 9 8 13 12 5 7 6 6 5 9 8 12 14 8 12
Table 4. Comparison of Pesticide Detection Limits Obtained with Different GC Plasma Sources and Detectorsa detection limit (pg of Cl/compd) pesticide
this workb
MIPAEDc
MIPAEDd
MIPAEDe
R-BHC β-BHC γ-BHC δ-BHC heptachlor aldrin heptachlor epoxide endosulfan I dieldrin 4,4′-DDE endrin endosulfan II 4,4′-DDD endrin aldehyde endsulfan sulfate 4,4′-DDT
40 40 40 190 70 70 50 60 60 30 70 50 40 30 70 80
146 146 242
540 540 540 540 180 108 144
183 183 117 226 86 81 73 81 50 50 390 94 90 90 100 125
464 560 440
120 360 540
440
360 360 540 540
500
ICPMSf
14
a All values in the table were corrected for the amount of Cl in each pesticide. b Coaxial MIP-MS detection limits are at 3 times the standard deviation of the noise; results from Table 3 were multiplied by the chromatographic peak width. c Reference 18; detection limit based on S/N ) 3. d Reference 19; detection limits are at 3 times the standard deviation of the noise. eReference 20; detection limit based on a S/N ) 3. f Reference 21; detection limit calculated by dividing 3 times the standard deviation of the background by the slope of the calibration plot.
et al.20 studied the MIP-AED response of 58 semivolatile organic compounds from different chemical classes. Detection limits were 50-390 pg for 18 chlorinated pesticides. In contrast, only one study, excluding ours, reports detection limits for pesticides by plasma MS. Castillano and co-workers21 reported the detection limit, 14 pg, for lindane using a low-pressure helium ICPMS. The limit of detection was calculated at 3 times the standard deviation of background signal divided by the slope of the calibration curve. (20) Gurka, D. F.; Pyle, S.; Titus, R. Anal. Chem. 1997, 69, 2411-2417. (21) Castillano, T.; Giglio, J.; Hywel Evans, E.; Caruso, J. J. Anal. At. Spectrom. 1994, 9, 1335-1340.
3106
35Cl
waveguide
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
DL (pg/s, n ) 7)
decades
r2
RF
% RSD
coaxial
waveguide
2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.3 2.3
0.994 0.994 0.994 0.998 0.997 0.996 0.997 0.998 0.999 1.000 1.000 1.000 0.996 1.000 0.998 0.999
1.28 2.75 2.72 1.01 0.960 1.05 1.06 0.823 0.780 0.878 0.617 0.709 0.596 0.707 0.316 0.471
18 18 17 19 17 9 12 16 11 11 9 10 15 17 14 17
4 4 4 19 7 7 5 6 6 3 7 5 4 3 7 8
46 21 10 43 34 20 30 41 37 27 36 36 41 25 43 25
In this study, we detected 30-190 pg for 16 organochlorine pesticides. Although it is difficult to compare instrument performance based on one compound, the MIP-MS detection limit for lindane was ∼3 times higher than that calculated for the ICPMS instrument. As can be seen from the table, measurement sensitivity is both instrument- and compound-specific. The performance of the tunable source was also evaluated by determining the percent recovery for seven pesticides found in a certified reference soil. The sample was obtained from a U.S. Environmental Protection Agency hazardous waste site. On the basis of an interlaboratory comparison study of 20 laboratories, the sample contained 8 chlorinated pesticides (70 ppm total), 17 polycyclic aromatic hydrocarbons (6000 ppm total), pentachlorophenol (1200 ppm), carbazole (40 ppm), and dibenzofuran (256 ppm) as well as an unknown mixture of petroleum contaminants. GC/MS and GC/MIP-MS (tunable) chromatograms are shown in panels a and b of Figure 4, respectively. Note that the GC/MS total ion current signal was 105 as compared to target pesticides, which were between 102 and 103. For example, peaks 3, 4, and 5 correspond to 4,4′-DDE, endrin, and endosulfan 2, respectively, and fall in the valley between PAH and petroleum constituents at 15.4, 16, and 16.5 min. It was extremely difficult to find these pesticides by EI MS without performing extensive sample cleanup. In contrast, the GC/MIP-MS 35Cl signals are clearly distinguishable. With the exception of γ-BHC, which coeluted with pentachlorophenol, all other chlorinated pesticides were easily identified. Recall that pentachlorophenol was present in the soil at 1200 ppm as compared to the pesticides at ∼10 ppm/compound. It appears from Figure 4b that some other chlorinated compound was also present in the sample. Percent recoveries, measurement precision, the certified reference values at the 95% confidence interval (CI), and the prediction intervals (PI) are listed in Table 5. Except for 4,4′-DDD (30% recovery), our results are within 10% of the certified values and well within both the confidence and prediction intervals. When the same extract was analyzed three times, measurement precision was
