EPA/600/R-16/124 I June 2016
www.epa.gov/homeland-security-research
United States
Environmental Protection
Agency
oEPA
Experiments with the
LECO Pegasus® Gas Chromatograph/
Time-of-Flight Mass Spectrometer
Phase 2: 2D Fast GC Separations
Office of Research and Development
National Homeland Security Research Center
-------
This page left intentionally blank
-------
&EPA
United States
Environmental Protection
Agency
EPA/600/R-16/124 I June 2016
www2.epa.gov/homeland-security-research
Experiments with the
LECO Pegasus® Gas Chromatograph/
Time-of-Flight Mass Spectrometer
Phase 2: 2D Fast GC Separations
Office of Research and Development
National Homeland Security Research Center
l
-------
Disclaimer
Neither the United States government nor Lawrence Livermore National Security,
LLC, nor any of their employees makes any warranty, expressed or implied, or
assumes any legal liability or responsibility for the accuracy, completeness, or
usefulness of any information, apparatus, product, or process disclosed, or
represents that its use would not infringe privately owned rights. The U.S.
Environmental Protection Agency does not endorse the purchase or sale of any
commercial products or services. Reference herein to any specific commercial
product, process, or service by trade name, trademark, manufacturer, or otherwise
does not constitute or imply its endorsement, recommendation, or favoring by the
United States government or Lawrence Livermore National Security, LLC. The
views and opinions of authors expressed herein do not necessarily state or reflect
those of the United States government or Lawrence Livermore National Security,
LLC, and shall not be used for advertising or product endorsement purposes.
Questions concerning this document or its application should be addressed to:
Romy Campisano
U.S. Environmental Protection Agency
Office of Research and Development
National Homeland Security Research Center (NG16)
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7016
campisano. romy @epa. gov
ii
-------
Acknowledgments
The research team wishes to acknowledge the support of all those who helped plan
and prepare this report. We thank Chris Retairdes, Eric Boring, Matthew
Magnuson, and Eric Graybill for their reviews of this work and their constructive
suggestions and comments. The U.S. Environmental Protection Agency, Office of
Research and Development, National Homeland Security Research Center
(NHSRC) funded this work and Romy Campisano of this organization provided
helpful conversations and contributions.
We thank Farai Rukunda, Spectrometer Training Specialist, and Scott Pugh,
Applications Laboratory Manager, of LECO Corporation for their support in
resolving hardware and software issues and for providing many helpful
conversations that allowed us to complete this study.
This document was prepared as an account of work sponsored by an agency of the
United States government. This work performed under the auspices of the U.S.
Department of Energy by Lawrence Livermore National Laboratory under
Contract DE-AC52-07NA27344. The research team was comprised of Heather
Mulcahy, Carolyn Koester, and Roald Leif.
111
-------
Abbreviations/Acronyms
2D - two-dimensional (chromatography)
CWA - Chemical Warfare Agent, in the context of this report, the CWAs of interest are HD,
GB, GD, GF, and VX
DFTPP - Decafluorotriphenylphosphine
EPA - United States Environmental Protection Agency
GB - sarin
GC - Gas Chromatography
GC/MS - Gas Chromatography/Mass Spectrometry. In the context of this report, mass
spectrometry is performed with a quadrupole mass spectrometer
GC/TOF-MS - Gas Chromatography coupled with Time-of-Flight Mass Spectrometry
GD - soman
GF - cyclosarin
HD - sulfur mustard
IAG - Interagency Agreement
i.d. - internal diameter (of a gas chromatography column)
IDL - Instrument Detection Limit; this figure of merit provides an indication of the optimal
capability of an instrument.
LLNL - Lawrence Livermore National Laboratory
m/z - mass-to-charge ratio for a specified ion produced by fragmentation in a mass spectrometer
NHSRC - National Homeland Security Research Center, Cincinnati, OH
NMR - Nuclear Magnetic Resonance Spectroscopy
PFTBA - Perfluorotributylamine
ppm - parts-per-million
QAPP - Quality Assurance Project Plan
RRF - Relative Response Factor
RSD - Relative Standard Deviation
RT - Retention Time (of a compound eluting from a chromatographic column)
SIM - Selected Ion Monitoring (operating mode of a mass spectrometer)
S:N - Signal-to-Noise ratio
TIC - Total Ion Chromatogram
iv
-------
TOC - Total Organic Carbon
TOF-MS - Time-of-Flight Mass Spectrometry
VX - 6>-ethyLY-[2-(diisopropylarnino)ethyl] methylphosphonothioate
-------
Executive Summary
Conventional analysis by gas chromatography/ quadrupole mass spectrometry (GC/MS)
can be time-consuming (30-60 minutes) and prone to interferences. The use of fast gas
chromatography coupled with time-of-flight mass spectrometry (GC/TOF-MS) offers the
advantages of faster (-10 min) analysis times, improved GC resolution afforded by the use of
narrower (0.1- 0.18 mm i.d.) columns, and improved mass resolution and scan speed provided
by the TOF-MS. In addition, GC/TOF-MS offers the promise of better instrument detection
limits than quadrupole GC/MS, while still providing full mass spectral data. If the GC/TOF-MS
can be operated to perform two-dimensional (2D) separations, the ability to resolve peaks of
interest from interfering compounds is improved. In 2D chromatography, a second GC column
with a different chemical phase than that of the first GC column is used to provide additional
separation of analytes as they elute from the first column. Because the chemistry of the second
GC column is different from the first, compounds that co-elute from the first column may be
easily resolved after a separation on the second GC column (i.e., the peak capacity of the system
is increased and the specificity of analyte detection is improved).
In this study, the LECO Pegasus® 4D GC/TOF-MS was used in the 2D mode to detect
chemical warfare agents (CWAs) and to compare this instrument's performance, with regards to
speed of analysis, instrument detection limits, and various matrices to conventional quadrupole-
based GC/MS. Analytes studied were sulfur mustard (HD), sarin (GB), soman (GD), cyclosarin
(GF), and 6>-ethyLY-[2-(diisopropylamino)ethyl] methylphosphonothioate (VX). Instrument
detection limits, as determined by 2D GC/TOF-MS, for these CWAs were compound-dependent
and ranged from 0.5-5 pg. 2D GC/TOF-MS was shown to provide fast separations and produce
lower instrument detection limits by approximately a factor of 10-20, compared to the speed of
analysis and instrument detection limits obtained by a quadrupole GC/MS (Agilent 5973 system)
operated in the SIM mode. In most cases, lower instrument detection limits, by approximately a
factor of 10-20, were obtained than the instrument detection limits obtained by GC/TOF-MS
operated with only a single separation.
2D GC/TOF-MS was also investigated for use in identifying factors associated with
inconsistent VX quantification. The quantification of VX has been found to be difficult in
previous studies in our laboratory. A calibration curve for VX was generated, with standards
ranging in concentration from 0.1 to 2 |ig/mL, and found to be linear (R2=0.99989). Using this
regression, the concentrations of the individual VX standards were determined and compared
with their known amounts. In all cases, calculated concentrations were within 5% of their
expected values.
Selected problematic matrices (those that consistently showed recoveries of greater than
100% for VX when analyzed by GC/MS) were spiked with VX, extracted, and analyzed using
the 2D GC/TOF-MS and quadrupole GC/MS. The matrices studied included samples of spiked
wipes, Virginia soil, and drywall. VX concentrations determined by 2D GC/TOF-MS and
GC/MS were compared. VX concentrations were consistently higher when measured by GC/MS
than when they were measured by 2D GC/TOF-MS (compared by Student's t-test, at a
significance level of a=0.05). 2D GC/TOF-MS was able to detect interferences in soil, drywall,
and wipe extracts that were not observable by a single separation with GC/MS. The observation
of interferences visualized by 2D GC/TOF-MS provided a partial explanation of why higher VX
vi
-------
concentrations were observed when GC/MS was used.
vii
-------
Table of Contents
Disclaimer ii
Acknowledgments iii
Abbreviations/Acronyms iv
Executive Summary vi
List of Tables ix
List of Figures x
1.0 Introduction and Background 1
2.0 Study Objectives 2
3.0 Experimental Conditions 2
3.1 Standards 2
3.2 Sample preparation 3
3.3 2D GC/TOF-MS conditions 4
3.4 GC/MS conditions (conventional quadrupole) 5
3.5 Quality Assurance (QA) 6
4.0 Results 7
4.1 Chromatographic separation 7
4.2 Instrument detection limits 10
4.3 VX calibration 12
4.4 Measurements of VX in sample extracts 12
5.0 Conclusions and Recommendations 22
Appendix A: Autosampler Method for LECO 2D GC/TOF-MS 23
Appendix B: GC Method for LECO 2D GC/TOF-MS 24
viii
-------
List of Tables
Table 1. Average Retention Times (± Standard Deviations) and Average Analyte
Responses (± Standard Deviations) for Seven Replicate Injections of 0.2 ng of CWAs
into the 2D GC/TOF-MS 10
Table 2. 2D GC/TOF-MS IDLs, Quantification (Quant) and Qualifying (Qual) Ions, and
Signal-To-Noise Ratios (S:N)11
Table 3. Comparative Data for Instrument Detection Limits (pg) Determined by Various
GC/MS Configurations 11
Table 4. Measured Amount of VX (|ig) (n=3) in Sample Extracts of Drywall Coupons
Spiked With 1 |ig of VX 21
IX
-------
List of Figures
Figure 1. Total ion chromatogram obtained when separating 0.5 ng each agent and 1 ng
each surrogate by ID GC/TOF-MS. 7
Figure 2. Extracted ion chromatogram (m/z 99) obtained when separating 0.2 ng each
CWA with 2D GC/TOF-MS. 9
Figure 3. Selected ion chromatogram of m/z 114 (VX quantitation ion) of Virginia soil
method blank (blue) and the overlaid chromatogram of m/z 114 for a 1 ppm VX standard
(black) 13
Figure 4. 2D chromatogram of extract produced from 0.1 |ig VX per gram Virginia soil.
14
Figure 5. 2D chromatogram of unspiked Virginia soil showing ions representing matrix
interferences for VX; no m/z 114 (VX quantitation ion) is present. 15
Figure 6. 2D GC contour plots of the TICs of the Virginia soil method blank (top) and
VX-spiked Virginia soil (bottom). 16
Figure 7. 2D GC contour plots of m/z 114 (VX quantification ion) of Virginia soil
method blank (top) and VX-spiked (0.1 jug/g) Virginia soil (bottom). 17
Figure 8. Magnified view of 2D GC contour plot for m/z 114 of VX-spiked (0.1 jug/g)
Virginia soil. 18
Figure 9. 2D GC/TOF-MS contour plot for m/z 114 (quantification ion for VX) from an
extract derived from a wipe spiked with 1 |ig of VX. 19
Figure 10. 2D GC/TOF-MS contour plot for m/z 114 (quantification ion for VX) from a
wipe extract containing 1 |ig VX. 19
Figure 11. 2D GC/TOF-MS contour plot of m/z 114 (the VX quantitation ion) for a
drywall extract, produced from 1 |ig VX on spiked drywall, which was filtered using
Whatman Autovial. 21
x
-------
1.0 Introduction and Background
Gas chromatography/mass spectrometry (GC/MS) is the method of choice for the
analysis of volatile and semivolatile organic compounds in environmental samples. Most
laboratories performing GC/MS analysis use quadrupole mass spectrometers. Quadrupole
GC/MS has been a laboratory standard technique for many years because the technology,
including data processing algorithms, is relatively mature and the instrumentation is relatively
inexpensive (-$85,000). In addition, quadrupole GC/MS systems are rugged, reliable, and
provide good detection limits-low- to sub-nanogram quantities of chemicals when operated in
the full scan mode and low picogram amounts of materials when operated in the selected ion
monitoring (SIM) mode.
The use of time-of-flight mass spectrometers (TOF-MSs) is becoming more common.
GC/TOF-MS is comparable in price (-$101,000 for a basic unit, with a single GC column) to
quadrupole GC/MS. GC/TOF-MS provides low picogram detection limits and retains complete
mass spectral data for each compound detected. Retention of complete mass spectral data is a
distinct advantage to increase confidence in analyte identifications. We have shown that
GC/TOF-MS can be used as an alternative to quadrupole GC/MS for the detection and
measurement of CWAs (1). Because its principle of operation offers fast data acquisition, the
TOF-MS is an ideal detector to couple with fast chromatographic separations [i.e., separations
that provide improved GC resolution afforded by the use of narrow, 0.1- 0.18 mm internal
diameter (i.d.), columns], including two-dimensional (2D) separations.
GC/TOF-MS operated in the 2D mode becomes a more valuable analytical tool in
comparison to GC/TOF-MS operated in ID mode because of increased detection specificity. 2D
separations are performed using two capillary columns of different phases which are connected
via a dual-stage thermal modulator. To perform 2D separations, a sample extract is first
introduced into the primary column, where the sample undergoes the initial separation. In the
work described here, a 15 m in length by 0.25 mm (or 0.18 mm) i.d. column was installed in the
GC injector. The end of this column was connected to a short, narrower bore (i.e., 1 m x 0.1 mm
i.d.) column, of a different stationary phase than the first column, that terminated in the TOF-
MS. This short column provided a second GC separation. Thus, all samples injected into the
GC passed through two GC columns. At the junction of the two columns was a dual-stage
thermal modulator that focused effluent from the primary column onto the secondary column
with cold provided by liquid nitrogen (i.e., a cold jet), and then quickly injected this effluent,
using heated air to desorb analytes from the modulation zone (i.e., hot jet), onto the secondary
column. By performing the modulation process rapidly (on the average of 3 seconds) and
performing a quick secondary separation, resolution on the first column was maintained and two
independent separations occurred from one injection. By performing 2D separation, the ability
to resolve compounds of interest from interferences was improved (i.e., the peak capacity of the
system is increased and the specificity of analyte detection is improved). In addition,
cryofocusing and modulation of the effluent from the first GC column, which was performed
prior to 2D separation, resulted in improved, narrower, peak shapes for most compounds and,
thus, lower instrument detection limits than could be obtained from a single GC separation. The
increased analytical power provided by 2D-GC/TOF-MS comes at double the cost of a basic
GC/TOF-MS system; the instrument used in this study is valued at approximately $215,000.
1
-------
In this study, the LECO Pegasus® 4D GC/TOF-MS, operated in 2D mode, was used to
detect chemical warfare agents (CWAs) and this instrument's performance, with regards to speed
of analysis and detection limits, was compared to conventional quadrupole GC/MS. Analytes
studied were sulfur mustard (HD), sarin (GB), soman (GD), cyclosarin (GF), and 6>-ethyLY-[2-
(diisopropylamino)ethyl] methylphosphonothioate (VX). The use of the 2D GC/TOF-MS was
also compared with quadrupole GC/MS for the analysis of VX in various matrices, including
soil, drywall, and wipes, that were spiked with VX. Please note that, for the remainder of this
report, the term "GC/MS" will be used when quadrupole GC/MS is indicated.
2.0 Study Objectives
The focus of this work was to determine how to best utilize the 2D GC/TOF-MS for the
analysis of CWAs and the value of the 2D mode. Specifically, the goals of this study were to:
a) establish separation conditions for the analysis of HD, GB, GD, GF, and VX by 2D
GC/TOF-MS
b) determine instrument detection limits (IDLs) by 2D GC/TOF-MS (electron ionization
mode) for HD, GB, GD, GF, and VX
c) investigate VX analysis in various matrices
Specifically, with regard to the VX analysis, the amount of VX in the extracts from
various spiked matrices, including soil, drywall, and wipes, was measured to determine if the use
of 2D GC/TOF-MS analysis would eliminate quantitation problems that might be attributed to
interferences encountered in quadrupole GC/MS analysis.
3.0 Experimental Conditions
The experimental strategy was to first optimize 2D separation conditions for the analysis
of HD, GB, GD, GF, and VX and then to determine IDLs. Next, the 2D GC/TOF-MS was used
to determine VX in various matrices and to compare the measured amounts of VX with those
determined by ID GC/TOF MS and by GC/MS analyses.
3.1 Standards
CWA standards used for this study were synthesized by Lawrence Livermore
National Laboratory (LLNL) and were characterized for purity by nuclear magnetic
resonance spectroscopy (NMR) and GC/MS analyses. As determined by proton NMR,
the purities for GB, GD, GF, HD, and VX were 97.2%, 92.9%, 94.4%, 94.0%, and
94.0%), respectively. Dilute standards were prepared gravimetrically from neat materials
and diluted in dichloromethane (stock solutions of 100 ppm concentrations of each
individual CWA were made new every 6 months; dilute solutions used for calibrations
2
-------
were made from the 100 ppm stock solutions as needed, but typically used within two
weeks).
Surrogate standards used were those of U.S. EPA Method 8270D and included
nitrobenzene-ds (NB-ds), 2-fluorobiphenyl (FBP), phencyclidine-ds (PCP-ds),
terphenyl-di4 (Ter-di4), and triphenyl phosphate (TPP). Specific solutions purchased for
this work included: Base/Neutrals Surrogate Standard, 1000 |ag/m L, in dichloromethane
(Catalog number ERB-076, Cerilliant, Round Rock, TX), Triphenylphosphate, 5000
|ig/mL, in methyl tert-butyl ether (Catalog number ERT-108S, Cerilliant), and PCP-ds
(phencyclidine-ds), 1000 |ag/m L, in methanol (Catalog number P-006, Cerilliant).
Internal standards used were those of U.S. EPA Method 8270D and included
l,4-dichloro-benzene-d4, naphthalene-dx, acenaphthene-dio, phenanthrene-dio,
chrysene-di2, and perylene-di2. These standards were purchased as a Semivolatile
Internal Standard Mix, 2000 |ag/m L in dichloromethane (Catalog number 861238,
Supelco, Bellefonte, PA). Internal standards were spiked into all sample extracts to
produce a final concentration of 1 ng/|iL for all analyses.
Decafluorotriphenylphosphine (DFTPP) was used to verify that the mass
spectrometer systems were functioning optimally. DFTPP was purchased as a solution
with a concentration of 1000 |ag/m L in acetone (Catalog number, 47941, Supelco,
Bellefonte, PA).
All standards and samples were stored at 4-8 0 C.
3.2 Sample preparation
All sample preparation procedures were consistent with a draft CWA analytical
protocol currently under development at EPA and the extraction materials and protocols
used have been previously described (2). Only VX was investigated in different
matrices.
Soil samples
Sand, purified, CAS No. 14808-60-7, Part No. 3382-05 (JT Baker, Inc.,
Phillipsburg, NJ) and Virginia soil, with composition of 64.5% sand, 28% silt, 7.5%
clay, and 2.6% Total Organic Carbon (TOC) and pH 4.1 in 1:1 soil:water (obtained
from National Exposure Research Laboratory, U.S. EPA, Las Vegas, NV) were used in
this study. Briefly, 10-g aliquots of sand and soil were spiked with 500 ng of each
surrogate (per Section 3.1), and extracted for one hour by water bath sonication with 25
mL of 25/50/25 (v/v/v) acetone/dichloromethane/ ethyl acetate. The resulting extract was
separated from the soil by centrifugation and the supernatant removed. The soils were
then extracted for a second time, as described above, with 5% triethylamine (TEA) in
ethyl acetate. Extracts from the two extraction procedures were kept separate, reduced in
volume to approximately 500 |iL [using nitrogen and a Rapid Vap unit, customized to
accommodate 40-mL vials (LabConco, Kansas City, MO) and a Pierce Reacti-Therm III,
#188 evaporation module (ThermoScientific., Hudson, NH)], reconstituted to 1.00 mL
with dichloromethane, and spiked with internal standards (per Section 3.1) prior to
3
-------
analysis.
Wipe samples
Wipes (3" x 3", Kendall-Curity, 12-ply, P/N 1903, available from Tyco
Healthcare Group LP, Mansfield, MA) were spiked with 500 ng surrogates (see Section
3.1 above), and extracted by waterbath sonication for 30 minutes, twice, with 15 mL
25/50/25 (v/v/v) acetone/dichloromethane/ethyl acetate. The resulting extracts were
combined, evaporated to 1.00 mL (with nitrogen, as previously described), and spiked
with internal standards (per Section 3.1) prior to analysis.
Drywall samples
Circular coupons (1.5" in diameter and 0.5" thick) of standard drywall (Home
Depot), painted with 1 coat Glidden commercial latex primer and 1 coat Glidden interior
eggshell paint, were spiked with l|ig of VX, and were extracted by waterbath sonication
for 15 minutes, twice, with 20 mL 50/50 (v/v) dichloromethane/acetone. The resulting
extracts were combined, evaporated to 1.00 mL (with nitrogen, as previously described),
and spiked with internal standards (per Section 3.1) prior to analysis.
All sample extracts were stored at 4-8 °C until the time of analysis and each batch
of sample extracts was analyzed with corresponding method blanks.
3.3 2D GC/TOF-MS conditions
2D GC/TOF-MS experiments were performed with an Agilent 6890 gas
chromatograph (Agilent Technologies, Inc., Santa Clara, CA) coupled with a LECO
Pegasus® 4D mass spectrometer (LECO Corp., St. Joseph, MI). Prior to use, the 2D
GC/TOF-MS was tuned with the vendor's standard protocols and perfluorotributylamine
(PFTBA) as a calibrant. An injection of 10 ng DFTPP was used to check the
performance of the instrument prior to analyzing samples. Experimental data were
collected using the exact same instrument conditions, including electron multiplier
voltages, as those used to analyze the DFTPP check samples.
The method parameters routinely used to collect data have been recorded in
Appendices A, B, and C of this report. Because the values of these parameters that are
entered into the LECO software are of great importance, screen capture images of the
method setup pages are included so that the conditions of the analyses can easily be
replicated.
Briefly, the following conditions were used to achieve 2D GC separations.
Injection size:
Injection type:
Pulse pressure:
Purge time:
Carrier gas:
GC injection port:
1 |iL
Split/splitless (pulsed)
40 psi for 0.5 min
35 sec at 30 mL/min
He (Ultrahigh purity, Air Products, Allentown, PA) with
constant flow of 1.2 mL/min
250 °C
4
-------
GC columns:
GC oven:
2nd GC oven:
GC transfer line:
Modulation period:
Hot pulse time:
Cool time:
15 m x 0.25 mm i.d. x 0.25 |im film thickness, HP5-MS UI
(Agilent Technologies, Inc, Santa Clara, CA)
1 m x 0.1 mm i.d. x 0.1 |im film, Rxi-17 (Restek,
Bellefonte, PA)
55 °C held for 0.5 min, 10 °C/min to 115 °C, 40 °C/min to
290 °C, held for 4.00 min
70 °C held for 0.5 min, 10 °C/min to 130 °C, 40 °C/min to
305 °C, held for 4.00 min
305 °C
3 sec
0.6 sec
0.9 sec
The following MS conditions were used for detection.
MS filament delay:
MS scan range:
MS source:
Electron energy:
1.5 min
35-500, at a rate of 200 spectra/sec
250 °C
70 eV
3.4 GC/MS conditions (conventional quadrupole)
GC/MS analysis was performed with an Agilent 6890 GC coupled with an
Agilent 5973 MS (Agilent Technologies, Inc., Santa Clara, CA). Prior to use, the GC/MS
was tuned, with the vendor's standard protocols and PFTBA as a calibrant. An injection
of 50 ng DFTPP was used to check the performance of the instrument prior to analyzing
samples. Continuing calibration checks were also performed as prescribed by EPA
protocols during the course of run sequences.
The standard GC parameters were:
Carrier gas:
Injection mode:
Injector temperature:
Sample injection volume:
GC Column:
Column dimensions:
GC temperature program:
He (Ultrahigh purity, Air Products, Allentown, PA),
at a constant flow of 32 cm/s
splitless for 0.75 min
250 °C
1 |iL
Agilent HP-5MS UI, (5%-phenyl)methyl
polysiloxane
30 m x 0.25 mm x 0.25 |im (length x i.d. x film
thickness)
40 °C held for 3 min, 10 °C/min to 150°C, 25
°C/min to 280 °C, held for 10.8 min
The standard MS conditions for full scan analyses performed in electron
ionization mode were:
MS transfer line temperature: 280 °C
5
-------
MS source temperature: 230 °C
MS quadrupole temperature: 150 °C
Solvent delay time: 3 min
Scan range: 35-500 m/z
Electron energy: 70 eV
Scan time: 3.15 scans/sec
Ionization polarity: Positive
The standard MS conditions for selected ion monitoring analyses performed in
electron ionization mode were:
MS transfer line temperature: 280 °C
MS source temperature: 230 °C
MS quadrupole temperature: 150 °C
Electron energy: 70 eV
Ion dwell time: 100 msec per ion (each analyte was assigned its
own SIM group; depending on the number of ions
monitored, cycle times ranged from 1.44 - 2.86
cycles/sec)
Ionization polarity: Positive
3.5 Quality Assurance (QA)
Data limitations. The study was performed per the approved quality assurance
project plan (3); Tasks 6 and 7 of that study plan were completed and discussed in this
report. The LECO Pegasus® 4D GC/TOF-MS was used to detect CWAs. While data
presented in this report are only valid for the specific instrument cited, the applicability of
this work to mass spectrometers of other vendors to the analysis of CWA can be inferred.
However, instrument and method detection limits will vary and will be instrument-
specific.
Deviations from study plan. Analytes studied were HD, GB, GD, GF, and VX.
According to the original QAPP, tabun (GA) was also to be studied. When the QAPP
was written, it was thought that there was sufficient stock of existing GA of acceptable
purity to include in this study. As GA was not available at the time the study was
initiated, it was decided to study only the five CWAs being provided to the
Environmental Response Laboratory Network (ERLN) laboratories.
The original QAPP called for the use of 2 ng DFTPP to document the good
performance of both the GC/TOF-MS and the GC/MS. During this study, 10 ng DFTPP
were used to access the performance of the GC/TOF-MS and 50 ng DFTPP were used to
access the performance of the GC/MS. This change was made to be consistent with the
quality checks used with these instruments for other EPA-sponsored work that was
ongoing at the time of this study. The change in concentration did not affect the ability to
demonstrate that the mass spectrometers were in working condition prior to commencing
analyses.
6
-------
The column configuration used for this study was a primary column of 15 m x
0.25 mm id x 0.25 |im film thickness, HP5-MS UI coupled with a secondary column of 1
m x 0.1 mm id x 0.1 |im film, Rxi-17; this column configuration was used to provide
optimal separation of the CWAs studied.
4.0 Results
4.1 Chromatographic separation
In order to understand the data produced by 2D GC/TOF-MS, it is important to
have a clear picture of the separation process. A sample is initially separated on the
primary GC column (15 m x 0.25 mm i.d. x 0.25 |im film, HP5-MS UI). At this point,
the resulting chromatogram can be represented as a plot of total ion current versus time,
as is common practice in the field of chromatography; see Figure 1.
Using the GC/TOF-MS, with conditions as described in Section 3.3,
chromatographic analysis of CWAs was completed in 13 minutes. A 13-minute analysis
time represents a considerable (factor of two) time savings over the 30-minute analysis
required by conventional GC/MS with a 30 m x 0.25 mm id GC column. By using this
2D GC/TOF-MS method, a calculated throughput of 72 analyses per 24 hours could be
achieved. This analysis time includes the time necessary for both chromatographic
separations and for post-run cooling of the GC to its initial temperature, but does not
include data processing, which is considerably more labor intensive than both ID
GC/TOF-MS or GC/MS. During this same time period, only 35 samples could be
analyzed by conventional GC/MS.
"2 iO -
is ^ Is
c
as
jr
Q_
O
JZ
a.
c
O)
01
>.
~>-
_c
c
-------
The first step in performing 2D chromatography is to produce the best possible
chromatographic separation using the primary GC column (15 m x 0.25 mm i.d. x 0.25
|im film thickness, HP5-MS UI) to ensure optimal separation of the analytes on the
second GC column, which is of a different stationary phase (1 m x 0.1 mm i.d. x 0.1 |im
film thickness, Rxi-17) than the primary column. After initial separation on the primary
column, the effluent from that column enters a dual-stage thermal modulator. The dual-
stage thermal modulator of the Pegasus® 4D consists of a series of two cryotraps, which
are used to transfer analytes from the primary column to the secondary column. The first
cryotrap collects the sample from the primary GC column. Then, the sample is released
(using heat) and collected on the second cryotrap. This second cryotrap holds the sample
while the first cryotrap cools. When the first cryotrap is cold, the second cryotrap
releases the sample onto the second GC column. The second cryotrap then cools so that
the first cryotrap can release the next trapped sample portion. The continual process of
temperature cycling by which the thermal modulator traps effluent from the primary
column and introduces it onto the secondary column (referred to as "modulation") is done
at a frequency that preserves the resolution achieved on the primary column (i.e.,
typically 3 to 4 samples per first dimension peak width are collected; each of these
samples is referred to as a "slice") and considers the time needed for separation on the
secondary column (3 seconds in this method). With such a short separation time, the
separation on the secondary GC column is essentially isothermal (i.e., the temperature
changes of the primary oven do not significantly impact the higher temperature of the
secondary column). By performing the modulation process rapidly and performing a
quick secondary separation, resolution on the first column is maintained and two
independent separations are obtained from one injection.
During a 2D chromatographic analysis, there are repeated cycles of collecting
effluent from a primary column and separating that effluent on the secondary column.
For this reason, when the system's software collects and represents 2D GC/TOF-MS
data, the time that is required for modulation and the time that is required to perform the
second chromatographic separation must be considered. For this reason, a data sampling
period is equal to the period of modulation and the separation time on the secondary GC
column (typically 2-10 seconds). As a result, 2D chromatographic data are displayed as
discrete "slices", each corresponding to an independent separation on the secondary GC
column; see Figure 2. Figure 2 shows the extracted ion chromatogram of m/z 99 obtained
when separating 0.2 ng of each CWA with 2D GC/TOF-MS. In order to illustrate the
discrete nature of the data clearly, we chose to simplify the presentation by displaying
only data for m/z 99 (however, data were collected for the entire mass range of m/z 35 to
m/z 500). As the data in Figure 2 suggest, in order to represent the area of a single
analyte, the sum of all of the individual slices must be obtained (i.e., in 2D GC/TOF-MS,
what would be considered a single analyte peak in a ID separation is composed of
multiple slices when represented after 2D separation).
8
-------
CE
o
Q
O
u.
O
1st Time (s)
2nd Time (s)
199
1
298
2
400
0
—I—
499
1
598
2
—I—
700
0
I 99
Figure 2. Extracted ion chromatogram (m/z 99) obtained when separating 0.2 ng each
CWA with 2D GC/TOF-MS.
799
1
The chromatographic peaks, or slices, generated by 2D GC separation were
typically 100 ms wide. These peaks were much narrower than the approximately 2-
second peaks that were produced by the ID separation. For this reason, care was taken to
ensure that a minimum of 20 data points were generated across each chromatographic
peak in order to adequately define the peak and to obtain reproducible data. This was the
reason that a fast data acquisition speed of 200 spectra per second was needed to
successfully implement the GC/TOF-MS method. If too few data points were collected
during an analysis, the chromatographic peak would be ill-defined, chromatographic
information would be lost, reproducibility would be poor, and LECO's deconvolution
software algorithms would not have been properly implemented.
Replicate analyses of CWA standards performed with the 2D GC/TOF-MS
provided both reproducible retention times (RT) and responses, providing assurance that
the data acquisition rate was appropriate for the application. Table 1 shows the retention
times and analyte responses for seven replicate injections of a standard containing 0.2 ng
of CWAs. Note that, because 2D chromatography was performed, retention time
reproducibility in both dimensions could be determined. When describing 2D retention
times, it is common convention to use the format of RT = (ID elution time, 2D elution
time). For example, the retention time for GB would be reported as (112, 1.191). As is
evident from the data, for the first dimension of chromatographic separation, retention
times agreed well, with a relative standard deviation of less than 2% for all the CWAs.
The relative standard deviations of retention times for the second dimension of separation
also showed good agreement, as the relative standard deviations in all cases were less
than 3%. Likewise, analyte responses were reproducible; with all relative standard
deviations of less than 5%. These results demonstrate that operation of the GC/TOF-MS
in 2D-mode can provide comparable reproducibility to that demonstrated by the
GC/TOF-MS operated in the ID-mode and to GC/MS (1).
9
-------
Table 1. Average Retention Times (± Standard Deviations) and Average Analyte
Responses (± Standard Deviations) for Seven Replicate Injections of 0.2 ng of CWAs
into the 2D GC/TOF-MS
Analyte
ID
Retention Time
(sec)
2D
Retention Time
(sec)
Analyte Response
(peak area in arbitrary units)
GB
112 ±2.07
1.191 ±0.013
280555 ± 9979
GD
255 ±4.14
1.546 ±0.022
151407± 5483
GF
399 ±2.27
1.821 ±0.048
259259± 10866
HD
379 ±2.70
1.931 ±0.024
245416± 10404
VX
568 ±0.00
0.851 ±0.002
18085 ±758
4.2 Instrument detection limits
Instalment detection limits (DDLs) were determined by making successive
injections of individual standards of decreasing analyte concentrations until a signal-to-
noise ratio (S:N) of approximately 3:1-5:1 was obtained for the analyte peak of the
second confirmation ion present in the chromatogram (detection of an analyte required
the presence of the quantitation ion and two qualifying ions). The analyte mass at which
a S:N of 3:1-5:1 was obtained for the second qualifying ion of three successive injections
was reported as the DDL. Blank samples (e.g., clean solvent) were analyzed before the
determination of the final DDLs to ensure that carryover of higher concentrations of
analytes did not influence DDL determinations. As shown in Table 2, 2D GC/TOF-MS
DDLs ranged from 0.5-2.5 pg and were lower than those obtained by ID GC/TOF-MS,
except for VX. The lower detection limits for 2D GC/TOF-MS can be attributed to the
cryofocusing that occurs before introduction of the analytes into the second GC column
and to the narrower dimensions of the second column (0.1 mm i.d. versus 0.25 mm i.d.
for the primary column), which produces a sharper, narrower peak for a given amount of
analyte. As a result of the above process, signal enhancement is achieved by the
focusing of the sample in the modulator and its injection into the second GC column as a
very narrow band. Analyte slices that are multiple seconds wide when eluting from the
first GC column are detected as narrower peaks (50 to 500 msec wide) when eluting from
the second GC column. Because analyte mass remains constant, the narrowing of peak
width means that peak height must increase. Depending on the conditions, the height of a
peak could be increased more than 50-fold. Increased peak heights lead to increased
signal-to-noise ratios, which translate to lower detection limits being achievable with 2D
GC/TOF-MS. Note that this is a somewhat simplistic scenario, as we acknowledge that
S:N also depends on other factors, such as detector data acquisition rate, and also the
stationary phase chosen for the 2D separation (which might explain why the IDL for VX
in 2D was the same as it was in ID - e.g. the stationary phase chosen for the second
dimension of separation was not optimal for VX separation).
10
-------
Table 2. 2D GC/TOF-MS IDLs, Quantification (Quant) and Qualifying (Qual) Ions,
and Signal-To-Noise Ratios (S:N)
1DIDL
2D IDL
1st
2nd
Avg. S:N
TOF
TOF
Quant.
Qual.
Qual.
(n=3)
Analyte
(pg)
(pg)
Ion
Ion
Ion
2nd Qual. Ion*
GB
25
0.5
99
125
81
7.6
GD
5
0.5
99
126
82
3.9
GF
2.5
0.5
99
67
81
6.3
HD
2.5
0. 5
109
111
63
5.1
VX
2.5
2.5
114
72
127
5.0
* S:N values were determined by manual integration of 2D GC/TOF-MS data for the second qualifying
ion of the relevant CWA.
DDLs determined by 2D GC/TOF-MS and GC/MS (quadrupole) were also
compared; see Table 3. IDLs determined by 2D GC/TOF-MS were approximately 10
fold better than IDLs determined by GC/MS operated in the selected ion monitoring
mode.
Table 3. Comparative Data for Instrument Detection Limits (pg) Determined by
Various GC/MS Configurations
Analyte
GC/MS Configuration
2D
GC/TOF-MSa
ID
GC/TOF-MS b
GC/MS, FSb
GC/MS, SIMb
GB
0.5
25
200
10
GD
0.5
5
50
10
GF
0.5
2.5
200
20
HD
0.5
2.5
50
10
VX
2.5
2.5
200
50
a -This study
b - data are from previous study, reference 1
11
-------
4.3 VX calibration
One goal of this study was to determine if 2D GC/TOF-MS would eliminate
problems, such as anomalously high recoveries, encountered when quantifying VX using
quadrupole GC/MS. For this reason, 2D GC/TOF-MS studies focused exclusively on the
quantification of VX. A calibration curve for VX was established in a manner that was
consistent with instructions of the CWA protocol under development and EPA Method
8000 C (4) and used to measure the amount of VX in sample extracts. The
concentrations of VX in selected sample extracts were determined by 2D GC/TOF-MS
and, then, compared with the concentrations measured, for the same extracts, by
quadrupole GC/MS.
The first step in comparative analyses was calibrating the GC/TOF-MS for VX
measurement. When performing VX calibration, phenanthrene-dlO was used as an
internal standard. The calibration range was 0.1 to 2 |ag/m L and five points within this
range were used to define the calibration curve. The average of the relative response
factors (RRFs) over the calibration range was 0.102 and the percent relative standard
deviation (RSD) of the RRFs was 35.3%. Quantification of VX by RRFs could not be
used because the calculated RSD was 35.3%, a violation of EPA Method 8000C, Section
9.3.1, which states that "the criteria for linearity of an initial calibration curve based on
the average of the response factors is an RSD of <20% for each compound that is
included in the calibration." Instead, linear regression, which provided an R2 value of
0.99989 for five calibration levels, was used. This approach was deemed acceptable as
Method 8000C requires R2 > 0.99. Using this regression, the concentrations of the
individual VX standards were determined and compared with their known amounts. In
all cases, calculated concentrations were within 5% of their expected values (percent
differences over the calibration range were 0.02 - 4.7%). It should be noted that the
calibration range for VX was chosen to bracket the concentrations of VX expected to be
in the sample extracts. Note also that the concentrations of VX selected were sufficiently
high so as to be detected by GC/MS.
4.4 Measurements of VX in sample extracts
In previous studies, high recoveries of VX (>300%) have been observed when
performing experiments in which VX was spiked on and extracted from different
surfaces. It has been speculated that high VX recoveries could be attributed to interfering
compounds present in the sample matrices and/or to matrix enhancement effects (e.g.,
matrix components shielding active sites in the injection port or GC column of the
analytical system). Because of its greater chromatographic resolution over a single-
column GC separation, the use of 2D GC/TOF-MS offered the opportunity to investigate
the possibility of interference(s) that co-eluted with VX during a conventional GC
separation.
Selected problematic matrices (identical to those used in previous studies that
consistently showed recoveries of greater than 100% for VX) were spiked with VX (from
a dichloromethane solution), extracted using procedures described in Section 3.2, and
analyzed using the 2D GC/TOF-MS. Sample extracts were also analyzed by quadrupole
12
-------
GC/MS. The matrices studied included samples of spiked wipes, Virginia soil, and
drywall. VX concentrations determined by 2D GC/TOF-MS and GC/MS were
compared.
The first matrix examined was Virginia soil. Three 10-g aliquots of this soil were
spiked with 1 |ig of VX (i.e., 0.1 |ig VX per gram of soil), extracted, and analyzed by
both 2D GC/TOF-MS and GC/MS. The soil extracts were also analyzed by ID GC/TOF-
MS approximately three weeks after extraction. The VX concentration in soil as
determined by 2D GC/TOF-MS was 104 ng ± 4 ng (n=3 replicates) and by GC/MS (SIM)
was 147 ng ± 8 ng (n=3 replicates). Using the Student's t-test, at a significance level of
a=0.05,p value=0.11, the hypothesis that the VX concentrations measured by 2D
GC/TOF-MS and by GC/MS are equivalent is rejected (i.e., the concentration of VX in
Virginia soil measured by GC/MS is higher than that measured by 2D GC/TOF-MS).
One possible explanation for the higher VX concentration measured in the soil
extract analyzed by GC/MS might be the presence of an interference(s) that is not
encountered when using 2D GC/TOF-MS, which provides better chromatographic
resolution. Figure 3 shows a selected ion chromatogram for m/z 114 (quantification ion
for VX) produced by GC/MS (SIM) for an unspiked Virginia soil. This figure shows an
elevated baseline for m/z 114 (indicating possible contributions in ion signal from the
presence of other compounds); when the sample is examined by GC/MS, however, no
distinct interfering compounds are observed in the blank soil. In contrast, the results of
2D chromatographic separation of 0.1 |ig of VX in the Virginia soil extract (orange
trace), displayed in Figure 4, show the presence of at least three interferences —one with
a predominant ion of m/z 107 (green trace), one with a predominant ion of m/z 121 (blue
trace), and another with a predominant ion of m/z 135(red trace). While the raw spectrum
at the peak apex of VX contains ions representative of the matrix (top mass spectrum of
Figure 4), the deconvoluted spectrum (middle mass spectrum of Figure 4) shows a good
match to the reference spectrum for VX (bottom mass spectrum of Figure 4).
Using 2D GC/TOF-MS, mass spectra of these potentially interfering compounds
were obtained; see Figure 5. Figure 5 displays the 2D chromatogram of an unspiked
Virginia soil extract showing ions representing the matrix interferences for VX. No m/z
114 (VX quantitation ion) is present. The full mass spectrum corresponding to the
compound represented by the ion chromatogram of m/z 135 is shown as the top mass
spectrum of Figure 5. The full mass spectrum corresponding to the compound
represented by the ion chromatogram of m/z 121 is shown as the bottom mass spectrum
of Figure 5. We were unable to identify these compounds (although based on their mass
spectra, they appear to be aromatic compounds). Note that the separation on the
secondary column was sufficient (-100 ms separation from interference) to allow
LECO's deconvolution algorithm to separate the interference(s) from VX and to produce
a good mass spectrum for VX. In this case, 2D GC/TOF-MS provided more accurate
quantification of VX and was able to detect interferences that were not observed with
GC/MS. The presence of these interferences provides a partial explanation of why
higher VX concentrations are measured by GC/MS than are measured by 2D GC/TOF-
MS.
13
-------
Abundance
25000
20000
1 5000
1 OOQO
5000
Ion 1 1 4.00 [1 1 3.7U to 11 4.70): K0009090. D\data.ms
Ion 11 4. OO (1 1 3 7"U to 1 1 4 ^OJ: K0003031 . D \data. ms (")
1 ppm VX standard
Virginia soil method
blank
16. OO 16.50 17.00 17.50 18.00 1S.SO
T ime-->
Figure 3. Selected ion chromatogram of m/z 114 (VX quantitation ion) of Virginia
soil method blank (blue) and the overlaid chromatogram of m/z 114 for a 1 ppm VX
standard (black).
40000
35000
30000
25000
20000
15000
10000
5000
0
1st Time (s) 506,5
2nd Time (s) 2
- 114
506.5
121
107
- sarnfe "TOP: 1022*, 506.5, 2.435 sec(sec to 506.5 , 2.435 sec, sec - 90, 0.000 sec , sec to90, 0AM0sec ,
1M Raw Spectrum
55 gg | I 86
ittM#
135
121
¦J., i.ii.'. J*
149 167 181 196
40 60 80 100 120 140
>eak True • sarrtfe "TOP: 1022", p«<*. 17r at 506.5 , 2.435 sec , sec
1000*1
43
'
72
160 180 200 220 240
Deconvotuted
Spectrum
167
40 60
jtrarv Ht - sirriarty 876, "YX"
80 100 120 140
160 180 200 220
Reference Spectrum
°]
41
! I"
¦H K
*1»
JL* *
•n*-
100
127 139
140 160 180 200 220 240
135
Figure 4. 2D chromatogram of extract produced from 0.1 fig VX per gram Virginia
soil.
14
-------
Peak True - sample T0F:lQ2Cr, peak 19, at 506.5 , 2.2® sec , sec
0 -1
1st Time (s)
2nd Time (s)
40000 -
35000 -
30000 -
25000 -
20000 -
15000 -
10000 -
5000
506.5
2.1
506.5
2.3
506.5
2.5
n p if. n ,1 [Vi I • ¦
50 100 153 200
"
250 300 350 '00 IS] 500
¦' i i 1 -i i i
250 300 330 400 450 500
114
121
107
135
Va Soil Blank
Figure 5. 2D chromatogram of unspiked Virginia soil showing ions representing
matrix interferences for VX; no m/z 114 (VX quantitation ion) is present.
A better depiction of the interferences observed in the extracts of the Virginia soil
can be seen in contour plots from the 2D GC/TOF-MS data; see Figure 6. A contour plot
is a representation that displays ion intensities resulting from the first separation (HP-
5MS UI column) as a function of retention time (in seconds) on the x-axis and ion
intensities resulting from the second separation with the Rxi-17 as a function of retention
time (also in seconds) along the y-axis. In a contour plot, the color and intensity of a
peak (represented as a shaded area) is related to its amount, or concentration. In the plots
shown here, the intensity scale ranges from pale blue (low) to red (high). Black squares
serve as markers to emphasize specific compounds. In the total ion chromatogram (TIC)
contour plot of the Virginia soil method blank, shown in Figure 6 (top), there is a large
conglomeration of matrix compounds present, as well as some less intense matrix
compounds present in the surrounding area. Figure 6 (bottom) shows that this cluster of
peaks co-elutes with VX (present at a concentration of 0.1 |ig/g in the original soil
sample) after the separation on the primary GC column. Note that despite the coelution
problem after the first chromatographic separation, VX is successfully resolved from
these interferences after separation on the second GC column.
However, the coelution of the previously discussed interference would be a
problem only if the interference contains ions that hamper the quantification of VX. VX
quantification is based on the signal from the m/z 114 ion. The contour plot of the m/z
114 ion produced from a VX-spiked Virginia soil extract showed that the interferences do
contain ions at m/z 114 that coelute with VX after separation on the primary GC column;
see Figure 7. However, these interferences were completely resolved from VX (-100 ms)
when a second separation was performed, which allowed proper identification and
quantitation of VX. Figure 8 shows a magnified view of the contour plot for m/z 114
derived from VX-spiked (0.1 |ig/g) Virginia soil and that complete separation of VX
from the nearest interferent was achieved using the second dimension of separation.
15
-------
jg Masses: TIC
Unknown
<485.5 495.5 505.5 515.5 525.5
1DRTHP-5MSUI
M
lasses: TIC
2D RT Rxi-17
06 2.26 2.46 2.66
i i i i
^85.5 t
195.5 5
05.5 515.5 525.5
1D RT HP^SMS Ul
Figure 6. 2D GC contour plots of the TICs of the Virginia soil method blank (top)
and VX-spiked Virginia soil (bottom).
16
-------
Figure 7. 2D GC contour plots of ni/z 114 (VX quantification ion) of Virginia soil
method blank (top) and VX-spiked (0.1 fig/g) Virginia soil (bottom).
Interpretation notes: The x-axis (time in seconds) represents separation achieved after
chromatography occurred on the primary GC column (HP5-MS UI) and the y-axis (time in seconds)
represents the separation achieved on the secondary (Rxi-17) GC column. Each black dot represents a
different compound, or peak, detected by the LECO analysis software. Shaded areas represent individual
compounds on an intensity scale of pale blue (low) to red (high). The x-axis of the contour plot shows that
VX coelutes with 3 interferences (three black dots at RT = -506 sec, as indicated by the white arrow) as it
exits the first analytical column. However, the separation on the shorter, smaller-diameter second
analytical column easily resolves the VX peak from other three peaks of interfering compounds.
The 2D contour plots display the compounds as shaded areas based on the data intensity in
relation to the other compounds present within the particular viewing region; for this reason, no direct
comparisons between concentrations (i.e.. color intensities) of the method blank (top) and spiked soil
(bottom) can be made. Note also that the y-axes of the top and bottom figures are different; the range of the
y-axis shown for the Virginia soil blank (top) is narrower to emphasize the complex nature of the multiple
interferences observed near the VX peak.
17
-------
Figure 8. Magnified view of 2D GC contour plot for rn/z 114 of VX-spiked (0.1 jig/g)
Virginia soil.
The data collected from 2D-GC/TOF-MS showed the presence of compounds in
Virginia soil that would adversely affect VX quantification if only one dimension of GC
separation had been performed. These compounds contained ions of m/z 114, which
would contribute signal to the measured signal from the m/z 114 ion of VX. Thus, if
these compounds were not adequately resolved chromatographically from VX, then this
observation explains, in part, VX recoveries greater than 100%, which were often
observed with GC/MS analyses.
Next, the quantification of VX in a wipe sample extract was compared with
analysis by 2D GC/TOF-MS and by GC/MS. In this experiment, three separate wipes
were each spiked with 1 |ig of VX, extracted by protocols of the CWA-SAP, and
analyzed by both 2D GC/TOF-MS and GC/MS. The mass of VX per wipe was
determined to be 1.32 |ig ± 0.03 |ig (3 replicates) by 2D GC/TOF-MS and 1.69 |ig ± 0.04
|ig (n=3 replicates) by GC/MS. At a significance level of a=0.05,/> value=0.006, the
hypothesis that the VX concentrations measured by 2D GC/TOF-MS and by GC/MS are
equivalent is rejected (i.e., the mass of VX on wipes measured by GC/MS was higher
than that measured by 2D GC/TOF-MS).
Because the measured amount of VX determined by GC/MS was higher than that
measured by 2D GC/TOF-MS, the presence of an interference(s) was suspected.
Examination of the 2D contour plot for m/z 114 (quantification ion for VX) indicates the
presence of trace amounts of interfering compounds present after separation on the
primary GC column; see Figures 9 and 10. While these compounds were easily resolved
from VX using the separation afforded by 2D GC/TOF-MS, they would create problems
with quantitation if not adequately resolved when only one-dimensional GC separation is
18
-------
available. This might, in part, explain the recoveries of >100% which are sometimes
observed when known amounts of VX are spiked on wipes, extracted, and analyzed by
GC/MS.
Figure 9. 2D GC/TOF-MS contour plot for nt/z 114 (quantification ion for VX)
from an extract derived from a wipe spiked with 1 fig of VX.
Interpretation notes: Two dimensions of separation are shown, with the first dimension retention time (in
seconds) shown on the x-axis, resulting from separation with the HP-SMS UI column and the second dimension
retention time (also in seconds) along the v -axis. resulting from a second separation with the Rxi-17 column. The
intensity of the peak is represented on a scale of pale blue (low) to red (high). The red peak is VX. The contour plot
shows that two other compounds coelute with VX from the first column (notice the two black dots with the same
first dimension retention time as VX). However, these compounds are easily resolved from VX using the separation
afforded by the second column and. thus, they would not show a response at m/z 114 that would interfere with the
quantitation of VX if 2D GC separation were used.
Figure 10. 2D GC/TOF-MS contour plot for m/z 114 (quantification ion for VX)
from a wipe extract containing 1 VX.
19
-------
The last matrix examined was drywall, as previous work for the Department of
Homeland Security (data unpublished) showed exceptionally high recoveries of VX
(>300%) in drywall extract produced from the same lot of drywall studied here. Three
drywall coupons were spiked with l|ig of VX, extracted as previously described, filtered,
spiked with internal standard (at a concentration of l|ig per 1.00 mL sample extract), and
analyzed by both 2D GC/TOF-MS and GC/MS. Because the extraction of drywall
creates fine particles of drywall, filtration of the sample extract prior to its analysis is
needed. Two filtering strategies were tested - the first used filtration of the sample
extract with an Acrodisc filter (Pall 0.45 |im PTFE Acrodisc CR, P/N 4219T) attached to
a 10-mL syringe and the second provided filtration of sample extract with a Whatman
Autovial syringeless filter device (0.45|im, PTFE membrane with glass microfiber
prefilter with polypropylene housing, P/N AV125UORG). The total amounts of VX
measured in these sample extracts are shown in Table 4. A slightly higher recovery for
VX was obtained when the Whatman Autovial was used for filtering than when the
Acrodisc was used. Using 2D GC/TOF-MS, the average amount of VX in the sample
extracts prepared with the Whatman Autovial was 1.01 |ig ± 0.04 |ig (n=3 replicates) and
the average amount of VX in the sample extracts prepared with the Acrodisc was 0.88 |ig
± 0.02 |ig (n=3 replicates). At a significance level of a=0.05,p value=0.02, the
hypothesis that the VX amounts in the sample extracts prepared by the Whatman
Autovial and the VX amounts in the sample extracts prepared by the Acrodisc (both
measured by 2D GC/TOF-MS) are equivalent is rejected (i.e., the mass of VX in the
sample extracts prepared by the Whatman Autovial is higher than that of sample extracts
prepared by the Acrodisc). Using a similar statistical treatment, the GC/MS data
comparing the amounts of VX in sample extracts prepared using the differing filtering
strategies also showed that higher recoveries were obtained with the Whatman Autovial.
The amounts of VX determined in drywall extracts by 2D GC/TOF-MS and by
GC/MS for each filtering treatment were compared. When the Acrodisc was used, 0.88
|ig ± 0.02 |ig of VX (n=3 replicates) were measured by 2D GC/TOF-MS and 1.12 |ig ±
0.06 |ig of VX (n=3 replicates) were measured by GC/MS. At a significance level of
a=0.05,p value=0.005, the hypothesis that the VX amounts in the drywall samples
measured by 2D GC/TOF-MS are equivalent to those measured by GC/MS is rejected
(i.e., the mass of VX in the sample measured by GC/MS is higher than that measured by
TOF-GC/MS). The same conclusion was reached for VX amounts in the drywall
samples prepared with the Whatman Autovials, a=0.05,p value=0.0006 (i.e., the mass of
VX in the sample measured by GC/MS is higher than that measured by GC/MS). As
seen in the other matrices, the presence of interfering compounds that may not be
resolved from VX when using a single dimension of separation was observed; see Figure
11. Data indicate that both sample preparation procedures and the instrument analysis
method chosen affect the amount of VX detected in drywall.
In this study, we did not observe the exceptionally high (>300%) recoveries of
VX measured in previous studies. Recoveries of VX greater than 100% were observed
for all of the sample extracts that were measured by GC/MS and for the wipe extracts
measured by 2D GC/TOF-MS. The use of 2D GC/TOF-MS demonstrated that matrix
interferences were partially responsible for high VX recoveries. However, the fact that
20
-------
VX recoveries >100% were detected by 2D GC/TOF-MS suggests that a combination of
interferences and matrix enhancement effects is most likely responsible for high VX
recoveries.
Table 4. Measured Amount of VX (jig) (n=3) in Sample Extracts of Drywall
Coupons Spiked With 1 ng of VX
Filtering
Strategy
Spike Amount on
Drywall (jig)
GC/MS Configuration
2D GC/TOF-MS
GC/MS (quadrupole)
Acrodisc
attached to a 10-
mL syringe
1.00
0.88 ±0.02
1.12 ±0.06
Whatman
Autovial, 0.45
(imPTFE
membrane with
glass microfiber
prefilter
1.00
1.01 ±0.04
1.54 ±0.06
Figure 11. 2D GC/TOF-MS contour plot of ni/z 114 (the VX quantitation ion) for a
drywall extract, produced from 1 jig VX on spiked drywall, which was filtered using
Whatman Autovial.
21
-------
5.0 Conclusions and Recommendations
2D GC/TOF-MS was found to be a useful tool for the analysis of CWAs, and VX
specifically. Advantages of 2D GC/TOF-MS include low CWA detection limits (0.5 - 5 pg for
the CWAs studied) and the retention of complete mass spectral data for each compound detected,
which provides greater confidence in analyte identifications. Because of its fast data acquisition
rate, a TOF-MS is ideal to couple with two-dimensional GC separations, which offer the
opportunity to resolve interferences present in complex matrices from quantitation ions.
Potential disadvantages of 2D GC/TOF-MS, as compared to quadrupole GC/MS, include the
higher level of training needed for the instrument operator, the more complex instrument
hardware (i.e., the modulator and liquid nitrogen needed to provide 2D separations), the less-
developed data analysis software, and the greater initial cost of instrumentation ($215,000 versus
$85,000).
In this study, the quantification of VX was investigated using 2D GC/TOF-MS.
Quantification was found to be reproducible and linear from VX concentrations of 0.1 to 2
|ig/mL, 2D GC/TOF-MS consistently measured VX concentrations that were significantly lower
than those determined by VX for spiked soil, wipe, and drywall samples. Examination of 2D
contour plots for these samples showed the presence of interferences possessing ions of m/z 114,
which interfered with accurate VX quantitation (VX has a quantitation ion of m/z 114). VX
quantification was affected by interferences when separation on a single HP-5MS UI GC column
was performed. However, the use of the Rxi-17 column to provide a second (2D) GC separation
was shown to sufficiently resolve VX from these interferences. Thus, the recommendation is
that 2D GC/TOF-MS be used to detect and quantify VX whenever interferences are suspected
when only a single chromatographic separation is performed.
6.0 References
1. Experiments with the LECO Pegasus® Gas Chromatograph/Time-of-Flight Mass
Spectrometer Phase 1: Fast GC Separations and Comparison of the GC/TOF-MS with
Conventional Quadrupole GC/MS and Fast Quadrupole GC/MS, Rev. 1.0, Heather
Mulcahy, Carolyn Koester, and Roald Leif, March 11, 2010, LLNL-TR-420785.
2. Verification of Methods for Selected Chemical Warfare Agents (CWAs), Rev. 4, Heather
Mulcahy, Roald Leif, and Carolyn Koester, December 18, 2009, LLNL-TR-415101.
3. Quality Assurance Project Plan and Study Plan for Lawrence Livermore National
Laboratory's Experiments with the LECO Pegasus® Gas Chromatograph/Time-of-Flight
Mass Spectrometer for EPA under IAG #DW89922616-01-0, Revision 1.5, LLNL-TR-
408100, Carolyn Koester, January 23, 2009.
4. Method8000C: Determinative Chromatographic Separations, Rev. 3, March 2003, U.S.
Environmental Protection Agency.
22
-------
Appendix A: Autosampler Method for LECO 2D GC/TOF-MS
Select auto sampler type:
•8- Agilent©
O Rail System (CTC. Gerstel, LEAP )
O Shimadzu©
H 7890, S890N or 6890 with nanoliter adapter enabled
~ Enable overlapped operation when connected to an Agilent 7890 GC
.
Pumps ( 0-15
7 sec
i 0-15 >:
¦h f 0-15 ):
¦h (0-15
Pre-tnjection De!ay: 0.00-1.00 min >:
:
Post-lnjectiop. Deiay ( 0.00-1.00 min >:
fo I
Solvent A Post-Wash ¦ 0-15 j:
[3
Solvent B Post-Wash { 0-15 ):
3
Slow plunger enable:
0 Yes @ No
1 .v.: r t ..v: e
O Yes ® No
Syringe bize ( pL i
10 I
Sample Volume i j.
1
Number of Sample
; : .rJ:. :
0
Sample Pre-Wash
1
Solvent A Fre-Vv'as
2
Solvent B Pre-Wa
23
-------
Appendix B: GC Method for LECO 2D GC/TOF-MS
Hardware control;
O Agilent© 7890
©Agilent© 6890
O Shimadzu© GC-2010
O Generic
O Direct Inlet
~ MACH/LTM Oven
V LECO© GCxGG
- r- V ~::I ' ~ - ¦¦ .V.-l j:.-"
:::! I-J1. I* r1 ¦.v.: r ..v: •
Shimadzu.-;- GC-2010
:: :• J - ¦¦ : ¦
.1 :
No problems detected with column configuration.
Flow Path 1:
# -
Type ;
Location
Length(m)
Int. Diameter^;*
Max Temp
Film Thickness! Phase
Bleed Masses
r
(Inlet |
Back
2
Capillary
GC Oven
15.000
250.00
350.0
0
HP5MS
73 149 207 281
3
Capillary
Modulator
0,100
100.00
325^0
0.11.'
Rxi-17
73 149 207 281
4
Capillary
Secondary
1.000
100.00
325.0
aio
Rxi-17
73 149 207 281
5
Capillary
Detector o
0.210
100.00
325.0
0.10
Rxi-17
73 149 207 231
6
Detector
TOF
Add
Delete
~ Enable Flow Path 2
Promote
Demote
Copy
Paste
Mass Selection for Auto Mass Defect ; racking
8> = ¦: , :.t: ' '::3 :--f = ' - Mass Defect Mode
. t: :: ¦=¦*¦=¦:
v-fr" :L*
'2 .•
:r;x. interferii
:r - •-
- -: . : .
. -r t :
: ' n *:
anci non-tarqet masses.
w Included Masses in Auto Mass Defect Mode
•oensrah
tarqet anal-
: t : . - T
:¦
•js -f:;,
24
-------
Appendix B, continued: GC Method for LECO 2D GC/TOF-MS
Helium
-niei I yoe:
Split / Splitless
z:::. ¦ '* -r" " ': :--r"
Pulsed Splitless
O Front
• Back
No problems detected with pressure / flow.
0 Corrected constant flow via pressure ramps
Column 2 / Back Inlet flow(s):
# J Rate (mL/mir
ia) j T arget Flow (mL/min)
| Duration (mfn)
1" iInitial
11.20
[ Entire Run
-I
Column 2 •' Sack rr.ie: Purge Time i sec
The time, after trie beg'irvng of the run. wnen the purge valve ..ill
¦o;-> from the purge vent, i his value cannot be used if your
column Is not defined
" L • -r' ::j - : ~ * y i y ± jr.- - ¦:: ; : -f-= . ¦
31.2 |
Column 2 i Back Inlet Pu se Pressure t psi>:
Column 2 / Back Inset Puise Duration i minutes >:
0.5
25
-------
Appendix B, continued: GC Method for LECO 2D GC/TOF-MS
1
Back Inlet temperature(s):
# i Rate PC/mm)
I" u :iet Temp f C)
j Duration (min)
1* ilnitial
1250.00
| Entire Run
J
No problems detected with oven temperatures.
Over, tauKibratiori Tim# i minutes ):
0,5
Not®: AM even temperature ramps ¦; except the secondary oven >
.'.ill have the same duration. This is accomplished by extending
:-e ¦" - ::: ¦¦¦=¦
i temperature ramp below:
# 1 Rate (°C/rnin)
1 .-v .jet Temp fC>
Duration fmin)
1* llnitial
155.00
10.50
... ¦ 0.00
¦
3 140,00
1290.00
4,00
Add
Remove
On ® Off
v;isi>v tin
12
0Er.-.= ::
# | Rate CC/fnin)
1 Target Temp (°C)
Duration (min)
1* linitial
""""l 70..0
0.50
2 f 10,00
0.00
3 j 40.00
1 305 00
4,00
Add
Remove
26
-------
Appendix B, continued: GC Method for LECO 2D GC/TOF-MS
i ransfer Line Temperature Equilibration i tme i sec >:
0
305
WEnabled
Moouiaioi 7emperati
^^serature ii:
::
i i
s Start
End
Modulation Period (s)
Hot Pulse Time
[ Cool Time Between Stages
1* istart'ofRun
End of Run
3.00
0.60
0.90
¦=¦ :: "¦ ¦j >1 :.r-
I t .r . e ::¦¦¦=¦ :t: \ : -r- =
27
-------
Appendix C: MS Method for LECO 2D GC/TOF-MS
Use GC method total time for MS method total time:
Yes No
acquisition delay
gg # Sec. The length of time from injection unti! the data system
Q yjn '.'.ill start storing data from the mass spectrometer.
Enter time(s) when the filament should be turned off {mm of 3 sec} in the grid below
Start of Run
lEnd
Filament
J 90s
Off
End of Run
On
Add
Remove
Required Disk Space
Enter the mass spectrometer settings:
Start Mass i u ;
35
tnd Mass t u
500
Acquisition Rate ( spectra ' second ]
200
Detector Voltage
1650
Electron tnergy * Volts '¦
-70
Mass defect mode
O Auto (Select masses for automatic tracking in column information section of GC method.)
® Manual
0 Verify offset before collecting data
Mass Defect ( mu / 100 u )
-------
Appendix C, continued: MS Method for LECO 2D GC/TOF-MS
Set the temperature for the ion Source.
ii5Q
V . ji t •.T-:T--r...'T5 :: it3c;- t—: it before starting acquisition
Source Temperature Equilibration Time (Seconds)
-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |