vvEPA
EPA 600/R-13/044 | May 2013 | www.epa.gov/ord
United States
Environmental Protection
Agency
Stability Study for
Ultra-Dilute Chemical
Warfare Agent Standards
I 14 15 ta v
II » » ,o ' 1
Office of Research and Development
National Homeland Security Research Center
-------
EPA/600/R-13/044
May 2013
Stability Study for
Ultra-Dilute Chemical Warfare Agent Standards
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
i
-------
Disclaimer
The United States Environmental Protection Agency through its Office of Research and
Development funded and managed the research described here under Interagency Agreement
DW-89-92328201 and DW-89-92261601 with the U. S. Department of Energy (DOE). It has
been subjected to Agency's administrative review and approved for publication. The views
expressed in this paper are those of the authors and do not necessarily reflect the views or
policies of the United States government or Lawrence Livermore National Security, LLC.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use. 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.
Auspices Statement
This work performed under the auspices of the U.S. Department of Energy by Lawrence
Livermore National Laboratory under Contract DE-AC52-07NA27344; the Lawrence
Livermore National Laboratory report number is LLNL-TR-573732. The research team was
comprised of Roald Leif, Carolyn Koester and Heather Mulcahy.
Questions concerning this document or its application should be addressed to:
Romy Campisano (EPA Project Officer)
U.S. Environmental Protection Agency
National Homeland Security Research Center
26 W. Martin Luther King Drive, MS NG16
Cincinnati, OH 45268
513-569-7016
Campisano.Romy@epa.gov
li
-------
Acknowledgments
The research team wish to acknowledge the support of all those who helped plan and prepare
this report. The contributions of Don MacQueen, Lawrence Livermore National Laboratory,
who performed statistical analyses, are greatly appreciated.
111
-------
Abbreviations/Acronyms
BSTFA - N,0-bis(trimethylsilyl)trifluoroacetamide
CWA - Chemical Warfare Agent. The CWAs of interest in this report are HD, GB, GD, GF, and VX.
DBT - dibenzothiophene
DCM - dichloromethane
(DES)2 -bis[2-(diisopropylamino)ethyl] disulfide, formula Ci6H36N2S2
DESH - 2-(N,N-diisopropylamino)ethanethiol, formula C8H19NS
EMPA - ethyl methylphosphonic acid
EPA - United States Environmental Protection Agency
ERLN - Environmental Response Laboratory Network
FPD - flame photometric detector
GB - Sarin, O-isopropyl methylfluorophosphonate, formula C4H10FO2P
GC - gas chromatograph
GC-FPD - gas chromatography coupled with a flame photometric detector
GC/MS - gas ghromatography/mass spectrometry
GD - Soman, O-pinacolyl methylphosphonofluoridate, formula C7Hi6F02P
GF - Cyclosarin, cyclohexyl methylphosphonofluoridate, formula C7H14F02P
HD - Distilled sulfur mustard, bis(2-chloroethyl)sulfide, formula C4H8C12S
Hex - hexane
IMPA - isopropyl methyl phosphonic acid
LLNL - Lawrence Livermore National Laboratory
MS - mass spectrometer
NMR - nuclear magnetic resonance (spectroscopy)
OPCW - Organisation for Prohibition of Chemical Weapons
ppm - parts-per-million, equivalent to ng/mL or ng/|iL
Pyro A - 0,0-diethyl dimethylpyrophosphonate, formula C6Hi605P2
Pyro B - O-ethyl, O-isopropyl dimethylpyrophosphonate, formula C7Hi805P2
TMP - trimethylphosphate
VX - 0-ethyl-S-(2-diisopropylaminoethyl) methylphosphonothioate, formula CnH26N02PS
iv
-------
Executive Summary
The purpose of this project was to
determine the stability over time of the
ultra-dilute chemical warfare agent (CWA)
analytical standard solutions in
dichloromethane and hexane for the five
CWAs under normal conditions of storage
and use. Ultra-dilute (10 ppm) CWA
standards (i.e., analytical standard
solutions) are being synthesized by
Lawrence Livermore National Laboratory
(LLNL) and supplied to the Environmental
Response Laboratory Network (ERLN)
laboratories. The Environmental Protection
Agency established the national ERLN to
support large scale environmental
responses for chemical, biological, and
radiological threats during nationally
significant incidents. These standards are
provided as authentic standards for the
analyses of CWA contaminants that could
remain in the aftermath of a terrorist attack,
for the unambiguous identification and
quantification of CWAs, and for analytical
method development by the ERLN
laboratories. Currently, data on the long-
term stability of the ultra-dilute standards
are lacking. The shelf-life data collected in
this study is intended to be used to estimate
reliable shelf lives for the ultra-dilute
standards.
CWAs in single-component and five-
component standard solutions, containing
5-10 ppm each CWA in hexane and
dichloromethane (DCM), were studied.
The CWAs studied included sarin (GB),
soman (GD), cyclohexylsarin (GF), sulfur
mustard (HD), and 0-ethyl-S-(2-
diisopropyl-aminoethyl)
methylphosphonothioate (VX). The
specific objectives were the following:
1. Measure the stability, over the
course of 12 months, of single-
component ultra-dilute (10 ppm)
CWA standards that were stored
at 4 °C in the dark
2. Measure the stability over the
course of 12 months of
multiple-component ultra-dilute
(5-10 ppm) CWA standards
that were stored at 4 °C in the
dark
3. Determine if CWA stability
differs in single- and multiple-
component mixtures,
4. Determine if solvent choice
(hexane versus DCM) affected
shelf life
5. Compare the stabilities of CWA
standards stored in flame-sealed
amber glass ampoules versus
CWA standards stored in amber
glass vials closed with Teflon®-
lined, silicone septa screw caps
that were opened and closed
periodically
After the CWA standards were prepared in
the desired solvent, 1-mL aliquots were
placed in flame-sealed, amber glass
ampoules. A predetermined number of
these ampoules were set aside for the
stability study (i.e., the "sealed ampoule
standards"). A predetermined number of
the ampoules were immediately cracked
open and transferred to screw-cap vials.
Duplicate sealed ampoules were sampled at
predetermined times and then discarded.
Duplicate vials were sampled at
predetermined times, resealed, and stored
for the next sampling. All ampoules and
vials were stored at 4 °C ± 2 °C in a
refrigerator. CWA concentrations were
measured using a gas chromatograph
coupled with a flame photometric detector.
CWA concentrations were plotted as a
function of time over the course of one
year. Selected samples were also analyzed
by gas chromatography/mass spectrometry
v
-------
(GC/MS) to identify contaminants and
degradation products.
Estimated shelf lives for CWA standards
under different conditions were determined
using Dunnett's Test (Hsu 1996) to identify
the time point before which a statistically
significant decrease in analyte
concentration occurred; see Table 1.
Results for individual standards stored in
sealed ampoules showed that the CWA
standards made in DCM were stable for a
year. Individual standards of CWAs stored
in hexane and in sealed ampoules showed
varying stabilities; GD and GF were stable
for a year, HD was stable for six months,
VX was stable for three months, and GB
was stable for only two months (although at
six months, GB concentration had
decreased by only 15% of its initial value).
Table 1. Summary of stability study data and estimated shelf lives for standards stored at 4 °C.
Estimated Shelf Lives* for CWAs in Single-Component Solutions
(months)
Dichloromethane
H
exane
Screw-cap
Vial
Sealed Ampoule
Screw-cap
Vial
Sealed Ampoule
GB
5
12
4
2
GD
12
12
12
12
GF
12
12
12
12
HD
6
12
4
6
VX
2**
12
5
3
Estimated Shelf Lives* for CWAs in Multiple-Component Solutions
(months)
Dichloromethane
H
exane
Screw-cap
Vial
Sealed Ampoule
Screw-cap
Vial
Sealed Ampoule
GB
9
12
0.3
9
GD
12
6
0.7
9
GF
12
12
12
12
HD
6
6
3
6
VX
0.2
12
Notes: * Estimated shelf life is defined as the time point prior to that for which a statistically
significant decrease in concentration was detected by Dunnett's Test.
** Because large variabilities between replicate analyses were noted (relative standard
deviation >50%), shelf-life was based on best judgement rather than the results of the
Dunnett's Test.
Data for individual standards stored in vials that most CWAs were stable in both DCM
that were periodically reopened showed and hexane solutions for six months.
vi
-------
Exceptions to this period of stability were
GB, which was stable for only four months
in hexane and five months in DCM, and
VX, which degraded after five months in
hexane and after two months in DCM.
Overall, the CWA standards were more
stable in the unopened ampoules when
compared to the screw capped vials.
Based on the results of Dunnett's Test,
multiple-component standards stored in
sealed ampoules showed that all CWAs in
DCM and hexane were stable for at least
six months. However, the research team
believes that the statistical analyses
overestimate the stability of VX because of
the higher variabilities in the concentrations
measured for sample replicates (>50%
relative standard deviations amongst
replicate measurements in some cases). In
these cases, VX stability may be less than
two months in hexane and DCM.
Multiple-component standards stored in
opened vials showed decreased stabilities
for most CWA. GD and GF in DCM were
stable for 12 months. In hexane, GF was
also stable for 12 months, but GD was
stable for only 0.7 months (although at one
month, the measured GD concentration had
decreased by only 10% of its initial value).
GB in hexane was not stable for two weeks
(although GB in DCM was stable for nine
months). HD in DCM stored in an opened
vial was stable for six months, while HD in
hexane was only stable for three months.
Multiple-component standards in DCM and
hexane showed estimated stability times for
VX of 0.2 and 12 months, respectively.
Overall, the CWA standards were more
stable in the unopened ampoules when
compared to the screw-capped vials.
Because VX was most prone to
degradation, its breakdown was examined
in some detail. The breakdown of VX in
the VX-only standards was initiated by the
presence of ethyl methylphosphonic acid
(EMPA); 0,0-diethyl
dimethylpyrophosphonate was formed by
the reaction of EMPA with VX. In the
multiple-component standards, the
breakdown of VX was initiated by the
presence of isopropylmethylphosphonic
acid (IMPA), a trace contaminant present in
the GB stock solution. VX reaction with
IMPA was observed to produce O-ethyl, O-
isopropyl dimethylpyrophosphonate. As
the IMPA impurity provided another
pathway for VX degradation, VX stability
in multiple-component standards was less
than that observed when VX was present as
a single component in a solvent.
This shelf life study demonstrates that the
stabilities of the CWAs vary greatly
between compound classes (e.g., G-, H-
and V-agents), and VX stability can be
affected by the presence of other CWAs.
Currently, CWA standards are being
supplied to the ERLN laboratories as two
solutions - the first, 10 ppm VX in DCM
and the second, a mixture of GB (10 ppm),
GD (5 ppm), GF (10 ppm), and HD (5
ppm) in DCM. Based on the results of this
study, the research team recommends that,
for convenience in planning and as a simple
rule of thumb, all CWA standards, prepared
in DCM, in sealed ampoules be used within
six months of receipt and that, once opened
and mixed into five-component solutions
containing GB, GD, GF, HD, and VX, all
CWA standards be kept for no longer than
one week. The team further recommends
that future work be performed to determine
how VX can be stabilized in the ultra-dilute
standards. Such stabilization strategies
could include the removal of EMPA and
IPMA from the standard solution,
implementation of a stringent water
removal strategy and the use of a stabilizer
to prevent VX degradation.
vii
-------
Table of Contents
Disclaimer
Abbreviations/Acronyms
Executive Summary
Introduction
Materials and Methods
Standard Preparation and Storage
Ampoule Sealing
Instrumentation
Analytical Procedure
Quantitation of Target Analytes
Results and Discussion
Single-Component Standards
Multiple-Component Standards
Comment on VX Stability (with regards to the ultra-dilute standards in dichloromethane) .
Conclusions
References
Appendix A: Mass Spectra of VX Degradation Products
.. ii
. iv
.. v
.. 1
.. 3
.. 3
..4
..4
..4
.. 5
.. 7
.. 8
14
18
26
29
30
viii
-------
List of Figures
Figure 1. Timeline for standard preparation, ampoulation or ampoulation.
Figure 2. Concentration (ppm) overtime (months) for chemical warfare agent single-
component standards in dichloromethane in reopened vials 9
Figure 3. Concentration (ppm) over time (months) for chemical warfare agents in
dicloromethane as a single-component standard in sealed ampoules 11
Figure 4. Concentration (ppm) over time (months) for chemical warfare agents present in
hexane as a single-component standard, in reopened vials 12
Figure 5. Concentration (ppm) over time (months) for chemical warfare agents present in
hexane as a single-component standard in sealed ampoules 13
Figure 6. Concentration (ppm) over time for chemical warfare agents present in
dichloromethane as a multiple-component standard, in reopened vials 15
Figure 7. Concentration (ppm) over time for chemical warfare agents present in
dichloromethane as a multiple-component standard in sealed ampoules 16
Figure 8. Concentration (ppm) over time (months) for VX in a multiple-component standard in
dichloromethane, in sealed ampoules 17
Figure 9. Concentration (ppm) over time (months) for chemical warfare agents present in
hexane as a multiple-component standard, in reopened vials 19
Figure 10. Concentration (ppm) over time (months) for CWAs present in hexane as a multiple-
component standard in sealed ampoules 20
Figure 11. GC-FPD chromatograms (both S and P channels) of single-component VX
standards in dichloromethane after one year of storage, at 4 °C, in a sealed ampoule (top) and
in a Teflon-lined, septum-capped vial (bottom) 22
Figure 12. GC-FPD chromatograms of multiple-component chemical warfare agent standards
in dichloromethane after one year of storage, at 4 °C, in a sealed ampoule (top) and in a Teflon-
lined, septum-capped vial (bottom) 24
IX
-------
Introduction
The purpose of this project was to determine
the stability over time of the ultra-dilute
CWA analytical standard solutions in
dichloromethane and hexane for the five
chemical weapons agents (CWAs) under
normal conditions of storage (i.e., 4 °C) and
use. Ultra-dilute (10 ppm) chemical warfare
agent (CWA) standards are being supplied
by Lawrence Livermore National
Laboratory (LLNL) to the Environmental
Response Laboratory Network (ERLN)
laboratories to allow the use of authentic
standards to assist in analyses required in
remediation scenarios involving CWAs.
The Environmental Protection Agency
established the national ERLN to support
large scale environmental responses for
chemical, biological, and radiological threats
during nationally significant incidents. It is
critical for the ERLN laboratories to be able
to work with authentic CWA standards to
allow the unambiguous identification and
quantification of CWAs. The ultra-dilute
standards are synthesized by LLNL for use
in analytical method development by the
ERLN laboratories. However, data
regarding the long-term stability of the ultra-
dilute standards are lacking. The data
collected in this study are intended to be
used to determine reliable shelf lives for the
ultra-dilute standards.
Shelf life for analytical standards is defined
as the length of time a standard can be
stored, from initial preparation to final use,
without significant changes to the original
analyte concentrations. The shelf life may
be affected by solvent type and storage
conditions. Analytical standards are used to
create the calibration curves that are used to
determine the concentrations of target
analytes in authentic samples. To generate
accurate calibration curves (and, therefore,
measure target analyte concentrations), it is
essential that the exact concentrations of
analytes in ultra-dilute standards be known.
If one adheres to the observed shelf lives
and conditions, one can reasonably expect
that the analyte concentrations in a standard
will remain at their specified values. For
this reason, it is important to have guidelines
to indicate how long and under what
conditions an analytical standard can be
stored and used before its concentration
changes significantly, adversely affecting
the ability to provide valid data.
The ultra-dilute CWA standards studied
were sarin (GB), soman (GD), cyclosarin
(GF), sulfur mustard (HD), and O-ethyl S-
[2-(diisopropylamino)ethyl]
methylphosphonothioate (VX). These
analytes were studied as single-component
standards and in a multiple-component
standard solution containing GB, GD, GF,
HD, and VX. The five-component standard
is desirable as a working standard because it
minimizes the number of separate CWA
standard solutions needed for the
quantitation of multiple CWA analytes. For
the ERLN laboratories, these five
compounds were initially combined in the
same working standard at concentrations of
5 to 10 |ig/mL (ppm). As part of this study,
the chemical stabilities of these five CWAs
were investigated in solutions prepared in
both dichloromethane (DCM) and hexane.
As noted above, the primary goal of this
study was to collect CWA stability data to
provide guidance for establishing shelf lives
for the ultra-dilute CWA standards, so that
the ERLN laboratories could establish
expiration dates for their ultra-dilute CWA
standards. Individual standards and
combined standard solutions kept at a
storage temperature of 4 °C were evaluated
and the concentrations were measured for a
period of one year. Standards were stored in
both sealed ampoules and in screw-capped
vials to assess the effect of reopening the
vials on the stability of the standards. A
secondary goal was to determine if CWA
standards could be provided to the ERLN
-------
laboratory as multiple-component standards single- and multiple-component standards
or if they needed to be provided as single- were tested for stability,
component standards. To that end, both
2
-------
Materials and Methods
To determine CWA stability and shelf lives,
aliquots of single- and multiple-component
standards were analyzed at various time
intervals and the concentration of each
CWA in the solutions was quantified.
Standard Preparation and Storage
CWAs were synthesized at LLNL. The
standards were checked for purity by nuclear
magnetic resonance spectroscopy (NMR)
and by gas chromatography/mass
spectrometry (GC/MS). Stock solutions of
each agent, at a concentration of 1000
|ig/mL, were gravimetrically prepared from
the purified neat agents in both
dichloromethane, DCM (Riedel-de Haen,
GC-grade, Lot # 7284M, with amylene as a
stabilizer) and hexane (Fluka, >99.0%
purity, GC-grade, Lot# 1351600, Filling
Code 1407326). Working standards were
prepared in both DCM and hexane from the
respective stock solutions by volumetric
dilution. The final concentrations of CWA
in solution reflected those currently shipped
to the ERLN laboratories. The
concentration of CWA in each single-
component standard was 10 |ig/mL. The
concentrations of CWAs in the mixed
standards were 10 |ig/mL for GB, 5 |ag/m L
for GD, 10 |ig/mL for GF, 5 |ig/mL for HD,
and 10 |ig/mL for VX. These
concentrations were derived from initial
experiments which considered the analyte
responses of the CWAs when analyzed by
GC/MS. All standards were flame-sealed in
ampoules two days after preparation; this
allowed a day to perform analyses to verify
that the concentrations of the CWAs were
correct and that no significant contaminants
were present in the standard solutions.
Standards were stored either in flame-sealed,
amber glass ampoules (P/N 176796,
Wheaton Science Products, Millville, NJ) or
amber glass vials, closed with Teflon®-
lined, silicone septa screw caps (P/N 5182-
0556, Agilent Technologies, Santa Clara,
CA). For the sealed ampoule standards,
aliquots of the prepared standards were
transferred to amber glass ampoules and
flame-sealed; the procedure is described in
the following section. The t=0 ampoules
were then immediately analyzed and the
remaining ampoules were immediately
refrigerated. A sufficient number of
standards in ampoules were prepared such
that duplicate sealed ampoules could be
sampled at predetermined times and then
discarded.
As indicated previously, all standards,
including those that would be stored in
screw-capped vials, were initially
transferred to flame-sealed ampoules. It was
deemed to be important to first transfer the
standards to flame-sealed ampoules to
accurately reproduce the procedure by
which the standards were prepared and
shipped to the ERLN laboratories (i.e., all
standards are placed in flame-sealed
ampoules prior to shipping to the
laboratories and being opened for the first
time). In addition, the attempt was made to
reproduce any changes that might occur to
the standards during the flame-sealing
process (e.g., introduction of water into the
standards). As soon as a batch of ampoules
were flame sealed, the ampoules were
opened and the aliquots of standards were
immediately transferred to screw-cap vials.
As soon as practical, the t=0 vials were
analyzed and the remaining vials were
immediately refridgerated until later
analysis. The vials were sampled at
predetermined times, resealed, and stored
for the next sampling.
3
-------
All standards were refrigerated (4 °C ± 2 °C)
until analysis to mimic storage conditions
expected to be used by the laboratories. The
refrigerator was equipped with a
thermometer (Item # 20700T, H-B
Instrument Company, Collegeville, PA),
which was visually checked periodically
during the course of the study. Had the
temperature of the refrigerator exceeded the
range of 2 °C to 6 °C at any point during the
study, the standards would have been moved
to a different refrigerator. Such temperature
excursions did not occur.
Ampoule Sealing
One-milliliter aliquots of the working
standards were transferred to 2-mL,
prescored, amber, borosilicate ampoules
using a variable volume pipettor tipped with
a long Pasteur pipet. The ampoules were
used as received from the vendor. The
ampoules were loosely covered with a
septum, while their headspace was flushed
with argon. The ampoules were then placed
in liquid nitrogen (to freeze the solvent and
prevent its evaporation) prior to flame
sealing. Flame sealing was done using an
Ampulmatic® automated ampoule sealing
device (Bioscience, Inc., Allentown, PA),
using a propane/oxygen flame. The
procedures used to seal the ampoules were
identical to those used by LLNL to seal
ampoules of ultra-dilute standards prior to
sending them to the ERLN laboratories.
Instrumentation
The analyses of the CWAs were performed
by gas chromatography coupled with a
flame photometric detector (GC-FPD) using
an Agilent 6890 gas chromatograph (GC)
equipped with an HP-5ms column (30 m x
0.25 mm i.d. x 0.25 |im film thickness). The
GC oven was heated using the following
program: isothermal for 1 min at 40 °C, 15
°C/min to 300 °C, and held isothermal for 1
min, with the injector and detectors at 250
°C, and helium at 1.4 mL/min as carrier gas.
The detector, a dual wavelength FPD,
enabled both sulfur (S) - and phosphorus (P)
- containing analytes to be quantified in the
same GC run. Instrument check samples
consisting of malathion, trimethylphosphate
(TMP), and dibenzothiophene (DBT) were
analyzed to assess GC-FPD response and to
perform calibrations. This calibration
standard has been used for several years to
test GC-FPD operation during Organisation
for Prohibition of Chemical Weapons
(OPCW) proficiency tests and has been
determined to have a long shelf-life (> 5
years). TMP and malathion were used to
test the response of the P-channel and DBT
and malathion were used to test the response
of the S-channel.
Analytical Procedure
Direct analysis of the 10 |ag/m L standards
could not be done on the GC-FPD because
the detectors would saturate at this level, so
the standards were diluted to a suitable level
prior to injection. For all of the individual
standard solutions, aliquots of the original
solutions were transferred to new vials and
diluted with equal volumes of the same
solvent prior to analysis. Because the five-
component mixes were prepared with
components at both 5 and 10 |ig/mL, the
multiple-component standard mixes were
analyzed both undiluted (to quantify the
components originally present at 5 |ig/mL)
and diluted with equal volumes of the
appropriate solvent (to allow the analysis of
components originally present at
concentrations of 10 |ig/mL).
Individual standards were analyzed on at the
start of the experiment (Day/Month 0) and
on Days 14 (Month 0.5), 27 (Month 1), 56
(Month 2), 82 (Month 3), 111 (Month 4),
140 (Month 5), 171 (Month 6), 273 (Month
9), and 365 (Month 12).
-------
Multiple-component standards were
analyzed on at the start of the experiment
(Day/Month 0) and on Days 9 (Month 0.3),
22 (Month 0.7), 37 (Month 1), 65 (Month 2),
93 (Month 3), 124 (Month 4), 148 (Month
5), 177 (Month 6), 285 (Month 9), and 386
(Month 13).
The timeline and sequence of events for
standard preparation, ampoulation or
ampoulation followed by transfer into a
glass vial, and initial analysis is shown in
Figure 1.
standard placed in
ampoules/vials,
immediately analyzed or
immediately stored at 4 :C
standard
preparation
1
analysistimes
i
t
12
7
standard
QA checks
10
Time (months)
Figure 1. Timeline for standard preparation, ampoulation or ampoulation
followed by transfer into a glass vial, and analysis.
Duplicate or triplicate ampoules/vials were
analyzed on each sampling day. Each
analysis that was performed over the
duration of the study for standards in sealed
ampoules was accomplished using a new,
freshly-opened ampoule. Analyses for vials
were performed using aliquots that had been
repeatedly opened and closed during the
course of the study; this was done to
simulate how standards would be stored in
vials and used by the CWA laboratories.
The specific numbers of replicates that were
analyzed during each time series are
provided in the paragraphs discussing
Figures 2-10.
Quantitation of Target Analytes
Quantitation was performed by external
standard method, using a seven-point
calibration curve. The calibration
compounds used for this study were TMP,
DBT, and malathion. These calibration
compounds are also part of a multi-
component standard that has been used for
over five years for instrument evaluation in
5
-------
support of another LLNL project and found
to be stable during this time. As these
compounds were known to be stable in that
test solution, it was assumed that the
calibration standard would be stable for the
one-year duration of this study. A stock
solution (2 mg/mL each analyte) of
calibration standard was made, divided into
aliquots, and the aliquots were sealed in
multiple ampoules at the beginning of this
study. Then, at the designated sampling
points, the required dilutions for the
calibration solutions were made from newly-
opened calibration stock solutions. Because
analyte responses obtained by GC-FPD for
standards in hexane and responses obtained
for standards in DCM had been
demonstrated to be different during the
initial phase of this study, calibration curves
were made using the solvent system of the
samples to be measured. TMP, DBT, and
malathion were selected for use as external
standards, permitting the analytes of interest
to have chromatographic retention times of
0.70 to 1.70 relative to one of the external
standards. GB, GD and GF, which all
contain a phosphorus atom, were analyzed
by P-FPD and were quantified using the
TMP calibration curve. Also using P-FPD,
VX was quantified using the malathion
calibration curve. HD, which contains a
sulfur atom, was detected by the S-FPD and
was quantified using the DBT calibration
curve.
Seven calibration levels were used for
quantitation, covering a range from 0.5
|ig/mL to 6 |ig/mL (ppm) so that the analyte
concentrations would fall within the
calibration range. All CWA standards to be
measured were analyzed at 5 jag/m L
(assuming 100% recovery of their known
t=0 concentration). The response of the
FPD was linear in the phosphorus mode.
The G-agents and VX were therefore
quantified using linear regression calibration
curves. In sulfur mode, the response of the
FPD was non-linear. A power series fit was
therefore used to characterize the S-FPD
response for the given calibration range. It-
squared values for all calibration curves
were >0.99 and continuing calibration
standards showed that the standard
responses remained within ±20% of their
expected values during analyses.
6
-------
Results and Discussion
The concentrations of CWAs were measured
at various times for single-component and
multiple-component standards stored at 4 °C
± 2 °C to determine whether degradation of
the CWAs occurred during the course of the
study. The standards were made in both
hexane and DCM. DCM was one solvent
of interest because ultra-dilute CWA
standards are currently prepared and shipped
to the ERLN laboratories in DCM.
Standards were also prepared in hexane, as
hexane may be a solvent of choice for future
use. Standards were stored in both sealed,
amber glass ampoules and in amber glass
vials with Teflon-lined, silicone septa screw
caps. The sealed glass ampoules represent
how the ultra-dilute CWA standards are
currently shipped to the ERLN laboratories
and how the standards would be stored prior
to opening. Screw-capped vials represent
the storage conditions of working solutions
in use by laboratories.
The initial concentration of CWA in each
single-component standard was 10 |ig/mL.
The initial concentrations of CWAs in the
mixed standards were 10 |ag/m L for GB, 5
|ig/mL for GD, 10 |ag/m L for GF, 5 |ag/m L
for HD, and 10 |ig/mL for VX. These
concentrations (referred to as "storage
concentrations") were derived from initial
experiments which considered the analyte
responses of the CWAs when analyzed by
GC/MS. Before analysis, the standards were
diluted to ensure that the linear range of the
GC-FPD was not exceeded. All single-
component standards were diluted by a
factor of two prior to analysis (i.e., yielding
an initial concentration of 5 ppm; the result
of this dilution is referred to as the "analysis
concentration"). The multi-component
standards were analyzed twice - the first
time they were analyzed without dilution to
measure GD and HD concentrations and the
second time they were analyzed at a twofold
dilution to quantify GB, GF, and VX. The
results of the analyses are presented in
Figures 2-10 as graphs of concentration
versus time for each CWA under varying
conditions. Note that the starting
concentration for each of the CWAs is 5
ppm, the target concentrations (t=0) at
which all of the CWAs were analyzed. Data
were displayed as analysis concentrations,
rather than storage concentrations, so that all
data could be displayed with the same y-
axis. Decreasing trends in CWA
concentrations with time will be apparent
regardless of the absolute magnitude of the
concentrations plotted.
The data were examined using Dunnett's
Test (Hsu, 1996). Dunnett's Test allows the
comparison of the means of several
experimental groups with the mean of a
control group in an analysis of variance
setting. In this study, one key experimental
factor is elapsed time, for which the
outcome of interest is a decrease in
concentration relative to the initial
concentration. Because there was no a priori
basis for any particular pattern of decrease
(e.g., linear, logarithmic) as a function of
time, the initial measurements (t=0) were
considered to be a control group, and
Dunnett's Test was used to examine whether
any subsequent time points had lower
average concentrations (a one-sided
statistical test of the null hypotheses that all
subsequent time point averages are greater
than or equal to that at the initial time). This
test was performed separately within each
combination of agent, solvent, storage
container (e.g., opened vial or sealed
ampoule), and standard type (i.e., single- or
mixed-component). Within each such
combination, Dunnett's Test was performed
at a significance level of alpha = 0.05, so
7
-------
that the probability of incorrectly declaring
any statistically significant differences
among the multiple comparisons with
control was 5% overall. In the subsequent
discussion, the phrase "statistically
significant" refers to rejection of the null
hypothesis under the conditions described
above.
Single-Component Standards
Single-component standards in
dichloroinethane, stored in screw-capped
vials.
Figure 2 shows the concentrations (ppm) as
a function of time (months) for the
individual CWA standards that were
prepared in DCM, sealed in ampoules, then
reopened and stored at 4 °C in vials that
were closed with Teflon-lined screw-caps.
Each point represents the average measured
concentration from two different vials (only
one vial was analyzed on t=0.5 months, or
14 days), and the error bars represent
plus/minus (±) the standard deviation of the
measurements. No statistically significant
loss in GB was observed until midway
through the study, at Month Six, when GB
was measured at 89% of its original
concentration. A gradual drop in GB
concentration was observed at the nine- and
twelve-month time points; the average GB
concentration was 66% of its starting
concentration after twelve months of storage
in a screw-capped vial. No losses were
observed for GD and GF during the one-
year period. HD was found to relatively
stable. The final average concentration of
HD represented an 81% recovery after one
year. Low concentrations for HD (that were
statistically significant) were measured at
the nine-month time point. These
abnormally low values were consistent with
chromatographic problems experienced at
that time and may be outliers due to poor
chromatographic conditions in the GC-FPD
that adversely affected the response of HD.
VX did not show a statistically significant
concentration decrease until Month Nine.
However, by Month Three, the average
concentration of VX dropped to 74% of its
initial concentration and continued to
decrease for the remainder of the study.
8
-------
GB in DCM, Single Standard, Open Vial
~~~ * ~ * i
2 4 6 8 10
Elapsed Time (months)
GD in DCM, Single Standard, Open Vial
3 4 ^ ~
4 6 8 10
Elapsed Time (months)
HD in DCM, Single Standard, Open Vial
6
- 5*
3 4
~ ~ **
4 6 8 10
Elapsed Time (months)
GF in DCM, Single Standard, Open Vial
~ ~ ~ ~ «
4 6 8 10
Elapsed Time (months)
VX in DCM, Single Standard, Open Vial
E 5
~ ~ ~
Q. 5
3 4
» .
£
° 3
•
« 2 -
I
?, 1
i
u
E 0 -
1 T
u
2 4 6 8 10 12 14
Elapsed Time (months)
Figure 2. Concentration (ppm) over time (months) for chemical warfare agent single-component standards in
dichloromethane in reopened vials
9
-------
Examination of the data in Figure 2 shows
large standard deviations in the average VX
measurements made for many of the
duplicate samples. Such large variabilities
in the data (approximately ±50% for the last
six timepoints of the study) hinder the
ability of the statistical analysis to discern
decreases in VX concentrations (i.e.,
decreases in VX concentration over time
actually occurred, but could not be detected
by Dunnett's Test). Possible reasons for the
large variations in measured VX
concentrations for replicate samples are
discussed later in this section (refer to
discussion of Figure 8).
Single-component standards in
dichloromethane stored in sealed ampoules.
Figure 3 shows the concentrations as a
function of time for the individual CWA
standards that were prepared in DCM,
sealed in ampoules, stored at 4 °C, then
opened and analyzed at the designated time.
Each point represents the average measured
concentration from two different ampoules
and the error bars represent ± the standard
deviation of the measurements. Overall, no
statistically significant losses of CWAs were
observed in any of these individual
standards. As observed in the previous data
set, low concentrations for HD were
measured at the nine-month time point; this
low concentration was attributed to
chromatographic problems experienced at
the time of analysis and, thus, this data point
was deemed to be an outlier.
Single-component standards in hexane,
stored in screw-capped vials.
Figure 4 shows the concentrations as a
function of time for the individual CWA
standards that were prepared in hexane,
sealed in ampoules, then opened, analyzed,
and stored at 4 °C, in vials closed with
Teflon-lined screw caps. Each point
represents the average measured
concentration from two different vials (only
one vial was analyzed at t=nine days) and
the error bars represent ± the standard
deviation of the measurements. The
concentration of GB was relatively stable,
with no statistically significant loss detected
until Month Five. After this time point, the
concentration of GB steadily decreased to
40% of its initial value after one year. No
statistically significant losses were observed
for GD and GF over the period of one year.
No statistically significant loss of HD
occurred until Month Five, when the HD
concentration was approximately 80% of its
initial value. For VX, no statistically
significant change in concentration was
observed until Month Six. The final
measured VX concentration at Month 12
represents approximately 50% of the initial
concentration.
Single-component standards in hexane,
stored in sealed ampoules.
Figure 5 shows the concentrations as a
function of time for the individual CWA
standards that were prepared in hexane,
sealed in ampoules, stored at 4 °C, then
opened and analyzed at the designated
times. Each point represents the average
measured concentration from two different
ampoules and the error bars represent ± the
standard deviation of the measurements. The
trends observed for this set of samples
match those of the previous set of opened
hexane standards shown in Figure 4. For
GB, a significant loss in concentration was
observed by Month Three (although the
decrease in GB concentration was only
20%). By one year, the GB concentration
dropped 26% from its initial value. No
statistically significant losses were observed
for GD and GF over the one year time
period. HD was stable for six months. A
steady and statistically significant decrease
in VX concentration was observed,
beginning at Month Four. The final VX
concentration at Month 12 was 2.69 |ig/mL,
representing a 47% drop from the starting
concentration.
10
-------
GB in DCM, Single Standard, Sealed Ampoule
— 6
E c
*—* *-
O 3
Elapsed Time (months)
GD in DCM, Single Standard, Sealed Ampoule
— 6
§ ,
Elapsed Time (months)
HD in DCM, Single Standard, Sealed Ampoule
o 3
Elapsed Time (months)
GF in DCM, Single Standard, Sealed Ampoule
— 6
§ ,
~ ~ ~
Elapsed Time (months)
VX in DCM, Single Standard, Sealed Ampoule
o 3
Elapsed Time (months)
Figure 3. Concentration (ppm) over time (months) for chemical warfare agents in dicloromethane as a single-component
standard in sealed ampoules.
li
-------
GB in Hex, Single Standard, Open Vial
^ 4
O 3
to 2
01
Elapsed Time (months)
GD in Hex, Single Standard, Open Vial
— 6
I 5
O 3
re 2
£ 1
3 o
~ f_
4 6 8 10
Elapsed Time (months)
HD in Hex, Single Standard, Open Vial
r
£
r
f A
Q.
' ~
~
C
I
£
0)
c
o
u
2 4 6 8 10
Elapsed Time (months)
12
14
GF in Hex, Single Standard, Open Vial
— 6
I 5
¦a 4
.1 3
2
? i
8 o
~ * ~ « *
Elapsed Time (months)
VX in Hex, Single Standard, Open Vial
£
r
r
~
T
Q.
t * ~
C
i
*
£
c
0)
c
o
u
2 4 6 8 10
Elapsed Time (months)
12
14
Figure 4. Concentration (ppm) over time (months) for chemical warfare agents present in hexane as a single-component
standard, in reopened vials.
12
-------
GB in Hex, Single Standard, Sealed Ampoule
. * ~ #
"£ 5
Q. q
O
re
c 1
g 0 -
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
GD in Hex, Single Standard, Sealed Ampoule
¥ 5
Q. q
n ^
• —
^0
1- .
c 1 "
g 0 -
O 0 2 4 6 8 10 12 14
Elapsed Time (months)
HD in Hex, Single Standard, Sealed Ampoule
. > . ~
S 5
~ ~ ~ ~ ~
Q. q
O
re
c 1
g 0 -
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
GF in Hex, Single Standard, Sealed Ampoule
- *
£ 5
>4* f ~ ~ * ~ ~
Q. q
O
re
c 1
g 0 -
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
VX in Hex, Single Standard, Sealed Ampoule
. * *
£ 5
* * « * f *
Q. q
O
*
re
c 1
g 0 -
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
Figure 5. Concentration (ppm) over time (months) for chemical warfare agents present in hexane as a single-component
standard in sealed ampoules.
13
-------
Multiple-Component Standards
Multiple-component standards in
dichloromethane stored in screw-capped
vials.
Figure 6 shows CWA concentrations as a
function of time for the ultra-dilute CWA
standard mix, containing GB, GD, GF, HD,
and VX, that was prepared in DCM, sealed
in ampoules, then opened and stored at 4 °C
in vials sealed with Teflon-lined screw caps.
Each point represents the average measured
concentration from three different vials
(only two vials were analyzed at t=0) and
the error bars represent ± the standard
deviation of the measurements. No
statistically significant loss in GB was
observed until the twelve-month time point,
where only 72% of the original
concentration of GB remained. No
statistically significant losses were observed
for either GD or GF for the duration of the
year-long study. For HD, only 44% of the
original concentration remained after one
year. A gradual statistically significant
decrease in HD began after Month Six. VX
was found to be quite reactive when
combined with the other CWAs and stored
in screw-capped vials. Almost 20% of the
VX was lost after only nine days. VX
continued to degrade throughout the study
period, with approximately 95% of the VX
lost after one year.
Multiple-component standards in
dichloromethane stored in sealed ampoules.
Figure 7 shows the CWA concentrations as
a function of time for the ultra-dilute CWA
standard mix that was prepared in DCM,
sealed in ampoules, stored at 4 °C, and
opened immediately prior to analysis. Each
point represents the average measured
concentration from two different ampoules
(three ampoules were analyzed at t = nine
days) and the error bars represent ± the
standard deviation of the measurements. No
statistically significant losses were observed
for either GB or GF for the duration of the
one year period. GD showed a statistically
significant loss after nine months and HD
was stable for only the first six months of
the study.
No statistically significant VX losses were
observed over the course of the study.
However, when the VX concentrations for
duplicate samples were averaged, large error
bars were observed and the power of the
statistical test to detect concentration
decreases was diminished (refer to previous
discussion). Upon examination of the
individual data points representing VX
concentrations, some of the ampoules
contained VX at its original concentration
while others contained lower concentrations
of VX (with decreasing VX concentrations
as study time increased); see Figure 8. For
example, at the four month time point, the
two replicate samples happened to be two
ampoules where VX did not degrade. At the
six month time point, the two replicate
samples happened to be two ampoules
where VX did degrade. The other time
points, with the exception of the initial
measurement at t = 0, consist of one
ampoule containing undegraded VX and the
other ampoule containing VX that had
undergone a substantial amount of loss. The
distribution between ampoules containing
degraded VX and those containing degraded
VX appears to be evenly split and random.
Because of the good precision of the VX
measurements at t=0 and t=4 months and
because the calibration curves for VX were
successfully constructed during the course
of this study, the VX concentration
differences between ampoules were
attributed to VX degradation and were not
attributed to inherent problems in the
reproducibility of the measurements.
14
-------
GB in DCM, Mixed Standard, Open Vials
6
~ i ~ i
!:E • •
•= 3
to
£ 2
Elapsed Time (months)
GD in DCM, Mixed Standard, Open Vials
£ 2
c
l
u
O 0
Elapsed Time (months)
HD in DCM, Mixed Standard, Open Vials
£ 2
-+—*-
~ ~
Elapsed Time (months)
GF in DCM, Mixed Standard, Open Vials
£ 2
I
Elapsed Time (months)
VX in DCM, Mixed Standard, Open Vials
£ 2
-f—f
_~ t_
Elapsed Time (months)
Figure 6. Concentration (ppm) over time for chemical warfare agents present in dichloromethane as a multiple-
component standard, in reopened vials.
15
-------
GB in DCM, Mixed Standard, Sealed Ampoules
~ ~
*- 2
¦M
Elapsed Time (months)
GD in DCM, Mixed Standard, Sealed Ampoules
Q.
' ~
~
E
O
~
•M
CD
C
u
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
HD in DCM, Mixed Standard, Sealed Ampoules
0 2 4 6 8 10 12 14
Elapsed Time (months)
GF in DCM, Mixed Standard, Sealed Ampoules
£ 5 -
,. ~ i ~
Q-
~ * ~ ~ +
.2
£ 2
c 1
c
o
u
2 4 6 8 10 12 14
Elapsed Time (months)
VX in DCM, Mixed Standard, Sealed Ampoules
~
l
•
1
•
0 2 4 6 8 10 12 14
Elapsed Time (months)
Figure 7. Concentration (ppm) over time for chemical warfare agents present in dichloromethane as a multiple-
component standard in sealed ampoules.
16
-------
VX in DCM, Mixed Standard, Sealed Ampoules
6.00
5.0d
4.0C
3.00
C
CU 2.00
u
c
<3 1-00
0.00
~ ~
10
Elapsed Time (months)
12
14
Figure 8. Concentration (ppm) over time (months) for VX in a multiple-component standard in dichloromethane, in
sealed ampoules.
17
-------
Multiple-component standards in hexane
stored in screw-capped vials.
Figure 9 shows the CWA concentrations as
a function of time for the ultra-dilute CWA
standard mix that was prepared in hexane,
sealed in ampoules, opened, then stored at 4
°C in vials sealed with Teflon-lined screw
caps. Each point represents the average
measured concentration from three different
vials (two vials were analyzed at t = 0) and
the error bars represent ± the standard
deviation of the measurements. GB was not
stable in this sample set. By three weeks,
statistically significant decreases in GB
concentrations were detected and subsequent
data collected at other time points revealed a
steady loss of GB. Only 10% of the GB
remained in the mixed standard in hexane
after a year. GD showed a significant
decrease in concentration at Month One.
However, at this time point, only an 8%
decrease in the initial GD concentration was
observed. No statistically significant loss of
GF was observed for the duration of the one-
year period. HD was stable for first three
months, but approximately 40% of the HD
was lost during the remainder of the study.
No statistically significant VX degradation
occurred during this study; 80% of the initial
VX remained after twelve months.
Multiple-component standards in hexane.
stored in sealed ampoules.
Figure 10 shows CWA concentrations as a
function of time for the ultra-dilute CWA
standard mix that was prepared in hexane,
sealed in ampoules, stored at 4 °C, and
opened immediately prior to analyses. Each
point represents the average measured
concentration from two different ampoules
(three vials were analyzed at t = 9 days) and
the error bars represent ± the standard
deviation of the measurements. No
statistically significant losses were observed
for either GB or GD for the first nine
months, but by the one year time point a
50% drop in GB concentration occurred and
a 30%) drop in GD concentration occurred.
GF was stable during the course of the study
and no loss was detected. No statistically
significant loss was observed for HD during
first six months of the study, but a 30%> loss
of HD had occurred by the one-year mark.
Loss of VX became statistically significant
after Month Six, with a 60%> loss of VX
measured at the study's end.
Comment on VX Stability (with regards
to the ultra-dilute standards in
dichloromethane)
VX was the most susceptible CWA to loss
during storage. This loss was attributed to
chemical degradation. VX was also the
analyte that the EPA's CWA laboratories
had found most prone to degradation (from
discussions during several of EPAs CWA
Protocol Teleconferences). Initially, LLNL
had supplied VX to EPA's CWA labs as a
multiple-component standard containing
GB, GD, GF, HD, and VX in DCM. After
the first CWA laboratories began working
with the multiple-component standards, they
too observed VX losses. LLNL therefore
began shipping VX to the CWA laboratories
as a single-component solution in DCM.
GB, GD, GF, and HD were shipped to the
CWA laboratories as a four-component
mixture, also in DCM. Below is a
discussion of the degradation issues that
were observed with ultra-dilute VX
standards. The discussion focuses on the
VX standard in DCM because the ultra-
dilute VX standard that is currently being
shipped to the ERLN laboratories is diluted
in this solvent.
18
-------
£ ¦;
GB in Hex, Mixed Standard, Open Vial
* i
Elapsed Time (months)
GD in Hex, Mixed Standard, Open Vial
5
Elapsed Time (months)
HD in Hex, Mixed Standard, Open Vial
£ 5 *
* *
4
~ ~ *
c
~
E ?
< ~
£ l
5 1
5 o -
0 2 4 6 8 10 12 14
Elapsed Time (months)
GF in Hex, Mixed Standard, Open Vial
E 5 ~
Elapsed Time (months)
VX in Hex, Mixed Standard, Open Vial
_ 6
Elapsed Time (months)
Figure 9. Concentration (ppm) over time (months) for chemical warfare agents present in hexane as a multiple-
component standard, in reopened vials.
19
-------
GBin Hex, Mixed Standard, Sealed Ampoule
6
1 5 4
•B 4
c
.2 3
+J
2 2
£ i
u
5 0
4 6 8 10
Elapsed Time (months)
GD in Hex, Mixed Standard, Sealed Ampoule
6
I 5 i
4 6 8 10
Elapsed Time (months)
HD in Hex, Mixed Standard, Sealed Ampoule
*
6
£ , L
4 6 8 10
Elapsed Time (months)
GF in Hex, Mixed Standard, Sealed Ampoule
6
"£
4 6 8 10
Elapsed Time (months)
VX in Hex, Mixed Standard, Sealed Ampoule
^ 6
E
*
~ T
4 ~ ~ ~ I
4 6 8 10
Elapsed Time (months)
Figure 10. Concentration (ppm) over time (months) for CWAs present in hexane as a multiple-component standard in
sealed ampoules.
20
-------
Some of the individual solutions of 10 ppm
VX in DCM that were stored in sealed
ampoules were stable for the duration of the
study (see Figure 3). This same VX solution,
when stored in screw-capped vials and
subjected to periodic openings at the
designated analysis times, experienced
statistically significant loss of VX after nine
months (see Figure 2). Using gas
chromatography/mass spectrometry
(GC/MS), the main breakdown products
observed in the ultra-dilute VX standards
were 2 -(dii sopropyl amino)ethanethi ol
(DESH), 0,0-diethyl
dimethylpyrophosphonate (Pyro A), and
/7/.s [2-(dii sopropyl ami no)ethyl] disulfide
[(DES)2], an oxidation product formed by
dimerization of two DESH molecules.
Derivatization of the VX standard using
N,0-/7/.s(tri methyl si lyl)-trifluoroacetarnide
(BSTFA) confirmed the presence of ethyl
methylphosphonic acid, another compound
produced from the hydrolysis of VX
(Buckles et al., 1977; Yang et al., 1996;
Yang, 1999). Mass spectra of the VX
degradation products have been compiled in
Appendix A.
The GC-FPD chromatograms of Figure 11
show the presence of the Pyro A degradation
product (doublet at 16.7 min) of VX (peak at
21.1 min) in one-year-old standards. Under
the separation conditions used (temperature
program of 40 °C for 3 min, ramped at 8
°C/min to 300 °C, and held at 300 °C for 3
min, with helium as a carrier gas at a
constant flow of 1.4 mL/min, with HP-5ms
GC column, 30 m x 0.25 mm i.d. x 0.25 |im
film thickness), the Pyro A was
chromatographically resolved into two
peaks, each representing stereoisomers
which arise because of the two stereogenic
phosphorous centers (Benschop and De
Jong, 1988). While other groups have also
observed this analyte as a VX degradation
product, Pyro A is usually reported as a
single peak in the literature because of the
particular GC conditions used. Separation
strategies used by other workers (e.g., larger
diameter GC columns, faster oven
temperature ramps) have produced
separations with lowered chromatographic
resolution. Other experimenters were
therefore not able to observe the
stereoisomers (e.g., Brevett et al., 2008;
D'Agostino et al., 1987; Rohrbaugh, 2000).
Symmetrical pyrophosphonates are known
to be formed through the breakdown of VX
(Yang et al., 1996) and present as impurities
of and/or formed during the breakdown of
nerve agents (Kumar et al., 2008).
Pyro A was formed by the reaction of ethyl
methylphosphonic acid (EMPA) with VX;
see Scheme 1. EMPA is a well-known
hydrolysis product of VX (Rohrbaugh
1998). It was confirmed, experimentally,
that Pyro A was produced by reaction of VX
and EMPA by preparing a DCM solution
containing both VX and EMPA, each at 50
ppm and then monitoring the solution
composition over the course of several days.
Pyro A was observed to be produced in this
reaction mixture.
Although Pyro A (doublet at 16.7 min)
appears in the chromatogram of the
individual VX standards that experienced
some degradation, another doublet (at 16.9
min), representing a second VX degradation
product, appeared in the five-component
solutions; see Figure 12. (GC conditions for
this analysis were: 40 °C for 3 min, ramped
at 8 °C/min to 300 °C and held at 300 °C for
3 min, with helium as a carrier gas at a
constant flow of 1.4 mL/min, with a HP-
5ms GC column, 30 m x 0.25 mm i.d. x 0.25
|im film thickness.) This compound was
identified as O-ethyl, O-isopropyl
dimethylpyrophosphonate (Pyro B) — a
pyrophosphonate formed by the reaction of
VX with isopropyl methyl phosphonic acid
(IMPA), a contaminant present in the GB
solution. This unsymmetrical
pyrophosphonate was unique to the
multiple-component standard and was not
present in the standards that contained VX
21
-------
only. Scheme 2 shows the reaction pathway
for the production of Pyro B,
S-channel
amco
VX
IflPOMO
DESH
J2!t_ <» «« tffi igw 12» lt» 149?
(DES)j
I
aoGccoo
P-channel VX
4000000
iccococ
2000000
Pyro A
1
T ^ n i | | J
r™ <¦» 4-6C eco IMP 1JL« M.« »,» nm HI 00 iZW MB)
S-channel
Signal: FPO 22S7.D>.FPD1B.CH
(DES)2
DESH
*
VX
Time 4.00
Response_
4500000
•1'JOKO'j
S5OCC00
3000000
;l=ij::::
200000Q
' •jlJCOlVj
IQOQOOO:
500000
8 00 10.00 12.00 14.00 16.00 1B.0O 20.00 22.00 24.00 26.00
P-channel
Siffiai: FPD-.2267_tKFP02A.eH
Pyro A
VX
10.M 12.00 14.00 16,00 18,00 2000 22,00 24,00 26.00
Figure 11. GC-FPD chromatograms (both S and P channels) of single-component VX standards in dichloromethane after
one year of storage, at 4 °C, in a sealed ampoule (top) and in a Teflon-lined, septum-capped vial (bottom).
22
-------
Scheme 1. Autocatalytic degradation mechanism for VX initiated by EMPA.
VX
HS J
T
DESH
[°]
DESH
(DES)2
OH
EMPA
0 O
1 I
PYROA
H,0
2 O—OH
I
EMPA
Scheme 2. Autocatalytic degradation mechanism for VX initiated by IMPA.
0
-P-s
1 v
„JL
VX
hs^n^
/k
DESH
[O]
DESH
Y
T~
.X
(DES)2
PYROB
O-P-OH
EMPA
Scheme 1
EMPA - Ethyl methylphosphonic acid
IMPA - Isopropyl methylphosphonic acid
Pyro A - 0,0-Diethyl-
dimethylpyrophosphonate
Pyro B - O-Ethyl, O-isopropyl-
dimethylpyrophosphonate
DESH - 2-(N,N-diisopropylamino)ethanethiol
(DES)2 - kis\ 2-(diisopropylamino)cthyl |
disulfide
23
-------
Signal: FP0-22S7.[W=PD1B.CH
300000Q
2500000
2CCOMO
1500000
1D00D00
500000
S-channel
HD
vx
6,00 8.06 10.00 12.00 14,00 15.00 18.00 20.00 22.C0 24.00 26.00
Signal: FPD-2237.CnfPD2ACH
P-channel
GF
VX
Pyro B
*
Time 4.M 6.00 8.00 10.00 12.00 14.K) 16.00 18.00 20.00 22.K) 24.00 26.00
RespKise_
2500000
Signal: FPD-22fi0.D\FPD1B.CH
HD
S-channel
DESH
(DES)2
|
VX +
Time 4 CO 6,00 8,00 1000 12.00 14.00 16.00 18.00 20.00 22X0 24.00 28.00
Response, Sigwl: FPO-22SO.[XfPD2ACH
5000000 _ , ,
P-channel
GB
GD
GF
Lil
Pyro A
Pyro B
i/
VX
6,00 10.00 12.00 14.00 18.00 18.00 20.00 22.00 24.D0 26.00
Figure 12. GC-FPD chromatograms of multiple-component chemical warfare agent standards in dichloromethane after
one year of storage, at 4 °C, in a sealed ampoule (top) and in a Teflon-lined, septum-capped vial (bottom).
24
-------
It was confirmed that the breakdown of VX
in the multiple-component standards was
initiated by the presence of IMPA by
performing a simple experiment in which
VX and IMPA, both at 50 ppm in DCM,
were allowed to react, at room temperature,
over the course of several days. Pyro B was
produced by this reaction. This reaction
provides an explanation for the faster
degradation of VX in the multiple-
component standards. In the multiple-
component standards, the initial attack of
VX by IMPA initiates the autocatalytic
breakdown cycle of VX (e.g., Yang et al.,
1996) and, as this cycle proceeds, the
production of both pyrophosphonate
compounds occurs; see bottom
chromatogram of Figure 12 and also Scheme
2. A review of the GC/MS data of the
freshly-prepared (t = 0), five-component
standard in DCM showed traces of Pyro B;
however, Pyro A, formed through the
autocatalytic breakdown cycle of VX
(Buckles et al., 1977; Yang et al., 1996), did
not appear until after the formation of Pyro
B.
The above discussion helps explain the
observations from this study. IMPA was not
present in the VX-only standard and,
therefore, VX was stable in a sealed
ampoule. However, the individual VX
standards exhibited loss of VX after being
opened to the atmosphere, perhaps through
hydrolysis from trace amounts of water
introduced during storage and handling. The
solvents used for the preparation of all the
standards were ultrapure solvents but not
anhydrous, and no additional water removal
was done prior to standard preparation.
Because these CWA standards contained
analytes at 5 or 10 |ag/m L (ppm), the trace
water present in the solvents was likely
present at concentrations much higher than
those of the CWAs and, therefore, at a
stoichiometric excess. For the solvents used
in this study, vendor labeling gave an upper
limit of 500 ppm water in hexane and 200
ppm water in the DCM (determined by Karl
Fischer titration). In such a situation,
hydrolytic degradation, once initiated, may
proceed to completion (Yang et al., 1996;
Brevett et al., 2009).
This study suggests that there are several
factors affecting the stability of VX in dilute
solutions. As discussed above, VX can react
with impurities and water present in the
standards. As the standard solutions age,
increasing amounts of water might also be
introduced into the solution from the
ambient environment, shortening the shelf
life of the standard. Clearly, potential
factors affecting the stability of VX should
be systematically studied and strategies for
providing ultra-dilute standards with longer
shelf lives should be studied.
25
-------
Conclusions
Statistically significant changes (as
determined by Dunnett's Test and described
in the third paragraph of the "Results and
Discussion" section) in concentrations
occurred with some CWAs during
refrigerated storage (4 °C ± 2°C) and the
stabilities of the CWAs were compound-
dependent. Results for individual standards
showed that all analytes (GB, GD, GF, HD,
and VX) were stable for 12 months when
prepared in DCM and stored in sealed
ampoules. GD and GF, as individual
standards, were stable for 12 months under
all of the conditions studied (i.e., both in
hexane and DCM in sealed ampoules and
screw-capped vials). Overall, the CWA
standards were more stable in the sealed
ampoules, when compared to the screw-
capped vials. Table 2 summarizes the study
findings and provides estimated analyte
shelf life. The estimated analyte shelf life is
defined as the time point before which the
decrease in analyte concentration was
determined to be statistically significant, by
Dunnett's Test, from its initial concentration
and was followed by other statistically
significant decreases in concentration in
subsequent month(s). For compounds that
showed no degradation, 12 months, which
represented the project duration, is used as
the default shelf life.
During the course of this study, VX
exhibited the most degradation. Anecdotal
evidence also suggested that the ERLN
laboratories observed VX degradation in the
CWA standards that were shipped to them.
These observations illustrate the reactivity
of VX and the potential difficulty of
preparing VX standards that will remain
stable over an extended period of time. VX
stability depends on the presence and,
presumably, also the concentrations, of
impurities (some of which can arise from
other CWAs). Future work could be
performed to determine how VX can be
stabilized in the ultra-dilute standards.
Stabilization strategies might include: (a)
the removal of EMPA and IMP A from the
solution, which would prevent the
mechanisms shown in Schemes 1 and 2, (b)
implementation of a stringent water removal
strategy, and (c) the use of a stabilizer to
prevent VX degradation (Buckles et al.,
1977). Of these strategies, the last two
might be the most feasible to implement
because the research team has detected
EMPA in newly-synthesized VX, even after
a washing process should have removed it;
(IMPA is an impurity of GB and only an
issue if VX is shipped in a mixture
containing GB).
The results of this study may be used to
guide the procurement and replacement
schedules of ultra-dilute CWA standards
used by the ERLN Laboratories. Currently,
all ultra-dilute CWA standards shipped to
the ERLN laboratories are made in DCM.
VX, at 10 ppm, is made as a single-analyte
solution in DCM; data from this study
suggest that VX in this standard remains
stable for one year. Once opened and/or
combined with other CWAs (as is frequently
done when GC/MS calibration curves are
made), this VX standard should be used or
replaced within one week. The remaining
CWAs, including GB (10 ppm), GD (5
ppm), GF (10 ppm), and HD (5 ppm), are
shipped as multiple-component standards,
also in DCM, to the ERLN laboratories.
While the multiple-component CWA
standard used in this study, which contained
VX, is different from the multiple-
component CWA standard currently being
shipped to the ERLN laboratories, the data
collected may still be used to provide
stability guidance. The reasonable
expectation is that the multiple-component
CWA standards will remain stable for six
months in sealed or opened ampoules.
However, if these standards were shipped as
26
-------
single-component solutions in DCM, they
could be kept for one year. For convenience
in planning and as a simple rule of thumb,
the research team recommends that all CWA
standards in sealed ampoules be used within
six months of production and that, once
opened, all CWA standards that do not
contain VX be kept for no longer than six
months and VX-containing mixed standards
be used within a week (individual VX
standards stored in screw capped vials can
be stable for up to six months).
27
-------
Table 2. Summary of stability study data and estimated shelf lives when standards are stored at 4 °C.
Estimated Shelf Livesa'b for CWAs in Single-Component Solutions
(months)
Dichloromethane
Hexane
Screw-cap Vial
Sealed Ampoule
Screw-cap Vial
Sealed Ampoule
GB
5 (0%)
12 (+6%)
4 (-6%)
2 (-4%)
GD
12 (0%)
12 (+9%)
12 (-4%)
12 (-9%)
GF
12 (0%)
12 (+2%)
12 (-11%)
12 (-9%)
HD
6 (+3%)
12 (+11%)
4 (+4%)
6 (+18%)
VX
2 (-3%)°
12 (+1%)
5 (-15%)
3 (-13%)
Estimated Shelf Livesa'b for CWAs in Multiple-Component Solutions
(months)
Dichloromethane
Hexane
Screw-cap Vial
Sealed Ampoule
Screw-cap Vial
Sealed Ampoule
GB
9 (+16%)
12 (+5%)
0.3 (-4%)
9 (-8%)
GD
12 (+1%)
6 (+21%)
0.7 (-0.6%)
9 (+8%)
GF
12 (-5%)
12 (+9%)
12 (-10%)
12 (+14%)
HD
6 (-7%)
6 (+13%)
3 (-3%)
6 (+15%)
VX
0.2 (-16%)
<1 (-11%)C
12 (-11%)
<1 (0%)c
Notes: a Estimated shelf life is defined as the time point prior to that for which a statistically significant decrease
in concentration was detected by Dunnett's Test.
Numbers in parentheses give the percent change in concentration from t = 0 to the concentration
measured at the estimated shelf life time; note that, on average, the percent standard deviation for
replicate measurements made at the shelf life time agreed within 5% (range of 0.4% to 24%, excluding
VX datapoints indicated with "c").
° Because large variabilities between replicate analyses were noted (relative standard deviation >50%),
shelf-life was based on best judgement rather than the results of the Dunnett's Test.
28
-------
References
Benschop H. P. and De Jong L. P. A. (1988)
Nerve agent stereoisomers: Analysis,
isolation and toxicology. Acc. Chem. Res.
21(10), 368-374.
Brevett C. A. S., Maclver B. K., Sumpter K.
B. and Rohrbaugh D. K (2008) SSMASNMR
Study ofHD, GD, and VX on Carbon Fiber
Textiles for Wipes. Report ECBC-TN-035,
Edgewood Chemical Biological Center,
Aberdeen Proving Ground, MD.
Brevett C. A. S., Sumpter K. B., Pence J.,
Nickol R. G., King B. E., Giannaras C. V.
and Durst H. D. (2009) Evaporation and
degradation ofVXon silica sand. J. Phys.
Chem. 113(16), 6622-6633.
Buckles L. C., Lewis S. M. and Lewis F. E.
(1977) S-(2-diisopropylamino-ethyl) O-ethyl
methylphosphonothiolate stabilized with
soluble carbodiimides. United States Patent
4,012,464.
D'Agostino P. A., Provost L. R. and
Visentini J. (1987) Analysis of O-ethyl S-[2-
(diisopropylamino)ethyl]
methylphosphonothiolate (VX) by capillary
column gas chromatography-mass
spectrometry. J. Chromatogr. 402, 221-232.
Kumar R., Pardasani D., Mazumder A. and
Dubey D. K. (2008) Microwave induced
synthesis of 0,0-dialkyl
dialkylpyrophosphonates under solvent free
conditions: Markers of nerve agents. Aust.
J. Chem. 61, 476-480.
Rohrbaugh D. K. (1998) Characterization of
equimolar VX-water reaction product by gas
chromatography-mass spectrometry. J.
Chromatogr. 809, 131-139.
Rohrbaugh D. K. (2000) Methanol chemical
ionization quandrupole ion trap mass
spectrometry of O-ethyl S-[2-
(diisopropylamino)ethyl]
methylphosphonothiolate (VX) and its
degradation products. J. Chromatogr. A
893(2), 393-400.
Yang Y.-C. (1999) Chemical detoxification
of nerve agent VX. Acc. Chem. Res. 32(2),
109-115.
Yang Y.-C., Szafraniec L. L., Beaudry W.
T., Rohrbaugh D. K., Procell L. R. and
Samuel J. B. (1996) Autocatalytic hydrolysis
ofV-type nerve agents. J. Org. Chem.
61(24), 8407-8413.
Hsu, J. C. (1996) Multiple Comparisons,
Theory and Methods (Chapter 3), Chapman
& Hall, NY (ISBN 0 412 98281 1).
29
-------
Appendix A: Mass Spectra of VX Degradation Products
The mass spectra and retention times
presented here were collected using an HP-
5ms column (30 m x 0.25 mm i.d. x 0.25 |im
film thickness). GC conditions for this
analysis were: 40 °C for 3 min, ramped at
8 °C/min to 300 °C, and held at 300 °C for 3
min, with helium as a carrier gas at a
constant flow of 1.4 mL/min. Retention
times (RT) and monoisotopic molecular
weights (MW) are provided for reference.
30
-------
Sea n 1
lata. ms
168
79
96
124
47
63
107
140
''''1 'fi 111 M 'i '| 'I'I'I'M'I ", , I I 111 ,11 ,' I" i"iL iV.'l'i , |111 ^1,111", |11I1
40 SO SO 70 SO 90 1 OO 1 1 O 1 20 1 30 1 40 1 SO 1 SO 1 70
Diethyl methylphosphonothioate. RT=10.6min. MW= 168.
„ 44
Sea n 1 ^
Jata. ms
114
86
72
i¦I','1,., , , i'mm
20 30 40 50 BO 70 80 90 100 110 120 130 140 150 160 170
2-(Diisopropylamino)ethanethiol (DESH). RT = ll.7 min. MW= I6l.
31
-------
143
Scan 20S
157
79
97
47
125
111
203
data . ms
175
186
230
so eo so -i oo i so -i 4o -i eo -i so soo
0,CM)iethyl dimethylpyrophosphonate (Pyro A). RT = 16.6 min. MW= 230.
157
79 97
143
41
125
175
203
185
20 40 SO SO 1 OO 120 1 40 ISO ISO 200 220 240
0-Ethyl, 0-isopropyl dimethylpyrophosphonate (Pyro B). RT = 17.0 min. MW= 244.
32
-------
157
175
sta . rn «
41
97
79
143
201
4-0 GO SO 1 OO ISO 1 -4 0 ISO ISO SOO SSO S-40 seo
Diisopropyl dimethylpyrophosphonate. RT=17.3min. MW= 258.
114
data. ms
72
129
,139
167
252
20 40 SO SO TOO 120 140 T SO "1 SO 200 220 240 260
VX. RT = 21.2 min. MW= 267.
33
-------
Scan 3^1
114
data. ms
70
84
127
144
157
193
20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320
Bis(diisopropylaminoethyl)disulfide (DESH-dimer). RT=25.7 min; MW=320.
34
-------
&EPA
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
------- |