Reaction of nerve agents with phosphate buffer at ph 7

Article abstract of DOI:10.1021/jp3024809

Chemical weapon nerve agents, including isopropyl methylphosphonofluoridate (GB or Sarin), pinacolyl methylphosphonofluoridate (GD or Soman), and S-(2-diisopropylaminoethyl) O-ethyl methylphosphonothioate (VX), are slow to react in aqueous solutions at midrange pH levels. The nerve agent reactivity increases in phosphate buffer at pH 7, relative to distilled water or acetate buffer. Reactions were studied using ³¹P NMR. Phosphate causes faster reaction to the corresponding alkyl methylphosphonic acids, and produces a mixed phosphate/phosphonate compound as an intermediate reaction product. GB has the fastest reaction rate, with a bimolecular rate constant of 4.6 × - 10^-3 M^-1s^-1[PO₄^3-]. The molar product branching ratio of GB acid to the pyro product (isopropyl methylphosphonate phosphate anhydride) is 1:1.4, independent of phosphate concentration, and the pyro product continues to react much slower to form GB acid. The pyro product has two doublets in the ³¹P NMR spectrum. The rate of reaction for GD is slower than GB, with a rate constant of 1.26 × - 10^-3 M^-1s^-1 [PO₄^3-]. The rate for VX is considerably slower, with a rate constant of 1.39 × - 10^-5 M^-1s^-1 [PO₄^3-], about 2 orders of magnitude slower than the rate for GD. The rate constant of the reaction of GD with pyrophosphate at pH 8 is 2.04 × - 10^-3 min^-1 at a concentration of 0.0145 M. The rate of reaction for diisopropyl fluorophosphate is 2.84 × - 10^-3 min^-1 at a concentration of 0.153 M phosphate, a factor of 4 slower than GD and a factor of 15 slower than GB, and there is no detectable pyro product. The half-lives of secondary reaction of the GB pyro product in 0.153 and 0.046 M solution of phosphate are 23.8 and 28.0 h, respectively, which indicates little or no dependence on phosphate.

Full text of DOI:10.1021/jp3024809

Article
pubs.acs.org/JPCA
Reaction of Nerve Agents with Phosphate Buffer at pH 7
William R. Creasy,*,† Roderick A. Fry,‡ and David J. McGarvey‡
†SAIC, P.O. Box 68 Gunpowder Branch, Aberdeen Proving Ground, Maryland 21010, United States
‡Edgewood Chemical Biological Center, Aberdeen Proving Ground-Edgewood Area, Maryland 21010, United States
S
* Supporting Information
ABSTRACT: Chemical weapon nerve agents, including isopropyl
methylphosphonofluoridate (GB or Sarin), pinacolyl methylphos-
phonofluoridate (GD or Soman), and S-(2-diisopropylaminoethyl)
O-ethyl methylphosphonothioate (VX), are slow to react in aqueous
solutions at midrange pH levels. The nerve agent reactivity increases
in phosphate buffer at pH 7, relative to distilled water or acetate
buffer. Reactions were studied using 31P NMR. Phosphate causes
faster reaction to the corresponding alkyl methylphosphonic acids,
and produces a mixed phosphate/phosphonate compound as an
intermediate reaction product. GB has the fastest reaction rate, with
a bimolecular rate constant of 4.6 × 10−3 M−1s−1[PO4 ]. The molar
3‑
product branching ratio of GB acid to the pyro product (isopropyl
methylphosphonate phosphate anhydride) is 1:1.4, independent of
phosphate concentration, and the pyro product continues to react much slower to form GB acid. The pyro product has two
doublets in the 31P NMR spectrum. The rate of reaction for GD is slower than GB, with a rate constant of 1.26 × 10−3 M−1s−1
[PO4 ]. The rate for VX is considerably slower, with a rate constant of 1.39 × 10−5 M−1s−1 [PO4 ], about 2 orders of magnitude
slower than the rate for GD. The rate constant of the reaction of GD with pyrophosphate at pH 8 is 2.04 × 10−3 min−1 at a
concentration of 0.0145 M. The rate of reaction for diisopropyl fluorophosphate is 2.84 × 10−3 min−1 at a concentration of 0.153
M phosphate, a factor of 4 slower than GD and a factor of 15 slower than GB, and there is no detectable pyro product. The half-
lives of secondary reaction of the GB pyro product in 0.153 and 0.046 M solution of phosphate are 23.8 and 28.0 h, respectively,
which indicates little or no dependence on phosphate.
3‑
3‑
values. These efforts are usually designed to develop less
corrosive decontamination methods for CW agents as
alternatives to strongly basic or oxidizing solutions that damage
substrate materials, such as polymers and electronics. Studies
are also underway of the reactions and degradation of CW
agents in the environment after exposure to materials such as
concrete, landfills, and marine environments.4,26−29 These types
of reactions are important in the various complex matrices in
the event of a CW agent release into the environment. Studies
of the stability of CW agents in simple aqueous solutions have
also been reported.30 A recent review by Churchill and co-
workers provides the recent work on reactions in decontami-
nation and environmental matrices.31
INTRODUCTION
■
Chemical weapon (CW) nerve agents, including isopropyl
methylphosphonofluoridate (GB or Sarin), pinacolyl methyl-
phosphonofluoridate (GD or Soman), and S-(2-diisopropyla-
minoethyl) O-ethyl methylphosphonothioate (VX), are highly
toxic cholinesterase inhibitors that react slowly in aqueous
solutions at midrange pH levels (6.5−7.5). The hydrolysis rate
of VX at 25 °C is reported by Epstein and co-workers as 7 ×
10−4 hr−1, or a half-life of 41 days.1 Somani reports the half-life
of VX as 350 days at 25 °C, and the half-life of GB as 5.4 h.2
The Organization for the Prohibition of Chemical Weapons
(OPCW) reports the half-lives of VX and GB as about 200 days
and 8 days, respectively, at 25 °C.3 Munro et al. reports that the
half-life for VX in water at 25 °C and pH 7 ranges from 17 to
42 days; for GB at 25 °C, the half-life ranges from 237 h (pH
6.5) to 24 h (pH 7.5); and GD has an estimated half-life of
approximately 60 h at pH 6 and 25 °C.4
GB has perhaps been studied in the most detail with regard
to hydrolysis5−7 and other displacement reactions,8 including
reactions with hypochlorite,7 amines,9 metal ions,10 catechols,11
and phenols.12,13 Hydrolysis rates seem to be affected by high
concentration of anions even if they are not strongly reactive.14
Studies are being done for testing of various CW
decontamination solutions,15−18 buffered enzymes,19−21 solid
formulations,22−25 or catalyst solutions,11 some at midrange pH
Reactions of nerve agents in phosphate buffers are important
for physiological studies of nerve agent exposures and for the
stability of standards prepared in phosphate buffers. Nerve
agent reactivity has been previously observed to increase in
phosphate buffer at pH 7 relative to distilled water or acetate
1
buffer. Gab and co-workers used a one-dimensional H−31P
̈
heteronuclear single quantum coherence (1D H−31P HSQC)
experiment to detect reaction products of GB and GD in
1
Received: March 14, 2012
Revised: June 4, 2012
Published: June 5, 2012
© 2012 American Chemical Society
7279
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
phosphate solution.32 Previous studies showed a lack of stability
of GD in phosphate buffer solutions.33,34 Buckles reported in
1947 that GD is hydrolyzed in less than 1 h in a solution of 0.2
M disodium phosphate/0.1 M citric acid at pH 7.4 and 37 °C,
measured by titrating HF.35
Reactions of VX to form symmetrical pyrophosphonates by
reaction with ethyl methylphosphonic acid are well-known.36,37
as well as in the presence of water or peroxide.8 In this study,
the kinetics of the reactions of the nerve agents with aqueous
phosphate buffer at pH 7 were measured, and the products are
reported. Preliminary results from this study were reported
previously.38
μL pipet. Varying concentrations of phosphate and buffer
solution were added in a 5-mm NMR tube (Wilmad-LabGlass,
Vineland, NJ, Part No. 527-PP-8) and vortexed to make the
aqueous solution. An amount of 50 μL D2O was added as an
NMR lock solvent. Reaction data was collected on a Bruker
Avance 300 Nuclear Magnetic Resonance Spectrometer
(NMR) using 31P detection at a frequency of 131 MHz,
using a quadruple nucleus probe (QNP). Most of the reactions
were slow, so the number of signal-averaging scans was adjusted
to improve the signal-to-noise, particularly later in the runs
when the reagent was low in concentration. The greater
number of scans increased the total signal acquisition time, but
the relaxation delay of each scan was not changed. Many of the
kinetic points were collected with automated spectra
acquisition spaced by the length of the run. The time value
of each data point was determined from the computer clock
time at the midpoint of the run. Other instrument parameters
include: sweep width (SW), 400 ppm; center frequency (O1P),
25 ppm; excite angle, 30°; recovery delay, 10 s; proton
decoupling; D2O solvent locking; and number of transform
points, 32K. The probe temperature was controlled by a
thermostat, which was externally calibrated according the
manufacturer’s instructions by using solutions of methanol and
ethylene glycol.39
Confirmation of the identity of the reaction products was
done by liquid chromatography/tandem mass spectrometry
(LC/MS and LC/MS/MS). Analysis was done on an Agilent
LC model 1100 system with a model 6490 triple quadrupole
mass spectrometer. Electrospray ionization in either positive or
negative ion mode was used. Gradient elution from 100%
aqueous 0.05 M ammonium acetate solution to 50% methanol
was used, with a Phenomenex Synergi Hydro RP column (150
mm × 2.0 mm × 0.4 μm, part number 00F-4375-B0).
The NMR spectrum of the phosphate solution indicated that
the solution contained a small amount of pyrophosphate at a
chemical shift of −4.3 ppm. Since it was considered possible
that the pyrophosphate was responsible for the observed
reaction, a solution of pyrophosphate was made from reagent-
grade sodium pyrophosphate basic decahydrate purchased from
Sigma Aldrich (99.0%, [CAS 13472−36−1], part no. S6422-
100G). A solution of 0.0145 M was prepared with 0.167 M
ammonium acetate to decrease the pH to 8.0 and tested for
reactivity.
EXPERIMENTAL SECTION
■
Aqueous solutions were made with distilled, deionized water
(DI water, CAS No. 7732-18-5, produced on demand using an
in-house source from Barnstead, Inc.) and a stock solution of
phosphate that was pH 7.0 0.1. The composition of the stock
3‑
solution was 1.61 M PO4 , made from 87.38 g K2HPO4 (Sigma
Aldrich >98%, CAS No. 7758-11-4) and 41.04 g KH2PO4 (EM
Scientific, crystals, CAS No. 7778−77−0) in 500 mL DI water.
The pH of the solution was measured to be 6.94, using a
Thermo Orion model 370 pH meter with a PrePHect
electrode, model No. 9272BN, glass combination pH sure
flow, calibrated with pH 4, 7, and 10 standard buffer solutions.
Most of the reaction solutions were generated by diluting this
stock solution in DI water. Since the reactions are usually slow,
an attempt was made to avoid introducing any other potential
contaminants that could add competing reactions. However, for
the most dilute solutions, it was found that the buffering
capacity of the dilute buffer solution was not sufficient to
neutralize the acid that was formed by the hydrolysis of the
agents. The production of acid by the reaction caused the rate
to change over time, causing curvature in the pseudo-first-order
rate plot. To eliminate this effect, ammonium acetate buffer was
used to generate a constant total buffering capacity of 0.1 M. A
stock solution of 0.2 M ammonium acetate (JT Baker # 0599-
08, CAS No. 631-61-8) was used for the solutions. The pH
measured for the 0.2 M solution was 7.00. Reaction rates that
are reported for some phosphate solutions were measured both
with and without 0.1 M ammonium acetate solution for extra
buffering, as discussed in the next section.
CW agents isopropyl methylphosphonofluoridate (Sarin or
GB, [CAS 107-44−8]), pinacolyl methylphosphonofluoridate
(Soman or GD, [CAS 96-64-0]), and S-(2-diisopropylami-
noethyl) O-ethyl methylphosphonothioate (VX, [CAS 50782-
69-9]) were obtained from the Chemical Agent Standard
Analytical Reference Material (CASARM) program at the
Edgewood Chemical Biological Center, Aberdeen Proving
Ground, MD. The agent purities were determined by liquid
NMR and confirmed by gas chromatography/thermal con-
ductivity detection. Purities of GB, GD, and VX were 97, 97,
and 95 wt %, respectively. All handling of the CWA was
conducted in a chemical surety laboratory certified for
supertoxic compounds. Diisopropyl fluorophosphate (DFP)
was purchased commercially from Aldrich ([CAS 55−91−4],
part number D12,600−4). Caution: The neat nerve agent
compounds are extremely toxic and must be handled in
accordance with all applicable Federal laws and international
treaties, using appropriate safety and security operating
procedures.
RESULTS
■
GB Reaction. The reaction product from the hydrolysis of
GB with water is isopropyl methylphosphonic acid (commonly
called GB acid). In the presence of phosphate, the rate of
reaction of GB to form GB acid is faster than in DI water or
acetate buffer. The additional “pyro” product is observed that is
a mixed phosphate/phosphonate compound identified as
isopropyl methylphosphonate phosphate anhydride (structure
1), which reacts slowly. The compound is distinctive because of
two widely spaced 31P NMR doublets corresponding to the two
distinct 31P species, with the chemical shifts and coupling
constants shown in Table 1. A sample NMR spectrum of a
reaction mixture is shown in Figure 1. 1 has two doublets each
with a small coupling from the 31P−31P interaction of 22.3 Hz,
consistent with NMR modeling results40 and literature reports
for the similar symmetric VX pyro compound bis[ethyl
methylphosphonate]ester.41,42
For kinetic studies, 1 μL of agent was added to a total of 500
μL of aqueous solution to make a 14 mM solution, using a 10
Gab and co-workers32 used a 1D 1H−31P HSQC experiment
̈
to detect the same reaction products of GB in phosphate
7280
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
The semilog kinetic plot of GB hydrolysis in 0.1 M
ammonium acetate buffer at pH 7 without phosphate is
shown in Figure 2. The half-life calculated from the slope is
Table 1. 31P Chemical Shifts for Compounds in Aqueous pH
a
7 buffer
chemical
shift (ppm)
coupling constant (Hz),
proton decoupled
compound
3−
PO4 at pH 7 (mixture of
protonated forms)
5.0
GB
38.38
30.02
1045.5 (31P−19F)
22.2 (31P−31P)
GB acid
Figure 2. Semilog kinetic plot of the reaction of GB with and without
isopropyl methylphosphonate
phosphate anhydride
26.8
3−
PO4 present. The half-life from the slope is calculated as 17.9 h for
−1.71
no phosphate buffer and 0.1 M acetate buffer. The reaction of GB in
phosphate buffer at a concentration of 0.031 M PO4 and 0.1 M
pyrophosphate
GD
−4.3
38.06
3−
1047.5 (31P−19F)
acetate buffer has a half-life of 194 min. The half-life for 0.015 M
PO43− with no acetate buffer is 278 min, using only the first four data
points. The curvature for the 0.015 M plot is due to the decreasing rate
from the changing pH of the solution due to insufficient buffering of
the acid reaction products.
37.41
GD acid
29.0
pinacolyl methylphosphonate
phosphate anhydride
26.13, 25.86
−2.31, −2.34
21.4 (31P−31P)
21.5 (31P−31P)
a
Shifts are externally referenced to concentrated H3PO4.
17.9 h at 23 °C, which falls in the range of the previous reports
of the GB reaction rate for pH 7 solution under these
conditions. The y-axis of all the kinetic plots is the natural log of
the ratio of molar concentration of remaining GB to the total of
GB and products, including the pyro and acid products, using
the signal from the 31P NMR. A question that was not studied
is whether this slow reaction is affected by the use of a glass
tube as a container, since the glass may react with the product
HF as a sink to eliminate the reverse reaction. This issue could
account for some of the variation in the reported rates of GB
hydrolysis.
solution. This technique works for molecules for which a 31P
1
atom was within three bonds of a H nucleus, so the 31P from
methylphosphonate is observed but the 31P from phosphate is
not. This experiment had the advantage of greater sensitivity
than 1D 31P NMR, so that lower concentrations of nerve agent
and products could be detected. It has the disadvantages that
31P chemical shifts or coupling constants are not detected, and
different compounds have different signal responses. In order to
obtain kinetic information, the authors had to correct for the
different signal responses of the reactants and products.
Figure 1. Sample NMR spectrum of GB and phosphate reaction solution, showing the residual reagent GB, the products GB acid and the pyro
compound 1, and the reagent phosphate (partially protonated in solution) and impurity pyrophosphate (partially protonated in solution).
7281
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
The rate increases substantially when phosphate was added,
and it has linear dependence on phosphate concentration. The
plot for GB in a 0.306 M phosphate buffer at 23 °C is shown in
Figure 3. For this run only, 3 μL of GB was added instead of 1
Figure 4. Linear plot of the first-order rates of GB reaction versus
△
concentration of phosphate. All data is from this work, except from
ref 32 (see text, not included in regression).
Figure 3. Semilog kinetic plot of the reaction of GB in phosphate
buffer at a concentration of 0.306 M PO43−. The half-life of the
reaction is 8.4 min.
both products arise from a mechanism with a common
intermediate. However, after the GB is consumed, 1 continues
to react slowly to GB acid, discussed in a later section.
The temperature dependence of the reaction of GB with
phosphate was measured at five temperatures from 23 to 63 °C,
at a phosphate concentration of 0.0307 and 0.1 M acetate
buffer. The rate data is plotted in Figure 5. The activation
energy Ea that is found from the slope of the plot for the
reaction is 63.3 kJ/mol.
μL, for a concentration of 42 mM, in order to increase the
signal to use shorter runs, but this concentration is still in the
pseudo-first-order condition. For these reaction conditions, the
half-life is 8.4 min, or a factor of 140 times faster than in acetate
buffer with no phosphate.
For the most dilute buffer solutions, it was found that the
buffering capacity of the dilute phosphate solution was not
sufficient to neutralize the acid that was formed by the
hydrolysis of the agents. The rate is pH dependent, so a change
in pH affects the rate, and the rate decreases from pH 7.5 to
6.5.4 Acetate buffer was added to increase the buffer strength.
Figure 2 shows the kinetic plots comparing reaction runs of GB
in phosphate buffer with and without acetate buffer. The
3−
solutions have a concentration of 0.031 M PO4 and 0.1 M
3−
acetate buffers, and 0.015 M PO4 and no acetate buffer. The
half-life of the reaction for 0.031 M PO43− is 194 min. The half-
3−
life for 0.015 M PO4 is 278 min, and only the first four data
points are used since the plot curves as the reaction slows
down. The curvature for the 0.015 M plot is due to the
decreasing rate, caused by the changing pH of the solution due
to insufficient buffering. By the end of the reaction, the pH of
the solution was measured as pH 3 (using pH test paper).
A previous report by Ellin and co-workers43 indicated that
GB and GD in acetate buffer solution at pH 4.5 has faster
degradation kinetics than other pH 4.5 buffer solutions, but the
present results do not indicate that acetate promotes faster
reaction at pH 7 compared to DI water, aside from buffering
the solution so that the rate does not slow down.
In order to get a second-order rate constant, the reaction was
done with a series of concentrations of phosphate. The plot of
the pseudo-first-order rates versus phosphate concentration is
shown in Figure 4. The bimolecular rate constant is found to be
4.6 × 10−3 M−1 s−1 [PO43−] at 23 °C from the regression to
this data. A data point in plotted using the kinetic data in ref
32.44
Figure 5. Rate data for the reaction of GB with phosphate at five
temperatures, using the natural log of the rate constant of loss of GB at
a concentration of 0.0307 M phosphate (with 0.1 M acetate buffer)
versus reciprocal absolute temperature.
In order to confirm the identity of the product, the reaction
solution was analyzed by LC/MS and LC/MS/MS. Products
that were observed are given in Table 2 and are consistent with
the assignment of the structure. In the positive ion electrospray
mode, the [M+H]+ peak at m/z 219 fragments to m/z 177 with
+
loss of propene, and to the m/z 99 ion H4PO4 . A minor peak
at m/z 137 corresponds to loss of P(OH)3. In the negative ion
mode, the [M−H]− peak at m/z 217 fragments to form the
Table 2. LC/MS/MS Fragment Ions That Were Observed
from Analysis of 1 and 2
The product molar branching ratio of GB acid to 1 is 1:1.4
during the course of the reaction of GB, independent of the
phosphate concentration. This ratio is in good agreement with
compound and polarity
parent ion, (m/z)
MS/MS fragment ions
1, positive ion
1, negative ion
2, positive ion
2, negative ion
219
217
261
259
177, 137, 99
79
177, 99
79
the ratio reported by Gab and co-workers for GB, GD, and
̈
GF.32 The fact that the branching ratio does not change and
does not depend on the phosphate concentration indicates that
7282
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
anion CH3P(OH)O− or (PO3)− at m/z 79. These results are in
good agreement with those of Gab and co-workers.32
̈
GD reaction. GD has an analogous reaction to GB, but GD
acid and structure 2 are formed. There is extra splitting of the
Figure 7. Linear plot of the first-order rates of GD reaction versus
△
concentration of phosphate. All data is from this work, except from
○
ref 32 and from ref 34 (not included in regression).
peaks for 2 in the NMR spectrum due to the diastereomers
from two chiral centers, on the P atom and the pinacolyl C
atom. Figure 6 shows the 31P NMR spectrum with expanded
insets of the spectral lines. The peak shifts and coupling
constants are given in Table 1.
In order to confirm the identity of the product, the reaction
solution was analyzed by LC/MS and LC/MS/MS. Figure 8
shows the negative ion LC/MS spectrum, with the peaks at the
expected masses. An interesting aspect of this spectrum is that
sodium and potassium adduct ions were observed even in the
negative ion spectrum, as [M+Na-2H]¯, [2M+Na-2H]¯, and
[2M+K-2H]¯. Fragment ions using LC/MS/MS are given in
Table 2. As with 1, the [M+H]+ peak fragments to m/z 177
+
with loss of alkene and to the m/z 99 ion H4PO4 . The [M−
H]− peak fragments to the anion CH3P(OH)O− or (PO3)− at
m/z 79.
GD Reaction with Pyrophosphate. Since there was a 2.2
mol % of pyrophosphate impurity in the phosphate, it was
considered that the source of the reaction could be from a fast
reaction of agent with the low concentration of pyrophosphate.
A solution of 0.0145 M pyrophosphate with 0.167 M
ammonium acetate was made which was pH 8. The semilog
kinetic plot for the reaction of GD with the solution of
pyrophosphate is linear and is given in the Supporting
Information. The rate constant of the reaction is 2.04 × 10−3
min−1 at a concentration of 0.0145 M. The reaction rate of GD
with phosphate at the same concentration, taken from the plot
of Figure 7, is 1.6 × 10−3 min−1, only slightly slower. A pyro
compound was observed in the NMR spectra, but the
branching ratio of pyro/acid products is 10%, much lower
than the reaction of GD with phosphate. As a result, the
reaction of pyrophosphate does not significantly contribute to
the production of 2. Since the concentration of pyrophosphate
is much lower than phosphate in the pH 7 buffer solution, this
reaction channel does not contribute significantly to the overall
reaction rate.
Figure 6. Product 31P NMR spectrum from GD reaction with
phosphate, with insets showing the mixed pyro product, 2.
VX Reaction. VX reacts with phosphate, but much slower
than GD. The analogous mixed pyro product is observed in the
31P NMR spectra. The semilog reaction rate plot for the VX
reaction at a phosphate concentration of 0.153 M is linear, and
the half-life is 54.3 h; the plot is given in the Supporting
Information. The plot of the pseudo-first-order rates versus
phosphate concentration is shown in Figure 9. The bimolecular
rate constant is found to be 1.40 × 10−5 M−1 s−1 [PO43−], 1% of
the rate for GD.
DFP Reaction. DFP is commonly used as a simulant to
substitute for GB and GD due to its similar chemistry and
slightly lower toxicity.46 The reactivity of DFP in phosphate
solution was faster than in acetate buffer, but the presence of a
pyro compound was not detected. The rate of reaction for DFP
The semilog reaction rate plots of GD with phosphate in 0.1
M acetate buffer are linear. An example for 0.0766 M
phosphate, with a half-life of 100.5 min, is given in the
Supporting Information. The plot of the pseudo-first-order
rates versus phosphate concentration is shown in Figure 7. The
bimolecular rate constant is found to be 1.26 × 10−3 M−1 s−1
[PO43−], which is 27% of the rate for GB. One data point for
0.1 M phosphate is plotted using the kinetic data in ref 32,44
and one point for 0.02 M phosphate is plotted from ref 34.45
These points were not used in the regression, but they are in
good agreement to the best fit line. The rate of reaction for GD
is probably slower than GB because of the greater steric
hindrance of the pinacolyl group on the GD compared to the
isopropyl group on the GB.
7283
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
Figure 8. LC/MS of 2 using negative ion electrospray ionization. The [M−H]− peak for the compound is at m/z 259, and the m/z 519 peak is the
[2M−H]− peak. It is interesting to observe alkali metal adducts at m/z 281 for [M+Na−2H]−, m/z 541 for [2M+Na−2H]−, and m/z 557 is [2M
+K−2H]−.
Figure 9. Linear plot of the first-order rates of VX reaction versus
concentration of phosphate.
Figure 10. Semilog plot of the first-order rates of 1 reaction versus
time for two concentrations of phosphate. The half-lives, calculated
from the slope of the line for times after 500 min when the product is
no longer forming, are 28.0 h for 0.046 M solution and 23.8 h for
was 2.84 × 10−3 min−1 at a concentration of 0.153 M
phosphate, compared to a rate of 4.3 × 10−2 min−1 for GB and
1.18 × 10−2 min−1 for GD. The rate for DFP is a factor of 4
0.153 M solution. Kinetic modeling of the data is given in the
slower than GD and a factor of 15 slower than GB.
Supporting Information.
Secondary 1 Reaction. The reaction rate of 1 was
measured. Figure 10 shows the kinetic plot for the reaction
Gab and co-workers say that their data supports a zero-order
̈
of 1 as a function of time at two different concentrations of
phosphate. The rate of reaction in 0.153 and 0.046 M solution
of phosphate are similar. The half-lives for these concentrations
are 23.8 and 28.0 h, respectively. The slopes for the half-lives
shown in Figure 10 are for times >500 min after the reaction
begins, and the GB is undetectable by this time, so 1 is no
longer forming. These results indicate that the secondary
reaction of 1 is not significantly dependent on the phosphate
concentration, but is likely to be a hydrolysis reaction with
water.
Simple modeling calculations of the kinetics of formation and
reaction of 1 were done, and sample plots are shown in the
Supporting Information. Rate constants that have been
reported from the regression plots are consistent with modeling
calculations of the concentration of 1 over the entire time
range. The most complex case was the rate of reaction in the
0.0153 M solution shown in Figure 2. The loss of GB is much
slower, and the concentration of 1 is almost flat to 14 000 min,
and this reaction was difficult to model due to the changing
rates as the pH values changed.
reaction rate for the decomposition of the pyro compounds.32
However, a zero-order rate does not make theoretical sense
under these conditions, and the present data is not consistent
with a zero-order rate law. A plot is shown in the Supporting
Information from the kinetic model that shows that the
concentration decrease appears to be linear (zero-order) over
the time range that was studied in that work, but it is first-order
over a longer time.
The rate of reaction of 2 is similar to the rate for 1, with a
half-life of 37 h. The uncertainty is large because the signal-to-
noise ratio of the NMR spectra is low because the signal is split
into many peaks. The branching ratio is 1:0.66 for the products
GD acid and 2, respectively, for a phosphate concentration of
0.0766 M.
DISCUSSION
■
Although this work has been done in synthetic phosphate
buffer solution, biological matrices are also phosphate buffers
that are near neutral pH. Standards for biological studies are
7284
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
often prepared in phosphate buffers, so it may be necessary to
be aware of the stability of nerve agent in such solutions
containing phosphate. Interactions of nerve agents with
phosphate could affect the reaction of nerve agents in biological
solutions such as blood or standards made for biological
studies.
under Contract Numbers DAAD13-03-D-0017 and W911SR-
10-D-0004.
REFERENCES
■
(1) Epstein, J.; Callahan, J. J.; Bauer, V. E. The Kinetics and
Mechanisms of Hydrolysis of Phosphonothiolates in Dilute Aqueous
Solution. Phosphorus 1974, 4, 157−63.
(2) Somani, S. M., ed. Chemical Warfare Agents; Academic: San
Diego, CA, 1992.
It may be possible to monitor biological fluids for
compounds 1 and 2 as a method for determining whether
individuals have been exposed to nerve agent, as an alternative
to current methods.47 However, the concentration of
orthophosphate in blood is rather low (1−2 mM) and
undergoes complex reactions,48 so the reactions with GB or
GD in blood would be slow and may be overwhelmed by other
loss mechanisms. Since the product compounds continue to
react, the monitoring would be necessary within days after the
exposure. The products of a reaction are likely to be difficult to
find, due to binding with proteins and the complex matrix.
The toxicity of 1 and 2 are unknown, but they may have
anticholinesterase activity, since the structures are similar to
tetraethyl pyrophosphate, a cholinesterase inhibitor (oral LC50
in rats of 500 μg/kg49). As a result, it is not possible to say that
the GB or GD has been detoxified by the reaction with
phosphate without testing of the toxicity of the reaction
product. The possible product toxicity as well as the slow speed
of the reactions will probably limit the application of these
reactions to routine nerve agent decontamination.
(3) Organization for the Prohibition of Chemical Weapons Web
Page. “Nerve Agents,” from A FOA Briefing Book on Chemical
Weapons: Threat, Effects, and Protection. http://www.opcw.org/resp/
html/nerve.html.
(4) Munro, N. B.; Talmage, S. S.; Griffin, G. D.; Waters, L. C.;
Watson, A. P.; King, J. F.; Hauschild, V. The Sources, Fate, and
Toxicity of Chemical Warfare Agent Degradation Products. Environ.
Health Perspect. 1999, 107, 933−74.
(5) Larsson, L. The Alkaline Hydrolysis of Isopropoxy-methyl-
phosphoryl Fluoride (Sarin) and Some Analogues. Acta Chem. Scand.
1957, 11, 1131−42.
(6) Lohs, K. Synthetische Gifte (Synthetic Poisons); Ministerium fur
̈
nationale Verteidigung: Berlin, 1963; p 308.
(7) Epstein, J.; Bauer, V. E.; Saxe, M.; Demek, M. M. The Chlorine-
Catalyzed Hydrolysis of Isoproyl Methylphosphonofluoridate (Sarin)
in Aqueous Solution. J. Am. Chem. Soc. 1956, 78, 4068−4071.
(8) Franke, S. Manual of Military Chemistry: Chemistry of Chemical
Warfare Agents; National Technical Information Service # AD-849
866, 1967; Vol. 1.
The mechanism of the reaction has not been determined.
The fact that the branching ratio of the reactions to form 1 and
GB acid does not change as a function of phosphate
concentration indicates that both products arise from a
common intermediate involving a phosphate complex that
decomposes to both products. There is a direct hydrolysis
reaction of GB involving only water, which may be acid or base
catalyzed, but it is much slower. The phosphate acts in a
catalytic role to consume the nerve agent, for the reaction
channel that produces the corresponding acid. There is no
direct evidence about the structure of the reactive intermediate,
in particular whether the P−F bond (or P−S for VX) is
eliminated before or after the intermediate is formed. However,
these results indicate that the destruction of toxic nerve agents
can be accomplished in a catalytic way. Further research in this
area of the fundamental chemistry of this class of compounds
may assist in improving the reactivity.
(9) Epstein, J.; Cannon, P. L., Jr.; Sowa, J. R. The Catalysis of the
Hydrolysis of Isopropyl Methylphosphonofluoridate in Aqueous
Solutions by Primary Amines. J. Am. Chem. Soc. 1970, 92, 7390−93.
(10) Epstein, J.; Rosenblatt, D. H. Kinetics of Some Metal Ion-
Catalyzed Hydrolyses of Isopropyl Methylphosphonofluoridate (GB)
at 25°C. J. Am. Chem. Soc. 1958, 80, 3596−98.
(11) Epstein, J.; Rosenblatt, D. H.; Demek, M. M. Kinetics of the
Reaction of Isopropyl Methylphosphonofluoridate with Catechols at
25°C. J. Am. Chem. Soc. 1956, 78, 341−43.
(12) Epstein, J.; Plapinger, R. E.; Michel, H. O.; Cable, J. R.;
Stephani, R. A.; Hester, R. J.; Billington, C., Jr.; List, G. R. Reactions of
Isopropyl Methylphosphonofluoridate with Substituted Phenols. I. J.
Am. Chem. Soc. 1964, 86, 3075−84.
(13) Epstein, J.; Michel, H. O.; Rosenblatt, D. H.; Plapinger, R. A.;
Stephani, R. A.; Cook, E. Reactions of Isopropyl Methylphosphono-
fluoridate with Substituted Phenols. II. J. Am. Chem. Soc. 1964, 86,
4959−63.
(14) Davisson, M. L.; Love, A. H.; Vance, A.; Reynolds, J. G.
Environmental Fate of Organophosphorus Compounds Related to
Chemical Weapons; Lawrence Livermore National Laboratory,
UCRL-TR-209748, 2005.
(15) Boone, C. M. Present State of CBRN Decontamination
Methodologies; TNO Report TNO-DV-2007-A028; Organisation for
Applied Scientific Research (TNO) Defence, Security and Safety:
Rijswijk, The Netherlands, 2007. This document is available from
http://stinet.dtic.mil.
(16) Wagner, G.; Yang, Y.-C. Rapid Nucleophilic/Oxidative
Decontamination of Chemical Warfare Agents. Ind. Eng. Chem. Res.
2002, 41, 1925−1928.
(17) Wagner, G. W.; Sorrick, D. C.; Procell, L. R.; Brickhouse, M. D.;
Mcvey, I. F.; Schwartz, L. I. Decontamination of VX, GD, and HD on a
Surface Using Modified Vaporized Hydrogen Peroxide. Langmuir
2007, 23, 1178−1186.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information includes additional kinetic plots,
and a discussion of numerical kinetic modeling of the change of
the concentration of the 1 and 2 products. This information is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: william.r.creasy.ctr@us.army.mil.
Notes
The authors declare no competing financial interest.
(18) Waysbort, D.; McGarvey, D. J.; Creasy, W. R.; Morrissey, K. M.;
Hendrickson, D. M.; Durst, H. D. A Decontamination System for
Chemical Weapons Agents Using a Liquid Solution on a Solid
Sorbent. J. Hazard. Mater. 2008, 161, 1114−1121.
(19) Wang, A. A.; Mulchandani, A.; Chen, W. Specific Adhesion to
Cellulose and Hydrolysis of Organophosphate Nerve Agents by a
Genetically Engineered Escherichia coli Strain with a Surface-
ACKNOWLEDGMENTS
■
The authors thank H. Dupont Durst and E. Michael
Jakubowski for support and helpful discussions. The work
was performed at the Edgewood Chemical Biological Center
7285
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286

The Journal of Physical Chemistry A
Article
Expressed Cellulose-Binding Domain and Organophosphorus Hydro-
lase. Appl. Environ. Microbiol. 2002, 68, 1684−89.
(41) Yang, Y.-C.; Szafraniec, L. L.; Beaudry, W. T.; Rohrbaugh, D. K.;
Procell, L. R.; Samuel, J. B. Autocatalytic Hydrolysis of V-Type Nerve
Agent. J. Org. Chem. 1996, 61, 8407−8413.
(20) Gopal, S.; Rastoqi, V.; Ashman, W.; Mulbry, W. Mutagenesis of
Organophosphorus Hydrolase to Enhance Hydrolysis of the Nerve
Agent VX. Biochem. Biophys. Res. Commun. 2000, 279, 516−19.
(42) The Supporting Information of ref 41 discusses the splitting of
the VX pyro peaks as due to diastereomeric isomers, due to the two
chiral P atoms in the compound. However, 1 has a similar splitting but
only has one chiral P center. The software in ref 40 calculates coupling,
not shifts from diastereomers, and it is in good agreement to the
splitting. As a result, we conclude that the splittings for 1 are from P−
P coupling, and the similar VX pyro splitting may also not be from
diastereomers.
(43) Ellin, R. I.; Groff, W. A.; Kaminkis, A. The Stability of Sarin and
Soman in Dilute Aqueous Solutions and the Catalytic Effect of Acetate
Ion. J. Environ. Sci. Health 1981, B16 (6), 713−717.
(44) Reference 32 does not report rate constants, but the effective
rate constant of [GB] reaction can be obtained from the data provided
using the pseudofirst order rate equation: [GB]t = [GB]0 exp(−kt), so
k = −t−1 ln(([GB]0 − Δ[GB])/[GB]0). [GB]0 = 7.1 mM, and Δ[GB]
= 83.6 μM/min, adding the corrected rates of reaction from both
adduct formation and acid formation.
(45) Reference 34 gives a half-life of 6.6 h at 0.02 M phosphate in pH
7.4 buffered saline solution at 27 °C, which gives a rate constant of
1.75 × 10−3 M−1 min.−1
(46) McGarvey, D. J.; Creasy, W. R.; Fry, R. A.; Durst, H. D.
Comparisons of Reactivity of Chemical Weapons Agents to Agent
Simulants, 238th American Chemical Society National Meeting,
Washington, D.C., Aug. 16−20, 2009.
(47) Renner, J. A.; Dabisch, P. A.; Evans, R. A.; McGuire, J. M.;
Totura, A. L.; Jakubowski, E. M.; Thomson, S. A. Validation and
Application of a GC-MS Method for Determining Soman Concen-
tration in Rat Plasma Following Low-Level Vapor Exposure. J. Anal.
Toxicol. 2008, 32, 92−98.
(48) Shoemaker, D. G.; Bender, C. A.; Gunn, R. B. Sodium-
Phosphate Cotransport in Human Red Blood Cells. J. Gen. Physiol
1988, 92, 449−474.
́
(21) Vayron, P.; Renard, P.-Y.; Taran, F.; Creminon, C.; Frobert, Y.;
Grassi, J.; Mioskowski, C. Toward Antibody-Catalyzed Hydrolysis of
Organophosphorus Poisons. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
7058−7063.
(22) Tucker, M. D.; Comstock, R. H. Decontamination Formulation
with Sorbent Additive, U.S. Patent #7,282,470, October 16, 2007
Tucker, M. D. Granulated Decontamination Formulations, U.S. Patent
#7,276,468, October 2, 2007.
(23) Hoffman, D. M.; McGuire, R. R. Oxidizer Gels for
Detoxification of Chemical and Biological Agents, U.S. Patent #
6,455,751, September 24, 2002.
(24) Hoffman, D. M.; Chiu, I. L. Solid-Water Detoxifying Reagents
for Chemical and Biological Agents, U.S. Patent # 7,030,071, April 18,
2006.
(25) McDonald, W. F.; Parsons, A. B.; Northrup, V. M.; Hill-Clark,
D.; O’Loughlin, M. E. Developments of New Reactive Sorbents, Battelle,
Dec. 1993, Edgewood Research, Development, and Engineering
Center Report ERDEC-CR-037.
(26) Bartelt-Hunt, S. L.; Barlaz, M. A.; Knappe, D. U. R.; Kjeldsen, P.
Fate of Chemical Warfare Agents and Toxic Industrial Chemicals in
Landfills. Environ. Sci. Technol. 2006, 40 (13), 4219−4225.
(27) Bizzigotti, G. O.; Castelly, H.; Hafez, A. M.; Smith, W. H. B.;
Whitmire, M. T. Parameters for Evaluation of the Fate, Transport, and
Environmental Impacts of Chemical Agents in Marine Environments.
Chem. Rev. 2009, 109 (1), 236−256.
(28) Brevett, C. A. S.; Sumpter, K. B.; Nickol, R. G. Kinetics of the
Degradation of Sulfur Mustard on Ambient and Moist Concrete. J.
Hazard. Mater. 2009, 162 (1), 281−291.
(29) Wagner, G. W.; O’Connor, R. J.; Edwards, J. L.; Brevett, C. A. S.
Effect of Drop Size on the Degradation of VX in Concrete. Langmuir
2004, 20, 7146−7150.
(49) U.S. Department of Labor Occupational Safety and Health
Administration Web Page, Occupational Safety and Health Guideline
for Tetraethyl Pyrophosphate (TEPP), http://www.osha.gov/SLTC/
healthguidelines/tepp/recognition.html#healthhazard.
(30) Morrissey, K.; Schenning, A.; Fouse, J.; Smith, P.; Nunes, R.;
Sumpter, K.; Durst, H. D. Degradation of Chemical Warfare Agents in
Potable Waters; 2009 Chemical and Biological Defense Science and
Technology Conference, Dallas, TX, Nov. 16−20, 2009.
(31) Kim, K.; Tsay, O. G.; Atwood, D. A.; Churchill, D. G.
Destruction and Detection of Chemical Warfare Agents. Chem. Rev.
2011, 111, 5345−5403.
(32) Gab, J.; John, H.; Blum, M.-M. Formation of Pyrophosphate-
̈
Like Adducts from Nerve Agents Sarin, Soman, and Cyclosarin in
Phosphate Buffer: Implications for Analytical and Toxicological
Investigations. Toxicol. Lett. 2011, 200, 34−40.
(33) Grunwald, J.; Raveh, L.; Doctor, B. P.; Ashani, Y. Huperzine A
as a Pretreatment Candidate Drug against Nerve Agent Toxicity. Life
Sci. 1994, 54 (14), 991−997.
(34) Broomfield, C. A.; Lenz, D. E.; MacIver, B. the Stability of
Soman and Its Stereoisomers in Aqueous Solution: Toxicological
Considerations. Arch. Toxicol. 1986, 59, 261−265.
(35) Buckles, L. C. The Hydrolysis Rate of GD; Technical Command
Informal Report TCIR 373, March 28, 1947.
(36) Yang, Y.-C.; Baker, J. A.; Ward, J. R. Decontamination of
Chemical Warfare Agents. Chem. Rev. 1992, 92, 1729−1743.
(37) Yang, Y.-C. Chemical Detoxification of Nerve Agent VX. Acc.
Chem. Res. 1999, 32, 109−115.
(38) Creasy, W. R.; Fry, R. A.; McGarvey, D. J. Reaction of Nerve
Agents with Phosphate Buffer at pH 7, presented at the 2008 Chemical
and Biological Defense Physical Science and Technology Conference,
New Orleans, LA, 17−21 November 2008.
(39) Van Geet, A. L. Temperature Measurement in Nuclear
Magnetic Resonance with a Spinning Thermistor. Rev. Sci. Instrum.
1969, 40, 177−178.
(40) ACD/XNMR, v. 14.00; Advanced Chemistry Development, Inc.:
Toronto, Canada, 2012.
7286
dx.doi.org/10.1021/jp3024809 | J. Phys. Chem. A 2012, 116, 7279−7286