Facile hydrolysis-based chemical destruction of the warfare agents VX, GB, and HD by alumina-supported fluoride reagents

Article abstract of DOI:10.1021/jo8019972

(Chemical Equation Presented) A facile solvent-free hydrolysis (chemical destruction) of the warfare agents VX (O-ethyl S-2-(diisopropylamino)ethyl methylphosphonothioate), GB (O-isopropyl methylphosphonofluoridate or sarin), and HD (2,2′-dichloroethyl sulfide or sulfur mustard) upon reaction with various solid-supported fluoride reagents is described. These solid reagents include different alumina-based powders such as KF/Al₂O₃, AgF/KF/Al₂O₃, and KF/Al₂O₃ enriched by so-called coordinatively unsaturated fluoride ions (termed by us as ECUF-KF/Al₂O₃). When adsorbed on these sorbents, the nerve agent VX quickly hydrolyzed (t_1/2 range between 0.1-6.3 h) to the corresponding nontoxic phosphonic acid EMPA as a major product (>90%) and to the relatively toxic desethyl-VX (<10%). The latter byproduct was further hydrolyzed to the nontoxic MPA product (t_1/2 range between 2.2-161 h). The reaction rates and the product distribution were found to be strongly dependent on the nature of the fluoride ions in the KF/Al₂O ₃ matrix and on its water content. All variations of the alumina-supported fluoride reagents studied caused an immediate hydrolysis of the highly toxic GB (t_1/2 < 10 min) to form the corresponding nontoxic phosphonic acid IMPA. A preliminary study of the detoxification of HD on these catalyst supports showed the formation of the nontoxic 1,4-thioxane as a major product together with minor amounts of TDG and vinylic compounds within a few days. The mechanisms and the efficiency of these processes were successfully studied by solid-state ³¹P, ¹³C, and ¹⁹F MAS NMR.

Full text of DOI:10.1021/jo8019972

Facile Hydrolysis-Based Chemical Destruction of the Warfare
Agents VX, GB, and HD by Alumina-Supported Fluoride Reagents
†
†
E. Gershonov, I. Columbus, and Y. Zafrani*
The Department of Organic Chemistry, Israel Institute for Biological Research, Ness-Ziona 74100, Israel
yossiz@iibr.goV.il
ReceiVed September 10, 2008
A facile solvent-free hydrolysis (chemical destruction) of the warfare agents VX (O-ethyl S-2-
(
diisopropylamino)ethyl methylphosphonothioate), GB (O-isopropyl methylphosphonofluoridate or sarin),
and HD (2,2′-dichloroethyl sulfide or sulfur mustard) upon reaction with various solid-supported fluoride
reagents is described. These solid reagents include different alumina-based powders such as KF/Al
AgF/KF/Al , and KF/Al enriched by so-called coordinatively unsaturated fluoride ions (termed by
us as ECUF-KF/Al ). When adsorbed on these sorbents, the nerve agent VX quickly hydrolyzed (t1/2
range between 0.1-6.3 h) to the corresponding nontoxic phosphonic acid EMPA as a major product
>90%) and to the relatively toxic desethyl-VX (<10%). The latter byproduct was further hydrolyzed to
the nontoxic MPA product (t1/2 range between 2.2-161 h). The reaction rates and the product distribution
were found to be strongly dependent on the nature of the fluoride ions in the KF/Al matrix and on its
2
3
O ,
2
O
3
2 3
O
2
O
3
(
2
O
3
water content. All variations of the alumina-supported fluoride reagents studied caused an immediate
hydrolysis of the highly toxic GB (t1/2 < 10 min) to form the corresponding nontoxic phosphonic acid
IMPA. A preliminary study of the detoxification of HD on these catalyst supports showed the formation
of the nontoxic 1,4-thioxane as a major product together with minor amounts of TDG and vinylic
compounds within a few days. The mechanisms and the efficiency of these processes were successfully
3
1
13
19
studied by solid-state P, C, and F MAS NMR.
Introduction
availability, and easy disposal. Inorganic oxide powders may
comply with these requirements due to their unique surfaces
that possess a large number of basic and acidic Br o¨ nsted and
Lewis sites, which can significantly accelerate hydrolysis
processes. They are also environmentally benign and benefit
from lack of corrosiveness and low cost. Following these
requirements, Wagner et al. have reported on the reactions of
neat liquid VX, GD (soman), and HD with the nanosize
Detoxification of extremely toxic chemical warfare agents
CWA) such as VX, GB (sarin), and HD (sulfur mustard) is a
(
current concern. Rapid removal of CWA from diverse surfaces
is required in order to regain use of the affected area and
equipment. Reactive decontaminants that chemically destroy
CWA rather than simple physical removal, as in the case of
the Fuller’s earth sorbent, are favored.
characterized by a capability to neutralize CWA in a rapid and
safe manner, ease of handling, stability in long-term storage,
1
-5
They should be
1
2
3
inorganic oxide particles MgO, CaO, NaY, and AgY zeolites
4
and alumina. Although these three CWAs were found to
undergo a room-temperature hydrolysis (chemical destruction)
on the surface of the above-mentioned inorganic oxides, the
reactions were limited by their diffusion capability, which
reduced the overall reaction rates (a t1/2 range between hours to
days). In addition, these metal oxides nanoparticles were found
to be unstable in the presence of air or moisture. Overcoming
†
These authors contributed equally.
(
1) Wagner, G. W.; Koper, O. B.; Lucas, E.; Decker, S.; Klabunde, K. J. J.
Phys. Chem. B 2000, 104, 5118.
2) Wagner, G. W.; Bartram, P. W.; Koper, O. B.; Klabunde, K. J. J. Phys.
Chem. B 1999, 103, 3225.
3) Wagner, G. W.; Bartram, P. W. Langmuir 1999, 15, 8113.
(
(
1
0.1021/jo8019972 CCC: $40.75  2009 American Chemical Society
J. Org. Chem. 2009, 74, 329–338 329
Published on Web 11/16/2008

Gershonov et al.
these drawbacks, the same author has recently reported that VX,
SCHEME 1
GD, and HD rapidly hydrolyzed (t1/2 between minutes to hours)
5
under the action of nanotubular titania. Spafford reported on
6
immediate decomposition of VX by using powder AgF;
however, formation of the toxic “G-analogue” byproduct is a
significant disadvantage of this method.
Supporting reagents on inorganic oxides, such as alumina,
silica, zeolites, and clays, is a well-known approach to increase
reactivity or selectivity of these solid supports catalysts. For
example, adsorption of potassium fluoride on the surface of
2 3
neutral alumina (KF/Al O ) results in the formation of a basic
heterogeneous catalyst. This catalyst was found to be a
convenient and efficient solid support for a variety of substitu-
7,8
tion, elimination, addition, and condensation reactions. How-
ever, the source of its basicity and the identity of its catalytic
sites are still ambiguous. It is well-known that a thermal reaction
analytical methods such as titration, IR, solid-state 19F MAS
NMR and XRD.
2 3 3 6
of KF with Al O in the presence of water can give K AlF as
a product together with KOH (eq 1). Weinstock et al. have
shown that in the base-catalyzed reaction of chloroform with
m-nitrobenzaldehyde no differences in the reaction yield/kinetic
3
H O
2
1
2KF + Al O
8 2K AlF + 6KOH
(1)
2
3
3
6
were observed when KF/Al
2
O
3
versus KOH/Al
as basic solid supports. Therefore, they claimed that the basicity
of KF/Al arises only from the KOH base formed in the initial
preparation of the solid support. On the other hand, Ando and
2 3
O were used
9
In this paper we wish to present our results on the effects of
alumina-based fluoride reagents (MF/Al powders) as con-
2 3
O
10
2 3
O
1
1
venient and efficient reactive sorbents for rapid hydrolysis and
detoxification of the CWAs VX, GB, and HD. The solid-state
Clark and their co-workers demonstrated that additional to
the formation of KOH on the surface of KF/Al , the remaining
coordinatively unsaturated fluoride ions (i.e., not bound to Al
2 3
O
31
19
13
P, F, and C MAS NMR was applied as a detection
3+
technique, which allowed sample study in real time and in a
12
ions) might play a crucial role in activity of this solid support.
Intensification to this argument was obtained by a series of
1-5,15,16
nondestructive manner.
13
14
experiments performed by the same authors and others using
Results and Discussion
(
4) (a) Wagner, G. W.; Procell, L. R.; O’Connor, R. J.; Munavalli, S.; Carnes,
2 3
Destructive Adsorption of VX on KF/Al O . Scheme 1
C. L.; Kapoor, P. N.; Klabunde, K. J. J. Am. Chem. Soc. 2001, 123, 1636. (b)
Wagner, G. W.; Procell, L. R.; Munavalli, S. J. Phys. Chem. C 2007, 111, 17564.
represents two possible routes in the basic hydrolysis of VX.
This study is focused both on the P-S cleavage route (fast and
favored process), leading to the nontoxic EMPA product, and
on the P-O cleavage route that leads to the relatively toxic
desethyl-VX byproduct (slow and not favored process). In our
(
(
5) Wagner, G. W.; Chen, Q.; Wu, Y. J. Phys. Chem. C 2008, 112, 11901.
6) Spafford, R. B. The DeVelopment of a ReactiVe Sorbent for Immediate
Decontamination, ERDEC-CR-218; U.S. Army ERDEC: Aberdeen Proving
Ground, MD, 1996.
(
7) Selected reviews: (a) Blass, B. E. Tetrahedron 2002, 58, 9301. (b)
Kabalka, G. W.; Pagni, R. M. Tetrahedron 1997, 53, 7999. (c) Clark, J. H.;
Kybett, A. P.; Macquarrie, D. J. Supported Reagents Preparation, Analysis, and
Application; VCH Publishers, Inc.: New York. 1992. (d) Clark, J. H. Catalysis
of Organic Reactions by Supported Inorganic Reagents; VCH Publishers, Inc.:
New York, 1994.
2 3
study, all the reactions of VX with KF/Al O led to EMPA as
a major product and in some cases together with the undesired
desethyl-VX byproduct (as previously observed in reactions
performed in basic aqueous solution). Nevertheless, this minor
byproduct was further degraded by a second slow hydrolysis
to the final nontoxic MPA product. Obviously, in basic
conditions both EMPA and MPA products are eventually
deprotonated to their corresponding phosphonate salts.
(
8) Selected examples: (a) Veldurthy, B.; Clacens, J.-M.; Figueras, F. Eur.
J. Org. Chem. 2005, 1972. (b) Yadav, V. K.; Babu, K. G.; Mittal, M. Tetrahedron
001, 57, 7047. (c) Boullet, F. T.; Villemin, D.; Ricard, M.; Moison, H.; Foucaud,
2
A. Tetrahedron 1985, 41, 1259. (d) Kabalka, G. W.; Wang, L.; Namboodiri,
V.; Pagni, R. M. Tetrahedron Lett. 2000, 41, 5151. (e) Kabalka, G. W.; Pagni,
R. M.; Hair, C. M. Org. Lett. 1999, 1, 1423. (f) Nakano, Y.; Shi, W.; Nishii, Y.;
Igarashi, M. J. Heterocycl. Chem. 1999, 36, 33. (g) Fraile, J. M.; Garcia, J. L.;
Matoral, J. A.; Figueras, F. Tetrahedron Lett. 1996, 37, 5995. (h) Alloum, B. A.;
Villemin, D. Synth. Commun. 1989, 19, 2567. (i) Yamawaki, J.; Ando, T.;
Hanafusa, T. Chem. Lett. 1981, 1143.
In the course of this study, several types of alumina powders
2 3 2 3 2 3
(i.e., neutral γ-Al O , basic Al O , and KF/Al O ) were initially
tested as simple reactive sorbents for chemical destruction of
VX, by means of a heterogeneous basic hydrolysis. The various
(
9) Weinstock, L. M.; Stevenson, J. M.; Tomellini, R. B.; Reinhold, D. F.
Tetrahedron Lett. 1986, 27, 3845.
10) Ando, T.; Clark, J. H.; Cork, D. G.; Hanafusa, T.; Ichihara, J.; Kinura,
T. Tetrahedron Lett. 1987, 28, 1421.
11) Ando, T.; Brown, S. J.; Clark, J. H.; Cork, D. G.; Hanafusa, T.; Ichihara,
J.; Miller, J. M.; Robertson, M. S. J. Chem. Soc., Perkin Trans. 2 1986, 1133.
12) Although the term ”coordinatively unsaturated” is frequently used to
KF/Al
function of their preparation conditions. Therefore, the identi-
fication of each KF/Al support will be expressed by its KF
2 3
O supports were found to be chemically different as a
(
(
2 3
O
(
loading percent, preparation medium, and drying temperature,
respectively, summarized in the attached parenthesis. For
describe the oxidation state of transition metal ions in organometallic complexes,
we decided to use the traditional terminology ”coordinatively unsaturated fluoride
ions” (as in refs 10 and 11) for the description of free (not bound to Al atom)
example, the expression KF/Al
Al that contains 20wt % of KF, prepared in water, and finally
dried at the temperature of 160 °C.
2 3 2
O (20, H O, 160) refers to KF/
fluoride ions in the KF/Al
2
O
3
surface.
2 3
O
(
13) (a) Clark, J. H.; Goodman, E. M.; Smith, D. K.; Brown, S. J.; Miller,
J. M. J. Chem. Soc., Chem. Commun. 1986, 657. (b) Duke, C. V. A.; Miller,
J. M.; Clark, J. H.; Kybett, A. P. J. Mol. Catal. 1990, 62, 233.
(
14) (a) Ono, Y.; Baba, T. Catal. Today 1997, 38, 321. (b) Handa, H.; Baba,
(15) (a) Shaw, J. A.; Harris, R. K.; Norman, P. R. Langmuir 1998, 14, 6716.
(b) Mizrahi, D. M.; Columbus, I. EnViron. Sci. Technol. 2005, 39, 8931. (c)
Columbus, I.; Waysbort, D.; Shmueli, L.; Nir, I.; Kaplan, D. EnViron. Sci.
Technol. 2006, 40, 3952.
T.; Sugisawa, H.; Ono, Y. J. Mol. Catal. A 1998, 134, 171. (c) Baba, T.; Kato,
A.; Takahashi, H.; Toriyama, F.; Handa, H.; Ono, Y.; Sugisawa, H. J. Catal.
1
998, 176, 488. (d) Kabashima, H.; Tsuji, H.; Nakata, S.; Tanaka, Y.; Hattori,
H. Appl. Catal. A: Gen. 2000, 194-195, 227.
(16) Zafrani, Y.; Gershonov, E.; Columbus, I. J. Org. Chem. 2007, 72, 7014.
3
30 J. Org. Chem. Vol. 74, No. 1, 2009

Destruction of the Warfare Agents VX, GB, and HD
TABLE 1. Hydrolysis Rates of VX Adsorbed on Alumina-Based Powders
desethyl-VX
a
b
run
solid used
VX (wt %)
H
2
O (wt %)
t1/2 (h)
EMPA (%)
%
t1/2 (h)
1
2
3
4
5
6
7
8
9
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
2
2
2
2
2
2
2
2
O
O
O
O
O
O
O
O
3
3
3
3
3
3
3
3
(40, H
(40, H
(40, H
(20, H
(20, H
(10, H
(20, H
(20, H
2
2
2
2
2
2
2
2
O, 160)
O, 160)
O, 160)
O, 160)
O, 160)
O, 160)
O, 160)
O, 160)
5
5
5
5
5
0
5
10
5
10
5
5
5
5
5
5
∼1200
6.3
11.2
3.0
2.9
6.1
50
91
91
93
93
88
93
84
9
9
7
7
12
7
161
c
nd
nd
c
50
5
330
c
10
25
5
5
5
2.6
nd
nd
c
c
nd
5
neutral alumina
basic alumina
KOH/Al
11.2
6.6
5.6
100
100
85
1
1
0
1
c
2
O
3
(20, H
2
O,160)
15
nd
a
b
c
Half-life of pseudo-first-order reaction. Half-life of the hydrolysis of desethyl-VX to MPA. Half-life time was not determined.
The real-time kinetic results of the hydrolysis process were
expressed by a more rapid rate reaction and lower percentage
of the desethyl-VX byproduct (runs 2, 4, and 6; Figures 1 and
S3 and S5, Supporting Information). It is known that the original
surface area of the alumina starting material is not considerably
affected by the preparation procedure in cases in which a low
loading of KF (<5 mmol/g) or a low drying temperature (<200
31
extracted from the solid-state P MAS NMR spectra and are
summarized in Table 1. In a typical experiment 2 µL of VX
was carefully added, without mixing or crushing, to the
examined powder (ca. 40 mg, wetted with 5% water) placed in
31
the NMR rotor. Monitoring the reaction by P MAS NMR
spectroscopy revealed that the signals assigned for VX and
desethyl-VX side product (at 55 and 38 ppm, respectively) were
fairly narrow and well separated from the relatively broader
signals assigned for the nontoxic products EMPA and MPA (at
ca. 25 and 20 ppm, respectively). Thus, the integration of the
NMR signals of both the starting material and the products
should be proportional to the concentration of the adsorbed
compounds and therefore can be monitored for determining the
rate constants of the reactions. The broadness of the solid-state
MAS NMR signals is determined by the nature of the tested
adsorbed compound. Namely, cases in which the adsorption is
physical (compounds are mobile on the surface), rather than
chemical, will lead to relatively narrow peaks, and vise versa.
It will be shown that in the present study, this feature was
strongly affected by the reaction conditions.
1
1,13,14
°C) are used.
It seems that higher loading of KF (40%,
8.4 mmol/g) causes a decrease in the surface area of the alumina
1
1
support, while lower loading of this salt (10%) results in a
matrix with less amount of active sites. In addition, we found
that the rate of the hydrolysis was not affected by the
2 3 2
concentration of VX, using KF/Al O (20, H O, 160) with 5%
water at ambient temperature (runs 4 and 7; Figures S3 and
S6, Supporting Information). At these conditions, the concentra-
tion limit for VX full hydrolysis was found to be somewhere
between 20% and 25%, a very high concentration for any
possible scenario.
In most cases, the profiles of the hydrolysis of both VX and
desethyl-VX were found to be compatible with a pseudo-first-
order reaction, suggesting that the diffusion of the reactant
toward the active sites is the rate-determining step (Figures 1
and S2-S6, Supporting Information).
Inspection of the data presented in Table 1 revealed that KF/
Al
at ambient temperature. A typical picture of P MAS NMR
spectra for the monitoring of VX hydrolysis onto KF/Al (40,
O, 160) and its reaction profile are depicted in Figure 1. Both
O
2 3
supports can easily hydrolyze VX, at different rates and
Interestingly, subjection of VX to a commercial neutral or
basic Al O , using water as an additive (5 wt %) resulted in the
2 3
formation of the nontoxic EMPA product in a quantitative yield,
without formation of desethyl-VX byproduct (runs 9 and 10).
31
O
2 3
H
2
in terms of the reaction kinetics and the nature of the products,
the reactions were dependent on three parameters: the water
content of the powder (Table 1, runs 1-5), the percentage
loading of KF on the alumina (runs 2, 4, and 6), and the
concentration of VX in the sorbent (runs 4, 7 and 8). It was
found that the rate of the hydrolysis reaction was slowed by
either a lack (0-2%) or an excess (10%) of water added to the
Moreover, the hydrolysis rate under the action of basic Al
was fairly close to that of the more basic KF/Al (40, H
160) at the same conditions (runs 2 and 10). In addition, from
a kinetic point of view, the performance of wetted neutral Al O
3
(5 wt %) in its ability to hydrolyze VX at ambient temperature
is impressive by itself. It seems that the difference between our
2 3
O
2
O,
O
2 3
2
experiments with neutral Al
2 3
O and the reactions of VX with
4
KF/Al
2
O
3
(40, H
2
O, 160) or KF/Al
2
O
3
(20, H
2
O, 160) powders
nanosize Al , reported by Wagner et al., stems from the
2 3
O
prior to the addition of VX (runs 1-5). It is reasonable to
assume that excess of water molecules may block the active
sites of the solid catalyst or enhance its hydrophilicity, in such
a way that the diffusion of VX is slowed down. On the other
hand, in the absence of water molecules the hydrolysis process
is very slow, despite the amount of hydroxide ions available in
presence or the absence of water as an additive. This feature
may affect the mechanism of the hydrolysis and therefore the
rate of the reaction and the nature of the products (vide infra).
One way to estimate the effectiveness of KF/Al
heterogeneous basic catalyst for organic transformations is to
compare it with KOH/Al support, considering both the
selectivity and the kinetics of the reaction. Thus, comparing
the hydrolysis of VX onto KF/Al (20, H O, 160) vs KOH/
Al (20, H O, 160) revealed that the former support is more
2 3
O as a
2 3
O
2 3 2
the KF/Al O (40, H O, 160) support (run 1). This observation
implies that the hydrolysis occurred by the assisting of water
additive in a cooperative mechanism (vide infra). Thus, the
addition of water (5%) as a reactant, which did not affect the
flowing nature of the powder, is necessary for obtaining optimal
reproducible results at ambient temperature.
O
2 3
2
O
2 3
2
effective not only by terms of reaction rate (ca. 2-fold higher)
but also in terms of products distribution (runs 4 and 11). Taking
into account that the reaction of KF with Al
0.5 equiv of KOH depending on the preparation conditions,
2 3
O may produce a
7
The reaction studies indicate that 20% KF loading (compared
to 10% and 40%) results in a more effective hydrolysis reagent,
KOH/Al O
2 3
(20, H
2
O, 160) contains more hydroxide ions per
J. Org. Chem. Vol. 74, No. 1, 2009 331

Gershonov et al.
FIGURE 1. Selected 31P MAS NMR spectra of adsorbed VX (5 wt %) on KF/Al
degradation profiles of VX (bottom left) and desethyl-VX (bottom right) onto this sorbent.
O
2
3
2 2
(40, H O, 160) using H O (5%) as an additive (top) and
weight than that of KF/Al
. Therefore, comparing the hydrolysis of VX onto KF/Al
40, H O, 160) versus KOH/Al (20, H O, 160) is more
2
O
3
(20, H
2
O, 160) according to eq
160) gave a single peak at -158 ppm that is attributed to K
3
AlF
6
13,14
1
(
O
2 3
(Figure 2a),
whereas in the spectrum of KF/Al
2
O
3
2
(20, H O,
2
O
2 3
2
100), an additional peak appeared at -121 ppm (Figure 2b).
The assignment for this signal is KF solvated by water (drying
at 100 °C leaves out sufficient amounts of water molecules for
this solvation), which means that not all of the fluorine atoms
indicative regarding the role of other species than hydroxide
ions (such as coordinately unsaturated fluoride ions), in deter-
mining both kinetics and products of the reaction. Inspection
of the data presented in Table 1 (runs 2 and 11) reveals that the
reaction rates on these two solid supports are fairly close, but
the product distribution is somewhat different (9% vs 15% of
the side product desethyl-VX, respectively). This observation
clearly shows that in terms of the reaction kinetics, there is no
difference between these two solid supports, and therefore, it
seems that the involvement of “free” fluoride ions in the
1
3
are bound to the aluminum atoms. Moreover, when the
2 3 2
hydrolysis of VX with KF/Al O (20, H O, 100) was monitored
by both 31P and F-MAS NMR spectroscopy, this signal
disappeared in parallel to the reaction progression (Figure 3,
right). Obviously, this is direct evidence for the participation
of the coordinatively unsaturated fluoride ions in an organic
19
reaction, catalyzed by KF/Al
2
O
2
3
. In view of the fact that KF/
(20, H O, 100) are different
acceleration of VX hydrolysis onto KF/Al
is particularly minor.
O
2 3
(40, H
2
O, 160)
Al (20, H O, 160) and KF/Al
O
2 3
2
O
3
2
not only by their performance to accelerate VX hydrolysis but
also by their chemical composition, we chose to term the latter
From the above-mentioned results one may deduce that
standard KF/Al compared to KOH/Al or even commercial
O
2 3
O
2 3
catalyst as ECUF-KF/Al
with coordinatively unsaturated F ions.
2 3 2 3
O , which means KF/Al O enriched
-
basic alumina as solid supports for VX hydrolysis showed no
significant advantages. Therefore, we decided to further inves-
tigate the performance of this inorganic catalyst by modification
of its preparation conditions.
In view of this result, we attempted a different preparation
approach in which dry alumina and dry KF are ground together
without the involvement of water. This may prevent the
KF/Al
2
O
3
Enriched with Coordinatively Unsaturated
F ions (ECUF-KF/Al ). Interestingly, the hydrolysis of VX
onto KF/Al (20, H O, 100), dried at 100 °C instead of 160
C, was found to be 4-fold faster than that performed onto its
KF/Al (20, H O, 160) counterpart (Table 2, runs 1 and 2).
These two solid supports were compared in their solid-state F-
MAS NMR spectroscopy. The spectrum of KF/Al (20, H O,
formation of K
support. Indeed, this KF/Al
characteristic peak for dry KF at -132 ppm (Figure 2e).
However, this solid support did not exhibit significant advantage
in the hydrolysis of VX, compared to the standard KF/Al
(20, H O, 160), except for the fact that the process was
completed with no formation of desethyl-VX (Table 2, runs 1
3
AlF
6
and enrich the KF component in the solid
-
O
2 3
2 3
O
(20, dry, 160) gave only the
1
3
2
O
3
2
°
O
2 3
2
2 3
O
1
9
2
O
2 3
2
3
32 J. Org. Chem. Vol. 74, No. 1, 2009

Destruction of the Warfare Agents VX, GB, and HD
2 3
TABLE 2. Hydrolysis Rates of VX Adsorbed on ECUF-KF/Al O Powders
desethyl VX
a
b
run
KF/Al
O
2 3
used
added H
2
O (%)
t
1/2 (h)
EMPA (%)
%
t1/2 (h)
c
1
2
3
4
5
6
7
8
9
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
KF/Al
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
O
2 3
(20, H
(20, H
(40, H
(40, MeOH, 160)
(40, MeOH, 160)
(40, MeOH, 160)
(40, MeOH, 160)
(20, MeOH, 160)
(20, EtOH, 160)
(20, dry, 160)
2
2
2
O, 160)
O, 100)
O, 160)
5
5
5
0
2
5
10
5
5
5
2.7
0.7
6.3
21
2.2
0.26
0.5
0.26
0.22
4.6
93
93
91
100
100
96
94
96
100
96
7
7
9
nd
nd
c
161
d
d
4
6
4
2.2
5.6
2.6
1
0
a
b
c
d
Half-life of pseudo-first-order reaction. Half-life of the hydrolysis of desethyl VX to MPA. Half-life time was not determined. This is the first
and short t1/2 originating from two half-life times observed by the reaction profile of these reactions (see Figures S8 and S9, Supporting Information).
direct evidence to the participation of coordinatively unsaturated
fluoride ions in the hydrolysis process was achieved by
monitoring the reaction with both 31P and F MAS NMR
19
(
Figure 3, left). Interestingly, in these reactions the compound
O-ethyl-O-methyl methylphosphonate (EMMP) was also ob-
served as a minor byproduct resulting from traces of methoxide
ions present in this solid support (see Figures 4 and S7,
Supporting Information).
Similarly, preparation of KF/Al
ethanol was used as a solvent, left almost all free fluoride ions
in the sorbent, and only a small quantity of K AlF was formed
Figure 2d). The hydrolysis of VX onto this support was found
to be faster than that of KF/Al (20, MeOH, 160), supposedly
2 3
O (20, EtOH, 160), in which
3
6
(
2 3
O
FIGURE 2. 19F MAS NMR spectra of several KF/Al
due to the present of large amounts of coordinatively unsaturated
fluoride ions (Table 2, run 9).
2 3
O sorbents.
Effect of Silver Ion by AgF Impregnation. These results
encouraged us to further improve the activity of the powder by
investigating the effect of silver ion impregnated on the alumina-
based fluoride reagent on the hydrolysis of VX. The rationale
of this approach is based on previous results reported by Wagner
3
et al., who demonstrated that AgY zeolites can chemically
destruct VX agent, by the well-known activation of the P-S
1
7
bond. Indeed, impregnation of AgF instead of KF in neutral
alumina resulted in the formation of an improved solid support
FIGURE 3. 19F MAS NMR spectra of KF/Al
and KF/Al (20, MeOH, 160) (left): (a) the dry sorbent, (b) after
addition of 5% water, and (c) after the reaction with VX.
2 3 2
O (20, H O, 100) (right)
to the VX hydrolysis purpose. Thus, when using AgF/Al
40, H O, 160), the hydrolysis of VX at the same conditions
was 3.9-fold faster compared to KF/Al (40, H O, 160) as
2 3
O
2 3
O
(
2
O
2 3
2
and 10). We assume that despite the stringent grinding of the
two dry solids, the dispersion of free fluoride ions in the pores
of the alumina matrix is limited intrinsically by the size of KF
particles. In order to enhance the dispersity of KF in the KF/
Al O support along with the prevention of K AlF formation,
2 3 3 6
the latter was prepared in dry methanol instead of water and
solid support (Table 3, runs 1-2). Moreover, a composition of
AgF (5 wt %), KF (35 wt %), and neutral alumina gave a very
effective solid support, in which the hydrolysis of VX occurred
more than 60-fold faster (with t1/2 < 0.1 h) when compared to
the usual KF/Al O (Figure 5; Table 3, run 4). Expectedly, the
2 3
hydrolysis of desethyl-VX was also enhanced using this solid
support (Figure 5, right; see also Figures S10 and S11,
then dried at 160 °C. The remaining KF peak at -132 ppm
implied that again this modification in KF/Al O preparation
2 3
Supporting Information). In dry conditions, the AgF/KF/Al
5, 35, H O, 160) powder also chemically destroyed VX, in a
relatively fast manner (Table 3, run 6).
The reaction profiles for the VX hydrolysis onto selected KF/
sorbents depicted in Figure 6 clearly show the influence
of these procedural modifications on the reaction kinetics. The
2 3
impressive performance of AgF/KF/Al and ECUF-KF/Al O
2 3
O
left out a significant amount of free fluoride ions dispersed in
the sorbent (Figure 2c). These catalysts, i.e., KF/Al (40,
MeOH, 160) and KF/Al (20, MeOH, 160), were found to
(
2
2 3
O
2 3
O
be much more reactive both in the hydrolysis of VX and
desethyl-VX (Figure 4), than their “aqueous preparation”
counterparts (runs 1, 3 vs 8, 6, respectively). The addition of
2 3
Al O
O
2 3
5
% water as additive is recommended in order to accomplish
in the hydrolysis-based chemical destruction of both VX and
desethyl-VX strongly emphasizes their potential utilization as
a destructive powder for the decontamination of CWA. To the
best of our knowledge, the hydrolysis process onto wetted (5%)
the hydrolysis process in a fast manner (Table 2, runs 4-7; see
also Figures S7-S9, Supporting Information). The same
reactivity of the higher loading KF/Al
compared to its lower loading counterpart KF/Al
60) revealed that this preparation method did not affect the
surface area of the support by high KF impregnation. Again,
O
2 3
(40, MeOH, 160)
2 3
O
(20, MeOH,
1
(
17) Cotton, F. A.; Wilkinson, G. In AdVanced Inorganic Chemistry, 5th ed.;
John Wiley & Sons: New York, 1988; pp 942-943.
J. Org. Chem. Vol. 74, No. 1, 2009 333

Gershonov et al.
FIGURE 4. Selected 31P MAS NMR spectra of adsorbed VX (5 wt %) on KF/Al
O
(40, MeOH, 160) using H
2
3
2
O (5%) as an additive (top) and
degradation profiles of VX (bottom left) and desethyl-VX (bottom right) onto this sorbent.
TABLE 3. Hydrolysis Rates of VX Adsorbed on Alumina-Supported AgF Powders
desethyl-VX
a
b
run
solid used
(40, H O, 160)
(40, H O, 160)
H
2
O (wt %)
t
1/2 (h)
EMPA (%)
(%)
t
1/2 (h)
161
1
2
3
4
5
6
KF/Al
AgF/Al
2
O
3
2
5
5
10
5
2
0
6.3
1.6
91
99
90
90
96
99
9
1
10
10
4
2
O
3
2
AgF/KF/Al
AgF/KF/Al
AgF/KF/Al
AgF/KF/Al
O
2 3
O
2 3
O
2 3
O
2 3
(5, 35, H
(5, 35, H
(5, 35, H
(5, 35, H
2
O, 160)
2
O, 160)
2
O, 160)
2
O, 160)
0.28
2.8
1.8
4.8
<0.1
0.53
22.4
a
b
Half-life of pseudo-first-order reaction. Half-life of the hydrolysis of desethyl-VX to MPA.
FIGURE 5. Selected 31P MAS NMR spectra of adsorbed VX (5 wt %) on AgF/KF/Al
and degradation profile for desethyl-VX byproduct onto this adsorbent (right).
2 3 2 2
O (5, 35, H O, 160) using H O (5%) as an additive (left)
KF/Al
2
O
3
(20, EtOH, 160) or AgF/KF/Al
2
O
3
(5, 35, H
2
O, 160)
Mechanism. It was concluded by Ando and Clark et al. that
organic reactions onto KF/Al might take place by three
possible mechanisms: (a) a direct reaction with free active
fluoride ion as a strong base, (b) a mechanism in which fluoride
ion replaces a hydroxyl group [Al-OH] that cooperatively reacts
supports is the fastest solid-supported chemical detoxification
of VX yet found. One should add that these powders are very
simple for preparation using the stable, environmentally benign,
and low-cost starting materials alumina and potassium fluoride.
2 3
O
3
34 J. Org. Chem. Vol. 74, No. 1, 2009

Destruction of the Warfare Agents VX, GB, and HD
FIGURE 7. (left) The line width (lw, in Hz) of the 31P-MAS NMR
EMPA signals (right) from the KF/Al
with different amounts of water.
2 3
O (40, MeOH, 160) sorbents
2 3
In the course of our study we have found that KF/Al O can
chemically destroy VX even in dry conditions in a relatively
fast manner. For example, the half-life of VX is only 21 h
without the formation of desethyl-VX when subjected to dry
2 3
FIGURE 6. Reaction profiles for VX hydrolysis on several KF/Al O
sorbents.
SCHEME 2
2 3
KF/Al O (40, MeOH, 160), as shown in Table 2 (run 4).
Wagner et al. recently reported on the reaction of VX with
4
nanosize-Al O
2 3
2 3
and with conventional γ-Al O . The reactions
were performed without the addition of water to the sorbents.
The products were characterized as aluminophosphonate com-
pounds, resulting from an attack of aluminum oxide on VX,
without the formation of desethyl-VX. We assume that similar
to Wagner’s results, VX reacts with the aluminum oxide species
2 3
in dry ECUF-KF/Al O , to form the corresponding alumino-
phosphonate product (Scheme 2, bottom).
Solid-state MAS NMR experiments of organic compounds
adsorbed on an inorganic matrix may differentiate efficiently
between mobile or soluble materials to immobile ones. When
these compounds are physically but not strongly bound to the
matrix, their signals in solid-state MAS NMR will appear as
narrow and sharp peaks. On the other hand, when organic
compounds react directly with the inorganic matrix, the result
of this chemisorption in solid-state MAS NMR may be
4a
characterized by broad peaks. A confirmation for the formation
of soluble and mobile EMPA product in the wetted experiments
versus the formation of the rigid aluminophosphonate compound
in the dry test was obtained by the line width 31P-MAS NMR
analysis of the processes. Figure 7 shows that the line width of
as a strong base, and (c) a reaction with the hydroxide ion
-
7,10,11
according to eq 1 or by [Al-O ] when water is added.
Our observations in the hydrolysis of VX promoted by wetted
KF/Al revealed that these three mechanisms might actually
be valid, depending on the KF/Al used (Scheme 2, top).
Sorbent that was prepared in water and dried in relatively high
temperature, such as KF/Al (20, H O, 160), did not contain
coordinatively unsaturated fluoride ions but KOH and K AlF
31
P-MAS NMR signals of the product obtained by reactions of
VX with KF/Al (40, MeOH, 160) was found to be strongly
2 3
O
2 3
O
2
O
3
dependent on the percentage of water added prior to the reaction.
The product signal at 24.4 ppm in the reaction of VX with dry
KF/Al O (40, MeOH, 160) was found to be much wider than
2 3
those of the wetted experiments, and therefore it may result
from direct reaction with the basic alumina support. The same
tendency was observed in the reaction of VX with AgF/KF/
O
2 3
2
3
6
,
as shown in Figure 2a, and thus, its reaction with VX could
comply with mechanism (c) in Scheme 2 (top). Therefore, its
performance in the hydrolytic decontamination of VX was close
2 3 2
Al O (5, 35, H O, 160) in the presence of 0, 2%, 5%, and 10%
to that of the usual KOH/Al
Al that was prepared in alcoholic solvents or dried at low
temperature, such as KF/Al (20, MeOH, 160), KF/Al (20,
EtOH, 160), or KF/Al (20, H O, 100), could also react via
mechanisms (a) and (b) and therefore might enhance the
2 3
O . On the other hand, ECUF-KF/
water (Table 3, see also Figures S10-S12, Supporting Informa-
tion). Nevertheless, we cannot exclude the possibility that this
broad signal may partially be attributed to an insoluble EMPA
salt (from deprotonation of free EMPA product when the
displacement occurred by a hydroxide ion).
2 3
O
O
2 3
2 3
O
O
2 3
2
1
8
reactivity of the sorbent. In mechanism (a) HF is formed but
immediately reacts with the surface of the alumina support to
Centrifugation Effect by the MAS. Recently we reported
that in some protodesilylation reactions using montmorillonite
1
1
3 6
give K AlF .
1
KSF, the rates measured by solid-state MAS H NMR were
found to be higher as compared to those measured by solution
H NMR of the same reaction mixture. We assumed that these
solvent-free processes, in which the diffusion of reactants toward
the active sites of the solid support plays a significant role in
the reaction kinetics, may be affected by a centrifugation effect
(
18) Recently it was reported by Sandhage and co-workers on a rapid
hydrolysis of some nerve agent mimics induced by nanostructural fluorine-doped
titania (TiOF
1
16
2
) under mild conditions (pH 4.5-7.9, 22 °C). However, this
transformation is catalyzed by Lewis acidic sites of the solid support. See: Lee,
S.-J.; Huang, C.-H.; Shian, S.; Sandhage, K. H. J. Am. Ceram. Soc. 2007, 90,
632.
1
J. Org. Chem. Vol. 74, No. 1, 2009 335

Gershonov et al.
SCHEME 3
SCHEME 4
caused by the high spinning rate of the sample (ca. 8000 Hz).
In decontamination context it is important to know whether the
above-mentioned real-time reaction rates measurements of VX
hydrolysis are not affected by the MAS. Therefore, we
performed two different examinations on the hydrolysis reaction
of VX with KF/Al
60). In one experiment, the hydrolysis rate was tested on two
identical samples of wetted (5%) KF/Al (20, H O, 160),
when one sample was studied under MAS NMR conditions
using 8000 Hz spinning rate and the other sample was measured
without spinning. The rates of both reactions were found to be
similar. In another experiment, the hydrolysis rate of VX was
2 3 2 2 3
O (20, H O, 160) and KF/Al O (20, MeOH,
1
feature of the sulfur atom. The high reactivity of HD toward
nucleophilic substitutions results from the equilibrium state with
its cyclic sulfonium (CS) salt. However, this feature may
enhance its environmental stability (in neutral pH) due to the
tendency to equilibrate with its toxic sulfonium products such
O
2 3
2
2
0
as CH-TDG (vide infra). In addition, the oxidation reactions
of HD would form the relatively toxic product bis-2-chloroethyl
2 3
tested onto a sample of wetted (5%) KF/Al O (20, MeOH, 160)
2
1
under standard MAS NMR conditions, as mentioned above.
Once the concentration of VX was totally decayed, a second
sample of this sorbent, that was left aside, was extracted. The
reaction time for full conversion of VX measured by the MAS
NMR was found to be similar to its counterpart without
sulfone, and therefore, hydrolysis processes seem to be
favored. These features emphasize the challenge and the
importance of a quest for active powders suitable for HD
detoxification.
Preliminary results revealed that HD (ca. 7%) reacted with
spinning. Thus, the hydrolysis rates of VX by KF/Al
O
2 3
2 3 2 2 3
wetted (5%) KF/Al O (20, H O, 160) and KF/Al O (20,
13
measured by the MAS NMR technique are indeed results of
real-time kinetics. In light of these experiments, we assume that
the centrifugation effect observed by us previously is not general
for all solid-state organic reactions and is highly dependent on
both the reactants and the solid supports used.
Destructive Adsorption of GB (Sarin) on Alumina-Sup-
ported Fluoride Reagents. When we tested the reaction of the
highly toxic GB on various alumina-supported fluoride powders
using water as an additive, the result was immediate hydrolysis
MeOH, 160) to form several products as shown by the C-
MAS NMR spectra (Scheme 4 and Figure 8). The spectral
picture of the kinetic monitoring was found to be similar for
22
both reactions, and their estimated half-life times were 96 and
24 h, respectively. The products peaks were evident for
2
-chloroethyl vinyl sulfide (CEVS, 132.3, 113.4, 43.7, 34.7
ppm), divinyl sulfide (DVS, 130.8, 115.4 ppm), 2-hydroxyethyl
vinyl sulfide (HOEVS, 133.5, 111.7 ppm, vinylic carbons),
thiodiglycol (TDG, 62.1, 35.3 ppm) and 1,4-thioxane (72.0, 33.0
3
ppm). The reaction initiated by a base-catalyzed elimination
(
t1/2 < 10 min). For these experiments GB (5%) was added to
wetted (5%) KF/Al (20, H O, 160), KF/Al (20, MeOH,
60), or AgF/KF/Al (5, 35, H O, 160) supports, without
of HCl or by a substitution reaction of chloride via the CS form,
to yield the appropriate monovinyl (CEVS) or chlorohydrin
O
2 3
2
2 3
O
1
2
O
3
2
(
CH) products, respectively (Scheme 5). The former could
mixing or crushing. After 20-40 min the hydrolysis reactions
were completed, so that no starting material or intermediate
could be detected. The single product in all cases was isopropyl
undergo two base-catalyzed alternative routes, i.e., another
elimination of HCl or a nucleophilic substitution of the
remaining chloride that led to DVS and HOEVS, respectively.
The CH intermediate could eventually lead to the nontoxic
product, 1,4-thioxane (major), by a base-catalyzed intramolecu-
lar substitution of the remaining chloride. The small amount of
the TDG product (7%) could be a result of the intramolecular
versus intermolecular reaction preferred in the heterogeneous
conditions that resembles highly diluted reactions. It has to be
noted that toxic sulfonium ion product, such as CH-TDG
31
methylphosphonic acid (IMPA), as emerged from the P MAS
NMR spectra at ca. 24.0 ppm (Scheme 3, see also Figure S13,
Supporting Information). These results are in accordance with
the high volatility and the good water-solubility of GB, as well
as its susceptibility to basic hydrolytic conditions.
It is well-known that fluoride ion impregnated on alumina
7
,8
reacts as a strong base. Therefore, the reaction of VX with
all variations of KF/Al supports by no means led to the
2 3
O
(
resulting from a nucleophilic attack of CH intermediate by TDG
corresponding ethyl sarin product (G-analogue), in opposition
3
,20
6
product),
which might be produced in aqueous solutions or
to its reaction with neat AgF salt. In any case, the very high
less basic solid supports was not detected in our sorbents.
rate of the hydrolysis of GB onto the above-mentioned alumina-
supported fluoride powders reveals that even if the undesired
G-analogue would partially form during the reaction, it may
actually appear as an unstable intermediate.
13
Despite the use of C-unlabeled HD in this preliminary study,
it was clearly shown that with long acquisition times, the spectral
picture of relatively low-rate reactions could be analyzed.
Detoxification of Sulfur Mustard (HD). Despite the struc-
tural simplicity of the blister agent HD (bis-chloroethyl sulfide),
its chemistry is abundant and includes some elementary
processes such as nucleophilic displacements, eliminations,
additions and oxidations, as a function of the reaction condi-
tions. This reactivity arises from the two electrophilic centers
activated by the sulfur moiety, relatively acidic protons R to
the sulfide group and ꢀ to the chlorides, and the nucleophilic
(
19) (a) Black, R. M.; Brewster, K.; Harrison, J. M.; Stanfield, N. Phosphorus,
Sulfur Silicon 1992, 71, 31. (b) Black, R. M.; Brewster, K.; Harrison, J. M.;
Stanfield, N. Phosphorus, Sulfur Silicon 1992, 71, 49.
(20) Yang, Y.-C.; Szafraniec, L. L.; Beaudry, W. T.; Ward, J. R. J. Org.
Chem. 1988, 53, 3293.
(21) Yang, Y.-C.; Baker, J. A.; Ward, R. Chem. ReV. 1992, 92, 1729.
1
9
(
2 3
22) The reactions of HD with KF/Al O supports were performed with
unlabeled HD. Therefore, the long acquisition times that were required for the
C NMR measurements may cause an inaccuracy in their half-life time
determination.
13
3
36 J. Org. Chem. Vol. 74, No. 1, 2009

Destruction of the Warfare Agents VX, GB, and HD
FIGURE 8. Selected 13C MAS NMR spectra of adsorbed HD (5 wt %) on KF/Al
recorded at (a) 3, (b) 48, (c) 96, (d) 168, and (e) 552 h after sample preparation.
2 3 2 2
O (20, H O, 160) with H O (5%) as an additive. Spectra were
SCHEME 5
However, in order to extend our insights in this process, such
as for example studying the effect of Ag+ ions on its kinetic/
products or the influence of water or metal ion content in the
called coordinatively unsaturated fluoride ions (termed by us
as ECUF-KF/Al O ) was found to be very reactive for a
2
3
hydrolysis-based chemical destruction of VX, GB, and HD.
While sarin (GB) reacted instantaneously with these solid
supports to yield nontoxic product IMPA, the less reactive CWA
VX gave under the same conditions the nontoxic EMPA as a
major product (>90%; t1/2 range 0.1-6.3 h) together with the
toxic desethyl-VX (<10%) byproduct. The latter was further
hydrolyzed to the nontoxic MPA product (t1/2 range 2.2-161
1
3
solid support, further investigation with C-labeled mustard
HD*) is underway.
(
Conclusions
The surface area of alumina-supported fluoride reagents such
as KF/Al O , AgF/KF/Al O , and KF/Al O enriched by so-
2 3 2 3 2 3
J. Org. Chem. Vol. 74, No. 1, 2009 337

Gershonov et al.
h). The enhanced activity of wetted ECUF-KF/Al
O
2 3
sorbents
addition of CWA. The same procedure was performed in the
2 3 2 3
preparation of AgF/Al O and AgF/KF/Al O .
was related to the free fluoride ions, which act as a strong base.
With well-dried alumina-supported fluoride reagents, the reac-
tion with VX proceeded slowly and gave immobile EMPA that
was strongly bound to the solid supports. The reactions of HD
3
1
13
19
NMR. P, C, and F MAS NMR spectra were obtained at
2
02, 125, and 471 MHz, respectively, on a 11.7 T (500 MHz)
spectrometer, equipped with a 0.4 cm standard CP-MAS probe,
using direct polarization (i.e., no cross polarization (CP) was used).
with both KF/Al
O
2 3
and ECUF-KF/Al
O
2 3
were found to be
31
Typical spinning rates were 8-10 kHz. Chemical shifts for P,
slower (estimated t1/2 are 96 and 24 h, respectively) and gave
the nontoxic product 1,4-thioxane as a major product together
with minor amounts of TDG and vinylic compounds. These
13
19
C, and F were referenced to external trimethyl phosphate (TMP),
3
1
TMS, and CFCl , respectively, as 0 ppm. For P spectra the pulse
3
delay was 2 s, which is considered sufficient for relaxation in OP
1
-5,15,16
esters on solid matrices. The number of transients per spectrum
synthetic and mechanistic studies as well as others
1
3
was 64. For C spectra the pulse delay was 1 s, and the number of
transients per spectrum varied between 20 000 and 100 000. For
clearly show that the solid-state MAS NMR technique can serve
as a powerful tool for real-time study of organic reactions
performed on the surface of solid supports. Inter alia we have
demonstrated that its unique features such as the capability to
differentiate between mobile to immobile components onto the
solid surface can efficiently be utilized to probe the mechanism
and the products of solid-state organic reactions.
1
9
F spectra pulse delay was 2 s, and the number of transients per
spectrum varied between 100 and 1000. For comparison purposes,
spectra were recorded under identical conditions.
Sample Preparation. Caution: These experiments should only
be performed by trained personnel using applicable safety proce-
dures. Samples of the appropriate powder (ca. 40 mg) were added
to the 0.4 cm ZrO
HD) was applied via syringe to the center of the sample. The rotor
was sealed with a fitted Kel-F cap. P and C MAS NMR spectra
were measured periodically to determine remaining starting material
and identify degradation products.
2
rotor, and 2-5 µL of the CWA (VX, GB, or
3
1
13
Experimental Section
Materials. KF/Al
of KF onto γ-alumina from aqueous or dry alcoholic solutions. For
example, KF/Al (20, H O, 160) was prepared by dissolving 60 g
of KF in water (ca. 400 mL) in a 1 L flask, then 240 g of γ-alumina
neutral) was added, and the bulk of water was removed by
2 3
O reagents were prepared by impregnation
2
Supporting Information Available: Figures S2-S12, 31P
O
2 3
MAS NMR spectra and reaction profiles of adsorbed VX onto
(
31
selected KF/Al
O
2 3
powders; Figure S13, P NMR spectra of
evaporation at ca. 90 °C. The powder was dried at 160 °C for 24 h.
In cases in which the reaction was performed with wetted solid
supports, the appropriate amount of water was added to the powder,
which then was crushed for 5 min (using a glass rod) prior to the
GB before and after addition to KF/Al
2 3
O . This material is
available free of charge via the Internet at http://pubs.acs.org.
JO8019972
3
38 J. Org. Chem. Vol. 74, No. 1, 2009