Degradation of sulfur mustard on KF/Al2O3 supports: Insights into the products and the reactions mechanisms

Article abstract of DOI:10.1021/jo901713c

(Chemical Equation Presented) The degradation of the warfare agent sulfur mustard (HD) adsorbed onto KF/Al₂O₃ sorbents is described. These processes were explored by MAS NMR, using ¹³C-labeled sulfur mustard (HD*) and LC-M

Full text of DOI:10.1021/jo901713c

pubs.acs.org/joc
agent fromdiversesurfaces isrequiredinorder toregainuse of
Degradation of Sulfur Mustard on KF/Al2O3
Supports: Insights into the Products and the Reactions
Mechanisms
the affected area and equipment. The use of reactive sorbents
that chemically destroy CWAs, rather than physically remov-
ing them, is an attractive approach to solve this problem.3
Oxidation-based decontamination of HD would form the
relatively toxic products bis-2-chloroethyl sulfoxide and sul-
fone2,4 (even though they possess reduced volatility and
improved water solubility). Therefore, hydrolysis processes,
which lead to nontoxic products, are favored.
Yossi Zafrani,*,† Michael Goldvaser,† Shai Dagan,‡
Liron Feldberg,‡ Dana Mizrahi,† Daniel Waysbort,†
Eytan Gershonov,† and Ishay Columbus*,†
†Department of Organic Chemistry and ‡Department of
Analytical Chemistry, Israel Institute for Biological Research,
Ness-Ziona 74100, Israel
Recently, we reported the facile hydrolysis-based detoxi-
fication of the CWAs VX (O-ethyl S-2-(diisopropylamino)-
ethyl methylphosphonothioate), GB (O-isopropyl methyl-
phosphonofluoridate or sarin), and HD upon reaction with
various solid-supported fluoride reagents such as KF/Al2O3,
AgF/KF/Al2O3, and KF/Al2O3 enriched by so-called coor-
dinatively unsaturated fluoride ions (ECUF-KF/Al2O3).3b
The difference between ECUF-KF/Al2O3 (prepared in dry
methanol or ethanol) and regular KF/Al2O3 (prepared in
water) is that the former contains large amounts of “free”
fluoride ions, which may react as a base, while the latter
sorbent mainly contains K3AlF6 (nonreactive) and KOH, as
previously observed by 19F SS MAS NMR.3b The research
was focused on solvent-free degradations of the nerve agents
VX and GB, which were effectively monitored and studied
by a real-time solid-state magic angle spinning (SS MAS) 31P
NMR. The reactions of HD with these active sorbents
were also preliminarily investigated. It was found that HD
(ca. 6.5 wt %) reacts with wetted (5% H2O) KF/Al2O3 (20,
H2O, 160)5 or KF/Al2O3 (20, MeOH, 160) (defined as
ECUF-KF/Al2O3) to form several products. The products
were identified by 13C-MAS NMR as 2-chloroethyl vinyl
sulfide (CEVS), divinyl sulfide (DVS), 2-hydroxyethyl vinyl
sulfide (HOEVS), thiodiglycol (TDG), and 1,4-thioxane.
These reactions were performed with unlabeled HD, result-
ing in low sensitivity and requiring long acquisition times for
13C NMR measurements, which prevent kinetics measure-
ments and mechanistic study. To extend our insights into this
process, such as, for instance, studying its products, kinetics,
and mechanism as a function of the water, metal ion, and HD
content or to examine the possibility of a centrifugation
effect,6 further investigation with 13C-labeled mustard
(HD*) was required. Herein, we wish to disclose our results
on the degradation of HD* adsorbed onto alumina sup-
ported fluoride reagents with regard to all of the above-
mentioned chemical applications and mechanistic aspects.
In the present study, two types of alumina-supported fluor-
ide reagents, i.e., KF/Al2O3 (20, H2O, 160) and KF/Al2O3
yossiz@iibr.gov.il; ishayc@iibr.gov.il
Received August 6, 2009
The degradation of the warfare agent sulfur mustard
(HD) adsorbed onto KF/Al2O3 sorbents is described.
These processes were explored by MAS NMR, using
13C-labeled sulfur mustard (HD*) and LC-MS tech-
niques. Our study on the detoxification of this blister
agent showed the formation of nontoxic substitution and
less-toxic elimination products (t1/2 = 3.5-355 h). Inter-
estingly, the reaction rates were found to be affected by
MAS conditions, i.e., by a centrifugation effect. The
products and the mechanisms of these processes are
discussed.
Detoxification of the chemical warfare agent (CWA) HD
(sulfur mustard) is a current concern. Its copious chemistry,
which includes some elementary processes such as nucleo-
philic displacements, eliminations, and oxidations, depends
on the specific reaction conditions.1 In spite of its high
reactivity, which is largely attributed to the anchimerically
assisted leaving groups, HD exhibits enhanced environmental
stability (in neutral pH) because of its tendency to equilibrate
with the corresponding toxic sulfonium products, such as
CH-TG ((HOCH2CH2)2SþCH2CH2SCH2CH2OH).1 Rapid
decontamination of this environmentally persistant2 blister
(3) For recent examples for adsorption/destruction of CWA onto the
surface of reactive powders see: (a) Wagner, G. W.; Procell, L. R.;
Munavalli, S. J. Phys. Chem. C 2007, 111, 17564. (b) Gershonov, E.;
Columbus, I.; Zafrani, Y. J. Org. Chem. 2009, 74, 329. (c) Wagner, G. W.;
Chen, Q.; Wu, Y. J. Phys. Chem. C 2008, 112, 11901. (d) Prasad, G. K.; Singh,
B.; Ganesan, K.; Batra, A.; Kumeria, T.; Gutch, P. K.; Vijayaraghavan, R.
J. Hazard. Mater. 2009, 167, 1192.
(4) Yang, Y.-C.; Baker, J. A.; Ward, R. Chem. Rev. 1992, 92, 1729.
(5) The expression KF/Al2O3 (20, H2O, 160) refers to KF/Al2O3 that
contains 20 wt % of KF, prepared in water, and finally dried overnight at
160 °C. See also ref 3b.
(6) Zafrani, Y.; Gershonov, E.; Columbus, I. J. Org. Chem. 2007, 72,
7014.
(1) (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.
(2) Munro, N. B.; Talmag, S. S.; Griffin, G. D.; Waters, L. C.; Watson,
A. P.; King, J. F.; Hauschild, V. Environ. Health Perspect. 1999, 107, 933.
8464 J. Org. Chem. 2009, 74, 8464–8467
Published on Web 10/09/2009
DOI: 10.1021/jo901713c
r
2009 American Chemical Society

Zafrani et al.
JOCNote
SCHEME 1. Degradation Pathways for HD on KF/Al2O3
Sorbents
(20, EtOH, 160), were tested as simple reactive sorbents for the
chemical destruction of HD. In a typical experiment, appro-
priate amounts of HD* were carefully added to the examined
powder (ca. 100 mg, dry or wetted with 5% water, placed in the
NMR rotor) without mixing or crushing. In most cases, the
13C MAS NMR spectra of the final reaction mixture revealed a
somewhat complicated mixture of vinylic (CEVS, HOEVS,
DVS), TDG, TDG-oligomers, and thioxane products, that
were obtained in various ratios, depending on the sorbent and
the reaction conditions used (Scheme 1). The identification
of these products was based on both 13C SS MAS NMR and
LC-MS methods (vide infra). The use of HD* enabled us to
examine its degradation process even at the relatively low
concentration of 1 wt %, which is somewhat more indicative
for the real decontamination scenario (100:1 proportion of
decontaminant/CWA).
As in other previous studies,3 the profiles of the HD*
degradation processes were found to be compatible with
a pseudo-first-order reaction (correlation coefficients R2
ranged between 0.98 and 0.99) as shown in Figure 1, as well
as Figures S3-S12 in the Supporting Information.
FIGURE 1. Selected 13C MAS NMR spectra of adsorbed HD*
(6.5 wt %) on wetted KF/Al2O3 (20, EtOH, 160) and its degradation
profile onto this sorbent.
The results of the real time kinetics are summarized in
Table 1. Inspection of the data revealed that both KF/Al2O3
(20, H2O, 160) and KF/Al2O3 (20, EtOH, 160) supports react
with HD* at ambient conditions. In our previous study on
the reactions of VX (5 wt %) with the same solid supports, we
found that the latter (a type of ECUF-KF/Al2O3) sorbent is
more effective.3b The present work shows that this pheno-
mena is also relevant for the reactions of HD* with the
above-mentioned wetted supports. Thus, the degradation
rate of 6.5 wt % HD* onto wetted KF/Al2O3 (20, H2O, 160)
is 2-fold slower than that onto wetted KF/Al2O3 (20, EtOH,
160) (Table 1, runs 2 and 5). We assume that the difference in
final products identity and distribution may be attributed to
a competition between substitution versus elimination pro-
cesses, that is strongly dependent on the reaction conditions
and the sorbent type used (vide infra). In addition to the
above-mentioned chemical difference, we found that these
two active sorbents differ in other features, such as surface
area and basicity value (Table 2). As previously reported, the
impregnation of KF on alumina sorbent mostly caused a
decrease in its surface area (especially with high loading
of KF) together with a significant increase in its basicity.7
As is clearly shown by Table 2, the original surface area of
neutral alumina was decreased to a lesser extent in ECUF-
KF/Al2O3 supports in comparison to regular KF/Al2O3. The
formation of K3AlF6 and KOH in the latter may block many
of the available active sites and chemically destroy the sur-
face area of the alumina. These compounds are barely
formed when the solid support is prepared in EtOH,3b and
therefore, the surface area of the resulting ECUF-KF/Al2O3
is larger. In addition, the latter possess basicity strength
similar to that of regular KF/Al2O3.
Consistent with our previous study on the reactions
between VX and KF/Al2O3,3b the reaction of HD* with
these sorbents was found to be strongly dependent on their
water content. That is, the reaction rates were slowed by
either a lack (0%) or an excess (10%) of water added to both
types of KF/Al2O3 powders prior to the addition of 6.5 wt %
HD (Table 1, runs 1, 3 and 4, 6). This effect is attributed to
the ability of water molecules to block the active sites of the
solid catalyst or to enhance its hydrophilicity, and conse-
quently to slow down the diffusion of HD. Even so, in dry
conditions, there is no substantial difference between the two
sorbents, both acting very slowly (runs 1 and 4). This
observation implies that the relatively rapid degradations
are largely assisted by water molecules. Interestingly, the
(7) (a) 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. (b) Clark, J. H.; Goodman, E. M.; Smith, D. K; Brown, S. J.;
Miller, J. M. J. Chem. Soc., Chem. Commun. 1986, 657. (c) Duke, C. V. A.;
Miller, J. M.; Clark, J. H.; Kybett, A. P. J. Mol. Catal. 1990, 62, 233.
(d) Handa, H.; Baba, T.; Sugisawa, H.; Ono, Y. J. Mol. Catal. A 1998, 134,
171. (e) Bautista, F. M.; Campelo, J. M.; Garcia, A.; Luna, D.; Marinas,
J. M.; Romero, A. A. J. Chem. Soc., Perkin Trans. 2 2002, 227.
J. Org. Chem. Vol. 74, No. 21, 2009 8465

Zafrani et al.
JOCNote
TABLE 1. Degradation Rates and Products of HD* Adsorbed on
Alumina-Based Powders
a
HD H2O
(wt %) (wt %)
t1/2
(h)
products (%)
distributionb
run
KF/Al2O3
1
2
3
4
5
6
7
8
9
(20, H2O, 160)
(20, H2O, 160)
(20, H2O, 160)
(20, EtOH, 160)
(20, EtOH, 160)
(20, EtOH, 160)
(20, EtOH, 160)
(20, EtOH, 160)
(40, H2O, 160)
(40, EtOH, 160)
neutral Al2O3
6.5
6.5
6.5
6.5
6.5
6.5
1
0
5
10
0
5
10
0
5
5
5
0
355
87.0
138
319
44.4
224
22.6
3.5
14, 60, 6, 0, 20
5, 18, 4, 71, 3
2, 8, 8, 80, 2
9, 5, 9, 74,c 3
8, 32, 7, 51, 2
1, 10, 8, 80, 1
7, 4, 7, 70,c 12
28, 14, 7, 48, 3
36, 53, 1, 8, 2
30, 53, 2, 12, 3
1
6.5
6.5
6.5
6.5
59.7
29.0
nr
10
11
12
neutral Al2O3
5
315d
sulfonium salts
aHalf life times of pseudo-first-order reaction. bProducts distribution
for DVS, CEVS (eventually degraded to DVS final product), HOEVS
elimination products (in bold) and TDG-oligomers (n = 1-7) and
thioxane, respectively. cBound TDG. dThis is the second and steady state
t1/2 observed by the reaction profile of this reaction.
FIGURE 2. Overlaid LC-ESI(-)/RSIM-MS scan chromatograms
of TDG-oligomers (n = 2-7) and daughter scan spectrum of the
dimer (n = 2, an example for structure elucidation).
in the NMR spectrum. On the other hand, the reaction of
HD* with wetted (5 wt %) neutral alumina mainly produced
the toxic sulfonium salts CH-TG and H-2TG,1,9,10 resulting
from a nucleophilic attack on HD or CH intermediate by
TDG (see Figure S12, Supporting Information).
TABLE 2. Physical/Chemical Properties of the Alumina Powders
surface area
(m2/g)a
fluoride
group
19F MAS NMR
KF/Al2O3
pHb
(ppm)c
In view of the complexity of the products mixture obtained
from the above-mentioned processes, especially those that
were derived from TDG, we decided to confirm their identity
by LC-MS methods. To obtain simpler LC-MS data for
structure elucidation, unlabeled HD was reacted with wetted
KF/Al2O3 (20, EtOH, 160), and the products mixture was
extracted by chloroform. In this analysis, TDG was detected
by LC-MS, using positive ESI. The mass spectrum exhibited
low intensity peaks corresponding to TDG ions at m/z 123
(M þ H)þ and 145 (M þ Na)þ and a significant fragment ion
at m/z 105 due to loss of H2O. This identification was
confirmed by comparative analysis of a reference material.
TDG-oligomers (n = 2-7) were detected by LC-MS, using
negative ESI (Figure 2). They eluted from the LC column
consecutively as expected from their increasing hydropho-
bicity. The dimer, the most polar oligomer, eluted after 7 min
while the heptamer eluted after 13 min. Their mass spectra
contained only the (M - H)- ions with the expected isotopic
ratios typical for sulfur-containing compounds. Structure
elucidations of these compounds were performed following
MS-MS experiments, where typical sequential losses of -18
(-H2O), -44 (-CH2CH2O), -60 (-CH2CH2S), -62
(-CH2CH2O-H2O), and -104 (-OCH2CH2S-CH2CH2)
were observed.
In light of these results, the significant variety in the
observed products leads to the conclusion that the degrada-
tion of HD onto the surface of alumina-supported fluoride
sorbents may proceed by few plausible mechanisms, as shown
in Figure 3. Clearly, it is influenced by the chemical differences
between the above-mentioned four possible active sorbents, i.
e., wettedvs. dried and regular- vs. ECUF-KF/Al2O3. Itseems
that the nature of the final products is strongly dependent on
the water content of the sorbents, which determines the
competition between the substitution vs. elimination pro-
cesses. The substitution processes may be carried out by an
SN1-type mechanism that occurs via a CS (cyclic sulfonium)
neutral Al2O3
(20, EtOH, 160)
(20, H2O, 160)
(40, EtOH, 160)
(40, H2O, 160)
148.7
100.7
57.7
72.5
10.9
8.4
10.6
10.9
11.4
11.8
F-
K3AlF6
F-
K3AlF6
-132
-158
-132
-158
aMultipoint surface areas measured by BET technique. bMeasured as
a slurry of 0.5 g of powder in 10 mL of double distilled water. cChemical
shifts of fluoride groups of the dry powders (see also ref 3b).
reaction of HD* with dry KF/Al2O3 (20, EtOH, 160) re-
sulted in bound TDG as a major product, as can be shown by
its exceptionally broad signals in the NMR spectrum, which
are attributed to its immobility feature (Table 1, runs 4 and 7
and Figures S7 and S9 in the Supporting Information).
This product implies that the dominant active group in dry
KF/Al2O3 (20, EtOH, 160) is the [Al-O-], which directly
reacts with the chloromethylene functions of HD (vide infra).
On the other hand, the reaction of HD* (6.5 wt %) with dry
KF/Al2O3 (20, H2O, 160) gave the appropriate vinylic sulfides
as major products together with thioxane (run 1). The domi-
nant active group in this regular KF/Al2O3 is [OH-],7 which
reacts as a strong base in a dehydrochlorination process.
The reaction studies indicate that 40% KF loading
(compared to 20%) results in slightly faster rates with both
regular- and ECUF-KF/Al2O3, but with a higher percentage
of vinylic products (runs 9 and 10). In addition, the reactions
of 1 wt % of HD*, with both dried and wetted KF/Al2O3 (20,
EtOH, 160), are considerably faster than those of 6.5 wt %
(runs 4, 5 and 7, 8). These results are in agreement with the
fact that the reaction rate is diffusion dependent, and may be
explained by a faster dispersion of the reactant toward the
active sites when the bulk of the drop is smaller.3a
When compared to neutral alumina,8 these results are
impressive in terms of both products and kinetics (runs 11
and 12). Thus, adsorption of HD* onto the surface of dry
neutral alumina did not yield any products, but did strongly
physisorbe to the support, according to the signal broadening
(8) Taking into account that commercial neutral alumina powders may
differ from batch to batch, we used a sample of alumina that was dried at the
same condition (160 °C, overnight) for all experiments.
(9) Yang, Y.-C.; Szafraniec, L. L.; Beaudry, W. T.; Ward, J. R. J. Org.
Chem. 1987, 52, 1637.
(10) Wagner, G. W.; Bartram, P. W. Langmuir 1999, 15, 8113.
8466 J. Org. Chem. Vol. 74, No. 21, 2009

Zafrani et al.
JOCNote
FIGURE 4. A centrifugation effect on degradation rates of HD*
(1 wt %) onto the surface of wetted (5%) KF/Al2O3 (20, EtOH, 160).
in which the reaction rates were tested onto a sample of wetted
(5%) KF/Al2O3 (20, EtOH, 160) under MAS NMR condi-
tions (5 kHz) and onto another duplicate sample that was left
aside (samples 1 and 2, respectively). Once the concentration of
HD of the first sample was totally decayed under spinning
conditions, the second duplicate sample was subjected to MAS
NMR measurement. The reaction rate continuously measured
by MAS NMR was found to be higher than its still counter-
part. Thus, the degradation rates of HD by KF/Al2O3 mea-
sured by the MAS NMR technique are not the intrinsic rates,
and the contribution of the spinning should be taken into
account when comparing sets of data. It seems that this effect is
correlated to the relationship between the hydrophilicity/
hydrophobicity of both reactant and sorbent. Namely, in
reactions occurring between hydrophobic reactants and inor-
ganic solids, in which the repulsive interactions are dominant,
the spinning effect would be more significant.
In conclusion, this work shed light on the products and the
mechanisms of the degradation of HD adsorbed onto
KF/Al2O3, using 13C-labeled HD and both SS MAS NMR
and LC-MS techniques. The significant variety in the ob-
served nontoxic products (schematically divided into elimina-
tion and substitution products) stems from the chemical
differences between the possibilities of wetted versus dried
and regular- versus ECUF-KF/Al2O3. The fact that the
degradationrate ofHDontothe wettedKF/Al2O3 (20, EtOH,
160) sorbent was affected by the high spinning rate of the
sample (a centrifugation effect), contrary to VX, should be
considered in every study related to this topic. This pheno-
menon is now specifically under further extensive study.
FIGURE 3. Proposed mechanisms of the degradation of HD onto
the surface of regular- and ECUF-KF/Al2O3.
intermediate,11 which is solvated and stabilized by water
(strong ionization power). The elimination process probably
proceeds by both E1- and E2-type mechanisms via CS or
neutral HD starting material. The presence of water encourages
the formation of CS, and therefore, the major products in the
wetted sorbents, both regular and ECUF-KF/Al2O3, are TDG
and TDG-oligomers, which originate from nucleophilic sub-
stitutions. On the other hand, in the absence of water, the
dominant equilibrium form is that of the neutral HD, but not a
solvated CS, and therefore, the major products in its reaction
with regular KF/Al2O3 (enriched with the relatively strong
base, KOH) are vinylic sulfides, via an E2 pathway. On dry
ECUF-KF/Al2O3 the major product is bound-TDG obtained
through a substitution of chloride by Al-O- (a relatively weak
base). This substitution, as expected, is much slower than its
SN1 counterpart, in which CS acts as an active intermediate.
The fact that adsorption of HD onto the surface of both wetted
KF/Al2O3 (40, H2O, 160) and KF/Al2O3 (40, EtOH, 160) leads
to vinylic sulfides as major products may be attributed to the
high loading (40%) of KOH or KF, respectively, which are
highly solvated by water molecules, therefore consuming the
water available for CS solvation.
Recently, we reported that solid supported reaction rates,
measured by SS MAS NMR, may be affected by a centrifuga-
tion effect (or possibly by a heat friction) caused by the high
spinning rate of the sample (ca. 5-8 kHz).6 This effect is
attributed to the fact that in such processes the diffusion of
reactants toward the active sites of the solid support plays a
significant role in the reaction kinetics. In the context of
decontamination it is important to know whether the real time
reaction rates onto KF/Al2O3 measurements are affected by
MAS. It was found that in contrast to the case with VX, which
reacts with similar rates with and without spinning,3b the
degradation kinetics of HD were indeed affected (ca. 3-fold
faster) by the high spinning rate during the measurement
(Figure 4). To examine this effect, we performed an experiment
Experimental section
Sample Preparation. Caution: These experiments should only
be performed by trained personnel, using applicable safety
procedures. Samples of the appropriate powder (ca. 100 mg)
were added to the 0.4 cm ZrO2 rotor. A 0.8-5 μL sample of HD*
was applied via a 5 μL syringe to the center of the sample. The
rotor was sealed with a fitted Kel-F cap. 13C MAS NMR spectra
were measured periodically to determine the remaining starting
material and identify degradation products.
(11) (a) Bartlett, P. D.; Swain, G. J. Am. Chem. Soc. 1949, 71, 1406.
(b) McManus, S. P.; Neamati-Mazraeh, N.; Hovanes, B. A.; Paley, M. S.;
Harris, J. M. J. Am. Chem. Soc. 1985, 107, 3393. (c) Yang, Y.-C.; Ward, J. R.;
Luteran, T. J. Org. Chem. 1986, 51, 2756. (d) Yang, Y.-C.; Szafraniec, L. L.;
Beaudry, W. T.; Ward, J. R. J. Org. Chem. 1988, 53, 3293. (e) Converso, A.;
Saaidi, P.-L.; Sharpless, K. B.; Finn, M. G. J. Org. Chem. 2004, 69, 7336.
Supporting Information Available: General experimental
details and Figures S3-S12 giving 13C MAS NMR spectra and
reaction profiles of adsorbed HD* onto selected KF/Al2O3
powders. This material is available free of charge via the Internet
at http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 21, 2009 8467