Nerve agent degradation with polyoxoniobates

Article abstract of DOI:10.1002/ejic.201400016

Polyoxoniobates are exceptional amongst polyoxometalates in that they can potentially perform base catalysis in water, a process in which a proton is bonded to an oxo ligand, and a hydroxyl is released. Catalytic decomposition of chemical warfare agents such as organofluorophosphates that were used recently in the infamous civilian attacks in Syria is one opportunity to employ this process. Upon evaluation of the polyoxoniobate Lindqvist ion, [Nb ₆O₁₉]^8-, fast neutralization kinetics was discovered for the breakdown of the nerve agent simulant diisopropyl fluorophosphate (DFP). The polyoxoniobates were also tested against the nerve agents Sarin (GB) and Soman (GD). It was determined that different Lindqvist countercations (Li, K, or Cs) affect the rate of decomposition of the organophosphate compounds in both aqueous media (homogeneous reaction), and in the solid state (heterogeneous reaction). Small-angle X-ray scattering analysis of solutions of the Li, K, and Cs salts of [Nb₆O₁₉] ^8- for concentrations at which the experiments were performed revealed distinct differences that could be linked to their relative reaction rates. This study represents the first demonstration of exploiting the unique alkaline reactivity of polyoxoniobates for nerve agent decontamination. Polyoxometalates, like small pieces of metal oxide, can be dissolved or fixed on a surface to perform homogeneous or heterogeneous catalysis, respectively. Here we exploit the alkaline nature of polyoxoniobates to neutralize nerve agents in both solution and the solid state. Solution studies correlate reaction efficacy to the association of the dissolved polyoxoniobate with its counterions. Copyright

Full text of DOI:10.1002/ejic.201400016

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
FULL PAPER
DOI:10.1002/ejic.201400016
Nerve Agent Degradation with Polyoxoniobates
Mark K. Kinnan,[a] William R. Creasy,[b] Lauren B. Fullmer,[c]
Heidi L. Schreuder-Gibson,[d] and May Nyman*[c]
Keywords: Polyoxometalates / Kinetics / Ion pairs / Small-angle X-ray scattering / Nerve agents
Polyoxoniobates are exceptional amongst polyoxometalates
in that they can potentially perform base catalysis in water,
a process in which a proton is bonded to an oxo ligand, and
a hydroxyl is released. Catalytic decomposition of chemical
warfare agents such as organofluorophosphates that were
used recently in the infamous civilian attacks in Syria is one
opportunity to employ this process. Upon evaluation of the
(GB) and Soman (GD). It was determined that different Lind-
qvist countercations (Li, K, or Cs) affect the rate of decompo-
sition of the organophosphate compounds in both aqueous
media (homogeneous reaction), and in the solid state (hetero-
geneous reaction). Small-angle X-ray scattering analysis of
8–
solutions of the Li, K, and Cs salts of [Nb6O19
]
for concen-
trations at which the experiments were performed revealed
distinct differences that could be linked to their relative reac-
tion rates. This study represents the first demonstration of
polyoxoniobate Lindqvist ion, [Nb6O19
kinetics was discovered for the breakdown of the nerve
]
8–, fast neutralization
agent simulant diisopropyl fluorophosphate (DFP). The poly- exploiting the unique alkaline reactivity of polyoxoniobates
oxoniobates were also tested against the nerve agents Sarin
for nerve agent decontamination.
veloping materials for decontamination applications, the
nerve agent simulant diisopropyl fluorophosphate (DFP) is
typically employed in view of its likeness to GB and GD
but lower toxicity. For polymeric materials, reactive moie-
ties such as guanidine,[6] hydroxamic acid,[7] o-iodosylcarb-
oxylate,[8] and oxime[9] groups have been demonstrated to
be effective for the breakdown of nerve agents and/or their
simulants. Inorganic materials such as metal chelates,[10]
TiO2,[11] Zr(OH)4,[12] and polyoxometalates (POMs)[13] have
also been investigated. The advantage of POMs is the con-
trol they provide for surface functionalization. They are
small and discrete, and they can be dissolved in aqueous or
nonaqueous solvents (depending on the counterion). This
provides an easy and versatile mechanism for attaching
them to a fabric or filtration media by means of electrostat-
ics, or even covalent bond formation if they are appropri-
ately functionalized. Furthermore, POMs are entirely inor-
ganic and thus robust and not subject to breakdown by
exposure to environmental factors such as ultraviolet light.
In this paper, we strategically use polyoxoniobates
(PONbs) as reagents for the breakdown of DFP (di-
isopropyl fluorophosphate), GD (3,3-dimethylbutan-2-yl
methylphosphonofluoridate), and GB (propan-2-yl methyl-
phosphonofluoridate), shown in Figure 1, given their alk-
aline behavior. Although a variety of POMs such as
transition-metal-substituted polyoxometalates,[14] iron-
Introduction
Nerve agents are a type of organofluorophosphate (OP)
compound that are used as chemical warfare agents
(CWAs). They release HF upon contact with moist skin or
lungs and inhibit neural function; these were used in the
recent infamous attacks on civilians in Syria.[1] The deacti-
vation, and thus decontamination, of nerve agents such as
Sarin (GB) and Soman (GD) is accomplished by cleavage
of the phosphorus–fluorine bond in the molecule and its
replacement with a P–OH bond.[2] Even though nerve
agents autohydrolyze in water and high-pH environments
with time,[3] decontaminating materials that can be attached
to a surface while maintaining their reactive nature are of
interest.[4] This is particularly important for the develop-
ment of fabrics or filtration media that can protect an indi-
vidual from CWAs.[4]
Over the years, a variety of different polymeric and inor-
ganic materials have been evaluated for their ability to
break down nerve agents and their simulants.[5] When de-
[a] Sandia National Laboratories,
1515 Eubank SE, Albuquerque, New Mexico 87123, USA
[b] Leidos Corp, P. O. Box 68, Edgewood Chemical Biological
Center,
Aberdeen Proving Ground, Maryland 21010, USA
[c] Oregon State University, Department of Chemistry,
153 Gilbert Hall, Corvallis, Oregon 97331-4003, USA
E-mail: may.nyman@oregonstate.edu
http://chem.science.oregonstate.edu/nyman
[d] U. S. Army Natick Soldier Research, Development &
Engineering Center,
substituted
H5PV2Mo10O40
Keggin
heteropolytungstate,[15]
have been tested for the breakdown of
and
[16]
CWAs, the focus has been on blister agents (e.g. mustard
gas), which use different decomposition mechanisms that
are not suitable for nerve agents. Blister agents are decon-
15 Kansas Street, Natick, Massachusetts 01760-5020, USA
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201400016.
Eur. J. Inorg. Chem. 0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
taminated by means of oxidation reactions, whereas nerve by a bridging oxygen ligand in a Lindqvist ion is illustrated
agents are neutralized by base hydrolysis.
in Figure 2. The breakdown products of OP compounds
can be conveniently measured in solution by using a
fluoride ion probe[23] or by 31P NMR spectroscopy.[2] Solu-
tion-based tests using OP compounds can be performed in
minutes and are ideal for rapidly screening materials for
potential decontamination activity. NMR spectroscopy, on
the other hand, takes significantly longer to collect data but
has the advantage of performing the reaction in a closed
system by means of a capped NMR rotor. NMR spec-
troscopy can also simultaneously measure the concentra-
tion of OP compounds and their breakdown products.
Figure 1. Top left: Ball-and-stick model of the Lindqvist ion
[Nb6O19]8–. Blue spheres are niobium; red spheres are oxygen. In
this representation, two protonated µ2-bridging oxo ligands are
shown (gray spheres). Top right: The nerve agent simulant di-
isopropyl fluorophosphate (DFP). Bottom left: The nerve agent
Sarin (GB), (RS)-propan-2-yl methylphosphonofluoridate. Bottom
right: The nerve agent Soman (GD), 3,3-dimethylbutan-2-yl meth-
ylphosphonofluoridate.
PONbs are the most basic of the known POMs as a re-
sult of their high charge density: they readily bind up to
three protons in aqueous solution.[17] As recent work has
shown, PONbs also exhibit size-, geometry-, composition-,
and charge-versatility, which can be exploited to tailor reac-
tivity in base catalysis reactions.[18] The Lindqvist ion or
hexaniobate, [Nb6O19]8–, is the most studied PONb, and its
compact and highly symmetric geometry, consisting of six
mutually edge-sharing octahedra forming a super-octahe-
dron, is illustrated in Figure 1. The Lindqvist ion is highly
stable and readily synthesized with all alkali countercations
and also tetramethylammonium, and thus is an ideal model
system for this study.[19] Moreover, this cluster has the high-
est charge density of all POMs,[18] which directly correlates
with its basicity. Both solid-state and solution[17a] studies
indicate that the μ2-bridging O2– ligand is the favored pro-
tonation site over the doubly bonded, terminal oxo ligand,
and this is depicted in Figure 1.
The polyoxometalates of W, Mo, and V have been uti-
lized in catalytic reactions quite extensively, as oxidation
and acid catalysts in particular.[20] On the other hand, cata-
lytic reactivity studies of polyoxoniobates are rare. Hetero-
geneous photocatalysis has recently been demonstrated for
PONbs, but it requires the aid of co-catalysts.[21] This cur-
rent study represents the first demonstrated reaction of
polyoxoniobates that can be carried out effectively both het-
erogeneously and homogeneously to exploit the unique al-
kaline reactivity of PONbs.
Figure 2. Illustration of the decomposition of diisopropyl fluoro-
phosphate by the oxygen bridging ligand of the polyoxoniobate. R
designates an isopropyl group.
The kinetic data for the breakdown of OP compounds by
using PONbs is presented in Table 2 and also in Figure 3, a
representative comparison of the activity of the [Nb6O19]8–
Li, K, and Cs salts. Upon solution-based testing with use
of the fluoride ion probe, fast neutralization kinetics was
discovered for the breakdown of DFP, but the rate of DFP
decomposition was significantly affected by the countercat-
ion associated with the Lindqvist ion.
Previously reported small-angle X-ray scattering (SAXS)
studies of the Lindqvist salts in solutions of 1–3 m alkali
hydroxide showed that contact ion-pairing of [Nb6O19]8–
with Cs+ dominates in aqueous media, whereas solvent-sep-
arated or solvent-shared ion-pairing of [Nb6O19]8– is the
predominant state when K+ is the countercation.[24] The ion
association between Li+ and the Lindqvist ion has not yet
been probed, but on the basis of the general trend of the
alkalis it is assumed the Li+ carries a large hydration sphere
and is not closely associated with [Nb6O19]8–. These prior
studies are good controls for understanding cluster–coun-
terion association, in that the complicating effect of proton-
Results and Discussion
Nerve agents like GD and GB, as well as the simulant ation of the clusters is eliminated. Here we perform SAXS
DFP, are decomposed by cleavage of the phosphorus–fluor- analyses representative of the DFP neutralization studies
ine bond to produce the fluoride ion and a less toxic or- presented in Figure 3. The top plot in Figure 4 shows the
ganophosphate compound.[22] The decomposition of DFP scattering curves for 2 mm solutions of Li+, K+, and Cs+
Eur. J. Inorg. Chem. 0000, 0–0
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
Table 1. SAXS analysis of solutions of Li, K, and Cs salts of
[Nb6O19], with and without DFP.
Counter- Rg from Guinier
ion
analysis[a]
R from
Rg
Rg from PDDF[c]
plot
[b]
2 mm [Nb6O19],without DFP
Li+
3.4 (4)
4.1 (2)
8.4 (2)
4.39
5.35
10.9
3.5 (1)
4.4 (2)
8.5 (2)
K+
Cs+
5 mm [Nb6O19], with DFP
Li+
3.0 (3)
3.3 (2)
3.7 (3)
3.87
4.26
4.77
3.08 (2)
3.31 (3)
3.76 (2)
K+
Cs+
Figure 3. Comparing DFP decomposition rates, and the effect of
the Li, K, or Cs counterion. Reactions were performed in aqueous
solutions with mol ratios of 3.6, 4.8, and 4.4 of DFP to the Li, K,
and Cs salts of PONb, respectively. A fluoride electrode was used
to monitor the concentration of fluoride ions generated from the
decomposition of DFP in solution.
[a] In the range q = 0.18–0.25 Å–1 from the ln(I) vs. q2 plot. [b]
Assuming approximate spherical radius; R ≈ 1.29ϫRg. [c] PDDF:
pair distance distribution function.
tering species in solution), the associated radius (R, ca.
1.29ϫRg assuming a spherical particle), and PDDF profile
all agree with a “nude” [Nb6O19]8– anion that is not associ-
ated with any counterions (the crystallographically deter-
mined radius of this species is 4.2 Å). This is likely the
reason why the Li8[Nb6O19] solution is most effective at cat-
alyzing the neutralization of DFP in solution: its unassoci-
ated state allows direct contact with the DFP molecule. The
2 mm Cs8[Nb6O19] solution clearly contains larger scat-
tering species, as determined by the Rg, which was obtained
by two methods (see Table 1). The PDDF profile of
Cs8[Nb6O19] presents a classic curve of a dimerized primary
scatterer in solution,[25] as we have observed in prior SAXS
studies of polyoxoniobates.[26] The primary scatterer is con-
siderably larger than the “nude” Lindqvist ion, with a dia-
meter of around 12 Å, and the long dimension of the dimer
is approximately double that length, 22 Å. The 12 Å dia-
meter is almost 4 Å bigger than the “nude” Lindqvist ion
and could indicate a shell of Cs+ cations directly bonded to
the cluster. Two clusters of the dimer form could be associ-
ated by mutual H-bonding of the protonated faces of the
clusters, as observed (and modeled from structural data as
done previously;[26] see Supporting Information). Without
a solid-state model, we cannot describe the exact solution
state of the Lindqvist ions, but a qualitative approximation
is shown in the Supporting Information. The Cs8[Nb6O19]
solutions are not monodisperse, as indicated by the shallow
slope of the high-q region of the curve, which could not be
reasonably fit (above q = 0.5 Å–1). Nonetheless, the curve-
fitting routines indicate extensive clustering and ion-pair-
ing. The K8[Nb6O19] solutions, like the Li8[Nb6O19] solu-
tions, consist primarily of the unassociated Lindqvist ions
but also contain a population of dimers (estimated to be up
to 30%), likely formed by mutual H-bonding of two dipro-
tonated faces of the cluster without any associated K+. This
is a motif we often observe in the solid state[19c] and simu-
lated[27] SAXS data from solid-state structures. Simulated
SAXS data are shown in the Supporting Information for a
monomeric Lindqvist ion, two Lindqvist ions dimerized by
salts of [Nb6O19]8–. Although the nerve agent neutralization
reactions were performed with 1 mm solutions of the Lind-
qvist salts, these solutions did not give enough scattering
intensity to interpret accurately, so higher concentrations
were necessary. There are distinct differences, and the form
factors obtained from curve-fitting routines are summa-
rized in Table 1. Additionally, the pair distance distribution
function (PDDF) plots derived from the scattering curves
are shown in the bottom plot in Figure 4.
Figure 4. Top: X-ray scattering curves for 2 mm Li8[Nb6O19],
K8[Nb6O19], and Cs8[Nb6O19] dissolved in water. Bottom: Pair dis-
tance distribution function (PDDF) profiles from curve-fitting of
the SAXS plots.
For Li8[Nb6O19], the radius of gyration (Rg, a shape-in- H-bonding, and linear combinations of pure monomer and
dependent parameter that describes the size of the scat- pure dimer forms.
Eur. J. Inorg. Chem. 0000, 0–0
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
We also analyzed by SAXS the solutions of the [Nb6O19] acting with OP molecules and ultimately decreases the effi-
salts containing the same concentration of DFP that was cacy of homogeneous reaction in water. The SAXS data
utilized in the degradation experiments. The scattering also lends evidence for the mechanism of OP decomposi-
curves for these are shown in the Supporting Information, tion illustrated in Figure 2, where the transition state in-
and the PDDF profiles are in Figure 5. Initially, we ob- volves direct association of the protonated Lindqvist ion
tained even weaker scattering curves from 2 mm Lindqvist with the OP molecule rather than association of the OP
ion solutions with DFP; and thus we increased the [Nb6O19] molecule with a hydroxide generated by protonation of the
concentration to 5 mm. Upon analysis of these data, the Lindqvist cluster.
reason for the weaker scattering was self-evident: the scat-
Further testing of the PONbs against DFP in a closed
tering species were smaller in the presence of DFP, the Cs environment with significantly less water was performed in
analogue in particular (see Table 1). The scattering intensity an NMR rotor monitored by 31P solid-state NMR spec-
decreases significantly with the size of the scattering species, troscopy. This provided information on reactive behavior
since intensity is proportional to the sixth power of the ra- on a surface, which is closer to the state that would be em-
dius. This alone was a significant observation, indicating ployed for filtration media or protective clothing. The re-
that: (1) the Lindqvist ions do not decompose upon expo- sults revealed slower kinetics for Li8[Nb6O19] and
sure to DFP or the HF byproduct and (2) the DFP mol- K8[Nb6O19] compared to that observed in solution, but
ecules partially disaggregate the alkali–[Nb6O19] assemblies. faster kinetics for Cs8[Nb6O19]. Additionally, the observed
If we add the DFP to the Lindqvist ion solution in great decontamination rates were similar for the Lindqvist salts
excess, we do indeed observe clouding and formation of with no noticeable trend with respect to counterion. In the
very large particles by SAXS analysis. The PDDF profiles 31P NMR spectroscopy experiment, the amount of water in
of the solutions containing DFP and Li or K salts of the system is insufficient to provide complete dissolution of
[Nb6O19] suggest mostly unassociated Lindqvist ions (see the Lindqvist salts, and this results in a more heterogeneous
Figure 5 and Table 1). PDDF analysis of the DFP and Cs– system. These reactions contained a Lindqvist salt and a
[Nb6O19] indicates that the dimerized state still dominates small quantity of water creating a saturated solution of
the solution speciation. However, the linear extent and Lindqvist salt and undissolved PONb. Figure 6 shows rep-
maxima of the curves (compared to those of the solutions resentative 31P NMR spectra of the breakdown of DFP to
without DFP) suggest that the Cs counterions are less asso- diisopropyl phosphate (DIPP) for a heterogeneous solution.
ciated. These results collectively suggest that the DFP mol- For the specific case of Cs–[Nb6O19], kobs and t1/2 are slower
ecules do in fact displace associated alkali metals, which for solution testing than those observed in NMR spectro-
would be important for the reaction with [Nb6O19] to occur. scopic testing. With solution testing, there is the additional
Not entirely, however, as the Rg values determined by two element of OP molecule diffusion, followed by interaction
methods (see Table 1) show the trend Li Ͻ K Ͻ Cs. As dis- with the Lindqvist ion. This effect, in conjunction with the
cussed earlier, the dimerization that occurs extensively for close association of Cs+ to the Lindqvist ion, hinders de-
aqueous Cs–[Nb6O19], with and without DFP, is likely contamination at high volumes of water. For small volumes
through mutual H-bonding of protonated faces. Thus we of water, diffusion of the OP molecule is a less important
may hypothesize that these protonated faces are not as ac- factor.
cessible to DFP in the Cs–[Nb6O19] solutions, and the neu-
tralization reaction is considerably hindered.
Figure 6. Representative 31P NMR spectra of the breakdown of
diisopropyl fluorophosphate (DFP) to diisopropyl phosphate
(DIPP) for a heterogeneous solution. The DFP peak is split by
Figure 5. PDDF profiles for 5 mm Li8[Nb6O19], K8[Nb6O19], and
Cs8[Nb6O19] dissolved in water with DFP.
1
19F–31P J-coupling.
The association of clusters in solution by both ion pair-
After testing DFP with the three Lindqvist salts, ad-
ing with the counterions and H-bonding of protonated clus- ditional breakdown testing was performed with GD and
ter faces increases with increasing counterion size GB at the Edgewood Chemical Biological Center (ECBC,
(Li Ͻ K Ͻ Cs). This leads to decreased availability for inter- see Table 2 and Figures S5 and S6). GD was only tested
Eur. J. Inorg. Chem. 0000, 0–0
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
Table 2. Kinetic data for the breakdown of organophosphate (OP) compounds with polyoxoniobates (PONbs).
Material
OP
Method of measurement Mol ratio
(OP/PONb)
React. volume
[mL]
kobs
t1/2
[h]
[s–1]
Li8Nb6O19
K8Nb6O19
Cs8Nb6O19
DFP[a]
DFP
DFP
Ion Probe
Ion Probe
Ion Probe
3.6Ϯ0.3
4.8Ϯ0.5
4.4Ϯ0.2
10.
10.
10.
6.0ϫ10–5 Ϯ0.8ϫ10–5
3.4ϫ10–5 Ϯ0.3ϫ10–5
8.3ϫ10–6 Ϯ3ϫ10–6
3.2Ϯ0.4
5.7Ϯ0.4
27Ϯ13
Li8Nb6O19
K8Nb6O19
Cs8Nb6O19
DFP
DFP
DFP
NMR
NMR
NMR
2.8
2.1
2.5
0.11
0.11
0.12
1.86ϫ10–5
1.92ϫ10–5
1.80ϫ10–5
10.38
10.04
10.69
Li8Nb6O19
Li8Nb6O19
Li8Nb6O19
GD[b]
GD
GD
NMR
NMR
NMR
0.50
5.55
22.18
0.22
0.23
0.23
7.04ϫ10–4
3.93ϫ10–4
1.25ϫ10–4
0.27
0.50
Ͼ 1.54[d]
Li8Nb6O19
K8Nb6O19
Cs8Nb6O19
GB[c]
GB
GB
NMR
NMR
NMR
0.97
1.62
1.56
0.66
0.11
0.12
6.0ϫ10–3
5.7ϫ10–3
5.5ϫ10–3
0.03
0.03
0.04
[a] DFP: Diisopropyl fluorophosphate. [b] GD: 3,3-Dimethylbutan-2-yl methylphosphonofluoridate. [c] GB: Propan-2-yl methylphospho-
nofluoridate. [d] The rate constant is shown for the first 0.5 h, but the rate slows down after 0.5 h. See text for discussion.
with Li8[Nb6O19]. In this case, the mol ratio of GD to
Comparing the kinetic data reported in this manuscript
PONb was different for each test to determine any effects to kinetic data from other neutralization materials in the
on the kinetics. It was observed that, as the mol ratio of literature is difficult. There are many studies of different
GD/PONb increased, kobs decreased and the half-life in- materials for neutralization of a variety of CWAs and simu-
creased. The results from those tests revealed a range in GD lants. Challenges associated with comparing data to pre-
half-life from 0.27 to more than 1.54 h, almost an order of viously reported studies include: (1) differences in analytical
magnitude, for a mol ratio from 0.5 to 22 for GD/LiPONb. techniques, (2) differences in CWAs and/or simulants, (3)
For the mol ratio of 22, the pseudo-first-order kinetics plot different molar ratios of CWA and/or simulant to the neu-
was not linear and showed a marked decrease in rate after tralization catalyst. In particular, many previous studies
0.5 h. The kinetic plots are shown in the Supporting Infor- have used much lower ratios of CWA to decontamination
mation. The reason for the rate change is not clear and agent. In some prior studies, a liquid phase reaction was
requires more research. One possibility is that, at lower con- mixed or agitated to allow a larger amount of decontami-
centrations of GD in solution, the diffusion of GD mol- nant to come into contact with the OP. For example, one
ecules to an active site on a Lindqvist ion is likely the rate- report used polyacrylamidoxime (PANOx)[29] for the neu-
limiting step. By increasing the concentration of GD, the tralization of GD. The kinetic observations for kobs and
time for diffusion of a GD molecule to a Lindqvist ion de- t1/2 are 3.6ϫ10–4 s–1 and 0.54 h, respectively. Another exam-
creases. At higher mol ratios, there is a higher rate of inter- ple is the breakdown of GD with Zr(OH)4,[12] for which
action between the GD molecules and the Lindqvist ion. t1/2 is reported to be 0.15 h. In both of these examples, the
Another possible explanation is that there is a pH change approximate calculated mol ratio of GD to neutralization
resulting from the degradation of GD, since the degrada- catalyst ranges from approximately 0.01 to 0.02. In con-
tion produces acid products, which can result in more pro- trast, the results in Table 2 for GD neutralization with
tonation of the Lindqvist ions and therefore higher reactiv- Li8[Nb6O19] were obtained with mol ratios of up to 22.
ity. However, excessive protonation eventually results in de-
composition of the Lindqvist ions. These results illustrate
the importance of evaluating decontamination materials at
several different mol ratios.
Conclusions and Future Directions
Tests with DFP and GB were performed by using similar
The alkaline nature of polyoxoniobates renders them
quantities of PONb and agent. The GB reaction kinetics unique amongst polyoxometalates, and this characteristic
were extremely fast with half-lives on the order of minutes. has yet to be exploited in practical applications. Reported
Since the chemical structures of GB and GD are similar, in this paper is the first demonstration of the use of poly-
one would expect the kinetic breakdown rates to be similar. oxoniobate Lindqvist ions, [Nb6O19]8–, for the breakdown
However, GB has a higher water solubility[28] than GD, of the CWAs GD and GB, and the CWA simulant DFP. As
which enables faster diffusion of the GD molecules to the a result of their basic behavior, these PONbs are most effec-
Lindqvist ions. As observed with DFP, the Li+ Lindqvist tive for this base-catalyzed reaction. On the other hand,
salt exhibited the fastest kinetics, an order of magnitude prior studies have shown that blister agents can be neutral-
faster than the Lindqvist salts of K+ and Cs+, again indicat- ized by redox-active POMs.[14] In combination, these two
ing the importance of counterion association. After reac- types of chemically unique POMs could potentially provide
tion, 19F NMR spectroscopy was used to determine the complete package for protective clothing or filtration
whether any new fluorine-containing compounds were against chemical warfare agents. Even more intriguing is
formed, but no new 19F peaks were observed.
the very recent developments of incorporating redox-active
Eur. J. Inorg. Chem. 0000, 0–0
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
transition metals into the alkaline POMs,[30] which may
provide opportunity for a POM cluster that could function
in both redox-active reactions and base catalysis.
Our next steps in this work include functionalizing sur-
faces with the alkaline PONbs and testing their surface re-
activity toward CWA decontamination. These studies will
also involve fundamental investigation of mechanisms of
POM attachment to the surface. New CWA decontami-
nation systems are becoming specialized to address specific
ment & Engineering Center (NSRDEC). An Orion Dual Star pH/
ISE Benchtop Meter equipped with an Orion 9409BN fluoride
electrode was used to monitor the concentration of fluoride ions
generated from the decomposition of diisopropyl fluorophosphate
in solution.[23] An example of an experiment run is as follows: An
aliquot of PONb equivalent to 8.0ϫ10–6 mol was weighed out in
a glass Petri dish. Water (9.0 mL) was added to the Petri dish fol-
lowed by the addition of a stirring bar to mix the solution. The
first 90 s of the run was to establish equilibrium between PONb
dissolved in water and the fluoride ion probe. After 90 s, an aque-
applications. Although a fast decontamination reaction rate ous DFP solution (1.0 mL) was added to the Petri dish. Data was
logged on a computer connected to the Orion Dual Star Meter by
a USB cable. The fluoride ion concentration was measured in parts
per million every 5 s. Average data for ion probe testing in Table 1
was calculated from triplicate runs.
is important, materials with properties that fit the applica-
tion are more critical. We will exploit the high anionic
charge of the PONb to electrostatically bind it to fabric
treated with cationic ammonium terminal groups. Recent
studies by Weinstock indicate that the counterions may play
a significant role in the POM attachment mechanism.[31]
Therefore, POM-surface functionalization studies will also
provide insight into the role of counterions in base catalysis
reactivity, as was distinctly observed in the solution studies
reported here.
Solid-State NMR Spectroscopy Studies of DFP (NSRDEC): The
use of solid-state NMR spectroscopy to monitor OP decontami-
nation has been described in detail previously.[2] A weighed quan-
tity of PONb equivalent to 1.0ϫ10–5 mol was added to a 4 mm
Bruker MAS zirconia rotor (Art. Nr.: B5679) followed by the ad-
dition of DFP (4 μL). An aliquot of water (100 μL) was then added
to solvate the PONb, and the rotor was capped by using Kel-F
Drive Caps (4 mm) (Art. Nr.: B5677). The degradation of DFP was
monitored by using a Bruker Avance 400 spectrometer, operating
at 161.92 MHz (31P). The high-resolution magic-angle spinning
HRMAS spectra were recorded with a 4 mm Bruker 31P HRMAS
standard bore gradient probe (BL4HPCD) at ambient temperature
(18 °C) by using a phosphorus observed single-pulse experiment
with the magic-angle spinning rate of 5000 Hz. The following pa-
rameters were applied: center frequency of the acquisition (O1P),
–16 ppm; sweep width, 62.7 ppm; number of scans, 100; excite pulse
length (pw90), 8 μs; relaxation delay, 2 s; total experiment run time,
6.2 min. For the kinetic determinations, it was necessary to trade
off precision with acquisition speed for fast data collection. The
HRMAS spectra were externally calibrated (zeroed) to the one
peak of triphenyl phosphate (–17.8 ppm, depending on solvent, rel-
ative to the phosphoric acid standard).
Experimental Section
Materials: Lithium hydroxide, potassium hydroxide, cesium
hydroxide, and diisopropyl fluorophosphate were purchased from
Sigma–Aldrich. Hazard: Diisopropyl fluorophosphate is an ex-
tremely toxic chemical and should be handled appropriately! Hydrous
niobium oxide was acquired from Reference Metals Company Inc.
All chemicals were used as received. Water with a resistivity greater
than 18 MΩcm was used for all solutions.
PONb Synthesis: The synthesis of lithium, potassium, and cesium
Lindqvist salts from hydrous niobium oxide and the respective
alkali hydroxides has been described previously.[19a,19c] Following
the synthesis, the Lindqvist salts were recrystallized by dissolving
in a minimal amount of water and precipitating by rapid addition
of methanol. This process was repeated three times to remove ex-
cess free base to ensure that the base hydrolysis is effected by the
Lindqvist ion rather than a hydroxide contaminant. Characteriza-
tion to determine protonation state, identification of phase, and
number of water molecules per formula unit was carried out by
powder X-ray diffraction and thermogravimetry.
NMR Spectroscopy Studies of GD and GB (ECBC)
These studies were carried out at the Edgewood Chemical Bio-
logical Center (ECBC). Initial experiments were performed by
using a Bruker Avance 500 spectrometer to reproduce the condi-
tions used at Natick. The instrument operated at 202.46 MHz (31P).
The HRMAS spectra were recorded with a 5-mm Bruker 31P gradi-
ent probe at ambient temperature (23 °C) by using a 1-D proton-
decoupled decay experiment with a magic-angle spinning rate of
5000 Hz. The following parameters were applied: center frequency
of the acquisition (O1P), 20 ppm (4049 Hz); sweep width,
200.8 ppm; number of scans, 128; excite pulse length (pw90), 8 μs;
recycle delay, 3 s; total experiment run time, 7.5 min. Small relative
distortions of the relative peak areas, due to the short recycle delay,
were about the same as or less than the spectral signal-to-noise
ratios. For the kinetic determinations, it was necessary to trade off
precision with acquisition speed to obtain fast data acquisition.
Small-Angle X-ray Scattering (SAXS) Analysis: Data were collected
with an Anton Paar SAXSess instrument utilizing Cu-Kα radiation
(1.54 Å) and line collimation. Solutions containing 2 or 5 mm Lind-
qvist salt, with or without DPF, were sealed in a 1.5 mm-diameter
glass capillary tube for SAXS measurements and measured for 2 h
each. Background solutions were measured likewise. SAXSquant
software was used for data collection and treatment (normaliza-
tion, primary beam removal, background subtraction, desmearing,
and smoothing to remove extra noise created by the desmearing
routine). Rg (radius of gyration in the Guinier region of the scat-
tering curve was also determined by utilizing SAXSquant). All
other analyses and fits to determine size, shape, size distribution,
and PDDF (pair distance distribution function) were carried out
with the IRENA macros within the IgorPro 6 (Wavemetrics) soft-
ware.[32]
Because of the limited availability of the Bruker Avance 500 instru-
ment, comparison studies were performed with a Bruker 300 NMR
instrument with a liquids probe. It was observed that the water
added to the sample was sufficient to dissolve the PONb; therefore,
a solids NMR probe was not needed, as the agent was not spiked
on a solid matrix. However, it was necessary to add extra water
(220 μL instead of 100 μL) to fill the 5 mm NMR tube to the detec-
tion coil of the liquids probe. The extra water affected the molar
concentration of the solution. Also, 10% D2O was added as a lock
Solution Ion Probe Studies of DFP (NSRDEC): These studies were
carried out at the U.S. Army Natick Soldier Research, Develop-
Eur. J. Inorg. Chem. 0000, 0–0
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
[11] A. Kiselev, A. Mattson, M. Andersson, A. E. C. Palmqvist, L.
Osterlund, J. Photochem. Photobiol. A: Chem. 2006, 184, 125–
134.
[12] T. J. Bandosz, M. Laskoski, J. Mahle, G. Mogilevsky, G. W.
Peterson, J. A. Rossin, G. W. Wagner, J. Phys. Chem. C 2012,
116, 11606–11614.
[13] a) F.-J. Ma, S.-X. Liu, C.-Y. Sun, D.-D. Liang, G.-J. Ren, F.
Wei, Y.-G. Chen, Z.-M. Su, J. Am. Chem. Soc. 2011, 133, 4178–
4181; b) D. M. Mizrahi, S. Saphier, I. Columbus, J. Hazard.
Mater. 2010, 179, 495–499.
solvent, and no studies were done to determine whether it affected
the kinetics.
Kinetics of OP Neutralization
Calculating the kinetic data as it relates to decontamination has
been described in detail previously[5,29,33] and will only be briefly
outlined. The time-dependent degree of conversion, Ft, of agent to
breakdown product(s) was calculated by using:
ΣIp
Ft =
[14] C. L. Hill, N. M. Okun, D. A. Hillesheim, Y. V. Geletii in Anti-
terrorism and Homeland Defense: Polymers and Materials, Vol.
980 (Eds.: J. G. Reynolds, G. E. Lawson, C. J. Koester), Ameri-
can Chemical Society, Washington DC, 2007, pp. 198–209.
[15] a) N. M. Okun, T. M. Anderson, C. L. Hill, J. Am. Chem. Soc.
2003, 125, 3194–3195; b) N. M. Okun, J. C. Tarr, D. A. Hillesh-
iem, L. Zhang, K. I. Hardcastle, C. L. Hill, J. Mol. Catal. A
2006, 246, 11–17.
ΣIa + ΣIp
where ΣIp and ΣIa are the sum of the integration of signals for
breakdown product(s) and agent, respectively. The observed rate
(kobs) can then be calculated on the assumption of a pseudo-first-
order rate equation:
ln(1 – Ft) = –kobs
t
[16] R. D. Gall, C. L. Hill, J. E. Walker, Chem. Mater. 1996, 8,
followed by the calculation of half-life by:
2523–2527.
[17] a) J. R. Black, M. Nyman, W. H. Casey, J. Am. Chem. Soc.
2006, 128, 14712–14720; b) T. M. Alam, M. Nyman, B. R.
Cherry, J. M. Segall, L. E. Lybarger, J. Am. Chem. Soc. 2004,
126, 5610–5620.
[18] M. Nyman, Dalton Trans. 2011, 40, 8049–8058.
[19] a) T. M. Anderson, S. G. Thoma, F. Bonhomme, M. A. Rodri-
guez, H. Park, J. B. Parise, T. M. Alam, J. P. Larentzos, M. Ny-
man, Cryst. Growth Des. 2007, 7, 719–723; b) C. A. Ohlin,
E. M. Villa, W. H. Casey, Inorg. Chim. Acta 2009, 362, 1391–
1392; c) M. Nyman, T. M. Alam, F. Bonhomme, M. A. Rodrig-
uez, C. S. Frazer, M. E. Welk, J. Cluster Sci. 2006, 17, 197–219.
[20] T. Okuhara, N. Mizuno, M. Misono, Appl. Catal. A 2001, 222,
63–77.
[21] a) Z. Y. Zhang, Q. P. Lin, D. Kurunthu, T. Wu, F. Zuo, S. T.
Zheng, C. J. Bardeen, X. H. Bu, P. Y. Feng, J. Am. Chem. Soc.
2011, 133, 6934–6937; b) Z. L. Wang, H. Q. Tan, W. L. Chen,
Y. G. Li, E. B. Wang, Dalton Trans. 2012, 41, 9882–9884.
[22] N. B. Munro, S. S. Talmage, G. D. Griffin, L. C. Waters, A. P.
Watson, J. F. King, V. Hauschild, Environ. Health Perspect.
1999, 107, 933–974.
ln(2)
t1/2
=
kobs
Supporting Information (see footnote on the first page of this arti-
cle): Structures, PDDF and kinetic plots, X-ray scattering curves.
Acknowledgments
L. B. F. and M. N. acknowledge Oregon State University for sup-
port on collection and interpretation of SAXS data. M. K. K. ac-
knowledges the Sandia National Laboratories internal funding for
support. Sandia is a multiprogram laboratory operated by Sandia
Corporation, a Lockheed Martin Company, for the United States
Department of Energy under Contract DE-AC04-94AL85000. Ad-
ditional funding was provided by the Defense Threat Reduction
Agency for live agent testing at the Edgewood Chemical Biological
Center (ECBC) by Science Applications International Corporation
(SAIC) under contracts DAAD13-03-D-0017 and W911SR-11-C-
0047. This article has been reviewed and approved for public re-
lease: SAND-2013-1728 J and PAO U13-083.
[23] F. C. Hoskin, J. E. Walker, C. M. Mello, Chem. Biol. Interact.
1999, 119–120, 399–404.
[24] M. R. Antonio, M. Nyman, T. M. Anderson, Angew. Chem.
Int. Ed. 2009, 48, 6136–6140; Angew. Chem. 2009, 121, 6252.
[25] C. D. Putnam, M. Hammel, G. L. Hura, J. A. Tainer, Q. Rev.
Biophys. 2007, 40, 191–285.
[1] a) M. B. D. Nikitin, P. K. Kerr, A. Feickert, Syria’s Chemical
Weapons: Issues for Congress, 2013, report # R42848; b) Or-
ganization for the Prohibition of Chemical Weapons, United
Nations Mission to Investigate Allegations of the Use of
Chemical Weapons in the Syrian Arab Republic.
[2] G. W. Wagner, P. W. Bartram, J. Mol. Catal. A 1999, 144, 419–
424.
[26] Y. Hou, L. N. Zakharov, M. Nyman, J. Am. Chem. Soc. 2013,
135, 16651–16657.
[27] D. M. Tiede, R. T. Zhang, L. X. Chen, L. H. Yu, J. S. Lindsey,
J. Am. Chem. Soc. 2004, 126, 14054–14062.
[28] Potential Military Chemical/Biological Agents and Compounds
FM 3–11.9, MCRP 3–37.1B, NTRP 3–11.32, and AFTTP(I)
3–2.55, Jan 2005, report.
[29] L. Bromberg, H. Schreuder-Gibson, W. R. Creasy, D. J.
McGarvey, R. A. Fry, T. A. Hatton, Ind. Eng. Chem. Res. 2009,
48, 1650–1659.
[3] G. Lunn, E. B. Sansone, Appl. Biochem. Biotechnol. 1994, 49,
165–172.
[4] H. L. Schreuder-Gibson, Q. Truong, J. E. Walker, J. R. Owens,
J. D. Wander, W. E. Jones, MRS Bull. 2003, 28, 574–578.
[5] Y. C. Yang, J. A. Baker, J. R. Ward, Chem. Rev. 1992, 92, 1729–
1743.
[6] L. Bromberg, T. A. Hatton, Polymer 2007, 48, 7490–7498.
[7] L. Chen, L. Bromberg, J. A. Lee, H. Zhang, H. Schreuder-Gib-
son, P. Gibson, J. Walker, P. T. Hammond, T. A. Hatton, G. C.
Rutledge, Chem. Mater. 2010, 22, 1429–1436.
[30] a) J. H. Son, C. A. Ohlin, R. L. Johnson, P. Yu, W. H. Casey,
Chem. Eur. J. 2013, 19, 5191–5197; b) J. H. Son, C. A. Ohlin,
E. C. Larson, P. Yu, W. H. Casey, Eur. J. Inorg. Chem. 2013,
1748–1753; c) J.-H. Son, C. A. Ohlin, W. H. Casey, Dalton
Trans. 2012, 41, 12674–12677.
[31] Y. F. Wang, I. A. Weinstock, Chem. Soc. Rev. 2012, 41, 7479–
[8] a) L. Bromberg, H. Zhang, T. A. Hatton, Chem. Mater. 2008,
20, 2001–2008; b) H. Morales-Rojas, R. A. Moss, Chem. Rev.
2002, 102, 2497–2522.
7496.
[32] J. Ilavsky, P. R. Jemian, J. Appl. Crystallogr. 2009, 42, 347–353.
[33] D. Waysbort, D. J. McGarvey, W. R. Creasy, K. M. Morrissey,
D. M. Hendrickson, H. D. Durst, J. Hazard. Mater. 2009, 161,
1114–1121.
[9] L. Bromberg, T. A. Hatton, Ind. Eng. Chem. Res. 2007, 46,
3296–3303.
[10] a) Y. W. Xie, B. N. Popov, Anal. Chem. 2000, 72, 2075–2079; b)
Y. W. Xie, B. N. Popov, Anal. Chim. Acta 2001, 448, 221–230.
Received: January 8, 2014
Published Online: ᭿
Eur. J. Inorg. Chem. 0000, 0–0
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I40016
/KAP1
Date: 10-04-14 18:20:39
Pages: 8
www.eurjic.org
FULL PAPER
Polyoxoniobates
M. K. Kinnan, W. R. Creasy,
L. B. Fullmer, H. L. Schreuder-Gibson,
M. Nyman* ....................................... 1–8
Polyoxometalates, like small pieces of metal
oxide, can be dissolved or fixed on a sur-
face to perform homogeneous or hetero-
geneous catalysis, respectively. Here we ex-
ploit the alkaline nature of polyoxoniob-
ates to neutralize nerve agents in both solu-
tion and the solid state. Solution studies
correlate reaction efficacy to the associ-
ation of the dissolved polyoxoniobate with
its counterions.
Nerve Agent Degradation with Poly-
oxoniobates
Keywords: Polyoxometalates / Kinetics / Ion
pairs / Small-angle X-ray scattering / Nerve
agents
Eur. J. Inorg. Chem. 0000, 0–0
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim