The room temperature decomposition mechanism of dimethyl methylphosphonate (DMMP) on alumina-supported cerium oxide - Participation of nano-sized cerium oxide domains

Article abstract of DOI:10.1021/jp035590c

The adsorption and decomposition reactions of dimethyl methylphosphonate (DMMP) on cerium oxide supported on aluminum oxide have been examined at 25?°C. Experiments were carried out that involved dosing the reactive adsorbent with small doses of DMMP, followed by quantitative determination of the decomposition products. The results suggest that the formation reactions of methanol and dimethyl ether are competitive processes involving the same surface intermediate, which is most likely a surface methoxy species. Based on the observed results, it is proposed that the formation of dimethyl ether is due to the combination of two surface methoxy groups, while an important, if not the dominant, reaction producing methanol involves a surface methoxy group interacting with a vapor phase or physisorbed DMMP molecule. The presence of significant amounts of methoxy fragments formed upon DMMP adsorption is supported by results from diffuse reflectance spectroscopy, which also show that those groups are primarily associated with the cerium oxide domains. FT-Raman spectroscopy shows that the most active cerium oxide domains are highly dispersed two-dimensional domains or very small (< 1 nm) crystallites. Somewhat larger (< 6 nm) three-dimensional crystallites add to the decomposition yield, but less strongly. The FT-Raman evidence also supports the formation of a relatively narrow particle size distribution of cerium oxide crystallites on the alumina support surface from the sample preparation method. The alumina-supported cerium oxide reactive adsorbents developed as part of this study are the most active that have been reported in the literature for ambient temperature applications, decomposing approximately 775 ??mol of DMMP per gram of adsorbent at 25?°C, and strongly or irreversibly adsorbing an additional 400 ??mol, for a total capacity at room temperature of 1.1-1.2 mmol of DMMP per gram.

Full text of DOI:10.1021/jp035590c

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1634
J. Phys. Chem. B 2004, 108, 1634-1645
The Room Temperature Decomposition Mechanism of Dimethyl Methylphosphonate
(DMMP) on Alumina-Supported Cerium Oxide - Participation of Nano-Sized Cerium
Oxide Domains
Mark B. Mitchell,* Viktor N. Sheinker, Woodrow W. Cox, Jr., Enid N. Gatimu, and
Aron B. Tesfamichael
Department of Chemistry and the Center for Surface Chemistry, Clark Atlanta UniVersity,
Atlanta, Georgia 30314
ReceiVed: June 5, 2003; In Final Form: September 19, 2003
The adsorption and decomposition reactions of dimethyl methylphosphonate (DMMP) on cerium oxide
supported on aluminum oxide have been examined at 25 °C. Experiments were carried out that involved
dosing the reactive adsorbent with small doses of DMMP, followed by quantitative determination of the
decomposition products. The results suggest that the formation reactions of methanol and dimethyl ether are
competitive processes involving the same surface intermediate, which is most likely a surface methoxy species.
Based on the observed results, it is proposed that the formation of dimethyl ether is due to the combination
of two surface methoxy groups, while an important, if not the dominant, reaction producing methanol involves
a surface methoxy group interacting with a vapor phase or physisorbed DMMP molecule. The presence of
significant amounts of methoxy fragments formed upon DMMP adsorption is supported by results from diffuse
reflectance spectroscopy, which also show that those groups are primarily associated with the cerium oxide
domains. FT-Raman spectroscopy shows that the most active cerium oxide domains are highly dispersed
two-dimensional domains or very small (<1 nm) crystallites. Somewhat larger (<6 nm) three-dimensional
crystallites add to the decomposition yield, but less strongly. The FT-Raman evidence also supports the
formation of a relatively narrow particle size distribution of cerium oxide crystallites on the alumina support
surface from the sample preparation method. The alumina-supported cerium oxide reactive adsorbents developed
as part of this study are the most active that have been reported in the literature for ambient temperature
applications, decomposing approximately 775 µmol of DMMP per gram of adsorbent at 25 °C, and strongly
or irreversibly adsorbing an additional 400 µmol, for a total capacity at room temperature of 1.1-1.2 mmol
of DMMP per gram.
Introduction
Dimethyl methylphosphonate (DMMP) I is widely used as a
simulant for pesticides and for chemical warfare agents,
particularly for the study of the reactions on surfaces.1,2 An
understanding of the mechanism of decomposition of DMMP
The mechanism typically used to describe the decomposition
of DMMP on metal oxide surfaces is that proposed by
Templeton and Weinberg in their examination of the interaction
and reaction of DMMP and other phosphonates on alumina.23,24
In those studies it was found that DMMP initially adsorbs at
both Brønsted and Lewis acid sites through the phosphoryl
oxygen. At room temperature, the interaction at Brønsted acid
sites resulted in molecular adsorption. Dissociative adsorption
at Lewis acid sites (coordinatively unsaturated aluminum sites)
led to the formation of an adsorbed methyl methylphosphonate,
with the second methoxy group presumably forming methanol
either directly through reaction with a surface hydroxyl group,
or via a two-step process with a surface methoxy group as an
intermediate (Scheme 1). Once formed, the adsorbed methyl
methylphosphonate fragment is not volatile and passivates the
adsorption site(s) to which it is bound. Zhanpeisov et al.
suggested in their computational study that gas-phase or
adsorbed water could take the place of surface hydroxyl groups
in the formation of methanol.35 Dimethyl ether, also observed
as a product, has been suggested to be the product of a
is critical for the development of materials such as powders,
reactive fabrics, and coatings that can decompose chemical
warfare agents at low temperatures, for the protection of
personnel in the field who may have been exposed to such
agents, for the detoxification of equipment and surfaces, and
for the development of sensors for chemical warfare agents.3
A number of groups have studied the interactions of DMMP
with oxide surfaces at high temperatures for applications such
as chemical warfare agent destruction and demilitarization of
chemical weapons.4-13 Others have examined the photochemical
reactions of DMMP on photoactive surfaces for similar
applications,14-17 and several groups have examined the surface
adsorption of DMMP for the development of chemical agent
sensors.18-22 Several studies have examined the decomposition
of DMMP at ambient temperatures and lower.23-31 Wagner and
co-workers have investigated the interactions of VX, HD, and
other chemical agents with nanosized Al2O3, MgO, and
CaO.32-34
* Corresponding author. E-mail: mmitchel@cau.edu
10.1021/jp035590c CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/09/2004

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Adsorption and Decomposition of DMMP
SCHEME 1
J. Phys. Chem. B, Vol. 108, No. 5, 2004 1635
aluminum oxide support with a maximum surface area. While
the two alumina materials show product formation from DMMP
decomposition in approximately the same ratio as their surfaces
(26.2 µmol of vapor-phase carbon formed for η-Al2O3 vs 19.1
µmol for the γ-Al2O3), there are some differences in the product
branching ratio. When the DMMP decomposition is examined
on pure η-Al2O3 with a continuous flow of DMMP, less than
3% of the total vapor-phase carbon produced is dimethyl ether,
with the rest of the decomposition yield accounted for by the
formation of methanol. Under the same conditions, the γ-Al2O3
DMMP decomposition yield is approximately 10% dimethyl
ether and 90% methanol. These differences are probably
accounted for by differences in surface acidity.
The impregnated γ-alumina samples were the same ones as
described in an earlier publication and were prepared and
analyzed as discussed in that reference.27 The impregnated
η-alumina samples were prepared using a method known as
precipitation deposition, which has been used with success for
the preparation of supported cerium oxide materials with high
dispersion.36,37 In this method, a dilute ammonia solution is
slowly added to an aqueous solution of cerium nitrate, containing
an appropriate amount of the substrate material, until all of the
cerium has precipitated from solution (pH 9.5-10.0). Once
filtered and dried, the supported oxide is formed upon calcining
the solid in air at 400 °C. The cerium content was calculated
from the amount of cerium and η-Al2O3 used in the impregna-
tion solution, assuming complete conversion of the aqueous
cerium precursor to supported oxide.
secondary, acid-catalyzed condensation reaction between two
product methanol molecules.26
Previous investigations examining the room-temperature
decomposition of DMMP in this laboratory26,27 and in others,
including Segal et al.16 and Rusu and Yates,29 have yielded
results that are consistent with the mechanism suggested
in Scheme 1. One mechanism requiring two DMMP molecules
to yield gas-phase products has been proposed by Li and
Klabunde10 and Li et al.,8 who suggested that, following
decomposition of DMMP on magnesia to give an adsorbed
methoxy group and an adsorbed methyl methylphosphonate, a
second DMMP molecule could oxidize the adsorbed methanol,
ultimately yielding formic acid as a reaction product.
The current study was undertaken to shed further light on
the mechanism of decomposition of DMMP on these materials
at room temperature, so that an accurate assessment of the
amount of DMMP either decomposed or irreversibly adsorbed
on the adsorbents could be developed from measurements of
gas-phase products. In this study, the decomposition products
from small aliquots of DMMP dosed onto a reactive adsorbent
are measured as a function of the amount of added DMMP.
The results observed are significantly different from those
expected, and suggest that an isolated DMMP molecule on a
metal oxide surface is unlikely to form gas-phase products. The
observations are consistent with the formation of the observed
dimethyl ether by the reaction between two surface methoxy
fragments, and the formation of methanol from an interaction
between a surface methoxy fragment and a mobile physisorbed
or vapor-phase DMMP molecule. Additionally, the formation
of significant amounts of surface methoxy fragments is indicated
by the results from in situ infrared diffuse reflectance spectros-
copy. FT-Raman spectroscopic results show that the form of
the supported cerium oxide most active for decomposition is in
the form of two-dimensional patches or very small (<1 nm)
three-dimensional crystallites with additional decomposition
activity provided by somewhat larger (<6 nm) crystallites that
are formed as the cerium loading increases.
Microreactor Studies. The microreactor system has been
described in detail previously, and only a short description is
given here.26 The microreactor consists of a 1/4 in. o.d. stainless
steel u-tube in a controlled-temperature furnace, which is
connected to a gas inlet system that carries the DMMP/He
mixture to the reactor, and an exit system that carries the
products to a 2.4 m long-path infrared gas cell for analysis.
The inlet helium gas flow is controlled by a Teledyne-Hastings
mass flow controller. The helium gas bubbles through liquid
DMMP at a flow rate of 30 mL/min. After modification of the
flow system, compared to that described in the earlier work,
and recalibration of the DMMP flow, the DMMP concentration
in the He stream has been determined to be 47.0 µmol/L. At a
flow rate of 30.0 mL/min, the flow rate of DMMP was
determined to be 1.41 µmol/min.
In the microreactor experiments, the adsorbent was placed
near the bottom of the u-tube but on the entrance side. The
adsorbent bed was composed of particles with a diameter of
90-250 µm, classified using standard testing sieves, and was
supported on a porous stainless steel disk with 2 µm diameter
pores (Mott Corporation). Prior to exposure to DMMP vapor,
the adsorbent was heated to 400 °C in 20% O2 in helium for 1
h, then in pure helium for 30 min. The adsorbent was then
cooled to 25 °C in a pure helium flow. The DMMP vapor-
generation system was contained in a temperature-controlled
enclosure maintained at 25 °C. After the microreactor, the gas
streamflowed to a long-path gas cell (Ultra-Mini Long Path Cell
from Infrared Analysis, Inc.) with an effective path length of
2.4 m and an internal volume of 100 mL. Infrared spectra
(Thermo-Nicolet Avatar 360 FT-IR), at 2 cm-1 resolution, were
collected approximately every two minutes. The data analysis
and product quantitation were carried out as in the earlier study.
New to this study is the completion of the mass balance for the
system. The DMMP breakthrough was calibrated by allowing
small doses of DMMP to flow, at a known flow rate and
concentration, into the infrared cell, and the infrared absorption
Experimental Section
The helium used for the study was a 99.9995% ultrahigh
purity grade from Holox. DMMP was purchased from Aldrich
and distilled under vacuum before use. The adsorbents were
prepared using incipient wetness, or other aqueous-phase
impregnation methods, with iron and cerium nitrates as the
precursors and deionized water as the solvent. The alumina used
was provided by UOP. Two different types were used, a γ-Al2O3
material called Versal GH (253 m2/g) and a Bayerite material
called Versal B that is transformed into an η-alumina on
calcining (360 m2/g). The η-alumina was chosen to provide an

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1636 J. Phys. Chem. B, Vol. 108, No. 5, 2004
Mitchell et al.
due to DMMP measured for a very long period of time, typically
24-28 h, so that all of the DMMP flows through the infrared
cell. The integrated infrared signal then corresponds to the dose
amount. A straight line fit for a series of integrated intensities
vs dose amounts then formed the response curve for DMMP.
The precision of the measurement is not as good as it is for the
measurement of methanol and DMMP, but it is good enough
to be able to draw important conclusions regarding the
decomposition yields.
and detected using a room-temperature InGaAs detector. The
neat powder samples were placed in capillary tubes for data
collection.
Results
Microreactor Studies. Prior to describing the dosing experi-
ments, it is useful to describe the results from several different
control experiments that were carried out to help evaluate the
results. In the first, a mixture of methanol in helium, 253 ppm
from Matheson Gas, was allowed to flow continuously through
the adsorbent bed at 25 °C. No conversion to dimethyl ether
was observed, although conversion to dimethyl ether can be
observed at 150 °C. Similarly, a mixture of dimethyl ether in
helium, 490 ppm from Matheson Gas, gave no conversion to
methanol at 25 °C. Finally, the DMMP in helium mixture
flowing into an empty reactor did not yield any measurable
decomposition products.
Initially, it was thought that 20-24 h would be more than
sufficient to allow for all of the reaction products to flow from
the adsorbent bed. In studies that we had carried out previously
using the microreactor setup, both methanol and dimethyl ether
were observed to flow from the bed relatively quickly.26,27
However, those experiments were carried out using a continuous
flow of DMMP, and the current experiments using small doses
yield significantly different results as described below.
Variable Dose/Fixed Adsorbent Amount. A series of experi-
ments was carried out on reactive adsorbents that had been
examined in an earlier study.27 These adsorbents were prepared
using incipient wetness impregnation and included 7.5 wt %
Ce and 5.0 wt % Fe co-impregnated on γ-Al2O3, 5.0 wt % Fe
on γ-Al2O3, 7.5 wt % Ce on γ-Al2O3, and γ-Al2O3. The results
for all of these adsorbents were similar and showed an apparent
maximum in product formation of 23-27 µmol of product after
exposure to an 80 µmol or larger dose of DMMP. The products
formed represented a combination of dimethyl ether and
methanol. For dose amounts less than 30 µmol of DMMP, the
only product measured was dimethyl ether. Exposure to a 40-
50 µmol dose of DMMP resulted, for most of the solids, in the
observation of large amounts of methanol and diminishing
amounts of dimethyl ether. These results led us to suspect that
the processes leading to gas-phase products might be more
complex than originally assumed.
A new series of reactive adsorbents was developed that
showed significantly higher decomposition yields per gram of
adsorbent. Our previous work showed that higher dispersions
of CeO2 yielded higher decomposition yields, and the precipita-
tion deposition method used to prepare the new adsorbents in
the current study had been shown by Soria and co-workers to
yield high dispersions of supported cerium oxide.36,37 The most
reactive adsorbent had a composition of approximately 20 wt
% Ce on η-Al2O3 and was used to obtain all of the microreactor
results discussed. Samples with lower amounts of cerium
showed lower decomposition yields as discussed in a later
section. The decomposition product yield of DMMP as a
function of DMMP dose for this adsorbent is shown in Figure
1A and B. These graphs show the amounts of methanol and
dimethyl ether produced as a function of dose amount. The
observation of products begins at somewhat lower doses on this
solid than on those studied earlier, and the new solid is capable
of forming approximately 44% more products than the previous
formulations. These results are tabulated in Table 1.
Experiments using the pure alumina materials were carried
out multiple (more than three) times to evaluate the experimental
setup and data analysis for reproducibility. Typically, an
uncertainty of approximately (2 µmol is found for the deter-
mination of the integrated amount of methanol formed. The
uncertainty found for dimethyl ether is less than that due to the
fact that there are no interfering absorptions for the dimethyl
ether gas-phase infrared absorption bands used for quantitation,
while those of methanol are overlapped somewhat by DMMP
absorptions.
Several different types of dosing experiments were carried
out, and for each the pretreatment was carried out as described
above. In the variable dose/fixed adsorbent experiment, 60 mg
of the adsorbent was exposed to different dose amounts of
DMMP. The dose amounts are formed by exposing the
adsorbent to a flow of DMMP in He for an amount of time
necessary to dose the sample with the predetermined amount
of DMMP. The flow through the DMMP liquid is then stopped,
and pure helium is directed to flow through the microreactor
for the remaining time, at a constant flow rate of 30.0 mL/min.
Typically, the flow of helium is continued for approximately
20 h (a total flow of helium of 36-37 L), compared to a dosing
time of about 7 min for a 10 µmol dose of DMMP.
In the fixed dose/variable adsorbent experiment, several
different amounts of adsorbent were exposed to fixed doses of
DMMP, typically 20 µmol. The dose was again followed by a
20-hour continuous flow of pure helium and monitoring of
products. The incremental dose/single adsorbent experiment was
performed by dosing 60 mg of the adsorbent with successive
10 µmol doses of DMMP. Approximately 20 h were allowed
to elapse between each dose, and infrared spectra were recorded
continuously.
In Situ FT-IR Studies. The infrared diffuse reflectance
studies were carried out in much the same way as described in
an earlier publication.38 The adsorbent to be examined was
sieved and placed in a Harrick Scientific HVC-DR infrared
diffuse reflectance controlled environment cell, the cell was
closed and evacuated using an Alcatel diffusion pump. The
sample was then exposed to a mixture of 20% O2 in helium at
400 °C for approximately 30 min, and subsequently cooled to
room temperature. All of the infrared spectra were collected
using a Harrick Scientific DRA-2 optical accessory. After
collection of a background spectrum, the dilute mixture of
DMMP in helium (1:1000) was allowed to flow through the
cell over the sample. The pressure in the cell during this time
was approximately 500 mTorr. Spectra were measured during
the adsorption process to monitor for DMMP adsorption using
a Nicolet Magna 750 FT-IR and a liquid nitrogen cooled
mercury-cadmium-telluride (MCT) detector. During adsorp-
tion, 250 scans at 8 cm-1 resolution (to minimize collection
time) were co-added to form the spectra shown.
FT-Raman Studies. FT-Raman spectra were obtained using
a Nicolet Raman 950 FT-Raman spectrometer using a laser
power of approximately 250 mW. Raman scattered radiation
was collected using a 180° backscattering collection geometry
The only product observed for the lowest doses of DMMP
onto the solid, 10 and 20 µmol, is dimethyl ether. Figures 2
and 3 show the formation of the products of the DMMP

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Adsorption and Decomposition of DMMP
J. Phys. Chem. B, Vol. 108, No. 5, 2004 1637
curve with a long tail. A significant change in the characteristics
of the breakthrough curve occurs for dose amounts between
100 and 150 µmol; a strong, relatively narrow peak due to
DMMP flow is superimposed on the leading edge of the broad
DMMP flow curve observed for the lower doses. These results
suggest that the surface can be very roughly characterized as
being composed of two kinds of DMMP adsorption site,
consistent with the results reported by Templeton and Wein-
berg23,24 and by others in earlier studies. The first type of site
is a very strong adsorption site that irreversibly adsorbs DMMP
at these temperatures, associated with Lewis acid sites, while
the second type of site is a more weakly adsorbing physisorption
site, associated with Brønsted acid sites. Saturation of the
chemisorption sites leads to the observation of the broad DMMP
breakthrough curve, while saturation of both the chemisorption
and physisorption sites leads to the strong, narrow peak.
From Figure 3B, it is clear that for DMMP doses of 70 µmol
and greater, methanol evolution begins at the nearly the same
time, regardless of dose. The results suggest that once the
strongest adsorption sites are occupied by DMMP and DMMP
breakthrough occurs, methanol produced by the decomposition
reaction evolves from the bed with little interaction with the
surface.
Figure 4 shows the amount of methanol observed as a
function of the DMMP dose, where the total amount of methanol
is taken to be the integrated methanol flow measured from the
beginning of the experiment to slightly more than 20 h after
the DMMP dose. The form of the data suggests that some
threshold amount of DMMP is needed before methanol is
observed. The data are fit to a function of the form
a(DMMP - DMMP0)
Figure 1. Decomposition products formed as a function of amount of
DMMP dose for 20 wt % Ce on η-Al2O3. The error bars represent the
estimated uncertainty associated with the determination of the amounts
of reaction products formed. The error bars associated with the dimethyl
ether are smaller since the infrared bands used to determine the dimethyl
ether yield are free of interferences, especially those due to DMMP,
whereas there is some overlap of the methanol infrared bands with
those of DMMP.
MeOH )
(1)
b + (DMMP - DMMP0)
where MeOH is the amount, in µmol, of methanol observed
after a dose, and DMMP is the number of µmol of DMMP dosed
onto the surface, while DMMP0 is the x-axis intercept in the
figure, determined to be 24.3 µmol. Equation 1 is functionally
equivalent to the Langmuir adsorption isotherm, indicating that
the amount of methanol produced is proportional to the surface
coverage of DMMP, after subtraction of the value DMMP0. The
limiting amount of methanol produced for exposure to an infinite
amount of DMMP is represented by the constant a in eq 1 and
is found to be 45.5 µmol. The constant b is related to the rate
of rise of MeOH with DMMP dose amount. Experimentally,
the amount of methanol observed after a flow of 467 µmol of
DMMP was found to be 41.5 µmol, compared to a value of
42.5 µmol predicted by eq 1.
Variable Adsorbent/Fixed Dose. To lend further insight into
the reactions occurring on the surface, a set of experiments was
carried out in which variable amounts of adsorbent were dosed
with a fixed amount of DMMP. The amounts of adsorbent tested
were 15 mg, 30 mg, and 45 mg. The 15 and 45 mg samples
were dosed with 20 µmol of DMMP, while two experiments
were carried out with 30 mg of sample, one dosing it with 20
µmol and one with a 40 µmole dose.
decomposition reaction as a function of time for different
DMMP doses. Included in Figure 2 is the measured DMMP
flow for a 10 µmol dose of DMMP into an empty reactor (the
long tail is due to the time needed to flush the DMMP from the
gas cell). This provides an estimate of the DMMP concentration
profile that the sample is exposed to. The results suggest that
dimethyl ether formation is a relatively low probability event,
because for a 20 µmol dose of DMMP, less than 1 µmol of
dimethyl ether is produced. Also, it takes a long period of time
for the dimethyl ether to evolve from the adsorbent bed, much
longer than would have been predicted from our earlier studies.
The integrated product flow curves for DMMP doses of 30
to 200 µmol are shown in Figure 3A. These curves show the
integrated product flow from the adsorbent as a function of time,
where integrated product is expressed as the equivalent number
of µmol of C1 product formed, equal to the number of µmol of
methanol plus two times the number of µmol of dimethyl ether.
The initial product observed is dimethyl ether in all cases, and
the time between the end of the dimethyl ether flow and the
beginning of the methanol flow leads to the odd shape of the
integrated product flow curves for the 30-50 µmol doses in
Figure 3A. Figure 3B shows the integrated flow of methanol
as a function of time.
For 15 mg of the adsorbent, 20 µmol of DMMP more than
saturated the adsorbent, DMMP breakthrough was observed,
and methanol was the only product observed; no dimethyl
ether was found. For the same dose on 30 mg of adsorbent,
no DMMP breakthrough was observed and both methanol and
dimethyl ether were observed, although slightly less methanol
was observed than for the 15 mg experiment. For 45 mg of
adsorbent, for the same 20 µmole dose, only dimethyl was
DMMP breakthrough is observed only for dose amounts of
60 µmol and greater. The DMMP breakthrough curves (DMMP
flow in µmol/min) are characterized by a broad asymmetric