Subscriber access provided by Washington University | Libraries

Article
Adsorption of Dimethyl Methylphosphonate
on MoO: The Role of Oxygen Vacancies
3

Ashley Rose Head, Roman V. Tsyshevsky, Lena Trotochaud, YI YU,

Line Kyhl, Osman Karsl#o#lu, Maija M. Kuklja, and Hendrik Bluhm
J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.6b07340 • Publication Date (Web): 07 Dec 2016
Downloaded from http://pubs.acs.org on December 7, 2016

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society.

1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
Adsorption of Dimethyl Methylphosphonate on
9
10
11
12 MoO3: The Role of Oxygen Vacancies
13
14
15
16
17 Ashley R. Head,1 Roman Tsyshevsky,2 Lena Trotochaud,1 Yi Yu,1,3 Line Kyhl,1,4 Osman
18
19 Karslıoǧlu,1 Maija M. Kuklja,2,* Hendrik Bluhm1,*
20
21
22 1
23
Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720
24
25
2
26 Materials Science and Engineering Department and 3 Department of Chemistry and
27 Biochemistry, University of Maryland, College Park, MD 20742
28
29 4
iNANO, University of Aarhus, DK-8000 Aarhus C, Denmark
30
31
32 KEYWORDS: ambient pressure X-ray photoelectron spectroscopy, density functional theory,
33
34
35 chemical warfare agent simulant, metal oxide hydroxylation, organophosphonate
36
37
38
39 Abstract
40
41
42
43
Dimethyl methylphosphonate (DMMP) is a common chemical warfare agent simulant and is
44
45 widely used in adsorption studies. To further increase the understanding of DMMP interactions
46
47 with metal oxides, ambient pressure X-ray photoelectron spectroscopy was used to study the
48
49
50 adsorption of DMMP on MoO3, including the effects of oxygen vacancies, surface hydroxyl
51
52 groups, and adsorbed molecular water. Density functional theory calculations were used to aid in
53
54 the interpretation of the APXPS results. An inherent lack of Lewis acid metal sites results in
55
56
57 weak interactions of DMMP with MoO3. Adsorption is enhanced by the presence of oxygen
58
59
60
ACS Paragon Plus Environment
1
 The Journal of Physical Chemistry Page 2 of 54

1
2
3
vacancies, hydroxyl groups, and molecular water on the MoO3 surface, as measured by
4
5
6 photoelectron spectroscopy. Computational results agree with these findings and suggest the
7
8 formation of methanol through several possible pathways, but all require a proton transferred
9
10
11
from a hydroxyl group on the surface.
12
13
14
Introduction
15
16
17 Organophosphonates represent an important class of molecules due to their effectiveness as
18
19 pesticides and nerve agents.1 Currently, there is great interest in understanding the interaction
20
21
22
between these molecules and surfaces so new materials can be developed for personal protective
23
24 devices, molecular sensors, and catalysts for the destruction of these compounds.1 Dimethyl
25
26 methylphosphonate (DMMP) is frequently used as a model compound in adsorption studies of
27
28
29 chemical nerve agents.1 While much less toxic than the nerve agent sarin, DMMP retains the
30
31 phosphoryl group (P=O), methyl group, and an oxy group (Figure 1a,b). In the search for new
32
33 materials for applications and a better understanding of DMMP surface chemistry, especially on
34
35
36 metal oxides, adsorption studies with a focus on analyzing surface species have been conducted
37
38 on TiO2,2-8 ZnO,2 Al2O3,2,9-11 MgO2,9 WO3,2,5 La2O3,9 Fe2O3,9,12,13 CeO2,14 SiO2,12,15 Y2O3,16 and
39
40
combinations of these as supported metal oxides.17,18 Several techniques have been used to
41
42
43 investigate the surface species, including infrared (IR) spectroscopy,2,4-10,12-18 Raman
44
45 spectroscopy,18 X-ray photoelectron spectroscopy (XPS),3,6,14,16,17 inelastic tunneling
46
47
48 spectroscopy,,11 temperature programmed desorption (TPD),3,12,14 Auger electron spectroscopy,12
49
50 X-ray diffraction,17 and ion chromatography.17 The results of these studies show that DMMP
51
52 binds to metal oxides at room temperature primarily through the phosphoryl oxygen to Lewis
53
54
55 acid metal sites on the surface, namely under-coordinated metal atoms inherent to the structure
56
57 or as surface defects. Hydroxyl groups on the surface are the second favored binding site.
58
59
60
ACS Paragon Plus Environment
2
Page 3 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 1. (a) Chemical nerve agent sarin and (b) dimethyl methylphosphonate (DMMP), (c)
30
31
32
structure of gas phase DMMP, (d) slab model of the MoO3(010) surface, and (e) surface
33
34 structure of MoO3, including an oxygen vacancy and a hydroxylated oxygen vacancy. P atom is
35
36 green, C atoms are brown, H atoms are beige, O atoms are red, and Mo atoms are lilac.
37
38
39
40
In light of the adsorption and reactivity of DMMP with other metal oxide surfaces, we
41
42 have chosen to explore the interaction of DMMP on a polycrystalline MoO3 oxide surface, which
43
44 contrasts with other metal oxides that have been studied since the MoO3 surface structure
45
46
47 inherently lacks under-coordinated metal atoms. As shown in Figure 1d, the bulk structure of
48
49 MoO3 consists of bilayers with hexagonally coordinated Mo atoms that are O-terminated and is
50
51 often considered rather chemically inert.19 Computational studies on the adsorption of small
52
53
54 molecules, such as oxygen,20 water,21 nitric oxide,22 and small hydrocarbons,23 show that the
55
56 interaction of adsorbates with a stoichiometric MoO3 surface are typically weak Van der Waals
57
58
59
60
ACS Paragon Plus Environment
3
 The Journal of Physical Chemistry Page 4 of 54

1
2
3
interactions. However, the presence of oxygen vacancies on the surface results in under-
4
5
6 coordinated Mo5+ atoms (cf. Figure 1e) and increases the adsorption and reactivity of molecules
7
8 at these sites.22 MoO3 serves as a catalyst in several industrial reactions, such as the
9
10
11
dehydrogenation of alcohols24-26 and the oxidation of propene,27 with studies suggesting that
12
13 oxygen vacancies play a role in the catalytic activity.19,24 While oxygen vacancies can exist as
14
15 point defects in MoO3, under certain conditions, they tend to aggregate to form plane defects or
16
17
18 structurally ordered, oxygen-deficient phases known as Magnéli phases.28
19
20 Here, we have used ambient pressure X-ray photoelectron spectroscopy (APXPS) to compare
21
22 the room temperature DMMP adsorption on a polycrystalline MoO3 surface and a MoO3 surface
23
24
25 containing oxygen vacancies and OH groups. This technique allows for the collection of in situ
26
27 X-ray photoemission (PE) spectra of a substrate under pressures in the Torr range and provides
28
29
insight on pressure-dependent adsorption behavior.29-31 We found that DMMP does not interact
30
31
32 strongly with the O-terminated surface of MoO3. Introducing oxygen vacancies results in plane
33
34 defects or possibly Magnéli phases, hydroxyl groups, and molecularly adsorbed water. These
35
36
37 surface features were found to enhance DMMP adsorption. Although the majority of the DMMP
38
39 molecules physisorb intact, the C 1s spectra indicate some carbon loss. Density functional theory
40
41 (DFT) calculations to simulate the adsorption process, decomposition reactions, and core level
42
43
44 spectra suggest that methanol is formed by reaction with surface hydroxyl groups and desorbs
45
46 from the surface, in agreement with the APXPS measurements.
47
48 Experimental Details
49
50
51 Materials and Methods
52
53 The experiments were conducted at the APXPS end station32 of the 11.0.2 beamline33 at
54
55
56
the Advanced Light Source at Lawrence Berkeley National Laboratory. The end station is
57
58
59
60
ACS Paragon Plus Environment
4
Page 5 of 54 The Journal of Physical Chemistry

1
2
3
equipped with a chamber for sample preparation and an analysis chamber. The base pressure of
4
5
6 both chambers was better than 5 x 10-9 Torr. A conical aperture with a diameter of 0.2 mm is the
7
8 entrance from the analysis chamber to the differentially-pumped electrostatic lens system34 of the
9
10
11
electron energy analyzer, allowing in situ XPS measurements up to pressures of several Torr.
12
13 In the preparation chamber, a piece of Mo foil (99.99% Alpha Aesar), was cleaned with
14
15 one Ar+ ion sputter cycle (5 mA emission current, 1.5 keV, 1 x 10-5 Torr Ar, <~6.2 x 1012 Ar+
16
17
18 cm-2 s-1, 5 min) followed by annealing for 5 min at 900 ºC. The MoO3 surface was prepared by
19
20 heating the foil to 430 ºC for 10 min in 35 Torr of O2 followed by cooling to 100 ºC under the
21
22 same oxygen environment. To promote the hydroxylation of surface defect sites, the surface was
23
24
25 exposed to 10 Langmuir (L) of water. Using the known inelastic mean free path of the
26
27 photoelectron in MoO3 at 200 eV kinetic energy and the ratio between the hydroxyl and oxide
28
29
peak intensities, the hydroxyl coverage was estimated to be about 10% of a monolayer. To
30
31
32 prepare MoO3 with more oxygen vacancies, a second sample was prepared, briefly sputtered (1 x
33
34 10-5 Torr Ar, 1 keV, 5 mA, 3 min), and exposed to 10 L of water. The hydroxyl coverage was
35
36
37 estimated to be about 30% of a monolayer for this sample. By collecting PE spectra at different
38
39 areas of the sample, the surface was found to be homogenously MoO3 with a consistent amount
40
41 of Mo5+. DMMP (>97%, Fluka) was degassed via freeze-pump-thaw cycles and was introduced
42
43
44 into the analysis chamber via a leak valve.
45
46 Electrons were collected at an angle 42º from the sample normal. To minimize photon-
47
48 induced changes to the substrate and adsorbed DMMP, each spectrum was taken at a new
49
50
51 location on the sample, so all spectra presented are free from photon-induced effects. The
52
53 Supporting Information further discusses these photon-induced changes. The photon energies
54
55
56
were chosen to yield a photoelectron kinetic energy of 200 eV for all elements, thus probing the
57
58
59
60
ACS Paragon Plus Environment
5
 The Journal of Physical Chemistry Page 6 of 54

1
2
3
same depth. The energies used were 340 eV for P 2p, 435 eV for Mo 3d, 490 eV for C 1s, and
4
5
6 735 eV for O 1s. Each spectrum on the MoO3 substrate was calibrated by the Mo 3d5/2 peak at
7
8 232.5 eV.35 Due to a large shift of the Mo 3d5/2 in the sputtered MoO3 sample (OH-MoO3), the C
9
10
11
1s and P 2p spectra of this system were shifted to align with those of the pristine MoO3 sample;
12
13 more details are provided in the Supporting Information. A Shirley background was removed
14
15 from the Mo 3d spectra, and a polynomial background from all others. The peaks were fit using
16
17
18 Voigt functions. The spin-orbit components of the P 2p spectra were constrained to have the
19
20 same peak widths, a 2:1 area ratio, and a binding energy separation of 0.85 eV.36 The
21
22 components of the C 1s spectra were constrained to have identical peak widths. The combined
23
24
25 beamline and electron energy analyzer resolution was better than 350 meV.
26
27 DFT Calculations
28
29
Among the various crystal structures and surface terminations of MoO3, the (010) termination
30
31
32 of the α-phase has been found to be the most thermodynamically stable35 in both
33
34 experimental37,38 and theoretical20,23,39,,41 studies. Hence, this surface was used in our
35
36
37 computational simulations. This common polymorph of molybdenum trioxide has an
38
39 orthorhombic unit cell with the lattice constants of a= 3.96 Å, b= 13.86 Å, c= 3.70 Å.42,43 The
40
41 atomistic structure consists of bilayers parallel to the (010) plane that are bound by mainly
42
43
44 electrostatic and weak van der Waals (vdW) interactions. Recent theoretical studies emphasized
45
46 the importance of vdW contributions in density functional theory (DFT)44,45 calculations to
47
48 reproduce an accurate crystal structure of α-MoO3.46,47 On the other hand, formation energies of
49
50
51 surface oxygen vacancies on α-MoO3 (010) surface were found to depend on the accurate
52
53 treatment of the electron localization.20
54
55
56
57
58
59
60
ACS Paragon Plus Environment
6
Page 7 of 54 The Journal of Physical Chemistry

1
2
3
In the current study, solid state periodic calculations were performed by employing DFT with
4
5
6 the optPBE-vdW48,50-52 functional, which includes corrections for van der Waals interactions, as
7
8 implemented in the VASP code.53,55 The projector augmented wave (PAW) pseudo-potentials56
9
10
11
were used. In calculations of a stoichiometric MoO3 crystal, the convergence criterion for the
12
13 total energy was set to 10-5 eV, and the maximum force acting on each atom in the periodic
14
15 supercell was set not to exceed 0.02 eV/Å. The 2×6×6 Monkhorst−Pack k-point mesh with a
16
17
18 kinetic energy cut-off of 520 eV was used. The optimized lattice constants of the orthorombic
19
20 unit cell with the Pbnm space group (Figure S7) were a= 3.99 Å, b= 14.3 Å, c= 3.73 Å, agreeing
21
22 with experimentally determined42,43 within 3%.
23
24
25 In the MoO3 surface calculations (cf. Figure 1d), the model surface slab was cut out of the bulk
26
27 MoO3 structure to form the (010) surface, with the supercell lattice vectors of a =14.92 Å, b
28
29
=15.97 Å, and c =33.50 Å. A vacuum layer of 20 Å placed on top of the MoO3 (010) surface
30
31
32 served to minimize interactions between the supercells in the z-direction and to avoid any
33
34 significant overlap between wave functions of periodically translated cells. An oxygen vacancy
35
36
37 on the MoO3 (010) surface was simulated by removing one of the terminal oxygen atoms from
38
39 the top layer of 256 atom supercell (cf. Figure 1e), which requires less energy than removing
40
41 bridging oxygen atoms.20,57
42
43
44 To model adsorption and decomposition of DMMP on a hydroxylated MoO3 (010) surface,
45
46 four terminal oxygen atoms were replaced by hydroxyl groups (Figure 1e), as it has been shown
47
48 previously that hydrogen adsorbs more strongly to the terminal rather than bridging oxygen40,41
49
50
51 due to the strong covalent interaction of H 1s and O 2p atomic orbitals41. Surface supercell
52
53 calculations were performed at the Gamma point only, with the convergence criteria for
54
55
56
electronic and ionic steps set to 10-5 eV and 0.03 Å/eV, respectively.
57
58
59
60
ACS Paragon Plus Environment
7
 The Journal of Physical Chemistry Page 8 of 54

1
2
3
Calculations with the PBE+U method58,59 were performed to reveal the effect of the
4
5
6 electron localization on the adsorption of DMMP on the MoO3 (010) surface. A value of 6.3 for
7
8 U parameter was taken from the study by Coquet and coworkers.20 Our calculations revealed that
9
10
11
PBE+U tends to significantly overestimate “b” lattice constant of the MoO3 unit cell by ~9% due
12
13 to its failure to describe correctly the vdW interactions between MoO3 bilayers. Therefore, to
14
15 probe the adsorption of DMMP on the MoO3 (010) surface with the PBE+U approach, we
16
17
18 employed the supercell that was previously built for optPBE-vdW calculations in which atoms
19
20 only from the four top atomic layers were allowed to relax. The positions of the rest atoms were
21
22 fixed (see the Supporting Information for more details). PBE+U calculations were performed at
23
24
25 the Gamma point only with the kinetic energy cut-off set to 450 eV.
26
27 Our simulations show that geometries of pristine MoO3 and defect-containing MoO3
28
29
(010) surfaces as well as energies of DMMP adsorption calculated using optPBE-vdW and
30
31
32 PBE+U functionals were found to be in satisfactory agreement. At the same time, adsorption
33
34 energies obtained from DFT+U strongly suffer from the lack of corrections for vdW interactions.
35
36
37 As the result, DFT+U does not perform for our task with the required accuracy, as detailed in
38
39 Supporting Information. Therefore, optPBE-vdW was chosen as the main tool for modeling
40
41 decomposition reactions of DMMP on MoO3 surfaces.
42
43
44 In test calculations to model the gas phase decomposition using the VASP code, a single
45
46 DMMP molecule was placed in the center of a large periodic box with the cell parameters
47
48 a=b=c=20 Å. Additionally, calculations with the B3LYP60,61 functional were performed with the
49
50
51 6-31+G(d,p) basis set, as implemented in the Gaussian09 code.62 Consistency between periodic
52
53 (VASP) and molecular (GAUSSIAN) calculations lends an additional support for accuracy of the
54
55
56
results and serves as validation of obtained conclusions.
57
58
59
60
ACS Paragon Plus Environment
8
Page 9 of 54 The Journal of Physical Chemistry

1
2
3
Minimal energy paths in the VASP periodic calculations were obtained with the nudged elastic
4
5
6 band method63 with five intermediate images. Atomic positions were relaxed using conjugate
7
8 gradient and quasi-Newtonian methods within a force tolerance of 0.05 Å/eV. A Hessian-based
9
10
11
Predictor-Corrector integrator algorithm64,65 was employed to conduct intrinsic reaction
12
13 coordinate computations in the Gaussian09 code. The stationary points corresponding to the
14
15 energy minimum were positively identified by having no imaginary frequencies and the
16
17
18 transition states had exactly one imaginary frequency. Calculations of vibrational frequencies in
19
20 VASP, especially for such a large system, is a tedious procedure. Therefore, we calculated
21
22 vibrational frequencies of transition state structures for several possible reaction pathways of
23
24
25 methanol elimination (see Table S1 of Supplementary Information). In calculations of vibrational
26
27 frequencies, only ionic positions of DMMP atoms and surface hydroxyl groups were allowed to
28
29
be changed, whereas positions of other atoms were fixed.
30
31
32 Results
33
34 Ambient pressure XPS
35
36
37 Before dosing DMMP, the oxidized Mo foil was characterized by measuring the Mo 3d and O
38
39 1s spectra (cf. Figures 2a and 3a), which agree with those of MoO3 found in the literature.35 No
40
41 peak corresponding to the metal substrate was detected. The binding energy of the small
42
43
44 shoulder in the Mo 3d spectra matches that of a Mo5+ oxidation state.35,66-69 While the Mo5+
45
46 component was found to be linked to photon-induced damage of the substrate,70 the peak was
47
48 present also in spectra where exposure to the beam was minimized (see Supporting Information).
49
50
51 Oxygen vacancies on the surface give rise to Mo5+ species, either in the form of point defects,
52
53 plane defects, or extended oxygen-deficient phases,28 and these Mo5+ species are feasibly due to
54
55
56
defects, such as grain boundaries, in the polycrystalline foil from which MoO3 was grown.
57
58
59
60
ACS Paragon Plus Environment
9
 The Journal of Physical Chemistry Page 10 of 54

1
2
3
Additionally, in a previous report of single crystal MoO3, it was difficult to completely oxidize
4
5
6 the surface to MoO3 and remove all oxygen vacancies.35 Therefore, we assume that the prepared
7
8 surface contained some oxygen vacancies prior to photon beam exposure.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33 Figure 2. Mo 3d PE spectra of (a) pristine MoO3, (b) pristine MoO3 (black) and under a
34
35
36 DMMP pressure of 1 x 10-2 Torr (blue), (c) OH-MoO3 as prepared, and (d) OH-MoO3 as
37
38 prepared (black) and under a DMMP pressure of 1 x 10-4 Torr (blue). Voigt functions of the
39
40
spectral fits (red lines) are shaded with the dark grey representing Mo5+ and the lighter grey
41
42
43 representing Mo6+. All spectra are normalized to their maximum intensity.
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
10
Page 11 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 3. O 1s PE spectra of (a) pristine MoO3 , (b) pristine MoO3 under a DMMP pressure of 1
27
28
29
x 10-4 Torr, (c) OH-MoO3 as prepared, and (d) OH-MoO3 under a DMMP pressure of 1 x 10-4
30
31 Torr. The spectra are normalized to their maximum intensity. The black dots are the data points,
32
33 the lilac line is the fit of the Voigt functions (yellow is bulk MoO3, blue is hydroxyl groups, light
34
35
36 blue is molecular water, dark green is phosphoryl oxygen, and light green is methoxy oxygen).
37
38
39
40
41 Oxygen vacancies often promote water dissociation to form a hydroxyl group at the vacancy
42
43
44 site and transferring the second proton to a neighboring surface oxygen atom. For many oxide
45
46 surfaces, this process is facile even at water vapor pressure below 10-8 Torr.31 The O 1s spectrum
47
48 of the pristine MoO3 foil in Figure 3a shows the metal oxide component and a small binding
49
50
51 energy shoulder about 1 eV higher than the bulk oxide, consistent with the presence of surface
52
53 hydroxyl groups.31,71,72 This hydroxyl peak has been assigned previously on MoO3 samples
54
55 prepared under UHV conditions.35,66,68,69,73 The other possibility is that this peak corresponds to
56
57
58 the oxygen deficient Magnéli phases. While there is little literature on the O 1s binding energy of
59
60
ACS Paragon Plus Environment
11
 The Journal of Physical Chemistry Page 12 of 54

1
2
3
Magnéli phases of MoO3, two factors discount this assignment. (i) The O 1s binding energy of
4
5
6 the Mo6+ and Mo4+ oxides are within a few tenths of an eV of each other,35,69 and it is likely that
7
8 the intermediate oxide state would be similar. (ii) A study of the growth of MoO3 clusters on an
9
10
11
Au substrate initially shows a significant amount of Mo5+ in the Mo 3d spectrum but no
12
13 corresponding peak in the O 1s spectrum, indicating there is no significant chemical shift
14
15 between Mo6+ and Mo5+ oxides. Therefore, we assign this high binding energy shoulder to
16
17
18 hydroxyl groups, formed from a background pressure of water.
19
20 Since hydroxyl groups are inherent to our sample preparation method, we wanted to test the
21
22 interaction of DMMP on a MoO3 surface with a greater coverage of hydroxyl groups. On
23
24
25 polycrystalline MoO3, water adsorption and hydroxyl group formation is known to be quite low
26
27 and is higher for sub-stoichiometric films.74 Therefore, another sample with an increased number
28
29
of oxygen vacancies was deliberately prepared by light sputtering in order to promote a higher
30
31
32 hydroxylation coverage. The amount of the Mo5+ component has increased, as shown in the Mo
33
34 3d spectrum in Figure 2c. Sputtering has been shown to create oxygen vacancies on MoO3,
35
36
37 which migrate to form extended defect structures.28,75 To encourage oxygen vacancy
38
39 hydroxylation, both the untreated and sputtered surfaces were exposed to 10 L of H2O.
40
41 Correspondingly, the O 1s spectrum of this surface in Figure 3c shows an increase in the OH
42
43
44 coverage. A third component in the O 1s spectrum of the hydroxylated MoO3 in Figure 3c is
45
46 consistent with molecular water absorbed on the surface,31,76,77 with the peak being positioned
47
48 2.1 eV higher than the bulk peak. Though water on a room temperature metal oxide surface
49
50
51 under UHV is uncommon,78 it is known to occur on RuO279 and MgO,77 for example. The
52
53 molecular water further suggests a larger concentration of surface hydroxyl groups since
54
55
56
hydroxyl groups are known to induce water adsorption to surfaces.80 Thus, this sputtered surface
57
58
59
60
ACS Paragon Plus Environment
12
Page 13 of 54 The Journal of Physical Chemistry

1
2
3
contains extended, oxygen deficient phases, hydroxyl groups, and molecularly adsorbed water.
4
5
6 To distinguish between the two surfaces, the MoO3 as prepared is referred to as pristine MoO3,
7
8 and the sputtered surface that contains more hydroxyl groups is labelled OH-MoO3.
9
10
11
The substrates were gradually exposed to increasing DMMP pressures at room temperature
12
13 while collecting PE spectra. The two peaks in the C 1s PE spectra (cf. Figure 4a,b) show that
14
15 DMMP is adsorbing on both surfaces, with the peak at 284.5 eV assigned to the methyl group,
16
17
18 and the higher binding energy peak at 286.3 eV assigned to the methoxy groups.36 Figure 4c,d
19
20 shows the P 2p spectra of DMMP absorbed onto both surfaces, with the 2p3/2 and the 2p1/2 spin-
21
22 orbit components resolved. The lack of additional peaks in the C 1s and P 2p spectra suggest that
23
24
25 DMMP could be adsorbing intact. The methoxy to methyl C 1s peak area ratio is close to the
26
27 expected value of 2 for the intact molecule adsorbed on pristine MoO3, as shown in Figure 5a.
28
29
However, this ratio is much lower for DMMP adsorption on OH-MoO3, especially at lower
30
31
32 pressures. This deviation indicates that either a volatile product containing carbon is being
33
34 formed, preferentially from the methoxy groups, or that methoxy carbons are being converted to
35
36
37 methyl carbons.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
13
 The Journal of Physical Chemistry Page 14 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 Figure 4. C 1s PE spectra of DMMP on (a) pristine MoO3 and (b) OH-MoO3 and P 2p PE
20
21 spectra of DMMP adsorbed on (c) pristine MoO3 and (d) OH-MoO3 under a DMMP pressure of
22
23
24 1 x 10-4 Torr. The spectra have been normalized to their maximum intensity. Spectra of the clean
25
26 surfaces are given in the Supporting Information (Figures S5a and S5b).
27
28
29
30
31
32 Peaks corresponding to the methoxy groups and phosphoryl oxygen atoms are evident in O 1s
33
34 spectra of pristine MoO3 under a DMMP pressure of 1 x 10-4 Torr in Figure 3b. In Figure 3d, the
35
36 overlap of the hydroxyl peak with the phosporyl peak complicates the analysis of the OH-MoO3
37
38
39 spectrum. The spectral fit in Figure 3d was obtained by restricting the bulk oxide (yellow curve)
40
41 and hydroxyl (blue curve) peaks to the same parameters (FWHM, BE) as in the pristine MoO3
42
43
spectrum in Figure 3a. The ratio of the DMMP peaks (green curves) were set equal to the C 1s
44
45
46 area ratio in Figure 5a, which assumes that even if some methoxy groups are leaving, the
47
48 remaining surface fragment will have a methyl and a P=O group. As discussed in the theoretical
49
50
51 section, this assumption is reasonable. More details of the spectral fits are given in the
52
53 Supporting Information. Hydroxyl groups often react with surface species, especially molecules
54
55 prone to hydrolysis such as DMMP.
56
57
58
59
60
ACS Paragon Plus Environment
14
Page 15 of 54 The Journal of Physical Chemistry

1
2
3
The O 1s spectra can be used to calculate an approximate coverage of DMMP on the pristine
4
5
6 MoO3 and OH-MoO3 surfaces, using the method described in Ref. 72, which assumes adsorption
7
8 in a 2D geometry. While we expect that in the case of pristine MoO3 and OH-MoO3 the surface
9
10
11
will be rough due to the polycrystalline nature of the starting material, the method still allows for
12
13 an estimate of the coverages that can be compared between the two surfaces. Details of the
14
15 calculations are reported in the Supporting Information, and the results are plotted in Figure 5b.
16
17
18 The coverage calculated from both the methoxy and phosphoryl O peaks for the pristine MoO3
19
20 agree with each other. Since there is some methoxy loss on the OH-MoO3 sample, the coverages
21
22 calculated with this peak are slightly lower than those calculated from the P=O O 1s peak. The
23
24
25 results in Figure 5b show that for all pressures DMMP absorption is higher on OH-MoO3 than on
26
27 pristine MoO3.
28
29
The Mo 3d spectrum of OH-MoO3 does not change upon DMMP adsorption (cf. Figure 2d).
30
31
32 However, the Mo 3d spectra of the pristine surface show broader peaks at higher DMMP
33
34 pressures and a higher abundance of Mo5+, as seen in Figure 2b. The increase in the Mo5+ peak
35
36
37 intensity is monotonic with DMMP pressure and was found not to be photon-induced, as
38
39 evidenced by faster acquisition times and reproducibility in a second data set (see Figure S2).
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
15
 The Journal of Physical Chemistry Page 16 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 5. (a) Pressure dependence of the C 1s peak area ratio (methoxy:methyl) of DMMP
27
28
29 adsorbed on pristine MoO3 (black circles) and OH-MoO3 (grey squares). (b) DMMP coverage on
30
31 OH-MoO3 calculated from the methoxy (blue triangles) and P=O (orange diamonds) O 1s peaks
32
33 and on pristine MoO3 calculated from the methoxy (black circles)P=O (red squares) O 1s peaks.
34
35
36 The error bars originate from the peak fitting of the spectra.
37
38
39 DFT Calculations
40
41 The adsorption of DMMP on MoO3 (010), the most stable surface, was modeled with
42
43
44 DFT calculations to aid in the interpretation of the APXPS results of the polycrystalline film.
45
46 Figure 6a and b shows that while DMMP will adsorb on the pristine MoO3 (010) surface (with
47
48 the adsorption energy of 20 kcal/mol), oxygen vacancies, hydroxyl groups, and a combination of
49
50
51 these defects will enhance the chemisorption. Figure 6 also indicates that the most energetically
52
53 favored interaction between DMMP and a non-hydroxylated MoO3 (010) surface is through the
54
55
56
phosphoryl oxygen (configuration C1, Figure 6a). In configuration C1, a new weak bond
57
58
59
60
ACS Paragon Plus Environment
16
Page 17 of 54 The Journal of Physical Chemistry

1
2
3
between the phosphoryl oxygen and the molybdenum atom is formed. The adsorption of the
4
5
6 DMMP on the metal atom is accompanied by the cleavage of one of the bonds between Mo atom
7
8 and surface oxygen. As a result, the oxidation state of the Mo remains 6+, which agrees with the
9
10
11
experiments. In the second most favorable configuration C2 (Figure 6b), the DMMP molecule is
12
13 weakly adsorbed (through physisorption) on the MoO3 (010) surface, and no chemical bond is
14
15 formed between the molecule and the oxide surface. The calculated adsorption energy for the
16
17
18 configuration C2 is 20.1 kcal/mol. The presence of terminal oxygen vacancies noticeably
19
20 enhance DMMP adsorption, as the calculated adsorption energy of DMMP atop the under-
21
22 coordinated Mo atoms in the configuration C3 (Figure 6c) is 36.7 kcal/mol.
23
24
25 For DMMP on OH-MoO3, there are two favorable configurations with adsorption energies of -
26
27 35.8 kcal/mol for configuration C4, shown in Figure 6d, and 33.4 kcal/mol for configuration C5,
28
29
shown in Figure 6e. In adsorption configurations C4 and C5, DMMP interacts either with one or
30
31
32 two surface OH groups through its phosphoryl oxygen. We note, that in the configuration C4
33
34 (Figure 6d), a proton migrates from the surface to the phosphoryl oxygen of DMMP leading to a
35
36
37 formation of H+-DMMP cation. In addition to configurations C4 and C5, we also simulated
38
39 adsorption of DMMP on the hydroxylated MoO3 (010) surface through its methoxy oxygen
40
41 atoms (configuration C6, Figure 6f). The calculated adsorption energy for configuration C6 is
42
43
44 24.6 kcal/mol.
45
46 Further, our simulations show that adsorption of DMMP on MoO3 (010) surface containing
47
48 both an oxygen vacancy and hydroxyl groups in close proximity to each other is even more
49
50
51 energetically advantageous than the adsorption on the surface with oxygen vacancies only or
52
53 hydroxyl groups only. Thus, the calculated energies of the DMMP adsorption energy on the
54
55
56
surface with an oxygen vacancy and one hydroxyl group is 42.3 kcal/mol (configuration C7,
57
58
59
60
ACS Paragon Plus Environment
17
 The Journal of Physical Chemistry Page 18 of 54

1
2
3
Figure 6g), and on the same surface with two hydroxyl groups, the adsorption energy 47.6
4
5
6 kcal/mol (configuration C7, Figure 6g). Overall, the trend in all the adsorption energies in Figure
7
8 6 agree with larger DMMP coverage of OH-MoO3 plotted in Figure 5b, i.e. all adsorption
9
10
11
geometries of DMMP on oxygen deficient and/or hydroxylated surfaces MoO3 (010) have more
12
13 favorable adsorption energies than the lowest energy geometry on the pristine MoO3.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 Figure 6. The structure of (a), (b) the energetically favorable configurations of DMMP adsorbed
40
41 on pristine MoO3, labelled C1 and C2, (c) preferred configurations of DMMP on the MoO3
42
43
44 surface with a terminal oxygen vacancy, labelled C3, (d), (e) and (f) three preferred
45
46 configurations of DMMP on OH-MoO3, labelled C4, C5 and C6, respectively, (g) and (h)
47
48 possible configurations of DMMP on MoO3 surface with an oxygen vacancy and one and two
49
50
51 hydroxyl groups, labelled C7 and C8, respectively. O is red, P is green, H is beige, C is brown,
52
53 and Mo is lilac.
54
55
56
57
58
59
60
ACS Paragon Plus Environment
18
Page 19 of 54 The Journal of Physical Chemistry

1
2
3
To investigate possible reasons for the loss of methoxy carbon on the OH-MoO3 seen in
4
5
6 Figure 5a, reasonable decomposition products and pathways were explored. We first
7
8 considered DMMP decomposition in the gas phase by modeling several plausible
9
10
11
decomposition channels, including the gas phase bond dissociation of the P-CH3 bond (Eq. 1),
12
13 the O-CH3 bond (Eq. 2), and the P-OCH3 bond (Eq. 3) and the elimination of methanol by
14
15 intramolecular hydrogen transfer (Eq. 4).
16
17
18 OP(CH3)(OCH3)2 → OP●(OCH3)2 + ●CH3 (1)
19
20 OP(CH3)(OCH3)2 → OP(CH3)(OCH3)(O●) + ●CH3 (2)
21
22 OP(CH3)(OCH3)2 → OP●(CH3)(OCH3) + ●OCH3 (3)
23
24
25 OP(CH3)(OCH3)2 → OP(CH2)(OCH3) + HOCH3 (4)
26
27 Table 1 shows that the elimination of methanol is the most energetically favorable path among
28
29
these processes. Requiring 69.9 kcal/mol, this process has a 12.4 kcal/mol, 19.2 kcal/mol and
30
31
32 31.3 kcal/mol lower activation barrier than the O-CH3, P-CH3, and P-OCH3 bond dissociation
33
34 energies, respectively. These findings are consistent with a recent study, reporting methyl and
35
36
37 methanol among other products of DMMP pyrolysis in the gas phase.81 The same study also
38
39 found that formaldehyde loss occurred via a two-step process81 that involves complex
40
41 isomerization of the molecule accompanied by an intramolecular hydrogen transfer and a
42
43
44 rotation of functional groups (see Figure S8). Due to its intramolecular nature, the presence of
45
46 surface defects hardly affects the activation barrier of the rate limiting reaction step (~65
47
48 kcal/mol) of formaldehyde elimination from DMMP proceeding through intramolecular
49
50
51 hydrogen shift from methoxy carbon to phosphoryl oxygen. Hence we did not consider this
52
53 mechanism as a potential decomposition channel on the ideal and hydroxylated MoO3 surface
54
55
56
despite the similar activation energy to the methanol elimination (see Figure S8).
57
58
59
60
ACS Paragon Plus Environment
19
 The Journal of Physical Chemistry Page 20 of 54

1
2
3
Table 1. Calculated decomposition Energies of DMMP (kcal/mol)
4
5
6 DMMP adsorbed on
7 Surface with
gas Surface with
8 pristine oxygen vacancy
Reaction phase oxygen OH-MoO3 surface
9 surface and hydroxyl
vacancy
10 group
11 Conf. Conf. Conf. Conf. Conf. Conf.
Conf. C6
12 C1 C2 C3 C4 C5 C8
13 P-CH3 (Eq 1) 89.1 97.8 90.4 - 69.7 - - -
14 O-CH3 (Eq 2) 82.3 96.5 85.8 59.7 30.1b 29.8b 38.5b 35.6b
15 P-OCH3 (Eq 3) 101.2 107.0 92.8 99.4 102.3 - - -
16 74.7c
76.2c 76.1c 82.9c
17 CH3OH loss (Eq 4) 69.9a 78.7a 69.1a 78.a 28.9e
44.9d 45.2d 40.1d
18 41.6f
19 a
Collected decomposition energies correspond to the highest activation barrier of the two-
20 stage reaction of methanol formation (see Figure 7).
21
22 b
Energies of the O-CH3 bond dissociation were obtained assuming the absence of the activation
23
24
barrier of the reverse reaction (Figure S9)
25 c
26 Decomposition energies were calculated as overall reaction energies of methanol elimination
27 (see path II2 in Figure 8a).d The activation barrier calculated for the methanol loss reaction via
28 surface-to-molecule hydrogen transfer shown in Figure 8a.
29
e
30 The activation barrier calculated for the methanol loss via surface-to-molecule hydrogen
31 transfer shown in Figure 8b, reaction pathway IV.
32
33 f
The activation barrier calculated for the methanol loss via the stepwise reaction pathway V of
34 Figure 8b.
35
36
37
38
39 The energies for the reactions in Equations 1 through 4 were calculated for DMMP on a
40
41 pristine MoO3 surface, starting from the configurations C1 and C2 in Figure 6a and b; the
42
43
44
dissociation energies are provided in Table 1. Methanol loss remains the most favorable path for
45
46 both configurations, with the calculated activation barrier of 78.7 and 69.1 kcal/mol (Figure 7),
47
48 respectively. Dissociation of the P-CH3, O-CH3 and P-OCH3 bonds in the lowest energy
49
50
51 configuration C1 requires energy that is 7-10 kcal/mol higher than that in the gas phase.
52
53 Dissociation of the P-CH3 and O-CH3 bonds in the configuration C2 requires about the same
54
55 amount of energy as in the gas phase, whereas the energy necessary for the P-OCH3 bond
56
57
58 cleavage on the pristine MoO3 surface is 8.4 kcal/mol lower than that in the gas phase. The
59
60
ACS Paragon Plus Environment
20
Page 21 of 54 The Journal of Physical Chemistry

1
2
3
relatively low adsorption energy (20-23 kcal/mol, Figure 6a and b) coupled with high activation
4
5
6 energies required for the DMMP decomposition on the stoichiometric MoO3 (010) surface
7
8 suggest that pristine MoO3 will likely demonstrate rather low catalytic activity towards DMMP
9
10
11
decomposition. This agrees well with XPS data, which shows low DMMP coverage on pristine
12
13 MoO3 and the nearly 2:1 methoxy to methyl C 1s peak area ratio (cf. Figure 5a).
14
15 Dissociation of PO-CH3 bond in DMMP adsorbed on the oxygen deficient MoO3 (010) surface
16
17
18 (configuration C3, Figure 6c) requires energy, which is 22.6 kcal/mol lower than in the gas phase
19
20 and 26.1 kcal/mol lower than on the pristine surface (Table 1). Furthermore, the relatively low
21
22 energy of the PO-CH3 bond cleavage (59.7 kcal/mol, Table 1) suggests that this mechanism
23
24
25 dominates over the P-OCH3 bond dissociation and methanol formation in the DMMP
26
27 decomposition on the vacancy-rich MoO3 surface due to the higher activation energies required
28
29
for the latter two mechanisms, 99.4 kcal/mol and 78.4 kcal/mol (Table 1), respectively.
30
31
32 Next, the decomposition of DMMP on OH-MoO3, the methanol loss through an intramolecular
33
34 hydrogen transfer (path II, Figure 8a), starting from the most favorable configuration, C4 in
35
36
37 Figure 6d, was modeled. The calculated overall activation barrier, 76.2 kcal/mol, is similar to
38
39 those obtained for the gas phase and the ideal stoichiometric surface. The surface hydroxyl
40
41 groups provide new channels for the methanol elimination from DMMP via surface-to-molecule
42
43
44 hydrogen transfer, which is illustrated as path III in Figure 8a. This reaction pathway requires
45
46 only 44.9 kcal/mol, which is ~30 kcal/mol lower than the methanol elimination through an
47
48 intramolecular hydrogen transfer. The significant drop in the activation energy for methanol loss
49
50
51 on the OH-MoO3 surface agrees with the XPS measurements (cf. Figure 5a). The DMMP
52
53 adsorption on OH-MoO3 also reduces the dissociation energy of the P-CH3 and O-CH3 bonds as
54
55
56
compared to the gas phase and the pristine MoO3 surface (cf. Table 1). The calculated energy of
57
58
59
60
ACS Paragon Plus Environment
21
 The Journal of Physical Chemistry Page 22 of 54

1
2
3
the P-CH3 bond cleavage on the hydroxylated surface, although ~20 kcal/mol lower than that on
4
5
6 the ideal surface, remains too high to consider this mechanism as a viable channel for the DMMP
7
8 decomposition on OH-MoO3. The O-CH3 bond cleavage of DMMP on OH-MoO3 is
9
10
11
accompanied by migration of one of the surface hydrogens to the oxygen with unpaired electron
12
13 (Figure S9) and has the lowest energy (30.1 kcal/mol, Table 1) among all studied pathways; this
14
15 could explain the decrease in methoxy C 1s signal. However, the bond dissociation energy in
16
17
18 Table 1 may be underestimated as it was obtained as the difference between total energies of
19
20 products and reactants assuming that reverse reaction has no activation barrier (Figure S9) A
21
22 large activation barrier would preclude the removal of the methyl from the surface, in which case
23
24
25 a change of the C 1s peak would not be expected. To refine modeling of this decomposition
26
27 pathway, more comprehensive information on the decomposition products is required, but is not
28
29
attainable at this point. Hence, the possibility of this pathway occurring cannot be ruled out for
30
31
32 the time being.
33
34 We now turn our attention to the mechanisms of methanol loss starting from the less
35
36
37 energetically favorable, but still viable, configurations C5 and C6 on OH-MoO3. To study
38
39 decomposition of DMMP from the configuration C5, we modeled three possible mechanisms
40
41 (Figure S10) to seek a possible explanation for the experimentally observed change in methoxy
42
43
44 C 1s signal. Elimination of the methanol via intramolecular hydrogen transfer requires slightly
45
46 higher energy (76.1 kcal/mol, Table 1) than in the gas phase and on the pristine surface. The
47
48 calculated activation barrier of the alternative mechanism via surface-to-molecule hydrogen
49
50
51 transfer (45.2 kcal/mol, Table 2) is 30.9 kcal/mol lower. Similarly to the configuration C4, the
52
53 presence of surface hydroxyl groups causes a reduction of the energy of O-CH3 bond breaking
54
55
56
by 52.5 and 56 kcal/mol as compared to gas phase and pristine surface, respectively (Table 1).
57
58
59
60
ACS Paragon Plus Environment
22
Page 23 of 54 The Journal of Physical Chemistry

1
2
3
For the configuration C6, we also considered three possible mechanisms of the methanol
4
5
6 formation. The first mechanism proceeds via a surface-to-molecule hydrogen transfer, shown as
7
8 path IV in Figure 8b, and requires only 28.9 kcal/mol, which is ~17 kcal/mol lower than starting
9
10
11
from the more energetically favorable configurations C4 and C5 on OH-MoO3. The second
12
13 possible channel of the methanol formation (path V, Figure 8b) proceeds via splitting of the O-
14
15 CH3 bond in DMMP and subsequent adsorption of the methyl group onto a surface hydroxyl
16
17
18 (path V1, Figure 8b). The calculated activation barrier of this reaction step is 41.6 kcal/mol. The
19
20 methanol molecule is weakly bound to the surface, requiring only 18.3 kcal/mol for its
21
22 desorption (path V2, Figure 8b). The obtained energy required for methanol elimination via
23
24
25 intramolecular hydrogen transfer (74.7 kcal/mol, Table 1) is significantly higher than the
26
27 activation barriers of the former two channels. Furthermore, we examined the mechanism of the
28
29
loss of a methoxy group formed through the O-CH3 bond cleavage and subsequent adsorption of
30
31
32 the methyl group onto one of the terminal surface oxygen atoms (path VI1, Figure 8b). The
33
34 calculated activation barrier for this reaction step is 40.2 kcal/mol. The large energy of the
35
36
37 following step (69.3 kcal/mol, path VI2, Figure 8b) precludes this mechanism from contributing
38
39 to the methanol loss observed in the experimental data. If this pathway does occur with the
40
41 surface methoxy remaining on the surface, the experimental binding energies would be
42
43
44 indistinguishable from those of intact DMMP adsorbed on the surface. The reaction of CH3● loss
45
46 through O-CH3 bond breaking in DMMP adsorbed in the configuration C6 requires 38.5
47
48 kcal/mol (Table 1).
49
50
51 Finally, we studied the methanol elimination reaction from the most preferred adsorption
52
53 configuration C8 (Figure 6h), in which DMMP adsorbs on the MoO3 (010) surface containing
54
55
56
both oxygen vacancies and hydroxyl groups (Figure 9). The first step involves splitting the O-
57
58
59
60
ACS Paragon Plus Environment
23
 The Journal of Physical Chemistry Page 24 of 54

1
2
3
CH3 bond in the DMMP molecule, transfer of the CH3 group to a surface hydroxyl group, and
4
5
6 surface-to-molecule hydrogen transfer (path VII1, Figure 9). This reaction step requires 40.1
7
8 kcal/mol with the calculated reaction energy of -16.2 kca/mol. The energy of the subsequent
9
10
11
methanol desorption from the surface (path VII2, Figure 9) is 28.2 kcal/mol. The methanol loss
12
13 via intramolecular hydrogen transfer requires 82.9 kcal/mol. The calculated energy of the O-CH3
14
15 bond dissociation reaction is 35.6 kcal/mol.
16
17
18
19
Overall, our modeling clearly demonstrates that the surface hydroxyl groups and oxygen
20
21 vacancies enhance the adsorption of DMMP on the MoO3 surface. A significant drop in the
22
23 calculated activation energies for both methanol loss and O-CH3 bond cleavage is consistent with
24
25
26 the reduction of the methoxy to methyl C 1s peak area ratio for the hydroxylated surface (Figure
27
28 5).
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
24
Page 25 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 7. Schematic representation of the methanol formation via intramolecular hydrogen
30 transfer on the pristine MoO3 (010) surface (configurations C1 and C2) and surface containing
31 oxygen vacancies (configuration C3). Obtained energies are shown in kcal/mol.
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
25
 The Journal of Physical Chemistry Page 26 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42 Figure 8. Possible mechanisms for the methanol elimination from DMMP adsorbed on the
43
44 hydroxylated (010) MoO3 surface starting from a) configuration C3 and b) configuration C5. All
45
46 obtained energies are shown in kcal/mol. The black numbers are in reference to the lowest
47
48
49 energy structure labelled 0.0 kcal/mol. The blue numbers over arrows are activation energies.
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
26
Page 27 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27 Figure 9. Mechanisms for the methanol elimination from DMMP adsorbed on the (010) MoO3
28
29 surface containing oxygen vacancy and hydroxyl groups (configuration C8, Figure 8h). All
30
31 obtained energies are shown in kcal/mol. The black numbers are in reference to the lowest
32
33
34 energy structure labelled 0.0 kcal/mol. The blue numbers over arrows are activation energies.
35
36
37
38
39 Discussion
40
41
42 The low DMMP coverage on MoO3 shown in Figure 5b suggests that adsorption is not
43
44 strongly occurring, especially on pristine MoO3 at lower pressures, which is also seen in the DFT
45
46 calculations that show a weak interaction of the phosphoryl oxygen with the terminal surface O
47
48
49 atoms (cf. Figure 6a). The most favored organophosphonate interaction with other metal oxides,
50
51 such as TiO2, Al2O3, Fe2O3, and Y2O3, is proposed to be between the phosphoryl group and
52
53
under-coordinated Lewis acid metal sites.4,6,7,9-12,16,18,82 The lack of Lewis acid metal sites
54
55
56 inherent to the structure of the stoichiometric MoO3(010) surface leads to weak adsorption of
57
58
59
60
ACS Paragon Plus Environment
27
 The Journal of Physical Chemistry Page 28 of 54

1
2
3
DMMP. Oxygen vacancies (that generate Mo5+ sites) would provide the ideal binding sites for
4
5
6 DMMP, and, indeed, the calculations here predict favorable energies for DMMP adsorption on
7
8 oxygen vacancies. However, although calculations show that DMMP binds to both Mo5+ and
9
10
11
hydroxyl groups (both present on OH-MoO3) more strongly than to pristine MoO3, it is difficult
12
13 to assess the contribution of the Mo5+ to DMMP adsorption in these experiments. . While it is
14
15 possible that DMMP is binding to Mo5+ sites, we suggest that hydroxyl groups are also playing a
16
17
18 role in the increased adsorption of DMMP. Hydroxyl groups are expected to be the next
19
20 favorable site for DMMP adsorption onto other metal oxides.4-7,9-11,15,16 Typically, hydroxyl
21
22 groups form at oxygen vacancies of metal oxides by the dissociation of water and oxidation of
23
24
25 the metal;71 it is likely that, after sputtering, this process is occurring on some of the Mo5+ sites
26
27 with water from the background pressure. Our calculated adsorption energies in Figure 6 support
28
29
hydroxyl groups as favorable adsorption sites. Figure 6e shows another possible adsorption
30
31
32 structure on a hydroxylated surface that has not been considered previously, where hydrogen
33
34 bonds between both methoxy oxygen atoms and two surface hydroxyl groups result in a
35
36
37 configuration with an adsorption energy that is 0.5 eV higher than the lowest energy structure.
38
39 The binding energy difference between the functional groups in the C 1s and O 1s spectra as
40
41 well as the presence of only a single P species in the P 2p spectrum demonstrate strong
42
43
44 similarities between adsorbed and gas phase DMMP.36 However, the C 1s spectra indicate
45
46 methoxy group loss, especially for OH-MoO3. DFT calculations (Table 1 and Figures 8 and 9)
47
48 show the plausible adsorption configurations of DMMP on OH-MoO3 have energetically
49
50
51 favorable pathways to form methanol, while decomposition on oxygen vacancy sites was found
52
53 to not be favorable unless there is a neighboring hydroxyl group. Evidence of the remaining
54
55
56
molecular fragments is not seen in any of the spectra. While the binding energies of the possible
57
58
59
60
ACS Paragon Plus Environment
28
Page 29 of 54 The Journal of Physical Chemistry

1
2
3
dissociation products are not known, it is likely that they are similar to those of the intact DMMP
4
5
6 molecule.3 Our calculations show that methanol should easily desorb from the surface.
7
8 TPD24,26,83-85 and IR spectroscopy83,85-87 studies indicate that methanol both adsorbs weakly intact
9
10
11
and forms surface methoxy groups, but the presence of hydroxyl groups on the starting surface is
12
13 not addressed. However, while the adsorption is known to be affected by the crystalline phase
14
15 and defects,24 methanol does not readily adsorb onto stiochiometric MoO3(010).83 The binding
16
17
18 energies of both adsorbed methanol and surface-bound methoxy groups are indistinguishable
19
20 from adsorbed DMMP; therefore, these possibilities cannot be excluded. The C 1s
21
22 methoxy:methyl peak area ratio is clearly less than the predicted 2:1 for the intact molecule,
23
24
25 indicating at least some methanol is desorbing. The adsorbed water on the OH-MoO3 surface
26
27 could influence the amount of methanol that is formed. DMMP is known to hydrolyze in water
28
29
to form methanol and methyl phosphonic acid.1 This reaction combined with desorbing methanol
30
31
32 product may be the reason the C 1s peak area ratio is the lowest at low DMMP pressure for OH-
33
34 MoO3. In addition to methanol loss, a cleavage of the O-CH3 bond requires noticeably lower
35
36
37 energy on the OH-MoO3 surface (Table 1); however, our experimental results neither confirm
38
39 nor exclude this possibility.
40
41 These results complement recent work by Tang and coworkers on DMMP adsorption on size-
42
43
44 selected (MoO3)3 clusters.88 They found that DMMP adsorbs onto the clusters in a similar
45
46 configuration and calculated enhanced adsorption onto hydroxylated clusters. Similarly, they
47
48 found that that stoichiometric clusters did not decompose DMMP but calculations suggest
49
50
51 hydroxyl groups lower the energy required for decomposition by four-fold. The O-CH3 bond
52
53 dissociation energy was also found to be one of the lowest energy processes among Equations 1
54
55
56
through 4.
57
58
59
60
ACS Paragon Plus Environment
29
 The Journal of Physical Chemistry Page 30 of 54

1
2
3
The amount of Mo5+ on the MoO3 sample increases with DMMP pressure, as is shown in the
4
5
6 Mo 3d spectra in Figure 2b. This increase is gradual with the pressure and not a result of damage
7
8 by the photon beam. Simple alcohols have been shown to reduce MoO389 and have been used in
9
10
11
reduced MoO3 sample preparations.90 MoO3 by itself, on other metal oxide supports, and mixed
12
13 in a ferric molybdenate is used in methanol oxidation catalysis where Mo is the oxidant. Chung
14
15 and coworkers found that Mo was not reduced until 110 ºC.86 However if methanol is being
16
17
18 formed and then further reacting to form surface methoxy groups, perhaps Mo reduction is
19
20 occurring in this system. The Mo 3d spectrum of OH-MoO3 does not change with DMMP
21
22 pressure; the larger extent of Mo reduction before DMMP exposure is likely the reason.
23
24
25 Conclusions
26
27 APXPS studies were conducted in parallel with theoretical calculations to determine the room
28
29
temperature adsorption behavior and decomposition of DMMP on a polycrystalline MoO3
30
31
32 surface, including elucidation of the role of oxygen vacancies, hydroxyl groups and molecularly
33
34 adsorbed water. APXPS data combined with DFT calculations have revealed that DMMP binds
35
36
37 weakly to MoO3. Introducing oxygen vacancies on the MoO3 results in an increase in hydroxyl
38
39 groups and adsorbed molecular water on the surface. DFT calculations and the surface coverages
40
41 calculated from the APXPS data suggest that the oxygen vacancies and hydroxyl groups enhance
42
43
44 the adsorption of DMMP, but it is difficult to separate the contributions of the two features in the
45
46 APXPS data. With the C 1s spectra indicating a loss of methoxy carbon, it is highly likely that
47
48 methanol is forming and desorbing from the surface, facilitated by the increase of hydroxyl
49
50
51 groups and molecular water on the surface. Methanol may also adsorb on the MoO3 surface,
52
53 reducing the Mo6+. These findings are consistent with previous proposals of under-coordinated
54
55
56
metal atoms and hydroxyl groups being favored surface adsorption sites for DMMP on metal
57
58
59
60
ACS Paragon Plus Environment
30
Page 31 of 54 The Journal of Physical Chemistry

1
2
3
oxides. The results of this study build on results from a parallel study of DMMP adsorption on
4
5
6 (MoO3)3 clusters,88 showing how the adsorption behavior scales from the nanoregime to bulk
7
8 oxide films.
9
10
11
12
13 Supporting Information. Additional spectra addressing photon-induced sample damage,
14
15 spectral fits, DMMP coverage calculations, and additional theoretical calculations. This material
16
17
18 is available free of charge via the Internet at http://pubs.acs.org.
19
20
21 Corresponding Author
22
23
24 * E-mail mkukla@nsf.gov, Tel 703-292-4940 (M.M.K.).
25
26
27 *E-mail hbluhm@lbl.gov, Tel 510-486-5431 (H.B.).
28
29
30
Notes
31
32
33 The authors declare no competing financial interest.
34
35 ACKNOWLEDGMENT
36
37
38
This work was funded by the Department of Defense (Grant HDTRA11510005). R.T. and
39
40 M.M.K acknowledge support from NSF XSEDE resources (Grant DMR-130077) and DOE
41
42 NERSC resources (Contract DE-AC02-05CH11231) and computational resources at the
43
44
45 Maryland Advanced Research Computing Center (MARCC). MMK is grateful to the Office of
46
47 the Director of National Science Foundation for support under the Independent Research and
48
49 Development program. Any appearance of findings, conclusions, or recommendations expressed
50
51
52 in this material are those of the authors and do not necessarily reflect the views of NSF. The
53
54 Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy
55
56
Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
57
58
59
60
ACS Paragon Plus Environment
31
 The Journal of Physical Chemistry Page 32 of 54

1
2
3
4
5
6
7 ABBREVIATIONS
8
9
10 APXPS, ambient pressure X-ray photoelectron spectroscopy; DMMP, dimethyl
11
12
13 methylphosphonate; L, Langmuir; PE, photoemission; XPS, X-ray photoelectron spectroscopy.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
32
Page 33 of 54 The Journal of Physical Chemistry

1
2
3
Table of Contents Graphic
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
33
 The Journal of Physical Chemistry Page 34 of 54

1
2
3
REFERENCES
4
5
6
7 1
8 Jang, Y. J.; Kim, K.; Tsay, O. G.; Atwood, D. A.; Churchill, D. G. Destruction and Detection
9
10 of Chemical Warfare Agents. Chem. Rev. 2015, 115, PR1-PR76.
11
12
13 2
Aurian-Blajeni, B.; Boucher, M. M. Interaction of Dimethyl Methylphosphonate with Metal
14
15
16 Oxides. Langmuir 1989, 5, 170-174.
17
18
3
19 Zhou, J.; Varazo, K.; Reddic, J. E.; Myrick, M. L.; Chen, D. A. Decomposition of Dimethyl
20
21 Methylphosphonate on TiO2(110): Principal Component Analysis Applied to X-ray
22
23
24 Photoelectron Spectroscopy. Anal. Chim. Acta 2003, 496, 289-300.
25
26
4
27 Rusu, C. N.; Yates Jr., J. T. Adsorption and Decomposition of Dimethyl Methylphosphonate
28
29 on TiO2. J. Phys. Chem. B 2000, 104, 12292-12298.
30
31
32 5
33
Kim, C. S.; Lad, R. J.; Tripp, C. P. Interaction of Organophosphorous Compounds with TiO2
34
35 and WO3 Surfaces Probed by Vibrational Spectroscopy. Sens. Actuators, B 2001, 76, 442-448.
36
37
38 6
Panayotov, D. A.; Morris, J. R. Uptake of a Chemical Warfare Agent Simulant (DMMP) on
39
40
41
TiO2: Reactive Adsorption and Active Site Poisoning. Langmuir 2009, 25, 3652-3658.
42
43 7
44 Panayotov, D. A.; Morris, J. R. Thermal Decomposition of a Chemical Warfare Agent
45
46 Simulant (DMMP) on TiO2: Adsorbate Reactions with Lattice Oxygen as Studied by Infrared
47
48
49
Spectroscopy. J. Phys. Chem. C 2009, 113, 15684-15691.
50
51 8
52 Moss, J. A.; Szczepankiewicz, S. H.; Park, E.; Hoffman, M. R. Adsorption and
53
54 Photodegradation of Dimethyl Methylphosphonate Vapor at TiO2 Surfaces. J. Phys. Chem. B
55
56
2005, 109, 19779-19785.
57
58
59
60
ACS Paragon Plus Environment
34
Page 35 of 54 The Journal of Physical Chemistry

1
2
3
4
5 9
6
Mitchell, M. B.; Sheinker, V. N.; Mintz, E. A. Adsorption and Decomposition of Dimethyl
7
8 Methylphosphonate on Metal Oxides. J. Phys. Chem. B 1997, 101, 11192-11203.
9
10
10
11 Templeton, M. K.; Weinberg, W. H. Adsorption and Decomposition of Dimethyl
12
13
Methylphosphonate on an Aluminum Oxide Surface. J. Am. Chem. Soc. 1985, 107, 97-108.
14
15
16 11
17 Templeton, M. K.; Weinberg, W. H. Decomposition of Phosphonate Esters Adsorbed on
18
19 Aluminum Oxide. J. Am. Chem. Soc. 1985, 107, 774-779.
20
21
22 12
Henderson, M. A.; Jin, T.; White, J. M. A TPD/AES Study of the Interaction of Dimethyl
23
24
25 Methylphosphonate with α-Fe2O3 and SiO2. J. Phys. Chem. 1986, 90, 4607-4611.
26
27
13
28 Tesfai, T. M.; Scheinker, V. N.; Mitchell, M. B. Decomposition of Dimethyl
29
30 Methylphosphonate (DMMP) on Alumina-Supported Iron Oxide. J. Phys. Chem. B 1998, 102,
31
32
33 7299-7302.
34
35
14
36 Chen, D. A.; Ratliff, J. S.; Hu, X.; Gordon, W. O.; Senanayake, S. D.; Mullins, D. R.
37
38 Dimethyl Methylphosphonate Decomposition on Fully Oxidized and Partially Reduced Ceria
39
40
41 Thin Films. Surf. Sci. 2010, 604, 574-587.
42
43
15
44 Kanan, S. M.; Tripp, C. P. An Infrared Study of Adsorbed Organophosphonates on Silica: A
45
46 Prefiltering Strategy of the Detection of Nerve Agents on Metal Oxide Sensors. Langmuir 2001,
47
48
49 17, 2213-2218.
50
51 16
52 Gordon, W. O.; Tissue, B. M.; Morris, J. R. Adsorption and Decomposition of Dimethyl
53
54 Methylphosphonate on Y2O3 Nanoparticles. J. Phys. Chem. C 2007, 111, 3233-3240.
55
56
57
58
59
60
ACS Paragon Plus Environment
35
 The Journal of Physical Chemistry Page 36 of 54

1
2
3
4
5 17
6
Cao, L.; Segal, S. R.; Suib, S. L.; Tang, X.; Satyapal, S. Thermocatalytic Oxidation of
7
8 Dimthyl Methylphosphonate on Supported Metal Oxides. J. Catal. 2000, 194, 61-70.
9
10
18
11 Mitchell, M. B.; Sheinker, V. N.; Cox Jr., W. W.; Gatimu, E. N.; Tesfamichael, A. B. The
12
13
Room Temperature Decomposition Mechanism of Dimthyl Methylphosphonate (DMMP) on
14
15
16 Alumina-Supported Cerium Oxide – Participation of Nano-Sized Cerium Oxide Domains. J.
17
18 Phys. Chem. B 2004, 108, 1634-1645.
19
20
21 19
Hermann, K.; Witko, M. Theory of physical and chemical behavior of transition metal
22
23
24 oxides: vanadium and molybdenum oxides. In Oxide Surfaces, Woodruff, D. P., Ed., Elsevier
25
26 Science B. V.: Amersterdam, 2001; Ch. 4, pp 136-198.
27
28
29 20
Coquet, R.; Willock, D. J. The (010) Surface of α-MoO3, a DFT + U Study. Phys. Chem.
30
31
32 Chem. Phys. 2005, 7, 3819-3828.
33
34
21
35 Yin, X.; Han, H.; Miyamoto, A. Structure and Adsorption Properties of MoO3: Insights from
36
37
Periodic Density Functional Calculations. J. Mol. Model 2001, 7, 207-215.
38
39
40 22
41 Yan, Z.; Zuo, Z. Lv, X.; Li, Z.; Li, Z.; Huang, W. Adsorption of NO on MoO3(010) Surface
42
43 with Different Location of Terminal Oxygen Vacancy Defects: A Density Functional Theory
44
45 Study. Appl. Surf. Sci. 2012, 258, 3163-3167.
46
47
48 23
49 Hermann, K.; Witko, M.; Michalak, A. Density Functional Studies of the Electronic
50
51 Structure and Adsorption at Molybdenum Oxide Surfaces. Catal. Today 1999, 50, 567-577.
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
36
Page 37 of 54 The Journal of Physical Chemistry

1
2
3
4
5 24
6
O’Brien, M. G.; Beale, A. M.; Jacques, S. D. M.; Buslaps, T.; Honkimaki, V.; Weckhuysen,
7
8 B. M. On the Active Oxygen in Bulk MoO3 During the Anaerobic Dehydrogenation of
9
10 Methanol. J. Phys. Chem. C 2009, 113, 4890-4897.
11
12
13 25
Rousseau, R.; Dixon, D. A.; Kay, B. D.; Dohnálek, Z. Dehydration, Dehydrogenation, and
14
15
16 Condensation of Alcohols on Supported Oxide Catalysts Based on Cyclic (WO3)3 and (MoO3)3
17
18 Clusters. Chem. Soc. Rev. 2014, 43, 7664-7680.
19
20
21 26
Bowker, M.; Carley, A. F.; House, M. Contrasting the Behaviour of MoO3 and MoO2 for the
22
23
24 Oxidation of Methanol. Catal. Lett. 2008, 120, 34-39.
25
26
27
27 Ressler, T.; Wienold, J.; Jentoft, R. E.; Girgsdies, F. Evolution of Defects in the Bulk
28
29 Structure of MoO3 During the Catalytic Oxidation of Propene. Eur. J. Inorg. Chem. 2003, 301-
30
31
32 312.
33
34
28
35 Tokarz-Sobieraj, R.; Hermann, K.; Witko, M.; Blume, A.; Mestl, G.; Schlögl, R. Properties
36
37 of Oxygen Sites at the MoO3(010) Surface: Density Functional Theory Cluster Studies and
38
39
40 Photoemission Experiments. Surf. Sci. 2001, 489, 107-125.
41
42 29
43 Starr, D. E.; Liu, Z.; Hävecker, M.; Knop-Gericke, A.; Bluhm, H. Investigation of
44
45 Solid/vapor Interfaces Using Ambient Pressure X-ray Photoelectron Spectroscopy. Chem. Soc.
46
47
48 Rev. 2013, 42, 5833-5857.
49
50 30
51 Head, A. R.; Bluhm, H. Ambient Pressure X-Ray Photoelectron Spectroscopy in “Reference
52
53 Module in Chemistry, Molecular Sciences and Chemical Engineering”, Jan Reedijk (Ed.),
54
55
56 Elsevier B. V. (2016)
57
58
59
60
ACS Paragon Plus Environment
37
 The Journal of Physical Chemistry Page 38 of 54

1
2
3
4
5
6
31 Bluhm, H. Photoelectron Spectroscopy of Surfaces under Humid Conditions. J. Electron
7
8 Spectrosc. Rel. Phenom. 2010, 177, 71-84.
9
10
11 32 Ogletree, D. F.; Bluhm, H.; Hebenstreit, E. D.; Salmeron, M. Photoelectron Spectroscopy
12
13
under Ambient Pressure and Temperature Conditions. Nuc. Instrum. Methods Phys. Res., Sect. A,
14
15
16 2009, 601, 151-160.
17
18
19 33 Bluhm, H.; Andersson, K.; Araki, T.; Benzerara, K.; Brown, G. E.; Dynes, J. J.; Ghosal, S.;
20
21 Gilles, M. K.; Hansen, H.-Ch.; Hemminger, J. C.; et al. Soft X-ray Microscopy and Spectroscopy
22
23
24 at the Molecular Environmental Science Beamline at the Advanced Light Source. J. Electron
25
26 Spectrosc. Related Phenom. 2006, 150, 86-104.
27
28
29 34
Ogletree, D. F.; Bluhm, H., Lebedev, G.; Fadley, C. S.; Hussain, Z.; Salmeron, M. A
30
31
32 Differentially Pumped Electrostatic Lens System for Photoemission Studies in the Millibar
33
34 Range. Rev. Sci. Instrum. 2002, 73, 3872-3877.
35
36
37 35
Scanlon, D. O.; Watson, G. W.; Payne, D. J.; Atkinson, G. R.; Egdell, R. G.; Law, D. S. L.
38
39
40 Theoretical and Experimental Study of the Electronic Structure of MoO3 and MoO2. J. Phys.
41
42 Chem. C 2010, 114, 4636-4645.
43
44
45 36
Head, A. R.; Tsyshevsky, R.; Trotochaud, L.; Eichhorn, B.; Kuklja, M. M.; Bluhm, H.
46
47
48 Electron Spectroscopy and Computational Studies of Dimethyl Methylphosphonate. J. Phys.
49
50 Chem. A 2016, 120, 1985-1991.
51
52
53 37
Smith, R. L.; Rohrer, G. S. An Atomic Force Microscopy Study of the Morphological
54
55
56 Evolution of the MoO3 (010) Surface during Reduction Reactions. J. Catal. 1996, 163, 12–17.
57
58
59
60
ACS Paragon Plus Environment
38
Page 39 of 54 The Journal of Physical Chemistry

1
2
3
4
5 38
6
Firment, L. E.; Ferretti, A. Stoichiometric and Oxygen Deficient MoO3(010) Surfaces. Surf.
7
8 Sci. 1983, 129, 155-176.
9
10
39
11 Hermann, K.; Michalak, A.; Witko, M. Cluster Model Studies on Oxygen Sites at the (010)
12
13
Surfaces of V2O5 and MoO3. Catal. Today 1996, 32, 321-327.
14
15
16 40
17 Lei, Y.-H.; Chen, Z.-X. DFT+U Studies of Properties of MoO3 and Hydrogen Adsorption on
18
19 MoO3(010). J. Phys. Chem. C 2012, 116, 25757 −25764.
20
21
22 41
Chen, M.; Waghmare, U. V.; Friend, C. M.; Kaxiras, E. A Density Functional Study of
23
24
25 Clean and Hydrogen-Covered α-MoO3 (010): Electronic Structure and Surface Relaxation. J.
26
27 Chem. Phys. 1998, 109, 6854-6860.
28
29
30 42
Negishi, H.; Negishi, S.; Kuroiwa, Y.; Sato, N.; Aoyagi, S. Anisotropic Thermal Expansion
31
32
33 of Layered MoO3 Crystals. Phys. Rev. B. 2004, 69, 064111.
34
35 43
36 Sitepu, H. Texture and Structural Refinement Using Neutron Diffraction Data from
37
38 Molybdite (MoO3) and Calcite (CaCO3) Powders and a Ni-rich Ni50.7Ti49.30 Alloy. Powder Diffr.
39
40
41
2009, 24, 315–326.
42
43 44
44 Hohenberg, P.; Kohn, W. Inhomogeneous Electron Gas. Phys. Rev. 1964, 136, B864-B871.
45
46
45
47 Kohn, W.; Sham, L. J. Self-Consistent Equations Including Exchange and Correlation
48
49 Effects. Phys. Rev. 1965, 140, A1133-A1138.
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
39
 The Journal of Physical Chemistry Page 40 of 54

1
2
3
4
5 46
6
Ding, H.; Ray, K. G.; Ozolins, V.; Asta, M. Structural and Vibrational Properties of α-MoO3
7
8 from van der Waals Corrected Density Functional Theory Calculations. Phys. Rev. B 2012, 85,
9
10 012104.
11
12
13 47
Inzani, K.; Grande, T.; Vullum-Bruer, F.; Selbach, S. M. A van der Waals Density
14
15
16 Functional Study of MoO3 and Its Oxygen Vacancies. J. Phys. Chem. C 2016, 120, 8959−8968.
17
18
48
19 Dion, M.; Rydberg, H.; Schröder, E.; Langreth, D. C.; Lundqvist, B. I. Van der Waals
20
21 Density Functional for General Geometries. Phys. Rev. Lett. 2004, 92, 246401.
22
23
24 49
25 Thonhauser, T.; Cooper, V. R.; Li, S.; Puzder, A.; Hyldgaard, P.; Langreth, D. C. Van der
26
27 Waals Density Functional: Self-Consistent Potential and the Nature of the van der Waals Bond.
28
29 Phys. Rev. B 2007, 76, 125112.
30
31
32 50
33
Román-Pérez, G.; Soler, J. M. Efficient Implementation of a van der Waals Density
34
35 Functional: Application to Double-Wall Carbon Nanotubes. Phys. Rev. Lett. 2009, 103, 096102.
36
37
38 51
Klimeš, J.; Bowler, D. R.; Michaelides, A. Chemical Accuracy for the Van der Waals
39
40
41
Density Functional. J. Phys.: Cond. Matt. 2009, 22, 022201-1-5.
42
43 52
44 Klimeš, J.; Bowler, D. R.; Michaelides. Van der Waals Density Functionals Applied to
45
46 Solids. Phys. Rev. B 2011, 83, 195131.
47
48
49 53
Kresse, G.; Futhmüller, J. Efficiency of ab-initio Energy Calculations for Metals and
50
51
52 Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6, 15-50.
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
40
Page 41 of 54 The Journal of Physical Chemistry

1
2
3
4
5 54
6
Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for ab initio Total-Energy
7
8 Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169-11186.
9
10
55
11 Kresse, G.; Hafner, J. Ab initio Molecular Dynamics for Liquid Metals. Phys. Rev. B 1993,
12
13
47, 558-561.
14
15
16 56
17 Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17953-17979.
18
19 57
20 Lei, Y.-H.; Chen, Z.-X. A Theoretical Study of Stability and Vacancy Replenishing of
21
22 MoO3(010) Surfaces in Oxygen Atmosphere. Appl. Surf. Sci. 2016, 361, 107-103.
23
24
25 58
Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple.
26
27
28 Phys. Rev. Lett. 1996, 77, 3865-3868.
29
30
59
31 Dudarev, S. L.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-Energy-Loss
32
33
Spectra and the Structural Stability of Nickel Oxide: An LSDA+U Study, Phys. Rev. B 1998, 57,
34
35
36 1505−1509.
37
38
60
39 Becke, A. D. Density-Functional Thermochemistry. 3. The Role of Exact Exchange. J.
40
41 Chem. Phys. 1993, 98, 5648-5652.
42
43
44 61
45 Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti Correlation-Energy
46
47 Formula into a Functional of the Electron- Density Phys. Rev. B, 1988, 37, 785-789.
48
49
50 62
M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G.
51
52
53
Scalmani, V. Barone, B. Mennucci, G.A. Petersson, et al Gaussian 09, in, Gaussian, Inc.,
54
55 Wallingford, CT, USA, 2009.
56
57
58
59
60
ACS Paragon Plus Environment
41
 The Journal of Physical Chemistry Page 42 of 54

1
2
3
4
5 63
6
Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image Nudged Elastic Band
7
8 Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901-
9
10 9904.
11
12
13 64
Hratchian, H. P.; Schlegel, H. B. Accurate Reaction Paths using a Hessian Based Predictor-
14
15
16 Corrector Integrator. J. Chem. Phys. 2004, 120, 9918-9924.
17
18
65
19 Hratchian, H. P.; Schlegel, H. B. Using Hessian Updating to Increase the Efficiency of a
20
21 Hessian Based Predictor-corrector Reaction Path Following Method. J. Chem. Theory Comput.
22
23
24 2005, 1, 61-69.
25
26
66
27 Ramana, C. V.; Atuchin, V. V.; Kesler, V. G.; Kochubey, V. A.; Pokrovsky, L. D.;
28
29 Shutthanandan V.; Becker, U.; Ewing, R. C. Growth and Surface Characterization of Sputter-
30
31
32 Deposited Molybdenum Oxide Thin Films. Appl. Surf. Sci. 2007, 253, 5368-5374.
33
34
67
35 Song, Z.; Cai, T.; Chang, Z.; Liu, G.; Rodriguez, J. A.; Hrbek, J. Molecular Level Study of
36
37 the Formation and the Spread of MoO3 on Au(111) by Scanning Tunneling Microscopy and X-
38
39
40 ray Photoelectron Spectroscopy. J. Am. Chem. Soc. 2003, 125, 8059-8066.
41
42 68
43 Choi, J.-G.; Thompson, L. T. XPS Study of As-prepared and Reduced Molybdenum Oxides.
44
45 Appl. Surf. Sci. 1996, 93, 143-149.
46
47
48 69
49
Brox, B.; Olefjord, I. ESCA Studies of MoO2 and MoO3. Surf. Inter. Anal. 1988, 13, 3-6.
50
51 70
52 Liao, X.; Jeong, A. R.; Wilks, R. G.; Wiesner, S.; Rusu, M.; Bär, M. X-ray Irradiation
53
54 Induced Effects on the Chemical and Electronic Properties of MoO3 Thin Films. J. Electron
55
56
Spectrosc. Rel. Phenom., 2016, 212, 50-55.
57
58
59
60
ACS Paragon Plus Environment
42
Page 43 of 54 The Journal of Physical Chemistry

1
2
3
4
5 71
6
Ketteler, G.; Yamamoto, S.; Bluhm, H.; Andersson, K.; Starr, D. E.; Ogletree, D. F.;
7
8 Ogasawara, H.; Nilsson, A.; Salmeron, M. The Nature of Water Nucleation Sites on TiO2(110)
9
10 Surfaces Revealed by Ambient Pressure X-ray Photoelectron Spectroscopy. J. Phys. Chem. C
11
12
13 2007, 111, 8278-8282.
14
15 72
16 Newberg, J. T.; Starr, D. E.; Yamamoto, S.; Kaya, S.; Kendelewicz, T.; Mysak, E. R.;
17
18 Porsgaard, S.; Salmeron, M. B.; Brown Jr., G. E.; Nilsson, A. et al. Formation of Hydroxyl and
19
20
21
Water Layers on MgO Films Studied with Ambient Pressure XPS. Surf. Sci. 2011, 605, 89-94.
22
23 73
24 Zeng, H. C.; Xie, F.; Wong, K. C.; Mitchell, K. A. R. Insertion and Removal of Protons in
25
26 Single-Crystal Orthorhomic Molybdenum Trioxide under H2S/H2 and O2/N2. Chem. Mater.
27
28
29
2002, 14, 1788-1796.
30
31 74
32 Sian, T. S.; Reddy, G. B. Effect of Stoichiometry and microstructure on Hydrolysis in MoO3
33
34 Films. Chem. Phys. Lett. 2006, 418, 170-173.
35
36
37 75
Dufour, L. C.; Bertrand, O.; Floquet, N. Chemical Reactivity of (010)MoO3: A Structural
38
39
40 Study of the MoO2 Formation in Molecular Hydrogen. Surf. Sci. 1984, 147, 396-412.
41
42 76
43 Ketteler, G.; Yamamoto, S.; Bluhm, H.; Andersson, K.; Starr, D. E.; Ogletree, D. F.;
44
45 Ogasawara, H.; Nilsson, A.; Salmeron, M. The Nature of Water Nucleation Sites on TiO2(110)
46
47
48 Surfaces Revealed by Ambient Pressure X-ray Photoelectron Spectroscopy. J. Phys. Chem. C
49
50 2007, 111, 8278-8282.
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
43
 The Journal of Physical Chemistry Page 44 of 54

1
2
3
4
5 77
6
Newberg, J. T.; Starr, D. E.; Yamamoto, S.; Kaya, S.; Kendelewicz, T.; Mysak, E. R.;
7
8 Porsgaard, S.; Salmeron, M. B.; Brown Jr., G. E.; Nilsson, A. et al. Formation of Hydroxyl and
9
10 Water Layers on MgO Films Studied with Ambient Pressure XPS. Surf. Sci. 2011, 605, 89-94.
11
12
13 78
Henderson, M. A. The Interaction of Water with Solid Surfaces: Fundamental Aspects
14
15
16 Revisited. Surf. Sci. Rep. 2002, 46, 1-308.
17
18
79
19 Kim, Y. J.; Gao, Y.; Chambers, S. A. Core-level X-ray Photoelectron Spectra and X-ray
20
21 Photoelectron Diffraction of RuO2(110) Grown by Molecular Beam Epitaxy. Appl. Surf. Sci.
22
23
24 1997, 120, 250-260.
25
26
80
27 Yamamoto, S.; Andersson, K.; Bluhm, H.; Ketteler, G.; Starr, D. E.; Schiros, T.; Ogasawara,
28
29 H.; Pettersson, L. G. M.; Salmeron, M.; Nilsson A. Hydroxyl-Induced Wetting of Metals by
30
31
32 Water at Near-Ambient Conditions. J. Phys. Chem. C 2007, 111, 7848-7850.
33
34
81
35 Liang, S.; Hemberger, P.; Neisius, N. M.; Bodi, A.; Grützmacher, H.; Levalois-
36
37 Grützmacher, J.; Gaan, S. Elucidating the Thermal Decomposition of Dimethyl
38
39
40 Methylphosphonate by Vacuum Ultraviolet (VUV) Photoionization: Pathways to the PO
41
42 Radical, a Key Species in Flame-Retardant Mechanisms. Chem. Eur. J. 2015, 21, 1073 – 1080.
43
44
45 82
Kuiper, A. E. T.; van Bokhoven, J. J. G. M.; Medema, J. The Role of Heterogeneity in the
46
47
48 Kinetics of a Surface Reaction I. Infrared Characterization of the Adsorption Structures of
49
50 Organophosphonates and Their Decomposition. J. Catal. 1976, 43, 154-167.
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
44
Page 45 of 54 The Journal of Physical Chemistry

1
2
3
4
5 83
6
Chowdhry, U.; Ferretti, A.; Firment, L. E.; Machiels, C. J.; Ohuchi, F.; Sleight, A. W.;
7
8 Staley, R. H. Mechanism and Surface Structural Effects in Methanol Oxidation over Molybdates.
9
10 Appl. Surf. Sci. 1984, 19, 360-372.
11
12
13 84
Farneth, W. E.; Ohuchi, F.; Staley, R. H.; Chowdhry, U.; Sleight, A. W. Mechanism of
14
15
16 Partial Oxidation of Methanol over MoO3 as Studied by Temperature-Programmed Desorption.
17
18 J. Phys. Chem. 1985, 89, 2493-2497.
19
20
21 85
Street, S. C.; Goodman, D. W. Chemical and Spectroscopic Surface Science Investigation of
22
23
24 MoO3 and MoO3/Al2O3 Ultrathin Films. J. Vac. Sci. Technol. A 1997, 15, 1717-1723.
25
26
86
27 Chung, J. S.; Miranda, R.; Bennett, C. O. Study of Methanol and Water Chemisorbed on
28
29 Molybdenum Oxide. J. Chem. Soc., Faraday Trans. 1 1985, 81, 19-36.
30
31
32 87
33
Groff, R. P. An Infrared Study of Methanol and Ammonia Adsorption on Molybdenum
34
35 Trioxide. J. Catal. 1984, 86, 215-218.
36
37
38 88
Tang, X.; Hicks, Z.; Wang, L.; Ganteför, G.; Fairbrother, D. H.; Bowen, K. H.; Tsyshevsky,
39
40
41
R.; Sun, J.; Kuklja, M. M. Decomposition of Dimethyl Methylphosphonate by Size-Selected
42
43 (MoO3)3 Clusters in preparation
44
45
89
46 Farneth, W. E.; Staley, R. H.; Sleight, A. W. Stoichiometry and Structural Effects in Alcohol
47
48
Chemisorption/Temperature-Programmed Desorption on MoO3. J. Am. Chem. Soc. 1986¸108,
49
50
51 2327-2332.
52
53
90
54 Li, T.; Beidaghi, M.; Xiao, X.; Huang, L.; Hu, Z.; Sun, W.; Chen, X.; Gogotsi, Y.; Zhou, J.
55
56
Ethanol Reduced Molybdenum Trioxide for Li-ion Capacitors. Nano Energy 2016, 26 100-107.
57
58
59
60
ACS Paragon Plus Environment
45
 The Journal of Physical Chemistry Page 46 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 1. (a) Chemical nerve agent sarin and (b) dimethyl methylphosphonate (DMMP), (c) structure of gas
30 phase DMMP, (d) slab model of the MoO3(010) surface, and (e) surface structure of MoO3, including an
31 oxygen vacancy and a hydroxylated oxygen vacancy. P atom is green, C atoms are brown, H atoms are
32 beige, O atoms are red, and Mo atoms are lilac.
33 Figure 1
34 152x101mm (150 x 150 DPI)
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 47 of 54 The Journal of Physical Chemistry

. ter
1
2

ils r
3

ta nve
4
5
6
7

de Co
8
9
10
11

or Doc
12
13
14
15
16
17

rm F
e
18
19

fo D
20
21

m veP
22
23
24
25
26
.co cti
27
28
DF f a

29
30
31
32
eP n o

33
34
35
36
tiv tio

37
38
39
ac di

40
41
w. n E

42
43
44
45
ww tio

46
Figure 2. Mo 3d PE spectra of (a) pristine MoO3, (b) pristine MoO3 (black) and under a DMMP pressure of 1
47
x 10-2 Torr (blue), (c) OH-MoO3 as prepared, and (d) OH-MoO3 as prepared (black) and under a DMMP
48 pressure of 1 x 10-4 Torr (blue). Voigt functions of the spectral fits (red lines) are shaded with the dark grey
sit lua

49 representing Mo5+ and the lighter grey representing Mo6+. All spectra are normalized to their maximum
50 intensity.
51 Figure 2
52 53x88mm (300 x 300 DPI)
Vi va

53
54
55
E

56
57
58
59
60
ACS Paragon Plus Environment
 The Journal of Physical Chemistry Page 48 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Figure 3. O 1s PE spectra of (a) pristine MoO3, (b) pristine MoO3 under a DMMP pressure of 1 x 10-4 Torr, (c)
47
OH-MoO3 as prepared, and (d) OH-MoO3 under a DMMP pressure of 1 x 10-4 Torr. The spectra are
48 normalized to their maximum intensity. The black dots are the data points, the lilac line is the fit of the Voigt
49 functions (yellow is bulk MoO3, blue is hydroxyl groups, light blue is molecular water, dark green is
50 phosphoryl oxygen, and light green is methoxy oxygen).
51 Figure 3
52 54x88mm (300 x 300 DPI)
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 49 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42 Figure 4. C 1s PE spectra of DMMP on (a) pristine MoO3 and (b) OH-MoO3 and P 2p PE spectra of DMMP
43 adsorbed on (c) pristine MoO3 and (d) OH-MoO3 under a DMMP pressure of 1 x 10-4 Torr. The spectra have
44 been normalized to their maximum intensity. Spectra of the clean surfaces are given in the Supporting
45 Information (Figures S5a and S5b).
46 Figure 4
47 55x59mm (300 x 300 DPI)
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 The Journal of Physical Chemistry Page 50 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Figure 5. (a) Pressure dependence of the C 1s peak area ratio (methoxy:methyl) of DMMP adsorbed on
47
pristine MoO3 (black circles) and OH-MoO3 (grey squares). (b) DMMP coverage on OH-MoO3 calculated from
48 the methoxy (blue triangles) and P=O (orange diamonds) O 1s peaks and on pristine MoO3 calculated from
49 the methoxy (black circles)P=O (red squares) O 1s peaks. The error bars originate from the peak fitting of
50 the spectra.
51 Figure 5
52 55x88mm (300 x 300 DPI)
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 51 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 6. The structure of (a), (b) the energetically favorable configurations of DMMP adsorbed on pristine
27 MoO3, labelled C1 and C2, (c) preferred configurations of DMMP on the MoO3 surface with a terminal oxygen
28 vacancy, labelled C3, (d), (e) and (f) three preferred configurations of DMMP on OH-MoO3, labelled C4, C5
29 and C6, respectively, (g) and (h) possible configurations of DMMP on MoO3 surface with an oxygen vacancy
30 and one and two hydroxyl groups, labelled C7 and C8, respectively. O is red, P is green, H is beige, C is
31 brown, and Mo is lilac.
Figure 6
32
337x192mm (150 x 150 DPI)
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 The Journal of Physical Chemistry Page 52 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Figure 7. Schematic representation of the methanol formation via intramolecular hydrogen transfer on the
47
pristine MoO3(010) surface (configurations C1 and C2) and surface containing oxygen vacancies
48 (configuration C3). Obtained energies are shown in kcal/mol.
49
50 163x202mm (150 x 150 DPI)
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 53 of 54 The Journal of Physical Chemistry

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38 Figure 8. Possible mechanisms for the methanol elimination from DMMP adsorbed on the hydroxylated
39 (010) MoO3 surface starting from a) configuration C3 and b) configuration C5. All obtained energies are
40 shown in kcal/mol. The black numbers are in reference to the lowest energy structure labelled 0.0 kcal/mol.
41 The blue numbers over arrows are activation energies.
42 Figure 8
43 327x308mm (150 x 150 DPI)
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 The Journal of Physical Chemistry Page 54 of 54

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Figure 9. Mechanisms for the methanol elimination from DMMP adsorbed on the (010) MoO3 surface
38 containing oxygen vacancy and hydroxyl groups (configuration C8, Figure 8h). All obtained energies are
39 shown in kcal/mol. The black numbers are in reference to the lowest energy structure labelled 0.0 kcal/mol.
40 The blue numbers over arrows are activation energies.
41 Figure 9
42 177x162mm (150 x 150 DPI)
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment