Oxidation of carbon monoxide cocatalyzed by palladium(0) and the H 5PV2Mo10O40 polyoxometalate probed by electron paramagnetic resonance and aerobic catalysis

Article abstract of DOI:10.1021/ic900868t

The H₅PV₂Mo₁₀O₄₀ polyoxometalate and Pd/AI₂O₃ were used as co-catalysts under anaerobic conditions for the activation and oxidation of CO to CO ₂ by an electron transfer-oxygen transfer mechanism. Upon anaerobic reduction of H₅PV₂Mo₁₀O₄₀ with CO in the presence of Pd(0) two paramagnetic species were observed and characterized by continuous wave electron paramagnetic resonance (CW-EPR) and hyperfine sublevel correlation (HYSCORE) spectroscopic measurements. Major species I (65-70%) is assigned to a species resembling a vanadyl cation that is supported on the polyoxometalate and showed a bonding interaction with ¹³CO. Minor species II (30-35%) is attributed to a reduced species where the vanadium(IV) atom is incorporated in the polyoxometalate framework but slightly distanced from the phosphate core. Under aerobic conditions, CO/O₂, a nucleophilic oxidant was formed as elucidated by oxidation of thianthrene oxide as a probe substrate. Oxidation reactions performed on terminal alkenes such as 1-octene yielded a complicated mixture of products that was, however, clearly a result of alkene epoxidation followed by subsequent reactions of the intermediate epoxide. The significant competing reaction was a hydro-carbonylation reaction that yielded a ～1:1 mixture of linear/branched carboxylic acids.

Full text of DOI:10.1021/ic900868t

Inorg. Chem. 2009, 48, 7947–7952 7947
DOI: 10.1021/ic900868t
Oxidation of Carbon Monoxide Cocatalyzed by Palladium(0) and the H5PV2Mo10O40
Polyoxometalate Probed by Electron Paramagnetic Resonance and Aerobic
Catalysis
Hila Goldberg,† Ilia Kaminker,‡ Daniella Goldfarb,‡ and Ronny Neumann*,†
†Department of Organic Chemistry and ‡Department of Chemical Physics and Weizmann Institute of Science,
Rehovot, Israel 76100
Received May 5, 2009
The H5PV2Mo10O40 polyoxometalate and Pd/Al2O3 were used as co-catalysts under anaerobic conditions for the
activation and oxidation of CO to CO2 by an electron transfer-oxygen transfer mechanism. Upon anaerobic reduction of
H5PV2Mo10O40 with CO in the presence of Pd(0) two paramagnetic species were observed and characterized by
continuous wave electron paramagnetic resonance (CW-EPR) and hyperfine sublevel correlation (HYSCORE)
spectroscopic measurements. Major species I (65-70%) is assigned to a species resembling a vanadyl cation that is
supported on the polyoxometalate and showed a bonding interaction with 13CO. Minor species II (30-35%) is
attributed to a reduced species where the vanadium(IV) atom is incorporated in the polyoxometalate framework but
slightly distanced from the phosphate core. Under aerobic conditions, CO/O2, a nucleophilic oxidant was formed as
elucidated by oxidation of thianthrene oxide as a probe substrate. Oxidation reactions performed on terminal alkenes
such as 1-octene yielded a complicated mixture of products that was, however, clearly a result of alkene epoxidation
followed by subsequent reactions of the intermediate epoxide. The significant competing reaction was a hydro-
carbonylation reaction that yielded a ∼1:1 mixture of linear/branched carboxylic acids.
Introduction
often proceed via electron transfer oxidation of the organic
substrate,6 which in some cases includes also oxygen transfer
from the polyoxometalate to the organic substrate via an
electron transfer-oxygen transfer.7 In such reactions, there is
In the 1970-1980s, research demonstrated Pd(II) catalyzed
oxidation reactions where phosphovanadomolybdates of the
Keggin structure, [PVxMo12-xO40](3þx)-, especially x = 2,
were used as co-catalysts to reoxidize the primary palladium
catalyst ultimately using molecular oxygen as terminal oxi-
dant.1 Over the years further reports of such aerobic catalytic
oxidation reactions catalyzed by Pd/H5PV2Mo10O40 have
appeared,2 together with many additional H5PV2Mo10O40
catalyzed aerobic transformations and syntheses.3 Recently,
H5PV2Mo10O40 has been used in metallorganic-polyoxome-
talate hybrid catalysts for aerobic oxidation of methane,4 and
formation of methyl ketones from alkenes with nitrous oxide.5
On the mechanistic side some understanding has been devel-
oped suggesting that H5PV2Mo10O40 catalyzed reactions
first
a electron transfer from the substrate to the
H5PV2Mo10O40 followed by oxygen transfer from the reduced
polyoxometalate to the substrate, Scheme 1. For example, it
has been shown in the case of anthracene oxidation that an
outer sphere electron transfer (anthracene to poly-
oxometalate) leads to formation of an organic cation radical
that should form cation radical-polyoxometalate ion pairs,8
although such intermediate ion pairs have not been directly
observed.
In parallel, there have been several recent reports wherein
the combination of CO/O2 has been used in the presence of
H5PV2Mo10O40 and Pd to catalyze the hydroxylation or
carboxylation of arenes to the corresponding phenols,9 and
the dicarboxylation of cylic alkenes.10 While Pd catalyzed
*To whom correspondence should be addressed. E-mail: ronny.neumann@
weizmann.ac.il.
(1) (a) Kozhevnikov, I. V.; Matveev, K. I. Appl. Catal. 1983, 5, 135–150.
(b) Grate, J. H.; Hamm, D. R.; Mahajan, S. In Catalysis of Organic Reactions;
Kosak, J. R., Johnson, T. A., Eds.; Marcel Dekker: New York, 1984; Chapter 16.
(2) Some key references: (a) Yokota, T.; Tani, M.; Sakaguchi, S.; Ishii, Y.
J. Am. Chem. Soc. 2003, 123, 1476–1477. (b) Yokota, T.; Sakaguchi, S.; Ishii, I.
Adv. Synth. Catal. 2002, 344, 849–854. (c) Yamada, T.; Sakaguchi, S.; Ishii, Y. J.
Org. Chem. 2005, 70, 5471–5474.
(3) (a) Hill, C. L.; Prosser-McCartha, C. M. Coord. Chem. Rev. 1995, 143,
407–455. (b) Kozhevnikov, I. V. Chem. Rev. 1998, 98, 171–198. (c) Neumann,
R.; Khenkin, A. M. Chem. Commun. 2006, 2529–2538.
(6) Neumann, R.; Levin, M. J. Am. Chem. Soc. 1992, 114, 7278–7286.
(7) (a) Khenkin, A. M.; Neumann, R. Angew. Chem., Int. Ed. 2000, 39,
4088–4090. (b) Khenkin, A. M.; Weiner, L.; Wang, Y.; Neumann, R. J. Am.
Chem. Soc. 2001, 123, 8531–8542. (c) Khenkin, A. M.; Neumann, R. J. Am.
Chem. Soc. 2008, 130, 14474–14476.
(8) Weinstock, I. A. Chem. Rev. 1998, 98, 113–170.
(9) (a) Tani, M.; Sakamoto, T.; Mita, S.; Sakaguchi, S.; Ishii, Y. Angew.
Chem., Int. Ed. 2005, 44, 2586–2588. (b) Ohashi, S.; Sakaguchi, S.; Ishii, Y.
Chem. Commun. 2005, 486–487.
(10) Yokota, T.; Sakaguchi, S.; Ishii, Y. J. Org. Chem. 2002, 67, 5005–
5008.
(4) Bar-Nahum, I.; Khenkin, A. M.; Neumann, R. J. Am. Chem. Soc.
2004, 126, 10236–10237.
(5) Ettedgui, J.; Neumann, R. J. Am. Chem. Soc. 2009, 131, 4–5.
r
2009 American Chemical Society
Published on Web 07/20/2009
pubs.acs.org/IC

7948 Inorganic Chemistry, Vol. 48, No. 16, 2009
Goldberg et al.
Scheme 1. General Pathway for an Electron Transfer-Oxygen Trans-
fer Reaction and a Ball and Stick Representation of the Five Isomers of
H5PV2Mo10O40 (black, Mo; green, V; yellow, P; red, O)
Figure 1. EDEPR spectrum(black) ofH5PV2Mo10O40 reacted with CO
in the presence of Pd/Al2O3 and simulations obtained with the para-
meters listed in Table 1. The individual traces of the two species (I-green,
II-blue) and their sum (red) are presented. Inset: expanded spectra at
340.3 mT.
oxidative carbonylation reactions have been often described
in the literature,11 such hydroxylation reactions are rare.
Isotope labeling experiments ruled out the in situ formation
of H2O2 as oxidant by reduction of O2 by CO so that the
nature and identity of the oxidizing species is still a ques-
tion.9a
These intriguing recent CO/O2 hydroxylation reactions
catalyzed by Pd(0)/H5PV2Mo10O40 led us to study (a) the
activation of CO in this catalytic system and (b) the reaction
of activated CO with O2 emphasizing the reactivity of the
oxidizing species obtained. We found by X-band HYSCORE
(hyperfine sublevel correlation) measurements, under anae-
robic conditions, that CO reacted with H5PV2Mo10O40 in the
presence of Pd(0) and formed VIV-CO complexes by electron
transfer. Furthermore, these VIV-CO complexes react via
oxygen transfer from the polyoxometalate to yield CO2.
Under aerobic conditions, these VIV-CO complexes appear
to yield a nucleophilic oxygen donating intermediate that can
also epoxidize terminal alkenes.
no evidence of a redox reaction between H5PV2Mo10O40
and Pd/Al2O3 in a reaction carried out in the absence of CO.
These results allows a conclusion consistent with previous
mechanistic studies on the oxidation of anthracene and
primary alcohols7 that CO activated by Pd(0) leads to an
electron transfer-oxygen transfer reaction wherein the poly-
oxometalate is reduced and CO is oxidized to CO2 via
oxygen transfer from H5PV2Mo10O40 to CO. It should be
noted that Pd/C showed similar reactivity to Pd/Al2O3; Pt/
Al2O3 showed significantly diminished activity while, Au/
Al2O3, Ru/Al2O3 and Rh/Al2O3 were all inactive.
EPR Measurements. The anaerobic reaction of 13CO
activated by Pd(0) with H5PV2Mo10O40 was further
probed by X-band echo detected (ED) EPR. A solution
of 5 mM H5PV2Mo10O40 and 5% Pd/Al2O3 (1 mg) in
8:1:1 AcOH/H2O/glycerol was gently heated (∼35 ꢀC, 10
min) under ∼1 bar 13CO to form the reduced polyoxo-
metalate; the mild conditions used prevented formation
of CO2. The ED-EPR spectrum of Pd activated 13CO with
H5PV2Mo10O40, shown in Figure 1 clearly shows the
formation of two reduced polyoxometalate species. The
EPR parameters and the relative amounts of the two
species were determined from simulations, also shown in
Figure 1; the simulation parameters are given in Table 1.
The EPR parameters of species I are similar to those of a
vanadyl cation, VO2þ from VO(acac)2, in aqueous solu-
tion and are listed in Table 1 as well.
HYSCORE measurements at different observer mag-
netic fields along the EPR powder pattern were carried
out to obtain details on the near environment of the
reduced VIV atom in the two species and more specifically
to observe interaction of 13CO with the polyoxometalate.
Figure 2a shows the spectrum recorded at 291.5 mT.
At this field only species I contributes to the EPR signal,
and the spectrum is “single crystal-like” (Figure 1).
The spectrum shows a ridge with a width of 1.4 MHz
centered at the 13C Larmor frequency, νI(13C)=3.12 MHz.
In addition, intense 1H ridges from the solvent and
H5PV2Mo10O40 are observed. The extent of the 1H ridges
(8.4 MHz) indicates a large dipolar interaction characteristic
Results and Discussion
Anaerobic Oxidation of CO. Dissolution of 5 μmol
H5PV2Mo10O40 (orange color, oxidized form) in 1 mL 9:1
AcOH/H2O as solvent followed by addition of 2 bar CO
in a 25 mL glass pressure tube at 90 ꢀC for 5 h showed
no evidence of oxidation of CO to CO2 as analyzed by
gas chromatography (GC) or formation of reduced
H5PV2Mo10O40 via appearance of the signature heteropoly
blue. Addition of 5% Pd/Al2O3 (1 μmol Pd) to such a
reaction mixture did lead to the formation of a stoichio-
metric amount of CO2 and formation of reduced, blue
H5PV2Mo10O40. This anaerobic oxidation of CO to CO2
over a period of 5 h in the presence of both H5PV2Mo10O40
and 5% Pd/Al2O3 was verified in two ways. First, by
carrying out the CO oxidation with ∼50% 18O labeled
H5PV2Mo10O40 to yield C16O18O as measured by GC-MS
and second by using 13CO and observing the formation of
13CO2 by 13C NMR at 125 ppm. Importantly also there was
(11) (a) Tsuji, J. Palladium Reagents, Innovations in Organic Synthesis;
Wiley: Chichester, 1995.(b) Herrmann, W. A. In Applied Homogeneous Cata-
lysis with Organometallic Complexes; Cornils, B., Herrmann, W. A., Eds.; VCH:
Weinheim, 1996; Volume II, pp 712-732.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7949
Table 1. EPR Parameters for H5PV2Mo10O40 Reacted with 13CO in the Presence
of Pd/Al2O3
species I
species II
VO(acac)2
% abundance
65-70
19.5
7.1
1.944
1.989
30-35
16.8
5.7
1.943
1.986
A , mT
A^, mT
g
18.9
7.1
1.938
1.978
g^
of equatorial water ligands.12-14 Interestingly, no signals
of 31P were observed indicating a long P-V distance. The
HYSCORE spectrum taken at 336.2 mT, Figure 2b, has
contributions from both species, but species I dominates.
Here we observe a clear ridge of a width of 2.8 MHz that
crosses the diagonal at νI(13C)=3.25 MHz. At this field
position the orientation selection is negligible, and the
width of the ridge can be considered as the largest 13C
hyperfine interaction, A . Neglecting any isotropic hy-
perfine interaction, aiso, and using the point dipole ap-
˚
proximation a minimum C-V distance of 2.42 A is
estimated indicating a direct coordination of 13CO to
VIV. The presence of a significant aiso value will of course
change the estimated distance on the one hand, but still
provide additional experimental evidence for direct co-
ordination. The spectrum also shows a peak on the
diagonal at 5.85 MHz, which corresponds to the 31P
Larmor frequency, νI(31P). It has a width of 0.6 MHz,
Figure 2. HYSCORE spectra recorded at (a) 291.5, (b) 336.2, and (c)
340.3 mT. The peaks on the diagonal of the (-, þ) quadrant are noise due
to incomplete removal of unwanted echoes by the phase cycle.
˚
from which a minimum distance P-V of 4.75 A is
estimated. Because of the absence of a 31P signal at 291.5
mT, where only species I contributes to the echo, we assign
the observed 31P signal at this field to species II. The extent
1
of the H ridges (12.3 MHz) indicates a large dipolar
interaction, typical of equatorial water ligands.12-14
At 340.3 mT, Figure 2c, where the contribution of
species II dominates (see Figure 1, inset) signals of 31P
and 1H are detected in the (þ, þ) quadrant. A close look
at the (-, þ) quadrant reveals two ridges parallel to the
diagonal with a width of 2.74 MHz and a frequency
spacing of 1.9 MHz, corresponding to the twice the
Larmor frequency of 95Mo, νI(95Mo), (I = 5/2, natural
abundance 15.9%). The assignment was made to the
95Mo isotope because although the 97Mo isotope has a
similar Larmor frequency, it has a lower natural abun-
dance and a significantly larger quadrapole moment that
should lead to extensive broadening. From the length of
the ridges and assuming an axial hyperfine interaction, A^
∼ 4.5 MHz and A ∼ 9 MHz can be estimated. In addition
a signal on the diagonal appears at 3.79 MHz, which is
close to νI(13C) at this field, but shifted upfield by about
0.1 MHz, suggesting that it could arise from 51V and not
13C. To substantiate this assignment we plotted the posi-
tion of this peak as a function of magnetic field for all
recorded HYSCORE spectra, and the result is shown in
Figure 3, along with the expected νI(13C) and 51V Larmor,
νI(51V), frequencies. At all fields where species I domi-
nates, the frequency of this peak falls on the νI(13C) line;
however, around 340 mT where species II dominates, it
deviated from this line, and at 340.3 mT it falls on the
νI(51V) line.
Figure 3. Plots of the frequency of the HYSCORE diagonal peak in the
(þ, þ) quadrant in the region of 3.5 MHz as a function of the observer
magnetic field. The solid and dashed lines correspond to νI(13C) and
νI(51V), respectively.
From the HYSCORE results we conclude that species I
has a vanadium(IV) site that is similar to a vanadyl cation
13
˚
and has a C atom in its close vicinity (∼2.4 A). The
proximity of the 13C atom to the VIV site is the first strong
spectral evidence for a previously hypothesized bonding
interaction of a substrate, in this case 13CO after activa-
tion with Pd, with H5PV2Mo10O40, Scheme 1. This bond-
ing interaction follows electron transfer from the 13CO to
H5PV2Mo10O40 to yield H5PVIVVVMo10O40 CO. The
3
vanadium(IV) is rather remote from the PO43- core and
the molybdenum atoms (no 31P and 95Mo signals) of the
polyoxometalate and can be considered to be supported
on the polyoxometalate. Furthermore, the observation of
the 13C extended ridges at the g^ region and a small
coupling along g suggests that the 13C atom is in the
equatorial plane. In contrast, in species II the vanadium-
(IV) atom has 31P, 95Mo, and 51V atoms in its close
vicinity, but not a 13C atom. Hence, species II is assigned
as one where the vanadium(IV) is incorporated in the
polyoxometalate. This assignment is also consistent with
(12) van Willigen, H. J. Magn. Reson. 1980, 39, 37–46.
(13) Atherton, N. M.; Shackleton, J. F. Mol. Phys. 1980, 39, 1471–1485.
(14) Dikanov, S. A.; Liboiron, B. D.; Orvig, C. J. Am. Chem. Soc. 2002,
124, 2969–2978.

7950 Inorganic Chemistry, Vol. 48, No. 16, 2009
Goldberg et al.
the smaller A of species II compared to species I. The
contribution of each of the equatorial ligands to the value
of A for 51V correlates approximately inversely with the
electron donor capacity of the ligand, with the most
donating ligands contributing the least to the hyperfine
coupling.15 In H5PVIVVVMo10O40 the equatorial ligands,
that is, the bridging oxo groups of the polyoxometalate,
are high in charge and electrondonating.16 The equatorial
ligands to vanadium(IV) incorporated in the polyoxome-
talate also should exhibit some deviation from planarity,
but it has been shown that the effect of such a distortion
on the A is rather small.17 One should, however, note that
from crystal structures of the oxidized species the V-P
Scheme 2. Oxidation of Thianthrene Oxide
Scheme 3. Benzylic versus Aromatic Oxidation of Toluene with CO/
O2 Catalyzed by Pd/Al2O3 and H5PV2Mo10O40
a
˚
distance is ∼3.5 A whereas the experimentally estimated
˚
distance for the reduced species II is longer (4.75 A). It is
likely that the strong tendency of vanadium(IV) to adapt
a square pyramidal rather than octahedral coordination
leads to its tendency to be removed from the polyoxome-
talate framework, yet the large 95Mo hyperfine couplings
confirms its equatorial position relative to vanadium(IV),
and the weak 31P coupling to vanadium(IV) confirms its
axial location,14 as expected from the Keggin structure.
Moreover, the weak 51V hyperfine coupling indicates that
species II corresponds to the 1,6 and 1,11 isomers of
H5PV2Mo10O40 where the two vanadium atoms are not
nearest neighbors, Scheme 1. Otherwise a hyperfine inter-
action comparable to that observed for 95Mo would have
been expected. Also notable in the context of formation of
two species in a wet solution in this study is a previous
study on the vanadium coordination environment in
hydrated and dehydrated solid H4PVMo11O40 where
water of hydration removed the vanadium(IV) atom from
the polyoxometalate to an interstitial position.18 There
a Values in parentheses are for the reaction without Pd/Al2O3.
Chart 1. Reaction Pathways for the Reaction of 1-octene (R = n-
pentyl) with CO/O2 Catalyzed by Pd/Al2O3 and H5PV2Mo10O40
˚
ENDOR of the dehydrated gave a P-V distance of 4.2 A
and the species was assigned to VO2þ directly attached or
coordinated to the suface oxygen atoms of the Keggin
species. Apparently, the solution and solid phase struc-
tures of reduced vanadium(IV) containing polyoxomo-
lybdates are different.
Aerobic Oxidations. As stated above, the identity and
mode of action of the active oxidant formed from a mixture
of CO/O2 in the presence of Pd and H5PV2Mo10O40
remains a question. First, to study whether a nucleophilic
or electrophilic oxidant is formed we used thianthrene
oxide as a probe substrate, Scheme 2.19 Thus, a reaction
of 270 μmol thianthrene oxide with 7 bar CO and 3 bar O2
catalyzed by 10 μmol H5PV2Mo10O40 and 1 μmol Pd
(5% Pd/Al2O3) in 1.5 mL 9:1 AcOH/H2O with 0.1 mmol
AcONa at 90 ꢀC for 15 h yielded only 95 μmol sulfone
indicating a reactive nucleophilic oxidant is operative.
Furthermore, using 1 mmol toluene as substrate
under otherwise identical reactions conditions as above
yielded 90% arene ring hydroxylation products at 80%
conversion with strong suppression of autoxidation at the
a Reaction conditions - 1 mmol 1-octene, 7 bar CO, 3 bar O2, 10 μmol
H5PV2Mo10O40, 1 μmol Pd (5% Pd/Al2O3) in 1.5 mL 9:1 AcOH/H2O,
0.1 mmol AcONa at 90 ꢀC for 15 h. Products were identified using
reference samples and GC-MS and quantified using GC. About 30% of
the CO is oxidized to CO2. Reactions with different palladium sources
and alternative solvent compositions, and various terminal alkenes are
presented in Supporting Information, Tables S1-S3.
benzylic position, Scheme 3. The importance of Pd in the
activation of CO and also the direction of reaction toward
ring hydroxylation versus benzylic autoxidation is pro-
minent. Finally, the reactivity of terminal alkenes such as
1-octene with this catalytic system was most informative
despite the rather complicated mixture of products ob-
tained, Chart 1.
The results show importantly that the Wacker reaction
previously reported to be catalyzed by Pd(II)/ H5PV2-
Mo10O40 in the presence of chloride was suppressed,2 as
was the autoxidative allylic oxidation. Two major path-
ways for the transformation of 1-octene were observed.
The first, about 40% of the products, yielded a ∼1:1 ratio
of nonanoic, 1a, and 2-methyloctanoic, 1b, acids. The
formation of the saturated carboxylic acids rather than
unsaturated acids or dicarboxylic acids indicates that a
(15) Smith, T. S.; LoBrutto, R.; Pecoraro, V. L. Coord. Chem. Rev. 2002,
228, 1–18.
(16) Hirao, H.; Kumar, D.; Chen, H.; Neumann, R.; Shaik, S. J. Phys.
Chem. C 2007, 111, 7711–7719.
(17) Cornman, C. R.; Geiser, K. M.; Rowley, S. P.; Boyle, P. D. Inorg.
Chem. 1997, 36, 6401–6408.
::
::
(18) Poppl, A.; Manikandan, P.; Kohler, K.; Maas, P.; Strauch, P.;
::
Bottcher, R.; Goldfarb, D. J. Am. Chem. Soc. 2001, 123, 4577–4584.
(19) (a) Adam, W.; Haas, W.; Lohray, B. B. J. Am. Chem. Soc. 1991, 113,
6202–6208. (b) Adam, W.; Golsch, D. Chem. Ber. 1994, 127, 1111–1113.

Article
Inorganic Chemistry, Vol. 48, No. 16, 2009 7951
Scheme 4. Formation of Carboxylic Acids by Hydrocarbonylation
temperature to be bonded only to the vanadium(IV) atom
supported on the polyoxometalate leads to the possibility,
but does not prove, that this species is preferably involved
in the activation and reaction of CO. At higher tempera-
tures, where oxygen transfer and oxidation occur, the
dynamic behavior of various catalyst species may involve
fast interchange of many possible intermediates. Still, in
such a scenario and since species II is associated with the
1,6 and 1,11 isomers of the oxidized H5PV2Mo10O40, one
may further speculate that the formation of species I
occurs upon reaction of CO activated by Pd(0) with the
1,2; 1,4 and 1,5 isomers of oxidized H5PV2Mo10O40. The
higher reactivity of the vicinal isomers of H5PV2Mo10O40
was already previously suggested in a different reaction,
the oxidation of aldehydes.23 One should, however, note
that calculations showed that one electron reduced forms
of all the isomers of H5PV2Mo10O40 have the same energy
although these calculations did not consider supported
species as observed in the study.16
By combining the EPR and catalytic results and based
on literature precedence where it has been shown that
aldehydes react with acidic polyoxometalates to yield
acetals or hemiacetals,24 we suggest a similar reductive
activation of CO. Therefore, in Scheme 5 we show the
reaction of Pd activated CO23 concomitant with electron
transfer at an equatorial oxygen atom of H5PV2Mo10O40
where the reactive species is a vicinal isomer.25 The mea-
sured vanadium-carbon distances and coupling constants
for species I support such a formulation. The formulation
of the EPR observed intermediate in Scheme 5 is shown as a
carbonyl complex (V-CO) rather than a carboxylate
complex V-C(O)OH because the conditions used to pre-
pare samples for the EPR spectra are mild (35 ꢀC, 10 min)
and do not yield CO2 which occurs at 90 ꢀC over 5 h.
Under anaerobic conditions and higher temperature
the carbonyl complex (V-CO) yields CO2 by oxygen
transfer from the polyoxometalate. Under O2 it may be
assumed that oxygen could either react at the vanadium-
(IV) atom to yield a vanadium(V)-superoxo species or at
the carbon atom to yield a peroxocarbonate species.
Under acidic conditions the vanadium(V)-superoxo spe-
cies would likely quickly disproportionate via second
order kinetics or abstract an allylic hydrogen of 1-
octene.26 Since allylic oxidation products were formed
only in small amounts we tend to discount this possi-
bility. On the other hand, peroxocarbonate species pre-
viously prepared in situ from bicarbonate and H2O2 have
been reported to epoxidize alkenes and oxidize sulfides
albeit atnearly neutral pH values.27 Itis possible or perhaps
likely that Pd also plays a role in the activation of alkenes
Pd(0) catalyzed hydrocarbonylation reaction rather than a
Pd(II) oxidative carbonylation reaction is operative,20
Scheme 4. Oxidative addition of the acidic H5PV2Mo10O40
and insertion of the alkene and CO in the presence of H2O
yielded the carboxylic acids. The second reaction pathway,
about 55% of products, was the result of epoxidation and
subsequent transformations, Chart 1. Thus, although the
original epoxidation product, 1-octene oxide, 4, was not
directly observed because of its apparent short lifetime
under reaction conditions, its use as a substrate under the
same reaction conditions gave exactly the same product
ratios for 5-9 as observed in the reaction of 1-octene. The
known transformations of 4 (ring-opening and esterifica-
tion, and C-C bond cleavage) are primarily due to the
acidic conditions dictated by the polyoxometalate.21
The formation of furanones deserved specific verifica-
tion and comment. Therefore, the reaction of heptanoic
acid as substrate (1 mmol heptanoic acid, 7 bar CO, 3 bar
O2, 10 μmol H5PV2Mo10O40, 1 μmol Pd (5% Pd/Al2O3) in
1.5 mL 9:1 AcOH/H2O, 0.1 mmol AcONa, 90 ꢀC, 15 h)
showed a 11% conversion with formation of 2-methyl-, 2-
ethyl-, and 2-propylfuranone in an approximately ∼3:1:1
ratio. Such activation of remote C-H bonds in aliphatic
carboxylic acids has been reported in the past to be
catalyzed by Pt(II) using Pt(IV) as oxidant.22 Assumingly,
H5PV2Mo10O40 serves here as oxidant in this reaction
catalyzed by Pd(II).4,5
Mechanistic Discussion. From the combined EPR spec-
troscopy and catalysis studies several hypotheses may be
suggested concerning (i) the reduced H5PV2Mo10O40
species, (ii) the reaction of H5PV2Mo10O40 with CO
activated by Pd(0), and (iii) the reactivity of
H5PV2Mo10O40/Pd(0) activated CO under anaerobic
and aerobic conditions. The EPR measurements show
that two different types of species were formed upon
reduction of H5PV2Mo10O40 with CO. Major species I
showed CO bonded to a vanadium(IV) atom that is
remote or supported on the polyoxometalate cluster as
indicated by the absence of coupling to phosphorus and
molybdenum atoms. Minor species II does not appear to
bind CO, and the vanadium(IV) atom appears to be
incorporated within the polyoxometalate cluster as in-
dicated by the presence of an axial phosphorus atom and
equatorial molybdenum atoms but slightly distanced
(23) Khenkin, A. M.; Rosenberger, A.; Neumann, R. J. Catal. 1999, 182, 82–91.
(24) (a) Konishi, Y.; Sakata, K.; Misono, M.; Yoneda, Y. J. Catal. 1982,
77, 169–179. (b) Popova, G. A.; Budneva, A. A.; Andrushkevich, T. V. React.
Kinet. Catal. Lett. 1997, 61, 353–362.
(25) It is possible that CO is activated on a Pd(0) atom on the Al2O3
surface or via formation of soluble Pd(0) carbonyl species.
(26) Sawyer, D. T. In Oxygen Complexes and Oxygen Activation by
Transition Metals; Martell, A. E., Sawyer, D. T., Eds.; Plenum: New York,
1988; pp 131-148.
(27) (a) Richardson, D. E.; Yao, H.; Frank, K. M.; Bennett, D. A. J. Am.
Chem. Soc. 2000, 122, 1729–1739. (b) Yao, H.; Richardson, D. E. J. Am. Chem.
Soc. 2000, 122, 3220–3221. (c) Bennett, D. A.; Yao, H.; Richardson, D. E. Inorg.
Chem. 2001, 40, 2996–3001.
3-
from the PO4 core. Species II can be associated with
H5PV2Mo10O40 isomers with distal vanadium sites in the
oxidized state. The fact that CO was observed at low
(20) (a) Tsuji, J. New J. Chem. 2000, 24, 127–135. (b) Tsuji, J. Acc. Chem.
Res. 1969, 2, 144–152. (c) Tsuji, J.; Morikawa, M.; Kiji, J. Tetrahedron Lett.
1963, 1437–1440. (d) Bittler, K.; Kutepow, N. V.; Neubauer, K.; Reis, H. Angew.
Chem., Int. Ed. 1968, 7, 329–335.
(21) Sheldon, R. A. J. Mol. Catal. 1983, 20, 1–26.
(22) Kao, L.-C.; Sen, A. Chem. Commun. 1991, 1242–1243.

7952 Inorganic Chemistry, Vol. 48, No. 16, 2009
Goldberg et al.
Scheme 5. Suggested Pathway for the Activation of CO on Vicinal Isomers of H5PV2Mo10O40 Mediated by Pd(0) and its Further Reaction under
Anaerobic and Aerobic Conditions
and arenes for epoxidation and hydroxylation reactions,
respectively.
analysis (Mettler 50) indicated 34 water molecules per polyoxome-
talate unit. Elemental analysis: experimental (calculated) % P 1.31
(1.34), V 4.38 (4.41), Mo 41.32 (41.56). IR: 1057, 960, 865, and 774
cm-1 31P NMR: (CD3COCD3, 85% H3PO4 external standard)
.
Conclusion
δ -3.96 (6), -3.42 (1), -3.37 (4), -3.31 (2), and -3.22 (2) ppm
(area of peak). 18O labeled H5PV2Mo10O40 (∼50% enrichment)
was prepared and quantified as reported.7b 13CO (99% 13C) was
purchased from Aldrich.
Carbon monoxide activated by Pd(0) was oxidized under
anaerobic conditions to CO2 by an electron transfer-oxygen
transfer mechanism with H5PV2Mo10O40 as oxygen donor.
EPR and HYSCORE measurements showed the formation
of two species upon reduction of H5PV2Mo10O40 with 13CO.
Major species I (65-70%) showed coupling with hydrogen
atoms of the solvent and polyoxometalate but not 31P and
95Mo, which together with the g and A parameters found,
leads to the assignment of species I to a species resembling a
vanadyl cation that probably is supported on the polyox-
ometalate. Additionally, the substantial 13C hyperfine cou-
pling indicates a bonding interaction with 13CO. Minor
species II (30-35%) showed weak axial coupling with 31P
Reactions. Reactions were carried out in Teflon lined auto-
claves. In a typical reaction, 10 μmol H5PV2Mo10O40 and 5%
Pd/alumina (2 mg) were dissolved in 1.5 mL AcOH 90%
containing 0.1 mmol NaOAc. One mmol substrate was added.
CO (7 bar) and O2 (3 bar) were charged, and the mixture was
allowed to react under stirring at 90 ꢀC for 15 h. After, the reaction
mixture was cooled, and GC and GC-MS analyses were performed.
EPR Measurements. EPR samples for 13CO measurements
were prepared in a quartz EPR tube containing 5 mM
H5PV2Mo10O40 and 5% Pd/Al2O3 (1 mg) in 8:1:1 AcOH:H2O:
glycerol that were degassed by three pump/thaw cycles. The
degassed sample was attached to a double switch lock attached
to a vacuum line and a 13CO lecture bottle. 13CO was allowed to
flow into the EPR tube. The reduced polyoxometalate (blue
color) was formed under mild heating, ∼35 ꢀC, 10 min, and slow
stirring. The tube was then sealed with flame. Pulse EPR
measurements were carried out at 9.8 GHz and 25 K on an
EleXsys E-580 Bruker spectrometer. X-band echo-detected
(ED) EPR spectra were recorded using the two pulse echo with
pulse durations of 16 and 32 ns and a time interval τ=300 ns.
For HYSCORE29 (π/2-τ- π/2-t1-π-t2- π/2-τ- echo) the pulse
durations were tπ/2=12 ns and tπ=24 ns and τ=140 ns. The dwell
time in t1 and t2 was 20 ns; 125 ꢀ 125 points were collected, and a
repetition time of 4.0 ms with 3 shots per point was used. Scans
were accumulated until sufficient signal-to-noise ratio was
achieved. Four-step phase cycle was applied. The time domain
data of HYSCORE experiments were manipulated as follows:
Initially the background decay was removed by a polynomial fit,
and then the data were apodized with a Sine Bell window, zero
filled, and Fourier transformation was carried out after which the
magnitude of the spectrum was calculated. Simulations of the ED-
EPR spectrum were carried out using EasySpin.30
˚
(P-V distance estimated as 4.75 A), strong equatorial cou-
pling with 95Mo, and no interaction with 13C. Thus, species II
is attributed to a reduced species where the vanadium(IV)
atom is incorporated in the polyoxometalate framework but
slightly distanced from the phosphate core that does not bind
CO. Under aerobic conditions, CO/O2, the oxidation of
thianthrene oxide to the sulfone rather than the disulfoxide
indicates the formation of a nucleophilic oxygen donor.
Oxidation reactions performed on terminal alkenes such as
1-octene yielded (∼40%) nonanoic and 2-methyoctanoic acid
in a 1:1 ratio as a result of hydrocarbonylation and a compli-
cated mixture of products (∼60%) that was, however, clearly a
result of alkene epoxidation followed by subsequent reactions
of the intermediate epoxide. One can speculate that the active
oxygen donor formed may be a peroxocarbonate species.
Experimental Section
General Methods and Materials. All the chemicals and sol-
vents were purchased commercially and used without further
ꢀ
purification. The IR spectra were measured on a Nicolet Protege
Acknowledgment. The research was supported by the
German-Israeli Project Cooperation (DIP-G7.1), the Israel
Science Foundation, the Kimmel Center for Molecular Design
and the Ilse Katz Center for Magnetic Resonance. R.N. is the
Rebecca and Israel Sieff Professor of Organic Chemistry. D.G.
holds the Erich Klieger professorial chair of Chemical Physics.
460 FTIR; solid samples were prepared as ∼3-5 wt % KBr
based pellets. The 31P NMR spectra were measured on a Bruker
Avance DPX 500 spectrometer. Reactions were carried out in
20 mL glass pressure tubes. Products were characterized using
GLC (HP-6890 gas chromatograph) with a flame ionization
detector and a 30 m ꢀ 0.32 mm 5% phenylmethylsilicone
(0.25 μm coating) capillary column and helium carrier gas.
The products were identified using a gas chromatograph equipped
with a mass-selective detector (GC-MS HP 5973) equipped with the
Supporting Information Available: Additional results (Tables
S1-S3) of catalytic reactions. This material is available free of
charge via the Internet at http://pubs.acs.org.
same column. The H5PV2Mo10O40 34H2Opolyoxometalatewaspre-
3
pared using a known literature method.28 Thermogravimetric
::
::
(29) Hofer, P.; Grupp, A.; Nebenfuhr, H.; Mehring, M. Chem. Phys. Lett.
1986, 132, 279–282.
(28) Tsigdinos, G. A.; Hallada, C. J. Inorg. Chem. 1968, 7, 437–441.
(30) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42–55.