Electron and oxygen transfer in polyoxometalate, H5PV2Mo10O40, catalyzed oxidation of aromatic and alkyl aromatic compounds: Evidence for aerobic Mars-van Krevelen-type reactions in the liquid homogeneous phase

Article abstract of DOI:10.1021/ja004163z

The mechanism of aerobic oxidation of aromatic and alkyl aromatic compounds using anthracene and xanthene, respectively, as a model compound was investigated using a phosphovanadomolybdate polyoxometalate, H₅PV₂Mo₁₀O₄₀, as catalyst under mild, liquid-phase conditions: The polyoxometalate is a soluble analogue of insoluble mixed-metal oxides often used for high-temperature gas-phase heterogeneous oxidation which proceed by a Mars-van Krevelen mechanism. The general purpose of the present investigation was to prove that a Mars-van Krevelen mechanism is possible also in liquid-phase, homogeneous oxidation reactions. First, the oxygen transfer from H₅PV₂Mo₁₀O₄₀ to the hydrocarbons was studied using various techniques to show that commonly observed liquid-phase oxidation mechanisms, autoxidation, and oxidative nucleophilic substitution were not occurring in this case. Techniques used included (a) use of ¹⁸O-labeled molecular oxygen, polyoxometalate, and water; (b) carrying out reactions under anaerobic conditions; (c) performing the reaction with an alternative nucleophile (acetate) or under anhydrous conditions; and (d) determination of the reaction stoichiometry. All of the experiments pointed against autoxidation and oxidative nucleophilic substitution and toward a Mars-van Krevelen mechanism. Second, the mode of activation of the hydrocarbon was determined to be by electron transfer, as opposed to hydrogen atom transfer from the hydrocarbon to the polyoxometalate. Kinetic studies showed that an outer-sphere electron transfer was probable with formation of a donor-acceptor complex. Further studies enabled the isolation and observation of intermediates by ESR and NMR spectroscopy. For anthracene, the immediate result of electron transfer, that is formation of an anthracene radical cation and reduced polyoxometalate, was observed by ESR spectroscopy. The ESR spectrum, together with kinetics experiments, including kinetic isotope experiments and ¹H NMR, support a Mars-van Krevelen mechanism in which the rate-determining step is the oxygen-transfer reaction between the polyoxometalate and the intermediate radical cation. Anthraquinone is the only observable reaction product. For xanthene, the radical cation could not be observed. Instead, the initial radical cation undergoes fast additional proton and electron transfer (or hydrogen atom transfer) to yield a stable benzylic cation observable by ¹H NMR. Again, kinetics experiments support the notion of an oxygen-transfer rate-determining step between the xanthenyl cation and the polyoxometalate, with formation of xanthen-9-one as the only product. Schemes summarizing the proposed reaction mechanisms are presented.

Full text of DOI:10.1021/ja004163z

J. Am. Chem. Soc. 2001, 123, 8531-8542
8531
Electron and Oxygen Transfer in Polyoxometalate, H5PV2Mo10O40,
Catalyzed Oxidation of Aromatic and Alkyl Aromatic Compounds:
Evidence for Aerobic Mars-van Krevelen-Type Reactions in the
Liquid Homogeneous Phase
Alexander M. Khenkin,† Lev Weiner,‡ Yan Wang,§ and Ronny Neumann*,†
Contribution from the Department of Organic Chemistry, Weizmann Institute of Science,
RehoVot, 76100 Israel, Chemical SerVices DiVision, Weizmann Institute of Science, RehoVot, 76100 Israel,
and Casali Institute of Applied Chemistry, The Hebrew UniVersity of Jerusalem, Jerusalem, Israel, 91904
ReceiVed December 4, 2000
Abstract: The mechanism of aerobic oxidation of aromatic and alkyl aromatic compounds using anthracene
and xanthene, respectively, as a model compound was investigated using a phosphovanadomolybdate
polyoxometalate, H5PV2Mo10O40, as catalyst under mild, liquid-phase conditions. The polyoxometalate is a
soluble analogue of insoluble mixed-metal oxides often used for high-temperature gas-phase heterogeneous
oxidation which proceed by a Mars-van Krevelen mechanism. The general purpose of the present investigation
was to prove that a Mars-van Krevelen mechanism is possible also in liquid-phase, homogeneous oxidation
reactions. First, the oxygen transfer from H5PV2Mo10O40 to the hydrocarbons was studied using various
techniques to show that commonly observed liquid-phase oxidation mechanisms, autoxidation, and oxidative
nucleophilic substitution were not occurring in this case. Techniques used included (a) use of 18O-labeled
molecular oxygen, polyoxometalate, and water; (b) carrying out reactions under anaerobic conditions; (c)
performing the reaction with an alternative nucleophile (acetate) or under anhydrous conditions; and (d)
determination of the reaction stoichiometry. All of the experiments pointed against autoxidation and oxidative
nucleophilic substitution and toward a Mars-van Krevelen mechanism. Second, the mode of activation of the
hydrocarbon was determined to be by electron transfer, as opposed to hydrogen atom transfer from the
hydrocarbon to the polyoxometalate. Kinetic studies showed that an outer-sphere electron transfer was probable
with formation of a donor-acceptor complex. Further studies enabled the isolation and observation of
intermediates by ESR and NMR spectroscopy. For anthracene, the immediate result of electron transfer, that
is formation of an anthracene radical cation and reduced polyoxometalate, was observed by ESR spectroscopy.
1
The ESR spectrum, together with kinetics experiments, including kinetic isotope experiments and H NMR,
support a Mars-van Krevelen mechanism in which the rate-determining step is the oxygen-transfer reaction
between the polyoxometalate and the intermediate radical cation. Anthraquinone is the only observable reaction
product. For xanthene, the radical cation could not be observed. Instead, the initial radical cation undergoes
fast additional proton and electron transfer (or hydrogen atom transfer) to yield a stable benzylic cation observable
by 1H NMR. Again, kinetics experiments support the notion of an oxygen-transfer rate-determining step between
the xanthenyl cation and the polyoxometalate, with formation of xanthen-9-one as the only product. Schemes
summarizing the proposed reaction mechanisms are presented.
Introduction
metal-catalyzed free radical autoxidation pathways.1 Additional
possibilities include activation of molecular oxygen through
formation of singlet oxygen;2 use of oxygen to reoxidize redox
active metals, as in Wacker-type reactions;3 or activation in the
presence of reducing agents, as is typical for monooxygenase
enzymes and their mimetic counterparts.4 There are also a few
examples of activation of molecular oxygen by splitting of the
oxygen-oxygen bond in the absence of a reducing agent.5
In classic terms, homogeneous catalysis may be defined as a
reaction in a liquid phase whereby a dissolved compound of
well-defined molecular structure and property is used as a
catalyst. Heterogeneous catalysis on the other hand usually
occurs at a solid-gas or solid-liquid interface. The solid
catalyst is often or usually ill-defined from the point of view of
molecular structure. In the area of catalytic oxidation of
hydrocarbons using molecular oxygen as oxidant, homogeneous
and heterogeneous catalysis often also differ entirely as to the
mechanism of oxidation and the mode of activation of molecular
oxygen. Typically, in classic homogeneous liquid-phase aerobic
oxidation, molecular oxygen reacts via the well-established
(1) (a) Emanuel, N. M.; Denisov, E. T.; Maizus, Z. K. Liquid Phase
Oxidation of Hydrocarbons; Plenum: New York, 1967. (b) Mayo, F. R.
Acc. Chem. Res. 1968, 1, 193. (c) Howard, J. A. In Free Radicals; Kochi,
J. K., Ed.; Wiley: New York, 1973; Vol. 2, p 3. (d) Betts, J. Q. ReV. Chem.
Soc. 1971, 25, 265.
(2) Wasserman, H. H.; Murray, R. W. Singlet Oxygen; Academic Press:
New York, 1978.
(3) (a) Jira, R.; Freiesleben, W. Organomet. React. 1972, 3, 1. (b) Gates,
B. C.; Katzer, J. R.; Schuit, G. A. Chemistry of Catalytic Processes;
McGraw-Hill: New York, 1969.
† Department of Organic Chemistry, Weizmann Institute of Science.
‡ Chemical Services Division, Weizmann Institute of Science.
§ The Hebrew University of Jerusalem.
10.1021/ja004163z CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/08/2001

8532 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
Scheme 1
In gas phase, heterogeneous oxidation catalysis autoxidation
pathways are significantly suppressed. Instead, especially with
prevalent metal oxide catalysts, molecular oxygen activation
often occurs via the well-known and established Mars-van
Krevelen mechanism, Scheme 1.6 This two-stage mechanism
involves, first, C-H bond activation of a hydrocarbon, for
example, by electron transfer and proton transfer from the
hydrocarbon to the metal oxide, followed by lattice oxygen
transfer from the oxide to yield the reaction product and the
reduced catalyst. Second, there is reoxidation of the catalyst by
molecular oxygen, coupled with formation of water. A typical
example of an oxidation reaction of this kind is the oxidation
of propylene to acrolein.7
Since the early 1970s, a significant body of catalysis research
has dealt with the heterogenization of homogeneous catalysts.8
The impetus of this research has been to combine the selectivity
and milder conditions attainable by use of homogeneous
catalysts with the catalyst-handling advantages inherent in
heterogeneous catalysis. Typical approaches include attachment
or support of homogeneous catalysts on insoluble matrixes9 or
use of biphasic liquid-liquid media,10 whereby ideally, one
phase contains the catalyst and the other contains the substrate
and product. Interestingly, the idea of homogeneous analogues
of heterogeneous catalysis is a considerably less developed area
of research. However, in the field oxidation catalysis using
transition metal oxides as catalysts, there are several conceivable
advantages in the development of homogeneous analogues that
can be identified. Thus, hydrocarbon activation and oxidation
with molecular oxygen only via the Mars-van Krevelen
mechanism would avoid the occurrence of radical chain, and
generally nonselective, autoxidation reactions. In addition,
heterogeneous oxidation reactions at high temperatures often
are of reduced selectively as a result of over-oxidation to COx
and are usually limited to simple and low-molecular-weight
hydrocarbon molecules. Homogeneous analogues of such hetero-
geneous catalysis could, in principle, lead to viable, selective
lower temperature aerobic oxidation of many hydrocarbons.
Because insoluble and often nonstoichiometric mixed metal
oxides catalyze heterogeneous gas-phase oxidation reactions, a
parallel homogeneous system would require the use of well-
characterized, discrete, and soluble (mixed) metal oxides.
Anionic polyoxometalates11 are such compounds and, therefore,
are attractive candidates for Mars-van Krevelen-type reactions
in a homogeneous liquid phase. In this context, polyoxometalates
have over the past decade or so attracted attention for their
potential as oxidation catalysts12 and in electron-transfer proc-
esses.13 One approach to homogeneous catalysis by polyoxo-
metalates has been to consider a lacunary polyoxometalate as
an inorganic ligand to an active transition metal center. In the
presence of a suitable oxidant, an active intermediate formed
at the transition metal center leads to oxygenation of a substrate.
Various oxidants or oxygen donors such as iodosobenzene,14
N-oxides,15 periodate,16 ozone,17 tert-butylhydroperoxide,18 mo-
lecular oxygen5c,d,19 and hydrogen peroxide,20 have been shown
to be effective for alkene and alkane oxidation. Another catalytic
use of polyoxometalates in the liquid phase, using almost
(3 + x)-
exclusively phosphovanadomolybdates, PVxMo12-xO40
(x ) 0, 1, 2) (Figure 1), has been for oxidative dehydrogenation.
Oxidative dehydrogenation of cyclic dienes, such as dihydro-
anthracene and R-terpinene,21 alcohols and amines,22 and
(11) Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer: Berlin,
1983.
(12) (a) Mizuno, N.; Misono, M. Chem. ReV. 1998, 98, 199. (b) Okuhara,
T.; Mizuno, N.; Misono, M. AdV. Catal. 1996, 41, 113. (c) Hill, C. L.;
Duncan, D. C.; Harrup, M. K. Comments Inorg. Chem. 1993, 14, 367. (d)
Hill, C. L.; Prosser-McCartha, C. M. Coord. Chem. ReV. 1995, 143, 407.
(e) Neumann, R. Prog. Inorg. Chem. 1998, 47, 317. (f) Kozhevnikov, I. V.
Chem. ReV. 1998, 98, 171. (g) Kozhevnikov, I. V. Catal. ReV. Sci. Eng.
1995, 37, 311.
(4) (a) Ortiz de Montanello, P. Cytochrome P-450; Marcel Dekker: New
York, 1986. (b) Sheldon, R. A. Metalloporphyrins in Catalytic Oxidations;
Marcel Dekker: New York, 1994. (c) Montanari, F.; Casella, L. Metallo-
porphyrin Catalyzed Oxidations; Kluwer: Dordrecht, 1994. (d) Meunier,
B. Chem. ReV. 1992, 92, 1411.
(5) (a) Groves, J. T.; Quinn, R. J. Am. Chem. Soc. 1985, 107, 5790. (b)
Groves, J. T.; Ahn, K. H. Inorg. Chem. 1987, 26, 3831. (c) Neumann, R.;
Dahan, M. Nature 1997, 388, 353. (d) Neumann, R.; Dahan, M. J. Am.
Chem. Soc. 1998, 120, 11969.
(13) Weinstock, I. A. Chem. ReV. 1998, 98, 113.
(14) (a) Hill, C. L.; Brown, R. B. J. Am. Chem. Soc. 1986, 108, 536. (b)
Mansuy, D.; Bartoli, J.-F.; Battioni, P.; Lyon, D. K.; Finke, R. G. J. Am.
Chem. Soc. 1991, 113, 7222. (c) Weiner, H.; Hayashi, Y.; Finke, R. G.
Inorg. Chem. 1999, 38, 2579.
(15) Zhang, X.; Sasaki, K.; Hill, C. L. J. Am. Chem. Soc. 1996, 118,
4809.
(16) (a) Neumann, R.; Abu-Gnim, C. J. Chem. Soc., Chem. Commun.
1989, 1324. (b) Neumann, R.; Abu-Gnim, C. J. Am. Chem. Soc. 1990, 112,
6025. (c) Steckhan, E.; Kandzia, C. Synlett 1992, 139. (d) Bressan, M.;
Morvillo, A.; Romanello, G. J. Mol. Catal. 1992, 77, 283.
(17) Neumann, R.; Khenkin, A. M. Chem. Commun. 1998, 1967.
(18) (a) Faraj, M.; Hill, C. L. J. Chem. Soc., Chem. Commun. 1987,
1487. (b) Neumann, R.; Khenkin, A. M. Inorg. Chem. 1995, 34, 5753.
(19) (a) Lyons, J. E.; Ellis, P. E.; Durante, V. A. Stud. Surf. Sci. Catal.
1991, 67, 99. (b) Mizuno, N.; Hirose, T.; Tateishi, M.; Iwamoto, M. J.
Mol. Catal. 1994, 88, L125. (c) Mizuno, N.; Tateishi, M.; Hirose, T.;
Iwamoto, M. Chem. Lett. 1993, 2137. (e) Katsoulis, D. E.; Pope, M. T. J.
Chem. Soc., Dalton Trans. 1989, 1483. (f) Neumann, R.; Dahan, M. J. Chem.
Soc., Chem. Commun. 1995, 171. (g) Neumann, R.; Dahan, M. Polyhedron
1998, 17, 3557. (h) Weiner, H.; Finke, R. G. J. Am. Chem. Soc. 1999, 121,
9831. (i) Mizuno, N.; Hirose, T.; Tateishi, M.; Iwamoto, M. Chem. Lett.
1993, 1839. (j) Mizuno, N.; Tateishi, M.; Hirose, T.; Iwamoto, M. Chem.
Lett. 1993, 1985. (k) Khenkin, A. M.; Rosenberger, A.; Neumann, R. J.
Catal. 1999, 182, 82. (l) Neumann, R.; Khenkin, A. M.; Dahan, M. Angew.
Chem., Int. Eng. Ed. 1995, 34, 1587. (m) Hayashi, T.; Kishida, A.; Mizuno,
N. Chem. Commun. 2000, 381. (n) Khenkin, A. M.; Neumann, R. Inorg.
Chem. 2000, 39, 3455.
(6) (a) Mars, P.; van Krevelen D. W. Chem. Eng. Sci. 1954, 3, 41. (b)
Satterfield, C. N. Heterogeneous Catalysis in Practice; McGraw-Hill: New
York, 1980.
(7) Burrington, J. D.; Kartisek, C. T.; Grasselli R. K. J. Catal. 1983, 81,
489.
(8) Cornils, B.; Herrmann, W. A. Applied Homogeneous Catalysis with
Organometallic Compounds; VCH: Weinheim, 1996; Chapter 3.
(9) (a) Hartley, F. R. Supported Metal Complexes; Reidal: Dordrecht,
1985. (b) Sheldon, R. A. Curr. Opin. Solid State Mater. Sci. 1996, 1, 101.
(c) Knops-Gerrits, P. P.; Vos, D. E.; Thibault-Starzyk, F.; Jacobs, P. A.
Nature 1994, 369, 543. (d) Parton, R. F.; Vankelecom, I. F. J.; Casselman,
M. J. A.; Bezoukhanova, C. P.; Uytterhoeven, J. B.; Jacobs, P. A. Nature
1994, 370, 541. (e) Vankelecom, I. F. J.; Tas, D.; Parton, R. F.; Van de
Vyver, V.; Jacobs, P. A. Angew. Chem., Int. Ed. Engl. 1996, 35, 1346. (f)
Rony, P. E. J. Mol. Catal. 1975, 1, 13. (g) Davis, M. E. Chemtech 1992,
498.
(10) (a) Cornils, B.; Wiebus, E. Chemtech 1995, 33. (b) Chaudhari, R.
V.; Bhanage, B. M.; Deshpande, R. M.; Delmas, H. Nature 1995, 373, 501.
(c) Chaudhari, R. V.; Bhattachyna, A.; Bhanage, B. M. Catal. Today 1995,
24, 123. (d) Horvath, I. T.; Rabai, J. Science 1994, 266, 72.

Electron, Oxygen Transfer in H5PV2Mo10O40
J. Am. Chem. Soc., Vol. 123, No. 35, 2001 8533
electron transfer, including observation of reaction intermediates,
and present some insight on the oxygen-transfer step from the
polyoxometalate to the activated hydrocarbon.
Results and Discussion
Oxygen Transfer from H5PV2Mo10O40 to Aromatic and
Alkyl Aromatic Hydrocarbons. To investigate the oxidation
of aromatic and alkyl aromatic compounds, relatively reactive
anthracene and xanthene, respectively, were used as model
substrates under aerobic, 1 atm pure O2, conditions using the
H5PV2Mo10O40 compound as catalyst. Thus, both anthracene
and xanthene (60 mM) and a Keggin-type polyoxometalate,
H5PV2Mo10O40 (1 mM), in acetonitrile were reacted under 1
atm molecular oxygen at 60 °C for 18 h. Analysis of the reaction
mixture showed highly selective oxidation (>99%) of an-
thracene to anthraquinone and xanthene to xanthen-9-one at
conversions of 92% and 96%, respectively. The polyoxometalate
is stable under the reaction conditions; the 31P NMR and IR
are unchanged after the reaction. As indicated in the Introduc-
tion, a Mars-van Krevelen-type oxidation, Scheme 1, implies
an oxygen transfer from the catalyst to an activated hydrocarbon
intermediate. On the other hand, the well-accepted mecha-
nism(s)1,26 for oxidation of alkyl aromatic compounds in the
liquid phase, Scheme 3, requires initiation of the reaction by
the formation of a radical intermediate either by direct hydrogen
atom abstraction (pathway a) or by electron transfer followed
by deprotonation (pathway b). The following formation of the
peroxo intermediate is very fast, in fact, diffusion-controlled.
The peroxo intermediate can increase the radical chain length
and, therefore, autocatalyze the reaction via formation of more
alkyl radical species. At secondary alkyl aromatic positions, both
alcohols and ketones are the major reaction products.
5-
Figure 1. Ball-and-stick representation of the PV2Mo10O40 poly-
oxometalate highlighting one octahedral vanadium (molybdenum)
position and noting the different types of oxygen atoms.
Scheme 2
oxidation of phenols,23 has been studied. In these cases, the
substrates are oxydehydrogenated (electron and proton transfer
from the substrate, P-H2, to the polyoxometalate) but not
oxygenated (oxygen transfer from the metal oxide) (Scheme
2).
In general, oxygen transfer through implementation of the
Mars-van-Krevelen-type mechanism using molecular oxygen
as the terminal oxidant24 has until recently not been reported in
liquid-phase homogeneous oxidation catalysis. In this connec-
tion, we have just recently published a preliminary communica-
tion describing such a catalytic process.25 Here, we now present
our more complete studies of our finding that vanadium contain-
The liklihood of formation of radical intermediates in aromatic
compounds without alkyl side chains is smaller as a result of
the stronger C-H bonds. Thus, an additional pathway is possible
via electron-transfer oxidation of the substrate coupled with
nucleophilic attack25,27 (Scheme 4).
An initially formed radical cation can react with a nucleophile,
resulting in oxidative nucleophilic substitution. Using water and
anthracene as the substrates, the tautomer of 9-hydroxyanthra-
cene anthrone is formed. Under aerobic conditions, water is
generated in a separate catalyst reoxidation step in a process
similar to the one outlined in Scheme 2. Further reaction at the
benzylic carbon of anthrone will yield anthraquinone. An alkyl
aromatic substrate may also be oxygenated via an oxidative
nucleophilic substitution. Here, initial electron transfer followed
by further proton and electron transfer or hydrogen transfer will
yield a benzylic cation. The latter would react with a nucleo-
phile; thus, the oxidation of xanthene would yield xanthen-9-ol
(Scheme 5).
According to the autoxidation pathway, Scheme 3, upon
oxidation of hydrocarbons the only source of the oxygen atoms
in the products is from atmospheric molecular oxygen; therefore,
the oxidation of anthracene and xanthene was carried out using
high-purity (96.1%) 18O2. The isotope enrichment of anthra-
quinone and xanthen-9-one as a function of the turnover number
is given in Figure 2. From the figure, one may observe that
after one turnover, no 18O was observed in the organic products;
however, as the reaction continued, the 18O isotope enrichment
increased as a function of the number of turnovers. This
5-
ing polyoxometalates of the Keggin structure, PV2Mo10O40
(Figure 1), can activate C-H bonds in aromatic and alkyl
aromatic compounds and oxygenate such substrates by a Mars-
van Krevelen-type mechanism. We will also show and discuss
the evidence for the activation of the hydrocarbon substrate by
(20) (a) Khenkin, A. M.; Hill, C. L. MendeleeV Commun. 1993, 140.
(b) Neumann, R.; Gara, M. J. Am. Chem. Soc. 1994, 116, 5509. (c)
Neumann, R.; Gara, M. J. Am. Chem. Soc. 1995, 117, 5066 (d) Neumann,
R.; Khenkin, A. M. J. Mol. Catal. 1996, 114, 169. (e) Neumann, R.; Juwiler,
D. Tetrahedron 1996, 47, 8781. (f) Neumann, R.; Khenkin, A. M.; Juwiler,
D.; Miller, H.; Gara, M. J. Mol. Catal. 1997, 117, 169. (g) Mizuno, N.;
Nozaki, C.; Kiyoto, I.; Misono, M. J. Am. Chem. Soc. 1998, 120, 9267. (h)
Zhang, Z.; Chen, Q.; Duncan, D. C.; Lachicotte, R. J.; Hill, C. L. Inorg.
Chem. 1997, 36, 4381. (i) Zhang, X.; Chen, Q.; Duncan, D. C.; Campana,
C. F.; Hill, C. L. Inorg. Chem. 1997, 36, 4208. (j) Mizuno, N.; Seki, Y.;
Nishiyama, Y.; Kiyoto, I.; Misono, M. J. Catal. 1999, 184, 550. (k) Mizuno,
N.; Nozaki, C.; Kiyoto, I.; Misono, M. J. Catal. 1999, 182, 285. (l) Ben-
Daniel, R.; Khenkin, A. M.; Neumann, R. Chem. Eur. J. 2000, 6, 3722.
(21) (a) Neumann, R.; Lissel, M. J. Org. Chem. 1989, 54, 4607. (b)
Neumann, R.; Levin, M. J. Am. Chem. Soc. 1992, 114, 7278.
(22) Neumann, R.; Levin, M. J. Org. Chem. 1991, 56, 5707.
(23) (a) Kholdeeva, O. A.; Golovin, A. V.; Maksimovskaya, R. A.;
Kozhevnikov, I. V. J. Mol. Catal. 1992, 75, 235. (b) Lissel, M.; Jansen
van de Wal, H.; Neumann, R. Tetrahedron Lett. 1992, 33, 1795.
(24) Oxygen transfer from stable polyoxometalates (CrV ) O) has been
described, but the latter is prepared with iodosobenzene. Khenkin, A. M.;
Hill, C. L. J. Am. Chem. Soc. 1993, 115, 8178.
(26) Sheldon, R. A.; Kochi, J. K. Metal Catalyzed Oxidation of Orgainic
Compounds; Academic Press: New York, 1981.
(25) Khenkin, A. M.; Neumann, R. Angew. Chem., Int. Ed. 2000, 39,
4088.
(27) (a) Taylor, E. C.; McKillop, A. Acc. Chem. Res. 1970, 3, 338. (b)
Fukuzumi, S.; Nakanishi, I.; Tanaka, K. J. Phys. Chem. A 1999, 103, 11212.

8534 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
Scheme 3
Scheme 4
Scheme 5
polyoxometalate is low, and the excess H216O has little effect.
As the reaction continues (TON 8-15), the excess H216O
(H216O/H218O ) 5 ( 1) dilutes the18O load in the polyoxo-
metalate, leading to the observed nonlinearity in Figure 2.
An additional test of a Mars-van Krevelen versus free radical
mechanism can be carried out by performing the reaction under
strictly anaerobic conditions. Thus, independent reactions of
equimolar amounts of anthracene or xanthene and H5PV2-
Mo10O40 (1 mM in CD3CN) were carried out at 60 °C for 24 h
1
under argon in a NMR tube and yielded by H NMR up to
48% anthraquinone and 71% xanthen-9-one, respectively.
Because each polyoxometalate molecule can be formally
considered a three-electron oxidant,28 the maximum theoretical
or possible conversions to anthraquinone and xanthen-9-one by
a Mars-van Krevelen mechanism are 50% and 75%, respec-
tively (Scheme 6).
Further support for showing that the source of oxygen in the
oxidation product is the polyoxometalate rather than molecular
oxygen can be found by carrying out similar independent
Figure 2. 18O incorporation as a function of the turnover number.
Reaction conditions: anthracene or xanthene (20 mM) and H5PV2-
Mo10O40‚34H2O (1mM) in CH3CN under 1 atm 18O2 (96.1% purity) at
60 °C. The amount of 18O in the product was determined by GC/MS.
Asterisk (*), anthracene; ^, xanthene.
experiment clearly indicates that the source of the 18O in the
products is not directly from the molecular oxygen. The results,
however, do comply with a Mars-van Krevelen mechanism
whereby the original catalyst contains only 16O. As the reaction
advances, the polyoxometalate is enriched with 18O, and thus,
the product is progressively labeled. In this context, it should
be noted that there is no exchange of lattice oxygen of H5PV2-
Mo10O40 with labeled molecular oxygen in the absence of
substrate, even after 2 weeks. Furthermore, after one turnover,
the expected amount of H218O was observed (1 ( 0.2 µmol).
18O incorporation in the polyoxometalate was also observed by
the isotope shifts in the IR spectrum. The curves in Figure 2
tend to level off instead of continuing in a linear fashion. This
is due to the fact that water does exchange oxygen atoms in
the polyoxometalate framework, and H216O is already present
at the beginning of the reaction (∼50 mol water from hydrate
and solvent/mol polyoxometalate). From separate measurements,
the equilibration reaction (water-polyoxometalate) is relatively
fast (within 30 min at 60 °C), as compared with the oxidation
reaction. At low turnover number (1-5), the 18O load in the
anaerobic reactions of equimolar amounts (1 mM in CH3CN)
18
of anthracene or xanthene and H5PV2Mo10
O
40
(75% enrich-
ment). GC/MS analysis of the products showed that the
anthraquinone was 70% 18O-labeled, whereas the xanthen-9-
one was 72% 18O-labeled. The reaction stoichiometries, mo-
lecular oxygen/anthracene and molecular oxygen/xanthene, must
be 1.5/1 and 1/1 to support a Mars-van Krevelen mechanism.
Indeed, O2 consumption was measured using a gas buret in an
aerobic reaction [anthracene or xanthene (60 mM), H5PV2-
Mo10O40 (1 mM) in acetonitrile, 1 atm O2, 60 °C, 18 h].
The ratios were found to be 1.45 ( 0.2 equiv/reaction
cycle for anthracene and 0.95 ( 0.3 equiv/reaction cycle for
xanthene.
The results presented above quite conclusively go against the
formulation of an autoxidation mechanism, Scheme 3, for the
oxidation of either xanthene or anthracene. It is also worth
(28) One electron is transferred in the initial polyoxometalate-
hydrocarbon interaction, and during the oxygen transfer, there is an
additional reduction of the polyoxometalate.

Electron, Oxygen Transfer in H5PV2Mo10O40
J. Am. Chem. Soc., Vol. 123, No. 35, 2001 8535
Scheme 6
mentioning in this context that a compound, such as cyclohex-
ene, that is highly sensitive to autoxidation does not react under
the given reaction conditions. The results, however, do not all
necessarily disprove an oxidative nucleophilic pathway (Schemes
4 and 5), whereby the reduced polyoxometalate is reoxidized,
as indicated in Scheme 2. This possibility should be especially
considered, because the H5PV2Mo10O40 catalyst was used as a
hydrate, because dehydrated H5PV2Mo10O40 is insoluble in
acetonitrile. In such a mechanistic scheme, the immediate source
of oxygen in the product is water, which is continually replaced
by molecular oxygen in the reoxidation process. The most
obvious way to discount the oxidative nucleophilic pathway is
to perform the reaction under anhydrous conditions. Therefore,
the nonhydrated acetonitrile soluble quaternary ammonium salt
Q5PV2Mo10O40 (Q ) tetrabutylammonium) was used as catalyst
in place of the hydrate of H5PV2Mo10O40. Unfortunately, the
oxidation potential of polyoxometalates is significantly affected
by the countercation and the acidity of the solvent. The nonprotic
form of the catalyst has a significantly reduced oxidation
potentia,29 and in acetonitrile, no reduction of the polyoxometa-
late or hydrocarbon oxidation was observed. However, with
anhydrous but somewhat acidic 1,1,1,3,3,3-hexafluoro-2-pro-
panol as solvent, the oxidation of 20 mM anthracene or xanthene
in the presence of 1 mM Q5PV2Mo10O40 (55 °C, 1 atm O2, 16
h) was possible. Both anthraquinone, 20% yield, and xanthen-
9-one, 33% yield, were observed as the only products. A second
obvious experiment is to perform the reaction using isotopically
enriched water in place of the usual hydrate, that is, H5PV2-
Mo10O40 in the presence of H218O. As noted above, this
experiment is complicated by the fact that there is relatively
fast oxygen atom exchange between water and the terminal,
edge-shared, and corner-shared oxygen atoms in the H5PV2-
Mo10O40 polyoxometalate. At 60 °C, one can estimate by the
IR isotope shifts, a 30 ( 5% exchange within 15 min. Despite
this limitation, reactions between equimolar amounts of an-
thracene or xanthene (1 mM) and H5PV2Mo10O40 (1 mM, dried),
500 mM H216O, and 500 mM H218O (1/1 H216O/H218O) in dry
acetonitrile at 60 °C under argon were carried out. The16O/18O
isotope ratios in the product were measured after 15 min (∼3-
5% conversion). The results showed 16O/18O isotope ratios of
4.2 for anthracene and 4.5 for xanthene versus the approximately
1/1 ratio that would have been expected if the oxygen atom in
the product came from water rather than an oxygen atom of
the polyoxometalate. A third experiment that pointed toward a
Mars-van Krevelen mechanism and against an oxidative
nucleophilic substitution was to perform the reaction in the
presence of acetic acid as an alternative nucleophile instead of
water. A nucleophilic attack by acetic acid would be expected
to yield some of the known acetates30 instead of only the
observed quinone or ketone (Schemes 4 and 5). Reactions under
the following conditions: anthracene or xanthene (20 mM),
H5PV2Mo10O40‚34H2O, and acetic acid (160 mM) at 60 °C
Scheme 7
under 1 atm O2 for 16 h, showed no formation of any acetylated
hydrocarbons. Only anthraquinone and xanthen-9-one (90 and
94% conversion, respectively) were obtained. Together, the three
experiments described disprove the possibility of an oxidative
nucleophilic mechanism to account for the product of the
formation of anthraquinone and xanthen-9-one. An additional
point very worth noting in this context is that the oxidative
nucleophilic mechanism first requires the formation of anthrone
and xanthen-9-ol. These products were never observed (see also
1
the H NMR experiments discussed latter on).
All of the experiments described above are consistent with a
Mars-van Krevelen mechanism; the experiments described are
inconsistent with autoxidation or oxidative nucleophilic substitu-
tion mechanisms.
Activation of Hydrocarbons with H5PV2Mo10O40. For
oxidation catalysis to occur, the catalyst must first activate the
hydrocarbon. One possible mode of activation is electron transfer
from the hydrocarbon to the H5PV2Mo10O40 polyoxometalate
to yield initially a radical cation of the hydrocarbon and a one-
electron reduced H5PV2Mo10O40 (Scheme 7.)31 Alternatively,
hydrogen abstraction and formation of a caged radical species
can be considered.32
The activation of the hydrocarbon was initially studied
kinetically by mixing the hydrocarbons and polyoxometalate
in solution under argon and following the appearance of a
reduced blue species, PVVVIVMo10O406-, at 750 nm in the
visible spectrum. Kinetic profiles were obtained from the
reaction of 1 µmol of H5PV2Mo10O40 and 20 µmol of hydro-
carbon in 3 mL acetonitrile under argon at 60 °C and showed
that the reactions were first-order in H5PV2Mo10O40 and pseudo-
zero-order in the substrate, which was in excess. The interaction
of H5PV2Mo10O40 with hydrocarbons having aromatic C-H
bonds, such as anthracene, 2-methoxyanthracene, 2-chloro-
anthracene, anthracene-d10, and phenanthrene, and hydrocarbons
with benzylic C-H bonds, such as xanthene, 2-methylanthra-
cene, fluorene, and diphenylmethane, was measured. The gas-
phase ionization potential,33 which was used as a measure of
the oxidation potential of the hydrocarbon, showed a very good
correlation (r2 ) 0.93) as a function of the rate constant (Figure
3). Because the nonbenzylic compounds have strong C-H bonds
at the reactive position, ∼105 kcal/mol, and the benzylic
compounds have weaker C-H bonds at the reaction site, 80-
85 kcal/mol, it is clear that there was no correlation of the rate
as a function of the C-H bond strength. The results appear to
support an electron-transfer reaction rather than a hydrogen atom
transfer process for activation of the hydrocarbons by the
polyoxometalate. For further support of this conclusion, the
interaction of H5PV2Mo10O40 with 4-methyl-, 4-ethyl-, and 4-i-
propylanisole was also investigated. The relative reactivity of
(29) (a) Carloni, P.; Eberson, L. Acta Chem. Scand. 1991, 45, 373. (b)
Keita, B.; Bouaziz, D.; Nadjo, L. J. Electrochem. Soc. 1988, 135, 87.
(30) (a) Koshitani, J.; Hiramatsu, M.; Ueno, Y.; Yoshida, T. Bull. Chem.
Soc., Jpn. 1978, 51, 3667. (b) Hanotier, J.; Hanotier-Bridoux, M.; de
Radadzitzky, P. J. Chem. Soc., Perkin Trans. 2 1973, 381.
(31) Shaik, S.; Shurki, A. Angew. Chem., Int. Ed. 1999, 38, 587.
(32) Mayer, J. M. Acc. Chem. Res. 1998, 31, 441.
(33) Taken from the NIST at http://webbook.nist.gov/chemistry

8536 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
Figure 4. Reaction of H5PV2Mo10O40 with several hydrocarbons as a
function of temperature. The reduction of H5PV2Mo10O40 was followed
by a diode array UV-vis spectrometer reading at 750 nm. Conditions:
1 µmol of H5PV2Mo10O40 and 20 µmol of hydrocarbon in 3 mL of
acetonitrile under argon at 30-60 °C. Asterisk (*), anthracene; ^,
xanthene; X in a box, 4-ethylanisole.
Figure 3. Reaction of H5PV2Mo10O40 with hydrocarbons as a function
of the gas-phase ionization potential. The reduction of H5PV2Mo10O40
was followed by a diode array UV-vis spectrometer reading at 750
nm. Conditions: 1 µmol of H5PV2Mo10O40 and 20 µmol of hydrocarbon
in 3 mL of acetonitrile under argon at 60 °C.
Table 2. Thermodynamic Data for the Electron-Transfer Step
between H5PV2Mo10O40 and Hydrocarbons
Table 1. Rate Constants for the Reduction of H5PV2Mo10O40 by
4-Alkylanisolesa
∆Hq
kcal/
mol
∆Sq
cal/
mol K
∆Gq
kcal/
mol
E°V
vs
SCE
1.37a
1.49b
1.52c
∆G°
kcal/
mol
∆G° ′
kcal/
mol
substrate
rate constant, sec-1
relative rate
compd
4-methylanisole
4-ethylanisole
4-i-propylanisole
3.00 × 10-5
2.59 × 10-4
4.51 × 10-5
0.12
1
0.18
anthracene
xanthene
4-ethylanisole
9.4
6.8
10.6
-42.5
-48.5
-39.6
22.0
21.3
22.3
15.7
18.5
19.1
7.9
10.7
11.4
a Reaction conditions: 1 µmol of H5PV2Mo10O40 and 10 µmol of
4-alkylanisole in 3 mL of acetonitrile under argon at 60 °C.
electrostatic term36 to give the corrected free energy value, ∆G°′
(eq 1).
the 4-alkylanisole compounds has also been used in the past as
a test for electron transfer versus hydrogen atom transfer
reactions.34 For electron-transfer reactions, the relative rates are
2° > 1° > 3° or 2° > 3° > 1°, whereas for hydrogen atom
transfer reactions, it is always found that 3° > 2° > 1°. Thus,
reactions of 4-alkylanisole with H5PV2Mo10O40 were studied
as above by reacting 1 µmol H5PV2Mo10O40 and 20 µmol
substituted anisole in 3 mL acetonitrile under argon at 60 °C.
From the kinetic profiles which also, as above, showed first-
order behavior for the reduction of H5PV2Mo10O40, one can
observe that the reaction is fastest for 4-ethylanisole > 4-i-
propylanisole > 4-methylanisole (Table 1). From the results
described above, which are also supported by examples to be
found in the literature describing oxydehydrogenation of other
substrates,13 such as dienes21b and phenols,23b and photoactivated
oxydehydrogenation,35 it would certainly appear that the initial
hydrocarbon-polyoxometalate reaction is an electron-transfer
process.
331.2B
r12D
∆G°′ ) ∆G° +
(Z1 - Z2 - 1)
12xµ/DT)
where B ) 10-(21.9r
(1)
B ) 1 assuming an ion strength, µ ) 0 and Z1 ) -5; Z2 ) 0;
D ) 35 (acetonitrile); r12 ) r1 + r2 ) 5.6 Å (H5PV2Mo10O40)
+ 1.7 Å (assuming side-on interaction between the polyoxo-
metalate and the aromatic ring) ) 7.3 Å.
In general, the free energy values, both ∆G° and ∆G°′ versus
∆Gq appear to support an outer-sphere electron transfer, as
opposed to an inner-sphere process for activation of the
hydrocarbon by H5PV2Mo10O40. However, as already pointed
out by others, the kinetic analysis of the reaction of polyoxo-
metalates with electron donors may be complicated by several
phenomena.13,37 These include effects of ionic strength, ion
pairing between the polyoxometalate and any cations in the
solution, and formation of stable complexes between the
polyoxometalate and the electron donor (hydrocarbon). First, it
should be noted that there is clear evidence for the formation
of stable complexes after electron transfer, as indicated from
the analysis of the ESR spectrum (see below) of the reaction
between anthracene and H5PV2Mo10O40. Also relevant are
further fast or possibly concurrent proton, electron, or atom
To gain more insight into the electron-transfer process, in
particular in order to distinguish between outer-sphere and inner-
sphere-type reactions, the electron-transfer reaction of H5PV2-
Mo10O40 with three select hydrocarbons, anthracene, xanthene,
and 4-ethylanisole, was measured as a function of temperature
(Figure 4). The plot enables the computation of the activation
parameters, and in Table 2 are summarized all of the thermo-
dynamic data for the electron-transfer step between H5PV2-
Mo10O40 and the three substrates. The free energy of reaction,
∆G°, was computed from ∆E° (∆G° ) 23.06∆E° kcal/mol)
using an oxidation potential of 0.69 V for H5PV2Mo10O40 in
acetonitrile, as measured by cyclic voltammetry. The free energy
under prevailing reaction conditions was corrected by an
1
transfers, as is indicated in the interpretation of the H NMR
spectrum (see below) of the reaction between xanthene and
H5PV2Mo10O40. The probable existence of preassociation com-
plexes between the polyoxometalate and the hydrocarbon
substrates may be inferred from calculated reorganization
energies. Assuming that there are no preassociation complexes,
the Marcus equation, eq 2, yields computed reorganization
(34) Baciocchi, E.; D’Acunzo, F.; Galli, C.; Lanzalunga, O. J. Chem.
Soc., Perkin Trans. 2 1996, 133.
(35) Renneke, R. F.; Hill, C. L. J. Am. Chem. Soc. 1988, 110, 5461.
(36) Eberson, L. AdV. Phys. Org. Chem. 1982, 18, 79.
(37) Eberson, L. New J. Chem. 1992, 16, 151.

Electron, Oxygen Transfer in H5PV2Mo10O40
J. Am. Chem. Soc., Vol. 123, No. 35, 2001 8537
Figure 5. ESR spectrum of a solution of Q5PV2Mo10O40 and an-
thracene. Top: 4 mM Q5PV2Mo10O40 and 2 mM anthracene in
1,1,1,3,3,3-hexafluoro-2-propanol under argon were reacted for 20 min
at 40 °C prior to the measurement of the ESR spectrum at room
temperature. Bottom: 4 mM H5PV2Mo10O40 and 2 mM anthracene in
acetonitrile under argon were reacted for 20 min at 40 °C prior to the
measurement of the ESR spectrum at room temperature.
energies, λ, of 102, 94, and 96 kcal/mol for anthracene,
xanthene, and 4-ethylanisole, respectively. These values are
seemingly and illogically high. If, however, electron transfer is
occurring within a stable preassociation complex, the Marcus
equation should be modified by dropping the Coulombic work
term, W(r), and not applying the electrostatic correction to ∆G°,
leaving only a quadratic dependence of ∆Gq on ∆G°, eq 3. The
reorganization energies, λ, which now may be calculated from
this latter expression are 52, 40, and 43 kcal/mol for anthracene,
xanthene and 4-ethylanisole, respectively. Thus, it would appear
that there is preassociation between the polyoxometalate and
the hydrocarbon substrate.
Figure 6. Expanded ESR spectrum of Figure 5 (top) at g ) 2.0027
(top) and computer simulation of the radical cation species (bottom).
practically identical I-V curves and an E° ) 1.58 V versus
SCE. In the past, significantly higher isotope effects, kH/kD )
∼6, have been observed in the electron transfer from alkyl
aromatic substrates such as 4-methylanisole. In that case, the
effect was ascribed to concurrent proton transfer during the
electron-transfer step.38
Observation of Reaction Intermediates. From the experi-
ments described above, one may conclude that the hydrocarbon
is activated by electron transfer to the polyoxometalate, resulting
in the formation of a radical cation and the mono-reduced
PVIVVVMo10O406-. We sought to observe this radical cation in
situ. In the case of anthracene and in acetonitrile as solvent,
the spectrum of the reduced paramagnetic polyoxometalate was
observed; however, we were unable to observe a radical cation
under these conditions. We turned to the use of 1,1,1,3,3,3-
hexafluoro-2-propanol (HFIP) as solvent, since this solvent is
known for its ability to stabilize radical cations.39 Therefore,
anthracene or xanthene and Q5PV2Mo10O40 in HFIP were
reacted under argon. These conditions allowed for the observa-
tion (overlapped) of both the reduced polyoxometalate (Figure
5, top) and the anthracene radical cation (Figure 6). First,
concerning the spectrum of the reduced paramagnetic
PVIVVVMo10O406-, one may observe that it is distinctly distorted
2
λ
4
∆G°′
λ
∆Gq ) W(r) + 1 +
(2)
(3)
(
)
λ
4
∆G° 2
∆Gq ) 1 +
(
)
λ
The isotope effect in the electron-transfer step was measured
by comparing the rate of reduction of H5PV2Mo10O40 with
nondeuterated and deuterated hydrocarbons. For anthracene/
anthracene-d10 the ratio, kH/kD ) 1.7 was measured. The source
of this isotope effect appears from the linear relationship
between the ionization potential and the rate observed in Figure
3 to be entirely due to the different ionization potentials of
anthracene and anthracene-d10, rather than to be due to any effect
of hydrogen or proton disassociation during the electron-transfer
step. A similar comparison between xanthene and xanthene-9-
d2 gave a ratio, kH/kD ) 2.3 ( 0.08. This higher kH/kD value
for xanthene, as compared with anthracene, suggests a different
explanation (see below), especially since only the benzylic
carbon, but not the aromatic carbons, is deuterated. Only
deuteration at the aromatic rings would most likely have an
effect on the ionization potential. Unfortunately, ionization
potentials were not available for xanthene-9-d2; however, cyclic
voltammetry measurements of xanthene and xanthene-9-d2 gave
6-
compared to the spectrum of PVIVVVMo10O40 formed in
acetonitrile (Figure 5, bottom), which does not involve associ-
ated radical cations or radical species.19k We attribute the
distortion to the interaction of the polyoxometalate with the
anthracene radical cation. Second, the spectrum of the species
(38) Eberson, L.; Wistrand, L.-G. Acta Chem. Scand. B 1980, 34, 349.
(39) Eberson, L.; Hartshorn, M. P.; Persson, O. J. Chem. Soc., Perkin
Trans. 2 1995, 1735.

8538 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
at g ) 2.0027 (Figure 6) is, indeed, indicative of the formation
of an anthracene radical cation.40 The spectrum consists of five
split components in an integrated ratio of 1:4:6:4:1. Additional
splitting of the quintet was also observed. Computer simulation
of the experimental spectrum yielded the hyperfine splitting
constants, a1,4,5,8 ) 3.4 G, a2,3,6,7 ) 0.95 G, and a9,10 ) ∼0.
Comparison of these hyperfine splitting constants with those
of a “pure” anthracene radical cation in which splitting constants
of a1,4,5,8 ) 3.08 G, a2,3,6,7 ) 1.39 G, and a9,10 ) 6.60 G were
measured reveals that there is a good agreement for a1,4,5,8 and
a2,3,6,7, but not for a9,10. The disappearance of the hyperfine
splitting associated with the 9,10 hydrogen atoms has also
previously been observed.41 For example, in trifluoracetic acid,
a splitting constant of a9,10 ) 0.25 G has been obtained and
associated with hydrogen bonding between the hydrogen of
anthracene at the 9(10) position and the trifluoroactetate. In this
case, we believe that the loss of hyperfine splitting is due to
the ion pairing between the polyoxometalate and the reactive
position of the anthracene radical cation. Even though we
observed a radical cation species in a medium that leads to
product formation, it should also be shown that the electron-
transfer step is on the reaction pathway to the anthraquinone
product. Thus, after reacting anthracene and Q5PV2Mo10O40 for
30 min, quantification of the anthracene radical cation by double
integration techniques shows ∼20 mol % conversion. A separate
UV-vis analysis of the degree of reduction of Q5PV2Mo10O40
indicated a similar 20% conversion. This correlation supports
the idea that the radical cation of anthracene is an intermediate
in the formation of anthraquinone. Although we were successful
in observing the anthracene radical cation, in similar experi-
ments with xanthene, only the spectrum of the reduced
PVIVVVMo10O406- species was observed. The radical cation of
xanthene was not observed by ESR spectrometry.
Figure 7. 1H NMR spectra of the reaction between xanthene and
H5PV2Mo10O40. Reaction conditions: 10 mM xanthene and 10 mM
H5PV2Mo10O40 in 0.5 mL of CD3CN under argon.
The reaction of anthracene or xanthene with H5PV2Mo10O40
in acetonitrile solutions was also followed by 1H NMR
spectroscopy. A reaction carried out in a NMR tube between
anthracene (2 mM anthracene, 4 mM H5PV2Mo10O40 in CD3CN
under argon at 60 °C), showed only the clean formation of
anthraquinone. No intermediate species or products such as
anthrone were observed. A similar experiment with xanthene
as substrate (2 mM xanthene, 4 mM H5PV2Mo10O40 in CD3CN
under argon at 60 °C) gave an entirely different result (Figure
7). In the NMR spectrum, there first appears after 20 min an
intermediate species (Figure 7, bottom). As the reaction
continues and after an additional 20 min, the formation of the
product, xanthen-9-one (marked p in the middle spectrum)
becomes apparent. The intermediate remains at an approximately
constant concentration. The spectrum of the intermediate species
appears to be very similar to the spectrum reported for a
xanthenyl cation, which was measured in 98% D2SO4.42 The
assignment of the peaks in this context is 10.37 ppm (s, 1H9),
8.65 ppm (bs, 4H1&3), 8.38 ppm (d, 2H4), and 8.10 ppm (t, 2H2).
Figure 8. Rate of reaction as a function of hydrocarbon concentration.
Reaction conditions: 2-9 mM substrate and 1 mM H5PV2Mo10O40 in
2 mL of acetonitrile, 1 atm O2, 60 °C. Asterisk (*), anthracene; ^,
xanthene.
cations are extremely strong acids43 and from calculations are
∼10 orders magnitude stronger acids than aromatic radical
cations.43 The electron-donating oxygen heteroatom in xanthene
stabilizes the benzylic cation. Interaction and stabilization of
the benzylic cation and polyoxometalate anion by ion pairing
or complex formation is certainly also very conceivable.
Reaction Kinetics in the Aerobic Oxidation. To shed more
light on the reaction mechanism and to determine the rate-
limiting step in the oxidation reaction, the kinetics of the aerobic
oxidation of both xanthene and anthracene was studied. First,
reaction order in each of the reaction components was deter-
mined by using the method of initial rates. Logarithmic van’t
Hoff plots, in which rate constants were obtained from series
of reaction profiles, were used to find the reaction order in the
hydrocarbon (2-9 mM) (Figure 8), polyoxometalate (0.25-2
mM) (Figure 9), and molecular oxygen (0.2-3 atm) (Figure
10). The reaction was found to be first-order in the hydrocarbon;
the slopes in the van’t Hoff plots were 0.93 (r2 ) 0.98) and
0.94 (r2 ) 0.99) for anthracene and xanthene, respectively.
Similarly, the reaction was observed also to be first-order in
H5PV2Mo10O40. Here, the slopes in the van’t Hoff plots were
0.92 (r2 ) 0.99) and 0.91 (r2 ) 0.99) for anthracene and
xanthene, respectively. Finally, van’t Hoff plots for the effect
of oxygen pressure yield slopes of 0.001 and 0.068 for
1
Notable also is the fact that xanthen-9-ol with H NMR peaks
at 7.48 ppm (d, 2H), 7.29 ppm (m, 2H) 7.10 ppm (m, 2H),
5.68 (d, 1H-benzylic) and 2.50 (d, 1H-hydroxyl) was not
detected. From the 1H NMR experiment for the case of xanthene,
it would be reasonable to conclude that a proton and then
additional electron transfer to yield a benzylic cation as an
intermediate quickly follow the initial electron transfer. The logic
behind this hypothesis stems from the fact that benzylic radical
(40) Sullivan, P. D.; Menger, E. M.; Reddoch, A. H.; Paskovich, D. H.
J. Phys. Chem. 1978, 82, 1158.
(41) Eloranta, J.; Sippula, A. Finn. Chem. Lett. 1975, 170.
(42) Dradi, E.; Gatti, G. J. Am. Chem. Soc. 1975, 97, 5472.
(43) Nicholas, A. M. P.; Arnold, D. R. Can. J. Chem. 1982, 60, 2165.

Electron, Oxygen Transfer in H5PV2Mo10O40
J. Am. Chem. Soc., Vol. 123, No. 35, 2001 8539
mM), anthracene-d10 (10 mM), and H5PV2Mo10O40 (1 mM) in
CH3CN at 60 °C and 1 atm O2, an inverse kinetic isotope effect,
kH/kD ) 0.85 ( 0.05, for the formation of anthraquinone was
measured. This result is consistent with a change of hybridiza-
tion from sp2 to sp3 at the 9(10) aromatic carbon of anthracene
in the rate-determining step.44 Therefore, in the rate-determining
step, there must be formation of a C-O bond at the 9(10)
aromatic carbon of anthracene. Similarly, the kinetic isotope
effect was measured for xanthene/xanthene-9-d2 (20 mM) and
H5PV2Mo10O40 (1 mM) in CH3CN at 60 °C and 1 atm O2. A
kinetic isotope effect, kH/kD ) 2.3 ( 0.1, for the formation of
xanthen-9-one was obtained. This result, combined with the
1
observation of the xanthenyl cation in the H NMR spectrum
upon reaction of xanthene and H5PV2Mo10O40 under anaerobic
conditions, and the identical isotope effect in the electron transfer
between xanthene and H5PV2Mo10O40, leads to the conclusion
that the rate-determining step in this case is the formation of a
xanthenyl cation.
Figure 9. Rate of reaction as a function of H5PV2Mo10O40 concentra-
tion. Reaction conditions: 9 mM substrate and 0.25-2.0 mM H5PV2-
Mo10O40 in 2 mL of acetonitrile, 1 atm O2, 60 °C. Asterisk (*),
anthracene; ^, xanthene.
Reoxidation of the Reduced Polyoxometalate. According
to the Mars-van Krevelen mechanism (Scheme 1), the original
catalyst is regenerated by reaction of the reduced and deoxy-
genated polyoxometalate and protons with molecular oxygen
with coformation of water. In the present case, the study of the
reoxidation step is complicated by two factors. First, from a
kinetic point of view, the reoxidation is not rate-determining
and is also very fast.19k Second, there are fast exchange reactions
between the oxygen atom in water and the terminal and bridging
oxygen atoms in H5PV2Mo10O40, as mentioned above and noted
in the literature.45 Taking this into account, two diametrical
scenarios can be pictured for the polyoxometalate regeneration.
In the first scenario, the deoxygenated polyoxometalate reacts
first with water to yield a reoxygenated, but still reduced,
polyoxometalate. The latter is then reoxidized by molecular
oxygen, and water is then formed again. Alternatively, as is
accepted for the Mars-van Krevelen mechanism in the gas
phase, the polyoxometalate framework oxygen atoms are
replaced directly by molecular oxygen with coformation of
water. A manifestation of these opposing views can also be
inferred from some reports described in the literature. Kawafune
was the first to study the reaction between triphenylphosphine
and PMo12O403- and claimed that an oxygen transfer was taking
Figure 10. Rate of reaction as a function of the oxygen pressure.
Reaction conditions: 9 mM substrate and 1.0 mM H5PV2Mo10O40 in 2
mL of acetonitrile, 0.2-3 atm O2, 60 °C. Asterisk (*), anthracene; ^,
xanthene.
anthrancene and xanthene, respectively. Statistical T-test analysis
on the slopes showed that these values are, in effect, zero at
greater than 99% confidence limits. Thus, the reactions can be
concluded to be zero-order in molecular oxygen. These results
clearly allow the conclusion that the rate-determining step is
during the hydrocarbon oxidation by the H5PV2Mo10O40 and
not during the oxidative regeneration of the catalyst, Scheme
1. Notable also in this context is the observation that under
anaerobic and stoichiometric conditions, the rate of formation
of anthraquinone and xanthen-9-one is the same, within 5%, as
it is under catalytic aerobic conditions. Subsequent reoxidation
of the reduced catalyst is very fast (see also previous research
in this area19k). Reactions as a function of temperature and the
relevant Eyring or Arrhenius plots yielded the following
complete empirical rate equations for the oxidation of anthracene
(eq 4) and xanthene (eqn 5).
place and claimed observation and isolation of the deoxygenated
q-
[PMo12O40-z
]
(z ) 1, 2, 3) species.46 Just recently,47 the
oxygen-transfer reaction mechanism has been challenged, and
the identification of [PMo12O39]3- has been disputed and revised
to [HPMo12O40]4- 45b or [H2PMo12O40]3- 45a
Unfortunately, in
.
none of the reports has isotope labeling been used. Rather, the
supposed polyoxometalate species have been inferred from
changes in the 31P NMR and IR spectra and different electro-
chemical properties of the various species. It should also be
pointed out that oxygen transfer from various organic substrates,
for example, sulfoxides to oxygen-deficient polyoxometalates,
has been demonstrated.48
For the oxidation of hydrocarbons, the presence of vanadium
in the polyoxometalate was obligatory; thus, H5PV2Mo10O40 was
more reactive than H4PVMo11O40, and H3PMo10O40 showed no
-d[RH]
) kobs[RH]1[POM]1[O2]0
dt
∆Hq ) 19.0 kcal/mol, ∆Sq ) -54 cal/mol/K (4)
-d[RH]
(44) Streitwieser, A., Jr.; Jagow, R. H.; Fahey, R. C.; Suzuki, S. J. Am.
Chem. Soc. 1958, 80, 2326.
(45) Duncan, D. C.; Hill, C. L. J. Am. Chem. Soc. 1997, 119, 243.
(46) (a) Kawafune, I. Chem. Lett. 1986, 1503. (b) Kawafune, I.;
Matsubayashi, G. Chem. Lett. 1992, 1869. (c) Kawafune, I.; Matsubayashi,
G. E. Inorg Chim. Acta 1991, 188, 33. (d) Kawafune, I.; Matsubayashi, G.
E. Bull. Chem. Soc. Jpn. 1994, 67, 694.
(47) (a) Artero, V.; Proust, A. Eur. J. Inorg. Chem. 2000, 2393. (b) Neier,
R.; Trajanovski, C.; Mattes, R. J. Chem. Soc., Dalton Trans. 1995, 2521.
(48) Piepgrass, K.; Pope, M. T. J. Am. Chem. Soc. 1989, 111, 753.
) kobs[RH]1[POM]1[O2]0
dt
∆Hq ) 7.5 kcal/mol, ∆Sq ) -46 cal/mol/K (5)
Having determined that the rate-determining step was during
the hydrocarbon oxygenation rather than the oxidative regenera-
tion of the catalyst, we sought to look at deuterium isotope
effects. First, in a competitive reaction between anthracene (10

8540 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
Figure 12. 17O NMR spectra of 17O-labeled H5PV2Mo10O40 before
(bottom) and after (top) reaction with anthracene and 16O2.
There is also a less significant decrease in the relative intensity
of the peak associated with V ) 17O. Again, it is difficult to
draw an exact conclusion from the changes in the 17O NMR.
However, with some degree of confidence, one can say that
deoxygenation of the polyoxometalate in the hydrocarbon
oxidation step followed by reoxygenation by 16O2 and “scram-
bling” by excess enriched H217O or vice versa leads to a different
distribution of 17O in the polyoxometalate. This would mean
that the oxygen transfer from the originally enriched polyoxo-
metalate is site-specific. Since the presence of vanadium is
necessary, one may conclude that most likely, oxygen transfer
is from positions adjacent to vanadium in the polyoxometalate.
Despite these experiments, it is at this point still impossible to
resolve the conflicting scenarios mentioned above for polyoxo-
metalate reoxidation.
Summary and Comments on the Reaction Mechanism.
A mechanistic scheme for the oxidation of anthracene is given
is Scheme 8. The reaction begins with an electron transfer from
anthracene to the polyoxometalate. The electron transfer appears
to be outer sphere with preassociation between the polyoxo-
metalate and hydrocarbon. From the ESR spectrum of the radical
cation-reduced polyoxometalate, one may deduce that a stable
donor acceptor complex is formed. Next, there is oxygen transfer
from the polyoxometalate to the initially formed radical cation.
There is extensive evidence that the source of oxygen in the
product is neither from molecular oxygen nor from water. The
rate-determining step appears to involve a change in hybridiza-
tion at the 9(10) carbon from sp2 to sp3. In other words, as
indicated, the oxygen-transfer step is also rate-determining.
Recalling that anthraquinone was the only observed reaction
product, the following reaction step(s) is faster; thus, we can at
this point only postulate on possible routes to the formation of
anthraquinone. First, it is possible that anthraquinone is formed
directly from the observed intermediate. Second, it is possible
that initially, the first oxidation step yields 9-hydroxyanthra-
cene or the tautomeric anthrone, which is then oxidized very
fast to anthraquinone. In separate experiments, the oxidation
of anthrone, even at room temperature, is complete in a few
seconds.
Figure 11. IR spectra of 18O-labeled H5PV2Mo10O40 before and after
reaction with anthracene and 16O2. Top, H5PV2Mo10 O40; middle,
16
18
H5PV2Mo10
O
40
with 50% 18O at the terminal positions and 90% 18O
at the corner and edge positions; bottom, spectrum after reaction with
anthracene and 16O2 (10 mM H5PV2Mo10
in 1 mL of dry acteonitrile, 1 atm 16O2).
O
40
and 60 mM anthracene
18
catalytic activity. Under anaerobic conditions in reactions carried
out with H5PV2Mo10O‚34H2O as catalyst in CD3CN, the 31P
NMR showed no formation of new peaks, but others disappeared
or slightly shifted because of the effect of the paramagnetic VIV
centers.19k The polyoxometalate reoxidation step was also
investigated by measurement of the IR spectra of 18O-enriched
H5PV2Mo10O40 after catalytic reaction with anthracene under
16O2 (Figure 11). The peaks for terminal, corner-, and edge-
18
shared oxygen atoms in H5PV2Mo10
O at 919, 831, and 758
40
cm-1 are shifted relative to peaks in unlabeled H5PV2Mo10
O
40
16
at 960, 865, and 774 cm-1. After aerobic (16O2) oxidation of
18
anthracene (60 mM anthracene, 1 mM H5PV2Mo10
O (75%
40
18O) in CH3CN, 60 °C, 1 atm 16O2), a 69% incorporation of
18O into anthraquinone was observed, along with an increase
in the peak ratio of 960 (16O) vs 919 cm-1 (18O). Shifts from
831 to 843 cm-1 and 758 to 761 cm-1 were observed for the
corner- and edge-shared positions, respectively. Clearly, there
appears to be direct or indirect reincorporation of 16O from 16O2
into the polyoxometalate.
17O-enriched H5PV2Mo10O40 was also prepared by treating
16
10 µmol of dehydrated H5PV2Mo10
O
40
with 500 µmol of
H217O (10.2% 17O) in 1 mL of dry acetonitrile for 6 days to
obtain an equilibrated system. The 17O NMR of the solution
obtained (Figure 12 bottom) shows the incorporation of 17O in
the terminal, VdO (1290 ppm), and ModO (935 ppm)
positions, and in the bridging Mo-O-Mo, Mo-O-V, and
V-O-V positions (540-590 ppm). After addition of 100 µmol
of anthracene and reaction under 16O2 (55% conversion), two
major changes in the 17O NMR spectrum (Figure 12, top) are
observed. First is the appearance of a peak at 515 ppm assigned
to [17O]anthraquinone.49 Second is the different “fingerprint”
associated with the peaks for oxygen in the bridging positions.
The oxidation of xanthene occurs by a slightly different
pathway (Scheme 9). As with anthracene, the reaction begins
(49) Boykin, D. W.; Baumstark, A. G. Tetrahedron 1989, 45, 3615.

Electron, Oxygen Transfer in H5PV2Mo10O40
J. Am. Chem. Soc., Vol. 123, No. 35, 2001 8541
Scheme 8
Scheme 9
Scheme 10
with an electron transfer from xanthene to the polyoxometalate.
However, here the radical cation is not observed; rather, as
deduced from the H NMR spectrum, a stable and observable
compounds. We are presently engaged in developing this
catalytic oxidation for synthetic applications, which will be
described and published separately.
1
benzylic cation is formed via additional proton and electron
transfer. The kinetics experiments, coupled with the kinetic
isotope effect, indicate that the formation of the xanthenyl cation
is rate-determining. As in the case of anthracene, there is a wide
body of evidence that indicates that there is direct oxygen
transfer from the polyoxometalate to the activated xanthene,
rather than reaction with molecular oxygen or water. Xanthen-
9-ol is not observed, an indication that if formed, it is also very
quickly oxidized, presumably by another molecule of polyoxo-
metalate. Alternatively, in the presence of additional catalyst,
xanthen-9-one may be formed directly.
The regeneration of the original oxidized form of the catalyst
by reaction with molecular oxygen occurs very fast. Thus, we
do not have, at this time, any real information concerning this
process, because both 31P NMR and IR measurements show
the presence of complete Keggin structures formed by the
reaction of the proposed oxygen-deficient Keggin species with
either water or molecular oxygen, as summarized in Scheme
10.
Experimental Section
Instruments. 1H NMR (250 MHz) and 31P NMR (101.27 MHz, 85%
H3PO4 external standard) measurements were taken on a Bruker Avance
250 DPX instrument, and 17O NMR (54.25 MHz, H217O internal
standard) measurements were taken on a Bruker Avance 400 instrument.
HPLC analyses were performed on a Merck-Hitachi L6200 equipped
with a L4000 UV detector. The column used was a reverse-phase
Lichospher 100 RP-18 (5 µm) 25 × 4 mm column using 80% methanol/
20% water as eluent at a flow rate of 0.8 mL/min with detection at
242 nm. Retention times: anthracene, 18.1 min; anthraquinone, 9.6
min; xanthene, 16.2 min; xanthen-9-one, 11.7 min.
GC/MS measurements were carried out on an instrument consisting
of a HP 5973 mass-selective detector and a HP 6890 GC. A 5% phenyl
methylsilicone 0.32 mm i.d., 0.25-µm coating, 30-m-length column
(Restek 5MS). The eluent gas was helium.
IR measurements were carried out on a Nicolet Protege 460 FTIR
instrument. Solutions of polyoxometalate samples were placed on KBr
disks, and the solvent was allowed to evaporate. UV-vis spectra and
kinetic experiments were carried out on a HP 8452A diode array UV-
vis spectrometer equipped with a stirring apparatus and temperature
bath (( 0.1 °C). Cyclic voltammetry was carried out on a BAS-1
instrument. The ESR spectra were recorded on a Bruker ER200D-SRC
spectrometer (X-band).
Finally, a few words should be presented on the scope of the
oxygen-transfer reaction described herein. In this manuscript,
we have concentrated on understanding the mechanism of the
reaction for oxidation of both aromatic and alkyl aromatic

8542 J. Am. Chem. Soc., Vol. 123, No. 35, 2001
Khenkin et al.
Materials. The H5PV2Mo10O40‚34H2O polyoxometalate was prepared
using a known literature method.50 Thermogravimetric analysis (Mettler
50) indicated 34 water molecules per polyoxometalate. Anal. found
(calculated): P, 1.31 (1.34); V, 4.38 (4.41); Mo, 41.32 (41.56). IR:
(10.2%) in dry CH3CN and mixing for 6 days in order to obtain a fully
equilibrated compound.
Oxidation Reactions. Reactions under aerobic conditions were
carried out using 25-mL Ace glass pressure tubes. The tubes were
charged with the appropriate amounts of substrate, catalyst, and solvent
and were gently flushed with oxygen for 5 min. For experiments with
18O2, two freeze-pump-thaw cycles instead of flushing were used to
fill the reaction tube. The pressure was brought to 1 atm, and the
reactions were initiated by placing the reaction vessel in a thermostated
oil bath, usually at 60 ( 0.5 °C. At the desired time interval, 10-µL
aliquots were removed for HPLC analysis, or 1-µL aliquots were
removed for GC-MS analysis. In kinetic experiments, the reaction tube
was immediately repressurized. In general, HPLC was used for kinetics
experiments and determination of product yields. The sample aliquots
for HPLC analysis were diluted in the eluent, and for GC/MS analysis,
the aliquots were directly injected. 18O incorporation was measured
from the mass spectra by comparing peak intensities. For experiments
using anthracene, anthraquinone product peaks at m/z ) 208, 210, and
222 were compared. For xanthene, xanthen-9-one intensities of peaks
at m/z ) 196 and 198 were compared. For water peaks (retention time,
1.12 min) at m/z ) 18 and 20 were compared. Reactions under
anaerobic conditions were carried out either in 25-mL Ace glass
pressure tubes or directly in 5-mm NMR tubes equipped with a vacuum/
pressure valve. The tubes were charged with the appropriate amounts
of substrate, catalyst, and solvent and were gently flushed with argon
for 5 min. The tubes were then frozen, pumped down, filled with argon,
and then thawed. This procedure was repeated three times until final
pressurization to 1 atm argon. HPLC analysis was done as described
1057, 960, 865, 774 cm-1 31P NMR (CD3COCD3, 85% H3PO4 external
.
standard): -3.96 (6), -3.42 (1), -3.37 (4), -3.31(2), -3.22(2) ppm
(area of peak). Q5PV2Mo10O40 (Q ) (C4H9)4N) was prepared by mixing
10 equiv of QBr dissolved in water to H5PV2Mo10O40 also dissolved
in water. The precipitate formed was filtered and dried overnight in a
vacuum oven at 80 °C. The thermogravimetric analysis showed no water
was present. Anal. found (calculated): C, 32.09 (32.64); H, 5.99 (6.16);
N, 2.13 (2.38). Solvents and substrates available from commercial
sources were of the highest purity available and were used without
further purification. 18O2 (Enritech) was 96.1% 18O-labeled. H218O
(Rotem) was 94.3% 18O-labeled. H217O (D-Chem) was 10.2% 17O-
labeled. Dry acetonitrile (Biolab) contained < 20 ppm water.
Xanthene-9-d2. Xanthene-9-d2 was prepared by reduction of xanthen-
1
9-one with AlCl3-LiAlD4. H NMR and GC/MS was used to verify
the absence of benzylic hydrogen atoms and the presence of deuterium.
4-Alkyl Anisole. 4-Ethyl-, 4-n-propyl-, and 4-i-propylanisole were
prepared by reacting 0.1 mol 4-ethyl-, 4-n-propyl-, and 4-i-propylphenol
with 0.12 mol methyliodide and 0.12 mol potassium carbonate
(anhydrous) in 100 mL acetone at reflux for 24 h. After filtration, the
acetone solvent and excess methyliodide was removed by evaporation;
the product was purified over a silica gel column by elution with
chloroform (Rf ) 0.90).
2-Chloroanthracene. A mixture of 2-chloroanthraquinone (4.95 g,
0.02 mol), zinc powder (33 g), 1,4-dioxane (165 mL), and 28% NH4OH
(165 mL) were mixed under reflux for 3 h. The mixture was carefully
poured into 500 mL of water and extracted twice with 25 mL of 1,2-
dichloroethane. After removal of the 1,2-dichloroethane by vacuum
evaporation, the crude product was twice recrystallized from ethanol
to yield 1 g (23%) of a yellowish product (>99% purity by GC/MS).
mp, 218-219 °C. Anal. found (calculated): C, 78.22 (79.06); H, 4.48
1
above, and H NMR spectra were measured directly. Oxygen uptake
was measured by use of a gas buret. Kinetic isotope effect for oxidation
of anthracene/anthracene-d10 was measured in a competitive reaction
by reacting anthracene (10 mM), anthracene-d10 (10 mM), and H5PV2-
Mo10O40‚34H2O (1 mM) in acetonitrile at 60 °C and 1 atm O2. Peaks
were quantified by comparing molecular peak intensities of anthra-
quinone and anthraquinone-d8. The kinetic isotope effect for oxidation
of xanthene was studied by the comparison of reaction rates of xanthene
and xanthen-9-d2 oxidized separately. Thus, xanthene or xanthen-9-d2
(20 mM) and H5PV2Mo10O40‚34H2O (1 mM) in acetonitrile at 60 °C
and 1 atm O2 were reacted, and the rate of xanthen-9-one formation
was measured by HPLC using benzoic acid as calibration standard.
Kinetic Profiles of Interaction of Hydrocarbons with H5PV2-
Mo10O40. A solution of 1 µmol of H5PV2Mo10O40 in 3 mL of acetonitrile
was purged with argon for 20 min in a quartz cuvette and then brought
to 60 °C. The reaction was initiated by the injection via GC syringe of
20 µmol of hydrocarbon, and the reduction of the polyoxometalate was
followed at 750 nm in the UV-vis spectrometer. First-order rate
constants (r2 ) 0.99) were computed from the data.
1
(4.23); Cl, 16.65 (16.70). H NMR (CDCl3): 8.56 (d, 2H), 8.11 (m,
4H), 7.54 (m, 2H), 7.45 (dd, 1H).
2-Methoxyanthracene. 2-chloroanthraquinone (10.52 g, 0.043 mol)
was dissolved in a solution derived from 10 g of sodium metal in 100
mL methanol that was heated under reflux for 72 h. The solvent was
removed by distillation; water was added, and the mixture was
neutralized with dilute H2SO4. The solid 2-methoxyanthraquinone was
collected and crystallized from ethanol. Yield, 6.1 g, 59%. 2-methoxy-
anthraquinone (6.0 g, 0.025 mol), zinc powder (40 g), 1,4-dioxane (200
mL), and 28% NH4OH (200 mL) were mixed under reflux for 3 h.
The yield was 0.5 g (10%) with a purity of >98% by GC/MS after
treatment as described above. mp, 182-183 °C. Anal. found (calcu-
1
lated): C, 86.93 (86.54); H, 5.89 (5.77). H NMR (CDCl3): 8.35 (s,
1H), 8.27 (s, 1H), 7.93 (m, 2H), 7.89 (d, 1H), 7.40 (m, 2H), 7.20 (d,
1H), 7.16 p(dd, 1H), 3.92 (s, 3H).
Cyclic Voltammetry. The electrochemical potentials of xanthene
and xanthen-9-d2 were measured using a glass carbon cathode, a
platinum anode and calomel reference electrode. The measurements
were of solutions of 10 mM substrate in 100 mM tetrabutylammonium
tetrafluorborate in dichloromethane at 50 mV/sec; I/V ) 500 µA/V.
Electron Spin Resonance. Spectra were measured by preparing
solutions of 2 mM anthracene or xanthene and 4 mM Q5PV2Mo10O40
in 1,1,1,3,3,3-hexafluoro-2-propanol or 4 mM H5PV2Mo10O40 in
acetonitrile under argon and heating at 40 °C for the indicated time
period in a quartz capillary tube (1 mm, 200 µL). Spectra were recording
using microwave frequency, 9.78 GHz; microwave power, 12 mW;
field modulation, 320mG; receiver gain, 4 × 105; time constant, 320
ms; and scan time, 200 s. The Public EPR Software Tools developed
by D. Duling (NIEHS/NIH) were used for the computer simulation.
H5PV2Mo10 O40. 18O-labeled H5PV2Mo10O40 was prepared by drying
18
the polyoxometalate at 120 °C for 24 h and then adding 50 equiv H218O
(94.3%) in dry CH3CN and mixing for 18 h. The drying-exchange cycle
was repeated 3 times. 18O enrichment was estimated to be ∼50% at
the terminal oxygen atoms and >90% at the edge- and corner-shared
oxygen atoms. The enrichment and the bridging oxygen atoms were
estimated by deconvolution of the IR spectra assuming a theoretical
40 cm-1 isotope shift and expected equal peak intensities for labeled
and unlabeled oxygen. Total enrichment for potentially transferable
oxygen atoms, that is, excluding the four internal oxygen atoms, was
therefore ∼75%.
H5PV2Mo1017O40. 17O-labeled H5PV2Mo10O40 was prepared by drying
the polyoxometalate at 120 °C for 24 h and then adding 50 equiv H218O
Acknowledgment. This research was supported by the Basic
Research Foundation administered by the Israeli Academy of
Science and Humanities.
(50) Tsigdinos, G. A.; Hallada, C. J. Inorg. Chem. 1968, 7, 437.
(51) Parker, V. D. J. Am. Chem. Soc. 1976, 98, 98.
(52) Zhang, X.; Bordwell, F. G. J. Org. Chem. 1992, 57, 4163.
(53) Freccero, M.; Pratt, A.; Albini, A. Gong, C. J. Am. Chem. Soc. 1998,
120, 284.
JA004163Z