From surface-inspired oxovanadium silsesquioxane models to active catalysts for the oxidation of alcohols with O2-The cinnamic acid/ metavanadate system

Article abstract of DOI:10.1002/chem.201000171

Silsesquioxane dioxovanadate(V) complexes were investigated with respect to their potential as a catalyst for the oxidative dehydrogenation of alcohols with O₂ as an oxidant. The turnover frequencies determined were comparatively low, but during the oxidation of cinnamic alcohol an increase in activity was observed in the course of the process, which was inspected more closely. It turned out that during the oxidation of cinnamic alcohol, not only was the aldehyde formed but also cinnamic acid, which in turn reacts with the silsesquioxane complex employed to give NBu_4- [O₂V(O ₂CC₂H₂Ph)₂], which can also be obtained from NBu₄VO₃ and cinnamic acid and represents a far more active catalyst, not only for cinnamic alcohol but also for other activated alcohols and hydrocarbons. The rate-determining step of the conversion corresponds to an hydrogen-atom abstraction from the C-H units, as shown by the determination of the kinetic isotope effect in case of 9-hydroxyfluorene, and the reoxidation of the reduced catalyst proceeds via a peroxo intermediate, which is also capable of oxidizing one alcohol equivalent. Furthermore the influence of the organic residues at the carboxylate ligands on the catalyst performance was investigated, which showed that the activity increases with decreasing pK_s value. Moreover, it was found that during the oxidation the catalyst slowly decomposes, but can be regenerated by addition of excessive carboxylic acid.

Full text of DOI:10.1002/chem.201000171

DOI: 10.1002/chem.201000171
From Surface-Inspired Oxovanadium Silsesquioxane Models to Active
Catalysts for the Oxidation of Alcohols with O2—The Cinnamic Acid/
Metavanadate System
Christian Ohde and Christian Limberg*[a]
Abstract: Silsesquioxane dioxovanada-
te(V) complexes were investigated
with respect to their potential as a cat-
alyst for the oxidative dehydrogenation
of alcohols with O2 as an oxidant. The
turnover frequencies determined were
comparatively low, but during the oxi-
dation of cinnamic alcohol an increase
in activity was observed in the course
of the process, which was inspected
more closely. It turned out that during
the oxidation of cinnamic alcohol, not
only was the aldehyde formed but also
cinnamic acid, which in turn reacts
with the silsesquioxane complex em-
ployed
AHCTUNGTRENNUNG
to
give
NBu4-
mination of the kinetic isotope effect
in case of 9-hydroxyfluorene, and the
reoxidation of the reduced catalyst pro-
ceeds via a peroxo intermediate, which
is also capable of oxidizing one alcohol
equivalent. Furthermore the influence
of the organic residues at the carboxyl-
ate ligands on the catalyst performance
was investigated, which showed that
the activity increases with decreasing
pKs value. Moreover, it was found that
during the oxidation the catalyst slowly
decomposes, but can be regenerated by
addition of excessive carboxylic acid.
obtained from NBu4VO3 and cinnamic
acid and represents a far more active
catalyst, not only for cinnamic alcohol
but also for other activated alcohols
and hydrocarbons. The rate-determin-
ing step of the conversion corresponds
to an hydrogen-atom abstraction from
ꢀ
the C H units, as shown by the deter-
Keywords: alcohols · homogeneous
catalysis · oxidation · silsesquiox-
ane · vanadium
Introduction
ture, function, or the spectroscopic signatures of correspond-
ing surface species. The construction of molecular models
for metal–oxo surface sites requires ligands that are capable
of mimicking the environments that metaloxo moieties ex-
perience on oxidic support or bulk materials, and for the
modeling of silica supports in particular silsesquioxane-
based ligands have established themselves.[5,6] Silsesquiox-
anes in general exhibit one-dimensionally extended or cage
structures that are saturated by organic residues. Some of
these, for example, the structure of the octamer RT8, resem-
bles skeletal frameworks found in crystalline forms of silica,
The oxidative dehydrogenation (ODH) of light alkanes and
methanol is currently a matter of intense research, especially
as the resulting products—alkenes and formaldehyde—are
important bulk chemicals in industry, and supported vanadi-
um oxides have proved to be among the best catalysts for
these processes.[1] The nature of the active sites and units on
the surfaces of these catalysts are still discussed controver-
sially,[2] though, and many different approaches have been
pursued to reveal new insights. Naturally, the results of cor-
responding studies can also provide a stimulus for the syn-
thesis of molecular model compounds.[3,4] These can then be
investigated in the homogeneous phase and may serve to
support or disprove proposals made concerning the struc-
and silsesquioxanes in general are therefore discussed as
3+
ꢀ
molecular sections of silica. If formally one of the R Si
R
corners is removed from the T8 cage, protonation of the re-
[a] C. Ohde, Prof. Dr. C. Limberg
Institut fꢀr Chemie, Humboldt-Universitꢁt zu Berlin
Brook-Taylor Str. 2, 12489 Berlin (Germany)
Fax : (+49)30-2093-6966
E-mail: christian.limberg@chemie.hu-berlin.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000171.
6892
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 6892 – 6899

FULL PAPER
R
sulting siloxide units yields in compound H3 T7, which pos-
sesses both structural and electronic similarities to hydroxyl-
ated silica surface sites.[7]
Correspondingly, with such compounds, for instance, pro-
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
cesses occurring during the grafting of molecular precursors
on silica surfaces can be investigated on the molecular
R
level.[7e,8,9] If on the other hand the protons of H3 T7 are re-
placed by transition-metal cations, the resulting complexes
simulate situations in which such cations are bound to silica
supports, but have the advantage of being treatable with the
methods of molecular chemistry.[5–7] A second ligand system
that is often employed for the more general modeling of an
oxidic coordination platform are calixarenes,[10] and in the
context of the above-mentioned supported oxovanadium
catalysts employed for ODH processes, we have recently re-
ported the results of structure–function analyses performed
for calixarene-based model compounds with respect to the
ODH of alcohols.[3,11] Furthermore we have utilized oxova-
nadium silsesquioxane complexes as spectroscopic models.[9]
In the next step those were to be tested as functional
models, especially, since—due to the more exact mimicking
of surface sites in comparison to the calixarene complexes—
a higher reactivity was expected. Here we report the find-
ings made in the course of corresponding ODH studies: Ser-
endipitously they led to a more simple system, which proved
to represent an efficient catalyst for the ODH of activated
alcohols.
Scheme 1. Formation of 1 (R=c-C5H9, Rꢃ=SiMePh2).
Figure 1. Molecular structure of the anion of 1. Selected bond lengths [ꢄ]
ꢀ
ꢀ
ꢀ
ꢀ
and angles [8]: V O1 1.626(6), V O2 1.609(5), V O3 1.819(5), V O4
ꢀ
ꢀ
1.838(5), Si1 O3 1.600(6), Si2 O4 1.594(5); O1-V-O2 108.0(3), O1-V-O3
109.7(3), O2-V-O4 110.3(3), O3-V-O4 109.0(2).
Results and Discussion
As substrates we chose aliphatic alcohols, activated alco-
hols and activated hydrocarbons, which were dissolved in
CH3CN (2 mL) and heated to 808C in the presence of mo-
lecular sieves and catalytic amounts (1 mol%) of the com-
plexes I, II, or 1, respectively, in an O2/Ar atmosphere.
The investigation was started by testing the complexes I and
II[9] as potential catalysts for the ODH of alcohols. More-
over, we were interested to complement these complexes by
1
After 3 h the reaction mixtures were analyzed by H NMR
spectroscopy, and the yields were determined by integration
and comparison of specific signals originating from sub-
strates and oxidation products. In case of a similar study per-
formed for the above-mentioned calixarene models the
highest activities had been observed for 9-hydroxyfluorene,
cinnamic alcohol, and benzylic alcohol (in this order),[3,11b]
and for a given catalyst a correlation between the TOFs and
ꢀ
the strengths of the alcoholic C H bonds had become evi-
dent; hence, hydrogen-atom abstraction has been proposed
as a rate-determining step, and this hypothesis has gained
support from the isolation of a catalysis intermediate. In
contrast, employing the silsesquioxane models I, II. and 1,
the highest activity was observed for cinnamic alcohol in
case of all three complexes (see Figure 2), which seemed to
point to a different type of mechanism.
However, the results proved difficult to reproduce (a fur-
ther difference to the calixarene chemistry), so that the
course of the product formation was inspected more closely
for the case of cinnamic alcohol. This showed that the cata-
lytic process required an induction period to develop its full
activity: If a solution of cinnamic alcohol in acetonitrile is
warmed to 808C in the presence of molecular sieves and
a derivative containing a larger siloxide dangling group in
order to study the influence of its size on the reactivity of
the metal site. Synthesis of the OSiMePh2 analogue 1 (see
Scheme 1) was achieved according to the route established
for the synthesis of I, but reacting the corresponding ligand
H2ACHTUNGTRENNUNG ACHTUNTGERN[NUGN VO2Cl2]. This led to the ex-
(Ph2MeSi)c-C H T7 with [PPh4]
5
9
+
pected replacement of the two protons by the VO2 group.
Its two V=O bond lengths and O-V-O angles within 1 are
almost identical to those observed for the other two com-
pounds, as revealed by a single-crystal X-ray analysis (see
Figure 1).
Chem. Eur. J. 2010, 16, 6892 – 6899
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6893

C. Limberg and C. Ohde
II. However, the above-mentioned acceleration of catalysis
after an approximate 10% oxidation of the alcohol to the al-
dehyde proved reproducible, and thus required an explana-
tion. Cinnamic aldehyde is known to undergo autoxidation
in air,[12] which raised the idea that beyond a certain critical
concentration of cinnamic aldehyde, such a process could be
triggered leading to a fast conversion of the alcohol via radi-
cal chains. To test whether such autoxidation occurs under
the conditions chosen here, experiments were performed in
the presence of radical starters, which did not lead to any al-
terations. Furthermore it turned out that the presence of ad-
equate amounts of the aldehyde right from the start
(10 mol%) neither made the vanadium component redun-
dant nor eliminated the induction period. We noted that
part of the cinnamic aldehyde formed in the course of the
oxidation of cinnamic alcohol was further oxidized to give
cinnamic acid; hence the influence of this product on the
catalytic performance was tested in the next step. Indeed,
after the addition of 5 mol% cinnamic acid to the system
with only 0.25 mol% II as catalyst no induction period was
required anymore for an efficient catalysis (TOF=250 hꢀ1,
see Figure 4).
Figure 2. TOF data for the oxidation of alcohols with I, II and 1 as cata-
lysts: Alcohol (1 mmol), catalyst (0.01 mmol) and molecular sieves (0.5 g,
3 ꢄ) in CH3CN (2 mL) were heated to 808C in an O2/Ar atmosphere.
After 3 h the reaction mixtures were analyzed by 1H NMR spectroscopy.
1 mol% of II in an O2/Ar atmosphere, the reaction proceeds
sluggishly within the first 1.3 h. Only after the formation of
about 10% of cinnamic aldehyde was a steeper increase of
the product yield observed (see Figure 3).
Figure 4. Yield versus time plot for the oxidation of cinnamic alcohol in
the presence of cinnamic acid (5 mol%) and II (0.25 mol%) as the cata-
lyst.
Figure 3. Yield versus time plot for the following reaction conditions: cin-
namic alcohol (2 mmol), II (0.02 mmol) and molecular sieves (1 g, 3 ꢄ)
in CH3CN (4 mL) were heated to 808C in an O2/Ar atmosphere.
Hence, now the reaction of II with cinnamic acid was in-
vestigated. On treatment of
a
solution of II in
[D3]acetonitrile with ten equivalents of cinnamic acid, the
signal initially observed for II at ꢀ587 ppm disappears in
favor of a new signal at ꢀ513 ppm, which can also be ob-
served if NBu4VO3 is reacted with cinnamic acid. Accord-
ingly, the presence of cinnamic acid leads to the detachment
This indicated that II as such is only slightly active, but
undergoes a transformation (which needs the aldehyde
product) to a more efficient catalyst. When 1 or I were em-
ployed as catalysts the yield/time plots looked nearly identi-
cal, but II was more active, so that all further investigations
were performed with II.
+
of the silsesquioxane ligand from the VO2 unit (i.e., to
leaching), and—bearing in mind that there are some exam-
+
ples of compounds containing VO2 units bound to two car-
boxylate functions as part of a chelating ligand known in the
literature[13]—the formation of NBu
4ACHTUNGTRNENUG[O2VAHCTUNGTRNEN(UGN O2CR)2] (R=
The active catalyst: When II was treated with cinnamic alco-
hol for 6 h under the conditions described above, but in the
absence of O2, no reaction took place, which showed that
the alcohol itself does not play a role in the conversion of
C2H2Ph) (2) seemed likely (Scheme 2), which would then
correspond to the active catalyst.[14]
This hypothesis was confirmed by isolation and character-
ization of this product from a corresponding conversion on
6894
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 6892 – 6899

Homogeneous Catalysis
FULL PAPER
Scheme 2. Formation of 2 (R=C2H2Ph).
the synthetic scale. The result of a single-crystal X-ray anal-
ysis is shown in Figure 5.
Figure 6. TOF data for the oxidation of alcohols with several catalysts:
Alcohol (2 mmol), catalyst (x mmol; x=0.02 for I, II, 1 and VO2acac2;
x=0.005 for 2) and molecular sieves (1 g, 3 ꢄ) in CH3CN (4 mL) were
heated to 808C in an O2/Ar atmosphere.
Figure 5. Molecular structure of the anion of 2. Selected bond lengths [ꢄ]
ꢀ
ꢀ
ꢀ
ꢀ
and angles [8]: V O1 1.618(4), V O2 1.614(5), V O3 1.985(5), V O4
ꢀ
ꢀ
2.380(5), V O5 1.998(4), V O6 2.314(4); O1-V-O2 105.7(3), O1-V-O6
153.2(2), O2-V-O4 156.0(2), O1-V-O5 97.3(2), O2-V-O3 99.7(3).
The oxygen atoms, which coordinate the vanadium atom
altogether form a distorted octahedron. The V=O bond
lengths are almost identical to those observed for 1. The car-
ꢀ
boxylate ligands bind in a bidentate fashion, but the V O4
ꢀ
(2.380(5) ꢄ) and the V O6 bonds (2.314(4) ꢄ) are signifi-
cantly longer than the distances of V to O3 and O5.
Indeed 2 turned out to be a much more effective ODH
catalyst in comparison to II, not only in case of cinnamic al-
cohol, but also for 9-hydroxyfluorene, benzyl alcohol, cin-
namic alcohol, crotyl alcohol, 1-phenyl-1-progargyl-alcohol,
1-phenyl-1-propanol and 9,10-dihydroanthracene (TOFs=
336, 35, 220, 12, 164, 25, and 64 hꢀ1, respectively; 9,10-dihy-
droanthracene was mainly oxidized to anthraquinone, 9,10-
dihydroynthracene-9,10-diol, and anthracene), and it is
worth noting that 2 can also be prepared in-situ by addition
of NBu4VO3 and cinnamic acid in a ratio of 1:2 to the reac-
tions. The TOFs reached for the above-mentioned series of
alcohols are even higher than those observed in case of the
most effective oxovanadium calixarene complexes, which in
Figure 7. Yield versus time plot for the following reaction conditions: 9-
hydroxyfluorene (2 mmol), 2 (5 mmol) and molecular sieves (1 g, 3 ꢄ) in
CH3CN (4 mL) were heated to 808C in an O2/Ar atmosphere; after
40 min six equivalents of cinnamic acid were added.
Consequently, if sufficient amounts of cinnamic acid are
added straight away, this guarantees continuous self-healing,
so that, for instance, 2 mmol 9-hydroxyfluorene can be
quantitatively oxidized to fluorenone in the presence of
0.25 mol% of NBu4VO3 and 5 mol% of cinnamic acid
within 2 h (it might be worth pointing out that fluorenone is
an important compound that is utilized as an element in or-
ganic solar cells and display devices, and also belongs to an
interesting class of compounds for biomedical applica-
tions).[16] Naturally the question arose whether cinnamic
acid is unique within this system or whether it can be re-
placed by other carboxylic acids, so that perhaps the cata-
lysts robustness or activity can be increased even further. A
corresponding investigation showed that indeed other acids
can be employed, but only a small range of pKs values (4–5)
comes into question: If the acids employed are too acidic,
they induce the formation of polyoxovanadates, and at pKs
turn had been superior to the reference system OVACTHUNRGTNEUNG(acac)2,
reported to represent an efficient catalyst for the aerobic ox-
idation of 1-phenyl propargylic alcohol (see Figure 6).[15]
Unfortunately, 2 is not indefinitely robust under the
chosen conditions and looses nearly all activity, for instance
within 40 min of catalytic oxidation of 9-hydroxyfluorene,
possibly due to CO2 elimination or C=C bond cleavage at
the carboxylate ligand. It can be regenerated, though,
through the addition of further equivalents of cinnamic acid,
as shown in Figure 7.
Chem. Eur. J. 2010, 16, 6892 – 6899
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6895

C. Limberg and C. Ohde
values above 5 the resulting catalysts are significantly less
active than 2 (the pKs value of cinnamic acid amounts to
4.44). Hence, acetic acid (pKs =4.76) and pivalic acid (pKs =
5.05) were successful candidates: An activity similar to the
one of cinnamic acid was reached with acetic acid, and pi-
Scheme 3. Illustration of the alcohol oxidation with VV via VIV.
ACHTUNGTRENNUNGvalACHTUNGTRENNUNGic acid only led to a slight decrease in activity, while the
resulting system showed an enhanced stability against water
(note that catalysis typically requires the presence of molec-
ular sieves to immediately remove water where formed and
As observed for the oxovanadium calixarene complexes,
the turnovers reached by 2 seamed to correlate with the
ꢀ
strengths of the C H bonds inherent to the alcohols em-
NBu
G
N
ployed (a strength of 350 kJmolꢀ1 still leads to acceptable
conversions), and in line with that a kinetic isotope effect of
molecular sieves). However, these two acids did not signifi-
cantly improve the long term stability, so that in the case of
2, probably decarbonylation and not reactions of the C=C
unit antagonize it.
3.9 (Figure 9) was observed for 9-deutero-9-hydroxyflu
ACHTUNGTRENNUNGo-
AHCTUNGTREGrNNNU ene: Considering that the maximum isotope effect reacha-
ble at 808C amounts to about 5, the value of 3.9 is certainly
indicative of a rate-determining step with significant in-
ꢀ
The catalytic cycle: As mentioned above, oxovanadium cal-
ixarene complexes were also found to be active in the cata-
lytic oxidation of alcohols. For these complexes we observed
that the oxidation of one alcohol molecule can be achieved
by one catalyst molecule (via VIII), but is preferably accom-
plished by two cooperating VV centers (via VIV). To check
whether the catalytic cycle of 2 involves one or two mole-
cules of 2, kinetic studies with varying catalyst concentra-
tions were performed, which unfortunately did not produce
unequivocal results. Therefore we decided to investigate the
oxidation state of the reduced catalyst. A mixture of 9-hy-
droxyfluorene, 2 (2 mol%), and excess molecular sieves in
CH3CN was heated to 808C and stirred for 10 min. After
cooling to room temperature an EPR measurement of the
filtrated solution at 77 K was performed (see Figure 8).
volvement of the C H bond, and a quite symmetric transi-
tion state along the hydrogen-transfer coordinate, O···H···C.
This finding is also consistent with a recent investigation
concerning the hydrogen-atom abstraction reactivity of cat-
ionic [O2VL2]+ complexes.[18]
Figure 9. Logarithmic plot of the oxidation of 9-hydroxyfluorene and 9-
deutero-9-hydroxyfluorene with catalyst 2 [alcohol (2 mmol), 2 (5 mmol)
and molecular sieves (1 g, 3 ꢄ) in CH3CN (4 mL) were heated to 808C in
an O2/Ar atmosphere]. Dashed line: best fit for 9-deutero-9-hydroxy-
fluorene, solid line: best fit for 9-hydroxyfluorene.
Combining these results suggests an initial hydrogen-atom
abstraction to yield a VIV species and an alkyl radical that
undergoes a single-electron oxidation (combined with a
proton abstraction) with a second equivalent of 2 to yield
the carbonyl compound. This poses the question, how is the
VIV species formed reoxidized by contact with O2? As water
is generated in this process, in-situ NMR studies to address
this point (that had to be carried out in the absence of mo-
lecular sieves) were performed with the more robust NBu4-
Figure 8. EPR spectrum of the reduced catalyst in CH3CN at 77 K and its
simulation. Simulation parameters:
Axx =190.0 G, Ayy =68.0 G, Azz =
68.0 G, gx =1.940, gy =1.975, gz =1.975, linewidth: x: 8 G, y: 10 G, z: 10 G.
ACHUTNGRENU[NG O2VACHTUNGTRNE(NUGN O2CR)2] system with R=tBu (2’; vide supra): 2’
The EPR spectrum unambiguously proves the presence of
a mononuclear VIV species after alcohol oxidation and thus
points to a mechanism featuring two molecules of 2 that per-
form single-electron oxidations (Scheme 3). Subsequently
the rate-determining step was further investigated.
(5 mmol) was dissolved in CH3CN (0.5 mL) and CD3CN
(0.1 mL) and treated with 9-hydroxyfluorene (50 equiv) at
808C to fully reduce it (20red). Then O2 was added at ꢀ108C
and a 51 V NMR spectrum was recorded, which is shown in
Figure 10.
6896
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 6892 – 6899

Homogeneous Catalysis
FULL PAPER
are reoxidized by O2 to give a 1:1 mixture of 2 and 2peroxo
,
which quickly oxidizes a second equivalent of alcohol to
yield back 2, thus closing the catalytic cycle.
This mechanistic proposal is plausible, as it accounts for
all four oxidation equivalents of O2, and at the same time it
is supported by 1) the EPR investigation of the reduced cat-
alyst, 2) the determination of the kinetic isotope effect for
the substrate 9-hydroxyfluorene, 3) the identification of
2peroxo as an intermediate, and 4) the investigation of the re-
activity of 2peroxo in contact with alcohols.
Figure 10. 51V NMR spectrum after reoxidation of the reduced catalyst
(R=CMe3).
Conclusion
In addition to a signal for 2’, a second one at ꢀ501 ppm
Our results show that the Si-O-V moieties of oxovanadium
silsesquioxane complexes are sensitive to an attack by car-
was observed (integral ratio of 1.3:1), which could be as-
ꢀ
signed to the peroxo compound NBu
4A
boxylic acids and get cleaved to yield Si OH functions and
A
vanadyl carboxylates. These findings should be transferable
to heterogeneous processes in which the over-oxidation of
hydrocarbons at silica-supported vanadiumoxide catalysts
produces carboxylic acids, that is, these should lead to leach-
ing. In the course of corresponding studies we were also
able to develop an efficient vanadium-based catalyst system
dependently from 2’ and H2O2 (see the Supporting Informa-
tion). Such kind of reactivity in contact with O2 has been
known for vanadium(IV) complexes since 1999.[17] In a fur-
ther NMR experiment it could be shown, that 20peroxo reacts
with an equimolar amount of 9-hydroxyfluorene to give flu-
orenone and 2’. This might be the pathway, by which the
catalytic cycle is closed.
Altogether, the following mechanism (see Scheme 4)
could explain these fresults: Doubtlessly, the rate-determin-
ing step is an hydrogen-atom transfer from the alcohol to
([O2VACHTNUGRTNEUNG
(O2CR)2]ꢀ salts in combination with HO2CR) for the
ODH of activated alcohols, which can be readily prepared
in-situ starting from easily accessible starting materials.
They react by means of a hydrogen-atom transfer from an
ꢀ
alcoholic C H unit to a V=O group in the rate-determining
step, and O2 is needed to reoxidize the resulting reduced va-
nadium(IV) species. This leads, inter alia, to a peroxide in-
termediate which is capable of oxidizing a second alcohol
equivalent.
Experimental Section
General remarks: All manipulations were carried out in a glove-box, or
else by means of Schlenk-type techniques involving the use of a dry
1
argon atmosphere. The H, 13C, 29Si, and 51 V NMR spectra were recorded
on a Bruker AV 400 NMR or DPX-300 spectrometer at 208C. The
1H NMR spectra were calibrated against the residual proton and natural
abundance 13C resonances of the deuterated solvent, the 29Si NMR spec-
tra against TMS, and the 51 V NMR spectra against VOCl3 as standards.
Coupling constants are given in Hz. Microanalyses were performed on a
Leco CHNS-932 or HEKAtech Euro EA 3000 elemental analyzer. Infra-
red (IR) spectra were recorded using samples prepared as KBr pellets
with a Shimadzu FTIR 8400S FTIR-spectrometer. ESR spectra were
measured on ERS 300 (ZWG/Magnettech GmbH, Berlin-Adlershof, Ger-
+
Scheme 4. Proposed mechanism (NBu4 ions have been omitted).
one of the terminal oxo atoms of 2,[18] employed as such or
prepared in in-situ. Perhaps prior to this step an addition of
the alcohol across a V=O bond of a second molecule of 2
occurs; such additions are often suggested in the literature
as initial, fast steps.[15,19] In the latter case a VV species with
a coordinated alkoxy radical would be formed that would
collapse into the carbonyl compound and NBu4[(O)(OH)V-
many). Pure [NBu4]
AHCTUNGRTEN(NGUN OSiPh2Me)(OH)2
G
[VO2Cl2],[21] and (C5H9)7Si7O9-
ACHTUNGTRENNUNG
[7e]
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
After addition of Et3N (1.0 mL, 7.2 mmol) the mixture was stirred for
15 h at room temperature. All volatiles were removed and the resulting
solid was extracted with THF (20 mL). Removing of all volatiles from
the extract afforded a grey powder. Recrystallization from a THF/n-
hexane mixture afforded 1 (0.24 g, 0.16 mmol; 43%) in form of colorless
crystals. 1H NMR (400 MHz, CDCl3, 258C, TMS): d=0.83 (s, 3H;
SiCH3), 0.77–1.09 (m, 7H; CH), 1.27–1.79 (brm, 56H; CH2), 7.20–7.28
(m, 6H; SiPh), 7.55–7.64 (brm, 12H; SiPh/PPh4+), 7.70–7.78 (m, 8H;
ACHTUNGTRENNUNG(O2CR)2] (2red); in the former case a free alcohol radical
would form that could be oxidized by a second equivalent
of 2. In both cases two equivalents of 2 react, one after an-
other, with one equivalent of alcohol to give the correspond-
ing carbonyl compound and two equivalents of 2red. These
Chem. Eur. J. 2010, 16, 6892 – 6899
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6897

C. Limberg and C. Ohde
SiPh/PPh4+), 7.79–7.87 ppm (m, 4H; PPh4+); 13C{1H} NMR (100 MHz,
b=22.3975(8), c=20.3594(10) ꢄ, a=b=g=908, V=8253.0(6) ꢄ3, Z=4,
1calcd =1.338 Mgmꢀ3, m=0.444 mmꢀ1; 53885 reflections measured, 10991
unique (Rint =0.0881); final R indices [I>2s(I)] R=0.0648, wR=0.1435
CDCl3, 258C, TMS): d=ꢀ0.5 (s, SiCH3), 22.6, 22.7, 23.6, 23.9, 24.5 (s,
1
CH), 27.0, 27.0, 27.1, 27.1, 27.4, 27.6, 27.6, 27.9 (CH2), 117.5 (d, J
A
90.1 Hz, PPh4+), 127.4 (s, SiPh), 128.8 (s, SiPh), 130.7 (d, 2J
G
Crystal data for 2: C34H50NO6V, Mr =619.69, T=100(2) K, l=0.71073 ꢄ,
monoclinic, space group P21/c, a=9.8460(8), b=9.9584(4), c=
3
13.4 Hz, PPh4+), 134.4 (s, SiPh), 134.4 (d, J
G
ACHTUNGTRENNUNG
(d, 4J(C,P)=3.6 Hz, PPh4+), 138.5 ppm (s, SiPh); 29Si{1H} NMR
33.969(2) ꢄ, a=90, b=93.604(6), g=908, V=3324.1(4) ꢄ3, Z=4, 1calcd
=
1.238 Mgmꢀ3, m=0.341 mmꢀ1; 33527 reflections measured, 4974 unique
(Rint =0.1434); final R indices [I>2s(I)] R=0.0911, wR=0.2491.
(79.5 MHz, CDCl3, 258C, TMS): d=ꢀ69.0, ꢀ66.6, ꢀ66.5, ꢀ65.7, ꢀ62.8
(s), ꢀ10.5 ppm (s, SiMePh2); 51V{1H} NMR (105 MHz, CDCl3, 258C,
VOCl3): d=ꢀ574 ppm; IR (KBr): n˜ =3058 (vw), 2948 (s), 2863 (m), 1437
(w), 1245 (vw), 1107 (vs), 1035 (m), 998 (sh), 947 (m), 935 (s), 756 (w),
722 (w), 526 (m), 488 cmꢀ1 (m); elemental analysis calcd (%) for
C72H96O14Si8V (1492.1): C 57.96, H 6.48; found: C 57.34, H 6.59.
Synthesis of 2: [NBu4]ACHTUNGTRENNUNG[VO3] (1.00 g, 2.96 mmol) was dissolved in CH2Cl2
(10 mL) and molecular sieves (3 ꢄ, 2.30 g) were added. Subsequently cin-
namic acid (0.88 g, 5.94 mmol) dissolved in CH2Cl2 (10 mL) was added
dropwise. After stirring of the mixture for 20 min at room temperature it
was filtrated. The solution was concentrated under reduced pressure to a
volume of 5 mL. Et2O (25 mL) was added to the resulting yellow solution
and the mixture was stirred for 30 min. The mixture was filtrated and the
beige residue was washed with Et2O (15 mL) and dried under high
vacuum over night. This yielded 2 (1.49 g, 2.40 mmol, 81%) in the form
Acknowledgements
We are grateful to the Deutsche Forschungsgemeinschaft, the CRC546,
the Fonds der Chemischen Industrie, and the BMBF for financial sup-
port. We thank C. Knispel and E. Hoppe for X-ray determination and C.
Jankowski for the preparation of starting materials. We also thank P.
Klꢁring, C. Lehmann, R. Schiwon, F. Pfaff and N. Manicke for their very
ambitious work in our laboratory.
1
of a beige powder. H NMR (300 MHz, CD3CN, 258C, TMS): d=0.95 (t,
3J
ACHTUNGTRENNUNG(H,H)=7.4 Hz, 12H; CH3), 1.34 (m, 8H; CH2-CH3), 1.59 (m, 8H; CH2-
CH2N), 3.09 (m, 8H; NCH2), 6.51 (d, 3J
ACTHNUTRGNE(NUG H,H)=15.9 Hz, 2H; CH), 7.35–
[1] a) A. Comite, A. Sorrentino, G. Capannelli, M. Di Serio, R. Tesser,
E. Santacesaria, J. Mol. Catal. A 2003, 198, 151–165; b) R. Zhou, Y.
Cao, S. Yan, J. Deng, Y. Liao, B. Hong, Catal. Lett. 2001, 75, 107–
112; c) T. Blasco, A. Galli, J. M. Lopez Nieto, F. Trifido, J. Catal.
1997, 169, 203–211; d) M. A. Chaar, D. Patel, H. H. Kung, J. Catal.
1988, 109, 463–467; e) A. Corma, J. M. Lopez Nieto, N. Paredes, J.
Catal. 1993, 144, 425–438; f) F. D. Hardcastle, I. E. Wachs, J. Mol.
Catal. 1988, 46, 173–186; g) G. Deo, I. E. Wachs, J. Haber, Crit. Rev.
Surf. Chem. 1994, 4, 141–187; h) I. E. Wachs, B. M. Weckhuysen,
Appl. Catal. A 1997, 157, 67–90; i) J. G. Eon, R. Olier, J. C. Volta, J.
Catal. 1994, 145, 318–326; j) T. Balsko, J. M. Lopez Nieto, Appl.
Catal. A 1997, 157, 117–142; k) A. Khodakov, B. Olthof, A. T. Bell,
E. Iglesia, J. Catal. 1999, 181, 205–216; l) A. Khodakov, J. Yang, S.
Su, E. Iglesia, A. T. Bell, J. Catal. 1998, 177, 343–351; m) E. V. Kon-
dratenko, M. Baerns, Appl. Catal. A 2001, 222, 133–143; n) M. D.
Argyle, K. Chen, E. Iglesia, A. T. Bell, J. Catal. 2002, 208, 139–149;
o) L. Owens, H. H. Kung, J. Catal. 1993, 144, 202–213; p) O. R.
Evans, A. T. Bell, T. D. Tilley, J. Catal. 2004, 226, 292–300.
[2] A classification of the literature according to the suggested active
oxo species is reviewed in: H. Fu, Z.-P. Liu, Z.-H. Li, W.-N. Wang,
K.-N. Fan, J. Am. Chem. Soc. 2006, 128, 11114–11123.
7.45 (m, 6H; CarH), 7.57–7.69 ppm (m, 6H; 4 CarH, 2 CH); 13C{1H} NMR
(75.5 MHz, CD3CN, 258C, TMS): d=13.7 (s, CH3), 20.3 (s, CH2), 24.2 (s,
CH2), 59.2 (t, 1J
ACHTUNGTRENNUNG(N,C)=2.8 Hz, NCH2), 121.6 (s, CH), 128.9 (s, 2 Car),
129.7 (s, 2 Car), 130.8 (s, Car), 135.9 (s, Car), 144.2 (s, CH), 177.3 ppm (s,
CO2); 51V{1H} NMR (105 MHz, CD3CN, 258C, VOCl3): d=ꢀ511 ppm;
IR (KBr): n˜ =2955 (m), 2930 (m), 2872 (m), 1707 (vw), 1641 (vs), 1596
(w), 1576 (s), 1551 (w), 1496 (w), 1487 (m), 1448 (w), 1382 (s), 1285 (vw),
1262 (vw), 1231 (m), 1179 (vw), 1153 (vw), 1109 (vw), 1073 (vw), 1027
(vw), 978 (m), 946 (s), 931 (vs), 886 (vw), 874 (vw), 868 (vw), 804 (vw),
777 (m), 745 (m), 722 (w), 687 (w), 620 (vw), 610 (w), 593 (vw), 490 (vw),
441 cmꢀ1 (w); elemental analysis calcd (%) for C34H50NO6V (619.7): C
65.90, H 8.13, N 2.26; found: C 66.05, H 8.10, N 1.98.
Synthesis of 2’: [NBu4]
(10 mL) and molecular sieves (3 ꢄ, 2.00 g) were added. Subsequently pi-
valic acid (0.60 g, 5.91 mmol) dissolved in CH2Cl2 (10 mL) was added
ACHTUNGTRENNUNG[VO3] (1.00 g, 2.96 mmol) was dissolved in CH2Cl2
A
ACHTUNGTRENNUNG
dropwise. After stirring of the mixture for 20 min at room temperature it
was filtrated. All volatiles were removed under reduced pressure, Et2O
(25 mL) was added, and the mixture was stirred for 30 min. It was then
filtrated, and the beige residue was dried under high vacuum over night.
This yielded 2’ (1.28 g, 2.43 mmol, 82%) in the form of a beige powder.
[3] E. Hoppe, C. Limberg, B. Ziemer, Inorg. Chem. 2006, 45, 8308–
8317.
[4] J. Canny, R. Thouvenot, A. Tꢅzꢅ, G. Hervꢅ, M. Leparulo-Loftus,
Inorg. Chem. 1991, 30, 976–981.
1H NMR (400 MHz, CD3CN, 258C, TMS): d=0.95 (t, 3J
ACTHNUGRTENUNG(H,H)=7.6 Hz,
12H; CH3), 1.14 (s, 18H; 2CACHTUNRTGNE(NUG CH3)3), 1.35 (m, 8H; CH2-CH3), 1.61 (m,
8H; CH2-CH2N), 3.10 ppm (m, 8H; NCH2); 13C{1H} NMR (75.5 MHz,
CD3CN, 258C, TMS): d=13.7 (s, CH3), 20.3 (s, CH2), 24.3 (s, CH2), 27.5
[5] a) C. Copꢅret, M. Chabanas, R. P. Saint-Arroman, J.-M. Basset,
Angew. Chem. 2003, 115, 164–191; Angew. Chem. Int. Ed. 2003, 42,
156–181; b) J. Meurig Thomas, R. Raja, D. W. Lewis, Angew. Chem.
2005, 117, 6614–6641; Angew. Chem. Int. Ed. 2005, 44, 6456–6482;
c) J. Zeller, J. Treptow, U. Radius, Z. Anorg. Allg. Chem. 2007, 633,
741–746; d) E. A. Quadrelli, J.-M. Basset, Coord. Chem. Rev. 2010,
254, 707–728.
[6] F. J. Feher, J. F. Walzer, Inorg. Chem. 1991, 30, 1689–1694.
[7] a) R. Duchateau, Chem. Rev. 2002, 102, 3525–3542; b) V. Lorenz,
F. T. Edelmann, Z. Anorg. Allg. Chem. 2004, 630, 1147–1157; c) F. J.
Feher, T. A. Budzichowski, Polyhedron 1995, 14, 3239–3253; d) S.
Lovat, M. Mba, H. C. L. Abbenhuis, D. Vogt, C. Zonta, G. Licini,
Inorg. Chem. 2009, 48, 4724–4728; e) F. J. Feher, D. A. Newman, J.
Am. Chem. Soc. 1990, 112, 1931–1936; f) R. Duchateau, U. Cremer,
R. J. Harmsen, S. I. Mohamud, H. C. L. Abbenhuis, R. A. van Sant-
en, A. Meetsma, S. K.-H. Thiele, M. F. H. van Tol, M. Kranenburg,
Organometallics 1999, 18, 5447–5459.
(s,
CACHTUNGTRENNUNG(CH3)3), 39.6 (s, CACHUTNGTENRNUG CAHUTNGTRENN(GU N,C)=2.8 Hz, NCH2),
(CH3)3), 59.2 (t, 1J
191.1 ppm (s, CO2); 51V{1H} NMR (105 MHz, CD3CN, 258C, VOCl3): d=
ꢀ512 ppm; IR (KBr): n˜ =2960 (s), 2874 (m), 1646 (w), 1586 (m), 1562
(m), 1481 (s), 1457 (m, sh), 1407 (m), 1375 (w), 1358 (m), 1261 (vw), 1220
(m), 1170 (vw), 1153 (vw), 1107 (vw), 1069 (vw), 1029 (vw), 948 (s), 927
(vs), 900 (s), 813 (w), 788 (vw), 743 (vw), 623 (m), 451 cmꢀ1 (w); elemen-
tal analysis calcd (%) for C26H54NO6V (527.2): C 59.18, H 10.32, N 2.65;
found: C 59.32, H 10.38, N 2.45.
X-ray diffraction studies: The crystals were mounted on a glass fiber and
then transferred into the cold nitrogen gas stream of the diffractometer
Stoe IPDS using MoKa radiation. The structures were solved by direct
methods (SHELXS-97)[22], refined versus F2 (SHELXL-97)[23] with aniso-
tropic temperature factors for all non-hydrogen atoms. All hydrogen
atoms were added geometrically and refined by using a riding model.
CCDC-761674 (1) and CCDC-761673 (2) contain the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[8] a) O. Viotti, G. A. Seisenbaeva, V. G. Kessler, Inorg. Chem. 2009,
48, 9063–9065.
Crystal data for 1·2CH2Cl2: C74H100Cl4O14Si8V, Mr =1661.97, T=
100(2) K, l=0.71073 ꢄ, orthorhombic, space group Pca21, a=18.0987(6),
[9] C. Ohde, M. Brandt, C. Limberg, J. Dçbler, B. Ziemer, J. Sauer,
Dalton Trans. 2008, 326–331.
6898
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 6892 – 6899

Homogeneous Catalysis
FULL PAPER
[10] a) D. M. Homden, C. Redshaw, Chem. Rev. 2008, 108, 5086–5130;
b) C. Limberg, Eur. J. Inorg. Chem. 2007, 3303–3314; c) D. Mendo-
za-Espinosa, T. A. Hanna, Inorg. Chem. 2009, 48, 7452–7456.
[11] a) E. Hoppe, C. Limberg, B. Ziemer, C. Mꢀgge, J. Mol. Catal. A
2006, 251, 34–40; b) E. Hoppe, C. Limberg, Chem. Eur. J. 2007, 13,
7006–7016.
[12] M. T. Bogert, D. Davidson, J. Am. Chem. Soc. 1931, 53, 3122–3130.
[13] a) G. Sꢀss-Fink, S. Stanislas, G. B. Shul’pin, G. V. Nizova, H. Stoeck-
li-Evans, A. Neels, C. Bobillier, S. Claude, J. Chem. Soc. Dalton
Trans. 1999, 3169–3175; b) G. Sꢀss-Fink, S. Stanislas, G. B. Shul’pin ,
G. V. Nizova, Appl. Organomet. Chem. 2000, 14, 623–628; c) J. Gꢁt-
jens, B. Meier, T. Kiss, E. M. Nagy, P. Buglyꢆ , H. Sakurai, K.
Kawabe, D. Rehder, Chem. Eur. J. 2003, 9, 4924–4935; d) L. Eng
Khoon, T. Siang Guan, B. M. Yamin, Acta Crystallogr. Sect. E 2006,
62, m207–m209; e) A. Kojima, K. Okazaki, S. Ooi, K. Saito, Inorg.
Chem. 1983, 22, 1168–1174; f) J. L. Hoard, W. R. Scheidt, D. M.
Collins, J. Am. Chem. Soc. 1971, 93, 3873–3877.
Rouen, F. DiNinno, M. L. Hammond, J. L. Huber, J. G. Sundelof,
G. G. Hammond, Bioorg. Med. Chem. Lett. 1999, 9, 3225–3230.
[17] M. Kosugi, S. Hikichi, M. Akitam, Y. Moro-oka, J. Chem. Soc.
Dalton Trans. 1999, 1369–1371.
[18] a) C. R. Waidmann, X. Zhou, E. A. Tsai, W. Kaminsky, D. A.
Hrovat, W. T. Borden, J. M. Mayer, J. Am. Chem. Soc. 2009, 131,
4729–4743; b) M. H. V. Huynh, T. J. Meyer, Chem. Rev. 2007, 107,
5004–5064; c) J. M. Mayer, Annu. Rev. Phys. Chem. 2004, 55, 363–
90.
[19] S. Rajasekhara Reddy, S. Das, T. Punniyamurthy, Tetrahedron Lett.
2004, 45, 3561–3564.
[20] V. W. Day, W. G. Klemperer, A. Yagasaki, Chem. Lett. 1990, 1267–
1270.
[21] a) P. Ehrlich, W. Engel, Z. Anorg. Allg. Chem. 1963, 322, 217–223;
b) D. Fenske, A.-F. Shihada, H. Schwab, K. Dehnicke, Z. Anorg.
Allg. Chem. 1980, 471, 140–146.
[22] SHELXS-97 Program for Crystal Structure Solution, G. M. Shel-
drick, University of Gçttingen, Gçttingen, 1997.
[14] Also compare: S. K. Hanson, R. T. Baker, J. C. Gordon, B. L. Scott,
A. D. Sutton, D. L. Thorn, J. Am. Chem. Soc. 2009, 131, 428–429.
[15] Y. Maeda, N. Kakiuchi, S. Matsumura, T. Nishimura, T. Kawamura,
S. Uemura, J. Org. Chem. 2002, 67, 6718–6724.
[23] SHELXL-97 Program for Crystal Structure Refinement, G. M. Shel-
drick, University of Gçttingen, Gçttingen, 1997.
[16] a) Y. Han , A. Bisello , C. Nakamoto, M. Rosenblatt, M. Chorev, J.
Peptide Res. 2000, 55, 230–239; b) M. L. Greenlee, J. B. Laub, G. P.
Received: January 21, 2010
Published online: May 12, 2010
Chem. Eur. J. 2010, 16, 6892 – 6899
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6899