Mechanism of hydride abstraction by cyclopentadienone-ligated carbonylmetal complexes (M = Ru, Fe)

Article abstract of DOI:10.1002/ejic.200800975

Cyclopentadienone-ligated ruthenium complexes, such as Shvo's catalyst, are known to oxidize reversibly alcohols to the corresponding carbonyl compounds. The mechanism of this reaction has been the subject of some controversy, but it is generally believed to proceed through concerted transfer of proton and hydride, respectively, to the cyclopentadienone ligand and the ruthenium center. In this paper we further study the hydride transfer process as an example of a coordinatively directed hydride abstraction by adding quantitative understanding to some features of this mechanism that are not well understood. We find that an oxidant as weak as acetone can be used to re-oxidize the intermediate ruthenium hydride without catalyst re-oxidation becoming rate-limiting. Furthermore, C-H cleavage is a significantly electrophilic event, as demonstrated by a Hammett reaction parameter of ρ = -0.89. We then describe how the application of our mechanistic insights obtained from the study have enabled us to extend the ligand-directed hydride abstraction strategy to include a rare example of an iron(0) oxidation catalyst. Wiley-VCH Verlag GmbH & Co. KGaA, 2009.

Full text of DOI:10.1002/ejic.200800975

FULL PAPER
DOI: 10.1002/ejic.200800975
Mechanism of Hydride Abstraction by Cyclopentadienone-Ligated
Carbonylmetal Complexes (M = Ru, Fe)
Megan K. Thorson,[a] Kortney L. Klinkel,[a] Jianmei Wang,[a] and Travis J. Williams*[a]
Keywords: Cyclopentadienone ligands / C–H bond activation / Oxidation / Hydrides / Reaction mechanisms
Cyclopentadienone-ligated ruthenium complexes, such as tone can be used to re-oxidize the intermediate ruthenium
Shvo’s catalyst, are known to oxidize reversibly alcohols to
the corresponding carbonyl compounds. The mechanism of
this reaction has been the subject of some controversy, but it
is generally believed to proceed through concerted transfer
of proton and hydride, respectively, to the cyclopentadienone
ligand and the ruthenium center. In this paper we further
hydride without catalyst re-oxidation becoming rate-limiting.
Furthermore, C–H cleavage is a significantly electrophilic
event, as demonstrated by a Hammett reaction parameter of
ρ = –0.89. We then describe how the application of our mech-
anistic insights obtained from the study have enabled us to
extend the ligand-directed hydride abstraction strategy to in-
study the hydride transfer process as an example of a coordi- clude a rare example of an iron(0) oxidation catalyst.
natively directed hydride abstraction by adding quantitative
understanding to some features of this mechanism that are
not well understood. We find that an oxidant as weak as ace-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
known to catalyze reversibly the interconversion of car-
bonyl and alcohol compounds (Scheme 1, C). Although 1
was first introduced as a hydrogenation catalyst, its utility
in the corresponding oxidation reaction has also been ex-
amined.[1f,3] Data collected to date regarding the mecha-
nism of oxidation reactions catalyzed by 1 are most consis-
tent with a mechanism involving transfer of both hydrogen
atoms (O–H and C–H) in or before a single rate-determin-
ing transition state when the reaction is conducted under
strongly oxidative conditions (tetrafluorobenzoquinone).[3b]
Thus, 3 is a bifunctional metal–ligand scaffold in which H+
(from O–H) acts as a coordinating direction element “L”
that disposes an electrophilic ruthenium center “M+” in the
proximity of a C–H bond.
We believe that coordination-directed hydride abstrac-
tion will be an excellent strategy for our program of
designing and developing efficient, selective, inexpensive,
and environmentally benign catalysts for C–H oxidation,
and we believe that identification of mild conditions of re-
oxidation of metal hydrides (such as 2) will be essential to
the identification of high-value catalytic systems because
this is a key to realizing catalytic turnover. Moreover, an
improved understanding of the polarization and energetics
of hydride abstraction mediated at a cyclopentadienone-
ligated metal center (such as 3) is essential to effective cata-
lyst design. In this paper we address these issues for the
case of some Shvo-related complexes, and add quantitative
understanding to some features of this mechanism that are
a key to our goal (Scheme 1, A): an oxidant as weak as
acetone can be used, and catalyst re-oxidation does not be-
come rate-limiting; replacement of a single phenyl group of
Introduction
C–H bond oxidation by hydride abstraction is an impor-
tant reaction for organic synthesis and has a central role in
applications ranging from utilization of hydrocarbon feed-
stocks to fine chemical synthesis,[1] yet its mechanism is un-
derexplored and its applications are scarcely exploited. Our
group is developing ligand–metal bifunctional catalysts for
hydride abstraction from general organic substrates to en-
able nucleophilic substitution reactions in which hydride is
activated as a leaving group (Scheme 1, A). Our strategy is
to devise a bifunctional catalyst in which the ligand “ad-
dresses” the catalyst to a particular C–H bond by placing
an electrophilic metal atom in its immediate proximity. Fol-
lowing hydride transfer to metal atom, the resulting M–H
group must be re-oxidized under mild conditions.
We have adopted the cyclopentadienone-ligated metal
scaffold (e.g. 3) as a starting point for development of coor-
dination-directed hydride abstraction reactions because of
the apparently significant role of the cyclopentadienone as
a redox non-innocent ligand. “Shvo’s catalyst” (Scheme 1,
1),[2] is an example of this type of reactivity. Compound 1
itself, an air-stable, commercially available, crystalline solid,
is the heterodimer of reduced (2) and oxidized (3) forms.
The mixture of 2 and 3, generated by dissociation of 1, is
[a] Loker Hydrocarbon Research Institute and Department of
Chemistry, University of Southern California,
837 Bloom Walk, Los Angeles, CA 90089-1661, USA
Fax: +1-213-740-6679
E-mail: travisw@usc.edu
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
Eur. J. Inorg. Chem. 2009, 295–302
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
295

M. K. Thorson, K. L. Klinkel, J. Wang, T. J. Williams
FULL PAPER
Scheme 1. Shvo’s system as a platform for hydride abstraction reactions.
the cyclopentadienone ring can reduce the efficiency of oxi- (3) forms (very small K1) for a reaction to occur. Alcohol
dation significantly; and C–H cleavage is a significantly elec- oxidation with 1 has isotope effects on both C–H (kH/kD
=
trophilic event. We then describe how the application of this 2.6) and O–H (kH/kD = 1.8).[3b] Bäckvall et al. have ex-
weakly oxidizing medium and our mechanistic insights ob- plained that these kinetic isotope effect (KIE) data, along
tained from the study have enabled us to extended the strat- with a kH/kD value of 4.61 for a fully deuterated substrate
egy to include a rare example of an iron(0)-based oxidation is consistent with a concerted transition state for hydride
catalyst.
and proton transfer (Scheme 2).[3b]
We started our study by switching to acetone, a milder,
environmentally more benign medium to effect oxidation
with catalyst 1. Upon dissolving dimer 1 in [D6]acetone
Results and Discussion
Studies on 1 conducted by Shvo et al.,[2] Casey et al.,[4] with alcohol 6a, we find dissociation and oxidation of di-
Bäckvall et al.,[1f,3] and others[5] have shown that in aro- mer 1 as is evident from the disappearance of the character-
matic solvents the catalyst resting state in these reactions is istic µ2-H peak (δ = –18.1 ppm) and appearance of [D6]2-
dimer 1, which must dissociate to reduced (2) and oxidized propanol (δ = +3.9 ppm), which indicates that the dissoci-
Scheme 2. Alcohol oxidation in Shvo’s system.
296
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 295–302

Hydride Abstraction by Cyclopentadienone-Ligated Carbonylmetal Complexes
ation equilibrium of 1 lies to the right under these condi- the rate of the overall reaction, and thus that [D6]acetone
tions. This transformation is complete within the first ac- is as efficient as a quinone in regenerating 3 from 2 in a
quisition of a kinetics run (see Supporting Information for solution.
an illustration). We perceive that the result is an acetone-
or substrate-ligated dicarbonylruthenium complex (5, L =
[D6]acetone or 6). Using this catalyst, we verify that 6a can
be converted to 7a in 97% isolated yield on a 1 mmol
scale.[6]
In cases in which we introduced benzoquinone to com-
pete with acetone in acetone solution (Table 1, Entries 4
and 5), acetone was the major oxidant because of its high
concentration. By comparison of 1H NMR peaks for
benzoquinone and [D6]2-propanol, we can estimate the ra-
Systematic kinetics experiments on the conversion of 6a
to 7a (Table 1) in [D6]acetone solution support a rate law
tio of rates of reduction of benzoquinone and acetone (kBQ
/
kacetone) as 414(4). Although we observe this ratio to be con-
stant over the course of the conversion, we must interpret
it as an upper bound because [D6]2-propanol can reduce
benzoquinone (by re-formation of [D6]acetone) in the pres-
ence of 1. Moreover, Table 1 (Entry 5) shows that an excess
of quinone can slow the reaction. We perceive that this is
a result of competitive binding of 3 by benzoquinone or
hydroquinone.
An internal competition isotope effect experiment with
[D]benzyl alcohol [PhC(D)HOH] afforded a value of kH/kD
= 3.7(2) at 323 K through the first 40% of conversion. We
assign our value to a primary KIE [as opposed to an equi-
librium isotope effect (EIE) or a combination of both] be-
cause it is constant over the course of conversion. This
value is higher than a literature value[3b] (kH/kD = 2.6 by
independent runs) observed for the conversion of 6d to 7d,
possibly because of the relative strength of the primary (ver-
sus secondary) C–H bond. The difference in these values
can also be attributed to the role of an EIE in the literature
measurement. By contrast, we observe full dissociation of
1.
of d[6a]/dt
=

–k2·[1]1/2·[6a]1·[oxidant]0 with k2
=
1.67(6)ϫ10–2
–1/2 s–1. This is analogous to what Bäckvall
et al. observed for quinone conditions (1, [D6]benzene, tet-
rafluorobenzoquinone), but is interesting because we ob-
serve no kinetic order in the re-oxidation step even though
it is a significantly milder oxidant than tetrafluorobenzo-
quinone. Kinetic order on [1] is illustrated in a plot of
ln(kobs) versus ln[1] (Figure 1, left). A slope of 0.40(6) in
this plot is consistent with a half-order dependence on [1].
This indicates that dimerization of 2 and 3 is occurring dur-
ing the catalysis even though this equilibrium lies well to
the right under the reaction conditions. The dependence of
the reaction rate on [acetone] is shown graphically in Fig-
ure 1 (right). The observed rate constants (kobs) decrease
only slightly with [acetone] in a mixed [D2]dichlorometh-
ane/[D6]acetone solution. The high y intercept of this plot
is inconsistent with the kinetic order on [acetone], but is
more likely interpreted as a medium effect: the data in Fig-
ure 1 (right) represent variation from 25% to 100% acetone.
This means that re-oxidation of 2 from 3 does not impact
We also show that the C–H cleavage is an electrophilic
event by measuring a Hammett ρ value of –0.89(5) by com-
paring the rates of oxidation of alcohols 6a–6d in indepen-
dent runs (Figure 2).[7] A Hammett reaction parameter of ρ
= –0.89 is larger than expected for a transition state involv-
ing β-hydride elimination or free-radical hydrogen atom
transfer. Kaneda et al. report ρ = –0.43 for the oxidation of
6 with a ruthenium catalyst immobilized on hydroxyapatite
and proposes a β-hydride elimination mechanism.[8] Kuria-
cose et al. report ρ = –0.3 for the same reaction with a
Table 1. Kinetics data for the 1-catalyzed conversion of 6a to 7a.[a]
Entry kobs (ϫ105 s–1) [1] [m] [6a] [m] BQ [m] kBQ/kacetone
1
2
3
4
5
7.55(26)
11.1(1)
7.46(3)
8.49(31)
4.13(7)
1.5
2.9
1.5
1.5
1.5
120
120
240
120
120
0
0
0
72
144
–
–
–
410
418
[a] kobs = k2·[6]. BQ = benzoquinone; k2 = 1.67(6)ϫ10–2 –1/2 s–1.

Figure 1. Left: plot of ln(kobs) versus ln[1]; slope = 0.40(6). Right: plot of kobs versus [acetone] in a mixed medium of [D6]acetone and
[D2]dichloromethane. Kinetics data were processed as described in the Exp. Sect.
Eur. J. Inorg. Chem. 2009, 295–302
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
297

M. K. Thorson, K. L. Klinkel, J. Wang, T. J. Williams
FULL PAPER
ruthenium trichloride catalyst in the presence of N-methyl- stability, and (3) rate in the C–H oxidation step (k2). We
morpholine N-oxide. These authors propose a hydrogen further hypothesized that a more electron-deficient catalyst
atom (radical) abstraction by an (oxoido)ruthenium(V) in- based on chloroalkyne 8c would facilitate hydride abstrac-
termediate.[9] Our observation is more in line with benzyl tion and afford a faster oxidation reaction, but this complex
alcohol oxidation by quinolinium chlorochromate (ρ = is apparently unreactive (Scheme 3).
–1.2).[10] Thus, with catalyst 1, we interpret that a signifi-
cant cationic character is evolved at the carbon atom in
Table 2. Conversion of 6a to 7a with catalyst precursors 1 (gray)
and 4b (black).[a]
the rate-determining transition state, which is only slightly
compensated by electron donation from the concurrent de-
protonation of the O–H bond.
Entry
[Ru] source
kobs (ϫ105 s–1)
Conversion
1
2
1 (gray)
6.3(2)
4.0(3)
92(3)%
46(1)%
Figure 2. Hammett plot for the 1-catalyzed conversion of 6 to 7.
Kinetics data were acquired and processed as described in the Exp.
Sec. ρ = –0.89(5).
4b (black)
[a] A solvent system of [D2]dichloromethane/[D6]acetone (1:1) was
used.
Furthermore, we have studied other ruthenium com-
plexes based on electronically differentiated cyclopen-
Having identified mild oxidative conditions in which cat-
tadienones to compare the reaction rates. We attempted to alyst 1 operates smoothly, and having developed an under-
prepare catalyst precursors analogous to 4 by treating standing of the mechanistic details of the reaction under
[Ru3(CO)12][4b] with the corresponding cyclopentadienone these conditions, we turned our attention to the possibility
9[11] (Table 2). Only methyl-substituted compound 9b par- of replacing the catalytic ruthenium atom by an iron atom.
ticipated in the formation of 4b cleanly. Surprisingly, al- To do so would involve hydride abstraction to a formal
though 4b appears monomeric under the reaction condi- iron(0) center. Although biological oxidases such as cyto-
tions used in Table 1, it reacts with 6a ca. 101 times slower chrome p450[12] rely on iron-based oxidation systems, these
than the parent (1). Because 4b is only sparingly soluble and related non-heme iron-based oxidation catalysts[13] gen-
acetone, the experiment was repeated in a solvent system of erally feature an iron(II) or iron(III) precursor that is sup-
[D2]dichloromethane/[D6]acetone (1:1). Under these condi- ported by several nitrogen ligands. These systems typically
tions, both 1 and 4b are completely dissolved and dissoci- activate C–H bonds through the generation of a high-valent
ated to active forms, and the parent system is only 60% (oxido)iron intermediate[13c] from which hydrogen atoms
faster: k2(Ph)/k2(Me) = 1.6(1). We observe, however, that (not hydride ions) are abstracted in a radical (or radical
the methyl-substituted system stalls at 46(1)% conversion, rebound) mechanism.[13a,13b] Several outstanding examples
whereas the parent reached 92(3)%. This, along with a dark of C–H oxidation have been reported recently based on this
color that develops over time, indicates that the cyclopen- strategy.[14] An iron homolog of 1, however, would be struc-
tadienone ligand is being lost in the course of the reaction. turally and mechanistically distinct: such a system would be
We thus assign the deficiencies of 4b as: (1) solubility, (2) supported by a redox-noninnocent cyclopentadienone li-
Scheme 3. Syntheses of 9 and 4.
298
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 295–302

Hydride Abstraction by Cyclopentadienone-Ligated Carbonylmetal Complexes
gand that would be an integral part of the catalyst’s ligand– ditions for re-oxidation of our catalyst is advantageous. In
metal bifunctional nature, and C–H abstraction would in- this case, a yield of 38% of 7a is realized (Entry 3). As
volve transfer of a hydride ion, rather than a hydrogen radi- predicted by Hammett analysis, rate, and thus conversion,
cal. More closely related to our strategy are some examples are higher with alcohol 6b: this substrate can be oxidized
of iron-catalyzed transfer hydrogenation catalysts in which in up to 79% yield of 7b.
a secondary alcohol is the hydrogen source. Some examples
include a porphyrin-ligated system,[15] [(PP3)Fe(H)(H2)]+
[BPh4]– [PP3 = P(CH2CH2PPh3)3],[16] and a cyclopen-
tadienone-based system.[17] The first two appear to have lit-
tle homology with our system, but the third suggests a
starting point.
Surprisingly, (cyclopentadienone)iron(0) complexes do
participate in productive hydride abstraction reactions in
oxidative media. We first investigated this by preparing bi-
cyclic complex 11 (Scheme 4)[18] and treating it with benzo-
quinone and alcohol 6a in [D6]benzene, which afforded only
a trace amount of 7a (Table 3, Entry 1). Curiously, we ob-
serve neither iron oxide formation nor ligand displacement
from 11 under these conditions. Data in Table 2 suggested
Scheme 4. Syntheses and X-ray structure of 11.[20] E = CO2Me.
that this system could be improved by switching from dou-
bly arylated complex 11 to fully phenylated complex 12.[19]
Although significant improvement remains to be made
This worked moderately (Entry 2) and enabled a yield of in the efficiency of this iron(0)-based oxidation system, we
24%. Much as in the case of 1, application of acetone con- have gathered some data regarding its mechanism. An iso-
Table 3. Iron-catalyzed conversion of 6 to 7.[a]
Entry
[Fe]
Alcohol
Oxidant
Solvent
T [°C]
Yield (conv.), time
1
2
11 (0.1 equiv.)
12 (0.1 equiv.)
6a
6a
BQ
BQ
C6D6
C6D6
65
65
1% (1%), 16 h
11% (14%), 17 h
24% (26%), 4 d
38% (44%), 4 d
52% (79%), 4 d
79% (97%), 2 d
3
12 (0.1 equiv.)
12 (0.2 equiv.)
12 (0.5 equiv.)
6a
6b
6b
–
–
–
(CD3)2CO
(CD3)2CO
(CD3)2CO
54
80
80
4[b]
5[b]
[a] Yield and conversion were determined by NMR spectroscopy. BQ = benzoquinone. All reactions were run in screw-capped NMR
tubes and prepared in air. [b] Experiments 4 and 5 were run with 1 equiv. (relative to 12) of D2O at 80 °C (bath temp.) in a J. Young
NMR tube under reduced pressure after rigorous degassing.
Scheme 5. Proposed mechanisms for alcohol oxidation with catalyst precursor 12.
Eur. J. Inorg. Chem. 2009, 295–302
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
299

M. K. Thorson, K. L. Klinkel, J. Wang, T. J. Williams
FULL PAPER
distilled on a small scale immediately prior to use. Compound 9a
is commercially available, and compounds 9b–e are known.[11]
Spectroscopy and data: NMR spectra were measured with a Varian
Mercury 400 NMR spectrometer. NMR spectroscopic data were
processed and analyzed with Varian vnmr 6.1c or Acorn NMR
NUTS. Spectra acquired for kinetics analysis were taken with a 5 s
pulse sequence (Ͼ 5ϫT1). Charts and graphs were generated with
Synergy Software KaleidaGraph 4. Error values were calculated
and propagated by using traditional methods.
tope effect (determined as above through 20% conversion
at 50 °C) of 3.6(9) indicates that C–H bond cleavage occurs
in the rate-limiting step. The relative facility of oxidation of
6b relative to 6a is consistent with the view that this cleav-
age is an electrophilic event, as is observed in the ruthenium
system. To the extent that direct comparison is possible,
iron-based system 12 is slower than ruthenium-based sys-
tem 1. In both cases, we perceive that the active catalytic
species is a dicarbonyl(cyclopentadienone)metal complex
(14 or 3) and that hydride abstraction by that species is rate-
determining. For alcohol 6a, we observe an initial turnover
frequency of 7.8(1) h–1 for ruthenium (1-based) and
0.10(1) h–1 for iron (12-based) under analogous conditions.
Kinetics: Sample procedure: A standard solution of Shvo’s catalyst
1 (700 µL, 1.5 m 1, in [D6]acetone) was added to an oven-dried
screw-cap NMR tube. No further precautions were taken to ex-
clude air or water. The sample was placed into the NMR spectrom-
eter pre-warmed to 323(1) K, and 10.2 µL (1.01 mmol, to make
Although we believe that this reflects the relative energetics 120 m, 2.5 mol-% Ru atom) of alcohol 6a was added by syringe,
and timing begun. The tube was reinserted into the NMR spec-
trometer. Integrations were recorded in comparison to a 1,2-dichlo-
roethane (4.7 µL) internal standard. Kinetic constants (pseudo-first
order kobs) were determined by statistical agreement of kobs values
measured individually for aryl, methyl, benzyl (6a), and isopropyl
C–H signals in multiple runs. Air sensitivity: For the case of ruthe-
nium complexes, independence of rate from air and water exposure
was verified by side-by-side runs conducted in [D6]benzene with
benzoquinone. One was prepared from solvent distilled from so-
dium/benzophenone ketyl and manipulated under air-free condi-
of hydride abstraction, an induction period for the initiation
of 12 cannot be excluded.
Because complexes 11 and 12 are coordinatively satu-
rated species (18 electrons), these are most likely precursors
that enter a catalytic cycle by ligand dissociation. We per-
ceive based on previous literature[17b] that the mechanism
of activation involves hydrolysis of one CO ligand to give
CO2 and intermediate 13 (Scheme 5). Thermal or photo-
chemical dissociation of CO from 12 (highlighted in gray)
is also a possibility. Once initiated, we believe that this tions, and the other was prepared with commercial materials on
the bench top. Rate constants observed in these experiments were
identical within error. This was not repeated in [D6]acetone because
anhydrous acetone is known to undergo dehydrative dimeriza-
tion.[21] Kinetic isotope effects: Internal competition KIEs (kH/kD)
were measured by comparison of [benzaldehyde] and [[D]benzalde-
mechanism proceeds analogously to Scheme 2.
Conclusions
1
hyde] as observed in H NMR integrations for benzaldehyde and
Herein we describe mechanistic paradigms regarding the
oxidation of alcohols with cyclopentadienone-ligated metal
catalysts that are a key to our ongoing program in directed
hydride abstraction: alcohol oxidation by 1 is a directed hy-
dride abstraction, and even with an oxidant as weak as ace-
tone, and under these mild conditions re-oxidation of inter-
mediate ruthenium hydride 2 is rapid relative to k2. We also
find that substituting even one phenyl group on parent
complex 1 significantly attenuates the oxidation reactivity.
Using these insights we have discovered a rare example of
C–H oxidation that occurs at a (formal) iron(0) center
within the ligand–metal bifunctional scaffold. This offers
significant potential advantages over ruthenium, such as
cost, toxicity, and environmental impact. Further studies re-
garding directed hydride abstraction reactions are ongoing
in our laboratory.
[D]benzyl alcohol. For catalyst 1, 25 points throughout the first
40% of conversion were used. For complex 12, 4 points through
20% conversion were used. Error was calculated as the standard
deviation of these measurements. To insure comparability of inte-
grations, spectra were acquired with a calibrated 90° pulse, 5 s ac-
Table 4. Kinetic dependence on [1] as shown in Figure 1 (left).
Entry
[1] [m]
kobs (ϫ105 s–1)
ln [1]
ln(kobs)
1
2
3
4
1.50
2.90
4.50
6.50
7.55(26)
11.1(1)
11.6(1)
14.1(1)
–6.50
–5.84
–5.40
–5.04
–9.49(3)
–9.11(1)
–9.06(1)
–8.87(1)
Table 5. Kinetic dependence on [acetone] as shown in Figure 1
(right).
Entry
[acetone] []
kobs (ϫ105 s–1)
1
2
3
4
13.6 (100 vol.-%)
10.2 (75 vol.-%)
6.80 (50 vol.-%)
3.40 (25 vol.-%)
7.55(26)
6.74(58)
6.12(12)
5.82(13)
Experimental Section
General: General procedures and instrumentation are defined fully
in the Supporting Information. Preparative details for all other ma-
terials and graphical 1H NMR spectra are provided in the Support-
ing Information. [D6]Benzene was vacuum-distilled from sodium/
benzophenone ketyl for kinetics experiments. [D6]Acetone was use
as received from Alfa Aesar. Shvo’s catalyst (1) was purchased from
Strem and used as received. Pentacarbonyliron(0) [Fe(CO)5] and
nonacarbonyldiiron(0) [Fe2(CO)9] were purchased from Strem, ma-
nipulated under air-free conditions and protected from light.
Benzoquinone was recrystallized from ethanol and used as a yellow
crystal. Alcohols 6a–d were purchased from common suppliers and
Table 6. Data for Hammett plot (Figure 2).
Entry Alcohol: R
σ[7]
kobs (ϫ105 s–1)
ln(kobs)
1
2
3
4
6b: OMe
6c: Me
6a: H
–0.268(77)
–0.170(46)
0.000(34)
12.8(2)
11.0(2)
7.55(26)
6.53(11)
0.228(15)
0.164(17)
0.000(20)
–0.063(32)
6d: F
+0.062(66)
300
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 295–302

Hydride Abstraction by Cyclopentadienone-Ligated Carbonylmetal Complexes
quisition time, and 45 s pulse delay. A control spectrum of
[H6]benzaldehyde showed Ͻ 5% error in 1H integrations under
these conditions (Tables 4, 5, and 6).
Petroleum Research Fund (grant 47987-G1). We thank Robert Bau
and Tim Stewart for assistance with X-ray crystallography. T. J. W.
thanks colleagues for insightful discussions: Kyung Jung, Surya
Prakash, Mark Thompson, George Olah, Charles McKenna, Rich-
ard Brutchey, Tom Flood, and Nicos Petasis.
Preparative Details: Sample procedure of the oxidation of 1-pheny-
lethanol: 1 (21.7 mg, 20.0 µmol), acetone (3 mL), and 6a (121 µL,
1.00 mmol) were combined in a test tube with a stir bar. The re-
sulting solution was subjected to reflux for 2 d, and the reaction
mixture was cooled to room temperature, concentrated, and puri-
fied by flash column chromatography to afford 7a (118 mg, 97%)
as a colorless liquid.
[1] Regarding hydride transfer, see: a) N. C. Deno, H. J. Peterson,
G. S. Saines, Chem. Rev. 1960, 60, 7–14. Some examples in-
clude: b) Z. Li, D. S. Bohle, C.-J. Li, Proc. Natl. Acad. Sci.
USA 2006, 103, 8928–8933; c) S. J. Pastine, K. M. McQuaid, D.
Sames, J. Am. Chem. Soc. 2005, 127, 12180–12181; d) W. H. N.
Nijhuis, W. Verboom, A. A. El-Fadl, S. Harkema, D. N. Rein-
houdt, J. Org. Chem. 1989, 54, 199–209. via SET:; e) S. Fuku-
zumi, K. Ohkubo, T. Okamoto, J. Am. Chem. Soc. 2002, 124,
14147–14155; f) for a review, see: J.-E. Bäckvall, J. Organomet.
Chem. 2002, 652, 105–111.
[2] Y. Blum, D. Czarkie, Y. Rahamim, Y. Shvo, Organometallics
1985, 4, 1459–1461.
[3] a) M. L. S. Almeida, M. Beller, G.-Z. Wang, J.-E. Bäckvall,
Chem. Eur. J. 1996, 2, 1533–1536; b) J. B. Johnson, J.-E.
Bäckvall, J. Org. Chem. 2003, 68, 7681–7684.
[4] a) C. P. Casey, S. E. Beetner, J. B. Johnson, J. Am. Chem. Soc.
2008, 130, 2285–2295; b) C. P. Casey, S. W. Singer, D. R. Pow-
ell, R. K. Hayashi, M. Kavana, J. Am. Chem. Soc. 2001, 123,
1090–1100.
[5] For a computational study of 1-catalyzed hydrogenation of
ketones, see: A. Comas-Vives, G. Ujaque, A. Lledós, Organo-
metallics 2007, 26, 4135–4144.
[6] Further information on the synthetic applications of 1 and its
related system [(PPh3)2RuCl2]/K2CO3 has been reported:
M. L. S. Almeida, M. Beller, G.-Z. Wang, J.-E. Bäckvall, Chem.
Eur. J. 1996, 2, 1533–1536.
Synthesis of 11: By cannulation, a 25 mL solution of 10[22] (0.42 g,
2.00 mmol) in NEt3 was added to a 25 mL solution of 4-iodotol-
uene (1.09 g, 5.01 mmol), [PdCl2(PPh3)2] (0.03 g, 0.04 mmol), and
[CuI] (0.02 g, 0.10 mmol) under nitrogen. The mixture was stirred
at room temperature for 18 h at which time all volatiles were re-
moved under reduced pressure. The resulting oil was dissolved in
CH2Cl2 and washed once each with ca. 20 mL of Na2S2O3 solution
and brine, respectively. The organic layer was then dried with
MgSO4. Upon removal of all volatiles under reduced pressure, a
golden-colored oil resulted. The oil was dissolved in 20 mL of hex-
anes and placed in a –15 °C freezer overnight. The product was
isolated and recrystallized from hexanes resulting in 0.37 g (57%
yield) of off-white crystals of 10a; m.p. 68–71 °C. 1H NMR ([D]-
3
chloroform, 400 MHz): δ = 7.28 (d, JH,H = 7.5 Hz, 4 H, Ar-H),
3
7.09 (d, JH,H = 7.5 Hz, 4 H, Ar-H), 3.80 (s, 6H. OCH3), 3.26 (s, 4
H, CH2), 2.34 (s, 6 H, CH3) ppm. 13C NMR ([D]chloroform,
100 MHz): δ = 169.5 (2 C), 138.2 (2 C), 131.6 (4 C), 129.0 (4 C),
120.0 (2 C), 83.9 (2 C), 83.2 (2 C), 57.3, 53.1 (2 C), 23.9 (2 C), 21.5
(2 C) ppm. FTIR (NaCl): ν = 3291 (w), 3030 (w), 2954 (m), 2923
˜
(w), 1743 (s), 1734 (s), 1511 (s), 1436 (m), 1327 (m), 1294 (m), 1263
(m), 1212 (br., s), 1107 (w), 1076 (m), 1058 (m), 1022 (w), 991 (w),
943 (w), 853 (w), 817 (s), 668 (w) cm–1. C25H24O4 (388.46): calcd.
C 77.30, H 6.23; found C 76.73, H 6.11. The reaction was run
air-free in a method similar to that described by Pearson et al.[18]
Specifically, 10a (0.11 g, 0.27 mmol) was dissolved in 600 mL of
toluene (distilled from calcium hydride) in a 50 mL Strauss flask
equipped with a stir bar. Next, [Fe(CO)5] (0.22 mL, 1.63 mmol),
previously degassed with N2 was syringed into the flask. The flask
was then flushed three times with CO gas and tightly sealed and
placed into a 110 °C oil bath for 4 d. All volitales were then re-
moved under reduced pressure, resulting in the crude product as a
powdery solid. The solid was treated with warm hexanes and fil-
tered, resulting in 80 mg (55% yield) of pure 11 as a mustard-yellow
solid; decomposition 201–204 °C. 1H NMR ([D]chloroform,
[7] L. P. Hammett, J. Am. Chem. Soc. 1937, 59, 96–103.
[8] K. Kaneda, K. Mori, T. Hara, T. Mizugaki, K. Ebitani, Catal.
Surv. Asia 2004, 8, 231–239.
[9] K. Vijayasri, J. Rajaram, J. C. Kuriacose, J. Mol. Catal. 1987,
39, 203–217.
[10] H. B. Özgün, N. Degirmenbas¸i, J. Chem. Res. (S) 1997, 32–
˘
33.
[11] P. A. Wender, T. J. Paxton, T. J. Williams, J. Am. Chem. Soc.
2006, 128, 14814–14815.
[12] J. T. Groves, Y.-Z. Han, in Cytochrome P450: Structure, Mecha-
nism, and Biochemistry, 2nd ed. (Ed.: P. R. Ortiz de Montel-
lano), Plenum Press, New York, 1995.
[13] a) M. Costas, M. P. Mehn, M. P. Jensen, L. Que Jr, Chem. Rev.
2004, 104, 939–986; b) M. M. Abu-Omar, A. Loaiza, N.
Hontzeas, Chem. Rev. 2005, 105, 2227–2252. Regarding the
first mechanistic evidence implicating the role of a non-heme
FeV=O species in alkane hydroxylation, see: c) K. Chen, L.
Que Jr, J. Am. Chem. Soc. 2001, 123, 6327 –6337.
[14] a) A. Company, L. Gómez, X. Fontrodona, X. Ribas, M. Co-
stas, Chem. Eur. J. 2008, 14, 5727–5731; b) P. C. A. Bruijnincx,
I. L. C. Buurmans, S. Gosiewska, M. A. H. Moelands, M.
Lutz, A. L. Spek, G. van Koten, R. J. M. Klein Gebbink,
Chem. Eur. J. 2008, 14, 1228–1237; c) M. S. Chen, M. C. White,
Science 2007, 318, 783–787; d) G. J. P. Britovsek, J. England,
S. K. Spitzmesser, A. J. P. White, D. J. Williams, Dalton Trans.
2005, 945–955.
[15] S. Enthaler, G. Erre, M. K. Tse, K. Junge, M. Beller, Tetrahe-
dron Lett. 2006, 47, 8095–8099.
[16] C. Bianchini, E. Farnetti, M. Graziani, M. Peruzzini, A. Polo,
Organometallics 1993, 12, 3753–3761.
[17] Regarding a cyclopentadienone-ligated iron(II) catalyst for
ketone hydrogenation, see: a) C. P. Casey, H. Guan, J. Am.
Chem. Soc. 2007, 129, 5816–5817; b) H.-J. Knölker, E. Baum,
H. Goesmann, R. Klauss, Angew. Chem. Int. Ed. 1999, 38,
2064.
[18] Prepared according to a known procedure: A. J. Pearson, R. J.
Shively, R. A. Dubbert, Organometallics 1992, 11, 4096–4104.
3
3
400 MHz): δ = 7.94 (d, JH,H = 7.9 Hz, 4 H, Ar-H), 7.23 (d, JH,H
2
= 7.5 Hz, 4 H, Ar-H), 4.22 (d, JH,H = 16.6 Hz, 2 H, CH2), 3.52
(d, JH,H = 16.6 Hz, 2 H, CH2), 3.92 (s, 3 H, OCH3), 3.89 (s, 3 H,
2
OCH ), 2.38 (apparent s, 6 H, CH ) ppm. FTIR (NaCl): ν = 3437
˜
3
3
(br., s), 2067 (s), 2029 (m), 1987 (s), 1739 (m), 1653 (m), 1645 (m),
1623 (s), 1309 (w), 1260 (w), 1206 (w), 1189 (w), 1166 (m), 1134
(w), 1050 (w), 872 (w), 822 (m), 752 (w), 732 (w), 711 (w), 615 (m),
588 (w), 568 (w) cm–1. C29H24FeO8 (556.34): calcd. C 62.61, H,
4.35; found C 62.76, H 4.35.
Supporting Information (see footnote on the first page of this arti-
cle): Complete general procedures, details of kinetics analysis (in-
cluding graphical spectra), preparative details and graphical spectra
for new compounds, and line-listed X-ray crystallographic data for
11.
Acknowledgments
This research was supported by the University of Southern Cali-
fornia, the Loker Hydrocarbon Research Institute, and the ACS
Eur. J. Inorg. Chem. 2009, 295–302
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
301

M. K. Thorson, K. L. Klinkel, J. Wang, T. J. Williams
[19] Prepared from Ph4C5O and [Fe2(CO)9]: G. N. Schrauzer, J. Am. [21] W. L. F. Armarego, C. L. L. Chai, Purification of Laboratory
FULL PAPER
Chem. Soc. 1959, 81, 5307–5310.
Chemicals, 5th ed., Elsevier, San Francisco, 2003.
[22] R. S. Atkinson, M. J. Grimshire, J. Chem. Soc. Perkin Trans. 1
1986, 1215–1224.
[20] CCDC-690601 contains the supplementary crystallographic
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.
Received: October 6, 2008
Published Online: December 2, 2008
302
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 295–302