Kinetic studies on sec-alcohol racemization with dicarbonylchloro(pentabenzylcyclopentadienyl)- and dicarbonylchloro(pentaphenylcyclopentadienyl)ruthenium catalysts

Article abstract of DOI:10.1002/cctc.201300163

Pentasubstituted cyclopentadienyl complexes of ruthenium R₅CpRu(CO)₂Cl (R=Ph, benzyl) form, upon activation with tBuOK, highly active catalysts for racemization of chiral sec-alcohols. In combination with suitable resolving enzymes, such catalyst systems can efficiently be utilized for dynamic kinetic resolution reactions providing chiral alcohols, after hydrolysis of the corresponding acetates, in high yields and high enantiomeric purities. Here, three such ruthenium complexes were first characterized by NMR spectroscopy and cyclic voltammetry analysis (CVA) for elucidating their electronic characteristics in detail. Then, accurate kinetic studies were performed providing for the first time the calculated racemization rate constants for such catalyst systems. Furthermore, the dependence of the racemization rate on the electronic structure of the catalyst was investigated from the Hammett constants, substitution patterns of the substrate, and by isotopic labeling studies. The results obtained support the earlier suggested racemization reaction mechanism and indicated that the electron-rich catalyst Bn₅CpRu(CO)₂Cl (Bn=benzyl) racemizes electron-rich substrates more efficiently and in most cases faster than its pentaphenyl substituted analogue, formerly often considered as the leading catalyst candidate for dynamic kinetic resolution applications. The electron-deficient catalyst Ph₅CpRu(CO)₂Cl, in turn, is more efficient for electron-poor substrates.

Full text of DOI:10.1002/cctc.201300163

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300163
Kinetic Studies on sec-Alcohol Racemization with
Dicarbonylchloro(pentabenzylcyclopentadienyl)- and
Dicarbonylchloro(pentaphenylcyclopentadienyl)ruthenium
Catalysts
Denys Mavrynsky,[a] Dmitry Yu. Murzin,[b] and Reko Leino*[a]
Pentasubstituted cyclopentadienyl complexes of ruthenium
R5CpRu(CO)2Cl (R=Ph, benzyl) form, upon activation with
tBuOK, highly active catalysts for racemization of chiral sec-al-
cohols. In combination with suitable resolving enzymes, such
catalyst systems can efficiently be utilized for dynamic kinetic
resolution reactions providing chiral alcohols, after hydrolysis
of the corresponding acetates, in high yields and high enantio-
meric purities. Here, three such ruthenium complexes were
first characterized by NMR spectroscopy and cyclic voltamme-
try analysis (CVA) for elucidating their electronic characteristics
in detail. Then, accurate kinetic studies were performed provid-
ing for the first time the calculated racemization rate constants
for such catalyst systems. Furthermore, the dependence of the
racemization rate on the electronic structure of the catalyst
was investigated from the Hammett constants, substitution
patterns of the substrate, and by isotopic labeling studies. The
results obtained support the earlier suggested racemization re-
action mechanism and indicated that the electron-rich catalyst
Bn5CpRu(CO)2Cl (Bn=benzyl) racemizes electron-rich sub-
strates more efficiently and in most cases faster than its penta-
phenyl substituted analogue, formerly often considered as the
leading catalyst candidate for dynamic kinetic resolution appli-
cations. The electron-deficient catalyst Ph5CpRu(CO)2Cl, in turn,
is more efficient for electron-poor substrates.
Introduction
Dynamic kinetic resolution (DKR) of racemic sec-alcohols by
combination of enzymes with transition-metal catalysts is
a useful method for the production of chiral building blocks
for fine chemicals and active pharmaceutical ingredients.[1] In
such processes, in contrast to conventional kinetic resolutions,
the resolving enzyme is combined with an appropriate homo-
geneous or heterogeneous metal catalyst, which allows for
a continuous in situ racemization of the slower reacting alco-
hol enantiomer, optimally providing quantitative conversions
of the initially racemic mixture into an enantiomerically pure
product. The first example of a homogeneous ruthenium cata-
lyst utilized for this purpose was the Shvo complex
1 (Figure 1).[2] Catalyst 1 requires, however, elevated tempera-
tures for activation, and, therefore, it is not fully compatible
with all enzymes. In recent years, several modified, highly
active catalysts, inspired by the Shvo system and capable of
racemizing alcohols at ambient temperature, have been devel-
Figure 1. Representative ruthenium-based catalysts 1–5 for racemization of
sec-alcohols.
oped. Notable earlier examples are the aminosubstituted (tet-
raphenylcyclopentadienyl)ruthenium catalyst 2 described by
Park and coworkers,[3] and the (pentaphenyl)cyclopentadienyl
complex 3a described by Bꢀckvall et al.,[4] the latter of which is
commercially available and has been shown to be potentially
applicable even for large-scale DKR and industrial applica-
tions.[5]
[a] D. Mavrynsky, Prof. Dr. R. Leino
Laboratory of Organic Chemistry
ꢀbo Akademi University
FI-20500 ꢀbo (Finland)
E-mail: reko.leino@abo.fi
[b] Prof. Dr. D. Y. Murzin
In 2009, we reported the synthesis and characterization of
ruthenium complex 4a,[6a] which, in combination with Candida
antarctica lipase B (CAL-B) and upon activation with tBuOK,
forms a highly efficient catalyst system for the DKR of a wide
range of sec-alcohols providing in most cases >99% ee values
Laboratory of Industrial Chemistry and Reaction Engineering
ꢀbo Akademi University
FI-20500 ꢀbo (Finland)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300163.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2436

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
and 91–99% isolated yields of the corresponding chiral ace-
tates.[6] Complex 4a is a (pentabenzyl)cyclopentadienyl ana-
logue of the Bꢀckvall catalyst 3a and, owing to its simple and
cost-efficient preparation,[7] enhanced solubility induced by the
benzyl substituents, and high racemization activity for a broad
substrate scope, a viable alternative to 3a even for large-scale
DKR applications.
suggestion. At the same time, exchange of labeled CO with
the chloro complex 3a was shown to be significantly slower.[8e]
Further detailed studies have shown that the mechanism of
sec-alcohol racemization with 18 electron 3a is based on (after
activation of the catalyst with tBuOK and the subsequent alk-
oxide exchange) dissociation of one CO ligand to generate
a vacant site for b-hydride elimination, subsequently yielding
a metal–hydride species with a coordinated ketone (step A,
Scheme 1).[4b,8] Re-addition of the hydride (step B) to the
ketone then, after another alkoxide exchange, releases the
racemized alcohol.[8] Recent studies on the influence of the
electronic contribution of the catalyst and substrate structure
on the rate of racemization have shown the racemization rate
to depend on the electronic properties of the substrate.[9,10]
With electron-deficient substrates the racemization rate is limit-
ed by the hydride abstraction step A, whereas the hydride re-
addition step B is faster. Electron-rich substrates, in turn, un-
dergo oxidation with ease, whereas the subsequent reduction
step is slowed down by the extensive electron-density contri-
bution from the carbonyl group. Consequently, electron-defi-
cient catalysts tend to facilitate hydrogen abstraction from
electron poor substrates, whereas electron-rich catalysts
become more efficient in hydrogen re-addition in combination
with electron-rich substrates.[9]
Earlier studies have shown that the active catalytic species 6
are formed upon treatment of the precatalyst with tBuOK
(Scheme 1).[4b] The obtained complex 6 is then able to undergo
Although some attempts to describe the racemization kinet-
ics of these systems have been reported earlier,[4b,6a,8a,9–11] accu-
rate determination of the detailed quantitative reaction kinet-
ics remains challenging. From the experimental point of view,
the difficulties are mainly associated with the typically small
scale of racemization and DKR reactions and the often ques-
tionable reproducibility of organometallic reactions on such
minimal reaction scales in the presence of trace amounts of
oxygen and/or moisture, particularly in the presence of the re-
solving enzymes that typically contain water molecules in their
reaction spheres. Furthermore, particularly with the most
active racemization catalysts, such as 3a and 4a, complete rac-
emizations of chiral alcohol starting materials take place within
minutes posing challenges for the collection of a sufficient
number of data points for determination of quantitative reac-
tion kinetics. Accordingly, the earlier published kinetic work on
these catalyst systems, and the conclusions based on the
same, have often been of semi-quantitative character at the
most.
Scheme 1. Mechanism for racemization of sec-alcohols by (cyclopentadi-
enyl)ruthenium catalysts.
fast exchange of the alkoxide ligand with other alcohols even
at low temperatures.[8a] It has been proposed that the resulting
(R)-7 undergoes reversible hydride transfer from the alkoxide
to ruthenium leading to formation of the corresponding
ketone and hydride complex 3b.[4b] A significant primary iso-
tope effect observed in the racemization of a-deuteroalcohols
supports this hypothesis.[8a] The absence of ketone exchange
or deuterium scrambling in cross-over experiments was initially
explained by a solvent–cage effect preventing the ketone in-
termediate from exchange with the environment.[4b,8a] To eluci-
date the validity of this mechanism, a cross-over experiment
with a substrate bearing both keto and alcohol groups in the
same molecule has been performed. In this experiment, deute-
rium scrambling was not detected, indicating that the ketone
intermediate remains coordinated to the ruthenium center.[8a]
Moreover, a proposed ruthenium hydride intermediate 3b was
found to be inefficient for racemization of (S)-1-phenyletha-
nol.[4b] Based on these observations, it was concluded earlier
that a vacant coordination site for ketone coordination is gen-
erated upon CO dissociation (Scheme 1). Experiments on ex-
change of labeled CO with alkoxy complexes confirmed this
In the present work, we performed a detailed kinetic analysis
of the two ruthenium catalysts 3a and 4a, based on penta-
phenyl- and pentabenzyl-substituted cyclopentadienyl ligands,
respectively, aiming to elucidate their earlier observed differen-
ces in racemization of electronically modified sec-alcohols (see
below) and for investigating their true activities, limits, and
scope for the use in chemoenzymatic dynamic kinetic resolu-
tion reactions. As an additional reference catalyst, complex
5,[6a] containing a monobenzyltetraphenyl-substituted cyclo-
pentadienyl ligand, and with a racemization behavior closer to
that of the (pentaphenyl)cyclopentadienyl system 3a than to
that of 4a, has been included.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2437

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Results and Discussion
Table 1. Selected 1H and 13C NMR chemical shifts for complexes 3a, 3b,
4a, 4b, and 5.[a]
Background
Group
Nuclei
C5R5
13C
CO
13C
Ru-H
1H
In our earlier work,[6b] the precatalysts 3a and 4a were shown
to be nearly equally efficient for racemization of the parent (S)-
1-(phenyl)ethanol [(S)-9c], whereas in DKR of the electron-rich
sec-alcohol 1-(para-methoxyphenyl)ethanol (9a), the penta-
phenyl-substituted catalyst 3a was less active than its (penta-
benzyl)cyclopentadienyl analogue 4a. Detailed experimental
analyses demonstrated that the conversion of 9a in 3 h under
identical reaction conditions varied from 61 to 70% (6 consec-
utive experiments) with 3a and from 98 to >99% (4 consecu-
tive experiments) in the case of 4a. The starting material re-
mained racemic during the course of the reaction in the case
of 4a and slightly enantioenriched (ee ꢀ50%) in the case of
3a. This indicates that 4a is more efficient for racemization of
electron-rich substrates than 3a. Simultaneously, DKR of the
highly electron-deficient substrate 1-(para-nitrophenyl)ethanol
was fast with 3a and significantly slower with 4a. The starting
material remained almost enantiopure in the case of 4a and
racemic with 3a. Generally, compared to racemization semire-
actions, DKR forms a much more complex reaction system.
Therefore, direct correlation of data obtained from DKR experi-
ments with racemization activities per se is difficult. Neverthe-
less, the degree of enantiopurity of the starting material in
DKR can help in deducing whether the racemization reaction
appears as a rate–limiting step.
3a
3b
4a
4b
5
107.12
108.94
105.78
105.47
102.73
107.11
117.28
198.04
202.04
198.44
202.36
197.92
–
ꢁ9.39
–
ꢁ9.94
–
[a] In C6D6 expressed in ppm.
Figure 2. Negative parts of the cyclovoltammograms of 3a, 4a, and 5 (com-
plex concentration 2 mmolLꢁ1; DMF, Bu4NPF6 0.1 molLꢁ1; scan rate
10 mVsecꢁ1; auxiliary and working electrodes Pt; reference electrode Ag/
AgCl, ferrocenium/ferrocene conversion at +0.33 V).
As evident from Scheme 1, a prerequisite for efficient race-
mization is the rapid oxidation of the coordinated sec-alkoxide
to the corresponding ketone and the subsequent readdition of
hydrogen for generating the racemic alcohol.[8] An earlier study
by Bꢀckvall and co-workers showed electron-rich catalysts to
efficiently racemize electron-rich substrates because of the
more efficient hydrogen re-addition.[9] Park and co-workers
have reported the faster racemization of (S)-9c in the presence
of electron-rich catalysts.[10]
ty on the Ru center relative to that of 3a and 5. Moreover,
complex 4a demonstrated an irreversible cyclic voltammetry
reduction, whereas the reduction of 3a was quasi-reversible.
Complex 5 exhibited quasi-reversible reduction at high scan
rates and irreversible at low scan rates (for details see the Sup-
porting Information). Geometrically and structurally, as shown
by earlier X-ray structure determinations,[4b,6a] complexes 3a,
4a, and 5 were very similar displaying the expected RuꢁC5
centroid distances 1.893 ꢂ for 3a, 1.875 for 4a and 1.884 for
5.
The (pentabenzyl)cyclopentadienylruthenium complex 4a
and its corresponding hydride analogue 4b are more electron-
rich than the pentaphenyl- and tetraphenyl-substituted com-
plexes 3a, 3b, and 5, respectively. The electronic differences
become evident by inspection of the IR stretching frequencies
of the CO ligands in both 3a and 5 (2049, 2000 cmꢁ1) and 4a
1
(2043, 1992 cmꢁ1), respectively, as well as from the selected H
Preparation of chiral sec-alcohol starting materials for the
kinetic experiments
and 13C NMR chemical shifts collected in Table 1. In particular,
the 0.55 ppm difference in the Ru–H hydride resonance
1H NMR chemical shifts (ꢁ9.39 ppm for 3b versus ꢁ9.94 ppm
for 4b) is significant, considering the presence of ruthenium
hydride intermediate species in the catalytic cycle of the race-
mization process.
For the kinetic and racemization studies herein, a series of
enantiomerically pure para-substituted 1-(phenyl)ethanols, rep-
resenting both electron-rich (9a and 9b) and electron-deficient
sec-alcohols (9d and 9e), together with the parent 1-(phenyl)e-
thanol (9c), were selected and prepared (Scheme 2). In addi-
tion, the ortho-methyl substituted 1-(phenyl)ethanol 9 f was in-
cluded for investigating the possible influence of steric contri-
butions. In earlier work,[7d] we have described the >100 g-
scale, chromatography-free preparation of the enantiomerically
pure (R)-1-(phenyl)ethanol 9c by DKR of rac-1-(phenyl)ethanol
using CAL-B and catalyst 4a, followed by alkaline hydrolysis of
The electronic differences can further be verified by cyclic
voltammetry (Figure 2). The cyclovoltammogram of 4a signifi-
cantly differs from those of 3a and 5. The reduction peak of
4a (ꢁ1.70 V) occurred at a more negative potential than for
3a (ꢁ1.35 V) and 5 (ꢁ1.38 V). Such shift of the RuII–RuI reduc-
tion to a more negative potential by 0.32–0.35 V indicates that
the five benzyl substituents in 4a increased the electron densi-
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2438

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
If the reaction components were kept in a nitrogen
or argon Schlenk-line, frequent piercing of the
septum would increase the permeability towards at-
mospheric air and moisture, potentially decomposing
the highly sensitive transition-metal catalyst. Further-
more, although possible lowering of the catalyst/sub-
strate ratio might also be utilized for slowing down
the racemization rate to a measurable level, a sub-
stantial decrease of the catalyst loading on an already
small scale of the reaction would bring the catalyst
concentration to a microscale, on which the predicta-
bility and reproducibility of air- and moisture-sensi-
tive organometallic reactions becomes difficult.[12]
Scheme 2. Preparative-scale DKR of sec-alcohols used for the racemization studies.
the obtained acetate (R)-10c to (R)-9c in 93% overall yield and
>99% ee over two steps. A similar method was employed
herein for preparation of all other enantiomerically pure alco-
hols 9a–f, obtained in 88–96% overall yields and, in most
cases, ee values equal to or exceeding 99%, further demon-
strating the feasibility of catalyst 4a for preparative scale DKR
applications (see the Supporting Information). In preliminary
racemization screening studies, the most electron-deficient
substrate 9e was shown to have the slowest racemization rate
in the presence of 4a, which makes it most suitable for de-
tailed kinetic investigation. For this reason, the deuterated ana-
logue d1-9e was also prepared for investigating the existence
of possible kinetic isotope effects.
In the initial screening studies for suitable experimental con-
ditions, as a feasible approach, an increase in the loading of
the optically active alcohol (R)-9c to 1 g (8 mmol) whereas re-
taining the catalyst loading on 20 mmol (0.25 mol%) level was
observed to increase the time required for complete racemiza-
tion to 10–15 min, allowing for sufficiently frequent, multiple
sampling to obtain reliable and reproducible kinetic data. Even
under such conditions, in a glove box, another critical issue
could be the possible decomposition or changes in the enan-
tiomeric excess of the product sample prior to chiral GC analy-
sis. In the case of relatively slow DKR experiments, this issue
can be solved by withdrawing the sample from the glove box
followed by irreversible quenching with a suitable derivatizing
agent that stops the racemization reaction (Scheme 3). For re-
actions lasting some hours such minute delays induced by
Experimental setup for the racemization reactions
Generally, the kinetic data obtainable from DKR reactions are
challenging to interpret because of the highly complex reac-
tion network consisting of a large number of components. Fur-
thermore, although routinely employing homogeneous racemi-
zation catalysts, chemoenzymatic DKR processes also contain
heterogeneous catalysts and reagents, including immobilized
enzymes and sodium carbonate, which both may induce diffu-
sion limitations to the rate of reaction. For simplifying the reac-
tion system, and for obtaining reliable data on the influence of
racemization catalyst on the overall DKR performance, we in-
stead turned into studying the actual racemization reaction.
The ruthenium-catalyzed racemization step of chemoenzymatic
DKR is homogeneous and, therefore, more suitable to be sub-
jected to a detailed and accurate kinetic study.
Scheme 3. Derivatization of the racemization reaction mixture for chiral GC
analysis.
time taken for withdrawal of the samples from the glove box
prior to analysis are not crucial. For the rapid 10–15 min race-
mization reactions, however, delays of even a few minutes in
quenching and derivatization would induce a significant exper-
imental error. As highly reactive and corrosive derivatizing
agents cannot be directly introduced into the glove-box, the
issue was here solved by tightly sealed GC vials containing the
derivatizing agent sealed with a septum. This procedure al-
lowed the withdrawal of samples directly from the reaction
mixture in the glove box by the use of a disposable syringe,
followed by immediate injection to the GC vial containing the
irreversible derivatization agent, thereby freezing the enantio-
meric ratio in the sample to the time point of sampling.
Owing to the generally fast reaction rates, investigation of
the racemization kinetics is, likewise, a challenging task. Under
typical reaction conditions using 2 mol% of homogeneous
ruthenium racemization catalysts, such as 3a or 4a, together
with 1 mmol of the sec-alcohol substrate, complete racemiza-
tion is observed within 2 min, severely limiting the possible
number of data points available for kinetic modeling. Addition-
ally, to ensure maximal reproducibility, and taking into account
the frequent sampling of the reaction mixture required for ac-
curate observation and determination of the reaction kinetics,
the racemization studies should ideally be performed in a nitro-
gen-filled glove box. This requirement, in turn, makes lowering
of the reaction temperature as means for slowing down the
rate of racemization to measurable level impractical.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2439

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Kinetic experiments
catalyst 4a. In full accordance with the earlier observations
and predictions of Bꢀckvall and coworkers on the dependence
of racemization rate on the electronic properties of both the
substratre and the catalyst,[9] the more electron-rich (pentaben-
zyl)cyclopentadienyl catalyst 4a racemized the electron-rich
substrates (R)-9a and (R)-9b at one order of magnitude faster
rates compared to the (pentaphenyl)cyclopentadienyl catalyst
3a (Figure 4 and Supporting Information). Time required to
achieve 50% ee in the presence of 3a was 20 and 10 min in
the case of (R)-9a and (R)-9b, correspondingly. Notably, under
the conditions employed, the racemization kinetics of (R)-9a
and (R)-9b in the presence of 3a did not fully corresponded to
the expected 1st order reaction and the estimated rate con-
stants for these experiments were, therefore, excluded from
Table 2. The observed behavior could possibly be explained by
The racemization reactions for the kinetic experiments were
performed by using an amount of 8 mmol of the correspond-
ing chiral alcohol 9a–f at 0.25 mol% ruthenium catalyst load-
ing. As shown earlier by Bꢀckvall and others, the dicarbonyl-
chlororuthenium precatalyst is first converted to the active
species 6 by alkoxylation reaction with tBuOK[4b] or some other
suitable alkoxide.[11a] Herein, the precatalysts 3a and 4a were
first reacted with 1.25 equivalents of tBuOK and allowed to stir
for 5 min before initiating the racemization reactions. All reac-
tions were performed in duplicate, unless indicated otherwise.
The rate expression for the racemization reaction rate can be
written as Equation (1), in which [R] and [S] are the concentra-
tions of the (R)- and (S)-alcohols, respectively.[13]
k
þ1kþ2k ½Rꢂ½COꢂ ꢁ kꢁ1kꢁ2k ½Sꢂ½COꢂ½Catꢂ
þ3
ꢁ3
ð1Þ
r ¼
k
þ1kþ2 þ ðkþ1kþ2 þ kꢁ1kꢁ2Þ½COꢂ þ ðkþ2k ½COꢂ þ kþ1kþ3 þ kþ1k ½COꢂÞ½Rꢂ þ ðkꢁ2k ½COꢂ þ kꢁ2kꢁ3 þ kꢁ1k ½COꢂÞ½Sꢂ
þ1
þ3
ꢁ3
ꢁ3
Owing to the identical racemization mechanisms for both sub-
strate enantiomers in the presence of an achiral catalyst, some
simplifications can be made. First, the rate constant for CO ab-
straction from (R)-7 to form 8 is equal to the rate constant of
Table 2. Rate constants of the racemization reactions.[a,b]
Substrate
Racemization Catalyst and Rate Constants[c]
4a
3a
5
[d]
[d]
[d]
(R)-9a
(R)-9b
(R)-9c
(R)-9d
(R)-9e
(R)-9 f
4.19ꢃ0.08
CO abstraction from (S)-7, i.e., k+1 =kꢁ2. Analogously, kꢁ1 =k+2
.
8.56ꢃ0.24
5.75ꢃ0.14
5.00ꢃ0.10
1.14ꢃ0.06
3.49ꢃ0.07
0.71ꢃ0.05
Both species (R)-7 and (S)-7, as well as the (R)- and (S)-alcohols,
1.30ꢃ0.06
2.42ꢃ0.05
5.40ꢃ0.12
6.73ꢃ0.20
6.85ꢃ0.28
0.84ꢃ0.06
1.66ꢃ0.06
3.08ꢃ0.06
have similar energies and thus k+3 =k . The mass balance
ꢁ3
analysis for CO generation is revealed in Equation (2):
d1-(R)-9e
d½COꢂ
ð2Þ
¼ rþ1 ꢁ rꢁ1 ꢁ rþ2 þ r
ꢁ2
dt
[a] Calculated according to Equation (5). [b] (R)-5 (8 mmol) with catalyst
(0.25 mol%) at ambient temperature. [c] 103 ꢃk secꢁ1. [d] Rate constants
not determined, owing to the reaction not fully following 1st order.
Equation (2) is equal to the steady-state approximation for 8.
Therefore, d[CO]/dt=0 and [CO] is constant. After simplifica-
tion of Equation (1), Equation (3) is obtained:
d½Sꢂ
d½Rꢂ
r ¼ k0ð½Rꢂ ꢁ ½SꢂÞ ¼
¼ ꢁ
ð3Þ
dt
dt
in which k’ is an apparent rate constant:
k
þ1kþ2k ½COꢂ½Catꢂ
k
þ1kþ2 þ 2kþ1k ½COꢂ þ C0ðþk3þ1k3 þ ½COꢂÞðkþ2kþ3 þ kþ1kþ3Þ
ð4Þ
þ2
C0 ¼ ½Rꢂ þ ½Sꢂ ¼ const:
Solving Equation (3) gives Equation (5) in which [R]0 =the initial
concentration and [R]f =the final concentration of the (R)-enan-
tiomer at equilibrium. Owing to identical reactions of both
enantiomers with the achiral catalyst, [R]f =[S]f =50%. Thus,
plot of ꢁln([R]ꢁ[R]f) versus time yields a straight line with the
slope equal to 2k’.
Figure 3. Plot for determination of the rate constant for racemization of (R)-
9a in the presence of catalyst 4a.
½Rꢂ0 ꢁ ½Rꢂf
¼ 2k0t
ð5Þ
ln
½Rꢂ ꢁ ½Rꢂf
the existence of an induction period (see the Supporting Infor-
mation). Another explanation could be the poor solubility of
catalyst 3a becoming a limiting factor with substrates race-
mized slower with this catalyst. Earlier, Bꢀckvall observed that
The calculated racemization rate constants are summarized
in Table 2. An example plot showing the goodness of fit is
shown in Figure 3 for racemization of (R)-9a in the presence of
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2440

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 6. Racemization of (R)-9d with catalysts 3a and 4a.
Figure 4. Racemization of (R)-9a and (R)-9b with catalyst 4a.
extended reaction time was in some cases required for activa-
tion of catalyst 3a with tBuOK.[9] To rule out the possible influ-
ence of an induction or activation period, additional experi-
ments with 30 min activation of 3a were performed. Differen-
ces in the racemization rates were, however, not observed be-
tween the racemization experiments on substrate (R)-9a using
the 5 versus 30 min activation periods for catalyst 3a. Never-
theless, (R)-9a, in the presence of 3a, clearly demonstrated the
slowest racemization rate. With the parent unsubstituted 1-
(phenyl)ethanol (R)-9c and the slightly electron-deficient (R)-
9d, differences in the racemization rate constants between 3a
and 4a were less pronounced, with 4a being four and two
times faster than 3a for substrates (R)-9c and (R)-9d, respec-
tively (Figure 5 and 6).
the deuterated 1-(para-trifluoromethylphenyl)-1-d-ethanol d1-
(R)-9e with 4a was 1.6ꢃ0.2 times smaller than the corre-
sponding rate constant for racemization of the nondeuterated
alcohol (R)-9e (Figure 7). This observation further supports the
hypothesis of hydride abstraction being the rate-limiting step
in the racemization of electron-deficient substrates.[14] With the
faster electron-deficient catalyst 3a, similar significant isotope
effect was not observed in the racemization of d1-(R)-9e vs. (R)-
9e (Figure 7), possibly owing to the much faster reaction rate
which makes rate measurements less accurate.
Figure 7. Racemization of (R)-9e and d1-(R)-9e with catalysts 3a and 4a.
In the case of the ortho-methyl substituted alcohol (R)-9 f,
the rate of racemization with 4a was slower than the rate of
racemization of the para-methyl substituted substrate (R)-9b
with the same catalyst (Figure 8). With the (pentaphenyl)cyclo-
pentadienyl catalyst 3a, however, the racemization of the
ortho-methyl substituted phenylethanol (R)-9 f was faster than
the racemization of the para-methyl substituted analogue (R)-
9b. These observations may possibly be explained by shielding
of the ruthenium center by the pendant benzyl substituent
from interactions with the sterically congested alcohol 9 f.
Direct, through-space coordination of a pentabenzylcyclopenta-
Figure 5. Racemization of (R)-9c with catalysts 3a and 4a.
With the electron-poor substrate (R)-9e, however, as expect-
ed, the more electron-deficient (pentaphenyl)cyclopentadienyl
catalyst 3a displayed a five times faster racemization rate than
4a (Figure 7). Interestingly, the rate constant for racemization
of (R)-9e with 3a was very close to the rate constant for race-
mization of the electron-rich substrate (R)-9a with the more
electron-rich catalyst 4a. The rate constant for racemization of
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2441

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 8. Racemization of (R)-9 f with catalysts 3a and 4a.
Figure 10. Racemization of (R)-9e with catalysts 3a and 5.
dienyl ligand to a metal center by stabilizing interaction of
a C=C bond in one of the benzylic phenyl groups of the ligand
with molybdenum and tungsten was described recently.[15]
Similar intramolecular arene coordination is also well known
for Ti- and Zr-based mono(cyclopentadienyl) olefin polymeri-
zation catalysts.[16] In principle, such interaction may even have
a stabilizing effect in the catalytic racemization cycle which,
prior to the currently prevailing mechanistical understanding
summarized in Scheme 1, was earlier proposed to involve an
h5!h3 ring slippage of the cyclopentadienyl ligand for genera-
tion of the vacant coordination site for b-elimination.[4b,17]
Finally, for comparison, the (monobenzyltetraphenyl)cyclo-
pentadienyl ruthenium catalyst 5 (Figure 1), likewise described
and characterized in our earlier work,[6a] was screened for race-
mization of several substrates (Figure 9 and 10). The structural
Figure 11. Hammett plots of the racemization rates of sec-alcohols with cata-
lysts 3a and 4a.
to that of the (pentaphenyl)cyclopentadienyl catalyst 3a, al-
though slightly lower racemization rates were generally
observed.
A Hammett plot of the racemization rates for catalysts 3a
and 4a is displayed in Figure 11. With the Bꢀckvall catalyst 3a,
as evident from the linear Hammett plot, the reaction rate in-
creased with the electron-withdrawing character of the para-
substituent of the phenylethanol substrate, consistent with hy-
dride re-addition as the rate-determining step for racemization
(see Scheme 1) with same reaction mechanism for all sub-
strates. Here, as discussed above, the data points for substrates
9a and 9b were omitted from the plot as the corresponding
reactions with these catalysts did not fully follow first order re-
action kinetics. Nevertheless, based on the experimental data
(for details, see the Supporting Information) these data points
can clearly be assumed to reside below the data points from
all other substrates. The nonlinear, concave downward plot
with catalyst 4a, however, can be associated with a change in
the rate-determining step when proceeding from electron-do-
nating to electron-withdrawing substituents on the phenyle-
thanol substrate with otherwise similar reaction mechanism.[18]
Figure 9. Racemization of (R)-9c and (R)-9d with catalysts 3a and 5.
features of this catalyst, as elucidated by earlier X-ray structure
determination,[6a] are similar to those of 3a[4b] and 4a,[6a]
whereas electronically this complex, with a single benzyl sub-
stituent, is closer to the (pentaphenyl)cyclopentadienyl catalyst
3a as evidenced by CO ligand IR stretching frequencies and
the cyclic voltamogramms (see above, Table 1, and Figure 2).
The racemization activity and behavior of 5 was found similar
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2442

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
1
With the electron-poor substrates, the racemization rate was
limited by the hydride abstraction (step A in Scheme 1), where-
as with the most electron-rich para-methoxy substituted sub-
strate (R)-9a the hydrogen re-addition (step B) became slow.
easily identifiable signals were not present in the H NMR spec-
trum. Moreover, with this complex, unidentified side product
or decomposition products precipitated from the solution
during the reaction with both lithium and potassium alkoxides
hampering the collection of any meaningful data. In the litera-
ture, the para-methylphenyl analogue of the Shvo catalyst
1 was reported to give characteristic methyl proton resonances
in 1H NMR.[19] Thus, for possible future mechanistical studies,
the tetraphenylbenzylcyclopentadienyl complex 5 could op-
tionally prove to be a suitable alternative and model com-
pound for 3a because of its similar racemization properties
but higher solubility, displaying a sharp resonance from the
NMR studies on catalyst activation
Additional NMR experiments were performed in an attempt to
elucidate the formation of intermediates in the catalytic race-
mization cycle. As shown earlier by Bꢀckvall and co-workers,[4-
b,8a,c]
the chlorodicarbonylruthenium precatalyst initially forms,
upon activation with tBuOK, a tert-butoxide complex which
then undergoes ligand exchange with the substrate alcohol to
be racemized (Scheme 1). Alternatively, as also described in
earlier work,[11a] a preformed ruthenium alkoxide of the sub-
strate undergoing racemization can be used directly without
the preactivation step. For preparative DKR purposes, however,
preactivation with the inexpensive tBuOK has obvious advan-
tages. For mechanistic work and NMR studies, on the other
hand, direct utilization of the ruthenium alkoxide of the sub-
strate to be studied simplifies the system by exclusion of the
unnecessary tert-butoxide–alcohol exchange step and by de-
creasing the number of components in the reaction mixture.
Here, the reactions between the (pentabenzylcyclopentadie-
nyl)ruthenium catalyst 4a and alkali-metal alkoxides generated
from racemic 1-(phenyl)ethanol were studied by 1H and
13C NMR. For these experiments, d6-benzene was used as a sol-
vent because of its simple NMR spectrum and high melting
point simplifying the freezing of the NMR samples prior to
flame sealing of the NMR tubes. The major products observed
in these experiments were the corresponding ruthenium hy-
dride and acetophenone (Scheme 4).
1
CH2 benzylic protons in the H NMR spectrum.
At first sight, the observation of the ruthenium hydride spe-
1
cies 4b in the H NMR spectrum could imply that such hydride
would also form as an intermediate in the racemization cycle.
However, a model racemization reaction of (R)-9c using
2 mol% of 4b as a catalyst proceeded very slowly, providing
<1% conversion in 3 h, whereas under identical conditions 4a
catalyzed a complete racemization within minutes. The addi-
tion of acetophenone as a possible hydride transfer agent[2b]
did not facilitate the racemization. Thus, the ruthenium hydride
complex 4b was, most likely, not an active intermediate or
species in the racemization reaction. The appearance of this
1
species in the H NMR spectrum was more likely owing to de-
composition of the intermediates formed from reactions be-
tween the starting lithium or potassium alkoxides and the pre-
catalyst 4a.
Conclusions
In this work, detailed kinetic studies on the racemization of
sec-alcohols with three homogeneous ruthenium cyclopenta-
dienyl catalysts have been performed with particular emphasis
on elucidating the differences between the two high-perfor-
mance catalysts Ph5CpRu(CO)2Cl (3a) and Bn5CpRu(CO)2Cl (4a)
studied earlier in detail for chemoenzymatic dynamic kinetic
resolution reactions. Herein, a reliable protocol for this compa-
rative study was developed, which allowed the determination
of accurate racemization rate constants for a range of electron-
ically modified, substituted 1-phenylethanols.
Scheme 4. Formation of Ru–H species from the reactions of catalyst precur-
sors 3a and 4a with alkali metal alkoxides of racemic 1-(phenyl)ethanol.
Comparison of the two catalysts 3a and 4a and their corre-
sponding hydride analogues by electrochemical and spectro-
scopic methods revealed clear differences between their elec-
tronic properties whereas structurally the two complexes were
closely related, as shown by earlier X-ray structure determina-
tions. Essentially, the (pentabenzyl)cyclopentadienyl-ligated
complex 3a possessed much higher electron density at the
ruthenium metal center compared to the more electron-defi-
cient (pentaphenyl)cyclopentadienyl analogue 4a. As shown
by the kinetic experiments, these electronic differences were
manifested in the racemization behavior of the two catalysts.
With this type of homogeneous ruthenium catalysts, the race-
mization of sec-alcohols proceeded through a two-step oxida-
tion–reduction mechanism in which, after generation of
a vacant coordination site through CO ligand dissociation, b-
hydride elimination from the alcohol alkoxide formed a rutheni-
The reaction of 4a with 1-(phenyl)ethoxylithium proceeded
very slowly at room temperature and required heating to 708C
at which temperature the reaction took place overnight. After
three days, a 6:1 ratio between the starting material 4a and
ruthenium hydride 4b was observed. Further heating to 708C
gave after 12 h a 1:2 ratio of 4a and 4b. Consequently, potassi-
um hydride was next used for preparation of 1-(phenyl)ethoxy-
potassium, which, compared with its lithium analogue, reacted
much faster with 4a resulting in complete transformation to
4b and acetophenone at room temperature in 2 h. The con-
1
version of 4a to 4b could be followed by H NMR spectrosco-
py, owing to the sharp and intense singlet resonance at
3.46 ppm (4a) and 3.62 ppm (4b), respectively, from the CH2
group of the five equivalent benzyl substituents. In the case of
the (pentaphenyl)cyclopentadienyl complex 3a, corresponding
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2443

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
hydride under argon atmosphere. Enantiomerically pure 1-phenyle-
thanol (R)-5c for racemization experiments was prepared as de-
scribed earlier.[7d] The racemic para-substituted 1-phenylethanols
rac-5a, rac-5b, rac-5d, rac-5e and rac-5 f were prepared by reduc-
tion of the corresponding ketones with sodium borohydride. The
deuterated racemic analogue d1-(R)-5e was prepared by reduction
of para-trifluoromethyl acetophenone with LiAlD4. The racemic
compounds used for the DKR reactions and the resolved enantio-
merically pure phenylethanols for racemization experiments were
redistilled over CaH2 under reduced pressure prior to use. Sodium
carbonate and the 4 ꢂ molecular sieves were dried at 1808C in
vacuum prior to use. C. antarctica lipase B in the form of Novozym-
435 was obtained from Novozymes A/S (Denmark). Racemization
catalyst 3a,[4a] Ph4BnC5H[6a] and Bn5CpH[7] ligands were prepared es-
sentially as previously described. The ruthenium hydride complex
3b was prepared according to a previously published method[4b]
except by using mesitylene as the reaction solvent instead of tolu-
ene–decane. Improved large-scale synthesis procedures for cata-
lysts 4a and 7 are described in the Supporting Information. NMR
spectra were recorded by using a 600 MHz NMR spectrometer
equipped with a BBI-5 mm-Zgrad-ATM probe or BBO-5 mm-Zgrad
um hydride species with coordinated ketone, which then, after
hydride re-addition and alkoxide exchange, released the race-
mized alcohol. In this process, the electronic properties of the
catalyst and the substrate influenced the rate-determining step
of the racemization process. As suggested earlier, and verified
by detailed kinetic studies in this work, with the electron-defi-
cient catalyst 3a, the hydride re-addition step was likely to be
rate-limiting whereas the hydride abstraction proceeds easily.
Therefore, with this catalyst for electronically modified 1-phe-
nylethanols, a close to linear Hammett plot was obtained with
a clear trend towards a positive slope. The electron-rich cata-
lyst 4a, however, showed a remarkable nonlinear Hammett be-
havior with difficult hydrogen abstraction from electron-defi-
cient sec-alcohols with simultaneously facilitated hydrogen
readdition. By switching from moderately electron-rich sub-
strates to highly electron-deficient ones, the Hammett-plot for
4a showed a negative slope. With highly electron-rich sub-
strates, the racemization rate with this catalyst decreased lead-
ing to a concave downwards Hammett-plot consistent with
similar reaction mechanism with gradual change in the rate-
limiting step. The rate-limiting role of hydride abstraction–re-
addition with catalyst 4a was further supported by a significant
primary isotope effect absent with the (pentaphenyl)cyclopen-
tadienyl analogue 3a.
1
probe at 298 K operating at 600.13 MHz for H and 150.92 MHz for
13C analysis. Conversions and enantiomeric excesses were mea-
sured by using a chiral GC flame ionization detector equipped with
a capillary b-dextrine chiral column (25.0 mꢃ250 mmꢃ0.25 mm).
The kinetic experiments were conducted in a thermostated glove
box at constant temperature (238C).
Furthermore, in the present work, a large-scale chromatogra-
phy-free procedure for preparation of the Bn5CpRu(CO)2Cl cata-
lyst was developed (for details see the Supporting Informa-
tion), exhibiting, consistent with our earlier reports, considera-
ble preparative advantages compared to syntheses of other
similar ruthenium catalysts including 3a, making catalyst 4a
an attractive candidate for large-scale chemoenzymatic dynam-
ic kinetic resolutions of sec-alcohols. For the present work, cat-
alyst 4a was successfully employed for multigram-scale prepa-
rations (20–140 g) of several chiral sec-alcohols in 91–96%
overall yields over two steps (enzymatic acylation followed by
hydrolysis) and 92–99% ee using only 0.05–0.2 mol% loadings
of the ruthenium catalyst (for details, see the Supporting Infor-
mation). In the enzymatic reactions, conversions close to quan-
titative allowed the conducting of the alkaline hydrolysis of
the acetates obtained without losses in enantiopurity. All chiral
sec-alcohol products were then isolated by distillation without
the use of chromatography.
Procedure for racemization reactions
All operations were performed in a glove box. In a typical racemi-
zation experiment, an amount of 20 mmol of the catalyst (3a,
12.8 mg; 4a, 14.1 mg; and 7, 13.0 mg), a magnetic stirring bar and
toluene (10 mL) were placed in a test tube (inner diameter
20 mm). After dissolution of the catalyst, a 0.25m solution of
tBuOK in THF (100 mL, 25 mmol) was added and the reaction mix-
ture was stirred for 5 min. Next, an amount of 8 mmol of the chiral
alcohol substrate was quickly added by a syringe and the reaction
started. Samples were withdrawn by using a single-use syringe and
injected into sealed GC vials containing a limited volume insert
with propionic anhydride (50 mL) and 1% solution of DMAP in pyri-
dine (20 mL). After injection of the sample into the vial, the vial was
sealed in a closed container to avoid evaporation of the harmful
derivatizing agents into the glovebox atmosphere.
NMR experiments
The results obtained in this work can be applied for further
optimization of ruthenium racemization catalysts for chemoen-
zymatic dynamic kinetic resolution reactions by better match-
ing of the electronic properties of the catalyst with the proper-
ties of the substrate. The accurate preparative kinetic method-
ology developed here for determination of rate constants for
fast catalytic reactions should also prove valuable for other re-
lated or similar catalytic systems.
A 0.5m solution of the alkoxide was prepared from 1-(phenyl)etha-
nol and nBuLi or KH in THF. The obtained solution was filtered
through a glass filter, after which 1.1 equivalents were transferred
into a Schlenk tube. The solvent was removed in vacuo and the
residue co-evaporated with benzene to eliminate THF signals from
the NMR spectrum. A solution of 1 equivalents of 4a in C6D6
(0.6 mL) was added, the mixture was stirred at RT and transferred
into an NMR tube, which was immediately sealed. If necessary,
a sealed NMR tube was heated in an oil bath at the desired tem-
perature.
Experimental Section
Acknowledgements
All glassware was oven-dried at 1508C overnight and cooled down
in a desiccator over phosphorus pentoxide. Solvents were dried ac-
cording to standard procedures. Potassium tert-butoxide was subli-
mated in vacuum. Isopropenylacetate was redistilled over calcium
Michal Wagner and Dr. Anton Tokarev are thanked for their assis-
tance with the electrochemical studies and Prof. Ari Ivaska for
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2444

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[8] a) B. Martꢆn-Matute, J. B. ꢂberg, M. Edin, J.-E. Bꢀckvall, Chem. Eur. J.
2007, 13, 6063–6072; b) J. Nyhlꢅn, T. Privalov, J.-E. Bꢀckvall, Chem. Eur.
J. 2009, 15, 5220–5229; c) J. B. ꢂberg, J. Nyhlꢅn, B. Martꢆn-Matute, T. Pri-
valov, J.-E. Bꢀckvall, J. Am. Chem. Soc. 2009, 131, 9500–9501; d) J.
Bosson, S. P. Nolan, J. Org. Chem. 2010, 75, 2039–2043; e) M. C. Warner,
O. Verho, J.-E. Bꢀckvall, J. Am. Chem. Soc. 2011, 133, 2820–2823.
[9] O. Verho, E. V. Johnston, E. Karlsson, J.-E. Bꢀckvall, Chem. Eur. J. 2011, 17,
11216–11222.
[10] J. H. Lee, N. Kim, M.-J. Kim, J. Park, ChemCatChem 2011, 3, 354–359.
[11] a) Q. Xi, W. Zhang, X. Zhang, Synlett 2006, 945–947; b) M. Eckert, A.
Brethon, Y.-X. Li, R. A. Sheldon, I. W. C. E. Arends, Adv. Synth. Catal. 2007,
349, 2603–2609; c) M. Merabet-Khelassi, N. Vriamont, L. Aribi-Zouioue-
che, O. Riant, Tetrahedron: Asymmetry 2011, 22, 1790–1796.
[12] A catalyst loading of 2 mol% on a 1 mmol scale corresponds to
20 mmol (10–15 mg) of the organometallic compound and a further de-
crease of the catalyst amount would be questionable in terms of repro-
ducibility. Possibilities to increase the amount of substrate are likewise
limited because of the relatively high price of enantiomerically pure
sec-alcohols required for the racemization studies.
[13] D. Murzin, T. Salmi, Catalytic Kinetics, Elsevier Science & Technology
Books, 2005.
[14] E. M. Simmons, J. F. Hartwig, Angew. Chem. 2012, 124, 3120–3126;
Angew. Chem. Int. Ed. 2012, 51, 3066–3072.
[15] M. A. Antunes, L. G. Alves, S. Namorado, J. R. Ascenso, L. F. Veiros, A. M.
Martins, Organometallics 2012, 31, 4387–4396.
[16] See, for example: a) J. Sassmannshausen, J. A. K. Powell, C. E. Anson, S.
Wocadlo, M. Bochmann, J. Organomet. Chem. 1999, 592, 84–94;
b) P. J. W. Deckers, B. Hessen, J. H. Teuben, Angew. Chem. 2001, 113,
2584–2587; Angew. Chem. Int. Ed. 2001, 40, 2516–2519; c) P. J. W. Deck-
ers, B. Hessen, J. H. Teuben, Organometallics 2002, 21, 5122–5135.
[17] See also: D. Mavrynsky, R. Sillanpꢀꢀ, R. Leino, Organometallics 2009, 28,
598–605.
[18] a) J. O. Schreck, J. Chem. Educ. 1971, 48, 103–107. For other recent ex-
amples of nonlinear Hammet effects in organometallic catalysis, see:
b) P. Fristrup, S. Le Quement, D. Tanner, P.-O. Norrby, Organometallics
2004, 23, 6160–6165; c) T. Furuya, D. Benitez, E. Tkatchouk, A. E. Strom,
P. Tang, W. A. Goddard III., T. Ritter, J. Am. Chem. Soc. 2010, 132, 3793–
3807.
fruitful discussions. Financial support from the Finnish Funding
Agency for Technology and Innovation (TEKES) (Technology Pro-
gramme SYMBIO project #40168/07: Developing New Chemoen-
zymatic Methods and Biocatalysts) is gratefully acknowledged.
Keywords: alcohols · cyclopentadienyl ligands · kinetics ·
racemization · ruthenium
[1] a) K. Bogꢄr, J.-E. Bꢀckvall, Tetrahedron Lett. 2007, 48, 5471–5474; b) P.
Krumlinde, K. Bogꢄr, J.-E. Bꢀckvall, J. Org. Chem. 2009, 74, 7407–7410;
c) E. V. Johnston, K. Bogꢄr, J.-E. Bꢀckvall, J. Org. Chem. 2010, 75, 4596–
4599; d) L. K. Thalꢅn, A. Sumic, K. Bogꢄr, J. Norinder, A. K. ꢂ. Persson, J.-
E. Bꢀckvall, J. Org. Chem. 2010, 75, 6842–6847; e) A. Trꢀff, R. Lihammar,
J.-E. Bꢀckvall, J. Org. Chem. 2011, 76, 3917–3921; f) R. Lihammar, R.
Millet, J.-E. Bꢀckvall, Adv. Synth. Catal. 2011, 353, 2321–2327. For recent
reviews, see: g) B. Martꢆn-Matute, J.-E. Bꢀckvall, Curr. Opin. Chem. Biol.
2007, 11, 226–232; h) J. H. Lee, K. Han, M.-J. Kim, J. Park, Eur. J. Org.
Chem. 2010, 999–1015.
[2] a) A. L. E. Larsson, B. A. Persson, J.-E. Bꢀckvall, Angew. Chem. 1997, 109,
1256–1258; Angew. Chem. Int. Ed. Engl. 1997, 36, 1211–1212; b) B. A.
Persson, A. L. E. Larsson, M. Le Ray, J.-E. Bꢀckvall, J. Am. Chem. Soc.
1999, 121, 1645–1650.
[3] a) J. H. Choi, Y. K. Choi, Y. H. Kim, E. S. Park, E. J. Kim, M.-J. Kim, J. Park, J.
Org. Chem. 2004, 69, 1972–1977.
[4] a) G. Csjernyik, K. Bogꢄr, J.-E. Bꢀckvall, Tetrahedron Lett. 2004, 45, 6799–
6802; b) B. Martꢆn-Matute, M. Edin, K. Bogꢄr, F. B. Kaynak, J.-E. Bꢀckvall,
J. Am. Chem. Soc. 2005, 127, 8817–8825.
[5] K. Bogꢄr, B. Martꢆn-Matute, J.-E. Bꢀckvall, Beilstein J. Org. Chem. 2007, 3,
No. 50.
[6] a) D. Mavrynsky, M. Pꢀiviç, K. Lundell, R. Sillanpꢀꢀ, L. T. Kanerva, R.
Leino, Eur. J. Org. Chem. 2009, 1317–1320; b) M. Pꢀiviç, D. Mavrynsky,
R. Leino, L. T. Kanerva, Eur. J. Org. Chem. 2011, 1452–1457.
[7] Pentabenzylcyclopentadiene can be prepared on large scale in high
yield by alkylation of cyclopentadiene with benzyl alcohol under basic
conditions facilitated by azeotropic water removal. For original proce-
dure, see: a) S. S. Hirsch, W. J. Bailey, J. Org. Chem. 1978, 43, 4090–4094.
For improved procedures, see: b) J. W. Chambers, A. J. Baskar, S. G. Bott,
J. L. Atwood, M. D. Rausch, Organometallics 1986, 5, 1635–1641; c) W.-
M. Tsai, M. D. Rausch, R. D. Rogers, Organometallics 1996, 15, 2591–
2594. See also, ref. 6a and: d) D. Mavrynsky, L. T. Kanerva, R. Sillanpꢀꢀ, R.
Leino, Pure Appl. Chem. 2011, 83, 479–487. The synthesis of other poly-
aryl-substituted cyclopentadienyl ligands such as (pentaphenyl)cyclo-
pentadiene commonly involve the use of expensive starting materials
and air-sensitive reagents (e.g., PhMgBr and LiAlH4).
[19] P. C. Casey, S. E. Beetner, J. B. Johnson, J. Am. Chem. Soc. 2008, 130,
2285–2295.
Received: March 7, 2013
Revised: April 20, 2013
Published online on May 24, 2013
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2436 – 2445 2445