Asymmetric C-C bond-formation reaction with Pd: How to favor heterogeneous or homogeneous catalysis?

Article abstract of DOI:10.1002/chem.201000833

The enantioselective allylic alkylation of (E)-1,3-diphenylallyl acetate was studied to clarify the heterogeneous or homogeneous character of the Pd/Al₂O₃-(R)-BINAP catalyst system. A combined approach was applied: the catalytic test

Full text of DOI:10.1002/chem.201000833

DOI: 10.1002/chem.201000833
À
Asymmetric C C Bond-Formation Reaction with Pd: How to Favor
Heterogeneous or Homogeneous Catalysis?
Sven Reimann,[a] Jan-Dierk Grunwaldt,[b, c] Tamas Mallat,[a] and Alfons Baiker*[a]
Abstract: The enantioselective allylic
alkylation of (E)-1,3-diphenylallyl ace-
tate was studied to clarify the heteroge-
neous or homogeneous character of the
considerable coverage of the Pd sur-
face by the bulky compound slows
down the initial reduction of the sur-
face oxides but BINAP itself may con-
sume surface oxygen (through its con-
version to BINAPO and BINAPO2)
and contribute to the maintenance of
the active metal surface during the re-
action. Carrying out the reaction under
pressure in an inert gas atmosphere is
important to minimize the oxygen dif-
fusion into the reaction mixture and to
avoid leaching. The (known) effect of
temperature is critical as well: our cat-
alyst system is inactive at room temper-
ature, which is a clear deviation from
the behavior of the corresponding ho-
mogeneous system. In contrast, halo-
genated solvents are easily dehalogen-
ated on Pd/Al2O3 and thus they favor
leaching of the metal and formation of
soluble compounds, analogous to classi-
cal metal corrosion in the presence of
halide ions. The frequently observed
dissolution of Pd in the presence of
halogenated substrates may be ex-
plained similarly.
À
Pd/Al2O3 (R)-BINAP catalyst system.
A combined approach was applied: the
catalytic tests were completed with in
situ XANES measurements to follow
the oxidation state of Pd as a function
of the reaction conditions. The study
revealed that the oxidized Pd (after ex-
posure to ambient air) is efficiently re-
duced by the solvents THF and diox-
ane, and by the nucleophile sodium di-
methyl malonate, and thus these condi-
tions prevent Pd leaching. The chiral
modifier BINAP plays a dual role: a
Keywords: asymmetric synthesis ·
heterogeneous catalysis
·
metal
leaching · palladium · X-ray diffrac-
tion
Introduction
sors, but for industrial applications heterogeneous catalysts
are often preferred.[2] An important obstacle for any practi-
cal application of homogeneous or heterogeneous catalysts
is the relocation of the active species, that is, reduction of
the dissolved metal complex to Pd0 (colloidal) particles or
the oxidation of the Pd0 surface sites and leaching into the
liquid phase during the reaction. There are numerous exam-
ples on the extremely high activity of homogeneous com-
plexes[3] that may be formed in situ from the solid catalyst,[4]
and identification of 1–2 nm particles as being the real
active sites or only spectator species is also a demanding
task (for recent reviews see Refs. [4c] and [5]).
À
À
C C and C N bond-forming reactions are a versatile tool in
synthetic chemistry.[1] Such reactions are usually catalyzed
by homogeneous complexes derived from palladium precur-
[a] S. Reimann, Dr. T. Mallat, Prof. Dr. A. Baiker
Institute for Chemical and Bioengineering
Department of Chemistry and Applied Biosciences
ETH Zurich, Hçnggerberg, HCI, 8093 Zurich (Switzerland)
Fax : (+41)44-632-116
Every endeavor has been made to distinguish between ho-
mogeneous and heterogeneous catalysis. Several methods
have been developed and refined in the past years for this
purpose. Besides TEM analysis, filtration tests, and poison-
ing experiments, also more complex methods such as kinetic
studies, the three-phase test, and the use of a membrane re-
actor belong to the available techniques.[5a] The great variety
of the commonly used tests seems to be an indication that it
is difficult to get an unambiguous conclusion from such ex-
periments. Classical examples are the poisoning experiments
E-mail: baiker@chem.ethz.ch
[b] Prof. Dr. J.-D. Grunwaldt
Department of Chemical and Biochemical Engineering
Søltofts Plads, Technical University of Denmark
DK-2800 Kgs. Lyngby (Denmark)
[c] Prof. Dr. J.-D. Grunwaldt
Present address: Institute for Chemical Technology an;
Polymer Chemistry, Karlsruhe Institute of Technology (KIT)
Kaiserstr. 12, 76128 Karlsruhe (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000833.
9658
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668

FULL PAPER
with Hg and CS2. Pd easily forms alloys with Hg even at
room temperature,[6] and thus the observed reactivity and
leaching behavior is characteristic of the bimetallic catalyst
and not of the original catalyst. In addition, this unintended
transformation may even open new reaction pathways.[7]
Mercury may react with homogeneous Pd0 complexes,[8] and
incorporation of sulfur into the catalyst support can retard
the dissolution of Pd.[9]
ported Pd catalyst modified with various phosphines in the
same reaction.[11] In contrast to other studies,[12] we conclud-
ed that truly heterogeneous catalysis was occurring based on
separation experiments, TEM investigations, and kinetic
comparisons with homogeneous catalysts. The current study
allows us to shed more light into this discrepancy in the
nature of the active species.
Another concept to discriminate homogeneous and
hetero
G
Results
which homogeneous catalysts are reduced and transformed
to a solid by forming (nano)particles. It has been shown for
the high-temperature Heck coupling reaction that actually
none of the homogeneous catalyst precursors are stable
under the applied conditions.[4c] A thoroughly investigated
field is hydrogenation with soluble complexes at high tem-
perature, in which in many cases the nature of the active
species is still not clear.[5a] It seems to be a general trend
that harsh reaction conditions, such as a reducing (hydro-
gen) atmosphere and high temperature, favor the formation
of Pd nanoparticles.
As a result of the interplay of reaction components and
conditions, already small changes may lead to an unexpected
dislocation of the active species. This complexity makes it
difficult to determine the factors that facilitate or prevent
leaching when using ex situ techniques. Therefore, comple-
mentary in situ investigations are needed to track down the
influences of the single reaction components on the catalyti-
cally active species and link them to the overall leaching be-
havior.
Here we used direct monitoring of the liquid and solid
phases to analyze systematically what reaction components
and conditions favor heterogeneous or homogeneous condi-
tions in the Pd-catalyzed asymmetric allylic alkylation. The
typical homogeneous catalytic cycle is initiated by an oxida-
tive addition of the substrate to the Pd0 catalyst, leading to
the formation of a (h3-allyl) PdII complex. The consecutive
attack of a nucleophile on this complex leads to the forma-
tion of an unstable Pd0-olefin complex from which the final
product is readily released.[10] We chose a commonly used
reaction, the transformation of (E)-1,3-diphenylallyl acetate
Oxidation state of Pd during asymmetric allylic alkylation—
X-ray absorption near edge structure (XANES) analysis:
We followed the oxidation state of Pd by in situ XANES
measurements through the catalyst bed of the XAS cell
during allylic alkylation of rac-1 (Scheme 1). In the first ex-
periment, the original reaction mixture, consisting of rac-1,
the sodium salt of 2, and (R)-BINAP in THF, was added to
the XAS cell containing the catalyst. The reduction of the
Al2O3-supported Pd particles already started during heating
of the reaction mixture to 608C and a steady state was
reached at this temperature after 30–40 min. Measurements
through the supernatant solution did not show any presence
of leached Pd.
For comparison, the reduction in THF with hydrogen was
also followed. The analysis of the normalized absorption in-
tensity of the Pd K-edge whiteline of the spectra is shown in
Figure 1a. The spectra were additionally analyzed by a
linear combination fit and this method gave very similar re-
sults (Figure 1b). Note that owing to the high dispersion of
the catalyst even the reduction of less than a monolayer of
oxygen should be recordable because it is >5% (see also
theoretical XANES spectra Ref.[13]). Unexpectedly, reduc-
tion with hydrogen led only to a slightly more reduced state
than that with the reaction mixture. Another interesting ob-
servation is that THF alone could efficiently reduce Pd (Fig-
ure 1a–c). Reduction by THF was slower but only a qualita-
tive comparison was possible because of the probably less
efficient mass transport in the XAS cell compared with that
in the Parr reactor used for allylic alkylation. Nevertheless,
the final reduction state of the catalyst in THF, the best sol-
vent found for this reaction,[11] was similar to the level ach-
ieved in the presence of hydrogen.
À
(rac-1, Scheme 1), and the Pd/Al2O3 BINAP catalyst
system. The oxidation state of Pd in the solid phase and the
concentration of Pd in the liquid phase were analyzed by in
situ X-ray adsorption spectroscopy (XAS). This report com-
plements our previous studies that used a commercial sup-
The fact that the reduction of the catalyst proceeds faster
in the presence of the reaction mixture than in pure THF
implies that the solvent is not the only reducing agent.
Therefore, reduction of Pd by
every single component of the
reaction mixture was followed
in THF by XANES (Figure 2).
The linear combination fit, by
using PdO and hydrogen-re-
duced Pd/Al2O3 in THF as ref-
erences, and analysis of the nor-
malized absorption intensities
of the Pd K-edge whiteline
Scheme 1. Allylic substitution of (E)-1,3-diphenylallyl acetate (rac-1) with dimethyl malonate (2).
again gave very similar results.
Chem. Eur. J. 2010, 16, 9658 – 9668
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9659

A. Baiker et al.
Figure 1. Reduction of 5% Pd/Al2O3 at 608C and 10 bar N2, as monitored by a) the analysis of the normalized absorption intensity of the Pd K-edge
whiteline or b) linear-combination fit by using PdO and hydrogen-reduced Pd/Al2O3 in THF. The XANES spectra of the reduction of Pd/Al2O3 by THF
and cyclohexane are shown in (c) and (d), respectively. Conditions: 5% Pd/Al2O3 (84 mg), rac-1 (1.4 mmol), sodium salt of 2 (2.8 mmol), (R)-BINAP
~
^
*
(6.4 mmol) in THF (6 mL) under N2
( =reaction mixture, & =THF (neat), =cyclohexane (neat), =dichloromethane (neat), N=5% H2 in He, in
THF (neat)). The reaction temperature was reached after 10 min.
Later, only the normalized absorption intensity at
24.352 keV is shown in Figure 2 and other similar figures.
A solution of sodium dimethyl malonate in THF reduced
the catalyst completely within a few minutes. The rate and
the final state of reduction are comparable to those ach-
ieved by hydrogen in THF (see also Figure 1a) and slightly
exceed those of the reaction mixture, leading to the conclu-
sion that sodium dimethyl malonate is probably the main re-
ducing agent in the reaction mixture. This assumption is sup-
ported by our former catalytic study in which a 100%
excess compared to the stoichiometric amount of malonate
was necessary to achieve full conversion of the substrate.[11a]
Reduction of the catalyst by pure THF was faster than in
the presence of (R)-BINAP and addition of the substrate
slowed down the reduction even more. Retardation of the
reduction process by these two reaction components can ex-
plain why the reduction of the catalyst by the reaction mix-
ture is less efficient than with the malonate solution in THF.
To exclude misinterpretation owing to beam damage by
the X-ray beam itself, we repeated a typical experiment
shown above by heating the catalyst in THF to 608C while
the X-ray beam was switched off. Subsequent XANES
Figure 2. Reduction of 5% Pd/Al2O3 at 608C and 10 bar N2 by the single
components (in THF) of a typical reaction mixture, as monitored by
analysis of the normalized absorption intensity of the Pd K-edge white-
line. Conditions: 5% Pd/Al2O3 (84 mg), rac-1 (1.4 mmol), sodium salt of
~
2 (2.8 mmol), (R)-BINAP (6.4 mmol), in THF (6 mL) under N2 ( =reac-
&
tion mixture, "=rac-1, "=sodium salt of 2, =(R)-BINAP, & =THF
(neat)). The reaction temperature was reached after 10 min.
9660
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668

À
Asymmetric C C Bond-Formation Reaction with Pd
FULL PAPER
Table 1. Allylic substitution of rac-1 in cyclic ethers.[a]
measurements after 30 min showed that Pd/Al2O3 was re-
duced to the same extent as in the case when the X-ray
beam was on. Another important confirmation, that the in
situ spectroscopic measurements are not distorted, is the
similar catalytic performance observed here and in previous
studies. After 3 h reaction time at 608C, a conversion of 3–
4% with an ee of 57% was achieved in the in situ reactor
cell, which are typical values at this temperature[11a] and
demonstrate that the reaction conditions chosen for the in
situ XANES measurements were close to those in the cata-
lytic reactor.
Entry Solvent Modifier t [h] Conv. [%] Chemosel. [%] ee [%]
1[b]
2[b]
3
THF
THF
(R)-BINAP
–
3
6
3
6
73
64
33
30
95
60
90
58
60
0
55
0
dioxane (R)-BINAP
dioxane
4
–
[a] Conditions: 5% Pd/Al2O3 (42 mg), 3.2 mmol of modifier, rac-1
(1.4 mmol), sodium salt of 2 (2.8 mmol), solvent (18 mL), 1208C, N2
(20 bar). Conv.=conversion, chemosel.=chemoselectivity. [b] Taken
from Ref. [11a].
solvents, addition of BINAP enhanced the reaction rate and
the chemoselectivity as compared with the unmodified reac-
tion (in the absence of BINAP).
According to XANES analysis, pure dioxane reduces the
catalyst somewhat faster than pure THF, but after 40–
50 min the oxidation states of Pd are very similar in the two
cyclic ethers (Figure 4). The negative effect of BINAP on
Reduction of Pd by the solvent: Reduction of Pd by sol-
vents is rarely described in the literature, with the important
exception of alcohols,[13–14] the dehydrogenation of which on
Pt-group metals is widely used as the source of hydrogen in
transfer hydrogenation reactions. Therefore, we studied here
in more detail the effect of different solvents.
Replacement of THF by cyclohexane resulted in a slug-
gish and incomplete reduction of Pd/Al2O3 (Figure 1a, b,
and d). Toluene was even less reactive: no significant reduc-
tion occurred at 608C, and even at 1008C Pd remained
partly oxidized (Figure 3). In the popular solvent dichloro-
*
Figure 4. Reduction of 5% Pd/Al2O3 in dioxane at 608C with ( ) and
~
without ( ) (R)-BINAP monitored by analysis of the normalized absorp-
tion intensity of the Pd K-edge whiteline. Neat THF (&) is shown for
comparison. Conditions: 5% Pd/Al2O3 (84 mg) in solvent (6 mL) under
N2 (10 bar), and (R)-BINAP (6.4 mmol). The reaction temperature of
608C was reached after 10 min.
Figure 3. Effect of temperature on the reduction of 5% Pd/Al2O3 by tolu-
^
^
ene with ( ) and without ( ) rac-BINAP. Analysis of the normalized ab-
sorption intensity of the Pd K-edge whiteline. Conditions: 5% Pd/Al2O3
(84 mg) in toluene (6 mL) under N2 (10 bar), and rac-BINAP
(12.8 mmol).
the rate of Pd reduction in dioxane is similar to that ob-
served in THF (Figure 2). However, if the original reaction
mixture including the modifier was present, the reduction
behavior of the catalyst in these two solvents was different.
The probable explanation for the reduction of Pd in the
most suitable solvents, THF and dioxane, is the consumption
of surface oxygen by the oxidation at the a-carbon atom of
the ether.[15] Palladium complexes[16] as well as supported
Pd,[17] Rh,[17] and Pt[18] are proven catalysts for the oxidation
of ethers to the corresponding lactones by using molecular
oxygen as the oxidant. In the specific case of THF, g-butyro-
lactone (main product), tetrahydrofuran-2-ol, and 4-
methane the rate of reduction was between those observed
in THF and cyclohexane (Figure 1). In general, increasing
the temperature to 100–1208C enhanced the rate and extent
of the reduction of Pd in all these solvents, as expected; an
illustration is shown for toluene in Figure 3.
À
In the Pd/Al2O3 catalyzed allylic alkylation of rac-1 the
best yield of 3 was achieved in THF.[11] The suitability of this
solvent is partly attributed to the good solubility of the
sodium salt of 2. Hence, we extended the study to dioxane,
which is chemically similar to THF. A comparative study re-
vealed that replacing THF by dioxane diminished the reac-
tion rate (conversion) to about one half but the chemo- and
enantioselectivities were barely affected (Table 1). In both
hydroxyACTHNUGTRNEUNGbutanal are formed (Scheme 2).
When a (supported) Pd catalyst is stored under air, the
outer surface layers of the metal particles are transformed
to metal oxide and reduction of this oxide layer is slow. This
Chem. Eur. J. 2010, 16, 9658 – 9668
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9661

A. Baiker et al.
Scheme 2. Products formed by catalytic oxidation of THF with molecular
oxygen.
slow reduction is seen in Figures 1–3. Before the use in allyl-
ic alkylation, the catalyst was always reduced in a fixed-bed
reactor in flowing hydrogen at 2008C, which increased the
reaction rate and the chemoselectivity, and also the reprodu-
cibility of the results.[11] Note that the pretreatment did not
lead to a change of the mean Pd particle size (typically
3 nm).[11a] During transferring the catalyst to the slurry reac-
tor, the surface of the Pd particles were reoxidized, but this
freshly formed oxide layer was thinner and more easy to
reduce, which explains the improved catalytic performance.
This explanation is supported by XPS measurements (see
the Supporting Information). Reduction of the surface by
THF regenerates the active metal surface, and the solvent
being present in large excess allows maintaining the reduced
state during the slow allylic alkylation reaction, despite of
the possible oxygen diffusion into the reactor.
Scheme 3. Possible mechanism for the reduction of surface PdII species to
Pd0 by dimethyl malonate anions.
neous complexes, but an example of a heterogeneously cata-
lyzed reaction is also described.[23]
The negative influence of the substrate on the reduction
of Pd may be attributed to a partial coverage of the surface
by the strongly adsorbing electron-rich molecule that slows
down the reduction by the solvent. Alternatively, some com-
peting reactions such as oxidative additions, analogous to
those described in homogeneous catalysis, may play a
role.[10b,24]
The slow reduction of oxidized Pd by cyclohexane
(Figure 1) and toluene (Figure 3) was attributed to dehydro-
genation of the solvents and reduction of the Pd surface by
the hydrogen produced in situ. The Pd-catalyzed dehydro-
genation of cyclohexane and toluene was accompanied by
the formation of benzene[19] and carbonaceous species on
the surface, respectively.[20] A control experiment with cyclo-
hexene and Pd/Al2O3 in THF led, as expected, to dehydro-
genation as well as hydrogenation of cyclohexene to ben-
zene and cyclohexane, respectively.[21] An analogous mecha-
nism for the cyclic ethers might also seem a possible reac-
tion pathway. However, significant dehydrogenation of the
ethers and the formation of surface hydrogen is unlikely to
occur because during allylic alkylation of rac-1 in ether sol-
vents no hydrogenation of the double bond in the substrate
was observed, even at 1208C.[11]
Interaction of the modifier with Pd: We have shown by UV/
Vis spectroscopy[11a] that BINAP adsorbs strongly on the
surface of alumina-supported Pd particles, which is a key
characteristic of a chiral modifier being present in trace
amounts.[25] The XANES analysis revealed that BINAP
slows down the reduction of Pd in THF (Figure 2) and diox-
ane (Figure 4), the two most suitable solvents for our test re-
action. In toluene the influence is negligible (Figure 3) but
this solvent proved to be the least effective in reducing the
oxidized Pd surface. The retardation by BINAP is, however,
less pronounced when the whole reaction mixture is present,
as monitored by the Pd K-edge XANES during allylic alky-
lation in the presence and absence of the modifier in THF
(Figure 5a) and dioxane (Figure 5b). Note that in these
cases, the analysis of the normalized absorption intensities
of the Pd K-edge whiteline gave a more distinct picture than
the linear-combination fit (not shown). In both cyclic ether
solvents, the rate of reduction of Pd and the steady state
during the reaction were similar with and without BINAP at
608C (Figure 5a and b, respectively) and also at 1008C (not
shown). Interestingly, in THF Pd was even more reduced in
the presence of a modifier, although the differences were
small. A feasible explanation is that the solvent adsorbs rel-
atively weakly on Pd and the presence of BINAP disturbs
the Pd–solvent interaction. In contrast, the electron-rich di-
methyl malonate anion is assumed to be adsorbed strongly
on the metal surface and BINAP does not hinder this inter-
action significantly. As discussed previously, the nucleophile
is the strongest reducing agent in the reaction mixture;
hence the effect of BINAP in the reaction mixture is minor.
Nevertheless, BINAP may play a role in the surface
(redox) processes by 1) modifying the electronic state of Pd,
2) displacement of surface-blocking species, and 3) stabiliza-
tion of the reduced state of the Pd surface. Based on the
Interaction of the nucleophile and the substrate with the Pd
surface: It is shown in Figure 2 that reduction of Pd in THF
occurs at similar rates with sodium dimethyl malonate or hy-
drogen. We assume that reduction of the oxidized surface
Pd sites by sodium dimethyl malonate occurs analogously to
the arylation of malonate anions via reductive elimination
from aryl Pd complexes.[22] A nucleophilic attack on the mal-
onate anion coordinated to the Pd surface leads to the re-
duction of PdII to Pd0 and the release of the corresponding
addition product of malonate (Scheme 3). Potential nucleo-
philic species are another malonate anion and alcohols
formed during reduction of the catalyst by the solvent
(Scheme 2). During allylic alkylation the acetate ion re-
leased by the substrate (Scheme 1) may as well react with
the adsorbed malonate to form the corresponding acetoxy
malonate. The reaction is usually carried out with homoge-
9662
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668

À
Asymmetric C C Bond-Formation Reaction with Pd
FULL PAPER
Table 2. Oxidation of rac-BINAP under nitrogen in the presence and ab-
sence of 5% Pd/Al2O3.[a]
Catalyst
Prereduction[b]
T [8C]
BINAP/BINAPO/BINAPO2
molar ratio
5% Pd/Al2O3
5% Pd/Al2O3
5% Pd/Al2O3
no catalyst
+
+
À
À
60
100:12:20
100:16:46
100:13:18
100:22:0
120
120
120
[a] Conditions: 84 mg of catalyst, rac-BINAP (16 mmol) in THF (10 mL),
N2 atmosphere, 20 bar, 3 h. [b] The catalyst was prereduced in hydrogen
at 2008C in a continuous flow reactor prior to use (+) or used as received
(À).
Scheme 4. Oxidation of BINAP leading to BINAP monoxide (BINAPO)
and BINAP dioxide (BINAPO2).
Next, the importance of the oxidation of BINAP during
allylic alkylation of rac-1 was investigated. The reactions,
carried out with different mixtures of (R)-BINAP and (R)-
BINAPO2, revealed that the influence of the oxidized modi-
fier is small as long as BINAP is present in the reaction mix-
ture (Table 3). When, however, only BINAPO2 was applied
Figure 5. Reduction of 5% Pd/Al2O3 by the reaction mixture in a) THF
~
~
or b) dioxane with ( ) and without ( ) (S)-BINAP at 608C; analysis of
the normalized absorption intensity of the Pd K-edge whiteline. Condi-
Table 3. Allylic substitution of rac-1 by using (R)-BINAP/(R)-BINAPO2
tions: 5% Pd/Al2O3 (84 mg), rac-1 (1.4 mmol), sodium salt of
2
mixtures as modifiers.[a]
(2.8 mmol), (S)-BINAP (6.4 mmol), in THF (6 mL) under N2 (10 bar).
The reaction temperature of 608C was reached after 15 min.
Entry (R)-BINAP/(R)-
BINAPO2
t [h] Conv. [%] Chemosel. [%] ee [%]
molar ratio
1
2
3
4
1:0
1:1
1:10
1:20
0:1
0:0
3
3
3
3
6
6
61
62
57
59
27
64
96
95
96
96
36
60
59
57
56
55
0
pertinent literature, BINAP is probably not inert on the Pd
surface. The facile transformation of the PdII/PIII system to
Pd0/PV is well-known and plays an important role, for exam-
ple in Heck coupling reactions.[26] The phosphine is thereby
transformed into the corresponding phosphine oxide and re-
sidual water or acetate is speculated to be the source of
oxygen.[27] In addition, phosphines are easily oxidized with
O2.[28]
5
6[b]
0
[a] Conditions: 5% Pd/Al2O3 (42 mg), modifier (3.2 mmol), rac-1
(1.4 mmol), sodium salt of 2 (2.8 mmol), solvent (18 mL), 1208C, 20 bar
N2. Conv.=conversion, chemosel. =chemoselectivity. [b] Taken from
Ref. [11a].
To confirm the possibility that BINAP is a potential re-
ducing agent for the oxidized Pd surface, transformation of
the modifier under nitrogen atmosphere was followed by
quantitative 31P NMR spectroscopy (Table 2). BINAP mon-
oxide (BINAPO) and BINAP dioxide (BINAPO2) were
identified as products (Scheme 4). In a control experiment
the catalyst was activated in hydrogen at elevated tempera-
ture prior to use. Although the catalyst was reoxidized upon
exposure to ambient air during transfer to the reactor, the
formation of BINAPO2 was enhanced compared to the non-
activated catalyst, and without catalyst this product did not
form at all. The results show that the formation of
BINAPO2 requires Pd as a catalyst.
as a modifier, no enantioselection was observed and the re-
action rate and chemoselectivity decreased. For comparison,
attempts to use BINAPO2 as a ligand for PdACTHNURGTNE(GNU OAc)2 failed,
and only traces of the racemic product were formed.
We can conclude from the results in Tables 2 and 3 that
BINAP may provide some protection of the active, metallic
Pd catalyst during allylic alkylation by consuming the sur-
face oxygen. This process is, however, detrimental to the
enantioselection as the co-product BINAPO2 is ineffective
as a chiral modifier. Oxidation of the P atoms of BINAP
probably weakens its adsorption on the Pd surface and the
Chem. Eur. J. 2010, 16, 9658 – 9668
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9663

A. Baiker et al.
molecule is replaced by an intact modifier from the solution
phase. This also explains the negligible effect of BINAPO2
on the alkylation reaction as long as BINAP is present. In
the light of these observations it seems essential to keep the
metal surface in a reduced state by the solvent and the nu-
cleophile that are present in large excess and thus protect
BINAP from oxidation and avoid the loss of enantioselec-
tion.
In control experiments, the catalyst was removed by cen-
trifugation under N2 after the reactions in Table 4, entries 1
and 3. After the addition of fresh sodium malonate, the fil-
trate was allowed to react further at room temperature.
After 18 h reaction time no further conversion was observed
in the case of pure THF as the solvent (reaction according
to Table 4, entry 1). However, when chloroform was present
during the catalytic reaction, a small conversion of addition-
al 4% was observed (reaction according to Table 4, entry 3).
This is an indication to the presence of soluble active species
in the filtrate.
In the allylic alkylation reaction two equivalents of nucle-
ophile 2 were used. In the experiment in Table 4, entry 2,
the amount of free chloride (0.23 mmol), and therefore the
amount of HCl that could deactivate the nucleophile by pro-
tonation, was much lower than the amount of nucleophile
(2.8 mmol). Therefore, deactivation of the nucleophile
through protonation with in situ formed HCl may have only
a small effect on the reaction (Table 4). Additionally, it may
be possible that the reaction is slower because the nucleo-
phile is consumed by the reduction of leached PdII ions.
Other reasons for the deactivation could be the degradation
of chloroform on the Pd surface similar to that of dichloro-
methane, leading to the formation of carbonaceous species
and site blocking.[35] Considering the large number of sur-
face sites necessary to accommodate the bulky modifier
BINAP, this site blocking can easily explain the dramatic
loss of catalytic performance. Already the adsorption of
chloroform on the Pd surface may diminish the activity by
blocking the surface and/or changing the electronic proper-
ties of the metal.[36]
Next, the role of chloroform in leaching of palladium and
the formation of a homogeneous active species was investi-
gated. The prereduced Pd/Al2O3 catalyst was shaken for
72 h in the presence of (R)-BINAP and chloroform at room
temperature under an inert atmosphere. After removal of
the solid catalyst and addition of the reaction components
rac-1 and 2 to the filtrate, the alkylation reaction proceeded
smoothly, even at room temperature (Table 5, entry 1). The
high reaction rate and chemoselectivity but moderate ee are
typical for the corresponding homogeneous catalyst.[11a,37]
When the catalyst was not prereduced prior to the leaching
period, the conversion in the subsequent alkylation reaction
decreased (Table 5, entry 2). As mentioned previously, after
prereduction in hydrogen at 2008C, the catalyst was reoxi-
dized during transferring into the slurry reactor, but prere-
duction diminishes the amount of surface oxide developed
during storage in open air and activates the noble metal par-
ticles for restructuring.[38] Upon increasing the amount of
(R)-BINAP, the conversion increased only in the case of the
prereduced catalyst (Table 5, entries 1 and 4, 2 and 5). In-
creasing the amount of chloroform led to higher conversion
(Table 5, entries 1 and 3) even when the catalyst was not
prereduced (Table 5, entries 5 and 7). In a control experi-
ment without chloroform, no conversion was detectable
(Table 5, entry 8).
Palladium leaching in chlorinated solvents: Chlorinated sol-
vents, such as dichloromethane and chloroform, are com-
monly used in allylic alkylation reactions.[1a,c] These solvents
are useful in the chemical transformation but represent a
potential source of corrosion of the metal catalyst. Support-
ed Pd is a highly active dehalogenation catalyst[29] that can
be applied also without hydrogen.[30] Interaction of the
chlorinated solvent with the metal surface may lead to the
formation of ClÀ,[31] which accelerates the metal corrosion
by shifting the redox potentials[32] and stabilizing the leached
Pd species as a chloro complex. A commonly neglected
point is that not only the metal but also the support can con-
tribute to the degradation of chlorinated hydrocarbons. It
has been found that methyl chloride interact with the sur-
face hydroxyl groups of Al2O3 to form adsorbed methoxy
species already below 1008C and the formation of HCl ac-
celerates at above approximately 1258C .[33] In addition, the
generation of HCl from chloroform can occur via an a,a-
elimination and the formation of dichlorocarbene in the
presence of a base[34] such as sodium dimethyl malonate.
To clarify the role of chlorinated solvents, we repeated
the experiment in Table 1 (entry 1) by replacing a small frac-
tion of THF with chloroform. The remarkable loss in reac-
tion rate (conversion) and the lower selectivities, shown in
Table 4, demonstrate that the chlorinated solvent is detri-
Table 4. The influence of CHCl3 on the allylic alkylation of 1.[a]
Entry Catalyst
CHCl3 ClÀ
Conv. Chemosel. ee
[mL]
[mmol] [%]
[%]
[%]
[b]
[c]
1
2
3
4
5% Pd/Al2O3
0
0.11
1
0
73
22
5
95
86
74
0
60
55
35
0
5% Pd/Al2O3
5% Pd/Al2O3
–
n.d.[d]
0.23
0.11
1
2
[a] Conditions: 5% Pd/Al2O3 (42 mg, prereduced), rac-1 (1.4 mmol),
sodium salt of 2 (2.8 mmol), (R)-BINAP (3.2 mmol), in THF (17 mL), 3 h,
1208C, under N2 (20 bar). Conv.=conversion, chemosel.=chemoselectiv-
ity. [b] 18 mL THF, taken from Ref. [11a]; [c] 2 h reaction time; [d] Not
determined.
mental to the performance of the heterogeneous catalyst.
The “poisoning” effect of chloroform increased with its
amount. The analysis of pure chloroform by argentometric
titration showed no presence of free chloride anions. In the
reaction mixture, however, the dehalogenation of chloro-
form proceeded already in the absence of the catalyst, prob-
ably by the action of the nucleophile 2 (Table 4, entry 4),
and free ClÀ ions were detected. With the catalyst present,
the amount of free ClÀ ions doubled.
9664
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668

À
Asymmetric C C Bond-Formation Reaction with Pd
FULL PAPER
Table 5. Leaching experiments with chloroform at room temperature.[a]
system and there is continuous oxygen diffusion into the re-
actor during reaction. Obviously, working under an inert at-
mosphere at elevated pressure minimizes the oxygen diffu-
sion rate into the reaction mixture and thus favors the main-
tenance of the active, reduced state of the Pd surface.
Another important factor is the presence of a “reducing
agent” in the reaction mixture. This component can reduce
the oxidized (supported) metal catalyst and restore the
active M0 sites at the early stage of the reaction and avoid
the reoxidation of the metal during reaction by, for example,
oxygen diffusion into the reactor. It is also important that
the amount of the reducing agent is high enough to keep Pd
in a reduced state throughout the reaction.
Entry Pre
G
CHCl3/
t
Conv. Chemosel. ee
[h] [%]
[%]
[%]
A
volume
ratio
1
52
20 100
20
20 73
100
100
100
100
100
100
100
100
100
100
100
100
100
98
21
27
23
26
30
30
25
26
17
22
25
27
24
28
–
1
2
3
4
5
6
7
8
+
À
+
+
À
+
À
+
6.4
6.4
6.4
16
1:4
1:4
1:0
1:4
1:4
1:0
1:0
0:1
1
1
100
20 100
1
91
20 100
15
20 53
93
20 100
60
20 93
1
16
1
The combined catalytic, in situ XANES study confirmed
our previous result that the enantioselective allylic alkyla-
16
2
16
À
tion of rac-1 with the Pd/Al2O3 BINAP catalyst system is a
truly heterogeneous reaction. We could show unambiguous-
ly that the original reaction mixture is able to reduce the Pd
surface and maintain this reduced state during reaction (Fig-
ure 1a), and no leached Pd was detected in the reaction
mixture by in situ XANES measurements. By using various
combinations of the reaction components, it was indicated
that sodium dimethyl malonate (2) is probably the main re-
ducing agent (Figure 2). The necessity of using more than a
stoichiometric amount of 2 to achieve high conversion had
already been recognized earlier but could not be rational-
ized at that time.[11a]
1
24
0
0
–
–
6.4
–
[a] Preleaching: 5% Pd/Al2O3 (42 mg) was shaken for 72 h in the solvent
mixture (THF/CHCl3, 5 mL ) under N2 at 1 bar. Then rac-1 (0.7 mmol),
sodium salt of 2 (1.4 mmol), and THF (9 mL) were added to the filtrate
for the allylic alkylation at RT. Conv.=conversion, chemosel.=chemose-
lectivity. [b] The catalyst was prereduced in hydrogen at 2008C in a con-
tinuous flow reactor prior to use (+) or used as received (À).
All these observations indicate that addition of chloro-
form favors homogeneous catalysis and is detrimental to the
heterogeneous route.
The proper choice of solvent is critical. The two best sol-
vents of the reaction, THF and dioxane, are strong reducing
agents of oxidized Pd (Figures 1a and 4). They reduced the
catalyst almost as fast as hydrogen, although the difference
in favor of hydrogen would probably have been significantly
higher under better mass-transfer conditions than those al-
lowed by the tiny XANES reactor cell (Figure 1a). In con-
trast, halogenated solvents such as chloroform promote
metal dissolution and homogeneous catalysis. Taking the ex-
ample of chloroform, the first step is probably the decompo-
sition and formation of chloride ions, catalyzed by Pd and
also by the support Al2O3 (Table 4). Halide ions are excel-
lent catalysts of metal corrosion in inorganic systems (lead-
ing to rate acceleration and stabilization of the Mn+ spe-
cies), and the same is assumed for the complex organic–inor-
ganic reaction mixture during allylic alkylation in the pres-
ence of chloroform. Note also that the corrosion effect of
halides increases in the order ClÀ <BrÀ <IÀ.[32,39] The effect
of halogenated substrates on the heterogeneity of Pd can
probably be explained similarly, and this speculation is sup-
ported by the numerous papers reporting Pd leaching and
homogeneous catalysis in the transformation of aryl bro-
mides and iodides.[4c]
Discussion
It is a demanding task to obtain convincing evidence for ho-
mogeneous or heterogeneous catalysis in the Pd-catalyzed
À
À
C C and C N bond-forming reactions, as illustrated by the
lively debate on the nature of active species in the past dec-
ade.[4c,5b,d] Instead of using the common (and frequently un-
reliable) heterogeneity tests, we considered the classical
metal corrosion in inorganic systems as an analogy to metal
leaching during allylic alkylation with a supported Pd cata-
lyst. Based on this analogy, we assumed that by choosing the
appropriate conditions and reaction components it is possi-
ble to favor either homogeneous (“molecular“) or heteroge-
neous (“surface”) catalysis. Here we tested this approach by
in situ monitoring of the oxidation state in the Pd-catalyzed
asymmetric allylic alkylation of rac-1 with Pd/Al2O3 by
using BINAP as the chiral modifier (Scheme 1).
In analogy to metal corrosion we assumed that the key re-
quirements for metal dissolution are the oxidation of the
surface metal atoms and the presence of an appropriate
anion or ligand that stabilizes the oxidized state in solution.
An important point in our considerations is that in the usual
laboratory reactor applied for organic synthesis, the condi-
tions cannot be considered as oxygen free. At the beginning,
the metal surface is in an oxidized state owing to storage of
the catalyst in ambient air. In addition, the reaction usually
starts with a certain level of oxygen concentration in the
The role of the chiral modifier BINAP is more complex.
It is present only in minor amounts. Because of the strong
adsorption on Pd, its effect on the oxidation state is impor-
tant. Coverage of the surface sites by the bulky modifier
slows down the redox processes and also modifies the elec-
tronic properties of the Pd surface (Figure 1a). On the other
hand, it is easily oxidized to BINAPO and BINAPO2 by Pd
Chem. Eur. J. 2010, 16, 9658 – 9668
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9665

A. Baiker et al.
ing to the data sheet provided by the manufacturer. The mean particle
size was determined to be 3.3 nm and the corresponding metal dispersion
was calculated to be 0.36, as described elsewhere.[11a]
in the presence of surface oxygen even in inert atmosphere
(Table 2). Although this transformation contributes to the
maintenance of the reduced state of Pd, the oxidized modifi-
er probably adsorbs weaker and loses its role as the source
of chiral information in the transformation of rac-1
(Table 3). This process is assumed to be the reason for the
necessity of applying relatively high BINAP/surface Pd
atoms ratio (typically 0.2–0.4) to obtain the optimum enan-
tioselectivity in the slow transformation of rac-1.[11a] It was
shown that under these conditions only a small fraction of
BINAP was actually adsorbed on the catalyst and the rest
was present in the solution phase functioning as a “reser-
voir” for the replacement of the oxidized modifier in the
course of the reaction.
Finally, the reaction temperature should be considered. It
has already been described in the literature that high reac-
tion temperatures usually lead to reduction of the soluble
active complex and to the formation of Pd nanoparticles.[5a]
This parameter is particularly important in our case because
the potential homogeneous catalyst is not stable at elevated
temperature.
(E)-1,3-diphenylallyl acetate (rac-1) was synthesized from (E)-1,3-diphe-
nylprop-2-en-1-ol (ꢀ 98%, Fluka) and acetic anhydride (>98%, Acros)
in dichloromethane and was purified by column chromatography (cyclo-
hexane/ethyl acetate 9:1).[40] The structure of rac-1 was confirmed by 1H
and 13C NMR spectroscopy and its purity (>99%) was determined by
GC analysis. (R)-BINAPO2 was synthesized by oxidizing (R)-BINAP
1
with H2O2 and identified by H and 31P NMR spectroscopy.[41]
Catalytic experiments: The allylic alkylation reactions were carried out in
a magnetically stirred stainless steel autoclave (Parr) equipped with a
glass liner. For a typical experiment, the glass liner was filled with solvent
(12 mL), 5 wt% Pd/Al2O3 (42 mg), (R)-BINAP (0.002 g, 3.2 mmol), rac-1
(0.353 g 1.4 mmol), and a solution of the sodium salt of 2 in the corre-
sponding solvent (0.47m, 6 mL). The malonate salt solution was always
freshly prepared from equimolar amounts of NaH and 2 and filtered
carefully before use. The reactor was pressurized with N2 (20 bar) and
heated to the desired reaction temperature (typically 1208C) by immerg-
ing it into an oil bath. The catalyst was prereduced prior to its use in a
fixed bed reactor in flowing H2 at 2008C for 60 min. After cooling to
room temperature in H2 (30 min), the catalyst was flushed with N2 for
10 min and transferred to the autoclave.
The leaching experiments at room temperature were carried out in a
glass vial under a N2 atmosphere. The proper amount of (R)-BINAP was
dissolved in a mixture of the solvent (5 mL) and catalyst (42 mg), which
was then sealed and shaken for 72 h. The solid catalyst was then removed
by centrifugation under N2, and the supernatant solution was transferred
into a N2 flushed 50 mL Schlenk flask. Finally, rac-1 (0.177 g, 0.7 mmol)
and a freshly prepared and filtered solution of the sodium salt of 2 in
THF (0.16m, 9 mL) were added to the reaction mixture. At the end of
the reaction, the reaction mixture was filtered and the filtrate was ana-
lyzed by GC and HPLC.
Conclusion
There are only very few reports on heterogeneous asymmet-
ric allylic alkylation reactions. Jansat et al. reported for the
first time the high activity and enantioselectivity of Pd nano-
particles in the addition of dimethyl malonate to rac-1
(Scheme 1).[12b] The nanoparticles were stabilized by a chiral
xylofuranoside diphosphite ligand. Some years later the au-
thors admitted that in fact the active species was a soluble
Pd complex formed during the reaction.[12a] Our present in-
vestigation supports this conclusion because chlorinated sol-
vents (in that case CH2Cl2) favor dissolution of metallic Pd,
and the high activity of the catalyst system under ambient
conditions is also a strong indication to homogeneous cataly-
sis.
The typical leaching experiments at 1208C were carried out according to
the procedure described previously for the standard catalytic experi-
ments. At the end of the reaction, the reactor was cooled down in an ice
bath and the solid catalyst was removed by centrifugation under N2. The
supernatant solution was transferred into a N2 flushed Schlenk flask, and
a freshly prepared and filtered solution of the sodium salt of 2 (0.47m,
6 mL) was added to the reaction mixture. After the desired reaction time
the reaction mixture was filtered and the filtrate was analyzed by GC
and HPLC.
The oxidation of rac-BINAP by 5 wt% Pd/Al2O3 was carried out in a
50 mL flask equipped with a cooler (608C) or in the reactor described
above (1208C). In both cases, 5 wt% Pd/Al2O3 (84 mg) and rac-BINAP
(0.010 g, 16 mmol) in THF (10 mL) under a N2 atmosphere (1 bar at
608C, 20 bar at 1208C) were used. The reaction time was set to 3 h.
Analysis: The 1H, 13C and 31P NMR spectra were recorded on a Bruker
Avance 200 or Avance 500 spectrometer and the signals were referenced
to TMS. The quantitative 31P NMR spectra were recorded on a Bruker
Avance 700 spectrometer using inverse gated decoupling.
In contrast, all observations of the present study corrobo-
À
À
rate the heterogeneity of the Pd/Al2O3 BINAP sodium di-
À
methyl malonate THF (dioxane) system. We hope that the
presented systematic approach of investigating directly the
effect of reaction conditions and reaction components on
the oxidation state and stability of the active metal species
will help discriminating between heterogeneous and homo-
geneous catalysis also in other cases.
The conversion of rac-1 was determined by using a Thermo Finningan
gas chromatograph equipped with an HP-5 capillary column (30 mꢁ
0.32 mmꢁ0.25 mm) and diethyl phthalate (>99%, Merck) as the internal
standard. The following conditions were used: 1.1 mLminÀ1 carrier gas
flow, 1508C initial temperature (1 min), 3008C final temperature (5 min),
258CminÀ1 heating rate. The retention times for 1 and 3 were 4.7 min
and 6.1 min, respectively. Structural identification of 3 was made by GC–
MS by using an HP-6890 gas chromatograph coupled with an HP-5973
mass spectrometer, and by 1H and 13C NMR measurements. Product 3
was isolated according to a known procedure.[42] The side products were
analyzed by GC–MS.
Experimental Section
Materials: Tetrahydrofuran (THF, 99.99%, Acros), dioxane (99.5%,
Acros) dichloromethane (99.5%, J. T. Baker), and chloroform (99%. J. T.
Baker) were dried and stored over activated molecular sieves. (R)-(+)-
2,2’-bis(diphenylphosphino)-1,1’-binaphtalene ((R)-BINAP; >99.5%,
Fluka), rac-BINAP (ꢀ98%, Fluka), dimetyl malonate (>99%, Aldrich),
and NaH (60% dispersion in mineral oil, Fluka) were used without fur-
ther purification. The 5 wt% Pd/Al2O3 (N8 40692) catalyst was purchased
from Engelhard. The BET surface of the catalyst was 130 m2 gÀ1 accord-
The ee of 3 ((R) enantiomer: 12.3 min, (S) enantiomer: 13.2 min) was de-
termined by HPLC (Merck LaChrom). The analysis was carried out on a
Chiracel OD chiral column (240 mmꢁ4.6 mm i.d., 10 mm particle size) at
258C with a liquid flow rate of 0.9 mLminÀ1 and a 99:1 n-hexane/isopro-
panol mixture as the eluent. The assignment of the peaks was made by
9666
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668

À
Asymmetric C C Bond-Formation Reaction with Pd
FULL PAPER
comparison with literature data.[43] The absolute configuration of 3 was
verified by comparing the sign of the optical rotation with literature
data.[44] Optical rotations were measured on a Perkin–Elmer 241 polarim-
eter using a 1 dm cell at room temperature in chloroform (l=589 nm,
Na d-line).
Before the argentometric determination of free ClÀ, the reaction mixture
was washed twice with water (15 mL). The combined water phases were
washed with cyclohexane (15 mL). The water phase was then diluted to a
volume of 60 mL. For a single measurement, 10 mL of the sample solu-
tion was again diluted to a volume of 50 mL. After adding a few drops of
K2CrO4/K2CrO7 indicator the solution was titrated with a 0.01m aqueous
solution of AgNO3. For each sample the measurement was repeated six
times.
[6] T. Mallat, Z. Bodnꢂr, S. Szabꢃ, J. Petrꢃ, Appl. Catal. 1991, 69, 85.
[7] G. M. Whitesides, M. Hackett, R. L. Brainard, J. Lavalleye, A. F. So-
winski, A. N. Izumi, S. S. Moore, D. W. Brown, E. M. Staudt, Orga-
nometallics 1985, 4, 1819.
[8] R. van Asselt, C. J. Elsevier, J. Mol. Catal. 1991, 65, L13.
[9] I. Takamiya, S. Tsukamoto, M. Shimoda, N. Miyashita, M. Arisawa,
Y. Arakawa, A. Nishida, Chem. Lett. 2004, 33, 1208.
[10] a) B. M. Trost, D. L. VanVranken, Chem. Rev. 1996, 96, 395; b) A.
Pfaltz, M. Lautens in Comprehensive Asymmetric Catalysis, Vol. 2,
1st ed. (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer,
Berlin, 1999, p. 834.
[11] a) S. Reimann, T. Mallat, A. Baiker, J. Catal. 2007, 252, 30; b) S. Re-
imann, T. Mallat, A. Baiker, J. Catal. 2008, 254, 79.
[12] a) M. Diꢄguez, O. Pꢅmies, Y. Mata, E. Teuma, M. Gꢃmez, F. Ribau-
do, P. W. N. M. van Leeuwen, Adv. Synth. Catal. 2008, 350, 2583;
b) S. Jansat, M. Gomez, K. Philippot, G. Muller, E. Guiu, C. Claver,
S. Castillon, B. Chaudret, J. Am. Chem. Soc. 2004, 126, 1592.
[13] J.-D. Grunwaldt, M. Caravati, A. Baiker, J. Phys. Chem. B 2006,
110, 9916.
[14] a) M. Caravati, J.-D. Grunwaldt, A. Baiker, Catal. Today 2003, 77–
88, 1; b) J.-D. Grunwaldt, M. Caravati, A. Baiker, J. Phys. Chem. B
2006, 110, 25586.
[15] C. R. A. Godfrey in Comprehensive Organic Synthesis Vol. 7 (Eds.:
B. M. Trost, I. Fleming), Pergamon Press, Oxford, 1991, p. 235.
[16] M. Sommovigo, H. Alper, J. Mol. Catal. 1994, 88, 151.
[17] M. Shi, J. Chem. Res. 1998, 592.
[18] a) A. Datta, M. Agarwal, S. Dasgupta, Catal. Surv. Asia 2004, 8, 171;
b) T.-J. Wang, Z.-H. Ma, M.-Y. Huang, Y.-Y. Jiang, Polym. Adv.
Technol. 1996, 7, 88.
[19] R. Raval, M. A. Chesters, Surf. Sci. 1989, 219, L505.
[20] C. Mondelli, J.-D. Grunwaldt, D. Ferri, A. Baiker, Phys. Chem.
Chem. Phys. 2010, 12, 5307.
[21] M. A. Aramenꢆa, V. Borꢂu, I. M. Garcꢆa, C. Jimꢄnez, A. Marinas,
J. M. Marinas, F. J. Urbano, J. Mol. Catal. A 2000, 151, 261.
[22] a) N. A. Beare, J. F. Hartwig, J. Org. Chem. 2002, 67, 541; b) J. P.
Wolkowski, J. F. Hartwig, Angew. Chem. 2002, 114, 4465; Angew.
Chem. Int. Ed. 2002, 41, 4289.
XANES measurements: In situ XANES experiments were carried out in
a specially designed XAS batch reactor cell possessing two paths for X-
ray adsorption measurements at the bottom and in the middle of the re-
actor to measure the solid/liquid interface at the catalyst bed and the
liquid phase, respectively, at temperatures up to 2208C and pressures up
to 250 bar. The reactor was equipped with inlet and outlet lines to com-
pressed liquids and gases, and magnetic stirring of the solution facilitated
mass transport. A polyetheretherketone (PEEK) inset in the 10 mL stain-
less steel reactor allowed corrosive materials to be used. The details of
the cell are described elsewhere.[45]
For a typical experiment, 5 wt% Pd/Al2O3 (84 mg; used as received,
without prereduction) was placed in the reactor followed by rac-1
(0.353 g, 1.4 mmol) and a solution of (R)-BINAP (0.004 g, 6.4 mmol) in
solvent (1 mL). Then a freshly prepared and filtered solution of the
sodium salt of 2 in the corresponding solvent (0.56m, 5 mL) was added.
After setting a N2 pressure of typically 10 bar, the cell was heated to the
desired temperature. For the measurements of a single reaction compo-
nent, the amounts described above in 6 mL of solvent were used. The
raw data were energy calibrated at the Pd K-edge by using the Pd K-
edge energy of a Pd foil, background corrected and normalized by using
the WINXAS 3.1 software;[46] in this way a difference in variation with
less than 5% could be analyzed. The linear-combination fit was also car-
ried out by using the WINXAS 3.1 software. As references, the XANES
spectra of PdO and the catalyst reduced by H2 in THF at 608C were
used.
[23] L. Djakovitch, K. Kçhler, J. Organomet. Chem. 2000, 606, 101.
[24] I. Pryjomska-Ray, A. Gniewek, A. M. Trzeciak, J. J. Ziꢃlkowski, W.
Tylus, Top. Catal. 2006, 40, 173.
[25] T. Mallat, E. Orglmeister, A. Baiker, Chem. Rev. 2007, 107, 4863.
[26] J. Tsuji, Palladium Reagents and Catalysis—New Perspectives for the
21st Century, Wiley, New York, 2004.
[27] a) C. Amatore, E. Carre, A. Jutand, M. A. Mbarki, Organometallics
1995, 14, 1818; b) C. Amatore, A. Jutand, M. A. Mbarki, Organome-
tallics 1992, 11, 3009; c) Z. Csꢂkai, R. Skoda-Fçldes, L. Kollꢂr,
Inorg. Chim. Acta 1999, 286, 93; d) F. Ozawa, A. Kubo, T. Hayashi,
Chem. Lett. 1992, 2177.
[28] P. van Leeuwen, Appl. Catal. A 2001, 212, 61.
[29] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation
for Organic Synthesis, Wiley, New York, 2001.
[30] S. Hironao, N. Hisamitsu, W. Sueo, K. Hiroya, M. Toshihiro, I.
Shinji, I. Akiko, (National Industrial Research Institute of Nagoya,
NAGARA BIONICS CO.), JP2009185025, 2009.
Acknowledgements
We thank Dr. Adam Webb and Dr. Maarten Nachtegaal for their assis-
tance during XANES measurements, Dr. Wolfgang Kleist for the XPS
measurements and HASYLAB at DESY (Germany) and the SLS at PSI
(Switzerland) for beam time. Financial support by the Swiss National
Foundation and the European Community (through the Integrated Infra-
structure Initiative “Integrating Activity on Synchrotron and Free Elec-
tron Laser Science”, Contract RII3-CT-2004-506008) is kindly acknowl-
edged.
[31] B. E. Bent, Chem. Rev. 1996, 96, 1361.
[32] D. V. Tripkovic, D. Strmcnik, D. van der Vliet, V. Stamenkovic,
N. M. Markovic, Faraday Discuss. 2009, 140, 25.
[33] T. P. Beebe, J. E. Crowell, J. T. Yates, J. Phys. Chem. 1988, 92, 1296.
[34] a) A. G. Cook, E. K. Fields, J. Org. Chem. 1962, 27, 3686; b) E. K.
Fields, J. M. Sandri, Chem. Ind. 1959, 1216.
[35] D. Ferri, T. Bꢇrgi, J. Am. Chem. Soc. 2001, 123, 12074.
[36] J. Rasko, J. Bontovics, F. Solymosi, J. Catal. 1993, 143, 138.
[37] M. Yamaguchi, T. Shima, T. Yamagishi, M. Hida, Tetrahedron:
Asymmetry 1991, 2, 663.
[38] R. Hess, F. Krumeich, T. Mallat, A. Baiker, Catal. Lett. 2004, 92,
141.
[39] A. H. M. de Vries, F. J. Parlevliet, L. Schmieder-van de Vondervoort,
J. H. M. Mommers, H. J. W. Henderickx, M. A. M. Walet, J. G. de V-
ries, Adv. Synth. Catal. 2002, 344, 996.
[1] a) Z. Lu, S. M. Ma, Angew. Chem. 2008, 120, 264; Angew. Chem.
Int. Ed. 2008, 47, 258; b) G. P. McGlacken, I. J. S. Fairlamb, Eur. J.
Org. Chem. 2009, 4011; c) B. M. Trost, J. Org. Chem. 2004, 69, 5813.
[2] L. X. Yin, J. Liebscher, Chem. Rev. 2007, 107, 133.
[3] a) F. Miyazaki, K. Yamaguchi, M. Shibasaki, Tetrahedron Lett. 1999,
40, 7379; b) M. Ohff, A. Ohff, D. Milstein, Chem. Commun. 1999,
357; c) M. Rohr, J.-D. Grunwaldt, A. Baiker, J. Catal. 2005, 229, 144.
[4] a) R. A. Sheldon, M. Wallau, I. W. C. E. Arends, U. Schuchardt,
Acc. Chem. Res. 1998, 31, 485; b) M. T. Reetz, E. Westermann, R.
Lohmer, G. Lohmer, Tetrahedron Lett. 1998, 39, 8449; c) J. G. de V-
ries, Dalton Trans. 2006, 421.
[5] a) J. A. Widegren, R. G. Finke, J. Mol. Catal. A. 2003, 198, 317;
b) A. Biffis, M. Zecca, M. Basato, J. Mol. Catal. A 2001, 173, 249;
c) N. T. S. Phan, M. Van Der Sluys, C. W. Jones, Adv. Synth. Catal.
2006, 348, 609; d) D. Astruc, Inorg. Chem. 2007, 46, 1884.
Chem. Eur. J. 2010, 16, 9658 – 9668
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9667

A. Baiker et al.
[40] I. D. G. Watson, S. A. Styler, A. K. Yudin, J. Am. Chem. Soc. 2004,
126, 5086.
[44] T. Hayashi, A. Yamamoto, T. Hagihara, Y. Ito, Tetrahedron Lett.
1986, 27, 191.
[45] J.-D. Grunwaldt, M. Ramin, M. Rohr, A. Michailovski, G. R.
Patzke, A. Baiker, Rev. Sci. Instrum. 2005, 76, 054104.
[46] T. Ressler, J. Synchrotron Radiat. 1998, 5, 118.
[41] M. Berthod, C. Saluzzo, G. Mignani, M. Lemaire, Tetrahedron:
Asymmetry 2004, 15, 639.
[42] U. Leutenegger, G. Umbricht, C. Fahrni, P. Vonmatt, A. Pfaltz, Tet-
rahedron 1992, 48, 2143.
Received: April 1, 2010
Published online: July 21, 2010
[43] S. R. Gilbertson, P. Lan, Org. Lett. 2001, 3, 2237.
9668
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9658 – 9668