Potential of ionic liquids as co-modifiers in asymmetric hydrogenation on platinum

Article abstract of DOI:10.1016/j.molcata.2012.01.029

The enantioselective hydrogenation of methyl benzoylformate on Pt/Al ₂O₃ has been studied in the presence of various ionic liquids (ILs) and three structurally related chiral modifiers: cinchonidine (CD), O-phenyl-cinchonidine (PhOCD

Full text of DOI:10.1016/j.molcata.2012.01.029

Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Potential of ionic liquids as co-modifiers in asymmetric hydrogenation on
platinum
Shogo Sano, Matthias Josef Beier, Tamas Mallat, Alfons Baiker∗
Department of Chemistry and Applied Biosciences, ETH Zurich, Honggerberg, HCI, CH-8093 Zurich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
The enantioselective hydrogenation of methyl benzoylformate on Pt/Al2O3 has been studied in the pres-
ence of various ionic liquids (ILs) and three structurally related chiral modifiers: cinchonidine (CD),
O-phenyl-cinchonidine (PhOCD), and O-naphthyl-cinchonidine (NaphOCD). Addition of ca. 1% IL to a
polar organic solvent – particularly alcohols – improved the enantioselectivity by up to 12% (to 92–93%)
in the presence of CD. On the other hand, ILs diminished the reaction rate sometimes dramatically,
by two orders of magnitude, and the poisoning effect was even stronger in cyclohexene hydrogenation
under identical conditions. Comparative studies revealed that only ILs possessing a heteroaromatic cation
interact strongly with Pt or the O-aryl function of PhOCD and NaphOCD. A tentative explanation is the
strong ␲-bonding interaction of ILs with the Pt surface (leading to site blocking and deactivation) and
with the O-aryl function of the modifier not involved in the adsorption onto Pt (resulting in a shift in
enantioselectivity).
Received 15 December 2011
Received in revised form 27 January 2012
Accepted 30 January 2012
Available online 8 February 2012
Keywords:
Enantioselective hydrogenation
␣-Ketoester
Ionic liquids
Cinchona alkaloids
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
addition of an IL in small amounts to the polar solvent of the reac-
tion [27,28].
There has been a rapid development in catalysis in the past
decade using ionic liquids (ILs) as solvents or catalysts of various
transformations [1–6]. The properties of ILs can be tuned in a broad
range by varying the cations and anions. Application of ILs in enan-
tioselective hydrogenations with chiral complexes is a thoroughly
investigated, successful approach [7–9]. In heterogeneous catalytic
hydrogenations on metal surfaces the use of ILs has yet been lim-
ited to the saturation of C C and C O bonds in simple molecules
[10,11] and to chemoselective transformations. The latter reactions
include the partial hydrogenation of various ␣,␤-unsaturated car-
bonyl compounds [12–17], halonitro-compounds [18,19], alkynes
[20,21], cyclooctadiene [22], and propionitrile [23].
The potential of ILs to influence the enantioselection is the
topic of the present study. The test reaction is the hydrogena-
tion of methyl benzoylformate (1, Scheme 1) on the commonly
used Pt/Al2O3 reference catalyst modified by cinchonidine (CD)
and its O-phenyl and O-naphthyl derivatives PhOCD and NaphOCD,
respectively (Scheme 1). A fundamental question addressed here is
the nature of interactions between the IL, the Pt surface, and the
reaction components. To clarify this point we studied the trans-
formation (i) in a weakly polar organic solvent and added various
ILs as a layer supported on the catalyst (catalyst modifier), (ii) in a
polar solvent mixture containing the IL in small amounts (reaction
modifier), and (iii) applied the IL as the solvent itself.
Despite of the excellent results achieved in various fields of het-
erogeneous and homogeneous catalysis, application of ILs has also
some disadvantages. ILs are still expensive for large scale applica-
tions; they are characterized by relatively high viscosity and low
substrate diffusion rates leading to mass transport limitations. An
attractive strategy to overcome some of these difficulties is the
application of a thin supported IL layer on a solid surface [24].
This method allows heterogenization of soluble metal complexes
in continuous flow fixed-bed reactors [25] and also biphasic hydro-
genation reactions [26]. A less common approach is the simple
The enantioselective hydrogenation of ␣-ketoesters has been
intensively studied in the past decades, mainly using the sim-
plest substrate, pyruvate esters [29,30]. Only after recognizing the
importance of side reactions involving the ␣-H atom of pyru-
vate esters [31–35] has the interest been shifted to the more
stable benzoylformates as suitable model substrates. The general
characteristics of this reaction have already been clarified. Under
optimized conditions, in an acetic acid–toluene mixture at 273 K,
98% ee was obtained by Bartók’s group [36] but the best ees in the
absence of acid and at room temperature were significantly lower,
in the range 84–88% [37,38].
The present study shows that ILs possessing a heteroaromatic
ring (Fig. 1) may strongly interact with the chiral modifier and the
metal surface, and thus can influence the reaction rate and enan-
tioselection.
∗
Corresponding author. Fax: +41 44 632 11 63.
E-mail address: baiker@chem.ethz.ch (A. Baiker).
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2012.01.029

118
S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
received (Fig. 1): [BMIm][BF4] (99%, ABCR), [BMIm][PF6] (98%,
OH
O
Acros), [BMIm][FEP] (99%, Merck), [BMIm][NTf2] (99%, ABCR),
[BMPy][BF4] (for synthesis, Merck), [BMPy][PF6] (99%, ABCR),
[BMPr][OTf] (for synthesis, Merck), [BMPr][FEP] (for synthesis,
Merck), [MMPp][NTf2] (98%, Merck), [MMMo][NTf2] (98%, Merck).
The 5 wt.% Pt/Al2O3 (E 4759) was purchased from BASF (Engelhard).
For the synthesis of catalysts modified with a supported IL layer,
50 mg freshly activated 5% Pt/Al2O3 was put in a Schlenk tube
together with a magnetic stir bar. The appropriate amount of IL
(15 mg for a 30% loading) was introduced as a solution in about 1 ml
of acetone. The mixture was gently stirred and acetone removed
slowly in a flow of N2 overnight. The catalyst was finally dried in
vacuum for several hours and stored under N2 atmosphere.
O
O
Pt/Al2O3, H2, Modifier
10 bar, r.t.
O
O
1
2
X
Modifier
Product
N
OX
H
CD
(R)-2
(S)-2
(S)-2
H
N
Phenyl
PhOCD
1-Naphthyl NaphOCD
Modifier
2.2. Catalytic hydrogenations
Scheme 1. Test reaction and the structures of the chiral modifiers.
As a standard reaction procedure, hydrogenation of 1 was car-
ried out in a magnetically stirred stainless steel autoclave with
50 ml glass inlet and PTFE cover. The pressure (usually 10 bar)
was controlled by a computerized constant-volume, constant-
pressure equipment (Büchi BPC 9901). Prior to the reaction, the
5 wt.% Pt/Al2O3 was pretreated in flowing N2 at 400 ◦C (673 K)
for 30 min, then in hydrogen at the same temperature for 60 min.
After cooling to room temperature in hydrogen, the activated
catalyst was flushed overnight by nitrogen. This step is neces-
sary to slowly reoxidize the surface of Pt particles by the oxygen
impurity in the nitrogen flow and thus avoid the uncontrolled
interaction of residual surface hydrogen with oxygen in open air.
After pretreatment, 10 mg of 5 wt.% Pt/Al2O3 was introduced in the
homogeneous reaction mixture containing 250 mg (1.5 mmol) of
1, 30 mg of IL (if any), 1 or 2 mg of modifier and 3 ml of solvent.
After stirring at 750 rpm for 60 min under nitrogen atmosphere
2. Experimental
2.1. Materials
Methyl benzoylformate (1, 99%, Acros), toluene (99.7%, Fluka),
3-pentanone (98%, Fluka), acetic acid (AcOH, 99.8%, Fluka),
acetonitrile (99.5%, Sigma–Aldrich), 1,2-dichlorobenzene (99%,
Acros), tetrahydrofuran (THF, 99.5%, Sigma–Aldrich), 2-propanol
(99.8%, Sigma–Aldrich), 2-butanol (99%, Acros), triethylamine
(99.5%, Sigma–Aldrich), piperidine (99%, Sigma–Aldrich), pyri-
dine (98%, J.T. Baker), morpholine (99%, Fluka), and cinchonidine
(CD, 96%, Aldrich) were used as received. Synthesis of O-
phenyl-cinchonidine (PhOCD) and O-(1-naphthyl)-cinchonidine
(NaphOCD) was described elsewhere [39]. The ILs were used as
OMe
N+
Cations:
N+
N
N+
O
1-butyl-1-methylpyrrolidinium
[BMPr]+
1-methoxy-1-methylmorpholinium
1-butyl-3-methylimidazolium
[BMIm]+
[MMMo]+
OMe
N+
N+
1-butyl-3-methylpyridinium
[BMPy]+
1-(2-methoxyethyl]-1-methylpiperidinium
[MMPp]+
Anions:
F
P-
O
S
F
F
F
F
F
F
F3C
O-
B-
F
F
F
O
tetrafluoroborate
[BF4]-
hexafluorophosphate
[PF6]-
trifluoromethanesulfonate
[OTf]-
F
P-
0
S
O
O
S
F
CF2CF3
F3C
N-
CF3
F3CF2C
CF2CF3
F
O
tris(pentafluoroethyl)trifluorophosphate
[FEP]-
bis(trifluoromethylsulfonyl)imide
[NTf2]-
Fig. 1. Cations and anions of ILs used in this study.

S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
119
Table 1
Effect of organic solvents and ILs on the hydrogenation of 1 to (R)-2 on CD-modified
Pt/Al2O3. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3, 1 mg CD, 3 ml solvent (organic or
IL); 10 bar, room temperature.
Solvents
Time (min)
Conv. (%)
ee (%)
TOF (h−1
)
Toluene
Acetic acid
60
15
15
15
15
60
15
15
180
120
120
180
120
1440
1440
30
35
54
69
75
43
63
39
82
56
95
91
94
14
100
100
9
88
78
92
87
90
90
76
84
9
4
0
0
0
71
87
83
86
46
64
210
1300
1700
1800
1100
380
910
2000
110
300
270
190
43
3-Pentanone
Tetrahydrofuran
Ethyl acetate
1,2-dichlorobenzene
2-Propanol
2-Butanol
Acetonitrile
Et3N
Piperidine
Pyridine
Morpholine
[BMPr][FEP]
[MMMo][NTf2]
[MMMo][NTf2]
[MMPp][NTf2]
[BMIm][BF4]
[BMIm][PF6]
–
–
110
140
3
Fig. 2. The influence of [BMIm][BF4] added to the reaction mixture before hydro-
genation of 1 in acetic acid. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3, 3 ml AcOH,
2 mg CD, 20 bar, room temperature, 10 min reaction time.
30
780
720
12
6
9
5
(ˇ = 0.71) (almost) ceased the enantioselection. The poor ee of 9% in
acetonitrile may be due to competing hydrogenation of the solvent
to ethylamines [42]. The detrimental effect of basic impurities of the
catalyst on the enantioselective hydrogenation of 1 was reported at
first by Orito et al. [43] and the very low ees in amine type organic
solvents in the hydrogenation of ethyl pyruvate by Blaser et al. [44],
but the phenomena could not be interpreted.
The ILs applied in Table 1 as solvents may be divided into
two categories. Those containing a cycloaliphatic cation ([BMPr]+,
[MMMo]+, [MMPp]+, Fig. 1) allow reasonably good ees, while the
presence of the heteroaromatic cation [BMIm]+ lowers the enan-
tioselection and in particular the reaction rate. As concerns the low
reaction rates in the presence of [BMIm]+, there is an increasing
amount of experimental indications to the poisoning effect of ILs,
including imidazolium-based salts, in hydrogenation reactions on
Pt-group metal catalysts [15,22,45,46]. It has been proposed that
imidazolium-based ILs may react with Ir(0) nanoclusters and form
“surface-attached carbenes”, although deactivation due to strongly
adsorbed, ␲-bonded imidazolium cations could not be excluded
[47]. Not only the [BMIm]+ cation but also some anions in the
order [PF6]− < [BF4]− < [CF3SO3]− may be the origin of deactivation
[48].
(pre-equilibration to improve the reproducibility), the reaction
started at room temperature by stirring after the introduction of
hydrogen. The Pt/Al2O3 modified with an IL layer was used directly,
without a second reductive treatment at 400 ◦C.
The conversions and enantiomeric excesses (ee) were deter-
mined by using a Thermo Finnigan trace gas chromatograph with a
CP-Chirasil-DEX CB (25 m × 0.25 mm × 0.25 ␮m) capillary column.
The reaction rates are characterized by the turn-over frequencies
(TOFs). TOFs are related to the total amount of Pt present in the
reaction mixture (mol substrate converted by 1 mol Pt in 1 h). The
estimated error of TOFs was about 10% and that of ee was about
1–2%. The reactions were repeated several times to reduce the
standard deviation. The sometimes poor reproducibility could be
improved by deactivating the glassware with trimethylsilyl chlo-
ride. The glassware was pretreated by immersing it into aqua regia
for 15 min. After washing with water and drying, it was treated with
a 10 vol.% toluenic solution of trimethylsilyl chloride for 30 min.
After careful washing with toluene, methanol and acetone, the
glassware was dried in vacuum.
Hydrogenation of cyclohexene was carried out according to the
standard procedure but at 3 bar, using 1.5 mmol of cyclohexene,
30 mg of IL and 3 ml of solvent.
3.2. Ionic liquids as solvent components (reaction additives)
3. Results and discussion
Due to the high cost of ILs as solvents, it is more attractive
to apply them in small amounts as additives. The influence of
[BMIm][BF4] additive in acetic acid solvent is illustrated in Fig. 2.
The poisoning effect is obvious: the higher the amount of IL in the
system, the lower is the reaction rate. After the fast initial drop
the deactivation was less pronounced. Multiplying the amount of
IL from 9 to 90 mg resulted in only a relatively small decrease of
TOF from 1400 to 1100 h−1 (not shown). Note that 90 mg of IL
corresponds to 2.8 mass% of the solvent acetic acid. The effect of
[BMIm][BF4] as reaction additive on the enantioselectivity was neg-
ligible: all values in Fig. 2 varied in a narrow range of 87 1%.
A significant decrease to 84% was observed only by the tenfold
increase (to 90 mg) of the amount of IL.
The effect of constant amounts of [BMIm][FEP] added to differ-
ent organic solvents is illustrated in Table 2, under partly different
conditions. In this series of experiments the amount of ILs (30 mg)
was increased to a relatively high level compared to those in Fig. 2
in order to suppress the effect of the amount of ILs on the cata-
lyst performance. In most solvents – excluding the alcohols – the
enantioselectivity barely changed. In the weakly polar, aromatic
3.1. Solvent effect in the hydrogenation of methyl benzoylformate
(1)
At first we investigated some commonly used ILs as solvents and
compared their properties to those of organic solvents (Table 1).
The reaction rates are characterized by the average TOF calculated
from the conversion and the total amount of Pt present in the reac-
tor. Although the effect of organic solvents on the reaction rate
and enantioselectivity is very strong, no general correlation with
the typical solvent properties (ETN, ˛, or ˇ [40,41]) could be estab-
lished. In the hydrogenation of benzoylformates [36] and other
␣-ketoesters the highest ees are commonly achieved in the pres-
ence of the good H-bond donor acetic acid (˛ = 1.12). Under the
conditions applied here, some basic (H-bond acceptor) solvents,
such as 3-pentanone (ˇ = 0.45) and ethyl acetate (ˇ = 0.45), allowed
significantly higher ees up to 92%. On the other hand, the ee was 90%
also in the very weakly basic and non-acidic solvent dichloroben-
zene (ˇ = 0.03, ˛ = 0), while strongly basic solvents, such as Et3N

120
S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
Table 2
Influence of small amounts of [BMIm][FEP] added to different solvents prior to the hydrogenation of 1, using CD-modified Pt. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3, 1 mg
CD, 3 ml solvent with or without [BMIm][FEP] (30 mg), 10 bar, room temperature.
Solvent
[BMIm][FEP]
Time (min)
Conv. (%)
ee (%)
TOF (h−1
)
Relative rate
3-Pentanone
−
15
150
69
76
92
92
1700
180
1.00
0.11
+
Tetrahydrofuran
Acetic acid
−
15
150
75
21
87
86
1800
50
1.00
0.03
+
−
15
15
54
29
78
79
1300
730
1.00
0.56
+
Ethyl acetate
1,2-Dichlorobenzene
2-Propanol
−
15
90
43
57
90
93
1100
240
1.00
0.22
+
−
60
60
63
65
90
92
380
400
1.00
1.05
+
−
15
15
39
33
76
88
910
820
1.00
0.90
+
2-Butanol
−
15
15
82
50
84
92
2000
1200
1.00
0.60
+
solvent 1,2-dichlorobenzene the influence of the IL is negligibly
small: the change in the reaction rate and ee does not exceed
the estimated experimental error. A tentative explanation for the
missing influence is the strong ␲-bonding interaction between the
aromatic ring of the solvent and the [BMIm]+ cation, as described
for imidazolium-based ILs and benzene or toluene [49–52]. This
interaction with the solvent probably minimizes the amount of IL
adsorbed on Pt. The TOFs are in general lower in the presence of
the IL, probably indicating some site blocking at the Pt surface,
as will be discussed later. In the extreme case, in tetrahydrofu-
ran, ca. 1 mass% IL in the solvent diminished the hydrogenation
rate of 1 by 97%. A surprising result is the 8–12% increase of enan-
tioselectivity in alcoholic solvents by addition of [BMIm][FEP]. For
comparison, hydrogenation of ␣-ketoesters in alcohols never pro-
vided outstanding ees, probably due to some hemiketal formation
[31,53].
Next, we choose 3-pentanone as solvent and varied the chem-
ical nature of the IL (Table 3). Clearly, both the cations and the
anions influence the enantioselective hydrogenation of 1. A signifi-
cant negative effect on the enantioselectivity was observed only in
the presence of [BF4]− and [PF6]− anions. Excluding these anions
the ee barely changed, independent of the nature of the cation.
It is reasonable to assume that the lower reaction rate, compared
to the reference reaction without IL, reflects the strong adsorp-
tion of the IL on the Pt surface resulting in site blocking. In other
words, the IL, the substrate, and the chiral modifier compete for the
same Pt0 active surface sites. Based on this assumption we can con-
clude that heteroaromatic cations adsorb much stronger than the
cycloaliphatic cations. The order of adsorption under these condi-
tions is: [BMIm]+ > [BMPy]+ > [BMPr]+ ≈ [MMPp]+ ≈ [MMMo]+. Also
the nature of the anion affects the adsorption strength, as can be
seen for example by comparing the effect of four different salts with
[BMIm]+ cation.
of enantioselection, independent of the chemical nature of the
cation. On the other hand, ILs containing heteroaromatic cations
adsorb strongly on Pt, cover a considerable fraction of surface sites
and diminish the reaction rate.
3.3. Catalyst poisoning in cyclohexene hydrogenation: a
comparison
To support our assumption on the competitive adsorption of ILs
on Pt, we repeated the experiments in Table 3 and substituted 1 and
CD with cyclohexene. Hydrogenation of the C C bond in cyclohex-
ene is a facile reaction that – in contrast to the enantioselective
reaction – does not require an extended ensemble of active sites
[55]. The results in Table 4 demonstrate that this simple test reac-
tion is poisoned stronger than the enantioselective hydrogenation
of 1. Here even the ILs possessing a cycloaliphatic cation diminish
the activity of Pt by 38–62% and the poisoning effect of heteroaro-
matic cations with various anions may be higher than 99%. For each
IL additive, the relative rate was lower in the hydrogenation of
cyclohexene than in case of 1.
An obvious explanation is the relatively weak adsorption of
cyclohexene, compared to that of 1 (possessing electron-rich
phenyl, carbonyl and ester groups) and in particular to CD [56–58].
A further probable component of catalyst poisoning by ILs is the
hydrophobic character of cyclohexene, while the ILs adsorbed on
the Pt surface are strongly hydrophilic and may hinder the adsorp-
tion of the olefin.
It has been found that addition of small amounts of ILs frequently
improved the chemoselectivity but at the cost of the reaction rate
[22,59]. Complete hydrogenation of highly reactive olefins such as
1-hexene [48] or cyclooctadiene [22] required 50–75 ◦C at medium
pressures. In the absence of ILs these transformations occur on Pt-
group metal catalysts already under ambient conditions [60].
We have to consider that catalyst deactivation might also be
caused by impurities. The striking, more than 100-fold deactivation
in the presence of [BF4]− anion might also be due to its hydrolysis
and the formation of HF [54]. This side reaction is less facile with the
[PF6]− anion. Note that a small amount of water is always present at
the catalyst surface at the beginning of the hydrogenation reaction
originating from the reduction of oxygen impurities and surface
oxygen on the catalyst. A further possibility is the hydrodehalo-
genation of the F-containing anions [BF4]− and [PF6]− that would
result in the loss of hydrogenation activity of Pt.
3.4. Ionic liquid–cinchona interaction: structural effects
The data in Table 3 revealed no sign of influence of chemically
stable ILs on the enantioselection, that is on the substrate–chiral
modifier interaction. To better understand this critical point, we
extended the study to O-arylated derivatives of CD. In the hydro-
genation of various activated ketones PhOCD and NaphOCD afford
the (S)-alcohols in excess, although the actual ee depended also on
the reaction conditions [61]. Inversion of the major product was
attributed to the steric hindrance by the bulky aryloxy group [62].
The use of these three modifiers provides an excellent opportunity
A feasible conclusion from the results in Table 3 is that (non-
reactive) ILs do not disturb the interaction of CD with 1, the source

S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
121
Table 3
The effect of small amounts of ILs added to 3-pentanone prior to hydrogenation of 1 on CD-modified Pt. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3, 1 mg CD, 3 ml solvent, 30 mg
IL, 10 bar, room temperature.
Ionic liquid
No
Time (min)
15
Conv. (%)
69
ee (%)
92
TOF (h−1
1700
)
Relative rate
1.00
Cycloaliphatic cation
[BMPr][OTf]
[BMPr][FEP]
[MMPp][NTf2]
[MMMo][NTf2]
15
15
15
15
40
72
67
63
89
92
92
92
980
1700
1600
1500
0.58
1.00
0.94
0.88
Aromatic cation
[BMIm][BF4]
[BMIm][PF6]
[BMIm][FEP]
[BMIm][NTf2]
[BMPy][BF4]
[BMPy][PF6]
900
180
150
150
180
60
32
60
76
69
56
56
77
92
92
92
65
79
13
120
180
170
120
340
0.008
0.07
0.11
0.10
0.07
0.20
the reaction rate, the data indicate a small but significant poisoning
effect of the [PF6]− anion.
N
N
For the interpretation of the results we assume that a ␲–␲
bonding interaction between the aryloxy group of the alkaloid
derivative and the heteroaromatic cation of the IL is responsible
for the shift in the product distribution. Adsorption of the sub-
strate on Pt involves the phenyl group and the two electron-rich
carbonyl groups [63]. During interaction with the ketone sub-
strate, also the quinoline ring of the alkaloid adsorbs strongly on
Pt in (close to) parallel position. Hence, the [BMIm]+ cation can-
not interact with these electron-rich groups and the ILs have only
minor or negligible influence on the ee when CD is applied. In
case of PhOCD and NaphOCD the situation is different: the ary-
loxy function does not adsorb parallel to the Pt surface but rather
in a tilted position [62] and thus it is available for interaction
with the imidazolium ring of the IL. This interaction changes the
shape and size of the chiral site available for adsorption of the
ketone substrate and thus controls the ee. The existence of this type
of ␲-bonding interactions has been evidenced between various
ILs containing 1-alkyl-3-methylimidazolium cations and aromatic
hydrocarbons such as benzene [49,51,52], toluene, [49,50] and 1-
methylnaphthalene [64].
O
H
O
Ph
C
COOEt
Fig. 3. Illustration of the N–H–O type interaction between the quinuclidine N atom
of the alkaloid and 1 adsorbed on Pt, and the tilted position of the O-phenyl ring
relative to the Pt surface.
to study structural effects, since aryl substitution at the OH func-
tion of CD induces only steric effects; the stereogenic centers of
the modifiers and the N–H–O type interaction between the quin-
uclidine N of the chiral modifier and keto-carbonyl group of the
substrate (Fig. 3) remain the same.
We repeated the experiments in Table 3 with the only difference
that CD was substituted with PhOCD. The results in Table 5 show
that ILs possessing cycloaliphatic cations barely reduced the ee to
(S)-2 (with the exception of [BMPr][OTf]), and the loss of catalyst
reactivity was always less than 50%. The effect of heteroaromatic
cations was much stronger: the ee dropped to one-half or even to
one-fifth and the rate decreased by up to 99%.
3.5. Supported ionic liquid layer or reaction modifier?
Table 6 presents a comparison involving all three chiral modi-
fiers. In this case the influence of only two typical ILs are compared,
both possessing the heteroaromatic cation [BMIm]+. There is no
influence on the ee in case of CD, while the (negative) effect of ILs
in the presence of PhOCD and NaphOCD is comparable. As concerns
We have mentioned previously the disadvantages of applying
ILs as solvents and in most cases we used them in small amounts
as reaction additives. This is a very simple approach and many
solvents dissolve small amounts of IL salts, but separation and recy-
cling of the expensive additive may be a limiting factor. To evaluate
Table 4
Hydrogenation of cyclohexene: the influence of IL additives containing cycloaliphatic and aromatic cations in 3-pentanone solvent. Conditions: 120 mg (1.5 mmol) cyclohexene,
30 mg IL, 10 mg 5% Pt/Al2O3, 3 ml solvent, 3 bar, room temperature.
Ionic liquid
No
Time (min)
15
Conv. (%)
66
TOF (h−1
1600
)
Relative rate
1.00
Cycloaliphatic cation
[BMPr][OTf]
[BMPr][FEP]
[MMPp][NTf2]
[MMMo][NTf2]
15
15
15
15
12
40
24
15
280
990
600
360
0.18
0.62
0.38
0.23
Aromatic cation
[BMIm][BF4]
[BMIm][PF6]
[BMIm][FEP]
[BMIm][NTf2]
[BMPy][BF4]
[BMPy][PF6]
120
120
120
120
30
2.5
4.2
6.1
3.3
14
7.5
13
19
10
180
280
0.005
0.008
0.012
0.006
0.11
30
23
0.18

122
S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
Table 5
The effect of small amounts of ILs on the hydrogenation of 1 to (S)-2 in 3-pentanone, using PhOCD-modified Pt. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3, 1 mg PhOCD, 3 ml
solvent, 30 mg IL, 10 bar, room temperature.
Ionic liquid
No
Time (min)
30
Conv. (%)
42
ee (%)
66
TOF (h−1
510
)
Relative rate
1.00
Cycloaliphatic cation
[BMPr][OTf]
[BMPr][FEP]
[MMPp][NTf2]
[MMMo][NTf2]
30
30
30
30
22
24
39
42
49
63
63
65
270
280
470
500
0.53
0.55
0.92
0.98
Aromatic cation
[BMIm][BF4]
[BMIm][PF6]
[BMIm][FEP]
[BMIm][NTf2]
[BMPy][BF4]
[BMPy][PF6]
900
900
900
900
180
120
8
15
18
19
29
26
13
22
33
32
18
22
4
6
7
8
59
79
0.008
0.012
0.014
0.016
0.12
0.15
Table 6
The critical role of the structure of the chiral modifier in the IL–modifier interactions during the hydrogenation of 1 in 3-pentanone. Conditions: 250 mg 1, 10 mg 5% Pt/Al2O3,
1 mg modifier, 3 ml solvent, 30 mg IL, 10 bar, room temperature.
Modifier
Ionic liquid
Time (min)
Conv. (%)
ee (%)
TOF (h−1
1700
120
180
510
6
7
270
7
)
Relative rate
CD
CD
CD
PhOCD
PhOCD
PhOCD
NaphOCD
NaphOCD
NaphOCD
No
15
180
150
30
900
900
60
69
60
76
42
15
18
45
18
21
92 (R)
92 (R)
92 (R)
66 (S)
22 (S)
33 (S)
44 (S)
20 (S)
24 (S)
1.00
0.07
0.11
1.00
0.012
0.014
1.00
0.026
0.033
[BMIm][PF6]
[BMIm][FEP]
No
[BMIm][PF6]
[BMIm][FEP]
No
[BMIm][PF6]
[BMIm][FEP]
900
900
9
the other commonly used approach, we deposited 30 mass% of an
IL as a separate layer onto the surface of the 5% Pt/Al2O3 catalyst
and carried out the enantioselective hydrogenations in the weakly
polar solvent toluene that is a poor solvent of ILs. An even less polar,
hydrophobic hydrocarbon solvent would be a better choice, but the
low solubility of cinchona alkaloids does not allow this choice. The
study involves three ILs with heteroaromatic and cycloaliphatic
cations and three different anions, and also two different chiral
modifiers. The effect of the same ILs used as reaction additives in
3-pentanone are also included in Table 7 for comparison.
layer approach, 13 mg catalyst contained 3 mg IL. This amount of
salt probably remained there during reaction in toluene, although
its distribution on the Pt and on the support is unknown. In the other
series 30 mg IL was added to the reaction mixture and probably only
a small fraction of this salt adsorbed onto the catalyst surface.
Analysis of the results reveals in Table 7 that independent of the
approach applied, [BMPr][FEP] possessing a cycloaliphatic cation
has negligible influence on the reaction, when CD is used as chiral
modifier. In the presence of PhOCD the catalyst deactivation is more
significant when [BMPr][FEP] is added to the solvent as a reaction
additive, but the enantioselectivities are barely affected. A similar
conclusion may be drawn for [BMIm] [FEP] and [BMIm][BF4]: with
An important difference between the two series of experiments
is the amount of IL on the catalyst surface. Using the supported IL
Table 7
Comparison of the effect of ILs as a supported layer on Pt/Al2O3 or as a reaction additive in the hydrogenation of 1. Conditions: 250 mg 1, 1 mg modifier, 3 ml solvent, 10 bar,
room temperature; catalyst: 10 mg 5% Pt/Al2O3 and 30 mg IL, or 13 mg 5% Pt/Al2O3 covered by 30 mass% IL layer.
Ionic liquid
Modifier
Solvent
Time (min)
Conv. (%)
ee (%)
TOF (h−1
)
Rel. rate
As a supported layer
No
[BMPr][FEP]
[BMIm][FEP]
[BMIm][BF4]
CD
CD
CD
CD
Toluene
Toluene
Toluene
Toluene
60
120
240
240
35
65
51
10
88 (R)
88 (R)
88 (R)
82 (R)
210
200
76
1.00
0.95
0.36
0.07
15
No
PhOCD
PhOCD
PhOCD
PhOCD
Toluene
Toluene
Toluene
Toluene
120
120
120
720
24
22
14
4
44 (S)
41 (S)
31 (S)
17 (S)
73
68
43
2
1.00
0.93
0.59
0.03
[BMPr][FEP]
[BMIm][FEP]
[BMIm][BF4]
As a reaction additive
No
[BMPr][FEP]
[BMIm][FEP]
[BMIm][BF4]
CD
CD
CD
CD
3-Pentanone
3-Pentanone
3-Pentanone
3-Pentanone
15
15
150
900
69
72
76
32
92 (R)
92 (R)
92 (R)
77 (R)
1700
1700
180
13
1.00
1.00
0.11
0.008
No
PhOCD
PhOCD
PhOCD
PhOCD
3-Pentanone
3-Pentanone
3-Pentanone
3-Pentanone
30
30
330
900
42
24
8
66 (S)
63 (S)
25 (S)
13 (S)
510
280
9
1.00
0.55
0.02
0.008
[BMPr][FEP]
[BMIm][FEP]
[BMIm][BF4]
8
4

S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
123
both chiral modifiers the (negative) effect on the reaction rate and
ee is stronger, when the IL is used as an additive in 3-pentanone. In
addition, comparison of the data allows differentiation between the
negative effect of the [BMIm]+ aromatic cation and the poisoning
of Pt by the [BF4]− anion.
tuning the enantioselectivity, especially for solvents usually allow-
ing inferior ees.
Acknowledgement
Financial support by the Swiss National Science Foundation
(Project 200020-131869) is kindly acknowledged.
4. Summary and conclusions
In the enantioselective hydrogenation of 1 to (R)-2 (Scheme 1)
over CD-modified 5% Pt/Al2O3 the highest enantioselectivities
(90–92%) were achieved in the organic solvents 3-pentanone, ethyl
acetate and 1,2-dichlorobenzene (Table 1). Strongly basic solvents
ceased the enantioselection and the commonly used solvents acetic
acid and toluene were not advantageous. In the six different ILs
tested as solvents, the ees varied in the range 46–87%. The ILs
containing the heteroaromatic cation [BMIm]+ (Fig. 1) reduced
the ee and diminished the hydrogenation rate typical for polar
organic solvents by more than two orders of magnitude. Interest-
ingly, application of them as reaction additives in about 1 mass%
improved the ee in some organic solvents, particularly in alco-
hols (Table 2, Fig. 2). The highest ee of 93% (without optimization)
was obtained in an ethyl acetate–[BMIm][FEP] mixture. The signifi-
cant shifts in the catalyst performance provide an indirect evidence
for the location of ILs on the Pt surface during the hydrogena-
tion reactions. In this respect, the effect of ILs on the reaction rate
and enantioselectivity was qualitatively the same, independent of
the method used to introduce them into the reaction mixture, i.e.
whether they were deposited onto the catalyst surface as a thin
supported layer prior to application in toluene that dissolves them
poorly, or they were simply added to a good polar solvent in larger
amount (Table 7).
An important target of the study was to understand the interac-
tion of ILs with the substrate, the chiral modifier and the Pt surface.
In general, IL additives possessing a cycloaliphatic cation (Fig. 1)
had only a small effect on the performance of the Pt–CD system
(Table 3). On the contrary, the heteroaromatic cations [BMIm]+ and
[BMPy]+ were detrimental. The most probable explanation is the
blocking of a fraction of surface Pt sites by the strongly adsorbed (␲-
bonded) cation [65,66]. This competition for the active surface sites
was even more pronounced in the hydrogenation of cyclohexene
under otherwise identical conditions (Table 4), which difference
is probably linked to the different adsorption strengths of the IL,
the substrate, and CD. Note that transformation (hydrolysis or
hydrogenolysis) of the anions [PF6]− and [BF4]− under reaction con-
ditions, and the resulting catalyst poisoning, may also contribute to
the observed deactivation. Nevertheless, all data indicate that the
changes in reactivity and enantioselectivity cannot be linked to a
direct interaction of the IL with the substrate (1 or cyclohexane) or
CD. According to our general knowledge, the electron-rich phenyl
and carbonyl groups of 1 and the extended aromatic ring system
of CD adsorb (close to) parallel to the Pt surface during hydrogen
uptake (Fig. 3) [30,67–70]. This strong interaction with the surface
Pt atoms prevents the significant interaction of these species with
the IL additive.
References
[1] Z.C. Zhang, Adv. Catal. 49 (2006) 153–237.
[2] V.I. Parvulescu, C. Hardacre, Chem. Rev. 107 (2007) 2615–2665.
[3] A. Winkel, P.V.G. Reddy, R. Wilhelm, Synthesis (2008) 999–1016.
[4] H. Olivier-Bourbigou, L. Magna, D. Morvan, Appl. Catal. A: Gen. 373 (2009) 1–56.
[5] F. Jutz, J.M. Andanson, A. Baiker, Chem. Rev. 111 (2011) 322–353.
[6] Q.H. Zhang, S.G. Zhang, Y.Q. Deng, Green Chem. 13 (2011) 2619–2637.
[7] K. Bica, P. Gaertner, Eur. J. Org. Chem. (2008) 3235–3250.
[8] M. Gruttadauria, S. Riela, P. Lo Meo, F. D’Anna, R. Noto, Tetrahedron Lett. 45
(2004) 6113–6116.
[9] U. Hintermair, T. Hofener, T. Pullmann, G. Francio, W. Leitner, ChemCatChem 2
(2010) 150–154.
[10] Y.L. Gu, G.X. Li, Adv. Synth. Catal. 351 (2009) 817–847.
[11] J.P. Hallett, T. Welton, Chem. Rev. 111 (2011) 3508–3576.
[12] K. Anderson, P. Goodrich, C. Hardacre, D.W. Rooney, Green Chem. 5 (2003)
448–453.
[13] R. Abu-Reziq, D. Wang, M. Post, H. Alper, Adv. Synth. Catal. 349 (2007)
2145–2150.
[14] Y. Kume, K. Qiao, D. Tomida, C. Yokoyama, Catal. Commun. 9 (2008) 369–375.
[15] J. Arras, M. Steffan, Y. Shayeghi, P. Claus, Chem. Commun. (2008) 4058–4060.
[16] Y. Hu, H.M. Yang, Y.C. Zhang, Z.S. Hou, X.R. Wang, Y.X. Qiao, H. Li, B. Feng, Q.F.
Huang, Catal. Commun. 10 (2009) 1903–1907.
[17] H. Hagiwara, T. Nakamura, T. Hoshi, T. Suzuki, Green Chem. 13 (2011)
1133–1137.
[18] D.Q. Xu, Z.Y. Hu, W.W. Li, S.P. Luo, Z.Y. Xu, J. Mol. Catal. A-Chem. 235 (2005)
137–142.
[19] X. Yuan, N. Yan, C.X. Xiao, C.N. Li, Z.F. Fei, Z.P. Cai, Y. Kou, P.J. Dyson, Green Chem.
12 (2010) 228–233.
[20] M. Ruta, G. Laurenczy, P.J. Dyson, L. Kiwi-Minsker, J. Phys. Chem. C 112 (2008)
17814–17819.
[21] R. Venkatesan, M.H.G. Prechtl, J.D. Scholten, R.P. Pezzi, G. Machado, J. Dupont,
J. Mater. Chem. 21 (2011) 3030–3036.
[22] U. Kernchen, B. Etzold, W. Korth, A. Jess, Chem. Eng. Technol. 30 (2007)
985–994.
[23] K. Obert, D. Roth, M. Ehrig, A. Schonweiz, D. Assenbaum, H. Lange, P. Wasser-
scheid, Appl. Catal. A-Gen. 356 (2009) 43–51.
[24] C.P. Mehnert, E.J. Mozeleski, R.A. Cook, Chem. Commun. (2002) 3010–3011.
[25] C.P. Mehnert, Chem. Eur. J. 11 (2005) 50–56.
[26] J. Dupont, G.S. Fonseca, A.P. Umpierre, P.F.P. Fichtner, S.R. Teixeira, J. Am. Chem.
Soc. 124 (2002) 4228–4229.
[27] M. Khodadadi-Moghaddam, A. Habibi-Yangjeh, M.R. Gholami, Appl. Catal. A-
Gen. 341 (2008) 58–64.
[28] J. Arras, M. Steffan, Y. Shayeghi, D. Ruppert, P. Claus, Green Chem. 11 (2009)
716–723.
[29] H.U. Blaser, H.P. Jalett, M. Müller, M. Studer, Catal. Today 37 (1997) 441–463.
[30] M. Bartók, Curr. Org. Chem. 10 (2006) 1533–1567.
[31] M. von Arx, T. Mallat, A. Baiker, Top. Catal. 19 (2002) 75–87.
[32] D. Ferri, T. Bürgi, K. Borszeky, T. Mallat, A. Baiker, J. Catal. 193 (2000)
139–144.
[33] J.M. Bonello, R.M. Lambert, N. Künzle, A. Baiker, J. Am. Chem. Soc. 122 (2000)
9864–9865.
[34] D. Ferri, T. Bürgi, A. Baiker, J. Phys. Chem. B 108 (2004) 14384–14391.
[35] D. Ferri, S. Diezi, M. Maciejewski, A. Baiker, Appl. Catal. A: Gen. 297 (2006)
165–173.
[36] M. Sutyinszki, K. Szöri, K. Felföldi, M. Bartók, Catal. Commun. 3 (2002) 125–127.
[37] K. Balázsik, I. Bucsi, S. Cserényi, G. Szöllösi, M. Bartók, J. Mol. Catal. A-Chem. 285
(2008) 84–91.
[38] F. Hoxha, B. Schimmoeller, Z. Cakl, A. Urakawa, T. Mallat, S.E. Pratsinis, A. Baiker,
J. Catal. 271 (2010) 115–124.
The situation is different, when an O-aryl function is introduced
into CD. The O-aryl function is adsorbed in a tilted position rel-
ative to the Pt surface (Fig. 3) and can easily interact with the
heteroaromatic cations of the ILs via ␲-bonding interactions. Sim-
ilar interactions involving a heteroaromatic cation of an IL and an
aromatic organic molecule have been confirmed by various spec-
troscopic methods [49–52,64]. As a result of this interaction the
chiral space available for adsorption of the ␣-ketoester substrate
changes, inducing a change in the substrate–chiral modifier inter-
action and in the enantioselectivity (Tables 5 and 6). This approach,
the double modification of the Pt surface with a strongly adsorbing
chiral molecule and an interacting IL offers a new possibility for
[39] S. Diezi, A. Szabo, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 14 (2003)
2573–2577.
[40] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, Wiley-VCH,
2003.
[41] P.G. Jessop, Green Chem. 13 (2011) 1391–1398.
[42] W.-R. Huck, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 200 (2001) 171–180.
[43] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1980) 670.
[44] H.U. Blaser, H.P. Jalett, J. Wiehl, J. Mol. Catal. 68 (1991) 215–222.
[45] L.S. Ott, S. Campbell, K.R. Seddon, R.G. Finke, Inorg. Chem. 46 (2007)
10335–10344.
[46] J. Arras, D. Ruppert, P. Claus, Appl. Catal. A-Gen. 371 (2009) 73–77.
[47] L.S. Ott, M.L. Cline, M. Deetlefs, K.R. Seddon, R.G. Finke, J. Am. Chem. Soc. 127
(2005) 5758–5759.
[48] L.M. Rossi, G. Machado, P.F.P. Fichtner, S.R. Teixeira, J. Dupont, Catal. Lett. 92
(2004) 149–155.

124
S. Sano et al. / Journal of Molecular Catalysis A: Chemical 357 (2012) 117–124
[49] J.D. Holbrey, W.M. Reichert, M. Nieuwenhuyzen, O. Sheppard, C. Hardacre, R.D.
Rogers, Chem. Commun. (2003) 476–477.
[60] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for Organic
Synthesis, Wiley, New York, 2001.
[50] T. Gutel, C.C. Santini, A.A.H. Padua, B. Fenet, Y. Chauvin, J.N.C. Lopes, F. Bayard,
M.F.C. Gomes, A.S. Pensado, J. Phys. Chem. B 113 (2009) 170–177.
[51] M.H. Kowsari, S. Alavi, M. Ashrafizaadeh, B. Najafi, J. Chem. Phys. 132 (2010)
0445071–0445079.
[61] S. Diezi, T. Mallat, A. Szabo, A. Baiker, J. Catal. 228 (2004) 162–173.
[62] N. Bonalumi, A. Vargas, D. Ferri, T. Bürgi, T. Mallat, A. Baiker, J. Am. Chem. Soc.
127 (2005) 8467–8477.
[63] A. Vargas, S. Reimann, S. Diezi, T. Mallat, A. Baiker, J. Mol. Catal. A: Chem. 282
(2008) 1–8.
[64] C. Hardacre, J.D. Holbrey, C.L. Mullan, M. Nieuwenhuyzen, T.G.A. Youngs, D.T.
Bowron, S.J. Teat, Phys. Chem. Chem. Phys. 12 (2010) 1842–1853.
[65] H.S. Schrekker, M.A. Gelesky, M.P. Stracke, C.M.L. Schrekker, G. Machado, S.R.
Teixeira, J.C. Rubim, J. Dupont, J. Colloid Interface Sci. 316 (2007) 189–195.
[66] H. Zhang, H. Cui, Langmuir 25 (2009) 2604–2612.
[67] T. Mallat, E. Orglmeister, A. Baiker, Chem. Rev. 107 (2007) 4863–4890.
[68] F. Zaera, J. Phys. Chem. C 112 (2008) 16196–16203.
[69] G. Kyriakou, S.K. Beaumont, R.M. Lambert, Langmuir 27 (2011) 9687–9695.
[70] D.Y. Murzin, P. Maki-Arvela, E. Toukoniitty, T. Salmi, Catal. Rev.-Sci. Eng. 47
(2005) 175–256.
[52] T. Shimomura, T. Takamuku, T. Yamaguchi, J. Phys. Chem.
8518–8527.
B 115 (2011)
[53] B. Minder, T. Mallat, P. Skrabal, A. Baiker, Catal. Lett. 29 (1994) 115–124.
[54] M.G. Freire, C. Neves, I.M. Marrucho, J.A.P. Coutinho, A.M. Fernandes, J. Phys.
Chem. A 114 (2010) 3744–3749.
[55] T. Bürgi, A. Baiker, Acc. Chem. Res. 37 (2004) 909–917.
[56] P.R.N. de Souza, D.A.G. Aranda, J.W.D. Carneiro, C.D.B. de Oliveira, O.A.C.
Antunes, F.B. Passos, Int. J. Quant. Chem. 92 (2003) 400–411.
[57] A. Vargas, A. Baiker, J. Catal. 239 (2006) 220–226.
[58] E. Schmidt, T. Mallat, A. Baiker, J. Catal. 272 (2010) 140–150.
[59] N. Wörz, J. Arras, P. Claus, Appl. Catal. A-Gen. 391 (2011) 319–324.