Structural requirements for substrate in highly enantioselective hydrogenation over the cinchonidine-modified Pd/C

Article abstract of DOI:10.1016/j.jcat.2008.11.022

Relationship between substrate structure and enantioselectivity is studied for the asymmetric hydrogenation of 42 different (E)-α, β-disubstituted acrylic acids (propenoic acids) over cinchonidine-modified Pd/C. The β-phenyl group is indispensable for high enantioselectivity of α-phenylcinnamic acid (2,3-diphenylpropenoic acid, 81% ee), and substitution on this group affects markedly the selectivity. The high ee up to 92% was achieved by the β - p-alkoxyphenyl substitution, and the selectivity is ascribed mainly to stronger interaction of the substrate with the chiral modifier on the catalyst surface. In contrast, substitution on the α-phenyl group does not affect notably the enantioselectivity (80-82% ee) or even the α-phenyl group itself is not indispensable but replaceable with a properly bulky group for the high enantioselectivity.

Full text of DOI:10.1016/j.jcat.2008.11.022

Journal of Catalysis 262 (2009) 57–64
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Structural requirements for substrate in highly enantioselective hydrogenation
over the cinchonidine-modified Pd/C
Takashi Sugimura a,∗, Takayuki Uchida a, Junya Watanabe a, Takeshi Kubota b, Yasuaki Okamoto b, Tomonori Misaki a,
Tadashi Okuyama a
a
Graduate School of Material Science, Himeji Institute of Technology, University of Hyogo, Kohto, Kamigori, Hyogo 678-1297, Japan
Department of Material Science, Shimane University, 1060 Nishikawatsu, Matsue 6908504, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Relationship between substrate structure and enantioselectivity is studied for the asymmetric hydro-
genation of 42 different (E)-α,β-disubstituted acrylic acids (propenoic acids) over cinchonidine-modified
Pd/C. The β-phenyl group is indispensable for high enantioselectivity of α-phenylcinnamic acid (2,3-
diphenylpropenoic acid, 81% ee), and substitution on this group affects markedly the selectivity. The
high ee up to 92% was achieved by the β-p-alkoxyphenyl substitution, and the selectivity is ascribed
mainly to stronger interaction of the substrate with the chiral modifier on the catalyst surface. In contrast,
substitution on the α-phenyl group does not affect notably the enantioselectivity (80–82% ee) or even
the α-phenyl group itself is not indispensable but replaceable with a properly bulky group for the high
enantioselectivity.
Received 29 September 2008
Revised 27 November 2008
Accepted 27 November 2008
Available online 30 December 2008
Keywords:
Asymmetric heterogeneous catalysis
Enantioselective hydrogenation
Palladium on carbon
Chiral modification
Olefin
© 2008 Elsevier Inc. All rights reserved.
Molecular recognition
1. Introduction
the high metal dispersion [12]. It was also reported more recently
◦
that a hydrogen-pretreated Pd/Al2O3 at 250 C was a suitable cat-
◦
alyst to give 73% ee (80% ee at 0 C) [13]. The newly introduced
Alkaloid-modified palladium is a unique heterogeneous cata-
lyst that can perform enantioselective hydrogenation of prochiral
olefins [1–3]. The enantioselectivity of the catalyst largely depends
on the chiral modifier employed, and cinchonidine (CD) and cin-
chonine (CN) are so far the best modifiers that show opposite
stereoselectivities to each other [4–8]. Choice of palladium cat-
alyst as a base catalyst for the modification is another impor-
tant factor. For the hydrogenation of a representative substrate,
α-phenylcinnamic acid (PCA), a 5% Pd/TiO2 prepared from a cer-
tain titania is so far the best palladium catalyst to give 72% ee of
the product [9], when it is modified with CD in a polar solvent,
such as water-containing (wet) DMF or wet dioxane (Scheme 1).
Low metal dispersion with a larger Pd particle size is believed to
be an essential factor to induce the good stereoselectivity [10].
Palladium-on-carbon (Pd/C) is a popular catalyst, but stereose-
lectivity of the hydrogenation is known to be poor in the range
of 20–50% ee when it is used with the CD modifier [4,11]. How-
ever, we have recently found that pretreatment under hydrogen at
catalysts are noteworthy because the reproducible stereoselectiv-
ities are obtained with commercially available catalysts, and any
skillful technique is no longer required for the catalyst prepara-
tion. With these catalysts in hand, high-throughput screening of
the CD-modified catalysis system became possible.
Several years ago, we have started a project to explore relation-
ship between the substrate structure and the enantioselectivity,
and found that the p-methoxy substitution of PCA brings about
improvement in the product ee [14]. When the improved cata-
lyst prepared from Pd/C was employed, the ee reached 92% with
ꢀ
p,p -dimethoxyphenylcinnamic acid (DMPCA) [12]. A report deal-
ing with the effects of methoxy substitution has recently been
published also by others [13]. In the present report, we would like
to present the results with PCA and its 41 analogues as hydrogena-
tion substrates to explore the electronic and steric requirements to
give the high ee [15]. Mechanisms for the stereoselective hydro-
genation will be discussed.
◦
80 C prior to the modification results in notable improvement in
the product ee, and the best ee value of 81% was obtained for the
hydrogenation of PCA with a certain ready-made Pd/C, in spite of
2. Experimental
STD-type 5% Pd/C was supplied from N.E. Chemcat in a wet
form (51%, w/w), the metal surface area of which is given as
339 m2 g−1, corresponding to metal dispersion of 76%. Our own
evaluation showed the dispersion is 66%, on the basis of the CO
Corresponding author. Fax: +81 791 58 0115.
E-mail address: sugimura@sci.u-hydogo.ac.jp (T. Sugimura).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.11.022

58
T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
Scheme 1. Enantioselective hydrogenation of PCA over CD-modified Pd/C.
adsorption capacity obtained by a pulse technique assuming a
CO/Pd ratio of 1:2 (for the catalyst pretreated with hydrogen at
200 C) [16]. AER-type 5% Pd/C and 5% Pd/Al2O3 (Engelhard 40692)
determined by the different chiral columns or by the repeated runs
were in agreement with one another ( 0.5%).
◦
Precise hydrogenation rates were determined by 1H NMR as
follows. In a 50 ml glass reactor with a small septum port, Pd/C
were also obtained from N.E. Chemcat (dispersion determined
by the producers = 67% and 19%, respectively). 5% Pd/TiO2 was
made by the reported procedure (dispersion = 26%) [10]. Phenyl-
and methylcinnamic acids were used after recrystallization from
acetone, but tiglic acid was used as received. α-Aryl substrates
were prepared by the Perkin reaction, and α-alkyl substrates were
prepared by the aldol condensation or the Reformatsky reaction.
1-Acenaphthylene and 9-phenanthrene carboxylic acids were pre-
pared from the corresponding bromides via the Grignard reaction.
Details of preparation and characterization of new compounds are
given in the supplementary material.
◦
(10 mg) in 2.5 ml of 2.5% (v/v) wet dioxane was heated at 80 C
◦
for 30 min, and after cooling to 23 C, a solution of cinchonidine
(3 mg, 0.02 mmol) in the wet dioxane (0.5 ml) was added. After
stirring for 30 min, a solution of a substrate or a mixture of two
substrates (0.5 mmol in total) in the wet dioxane (1 ml) and then
ethanolamine (0.5 mmol) were added. A portion of the reaction
mixture was taken out via a syringe during the hydrogenation and
subjected to 1H NMR analysis (600 MHz, CDCl3) to determine the
reaction conversion of each substrate. The ee value at the end of
reaction was determined by the HPLC.
Most hydrogenations were carried out with the STD-type 5%
Pd/C catalyst after the pretreatment as follows [12]. In a 50 ml
glass reactor with a small septum port, Pd/C (20 mg in dry form)
and 5 ml of 2.5% (v/v) wet dioxane were placed. Hydrogen of at-
mospheric pressure was charged to the flask, and the mixture was
3. Results
Unsaturated carboxylic acids employed for the substrates of hy-
drogenation are classified in seven groups A–G, as illustrated in
Fig. 1. Groups A–C include methyl-, methoxy-, and trifluoromethyl-
(fluoro-) substituted analogues, respectively. Group D is higher
analogues of DMPCA (B-6) and a naphthyl analogue (D-4). Group E
is α-alkyl analogues, while Group F is β-alkyl analogues. Group G
includes other types of substrates. The catalyst employed for the
modification was the pretreated Pd/C, the best catalyst for the PCA
hydrogenation [12]. The hydrogenation was carried out in a water-
◦
heated at 80 C for 30 min under stirring (1200 rpm). The suspen-
◦
sion was cooled to 23 C, and a solution of cinchonidine (6 mg,
0.02 mmol) in the wet dioxane (1 ml) was added. After stirring
for 30 min, a solution of a substrate (0.5 mmol, 100–150 mg) in
the wet dioxane (4 ml) was added, followed by addition of benzy-
lamine (BA, 55 μl, 0.5 mmol) [17]. The same hydrogenations were
also performed in the absence of BA. Hydrogenations over other
catalysts were performed under the same conditions except for the
pretreatment.
The reactivity of each substrate was roughly evaluated by the
initial hydrogenation rates (r, mol g−1 h−1) calculated from the hy-
drogen uptake around 25% conversion. After the hydrogen uptake
was complete, the reaction was continued for additional 2 h, and
2 M HCl (1 ml) was added to the solution. The mixture was filtered
with a small amount of ethyl acetate to remove the catalyst. It was
confirmed in every run that the product was completely dissolved
in the solution prior to filtration, and that no detectable product
remained in the filtered solid. The filtrate was extracted with ethyl
acetate (5 ml) and then washed with water (5 ml × 2). The extract
was analyzed by 1H NMR (600 Hz) in CDCl3 to confirm the com-
pletion of hydrogenation, and by HPLC with a chiral column (OJ-H,
AD, or QD, Daicel, 25 cm × 4.6 mm) to determine the enantiomer
ratio of the product. The enantiomer excess (ee) value was calcu-
lated from the ratio according to % ee = 100 × |S − R|/(S + R). The
analysis conditions and the retention time for each compound are
given in the supplementary material.
◦
containing dioxane (2.5%) at 23 C under the atmospheric pressure
of hydrogen. Since added amine is known to give better prod-
uct ee [17], benzylamine (BA) was added to the reaction mixture.
Fig. 2 shows the effect of the amount of BA for five selected sub-
strates, PCA, B-4, DMPCA, C-3, and E-4. The effects of added BA
reach close to the maximum for all the substrates around 0.6 mo-
lar ratio (equivalent) to the substrate both in the product ee and
in the hydrogenation rate, but additional BA induces some further
change in ee, depending on the substrate. To compare the results
with different substrates under the common reaction conditions,
hydrogenation was performed at 1.0 equivalent of BA. Except for
Group G, the substrate was consumed completely, and no side
products were detected by 1H NMR. The values shown under each
substrate in Fig. 1 are the ee value of the hydrogenation prod-
uct and the initial reaction rate (r, mol g−1 h−1) in the presence
of BA.
The hydrogenations were also performed in the absence of BA
for all the substrates. Fig. 3 illustrates the ee values obtained in
the reactions with BA vs. without BA for the substrates of Groups
A–E (see the supplementary material for the effects on the rate).
In the absence of BA, hydrogenations of some substrates were very
slow, which can be a reason for the low ee values due to the loss
in the catalyst performance during long reaction time. The slow
reaction must be due to the low solubility of the substrate acid in
The absolute stereochemistry was determined by the optical ro-
tations for stereochemically known compounds; the products from
PCA, B-4, E-1, E-2, E-4, F-1, F-2, and F-3 (Fig. 1). The other products
are assumed to have the same stereochemistry among the series of
the substrates, deduced from their optical rotations. The ee values

T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
59
(a)
Fig. 1. Substrate structures in Groups A–G and the values of the product ee and the initial rate (mmol g−1 h−1, determined by the H2 uptake) of their hydrogenation in the
presence of BA.
the reaction solvent. A representative substrate is D-4, which gives
41% ee at the slow initial rate of r = 8 in the absence of BA, while
the ee is high 83% in the presence of BA (r = 92). Except for these
poor substrates, the BA addition was found to increase the product
ee by 1.2–1.5 times (15–25%) for most of the substrates.
Under the present conditions, the reference substrate PCA re-
sulted in 81% ee in the presence of BA (65% ee in the absence of
BA) as shown in Fig. 1. As seen in the results for Group A, effects
of the methyl substitution are not large, but some improvement
was observed in the product ee. The substitution on the α-phenyl

60
T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
(b)
Fig. 1. (continued)
group (A-1 to A-3) does not affect much the stereoselectivity, but
only decrease in the hydrogenation rate was observed by the o-
substitution. In contrast, that on the β-phenyl group affects the
product ee depending on the position; p-methyl gives higher ee
of 86%, and o-methyl gives lower ee of 62%, while m-methyl does
not affect the ee (A-4 to A-6). Among the methyl-substituted sub-
strates studied so far, the p,p -dimethyl-substituted substrate A-7
gives the best result of 89% ee.
Methoxy substitution on the α-phenyl affects little the enan-
tioselectivity at any positions (B-1 to B-3), while the effects of the
substitution on the β-phenyl are different depending on the po-
sitions; p-methoxy-substitution results in a notable improvement
(B-4, 90% ee), while the substitution at the m-position does not
lead to a higher enantioselectivity (B-5). These substituent effects
are well consistent with the reported results with Pd/Al2O3, except
for B-3 [13]. The p-methoxy-substitution is still effective with ad-
ditional substitutions (B-6 to B-8 and B-12). Dimethylamino group
is a stronger donor than methoxy, but the substitution at the β-p-
position does not result in further improvement (B-9). DMPCA
(B-6) is, so far, the simplest analogue to give the best ee of 92%. In
ꢀ

T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
61
(a)
(b)
−1
Fig. 2. Effects of the amount of added BA for the hydrogenation of five representative substrates (a) on the product ee and (b) on the initial hydrogenation rate (mmol g−1 h
)
determined by the H2 uptake.
may not indicate an essential structural drawback but can be due
to the low solubility under the present reaction conditions. Naph-
thyl group should be adsorbed more strongly than phenyl group
on the palladium surface, but this nature does not bring about any
notable effects on the product ee (D-4 vs. B-2).
The hydrogenation of commercially available E-1 under the
present conditions gives 46% ee of the same stereochemistry as
the product of PCA. The product ee is lower than that of PCA, but
still higher than the reported values for E-1 using other palladium
catalysts [4,18,19]. Chain elongation increasing the lipophilicity of
the α-substituent as E-2 and E-3 results in a slight improvement in
the product ee. In contrast, substitution by the sterically bulky iso-
propyl group yields the product in higher stereoselectivity to give
80% ee in combination with β-phenyl (E-4) and 86% ee with β-p-
methoxyphenyl (E-5). Cyclohexyl group may be too bulky and the
reaction of E-6 gives lower 81% ee at a slower reaction rate. Hy-
drogenation of E-7 carrying a tBu group was so slow that it took
five days to complete the reaction.
In the reverse, the β-aryl group is indispensable to perform
sufficient stereocontrol under the present conditions. As shown in
Group F, both normal (F-1) and branched (F-2) β-alkyl substrates
react slowly and do not induce notable stereocontrol, as does a
conventional aliphatic example of tiglic acid (F-3). It should be
noted that the stereochemistries of the products from F-1 and F-2
determined by the optical rotations are reverse of those of PCA and
its analogues of Groups A–E, but are the same as those of aliphatic
substrates represented by tiglic acid [18,20].
When both α- and β-aryl groups in the substrate are fixed to
be coplanar with the olefinic bond as in G-1 and G-2, enantios-
electivity is lost (G-1) or the hydrogenation itself is interrupted
(G-2). Lack of the β-substituent results in poorer enantioselec-
tivity (G-3). Application of the present hydrogenation system to
N-acetyldehydrophenylalanine (G-4) was unsuccessful.
Fig. 3. Plots of the product ee for substrates of Groups A–E in the presence vs.
absence of BA.
Group C, trifluoromethyl (CF3) substitution at the α-phenyl does
not show any obvious effects (C-1, C-2). On the β-phenyl, p-CF3
(C-3) does not affect the product ee, but m-CF3 shows negative ef-
fects to give 56% ee (C-4). Notable improvement to give 86% ee
is observed with the β-p-fluoro group (C-5), which is electron-
withdrawing but also electron-donating in a conjugative sense.
Comparing with DMPCA, some loss in the product ee is ob-
served with a lipophilic analogue D-1 (79% ee), but the other two
analogues give similar ee to DMPCA irrespective of the alkoxy
structure (D-2: 91% ee and D-3: 92% ee). The lower ee with D-1
To disclose the role of the electron-donating group at the p-
position of the β-phenyl ring, properties of DMPCA during the
hydrogenation are compared with those of unsubstituted PCA. As

62
T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
Table 1
Structure dependency on the product ee with various catalysts.a
D (nm)c
Additive
PCA
DMPCA
% ee
b
Catalyst
% DM
d
d
% ee
rm and ru
(rm/ru
)
rm and ru
(rm/ru)
5% Pd/Al2O3
5% Pd/C (STD)
5% Pd/C (AER)
5% Pd/TiO2
19
76
67
52
76
5.8
1.5
1.7
2.1
1.5
BA
–
BA
–
BA
–
BA
–
39
35
46
43
59
52
71
59
81
65
9/74
15/120
40/96
12/250
78/110
39/360
33/58
7/110
105/240
55/340
(0.12)
(0.12)
(0.42)
(0.05)
(0.74)
(0.11)
(0.57)
(0.06)
(0.44)
(0.16)
58
55
65
56
72
66
82
70
92
69
4/21
5/74
24/5
8/150
32/62
11/240
12/–
11/–
43/60
12/120
(0.18)
(0.07)
(0.47)
(0.05)
(0.47)
(0.05)
5% Pd/C (STD)
(pretreated)
BA
–
(0.72)
(0.10)
a
◦
The hydrogenation was performed in a 2.5% water-containing dioxane on a 0.5 mmol scale with 23 mg of Pd catalyst and 6 mg (20 μmol) of cinchonidine at 23 C in the
presence or absence of benzylamine (BA, 1 equiv.).
b
Degree of palladium dispersion calculated from the metal surface.
c
Mean particle size of palladium estimated form the metal dispersion.
d
rm and ru (mmol g−1 h−1) are the initial reaction rates determined by the H2 uptake with CD-Pd/C and with unmodified Pd/C, respectively.
(a)
(b)
Fig. 4. Reaction conversion of PCA and DMPCA in the presence of ethanolamine over (a) CD-modified catalyst and (b) unmodified catalyst. The hydrogenation was carried out
with an individual substrate (shown by circles) and with a 1-to-1 mixture of the substrates (shown by squares).
shown in Table 1, DMPCA always gives better product ee than PCA
using palladium catalysts supported on different supports [21]. Im-
provement of the enantioselectivity by added BA was observed
for the both substrates, and with different catalysts. Effects of the
CD-modification on the hydrogenation rate are always negative, ir-
respective of the substrate, as seen in rm/ru values (<1), which
are obviously larger in the presence of BA. The rm/ru value is
also larger for good palladium catalysts, but detailed comparison
of the rm/ru values needs more accurate kinetic study. Comparing
the two substrates, DMPCA tends to have larger rm/ru values than
PCA in the presence of BA, while the differences are smaller in the
absence of BA. Under the best conditions (line 9 in Table 1), the
rm/ru values are 0.44 for PCA and 0.72 for DMPCA.
Precise kinetic studies with PCA and DMPCA were performed by
monitoring the reaction by 1H NMR. Ethanolamine (pKBH+ = 9.50
in aqueous solution) was employed in place of BA (pKBH+ = 9.34)
to avoid precipitation of the product salt during the reaction. The
substrate/catalyst ratio was doubled from the original procedure.
Fig. 4a illustrates conversion profiles of hydrogenation of PCA and
DMPCA over the CD-modified catalyst in individual reactions as
well as in the reactions with a 1-to-1 mixture of the two substrates
(one half in concentration of each substrate). Fig. 4b also illus-
trates similar profiles with the CD-unmodified catalyst. By mixing
the two substrates, the hydrogenation of DMPCA is suppressed by
the co-existing PCA, and becomes faster as the consumption of
PCA proceeds. This suppression phenomenon is larger with the un-
modified catalyst (Fig. 4b) than that with the CD-modified catalyst
(Fig. 4a). The extent of suppression is seen in Table 2, where the
reaction rates obtained from the data of initial regions (10–40%
conversion) are given. Since PCA is consumed in a short time with
a 1:1 mixture, the same experiment was also performed with a
mixture of PCA and DMPCA in 5:1 ratio to emphasize the sup-
pression effect by PCA (see the supplementary material for the
conversion profiles). As a matter of fact, the tendency observed
with the 1:1 mixture becomes obvious, and the hydrogenation of
DMPCA with the CD-modified catalyst becomes even faster than
that with the unmodified catalyst in the presence of a sufficient
amount of PCA (Table 2, entry 3, rm/ru > 1). The ee values at
the end of reaction are not affected by the change in the reac-
tion conditions or by mixing the two substrates. The rm/ru values
determined from the hydrogen consumption rate given in Table 2
are 0.56 for PCA and 0.81 for DMPCA.

T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
63
Table 2
The hydrogenation rate (mmol g−1 h−1, determined by 1H NMR) of PCA and DMPCA in the presence of ethanolamine.
Entry
PCA
DMPCA
CD-modified
Unmodified
r
m/ru
CD-modified
Unmodified
r
m/ru
1
2
3
Separate experiment
1:1 mixture
5:1 mixture
21 (82%)a
17 (81%)a
12
37
35
23
0.56
0.48
0.53
17 (92%)a
4.8 (92%)a
2.3
22
6.8
1.2
0.81
0.7
1.9
a
The product ee values determined by HPLC after the complete conversion.
4. Discussion
The enantioselectivities observed for the hydrogenation of 42
different α,β-unsaturated carboxylic acids are summarized as fol-
lows. The α-phenyl group is not electronically important but is
necessary only as a steric fence; the α-phenyl group must be
twisted from the olefinic plane of PCA. This non-planarity should
work to differentiate the enantioface of the olefin. This presump-
tion is supported by the high enantioselectivity with the substrates
having a properly bulky α-alkyl substituent (E-4 and 5) and the
low enantioselectivity with the coplanar α-aryl substrate (G-1).
In contrast, the β-aromatic group is indispensable for the stere-
oselectivity and its electronic effect is large. The β-aryl group can
directly interact with the palladium catalyst surface during the
hydrogenation of the olefinic part, and the interaction can be a rea-
son for the big difference in selectivities between the β-aryl and
β-alkyl substrates. However, the effect of different β-aryl groups
cannot be simply attributable to change in the direct interaction,
because the m-substituted substrate does not show the same ef-
fect as the p-substituted one. This presumption is confirmed by
the results with Group D, where the β-aryl substrates expected to
have different adsorption ability do not show different stereoselec-
tivity.
A key to obtain the high stereoselectivity is a conjugative
electron-donating substituent on the β-phenyl group of PCA, rep-
resented by DMPCA. The conjugative electron-donating group at
the p-position causes stronger through-conjugation with the car-
boxylic acid via the aromatic and olefinic π bonds, which makes
flatter conformation of the whole β-arene–olefin–carboxylic acid
conjugation system. In the reverse, β-o-methyl substituent inter-
feres with the flat conformation for steric reasons, and thus, the
low ee values are obtained with A-6 and B-10. The conjugation
with an electron-donating group also causes reduction of acidity
of the substrate. The lower acidity makes the carboxylate anion
more basic, and then ionic interaction between the N-protonated
CD of the chiral modifier and the carboxylate anion of the sub-
strate becomes stronger. The stronger CD–substrate interaction is
experimentally supported by the results of the competitive hy-
drogenation given in Table 2. That is, adsorption of DMPCA on
the catalyst is weaker than PCA, but the difference between the
substrates is smaller when the catalyst is modified by CD. Thus,
CD–DMPCA interaction must be stronger than CD–PCA interaction.
Fig. 5 illustrates four geometries of the carboxylate adsorbed by
the aid of the chiral CD on catalyst surface by Newman projec-
(a)
(b)
(c)
(d)
Fig. 5. Newman projections of the substrate looking from the chiral amine adsorbed
on the Pd surface to give (a) a major enantiomer and (b–d) a minor enantiomer,
the latter of which is caused by (b) misidentification of α-aryl as the olefin part,
(c) loss of the conformational planarity, or (d) lack of the stereo recognition of the
carboxyl group.
In general, chiral modification of a heterogeneous catalyst is
not uniform, but some unmodified sites remain active to pro-
duce racemic product. Thus, the ee value is not only governed by
the stereodifferentiating ability of the chiral modifier. A simplified
model to express the product ee with consideration of the catalyst
uniformity is given as % ee = factor-i × E/(E + N), where E is con-
tribution from enantioselective site (E-site) by the aid of the chiral
modifier and N is that from non-enantioselective site (N-site) per-
formed on remaining unmodified surface [22]. The contribution
from the E-site in the total reaction is expressed as E/(E + N). The
other factor is factor-i, which is the intrinsic stereocontrollability
of the chiral modifier on the catalyst and is equal to an ee value
produced specifically at the E-site. Here, the stronger modifier-
substrate interaction can induce both the higher E/(E + N) and
the larger factor-i.
Among the reactions using the same catalyst, relative magni-
tudes of the E-site contribution, E/(E + N), can be evaluated by
the kinetic parameters, rm and ru [23]. If the hydrogenation rate
over the unmodified catalyst (ru) represents the catalytic activity of
the unmodified site (N-site) on the CD-modified catalyst, changes
in a ratio of ru/rm parallel changes in the N/(E + N) value. The
most reliable rm and ru values to discuss the substrate structure
effect are those in entry 1 of Table 2. The ru/rm values of 1.79
for PCA and 1.23 for DMPCA indicate that contribution of the N-
site catalysis in DMPCA is smaller than PCA by 0.7 times. Since
the N-site catalysis in the reaction with DMPCA is in a range of
0–8% (= 100% − % ee), the N-site catalysis with PCA is calculated
to be in a range of 0–11%, which means the factor-i of PCA is less
than 92% (<0.82/0.89) [23]. As a result, the ee improvement by the
β-p-donor substitution on PCA must induce the larger factor-i by
the stronger substrate–modifier interaction.
tion through CO–C bond (views from CD), where the geometry
α
5a leads to the major enantiomer, while 5b–5d give the minor
enantiomer. Selection of 5a can be achieved when C cannot get
α
close to the Pd surface, but C can. 5b has just opposite ten-
β
dency to 5a. The minor enantiomer formation cannot effectively
be suppressed if the olefin and the carboxylate are not coplanar as
shown in Fig. 5c, or the primary chiral recognition at the carboxy-
late is looser as shown in Fig. 5d. This model is compatible with
the structural requirements for the substrate experimentally ob-
served to obtain the high enantioselectivity; coplanar arene group
at the β-position, bulkiness at the α-position, the flat conforma-
tion of the whole conjugation system, and the stronger conjugate
base (carboxylate).
5. Conclusion
We have studied the hydrogenation of a range of analogous
substrates originating from PCA, and found that α,β-unsaturated

64
T. Sugimura et al. / Journal of Catalysis 262 (2009) 57–64
acids having
a
properly bulky α-substituent and an electron-
[3] T. Mallat, A. Baiker, in: R.A. Sheldon, H. Bekkum (Eds.), Fine Chemicals through
Heterogeneous Catalysis, Wiley–VCH, Weinheim, 2001, p. 449.
[4] J.R.G. Perez, J. Malthête, J. Jacques, C. R. Acad. Sci. Catal. 300 (1985) 169.
[5] Y. Nitta, A. Shibata, Chem. Lett. (1998) 161.
[6] K. Borszeky, T. Bürgi, Z. Zhaohui, T. Tallat, A. Baiker, J. Catal. 187 (1999) 160.
[7] A. Tungler, Y. Nitta, K. Fodor, G. Farkas, T. Máthé, J. Mol. Catal. A Chem. 149
(1999) 135.
[8] W.-R. Huck, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 219 (2003) 41.
[9] Y. Nitta, K. Kobiro, Chem. Lett. (1996) 897.
[10] Y. Nitta, T. Kubota, Y. Okamoto, Bull. Chem. Soc. Jpn. 74 (2001) 2161.
[11] Y. Nitta, Y. Okamoto, Chem. Lett. (1998) 1115.
donating β-aryl group are suitable substrates for the enantioselec-
tive hydrogenation over the CD-modified palladium catalyst. This
substrate design is effective irrespective of the type of the palla-
dium catalyst and the reaction conditions. The electron-donating
β-aryl group with the p-alkoxy substitution realized improvement
of the product ee up to 92%. Stronger interaction between the
chiral modifier CD and such a substrate was experimentally sup-
ported.
[12] Y. Nitta, J. Watanabe, T. Okuyama, T. Sugimura, J. Catal. 236 (2005) 164.
[13] G. Szöllösi, B. Hermán, K. Felföldi, F. Fülöp, M. Bartók, J. Mol. Catal. A Chem. 290
(2008) 54.
Acknowledgments
The authors thank N.E. Chemcat Co. for donating the Pd/C cat-
alysts. The authors also thank Emeritus Professor Yuriko Nitta for
gift of Pd/TiO2 and helpful discussions.
[14] A preliminary communication dealing with three methoxy-substituted PCA
analogues, see: T. Sugimura, J. Watanabe, T. Okuyama, Y. Nitta, Tetrahedron:
Asymmetry 16 (2005) 1573.
[15] T. Sugimura, J. Watanabe, T. Uchida, Y. Nitta, T. Okuyama, Catal. Lett. 112 (2006)
27.
Supplementary material
[16] Y. Nitta, T. Kubota, Y. Okamoto, Bull. Chem. Soc. Jpn. 73 (2000) 2635.
[17] Y. Nitta, Chem. Lett. (1999) 635.
[18] R.E. Padgett Jr., R.L. Beamer, J. Pharm. Sci. 53 (1964) 689.
[19] G. Szöllösi, T. Hanaoka, S. Niwa, F. Mizukami, M. Bartók, J. Catal. 231 (2005)
480.
[20] K. Borszeky, T. Mallat, A. Baiker, Tetrahedron: Asymmetry 8 (1997) 3745.
[21] Y. Nitta, K. Kubota, Y. Okamoto, in: H.U. Blaser, A. Baiker, R. Prins (Eds.), Het-
erogeneous Catalysis and Fine Chemicals IV, Elsevier, Amsterdam, 1997, p. 191.
[22] T. Sugimura, S. Nakagawa, A. Tai, Bull. Chem. Soc. Jpn. 75 (2002) 355.
[23] For the relation between ru/rm and N/(E +N), see p. S20 in the supplementary
material.
The online version of this article contains additional supple-
mentary material.
Please visit DOI: 10.1016/j.jcat.2008.11.022.
References
[1] T. Mallat, E. Orglmeister, A. Baiker, Chem. Rev. 107 (2007) 4863.
[2] Y. Nitta, J. Synth. Org. Chem. 64 (2006) 827.