Adsorption and performance of chiral cinchona alkaloid modifiers over Pd/C catalyst for enantioselective hydrogenation of α-phenylcinnamic acids

Article abstract of DOI:10.1246/bcsj.20160162

The enantioselective hydrogenation reactions of α-phenylcinnamic acid (PCA) and 4,4′-dimethoxy α-phenylcinnamic acid (DMPCA) were carried out over chiral cinchona alkaloidmodified Pd/C. Two sets of the modifiers were employed to get deeper insights into the effects of relative adsorption strength between the modifier and the substrate on the enantioselectivity; cinchonidine (CD)/cinchonine (CN) and quinine (QN)/quinidine (QD). The performances of the two sets of modifiers were compared by systematically varying the modifier concentration over a wide range. It was clearly substantiated that the origin of the low selectivity observed with QN/QD at an ordinary concentration is primarily due to its weak adsorption strength on Pd metal surface, which originates from the steric hindrance of the methoxy substituent at C6. A new modifier, 6-hydroxy CD, was found to exhibit a performance comparable to that of CD, implying that the steric hindrance of the 6-methoxy group of QN/QD is much more influential than the electronic effects.

Full text of DOI:10.1246/bcsj.20160162

Advance Publication Cover Page
Adsorption and Performance of Chiral Cinchona Alkaloid Modifiers over Pd/C
Catalyst for Enantioselective Hydrogenation of -Phenylcinnamic Acids
Makoto Nakatsuji, Tomonori Misaki, Yasuaki Okamoto, and Takashi Sugimura*
Advance Publication on the web July 4, 2016 by J-STAGE
doi:10.1246/bcsj.20160162
© 2016 The Chemical Society of Japan

Adsorption and Performance of Chiral Cinchona Alkaloid Modifiers over Pd/C
Catalyst for Enantioselective Hydrogenation of α-Phenylcinnamic Acids
Makoto Nakatsuji, Tomonori Misaki, Yasuaki Okamoto, and Takashi Sugimura＊
Graduate School of Material Science, University of Hyogo, Kohto, Kamigori, Hyogo 678-1297, Japan
E-mail: <sugimura@sci.u-hyogo.ac.jp>
Abstract
The enantioselective hydrogenation reactions of
α-phenylcinnamic
acid
(PCA)
and
4,4’-dimethoxy
α-phenylcinnamic acid (DMPCA) were carried out over chiral
cinchona alkaloid-modified Pd/C. Two sets of the modifiers
were employed to get deeper insights into the effects of relative
adsorption strength between the modifier and the substrate on
the enantioselectivity; cinchonidine (CD)/cinchonine (CN) and
quinine (QN)/quinidine (QD). The performances of the two sets
of modifiers were compared by systematically varying the
modifier concentration over a wide range. It was clearly
substantiated that the origin of the low selectivity observed
with QN/QD at an ordinary concentration is primarily due to its
weak adsorption strength on Pd metal surface, which originates
from the steric hindrance of the methoxy substituent at C6. A
new modifier, 6-hydroxy CD, was found to exhibit a
performance comparable to that of CD, implying that the steric
hindrance of 6-methxy group of QN/QD is much more
influential than the electronic effects.
Scheme 1. Enantioselective hydrogenation over chiral cinchona
alkaloid-modified Pd/C
1. Introduction
Precious metals, especially Pt and Pd, can be
effective asymmetric hydrogenation catalysts when their
surfaces are modified with proper chiral organic molecules.1
The modifier molecules adsorbed onto the metal surface
generate enantio-differentiating ability as well as covering the
achiral catalytic active sites, thereby controlling the reaction
enantioselectivity. Cinchonidine (CD)-modified Pd is such an
example; numerous researchers have attempted to increase the
performance of this catalyst system, yielding 82% ee of the
(S)-product in the hydrogenation of 2,3-(E)-diphenylpropenoic
acid (α-phenylcinnamic acid, PCA) over Pd/C in the presence
of benzylamine (BnNH2) (Scheme 1).2 In highly selective
reactions, the modifier should suitably cover the catalyst
surface as much as possible while maintaining the ability of the
metal surface to adsorb the substrate. Accordingly, this
The low selectivity with QN/QD was ascribed to weak
adsorption of QN/QD due to an unfavorable adsorption mode: a
parallel adsorption mode of the 4-substituted quinoline unit of
CD/CN to the Pd surface (Figure 1a), in contrast to a tilted
adsorption mode of the 4,6-disubstututed quinoline unit of
QN/QD where the quinoline nitrogen faced the Pd surface
(Figure 1b).6-8 Parallel adsorption of the modifier over Pd metal
surface may be a critical factor in the enantio-differentiation
ability of the chiral modifier to catalyse asymmetric
hydrogenation reaction under balanced competitive adsorption,
but the details of this mechanism are still unknown. In the
situation could be achieved in
a dynamic adsorption
equilibrium among the modifier, the substrate (or
substrate-BnNH2 salt), and hydrogen. In our previous report,
we have shown that the difference between CD and cinchonine
(CN, 54% ee of the (R)-product) as a modifier is attributable to
different intrinsic stereocontrollability and catalytic activity of
the modified sites and not to the adsorption strength of the
modifiers.3 We have also shown that ligand acceleration is a
key to yield a high product ee;3,4 thus, the modifier adsorbed
b)
a)
onto the Pd surface serves as
a co-catalyst in the
enantio-differentiating hydrogenation catalysis.
Here we describe another series of cinchona
Figure 1. Adsorption model of CD on Pd/C with 76%
dispersion; CD on Pd55 in parallel (a) and tilted (b)
configurations. The lower part of the particle should be buried
below the active carbon. An approximate position of 6-X group
is shown as a red circle.
alkaloids: quinine (QN) and quinidine (QD), which are the
6-methoxy analogues of CD and CN, respectively. Despite its
structural similarity to CD/CN, QN/QD is known to be a poor
chiral modifier (ee < 20%, PCA hydrogenation over Pd/TiO2).5

present study, we systematically varied the cinchona alkaloid
modifier concentration over a wide range to obtain deeper
insights into the adsorption and performance of the modifier
(CD/CN and QN/QD) in the enantioselective hydrogenation of
the substrate (PCA and 4,4’-dimethoxy α-phenylcinnamic acid,
DMPCA) over Pd/C. It was clearly demonstrated that the origin
of the low selectivity observed with QN/QD is due to its weak
adsorption strength on Pd metal surface, which originates from
the steric hindrance of the 6-methoxy substituent. Figure 1
illustrates the Pd/C catalyst in our reaction system containing
small Pd metal particles with an average diameter of 1.4 nm.9
2. Experimental
All chemicals were purchased from commercial
sources and purified via distillation when necessary. 6-hydroxy
CD (6-OH-CD in Scheme 1) was prepared from QN via
de-methylation
(55%
yield).
The
enantioselective
hydrogenation reactions of PCA and DMPCA were carried out
over chiral modified Pd/C as follows.4,9 A stirring suspension of
5% Pd/C (43 mg as a 50% content, wet form, STD-type
supplied by N.E. Chemcat) and 5 ml of 2.5 % H2O-containing
dioxane was heated under 105 Pa of hydrogen at 353 K for 30
min. After cooling to 296 K, a solution of modifier (0.0015−12
mg) in the solvent (1 mL) was added. After 30 min, 0.5 mmol
of the substrate acid in the solvent (4 mL) and then
benzylamine (BnNH2, 55 μL) was added. The Pd/substrate
molar ratio was 10/500 (in μmol in 10 mL of polar wet
dioxane). The reaction temperature and pressure were 296 K
and 105 Pa of hydrogen, respectively. The catalytic activity was
calculated from the rate of hydrogen consumption at 25%
conversion. The hydrogen consumption continued for 1-3 h.
After additional 2 h, a 2 M HCl aqueous solution (1 mL) was
added to the solution, followed by filtration of the reaction
mixture to remove the catalyst. The filtrate was extracted with
ethyl acetate (2 mL) and washed with water (2 mL) .The
extract was analyzed by HPLC with a chiral column (Daicel
OJ-3 for PCA and AD-3 for DMPCA). The enantioselectivity is
expressed as the enantiomeric excess (ee) of the (S)-product,
%ee = 100 x ([S] – [R])/([S] + [R]),
Figure 2. Product ee (%) as a function of the modifier
concentration as expressed in logarithm for the hydrogenation
of PCA on Pd/C. The modifiers used were (○) CD, (●) CN,
(□) QN, and (■) QD. The results with CD and CN were
recalculated from reference 3 for comparison.
where [S] and [R] represent the concentrations of (S)- and
(R)-enantiomers, respectively.
3. Results and Discussion
The product ee values are shown in Figures 2 and 3 as a
function of the modifier concentration expressed in logarithm,
log [modifier], for the hydrogenation of PCA and DMPCA,
respectively. The ee values are low when the concentration (or
amount) of the modifier is insufficient and are saturated when
the concentration is higher than required.3 The modifier
concentration in our previous regular experiments was fixed at
2 mM,2-4 which is indicated by a dotted line in Figs.2 and 3. On
the basis of the metal dispersion (0.76)9 of Pd/C used and an
estimated Pd metal surface area occupied by the modifiers (ca.
10 Pd atoms per modifier), the amount of modifier (2 mM) is
greater than 50 times the requirement for a monolayer
formation. Although possible adsorption of the modifier on the
support renders exact evaluation of the adsorption amount of
the modifier on Pd metal surface difficult, the observed
Figure 3. Product ee (%) as a function of the modifier
concentration as expressed in logarithm for the hydrogenation
of DMPCA on Pd/C. The modifiers used were (○) CD, (●)
CN, (□) QN, and (■) QD.
reaches a plateau value (82 %ee of the (S)-product)
around 2 mM of CD (or log [CD] = -2.7), the optimum surface
coverage of CD on Pd metal surface where adsorbed PCA
effectively forms interaction intermediates with adsorbed CD in
a dynamic equilibrium and thus the hydrogenation of PCA on
unmodified sites is minimized. The ee value in the PCA
hydrogenation over CN-modified Pd/C obviously shows a
mirror image correlation of the CD-modified Pd/C, although
the maximum ee value (54%ee of the (R)-product) attained
around 2 mM of CN is lower than that on the CD-modified
catalyst. It is considered that both CD and CN are adsorbed on
Pd metal surface in a parallel geometry via aromatic π-bonding
of the quinoline moiety and thereby the adsorption energies of
the diastereomers are very similar. Actually, Meemken et al.
found by ATR-IR with Pd/TiO2 that π-bonded CD was formed
sigmoidal curves are well described by
a
simple
Langmuir–Hinshelwood reaction mechanism assuming
competitive adsorption of the modifier and substrate.4
As shown in Fig.2, the enantioselectivity in the
hydrogenation of PCA over CD-modified Pd/C becomes
apparent as low as around 3x10-3 mM of CD (log [CD] = -5.5).
The ee value increases with the amount of the modifier and

on Pd metal surface, in particular, in the coexistence of CD and
BA.8 In addition, Hahn et al. showed by DFT calculations with
Pt(111) that the adsorption energy of CN was very close to that
of CD.10 The mirror image correlation between CD- and
CN-modified Pd/C in Fig.2 prompts us to conclude that the
higher enantioselectivity of the CD-modified catalyst is not
ascribed to a higher surface coverage of CD than CN, but to a
higher enantio-differentiating ability of adsorbed CD. It is
suggested that the absolute configurations at C8 and C9 of
CD/CN primarily elucidate the enantiomer distribution.
Figure 2 shows the ee values as a function of the modifier
concentration for the PCA hydrogenation over QN- and
QD-modified Pd/C. With QN-modified Pd/C, enantioselectivity
becomes observable only at a concentration higher than 1 mM
(log[QN] = -3.0) and increases to 56%ee of the (S)-product as
the QN concentration increases to 30 mM (log[QN] = -1.6). A
plateau or maximum ee value was not attained because of
experimental limitations. With PCA hydrogenation over
modifier, the substrate, and hydrogen, these results lead us to
conclude again the similarity of the adsorption energies
between CD and CN and between QN and QD, with the former
set of the modifiers being more strongly adsorbed than the
latter one. However, there are noticeable differences between
PCA and DMPCA in the modifier concentrations where the
enantioselectivity becomes apparent and the maximum ee
values are attained; the enantioselectivity with DMPCA is
detected as low as 10-2-10-3 mM (log [modifier] = -6 ~ -5) of
CD/CN, 10 times lower concentration than PCA and the
maximum ee value is observed around 0.5 mM (log [CD/CN] =
-3.3) in contrast to 2 mM with PCA. The maximum ee value
with DMPCA is attained around 10 mM (log [modifier] = -2) of
QN/QD, while at > 30 mM with PCA. These findings are
interpreted by assuming weaker adsorption strength of DMPCA
on Pd/C than that of PCA due to steric hindrance of the
methoxy groups. It is considered that the weakened adsorption
strength of DMPCA induces more preferential interactions with
adsorbed chiral cinchona modifiers compared with Pd metal
surface. It is noted that the electron releasing effects from the
methoxy groups of DMPCA enhance the hydrogen bond
interactions between the modifier and the substrate,5 thereby
resulting in the higher enantioselectivity of DMPCA compared
to PCA, as suggested by Szöllösi.11
QD-modified Pd/C,
a mirror image correlation of the
QN-modified Pd/C is observed, although the ee value is very
limited (17 %ee of the (R)-product). At the regularly employed
concentration (2 mM), QN and QD resulted in very low ee
values (15% and 5%, respectively, with Pd/C), in agreement
with the literature.2 It is suggested from Fig.2 that these ee
values on QN/QD-modified Pd/C could be improved further if
higher modifier concentrations were achieved. The mirror
image correlation with QN/QD-modified Pd/C in Fig.2 shows
that the adsorption strength of the diastereomers on Pd metal
surface are very close each other and that the different ee
values attained are attributable to slightly different interaction
modes between the modifier and the substrate. It is considered
that the diastereomers, QN and QD, are adsorbed on Pd surface
in a tilted configuration via the N-lone pair of the quinoline
moiety with very close adsorption energies.
Comparing the ee value-modifier concentration mirror
image correlations in Fig.2 for the PCA hydrogenation over
CD/CN- and QN/QD-modified Pd/C, it is evident that the latter
set of modifiers generates enantio-differentiating ability and
reaches the plateau ee value at much higher modifier
concentrations than the former set of modifiers. It is considered
that this difference is resulted from the different adsorption
strength of the modifier between the two sets of the modifiers.
CD/CN is more strongly adsorbed on Pd surface with a parallel
geometry via π-electrons of the quinoline moiety than QN/QD
with a tilted configuration via the N-lone pair of the quinoline
moiety due to steric hindrance of 6-methoxy group. In a
dynamic adsorption equilibrium among the modifier, the
substrate, and hydrogen during the reaction, stronger adsorption
of the modifier results in a higher surface coverage of adsorbed
modifier to generate a higher enantio-differentiating ability
through more abundant interactions with the substrate, PCA, as
well as resulting in more extensive suppression of the
hydrogenation on unmodified sites or hydrogenation without
interactions with the modifier. Another difference among the
two sets of the modifiers is the maximum ee values attained. It
is considered that the parallel adsorption of CD/CN via
π-bonding of the quinoline moiety provides more favorable
configurations for enantio-differentiation in PCA-modifier
interaction intermediates than the tilted configuration of
QN/QD.
Figure 4. Product ee (%) as a function of the log of the modifier
concentration for the hydrogenation of PCA on Pd/C. The
modifiers used were (x) 6-hydorxy CD, (○) CD, and (□) QN.
The results with CD were recalculated from reference 3 for
comparison.
Another significant difference between PCA (Fig.2) and
DMPCA (Fig.3) resides in the maximum ee values achieved;
the maximum ee values of DMPCA are considerably higher
than those of PCA with all the modifiers examined here. It is
noted, in particular, that the ever highest ee values of 82% ee
and 42% ee are achieved for QN and QD, respectively, with the
enantioselective
QN/QD-modified Pd metal catalysts. On the basis of these
results, we presume that QN exhibits high intrinsic
hydrogenation
of
DMPCA
over
Figure 3 presents the dependences of the ee values on the
modifier concentration in the enantioselective hydrogenation of
DMPCA over the two sets of CD/CN- and QN/QD-modified
Pd/C. As shown in Fig.2 for the hydrogenation of PCA, two
sets of mirror image correlations are observed. Assuming a
dynamic equilibrium in competitive adsorption among the
a
stereo-controllability close to that of CD; however, the low
adsorption constant of QN is a decisive drawback. The
adsorption constant of QN/QD in common logarithm (log K)
should be lower approximately by 1.2–1.5 than that of CD/CN,

assuming relatively weak adsorption of the modifiers (i.e.,
QN/QD should exhibit a ca. 1/20–1/30 times smaller adsorption
constant).
logarithm (log K) is lower by 1.2–1.5 than that of CD/CN. The
new modifier 6-OH-CD was found, for the first time, to exhibit
a performance comparable to that of CD, indicating that the
steric hindrance of 6-methxy group of QN/QD is much more
influential than the electronic effects. The design of an aromatic
anchor is possible to control the modifier adsorption over Pd
catalysts.
During a further study on how the 6-methoxy group
of QN/QD hinders the modifier from parallel adsorption via the
quinoline ring, we found in the present study a new suitable
modifier, 6-hydroxy CD (6-OH-CD in Scheme 1), which was
prepared from QN via de-methylation. The performance (ee) in
the enantioselective hydrogenation of PCA over 6-OH-CD is
compared with CD and QN (or 6-CH3O-CD) in Figure 4. The
initial reaction rates relative to that over unmodified Pd/C are
also compared in Figure 5 with CD- and CN- modified Pd/C. It
is apparent that the properties of 6-OH-CD are very similar to
or even almost the same as CD with respect to both the
References
1.
a) D. Y. Murzin, P. Maki-Arela, E. Tukoniitty, T. Salmi,
Catal. Rev. 2005, 47, 175-256. b) E. Klabunovskii, G. V.
Smith, A. Zsigmond, in Heterogeneous Enantioselective
Hydrogenation, Springer, Dordrecht, 2006, Chapters 4
and 5. c) M. Heitbaum, F. Glorius, I. Escher, Angew.
Chem. Int. Ed. 2006, 45, 4732-4762. d) T. Mallat, E.
Orglmeister, A. Baiker, Chem. Rev. 2007, 107, 4863-4890.
e) T. Sugimura, in Handbook of Asymmetric
Heterogeneous Catalysis, ed. by K. Ding, Y. Uozumi,
Wiley-VCH, Verlagsgesellschaft, 2008. pp. 357-382. f) F.
Zaera, Acc. Chem. Res., 2009, 42, 1152-1160.
2.
a) Y. Nitta, J. Watanabe, T. Okuyama, T. Sugimura, J.
Catal. 2005, 236, 164-167. b) T. Sugimura, T. Uchida, J.
Watanabe, T. Kubota, Y. Okamoto, T. Misaki, T.
Okuyama, J. Catal. 2009, 262, 57-64.
3.
4.
T. Sugimura, H. Ogawa, Chem. Lett. 2010, 39, 232-234.
H. Ogawa, T. Mameda, T. Misaki, Y. Okamoto, T.
Sugimura, Chem. Lett., 2013, 42, 813-815.
5.
6.
Y. Nitta, A. Shibata, Chem. Lett. 1998, 161-162.
W. R. Huck, T. Mallat, A. Baiker, Catal. Lett. 2003, 87,
241-247.
7.
8.
9.
G. Szöllösi, K. Balásik, M. Bartók, Appl. Catal. A: Gen.,
2007, 319, 193-201.
F. Meemken, N. Maeda, K. Hungerbühler, A. Baiker, ACS
Catal. 2012, 2, 464-467.
T. Y. Kim, T. Uchida, H. Ogawa, Y. Nitta, T. Okuyama, T.
Sugimura, S. Hirayama, T. Honma, M. Sugiura, T.
Kubota, Y. Okamoto, Top. Catal. 2010, 53, 116-122.
K. P. Hahn, A. P. Seitsonen, A. Baiker, Phys. Chem.
Chem. Phys., 2015, 17, 27615-27629.
Figure 5. Initial hydrogenation rate (mmol/g·h) as a function of
the log of the modifier concentration for the hydrogenation of
PCA over Pd/C. The modifiers used were (x) 6-hydorxy CD,
(○) CD, and (●) CN. The reaction rates are normalized with
respect to the rates over the unmodified Pd/C. The results with
CD and CN were recalculated from reference 3 for comparison.
10.
11. Szöllösi, G., Catal. Lett., 2012, 142, 345-351
enantioselectivity and ligand acceleration effects.4 Thus, it is
revealed that the hydroxy-substitution at C6 of CD does not
affect the properties of CD as a modifier. Since 6-CH3O and
6-OH substituents have similarity in the electronic effects, a
slight steric difference between OH- and CH3O-groups results
in the profound difference in the performance, presumably, due
to the adsorption configuration of the modifier; 6-CH3O
substituent prefers the tilted adsorption (Figure 1b), while
6-OH substituent still favors the parallel adsorption (Figure 1a).
It is also deduced that the electronic effects of 6-CH3O
substituent do not affect the interaction mode and strength with
the substrate (PCA).
4. Conclusion
In the present study, the performances of the two sets of
modifiers, CD/CN and QN/QD, in the enantioselective
hydrogenation reactions of PCA and DMPCA over Pd/C were
compared by systematically varying the modifier concentration
over a wide range. It was clearly demonstrated that the origin
of the low selectivity observed with QN/QD is due to its weak
adsorption strength on Pd metal surface, which originates from
the steric hindrance of the methoxy substituent at C6. It is
estimated that the adsorption constant of QN/QD in common

Graphical Abstract
<Title>
Adsorption and Performance of Chiral Cinchona Alkaloid Modifiers over Pd/C Catalyst for Enantioselective Hydrogenation
of α-Phenylcinnamic Acids
<Authors' names>
Makoto Nakatsuji, Tomonori Misaki, Yasuaki Okamoto, and Takashi Sugimura＊
<Summary>
The enantioselective hydrogenations of PCA and DMPCA were carried out over chiral cinchona alkaloid-modified Pd/C by varying
the modifier concentration over a wide range. The low selectivity with QN/QD relative to CD/CN is primarily attributable to its weak
adsorption strength on Pd metal surface. A new modifier, 6-hydroxy CD, was found to exhibit a performance comparable to that of
CD.
<Diagram>
Low ee
High ee
CD/QN‐modified Pd/C