Over 98% optical yield achieved by a heterogeneous catalysis. Substrate design and analysis of enantio-differentiating factors of tartaric acid-modified Raney nickel hydrogenation

Article abstract of DOI:10.1246/bcsj.75.355

Tartaric acid-modified Raney nickel (TA-MRNi) is a chiral heterogeneous catalyst for the hydrogenation of prochiral ketones. An optical yield (OY) of 86% with methyl acetoacetate (1) as a substrate was improved to 94-96% by employing β-keto esters having a proper bulkiness at the γ-position. The γ-bulkiness effect contributes to a high intrinsic enantio-differentiating ability (factor-i) of the TA-MRNi catalysis. Through the study, we found the best substrate, γ-cyclopropyl-β-keto ester, the hydrogenation of which resulted in 98.6% OY. This further improvement in the OY was ascribed to a smaller contribution of non-enantio-differentiating hydrogenation (N-site catalysis) due to the substrate-specific activation of the enantio-differentiating hydrogenation by the chiral modifier. The OY of the hydrogenation of 1 was analyzed by comparing with well-behaved β-keto esters, and the contribution of the factor-i and the N-site to the OY value was evaluated to deduce the origin of the enantiodifferentiation.

Full text of DOI:10.1246/bcsj.75.355

c. Jpn.
© 2002 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 75, 355–363 (2002)
355
e Chemical
ciety of Ja-
n
e Chemical
ciety of Ja-
n
Over 98% Optical Yield Achieved by a Heterogeneous Catalysis.
Substrate Design and Analysis of Enantio-Differentiating Factors of
Tartaric Acid-Modified Raney Nickel Hydrogenation
Tartaric Acid-Modified Nickel Hydrogenation
T. Sugimura et al.
02
5
Takashi Sugimura,* Satoshi Nakagawa, and Akira Tai
3
Faculty of Science, Himeji Institute of Technology, Kohto, Kamigori, Ako-gun 678-1297
(Received September 17, 2001)
SJA8
09-2673
330
Tartaric acid-modified Raney nickel (TA-MRNi) is a chiral heterogeneous catalyst for the hydrogenation of
prochiral ketones. An optical yield (OY) of 86% with methyl acetoacetate (1) as a substrate was improved to 94–96% by
employing β-keto esters having a proper bulkiness at the γ-position. The γ-bulkiness effect contributes to a high intrinsic
enantio-differentiating ability (factor-i) of the TA-MRNi catalysis. Through the study, we found the best substrate, γ-cy-
clopropyl-β-keto ester, the hydrogenation of which resulted in 98.6% OY. This further improvement in the OY was as-
cribed to a smaller contribution of non-enantio-differentiating hydrogenation (N-site catalysis) due to the substrate-spe-
cific activation of the enantio-differentiating hydrogenation by the chiral modifier. The OY of the hydrogenation of 1
was analyzed by comparing with well-behaved β-keto esters, and the contribution of the factor-i and the N-site to the OY
value was evaluated to deduce the origin of the enantiodifferentiation.
01
01
0
3.6
An enantio-differentiating (asymmetric) reaction using a
heterogeneous catalyst (solid catalyst) was first disclosed in
1956 using silk-supported palladium, which catalyzes the hy-
drogenation of a CwN bond to result in chiral amino acids in
6–35% optical yield (OY).1 Many efforts were devoted there-
after to obtain higher OY using various heterogeneous cataly-
ses, and have resulted in only two reaction systems to give one
enantiomer produced over 10-times faster than the other
(> 82% OY). The Orito system, first reported in 1979, used
cinchona-modified palladium for the hydrogenation of α-keto
esters to result in 80–89% OY.2 By modifying this system, two
groups have achieved high OYs during the last decade,3 where
the highest OY of 97.6% indicates an 82-times faster produc-
tion of one enantiomer than the other. The success of this sys-
tem is mainly due to several hundred-fold activation of the ca-
talysis by the chiral modifier. The other system developed by
Izumi and co-workers uses tartaric acid-modified Raney nickel
(TA-MRNi) for the enantio-differentiating hydrogenation of
prochiral ketones,4 where they employed methyl acetoacetate
(1) as a simple β-keto ester substrate for the survey of a proper
chiral modifier for Raney nickel catalysts, and found the best
modifier, tartaric acid (TA).5 By using a combination of nickel
(base catalyst), TA (chiral modifier) and 1 (substrate), the ca-
talysis system has been investigated in detail by many research
groups to improve the OY;6,7 the OY had reached 80–91%, and
86% OY using our catalyst.8
tion from the chiral source (non-enantio-differentiating region
= N-site). Thus, the overall OY can be expressed as
%OY = 100×(factor-i)×E/(E + N)
(1)
where E and N are relative contributions of the E- and N-site
catalyses, respectively.
N-site catalysis is an inherent problem in enantio-differenti-
ating heterogeneous catalyses due to a difficulty in making a
uniform chiral catalyst surface. Although the origin of the
lowering OY was not known, most of the studies on the TA-
MRNi system had been intended to remove the N-site so as to
increase the value of E/(E + N). However, through studies
with various keto esters we have come to believe that factor-i
should be considered as being the reason for the low OY with
1, and have proposed an extended-stereochemical model,
which expresses the mechanism controlling the factor-i.9 This
model was suggested to us as a working hypothesis to improve
the OY of the TA-MRNi system. That is, the OY with 1 would
be increased with increasing factor-i when the γ-position is
bulky. Based on this substrate design, we have reported 96%
OY with methyl 4-methyl-3-oxopentanoate (3), the highest OY
of a heterogeneous catalysis at that time.9
The present report describes our efforts to improve the OY
of the TA-MRNi system by designing the substrate and by tun-
ing the reaction conditions based on the extended stereochemi-
cal model. The study resulted in a group of substrates, includ-
ing 3, giving 94–96% OY, and methyl 3-cyclopropyl-3-oxo-
propanoate (4) and its analogues, resulting in over 98% OY,
the first example of over a 100-times faster production of one
enantiomer against the antipode by a heterogeneous cataly-
sis.10 With this almost perfect enantio-differentiation as a clue,
the OY control factors, factor-i and E/(E + N), were separated,
The OY of the TA-MRNi system as well as other heteroge-
neous catalyses can be expressed in terms of the intrinsic enan-
tio-differentiating ability of the catalyst (factor-i) and the het-
erogeneity of the catalyst surface. The chiral catalyst has an
enantio-differentiating region (E-site) to give an optically ac-
tive product in some enantiomeric excess, that is factor-i, while
the other region produces a racemic product without perturba-

356 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Tartaric Acid-Modified Nickel Hydrogenation
and then a black box of the TA-MRNi system has been dis-
closed (Scheme 1).
β-keto ester does not depend on the temperature,12 has been
corrected (except for 1 itself). The OY was affected even by a
change in the partial structure of the substrate remote from the
reaction site, and thus the enantioface differentiation of the β-
keto esters by TA-MRNi does not simply depend on the struc-
ture around the reaction site, but on the whole substrate struc-
ture. A similar temperature dependency among entries 2–7 in-
dicates that their enantio-differentiating modes are essentially
the same, and some other factors, such as the unfavorable in-
teraction between the hydrophobic alkyl chain and the hydro-
philic nickel surface, slightly affect the adsorption mode of the
substrate and diminish the enantio-differentiating ability of
TA-MRNi.
Scheme 1.
Results and Discussion
As expected from the extended stereochemical model and
the results of the former report (entry 8),9 the bulkiness at the
γ-position induced both a larger temperature dependency and a
higher OY at 60 °C. Substrates having a branch at the γ-posi-
tion (entries 8–11) resulted in a moderately high OY at 100 °C,
but the OY increased and reached a higher level of 94–96% at
60 °C. When the δ-position was branched (entry 12), the OY
at 60 °C became lower, 93%, but again a 96% OY was ob-
tained when the δ-substituent was sufficiently bulky as a t-bu-
tyl group (entry 13).
Although the substrate design using the γ-steric effect effi-
ciently improved the OY up to 96%, this could not be extended
to a further improvement of the OY. When the γ-position was
too bulky like a t-butyl group (entry 15), the hydrogenation did
not proceed at all at 100 °C or even at 120 °C after 2 days.
Among the γ-mono-branched substrates, the (1-ethyl)propyl
analogue is also an unsuitable substrate resulting in a slow and
sluggish hydrogenation (20% conversion at 60 °C after 4 days,
entry 14). Although proper bulkiness at the γ-position in a
Taft’s steric constant (−Es*) is in the range of 0.06–1.9, there
is a notable non-linearity between the OY and the −Es* value.
A part of this inconsistency may be due to the hydrophobicity
of the substituent group, e.g. isopropyl vs cyclohexyl, but a
main part should be attributed to the speciality of molecular
recognition on the surface; the adsorption and reactions on the
surface are strictly regulated by geometrical demands.
γ
Steric and Hydrophobic Effects at the -Position of the
Substrate. TA-MRNi prepared from the ultrasonicated W-1
type Raney nickel has a minimum amount of the N-region as
well as high hydrogenation activity,8 which made it possible
for us to use a variety of β-keto esters, not only at 100 °C of
the commonly used temperature, but also at 60 °C. The hydro-
genation of the β-keto esters (R-COCH2COOMe) listed in Ta-
ble 1 at 60 and 100 °C gave essentially pure β-hydroxy esters
only by filtration and concentration of the reaction mixture
(except for entries 14 and 15). The OY values were calculated
from the ratios of the enantiomers determined by GLC analy-
sis, and are summarized in Table 1. Improved Taft’s steric
constants, −Es* for the R groups11 are also given.
Among the substrates having a linear alkyl chain at the γ-
position (entries 2–7), 2, having an ethyl group, gave the high-
est OY of 94% at 60 °C. The OY gradually diminished accord-
ing to the extension of the alkyl chain. Different from 1, they
showed a temperature dependency of ca. 3% increment in the
OY from 100 to 60 °C. The results indicate that 1, a common-
ly employed standard substrate, is not the best substrate for the
TA-MRNi system; regarding the temperature dependency, 1 is
an exception among the β-keto esters. From these findings, the
common expertise in the TA-MRNi system, that the OY with a
Table 1. Optical Yields (%OY) of Hydrogenation of Methyl
β-Keto Esters (R–COCH2COOMe) over TA-MRNi at 100
and 60 °C, and Improved Taft’s Steric Constants (−Es*)
A study of this series produced a group of the substrates, re-
sulting in an enantiomer production ratio of 32–49. The γ-ster-
ic effect should affect factor-i, but not E/(E + N), based on the
extended stereochemical model, which was supported by a ki-
netic study, which is discussed later.
Entry
R
−Es* a)
100 °C
60 °C
1
2
3
4
5
Me (1)
Et (2)
n-C3H7
n-C4H9
n-C6H13
n-C7H15
0
86b)
91b)
90
86b)
94b)
93
0.08
0.36
0.39
—
Further Increment of the OY toward over 98%. Since
most chiral pharmaceuticals require starting materials having
over 98% optical purity, an enantio-differentiating reaction be-
comes profitable when it can result in over 98% OY as well as
over 98% chemical yield. Hence, an enantiomer production
ratio of > 100 is a goal for studies for improving the OY of the
TA-MRNi system. Since a system resulting in 96% OY was
already in hand, > 98% OY can be achieved by reducing the
antipode production (or racemic production) by less than half.
During a further survey of a suitable substrate for TA-MRNi
system, we obtained methyl 3-cyclopropyl-3-oxopropanoate
(4). The cyclopropyl group in 4 is similar to the isopropyl
group in 3 regarding the steric bulkiness and hydrophobicity.
It was also expected that the cyclopropyl group survives the
hydrogenation conditions with TA-MRNi, while the resulting
88
91
87b)
87
90b)
89
6
—
7
8
9
10
11
12
13
14
15
n-C9H19
—
83
86
Iso-C3H7 (3)
Cyclo-C4H7
Cyclo-C5H9
Cyclo-C6H11
Iso-C4H9
Neo-C5H11
Et2CH
0.36
0.06
0.51
0.71
0.93
1.84
1.98
4.22
88b)
—
96b)
94
90
88
—
84
—
—
95
94
93
96
—
c)
d)
t-C4H9
—
a) Ref. 11. b) Ref. 9. c) 20% Conversion after 4 days (82%
OY). d) No reaction after 2 days.

T. Sugimura et al.
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 357
Table 2. Enantio- and Diastereodifferentiation of 3-Cyclo-
propyl-3-oxopropanoate and Its Analogues at 60 °C
product can be converted to more valuable optically active
compounds through cleavage of the cyclopropane ring.
The hydrogenation of 4 proceeded smoothly up to 100%
conversion, and none of the side reactions, like isomerization
or hydrogenation of the cyclopropane unit, were detected.
When the reaction was carried out at 100 °C, the minor enanti-
omer produced was only 2%, and thus the OY was as high as
96%. The high OY at 100 °C indicates the speciality of 4
among the γ-branched substrates. At 60 °C, the produced anti-
pode was diminished to be less than 1%. Judging from seven
repeated runs and the GLC integration errors, the amount of
the antipode was determined to be in a range of 0.6–0.8%. The
calculated OY of 98.6 0.2% is concluded to be > 98% (or
99%). The absolute structure of the product was determined
by a chemical correlation with (3S)-methyl 3-hydroxy-4-meth-
ylpentanoate by a PtO2-catalyzed hydrogenation of the prod-
uct.
Entry
1
Substrate
%OY
> 98
2
3
4
> 98a)
> 98b)
> 98
5
91
94c)
6
Here, we achieved the TA-MRNi system producing the opti-
cal active product needless of purification or optically enrich-
ment as a chiral synthon. The further increment of the OY in 4
from 3 is not reasonably understood based on the γ-steric effect
or the hydrophobicity. Based on a kinetic study, the difference
is attributed to a smaller contribution of the N-site catalysis
producing an unfavorable racemic product, but not to an in-
crease of the factor-i.
7d)
8d)
9d)
> 98 and 97
50 and 50e)
94 and 39
β
Hydrogenation of a Family of Cyclopropyl -Keto Es-
ters. The high performance of the TA-MRNi system with 4
at 60 °C (Table 2, entry 1) was continued at a lower reaction
temperature of 40 °C, where completion of the reaction took a
longer time of 78 h, but again the product of > 98% OY was
obtained in a quantitative yield (entry 2). The use of TA-
MRNi prepared with (S,S)-tartaric acid instead of the (R,R)-
modifier resulted in the antipode in the same OY (entry 3).
The use of ethyl ester instead of methyl of 4 also resulted in
over 98% OY (entry 4).
10d)
71 and 62f)
a) Hydrogenation temperature = 40 °C.
b) (S,S)-TA-MRNi as catalyst.
Although the introduction of a cyclopropyl group into the
substrate produced a fruitful result, the existence of other func-
tional groups as well as a proper steric bulkiness is still a pref-
erential factor in controlling the OY. The substrates of entries
5 and 6 are analogues of those of entries 12 and 15 in Table 1,
respectively. The moderate OY in entry 5 (91%) suggests that
a cyclopropyl group should be placed at the vicinal position of
the ketone to receive an extra benefit. The (1-methyl)cyclopro-
pyl group in entry 6 did not completely interrupt the hydroge-
nation, though its steric bulkiness is similar to that of the t-bu-
tyl group, and thus the cyclopropane at the γ-position was sug-
gested to have a function to promote hydrogenation of the ke-
tone at the β-position.
Chiral derivatives of 4 having substituent(s) at the 2-position
of the cyclopropyl unit, shown in entries 7–10, were employed
as racemic mixtures. The hydrogenation of each enantiomer of
them gives diastereomeric alcohols, and thus, stereo-differenti-
ations of those substrates in TA-MRNi catalysis can be evalu-
ated by the ratios of the diastereomers, which were converted
to the OY as an expedient (Table 2, entries 7–10). The methyl
substitution in the trans form slightly affected the OY, the de-
gree of which depended on the stereochemistry of the substrate
(entry 7). When the substitution was gem-dimethyl, the hydro-
genation became very sluggish (entry 8), which again indicates
c) 93% Conversion after 69 h.
d) Employed substrates are racemic mixtures. The values
of the OY for each enantiomer of the substrate were calcu-
lated as %de.
e) 29% Conversion after 30 h.
f) 44% Conversion after 67 h.
the limit of the proper bulkiness of the substrate. The ester or
phenyl substitution reduced the OY drastically, but differently,
between the isomers. The moderately high OY of 94% in the
hydrogenation of one of the ester-substituted enantiomers (en-
try 9) indicates that an extra ester group demands a certain
geometrical placement to reduce the stereo-differentiating
ability of TA-MRNi.
Kinetic Study to Evaluate the Contribution of the N-Site
Catalysis From the present survey of suitable substrates for
the TA-MRNi catalysis, we found two groups of β-keto esters,
one giving 94–96% OY and the other accessible to 98% OY. If
the extended stereochemical model explaining the γ-steric ef-
fect is appropriate, the OY differences among 1–3 are due to
factor-i, and the difference between 3 and 4 should be attribut-
able to the other remaining factor, E/(E + N). The contribu-
tion of the N-site catalysis in the total hydrogenation of the
TA-MRNi system is difficult to determine by physicochemical

358 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Tartaric Acid-Modified Nickel Hydrogenation
Table 3. Conversion (%) of β-Keto Esters 1–4 at 4 h over MRNi Prepared by Varied Modifiers and
Calculated Relative Reactivities
Entry
Methoda)
Catalyst modifier b)
1
2
3
4
Relative reactivity of 1:2:3:4
1
2
3
4
5
6
7
8
9
A
A
B
Bꢀ
Bꢀꢀ
B
B
B
B
B
TA-NaBr
NaBr
TA-NaBr
TA-NaBr
TA-NaBr
TA
MA-NaBr
NaBr
SA-NaBr
none
71
47
73
63
67
76
18
48
19
65
45
23
48
—
—
53
8
23
8
36
13
6
10
7
12
16
2
7
2
13
45
9
1:0.63:0.19:0.63
1:0.49:0.13:0.19
1:0.65:0.14:0.57
1: — :0.12: —
1: — :0.18: —
1:0.70:0.21:0.49
1:0.45:0.11:0.20
1:0.49:0.14:0.15
1:0.39:0.11:0.10
1:0.55:0.20:0.20
42
—
—
37
4
7
2
13
10
a) method A: The substrates (17 mmol/1 g of catalyst) were hydrogenated, separately. method B: A
mixture of the four substrates (4.25 mmol each/1 g of catalyst) were hydrogenated. method Bꢀ: A
mixture of the two substrates (8.5 mmol each/ 1 g of catalyst). method Bꢀꢀ: A mixture of the two
substrates (4.25 mmol of 1 and 12.75 mmol of 3/1 g of catalyst).
b) TA: tartaric acid, MA: malic acid, SA: succinic acid.
methods, but we established a reasonable system to estimate
per catalyst (17 mmol/g) constant. As shown in Table 3, both
the chiral and achiral catalyses showed conversions indepen-
dent of the existence of the other substrates (entries 1 vs 3 and
2 vs 8). The similar results between methods A and B indicate
that the adsorption site of 1–4 is common, and evenly co-occu-
pied in both the chiral and achiral catalyses. Thus, the sub-
strate-specific activation for 4 can be concluded to be a faster
reaction of the adsorbed 4 on the TA-MRNi surface. Other
chiral and achiral catalyses using method B and its variations
(methods Bꢀ and Bꢀꢀ) show a similar tendency, and support the
above conclusion. The malic acid-modified catalyst has a
smaller enantio-differentiating ability than TA-MRNi (60%
OY with 1). Interrelated to the lower enantio-differentiating
ability, the degree of the substrate-specific activation of 4 by a
malic acid-modification is smaller than that with TA-MRNi
(entry 7, see the difference between 3 and 4).
More detailed kinetic experiments were examined using an
autoclave fitted with an outlet to remove the sample during hy-
drogenation under high pressure. A mixture of four substrates,
1–4, was hydrogenated over the chiral TA-MRNi catalyst (TA
and NaBr-modified nickel) and the achiral counter catalyst
(NaBr-modified nickel) at 50, 60, 70, 80, and 100 °C; the con-
versions were determined by GLC at one-hour intervals. The
conversions at 60 °C as a function of the reaction time are
shown in Figs. 1 and 2. The initial reaction rates as a specific
rate of the reaction (rm, mmol g−1 h−1) were obtained by curve
fitting using data below the 20% conversion (Table 4). Since
some of the conversions at 80 and 100 °C were already over
20%, even in the initial 1 h, accurate values of rm could not be
obtained.
the value of E/(E + N). The hydrogenation rates over TA-
MRNi with β-keto esters are considered to be approximately
equal to the rate over the E-site catalysis, since the lowest OY
is as high as 86%, and thus most of the hydrogenation should
proceed through the E-site catalysis. As a representative mod-
el for the N-site catalysis, we employed achiral catalysts pre-
pared without any chiral modifiers. The achiral catalysis pro-
ceeds with no perturbation from the chiral modifier (tartaric
acid) and, thus, the hydrogenation rate over the achiral catalyst
can mimic the N-site catalysis rate in the TA-MRNi system.
The four representative substrates 1–4 (17 mmol/1 g cata-
lyst, method A) were hydrogenated for 4 h at 60 °C (107 Pa)
with TA-MRNi (modified with TA and NaBr) and with its
achiral counter catalyst (modified only with NaBr). From the
observed reaction conversions, relative reactivities of 2–4 in
reference to 1 were roughly estimated, as listed in Table 3 (en-
tries 1 and 2). Although the hydrogenation rates were varied
with the substrates and the catalysts employed, the relative re-
activities of 1–3 are similar between the chiral and achiral ca-
talyses if the higher conversion of 1 in entry 1 (71%) is consid-
ered to be a factor to increase the apparent relative reactivities
of 2 and 3. In contrast, the hydrogenation of 4 was much faster
over the chiral catalyst (entry 1) than that over the achiral cata-
lyst (entry 2). As for the similar reactivities of 3 and 4 in the
achiral catalysis, it can be reasonably understood from the sim-
ilarity in their steric bulkiness; the hydrogenation of 4 is con-
cluded to be accelerated by the tartaric acid modification.
The substrate-specific activation of the hydrogenation of 4
by the chiral modifier can be induced by two possible mecha-
nisms: the larger adsorption constant for 4 onto the TA-MRNi
surface and the faster surface reaction of the adsorbed 4 with
hydrogen. When the substrates are hydrogenated at a common
adsorption site, the hydrogenation rates for the coexisting sub-
strates are affected by each other, the degree of which depends
on the adsorption constants of the substrates. In method B, a
mixture of substrates 1–4 in an equimolar amount (4.25 mmol/
g each) was hydrogenated under the same hydrogenation con-
ditions while keeping the total molar amount of the substrates
From the rm values for each substrate, the ratios of the chiral
and achiral catalysis (shown as C/A in Table 4) were calculat-
ed. The C/A value represents the ratio of the E-site catalysis to
the N-site catalysis. When the contribution ratio of the E-site
and N-site catalyses in the hydrogenation of 1 is defined as a
reference at each temperature, the relative contribution of the
E-site for other substrates can be calculated, the values of
which are summarized in Table 5.
The calculated values, 1.2–1.3 and 1.1–1.3, for 2 and 3 indi-

T. Sugimura et al.
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 359
Table 4. Calculated Specific Rate (rm, mmol g−1 h−1) of the
Chiral and Achiral Catalysis with the Substrates 1–4 at
Various Temperatures, and the Rate Ratio of the Chiral and
Achiral Catalyses (= C/A) a)
Temp/˚C Catalyst
1
2
3
4
50
60
70
80
100
Chiral
Achiral
Ratio of C/A
Chiral
Achiral
Ratio of C/A
Chiral
Achiral
Ratio of C/A
Chiral
Achiral
0.065 0.033 0.008 0.022
0.080 0.034 0.009 0.009
0.81
0.97
0.9
2.4
0.146 0.080 0.018 0.056
0.176 0.073 0.020 0.020
0.83
1.1
0.9
2.8
0.248 0.159 0.043 0.121
0.30
0.83
(0.4)
(0.4)
—
0.143 0.038 0.041
1.1
1.1
3.0
(0.3)
0.085 0.219
0.157 0.072 0.079
Ratio of C/A
Chiral
Achiral
—
1.2
3.6
Fig. 1. Reaction conversion with the chiral (TA and NaBr-
modified) catalyst at 60 °C.
(0.7)
(0.6)
(0.6)
(0.4)
(0.2)
0.12
(0.5)
0.13
a) The values in parentheses are estimated from data over
20% conversions.
Table 5.
Calculated Relative E-site Contribution of the
Hydrogenation of 2–4 Normalized with 1
Temp/ °C
2
3
4
50
60
70
1.2
1.3
1.3
1.1
1.1
1.3
3.0
3.4
3.6
by 1/3.1. By this calculation, 4% formation of a racemic prod-
uct from 3 is reduced to be 1.3% by activation with 4, the value
of which well simulates the actual racemic production in the
hydrogenation of 4 (1.2–1.6%). Thus, the imperfect OY for
both 3 and 4, and the improvement of the OY from 3 to 4 most-
ly originated from the contribution of the N-site catalysis, and
the values of factor-i for both 3 and 4 were calculated to be al-
most unity. This result is consistent with an earlier expecta-
tion; the factor-i values for 3 and 4 should be approximately
equal due to the similarity in their steric bulkiness and hydro-
phobicity.
Fig. 2. Reaction conversion with the achiral (NaBr-modi-
fied) catalyst at 60 °C.
cate that the effects of the tartaric acid-modification to the hy-
drogenation rates are approximately equal to that for 1. Simi-
lar kinetic effects among the hydrogenations of 1–3 due to the
tartaric acid modification imply that the E/(E + N) values in
their TA-MRNi hydrogenations are almost unchanged and,
thus, the differences in the OY among them arise from the fac-
tor-i. This result is consistent with the expectation from our
working hypothesis based on the extended stereochemical
model.
In contrast, the values for 4 indicate that the activation of the
hydrogenation by the tartaric acid modification is 3.0–3.6 fold.
The activation of 3-fold at 60 °C, estimated by the preliminary
experiments. was refined to be 3.4. The substrate-specific acti-
vation for 4 indicates increment of the contribution of the E-
site catalysis over the N-site catalysis; thus, the E/(E + N) val-
ue is larger and closer to unity than those for other substrates.
The 3.4-fold activation of the E-site catalysis for 4 reduces the
contribution of the N-site catalysis for 1 by 1/3.4 and that for 3
Separation of the Factor-i and E/N Ratio in the Hydro-
genation of 1. The remaining task for the TA-MRNi/β-keto
ester system regarding the OY should be an improvement of
the OY for acetoacetate esters (86% OY for 1, methyl ester),
since the products of optically active 3-hydroxybutanoates are
industrially important chiral building blocks.13 The contribu-
tion of the N-site catalysis to the hydrogenation of 1 at 60 °C
was calculated to be 4–5% (4% of 3 × 1.1 of its relative contri-
bution of the E-site = 4.4%). Thus, factor-i for 1 was evaluat-
ed to be 0.90 (0.86/(1–0.044)). This analysis shows direction
toward a further improvement of the OY for 1. By optimizing
the base catalyst or the modification conditions, the N-site ca-
talysis will be reduced, but the expected maximum improve-
ment in the OY is only 4–5%. On the other hand, any variation
of the reaction conditions as well as the choice of the chiral
modifier affects factor-i; such an investigation has a potential
to improve the OY by 9–10%. The best MRNi for 1 in the OY,
though it has poorer activity, is the deep modified catalyst,

360 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Tartaric Acid-Modified Nickel Hydrogenation
which was reported to give 91% OY,14 the value of which is
consistent with our analysis and indicates that the deep modi-
fied catalyst has no more N-sites.
Summary. In this report, we presented an achievement of
over 98% OY by fine tuning the TA-MRNi system, and we
have proved that a heterogeneous catalysis can perform strictly
controlled enantio-differentiation. Together with 100% con-
version, a quantitative yield, and easy recovery and re-use of
the catalyst, the TA-MRNi system will become a promising
process for a large-scale production of sufficiently optically
pure alcohols.
Methyl 3-Oxobutanoate (1). The hydrogenation of 1 was
carried out at 100 °C for 4 h or at 60 °C for 31 h to give (R)-meth-
yl 3-hydroxybutanoate.19 An analytical sample was subjected to
the chiral GLC after acetylation with Ac2O/pyridine. The reten-
tion times of the GLC (90 °C) were 5.5 min for (S) and 12 min for
(R). Distillation of the product gave a colorless oil, [α]D20
=
−19.7° (neat) for the product at 100 °C, lit.19 −22.95° (neat,
100% ee).
Methyl 3-Oxopentanoate (2) The hydrogenation of 2 was
carried out at 100 °C for 18 h or at 60 °C for 34 h to give (R)-me-
thyl 3-hydroxypentanoate.20,21,22 The retention times of the GLC
(80 °C) were 10.2 min for (S) and 10.4 min for (R). The calculat-
ed OY values were 91.4% (at 100 °C) and 94.2% (at 60 °C). Fil-
tration of the mixture through a silica-gel column gave a colorless
oil, [α]D20 = −16.0° (neat) for the product at 60 °C, lit.21 −10.19°
(neat, 55% ee).
Methyl 3-Oxohexanoate. The hydrogenation of methyl 3-
oxohexanoate was carried out at 100 °C for 29 h or at 60 °C for 45
h to give (R)-methyl 3-hydroxyhexanoate.23 The retention times
of the GLC (90 °C) were 8.4 min for (S) and 8.6 min for (R). The
calculated OY values were 89.9% (at 100 °C) and 92.5% (at 60
°C). Filtration of the mixture through a silica-gel column gave a
colorless oil, [α]D20 = −9.0° (neat) for the product at 60 °C, lit.23
for antipode +13.8° (c 3.5, CHCl3, 56% ee).
Methyl 3-Oxoheptanoate. The hydrogenation of methyl 3-
oxoheptanoate was carried out at 100 °C for 31 h or at 60 °C for
67 h to give (R)-methyl 3-hydroxyheptanoate.24 The retention
times of the GLC (100 °C) were 12.6 min for (S) and 12.9 min for
(R). The calculated OY values were 88.4% (at 100 °C) and 91.2%
(at 60 °C). Filtration of the mixture through a silica-gel column
gave a colorless oil, [α]D20 = −7.6° (neat) for the product at 60 °C,
lit.24 for antipode, +26.2° (c 1.5, CHCl3, 99.5% ee).
Experimental
General. All of the substrates and the hydrogenation prod-
ucts were characterized by 1H NMR using a JEOL EXcaliber-400
spectrometer, and IR with a JASCO IR-88 spectrometer. The opti-
cal rotations were measured with a Perkin-Elmer 243B polarime-
ter. Analytical GLC was concluded on a Shimadzu GC17A.
Deionized water for preparating the catalyst was obtained from
Kanto Chemicals, Co., Ltd. All solvents were purified by distilla-
tion with proper drying agents.
Substrates. Methyl 3-oxobutanoate (1), methyl 3-oxopen-
tanoate (2), methyl 3-oxohexanoate, methyl 3-oxoheptanoate, and
methyl 4,4-dimethyl-3-oxopentanoate were obtained from com-
mercial sources. Methyl 3-oxononanoate, methyl 3-oxode-
canoate, methyl 3-oxododecanoate, methyl 4-methyl-3-oxopen-
tanoate (3), methyl 3-cyclohexyl-3-oxopropanoate, methyl 5-
methyl-3-oxohexanoate, methyl 5,5-dimethyl-3-oxohexanoate,
and methyl 3-cyclopropyl-3-oxopropanoate (4) were prepared
from Meldrum’s acid by the reported methods.15,16 The other sub-
strates were obtained by applying reported methods.17,18
Methyl 3-Oxononanoate. The hydrogenation of methyl 3-
oxononanoate was carried out at 100 °C for 36 h or at 60 °C for 52
h to give (R)-methyl 3-hydroxynonanoate. The retention times of
the GLC (130 °C) were 9.1 min for (S) and 9.3 min for (R). The
calculated OY values were 86.8% (at 100 °C) and 90.2% (at 60
°C). Filtration of the mixture through a silica-gel column gave a
Catalysts. All of the catalysts were prepared from Raney
nickel prepared by the W-1 type development of the Ni–Al alloy
(42/58, Kawaken Fine Chemicals, Ltd. Japan), followed by wash-
ing three times with ultrasonic irradiation in deionized water. The
modification was performed by heating the Raney nickel in a solu-
tion including the modifiers at 100 °C for 1 h. For preparing 0.4 g
of TA-MRNi, a solution of (R,R)-tartaric acid (0.5 g) and NaBr (5
g) in water (50 mL) was employed after adjusting the pH with
NaOH at 3.2. See the Ref. 8 for details of this catalyst. The other
catalysts were obtained by the same procedure, except for the use
of other acids (pH 3.2) or without using an acid (no pH adjust-
ment).
Preparative Hydrogenation. In a 100 mL autoclave (i.d. 36
mm × 100 mm), (R,R)-TA-NaBr-MRNi (0.4 g) and a solution of
the substrate (1.5 g) in THF (10 mL) were placed. Hydrogen was
charged into the autoclave under an initial pressure of ca. 107 Pa
(100 1 kg cm−2), and the autoclave was heated to 60 1 or 100
1 °C under reciprocating shaking until the end of the hydrogen
uptake. The autoclave was cooled and the excess hydrogen was
released from it. The products were obtained by filtration and
concentration of the reaction mixture, and were essentially chemi-
cally pure, judged by 1H NMR, except for the several sluggish re-
actions mentioned in the text. Enantiomeric ratios of the products
were determined by GLC equipped with a CP-Chirasil DEX CB
capillary column (25 m, 0.25 mm id, GL Science, Japan, flow rate:
30 cm/s). As for authentic samples of the GLC analysis, racemic
alcohols were prepared by reduction of the substrates with NaBH4
in methanol. The absolute configurations of the hydrogenation
products were determined by optical rotation and by a chemical
correlation in several key reactions.
colorless oil, [α]D20 = −5.5° (neat) for the product at 60 °C; H
1
NMR (CDCl3) δ 0.89 (t, J = 6.8 Hz, 3H), 1.21–1.57 (m, 10H),
2.40 (dd, J = 16.4, 9.0 Hz, 1H), 2.51 (dd, J = 16.4, 3.2 Hz, 1H),
2.89 (brs, 1H), 3.71 (s, 3H), 3.98 (m, 1H).
Methyl 3-Oxodecanoate. The hydrogenation of methyl 3-
oxodecanoate was carried out at 100 °C for 24 h or at 60 °C for 66
h to give (R)-methyl 3-hydroxydecanoate.16, 21 The retention times
of the GLC (140 °C) were 9.3 min for (S) and 9.5 min for (R). The
calculated OY values were 87.4% (at 100 °C) and 89.0% (at 60
°C). Filtration of the mixture through a silica-gel column gave a
colorless oil, [α]D20 = −4.5° (neat) for the product at 60 °C; H
1
NMR (CDCl3) δ 0.87 (t, J = 6.7 Hz, 3H), 1.18–1.57 (m, 12H),
2.40 (dd, J = 16.4, 9.0 Hz, 1H), 2.51 (dd, J = 16.4, 3.2 Hz, 1H),
3.70 (s, 3H), 4.00 (m, 1H).
Methyl 3-Oxododecanoate. The hydrogenation of methyl 3-
oxododecanoate was carried out at 100 °C for 24 h or at 60 °C for
69 h to give (R)-methyl 3-hydroxydodecanoate.16 The retention
times of the GLC (155 °C) were 11.7 min for (S) and 12.0 min for
(R). The calculated OY values were 83.1% (at 100 °C) and 86.4%
(at 60 °C). Filtration of the mixture through a silica-gel column
gave a colorless oil, [α]D20 = −1.0° (c 1.2, MeOH) for the product
at 60 °C; 1H NMR (CDCl3) δ 0.87 (t, J = 6.6 Hz), 1.25–1.57 (m,
16H), 2.40 (dd, J = 16.4, 9.0 Hz, 1H), 2.51 (dd, J = 16.4, 3.2 Hz,
1H), 3.70 (s, 3H), 3.99 (m, 1H).

T. Sugimura et al.
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 361
Methyl 4-Methyl-3-oxopentanoate (3). The hydrogenation
of 3 was carried out at 100 °C for 55 h or at 60 °C for 71 h to give
(S)-methyl 3-hydroxy-4-methylpentanoate.20,22,23 The retention
times of the GLC (80 °C) were 17.2 min for (R) and 17.5 min for
(S). The calculated OY values were 87.9% (at 100 °C) and 95.9%
(at 60 °C). Filtration of the mixture through a silica-gel column
gave a colorless oil, [α]D20 = −26.5° (neat) for the product at 60
°C, lit.23 for antipode +14.8° (c 1.2, EtOH, 52% ee).
to give (S)-methyl 4-ethyl-3-hydroxyhexanoate. The conversion
of the hydrogenation was 20%. An analytical sample was subject-
ed to the chiral GLC after acetylation with Ac2O/pyridine. The re-
tention times of the GLC (100 °C) were 16.7 min for (S) and 17.3
1
min for (R). The calculated OY value was 82.1%. H NMR (race-
mic form, CDCl3) δ 0.82–0.87 (m, 6H), 1.35–1.39 (m, 4H), 2.40
(dd, J = 16.6, 8.6 Hz, 1H), 2.47 (dd, J = 16.6, 3.8 Hz, 1H), 3.65
(s, 3H), 4.01 (m, 1H).
Methyl 3-Cyclobutyl-3-oxopropanoate. The hydrogenation
of methyl 3-cyclobutyl-3-oxopropanoate was carried out at 60 °C
for 72 h to give (S)-methyl 3-cyclobutyl-3-hydroxypropanoate.
An analytical sample was subjected to the chiral GLC after acety-
lation with Ac2O/pyridine. The retention times of the GLC (100
°C) were 20.4 min for (S) and 22.0 min for (R). The calculated
OY value was 94.0%. Filtration of the mixture through a silica-
Methyl 3-Cyclopropyl-3-oxopropanoate (4). The hydroge-
nation of 4 was carried out at 100 °C for 24 h or at 60 °C for 44 h
to give (S)-methyl 3-cyclopropyl-3-hydroxypropanoate.25 An ana-
lytical sample was subjected to the chiral GLC after acylation with
(S)-MTPA chloride/pyridine. The retention times of the GLC
(145 °C) were 42.0 min for (S) and 42.7 min for (R). The calculat-
ed OY values were 95.9% (at 100 °C) and 98.6% (at 60 °C). Fil-
tration of the mixture through a silica-gel column gave a colorless
gel column gave a colorless oil, [α]D20 = +2.1° (neat); H NMR
1
oil, [α]D20 = +11.0° (neat) for the product at 60 °C; H NMR
1
(CDCl3) δ 1.71–2.06 (m, 7H), 2.26 (dd, J = 16.2, 9.3 Hz, 1H),
2.41 (dd, J = 16.2, 2.9 Hz, 1H), 2.29 (m, 1H), 3.68 (s, 3H), 3.72
(m, 1H), 3.90 (m, 1H).
(CDCl3) δ 0.21 (m, 1H), 0.39 (m, 1H), 0.46–0.58 (m, 2H), 0.94
(m, 1H), 2.56 (dd, J = 16.0, 7.8 Hz, 1H), 2.62 (dd, J = 16.0, 4.2
Hz, 1H), 3.31 (m, 1H) 3.70 (s, 3H).
The absolute configuration of the product was determined to be
(S) by a chemical correlation. When the product from 4 was hy-
drogenated over PtO2 under H2 (1 kg cm−2, rt, in acetic acid), (S)-
methyl 3-hydroxy-4-methylpentanoate was obtained in a quantita-
tive yield.
Methyl 3-Cyclpentyl-3-oxopropanoate. The hydrogenation
of methyl 3-cyclpentyl-3-oxopropanoate was carried out at 100 °C
for 60 h or at 60 °C for 72 h to give (S)-methyl 3-cyclopentyl-3-
hydroxypropanoate. An analytical sample was subjected to the
chiral GLC after acetylation with Ac2O/pyridine. Retention times
of the GLC (110 °C): 23.1 min for (S) and 24.5 min for (R). The
calculated OY values were 89.7% (at 100 °C) and 95.2% (at 60
°C). Filtration of the mixture through a silica-gel column gave a
Ethyl 3-Cyclopropyl-3-oxopropanoate. The hydrogenation
of ethyl 3-cyclopropyl-3-oxopropanoate was carried out at 60 °C
for 73 h to give (S)-ethyl 3-cyclopropyl-3-hydroxypropanoate. An
analytical sample was subjected to the chiral GLC after acetyla-
tion with Ac2O/pyridine. The retention times of the GLC (100 °C)
were 16.3 min for (R) and 16.6 min for (S). The calculated OY
value was 98.2%. Filtration of the mixture through a silica-gel
colorless oil: [α]D20 = −9.6° (neat) for the product at 60 °C; H
1
NMR (CDCl3) δ 1.14–1.94 (m, 9H), 2.41 (dd, J = 16.4, 9.5 Hz,
1H), 2.55 (dd, J = 16.4, 2.7 Hz, 1H), 3.70 (s, 3H), 3.79 (m, 1H).
Methyl 3-Cyclohexyl-3-oxopropanoate.
The hydrogena-
tion of methyl 3-cyclohexyl-3-oxopropanoate was carried out at
100 °C for 60 h or at 60 °C for 75 h to give (S)-methyl 3-cyclohex-
yl-3-hydroxypropanoate.20,24 An analytical sample was subjected
to the chiral GLC after acetylation with Ac2O/pyridine. The reten-
tion times of the GLC (135 °C) were 13.4 min for (S) and 14.1
min for (R). The calculated OY values were 87.6% (at 100 °C)
and 94.0% (at 60 °C). Filtration of the mixture through a silica-
gel column gave a colorless oil, [α]D20 = −21.2° (neat) for the
product at 60 °C, lit.24 for antipode, +32.1° (c 1.4, CHCl3, 99.7%
ee).
column gave a colorless oil, [α]D20 = +8.5° (neat); H NMR
1
(CDCl3) δ 0.21 (m, 1H), 0.38 (m, 1H), 0.46–0.57 (m, 2H), 0.94
(m, 1H), 1.27 (t, J = 6.0 Hz, 3H), 2.57 (dd, J = 16.0, 8.0 Hz, 1H),
2.64 (dd, J = 16.0, 3.9 Hz, 1H), 2.84 (brs, 1H), 3.30 (m, 1H), 4.16
(q, J = 6.0 Hz, 2H).
Methyl 4-Cyclopropyl-3-oxobutanoate. The hydrogenation
of methyl 4-cyclopropyl-3-oxobutanoate was carried out at 60 °C
for 53 h to give (R)-methyl 4-cyclopropyl-3-hydroxybutanoate.
An analytical sample was subjected to the chiral GLC after acyla-
tion with (S)-MTPA chloride/pyridine. The retention times of the
GLC (170 °C) were 17.7 min for (R) and 18.2 min for (S). The
calculated OY value was 91.3%. The optical rotation of this prod-
uct was not determined. 1H NMR (CDCl3) δ 0.03–0.11 (m, 2H),
0.44–0.47 (m, 2H), 0.70–0.76 (m, 1H), 1.24–1.31 (m, 1H), 1.49–
1.56 (m, 1H), 2.45 (dd, J = 16.0, 8.0 Hz, 1H), 2.59 (dd, J = 16.0,
3.9 Hz, 1H), 3.69 (s, 3H), 4.07–4.13 (m, 1H).
Methyl 5-Methyl-3-oxohexanoate. The hydrogenation of
methyl 5-methyl-3-oxohexanoate was carried out at 60 °C for 76 h
to give (R)-methyl 3-hydroxy-5-methylhexanoate.24 The conver-
sion of the hydrogenation was 71%. The retention times of the
GLC (100 °C) were 10.2 min for (S) and 10.6 min for (R). The
calculated OY values were 92.8%. The product was separated by
a silica-gel chromatography to give a colorless oil, [α]D20 = +5.1°
1
(c 0.9, MeOH); H NMR (CDCl3) δ 0.91 (d, J = 6.4 Hz, 6H),
Methyl 3-(1-Methylcyclopropyl)-3-oxopropanoate.
The
1.13–1.22 (m, 2H), 1.72–1.83 (m, 1H), 2.38 (dd, J = 16.4, 8.8 Hz,
1H), 2.49 (dd, J = 16.4, 3.8 Hz, 1H), 3.71 (s, 3H), 4.08 (m, 1H).
Methyl 5,5-Dimetyl-3-oxohexanoate. The hydrogenation of
methyl 5,5-dimetyl-3-oxohexanoate was carried out at 100 °C for
55 h or at 60 °C for 82 h to give (R)-methyl 3-hydroxy-5,5-di-
methylhexanoate. The retention times of the GLC (110 °C) were
7.3 min for (S) and 7.6 min for (R). The calculated OY values
were 84.4% (at 100 °C) and 96.0% (at 60 °C). The optical rota-
tion of this product was not determined. 1H NMR (CDCl3) δ 0.98
(s, 9H), 1.41–1.52 (m, 2H), 2.38 (dd, J = 16.4, 8.8 Hz, 1H), 2.49
(dd, J = 16.4, 3.8 Hz, 1H), 3.71 (s, 3H), 4.16 (m, 1H).
hydrogenation of methyl 3-(1-methylcyclopropyl)-3-oxopro-
panoate was carried out at 60 °C for 69 h to give (S)-methyl 3-hy-
droxy-3-(1-methylcyclopropyl)propanoate. The conversion was
1
determined by H NMR to be 93%. An analytical sample was
subjected to the chiral GLC after acetylation with Ac2O/pyridine.
The retention times of the GLC (100 °C): 13.0 min for (R) and
13.2 min for (S). The calculated OY was 94.3%. 1H NMR in ra-
cemic form (CDCl3) δ 0.27–0.30 (m, 1H), 0.33–0.40 (m, 2H),
0.47–0.50 (m, 1H), 1.03 (s, 3H), 2.52 (dd, J = 16.0, 3.9 Hz, 1H),
2.58 (dd, J = 16.0, 9.3 Hz, 1H), 3.33–3.36 (m, 1H), 3.69 (s, 3H).
Methyl 3-[(1RS,2RS)-2-Methylcyclopropyl]-3-oxopropano-
ate. The hydrogenation of methyl 3-[(1RS,2RS)-2-methylcyclo-
propyl]-3-oxopropanoate was carried out at 60 °C for 28 h to give
Methyl 4-Ethyl-3-oxohexanoate.
The hydrogenation of
methyl 4-ethyl-3-oxohexanoate was carried out at 60 °C for 4 days

362 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Tartaric Acid-Modified Nickel Hydrogenation
(3S)-methyl 3-hydroxy-3-[(1RS,2RS)-2-methylcyclopropyl]pro-
panoate. An analytical sample was subjected to the chiral GLC af-
ter acetylation with Ac2O/pyridine. Enantiomeric pairs were as-
signed using the reduction product with NaBH4. The retention
times of the GLC (115 °C) for one pair of enantiomers were 11.8
and 12.5 min, and the other ones were 12.7 and 13.4 min. The
compositions of the hydrogenation products were 49.0, 0.9, 49.7,
and 0.4% (in order of the retention times). The two major prod-
ucts were empirically assigned to be (3S). Thus, the diastereomer-
ic excesses at the 3-position were 98.3 and 96.5%. Filtration of
the mixture through a silica-gel column gave a colorless oil (a
mixture of the diastereomers): 1H NMR (CDCl3) δ 0.20–0.78 (m,
4H), 1.00 and 1.03 (d, J = 4.0 Hz, 3H), 2.05–2.62 (m, 2H), 3.29–
3.35 (m, 1H), 3.68 and 3.68 (s, 3H).
ucts were empirically assigned to be (3S). Thus, the diastereomer-
ic excesses at the 3-position were 71.4 and 61.5%. Filtration of
the mixture through a silica-gel column gave a colorless oil (a
mixture of the diastereomers), 1H NMR (CDCl3) δ 0.90–0.99 (m,
1H), 1.22–1.31 (m, 1H), 1.96–2.02 (m, 1H), 2.35–2.48 (m, 1H),
2.58–2.68 (m, 1H), 3.69 and 3.70 (s, 3H), 3.97–4.04 (m, 1H),
7.04–7.29 (m, 5H).
Reaction Conversion by the Method A. In a 100 mL auto-
clave were placed a catalyst (0.4 g), a substrate (6.8 mmol), acetic
acid (0.05 mL), and methyl hexanoate (ca. 50 mg) as an internal
standard in THF (10 mL). Hydrogen was charged into the auto-
clave under an initial pressure of ca. 107 Pa. The autoclave was
heated to 60 °C before starting to shake. After the shaking had
continued for 4 h, the reaction mixture was cooled to room tem-
perature, and filtered to remove the catalyst. An analytical sample
was directly subjected to the GLC equipped with a TC-WAX cap-
illary column (120 °C). The retention times of the four products
were 13.7 min for 1, 17.7 min for 2, 20.1 min for 3 and 43.1 min
for 4, and that of the internal standard was 6.6 min.
Reaction Conversion by the Method B. Into a 100 mL au-
toclave were placed a catalyst (0.4 g), a mixture of four substrates
1–4 (1.7 mmol × 4), acetic acid (0.05 mL), methyl hexanoate (ca.
50 mg) as an internal standard, and THF (10 mL). The reaction
conditions and analytical method were the same as those for the
method A.
Kinetic Study of the Hydrogenation. The same procedure
as method B was adapted, except for the use of 40 mL of THF and
varied reaction temperatures (50–100 °C). An experiment at 100
°C was carried out with 0.08 g of the catalyst. The hydrogenation
was started with shaking after the temperature became constant.
Analytical samples were taken out every hour through a side tube
connected to the autoclave. The sample was analyzed under the
same conditions as in methods A and B.
Methyl 3-[(1RS)-2,2-Dimethylcyclopropyl]-3-oxopropano-
ate. The hydrogenation of methyl 3-[(1RS)-2,2-dimethylcyclo-
propyl]-3-oxopropanoate was carried out at 60 °C for 30 h to give
(3S)-methyl
3-[(1RS)-2,2-dimethylcyclopropyl]-3-hydroxypro-
panoate. The conversion was determined by 1H NMR to be 29%.
An analytical sample was subjected to the chiral GLC after acety-
lation with Ac2O/pyridine. Enantiomeric pairs were assigned us-
ing the reduction product with NaBH4. The retention times of the
GLC (100 °C) of one pair of enantiomers were 36.6 and 37.8 min,
and the other ones were 39.8 and 41.5 min. The compositions of
the hydrogenation products were 42.8, 10.7, 32.1, and 14.5% (in
order of the retention times). The two major products were empir-
ically assigned to be (3S). Thus, the diastereomeric excesses at
the 3-position were 50.0 and 50.0%. Filtration of the mixture
through a silica-gel column gave a colorless oil (mixture of the di-
1
astereomers): H NMR (CDCl3) δ 0.03–0.06 (m, 1H), 0.43–0.46
(m, 1H), 0.69–0.75 (m, 1H), 0.88 and 0.90 (s, 3H), 1.02–1.05 (s,
3H), 2.55–2.60 (m, 2H), 3.52–3.61 (m, 1H), 3.68 and 3.69 (s, 3H).
Ethyl 3-[(1RS,2RS)-2-Ethoxycarbonylcyclopropyl]-3-oxo-
propanoate. The hydrogenation of ethyl 3-[(1RS,2RS)-2-ethoxy-
carbonylcyclopropyl]-3-oxopropanoate was carried out at 60 °C
for 28 h to give (3S)-ethyl 3-[(1RS,2RS)-2-ethoxycarbonylcyclo-
propyl]-3-hydroxypropanoate. An analytical sample was subject-
ed to the chiral GLC after acylation with butanoic anhydride/pyri-
dine. Enantiomeric pairs were assigned using the reduction prod-
uct with NaBH4. The retention times of the GLC (120 °C) for one
pair of enantiomers were 130.3 and 133.3 min, and the other ones
were 136.1 and 146.6 min. The compositions of the hydrogena-
tion products were 52.7, 13.9, 31.7, and 1.7% (in order of the re-
tention times). The two major products were empirically assigned
to be (3S). Thus, the diastereomeric excesses at the 3-position
were 93.9 and 38.9%. Filtration of the mixture through a silica-
gel column gave a colorless oil (a mixture of the diastereomers),
1H NMR (CDCl3) δ 0.88 and 1.01 (m, 1H), 1.12–1.17 and 1.52–
1.57 (m, 2H), 1.20–1.28 (m, 6H), 1.61 and 1.67 (m, 1H), 2.51–
2.63 (m, 2H), 3.60 and 3.70 (m, 1H), 4.06–4.19 (m, 4H).
The authors thank Professors Y. Izumi (Osaka, Emeritus), T.
Harada (Ryukoku), T. Osawa (Toyama), Y. Nitta (Niihama), A.
Onishi (Toyo Kasei), and T. Okuyama (Himeji) for their help-
ful suggestions. This work was partially supported by a Grant-
in-Aid for Scientific Research (No. 0660697) from the Minis-
try of Education, Science, Sports and Culture.
References
1
S. Akabori, Y. Izumi, Y. Fujii, and S. Sakurai, Nature, 178,
323 (1956).
2
Y. Orito, S. Imai, and S. Niwa, Nippon Kagakukaisi, 1979,
1118; 1980, 670; and 1982, 137; Y. Orito, S. Imai, S. Niwa, and G.
H. Nguyen, J. Synth. Org. Jpn., 37, 173 (1979).
3
For improved Orito systems, see: H. U. Blaser, Tetrahe-
dron: Asymmetry, 2, 843 (1991). H. U. Blaser and H. P. Jalett,
Stud. Surf. Sci. Catal., 78, 139 (1993); A. Baiker, J. Mol. Catal.
A., 115, 473 (1997); H. U. Blaser, H. P. Jalett, M. Muller, and M.
Studer, Catal. Today, 37, 441 (1997); X. Zuo and M. Liu, Tetrahe-
dron Lett., 39, 1941 (1998).
Methyl 3-Oxo-3-[(1RS,2RS)-2-phenylcyclopropyl]propano-
ate. The hydrogenation of methyl 3-oxo-3-[(1RS,2RS)-2-phe-
nylcyclopropyl]propanoate was carried out at 60 °C for 67 h to
give (3S)-methyl 3-hydroxy-3-[(1RS,2RS)-2-phenylcyclopropyl]-
1
propanoate. The conversion of 7 was determined by H NMR to
4
5
Y. Izumi, Angew. Chem., Int. Ed. Engl., 10, 871 (1971).
For a review of varied MRNi’s, see: Y. Izumi, Adv. Catal.,
be 44%. An analytical sample was subjected to the chiral GLC af-
ter acetylation with Ac2O/pyridine. Enantiomeric pairs were as-
signed using the reduction product with NaBH4. The retention
times of the GLC (170 °C) for one pair of enantiomers were 26.5
and 27.3 min, and the other ones were 28.3 and 29.0 min. The
compositions of the hydrogenation mixture were 45.6, 9.0, 37.8,
and 7.6% (in order of the retention times). The two major prod-
32, 215 (1983).
6
For reviews of TA-MRNi system, see: A. Tai and T.
Harada, in “Tailored Metal Catalyst,” ed by Y. Iwasawa, Reidel,
Dordrecht (1986), p. 265; A. Tai and T. Sugimura, in “Chiral
Catalyst Immobilization and Recycling,” ed by D. E. De Vos, I. F.
J. Vankelecom, and P. A. Jacobs, Wiley, Weinheim (2000), p. 173.

T. Sugimura et al.
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 363
7
For the hydrogenation of simple ketones with TA-MRNi,
Osawa, and A. Tai, Abstract for the 8th Int. Symp. Relations
between Homogen and Heterogen. Catal., Balatonfured, (1995), p.
78.
15 Y. Oikawa, K. Sugano, and O. Yonemitsu, J. Org. Chem.,
43, 2087 (1978).
see: T. Osawa and T. Harada, Bull. Chem. Soc. Jpn., 57, 1618
(1984); T. Osawa, T. Harada, and A. Tai, J. Mol. Catal., 121, 7
(1990); T. Harada and T. Osawa, in “Chiral Reaction in Heteroge-
neous Catalysis,” ed by G. Jannes and D. Dubois, Plenum, New
York (1995), p. 83; T. Osawa, T. Harada, and A. Tai, Catal. Today,
37, 465 (1997).
16 M. Nakahata, M. Imaida, H. Ozaki, T. Harada, and A. Tai,
Bull. Chem. Soc. Jpn., 55, 2186 (1982).
8
A. Tai, T. Kikukawa, T. Sugimura, Y. Inoue, T. Osawa, and
17 S. Benetti and R. Romagnoli, Chem. Rev., 95, 1065 (1995).
18 D. F. Taber, J. C. Amedio, and Y. K. Patel, J. Org. Chem.,
50, 3618 (1985).
19 T. Harada and Y. Izumi, Chem. Lett., 1978, 1195.
20 M. J. Burk, T. Gregory P. Harper, and C. S. Kalbelg, J. Am.
Chem. Soc., 117, 4423 (1995).
S. Fujii, J. Chem. Soc., Chem. Commun., 1991, 795; A. Tai, T.
Kikukawa, T. Sugimura, Y. Inoue, S. Abe, T. Osawa, and T.
Harada, Bull. Chem. Soc. Jpn., 67, 2473 (1994).
9
T. Sugimura, T. Osawa, S. Nakagawa, T. Harada, and A.
Tai, Stud. Surf. Sci. Catal., 101, 231 (1996).
10 For a preliminary report, see: S. Nakagawa, T. Sugimura,
and A. Tai, Chem. Lett., 1997, 859 and 1998, 1257; See also: T.
Sugimura, Catal. Surv. Jpn., 3, 37 (1999).
21 T. Tanabe and Y. Izumi, Bull. Chem. Soc. Jpn., 46, 1550
(1973).
22 A. Tai, T. Harada, Y. Hiraki, and S. Murakami, Bull. Chem.
Soc. Jpn., 56, 1414 (1983).
11 R. Fellous, J. Am. Chem. Soc., 95, 5593 (1973).
12 T. Osawa, T. Harada, and A. Tai, J. Catal., 121, 7 (1990).
13 M. Fieser, “Organic Synthesis,” Wiley, New York, (1990),
Vol. 15, p. 207 and (1992), Vol. 16, p. 218; T. Ohashi and J.
Hasegawa, in “Chirality in Industry,” Wiley, New York, ed by A.
N. Collings, G. N. Sheldrake, and J. Crosby, (1992), p. 269.
14 T. Harada, Y. Sasaki, T. Kitamura, H. Hatta, K. Niwa, T.
23 R. Devant, U. Mahler, and M. Braun, Chem. Ber., 119,
2191 (1986).
24 W. Oppolzer and C. Starkemann, Tetrahedron Lett., 33,
2439 (1992).
25 H. Wolf, U. Mätzel, E. J. Brunke, and E. Klein, Tetrahe-
dron Lett., 25, 2339 (1979).