Enantioselective hydrogenation of tiglic acid in methanol and in dense carbon dioxide catalyzed by a ruthenium-BINAP complex substituted with OCF 3 groups

Article abstract of DOI:10.1016/j.molcata.2003.10.016

A fluorinated analog of the 2,2′-bis(diphenylphosphino)-1,1′- binaphthyl (BINAP) ligand was synthesized with OCF₃-substitution of the aryl groups in BINAP skeleton (p-OCF₃-BINAP). Ruthenium complexes of both BINAP (Ru-BINAP) and (p-OCF₃)-BINAP (Ru-[(p-OCF₃)-BINAP]) were also synthesized and investigated as catalysts for hydrogenation of tiglic acid in methanol. Typically, Ru-[(p-OCF₃)-BINAP] had lower activity but had higher enantioselectivity at high hydrogen pressures than Ru-BINAP at the same condition. The effect of OCF₃ groups on the catalytic properties was discussed on the basis of NMR spectra and kinetic data. Ru-[(p-OCF ₃)-BINAP] was found to have sufficiently high solubility in dense CO₂ for homogeneous catalytic reactions and was investigated for hydrogenation of tiglic acid in CO₂. The results showed that CO ₂ had a great influence on both activity and enantioselectivity. Addition of methanol to CO₂ was found to increase the enantioselectivity.

Full text of DOI:10.1016/j.molcata.2003.10.016

Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
Enantioselective hydrogenation of tiglic acid in methanol and in
dense carbon dioxide catalyzed by a ruthenium–BINAP
complex substituted with OCF3 groups
Xing Dong, Can Erkey∗
Department of Chemical Engineering, Environmental Engineering Program, University of Connecticut, 191 Auditorium Road, Storrs, CT 06269, USA
Received 7 August 2003; received in revised form 14 October 2003; accepted 14 October 2003
Abstract
A fluorinated analog of the 2,2ꢀ-bis(diphenylphosphino)-1,1ꢀ-binaphthyl (BINAP) ligand was synthesized with OCF3-substitution of the aryl
groups in BINAP skeleton (p-OCF3–BINAP). Ruthenium complexes of both BINAP (Ru–BINAP) and (p-OCF3)–BINAP (Ru–[(p-OCF3)–
BINAP]) were also synthesized and investigated as catalysts for hydrogenation of tiglic acid in methanol. Typically, Ru–[(p-OCF3)–BINAP]
had lower activity but had higher enantioselectivity at high hydrogen pressures than Ru–BINAP at the same condition. The effect of OCF3
groups on the catalytic properties was discussed on the basis of NMR spectra and kinetic data. Ru–[(p-OCF3)–BINAP] was found to have
sufficiently high solubility in dense CO2 for homogeneous catalytic reactions and was investigated for hydrogenation of tiglic acid in CO2.
The results showed that CO2 had a great influence on both activity and enantioselectivity. Addition of methanol to CO2 was found to increase
the enantioselectivity.
© 2003 Published by Elsevier B.V.
Keywords: BINAP; Fluorinated ligands; Tiglic acid; Hydrogenation; Supercritical CO2
1. Introduction
liquids [4] and liquid biphasic systems [5] are being inves-
tigated for chiral catalysis. For example, Pozzi et al. [6]
reported the first chiral fluorous biphasic catalysis in 1998
by using C2 nitrogen-based ligands. Subsequently, several
promising results with different fluorous catalysts were
described [7–9]. Francio et al. [10] showed that by using
rhodium complexes of the perfluoroalkyl-substituted ligand,
(R,S)-3-H2F6–BINAPHOS, a large variety of substrates
could be hydroformylated in scCO2 with rates and enan-
tioselectivities comparable to those obtained in a benzene
solution with the conventional catalyst.
An important issue that needs to be addressed in develop-
ment of fluorinated complexes is the effect of highly elec-
tron withdrawing property of fluorous groups on catalytic
activity and selectivity. A few studies have been reported in
the literature on this effect for fluorinated catalysts. It was
found that the activity of fluorinated rhodium complexes
in the hydroformylation of olefins in scCO2 increased with
decreasing basicity of phosphines [11,12]. Further investi-
gation by in-situ FTIR spectroscopy showed that this en-
hancement was due to a shift in the equilibrium distribution
of the catalytically active intermediates [13]. On the other
Chiral chemistry is a rapidly growing field since chiral
compounds play a key role in many fields such as agri-
culture, biology, materials science and life sciences. For
example, the worldwide sales for chiral drugs are growing
at an annual rate of 13%, and the sales are expected to reach
US$ 200 billion in 2008 [1]. Over the past several decades,
catalysis by chiral organometallic complexes has emerged
as one of the most attractive approaches in the production of
chiral chemicals. This method can multiply chirality with a
catalytic amount of chiral source contained in the catalysts.
Despite its advantages and efficiency, industrial appli-
cations of chiral catalysis are limited partly due to the
difficulties associated with recovery and recycle of the ex-
pensive catalysts from reaction mixtures. Therefore, novel
reaction media for organometallic chemistry, which include
supercritical fluids (SCFs) [2], ionic liquids [3], fluorous
∗
Corresponding author. Tel.: +1-860-486-4601;
fax: +1-860-486-2959.
E-mail address: cerkey@engr.uconn.edu (C. Erkey).
1381-1169/$ – see front matter © 2003 Published by Elsevier B.V.
doi:10.1016/j.molcata.2003.10.016

74
X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
O
O
H2
*
OH
Ru-[BINAP]
Scheme 1. Tiglic acid hydrogenation reaction.
hand, Horvath et al. [14] found that arylphosphine ligands
gave more reactive rhodium catalysts than alkylphosphine
ligands, and the fluorine substituents in the ligands further
retarded the reaction rate in the hydroboration of norbornene
in the fluorous solvent CF3C6H5. These studies indicate that
there is a need to study the influence of the fluorous groups
on activity and enantioselectivity of chiral organometallic
catalysts for a wide variety of reactions.
mixture was cooled under ice bath and 10% hydrochloric
acid (48 ml) was added slowly. The ether layer was ex-
tracted with 2% hydrochloric acid (200 ml), and the whole
was dried by Na2SO4 and evaporated, giving yellowish
bis (4-triflouromethoxy)benzene phosphine oxide(13.9 g,
79%). The spectral properties of the product are as fol-
lows: 1H NMR (400 MHz, CDCl3) δ: 8.48 ppm (d, 1H,
1JH,P = 495.8 Hz), 8.08 ppm (dd, 4H, JH,H = 8.4 Hz,
3
The ruthenium complexes containing 2,2ꢀ-bis(diphenyl-
phosphino)-1,1ꢀ–binaphthyl (BINAP) are the most exten-
sively used catalysts in many enantioselective reactions [15].
Among those, the enantioselective hydrogenation of unsat-
urated acids such as tiglic acid (Scheme 1) is of significant
interest because of the unique efficacy of the Ru–BINAP
complexes in these reactions and the high economic value
of the products, which afford useful building blocks for the
synthesis of non-steroidal anti-inflammatory agents [16]. In
this article, we report the synthesis of a fluorinated BINAP
ligand and its ruthenium complex, as well as our results on
the hydrogenation of tiglic acid in methanol and in dense
CO2 catalyzed by this new complex and the conventional
ruthenium–BINAP complex.
3JH,P = 13.2 Hz), 7.63 ppm (d, 4H, JH,H = 7.8 Hz). 31P
3
NMR (162 MHz, CDCl3) δ: 20.43 ppm (s, decoupled).
2.3. Synthesis of 2b
Trichlorosilane (2.7 ml) was added slowly with stirring
to 1b (to 2.25 g, 6.64 mmol), triethyl amine (3.84 ml) and
toluene (30 ml) under ice bath, and heated under reflux for
4 h. After cooling, 2N sodium hydroxide solution (133 ml)
was added slowly under ice bath. The organic layer was
combined with ether and evaporated under vacuum, giving
the yellow liquid (1.86 g, 87%). The spectral properties of
1
2 are as follows: H NMR (400 MHz, CDCl3) δ: 7.53 ppm
(dd, 4H, 3JH,H = 6.8 Hz, 3JH,P = 6.6 Hz), 7.24 ppm (d, 4H,
3JH,H = 7.6 Hz), 5.30 ppm (d, 1H, JH,P = 219.7 Hz). 31P
1
NMR (162 MHz, CDCl3) δ: −41.67 ppm (s, decoupled).
2.4. Synthesis of 3b
2. Experimental section
2.1. General methods for preparation of Ru–BINAP and
Ru–[(p-OCF3)–BINAP] complexes
To a solution of NiCl2dppe (355 mg) in DMF (10 ml) at
room temperature was added 2b (2.5 g, 7.74 mmol). The so-
lution was heated to 110 ◦C and kept at this temperature for
30 min. (R)-(−)-1,1ꢀ-Bi-2-naphthol bis (trifluoro methane-
sulfonate) (3.6 g, 6.55 mmol) and DABCO (2.9 g) in DMF
(20 ml) was added slowly to the above solution. Three addi-
tional equal portions of 2b were added after 1, 3 and 7 h. The
reaction was kept at 110 ◦C until (R)-(−)-1,1ꢀ-Bi-2-naphthol
bis (trifluoro methanesulfonate) was completely consumed
(3 days). 3b (3.2 g, 43%) was obtained by recrystallizion
from a mixture of methanol and DMF. The spectral prop-
All phosphine compound syntheses were carried out un-
der a nitrogen atmosphere. (R)-(−)-1,1ꢀ-Bi-2-naphthol bis
(trifluoro methanesulfonate), 1-bromo-4-(triflouromethoxy)
benzene, diethyl phosphite, trichlorosilane, triethyl amine
were purchased from Aldrich and magnesium turnings were
obtained from Acros. All the chemicals were used as re-
ceived. The synthetic scheme for fluorinated BINAP and
Ru–BINAP type complexes is shown in Scheme 2.
1
erties of 3b are as follows: H NMR (400 MHz, CDCl3) δ:
2.2. Synthesis of 1b
7.98 ppm (d, 2H, J = 8.3 Hz), 7.89 ppm (d, 2H, J = 8.0 Hz,
3JH,P = 13.2 Hz), 7.43 ppm (m, 4H), 7.15 ppm (m, 4H),
7.07 ppm (m, 8H), 6.97 ppm (m, 6H), 6.75 ppm (d, 2H, J =
8.8 Hz). 31P NMR (162 MHz, CDCl3) δ: −14.89 ppm (s, de-
coupled). Elemental analysis: C48H28F12O4P2 requires C:
60.12%; H: 2.93%; F: 23.80%, found C: 59.87%; H: 2.49%;
F: 24.80%.
1-bromo-4-(triflouromethoxy)benzene (25 g, 0.104 mol)
was added slowly with stirring to magnesium turnings
(2.74 g, 0.115 mol) and ether (25 ml), and heated under re-
flux for 30 min. To the cooled reagent, diethyl phosphite
(7 ml, 0.052 mol) in ether (5 ml) was added slowly with
stirring, and the whole heated under reflux for 60 min. The

X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
75
O
R
R
PH
Mg
Diethyl Phosphite
Diethyl Ether
Diethyl Ether
MgBr
Br
R
R
1a
1b
O
H
P
PH
HSiCl , Triethyl amine
3
Toluene
R
R
R
R
2a
2b
R
H
P
O
O
O
S
S
CF3
CF3
R
R
P
P
NiCl2(dppe)
+
DABCO, DMF
O
O
R
R
O
R
3a
3b
a R=H
b R= OCF3
O
O
P
P
P
3
COOH
CH3
Ru
O
Ru
Ru
Toluene
Diethyl ether
O
P
Scheme 2. Synthetic procedure for BINAP and (p-OCF3)–BINAP and their complexes with ruthenium.
2.5. Synthesis of Ru–BINAP type complexes
NMR (162 MHz, CDCl3) δ: 65.06 ppm (s, decoupled). The
spectral properties of Ru–[(p-OCF3)–BINAP] are as fol-
lows: H NMR (400 MHz, CDCl3) δ: 7.84 ppm (m, 4H),
7.55 ppm (t, 4H, J = 8.9 Hz), 7.32 ppm (m, 8H), 7.14 ppm
(m, 4H), 7.00 ppm (t, 2H, J = 8.4 Hz), 6.67 ppm (d, 2H,
J = 8.4 Hz), 6.18 ppm (d, 4H, J = 10.8 Hz). 31P NMR
(162 MHz, CDCl3) δ: 64.88 ppm (s, decoupled).
1
A solution of (COD)Ru(methylallyl)2 (168 mg, 0.527
mmol) and 3b (501 mg, 0.523 mmol) or 3a (325 mg,
0.523 mmol) in toluene (20 ml) was heated under reflux
for 4 h and the toluene was evaporated under vacuum. A
solution of acetic acid (2 ml) in ether (15 ml) was added
at room temperature and the resulting solution was stirred
for 10–14 h. Ru–BINAP complexes (402 mg, 91%) and
Ru–[(p-OCF3)–BINAP] (527 mg, 86%) were obtained as
yellow-brown solid. The spectral properties of Ru–BINAP
2.6. General procedure for investigation of the tiglic
acid hydrogenation reactions
1
are as follows: H NMR (400 MHz, CDCl3) δ: 7.84 ppm
The experimental set-up is shown in Fig. 1. The reactor
is a custom manufactured, 54 ml stainless steel vessel fitted
with two sapphire windows (1 in. i.d., Sapphire Engineer-
ing), polyether–ether–ketone o-rings (Valco Instruments),
(m, 4H), 7.45 ppm (m, 12H), 7.21 ppm (t, 2H, J
=
7.2 Hz), 7.09 ppm (m, 4H), 6.89 ppm (t, 2H, J = 8.4 Hz),
6.63 ppm (d, 2H, J = 8.7 Hz), 6.50 ppm (m, 6H). 31P

76
X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
0.12
3.4 bar
10.2 bar
13.6 bar
27.2 bar
95.3 bar
0.10
6
8
7
PT
0.08
2
0.06
10
0.04
3
TC
5
9
1
11
0.02
4
12
0.00
0
5
10
15
20
25
30
Fig. 1. Experimental set-up. (1) Hydrogen cylinder, (2) syringe pump, (3)
windowed reactor, (4) stir plate, (5) thermocouple assembly, (6) pressure
transducer, (7) vent line, (8) solvent reservoir, (9) rupture disk assembly,
(10) sample loop, and (11) sample collection vial.
Time / hr
Fig. 2. Hydrogen pressure effect on the tiglic acid hydrogenation
reactions in methanol catalyzed by [(p-OCF3)–BINAP]Ru[O2CCH3]2
{[tiglic acid]0 = 0.125 M, [catalyst] = 0.3 mM, T = 25 ◦C}.
a T-type thermocouple assembly (Omega Engineering,
DP41-TC-MDSS), a pressure transducer (Omega Engi-
neering, PX300-7.5KGV), a vent line, and a rupture disk
assembly (Autoclave Engineers). When methanol was used
as a solvent, a certain amount of tiglic acid and catalyst was
placed into the reactor. Then the reactor was sealed and the
air was removed by flushing with nitrogen several times. A
certain amount of degassed methanol was charged into the
reactor, and the vessel was heated to the desired tempera-
ture. The reaction was started by charging the reactor with
hydrogen. When CO2 was used as a solvent, the procedure
was the same as above except that the catalyst was isolated
in an ampoule, and the air was removed by flushing with
hydrogen several times. Subsequently, a certain amount
of hydrogen was charged, followed by charging with CO2
using a high-pressure syringe pump (ISCO 260D). During
this period, the ampoule broke and the reaction started.
The samples were taken periodically using the sampling
system shown in Fig. 1. The valve between high-pressure
sample loop and the reactor was opened periodically to take
samples. The sample was then depressurized into a sample
vial. Subsequently, the sample loop was flushed with a small
amount of acetone from a reservoir. The sample loop was
further dried with compressed air. The collected samples
were analyzed by gas chromatography (HP 6890) equipped
with a CP-Chiralsil-DEX CB column.
The results also show that the reaction rate with Ru–BINAP
was much higher than that with Ru–[(p-OCF3)–BINAP] at
the same hydrogen pressure. The hydrogen pressure effect
on enantioselectivity is given in Fig. 4 and shows that the
enantioselectivity did not change with pressure for hydrogen
pressures less than 13.6 bar. However, the enantioselectivity
decreased with further-increasing hydrogen pressure.
The effect of the tiglic acid concentration on reaction rate
was also investigated in methanol in the range of 1.2 ∼
500 mM. In all cases, the catalyst concentration was kept
constant as 0.3 mM. The data are listed in Figs. 5 and 6 and
show that the reaction rate increased with increasing tiglic
acid concentration. As shown in Table 1, the initial tiglic
acid concentration had no appreciable effect on enantiose-
lectivity.
The rate expression was developed on the basis of above
kinetic data. The experimental reaction rate at each data
0.25
3.4 bar
6.8 bar
12.6 bar
0.20
25.2 bar
0.15
0.10
0.05
0.00
3. Results
The effect of hydrogen pressure on reaction rate for hy-
drogenation of tiglic acid in methanol was investigated in
the pressure range of 3.4 ∼ 95.3 bar and 3.4 ∼ 25.2 bar with
fluorinated and conventional Ru–BINAP complexes, respec-
tively. In all cases, the catalyst concentration was kept con-
stant as 0.3 mM. The data are given in Figs. 2 and 3 indicate
that the rate increases with increasing hydrogen pressure.
0
1
2
3
4
5
Time / hr
Fig. 3. Hydrogen pressure effect on the tiglic acid hydrogenation reactions
in methanol catalyzed by [BINAP]Ru[O2CCH3]2 {[tiglic acid]0 = 0.25 M,
[catalyst] = 0.3 mM, T = 25 ◦C}.

X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
77
0.4
Ru-[(P-OCF )-BINAP]
3
Ru-BINAP
0.125M
0.25M
0.5M
100
80
60
40
20
0
0.3
0.2
0.1
0.0
0
1
2
3
4
Time / hr
3.4
6.8
10.2
13.6
27.2
95.3
Hydrogen pressure/bar
Fig. 6. Tiglic acid concentration effect on the tiglic acid hydrogena-
tion reactions in methanol [BINAP]Ru[O2CCH3]2 {H2 = 95.3 bar,
[catalyst] = 0.3 mM, T = 25 ◦C}.
Fig. 4. Hydrogen pressure effect on the enantioselectivity {[tiglic acid]0
= 0.125 M, [catalyst] = 0.3 mM, T = 25 ◦C}.
0.4
lated from the solubility of hydrogen in methanol [17] and
the reaction stoichiometry) were used to determine the rate
expression by using the Polymath software. The rate expres-
sion for the reactions catalyzed by Ru–BINAP in methanol
was:
0.125 M
0.25 M
0.5 M
0.3
−r = 5.672 × Ca0.881Cb0.599
0.2
0.1
0.0
where the rate is given in units of mol dm−3 h−1 and the
concentrations are expressed in mol dm−3. Ca represents the
concentration of tiglic acid and Cb represents the concen-
tration of hydrogen. Similar kinetic features were also ob-
served with Ru–[(p-OCF3)–BINAP] complex and the rate
expression was found to be:
−r = 0.875 × Ca0.929Cb0.627
0
2
4
6
8
10
12
Time / hr
The degree of fit for the rate expressions is illustrated in
Figs. 2, 3, 5 and 6 where the solid curves represent the pre-
dicted values that were calculated using the rate expression
and the dots represent the experimental data.The reaction
was also studied at different temperatures. The experimen-
tal data are provided in Figs. 7 and 8. The data were used
to determine the activation energy of the reaction from the
slope of the plot of ln k versus 1/T. The calculated activation
Fig. 5. Tiglic acid concentration effect on the tiglic acid hydrogena-
tion reactions in methanol catalyzed by [(p-OCF3)–BINAP]Ru[O2CCH3]2
{H2 = 95.3 bar, [catalyst] = 0.3 mM, T = 25 ◦C}.
point was calculated by differentiating the concentration of
tiglic acid versus time curves. These experimental reaction
rate data together with the experimental concentration data
(the concentration of hydrogen at each data point was calcu-
Table 1
[Tiglic acid]0 and temperature effect on enantioselectivity in methanol
Catalyst
ee (%)
Temperature (◦C)
[Tiglic acid]0 (M)
0
8.6
10.5
22.3
23.5
35.1
36.2
0.125
0.250
0.500
Ru–BINAP
Ru–[(p-OCF3)–BINAP]
–
82.9a
–
–
–
83.8a
–
84.0a
–
–
83.8c
77.8d
83.1c
78.6d
84.5c
80.0d
76.0b
75.3b
77.8b
75.5b
a
H2 = 13.8 bar, catalyst = 0.3 mM, [tiglic acid]0 = 0.25 M.
H2 = 96.5 bar, catalyst = 0.3 mM, [tiglic acid]0 = 0.125 M.
H2 = 13.8 bar, catalyst = 0.3 mM, T = 25 ◦C.
b
c
d
H2 = 96.5 bar, catalyst = 0.3 mM, T = 25 ◦C.

78
X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
Table 2
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
oC
oC
oC
Variation of enantioselectivity in CO2
22.3
10.5
36.2
Conditions
ee (%)
oC
0
[H2]0 (bar)a
0.7
30
50
57.3
54.0
55.1
[Tiglic acid]0 (mM)b
37.0
92.5
185.5
51.6
56.6
62.5
[Catalyst] (mM)c
0.08
0.24
0.37
55.3
56.6
53.2
a
[Tiglic acid]0 = 92.5 mM, [catalyst] = 0.24 mM, P (total) = 166 bar,
T = 25 ◦C, methanol = 1 ml.
0
2
4
6
8
10
12
b
Time / hr
[H2]0 = 30 bar, [catalyst] = 0.24 mM, P (total) = 166 bar, T =
25 ◦C, methanol = 1 ml.
Fig. 7. Temperature effect on the tiglic acid hydrogenation reactions in
methanol catalyzed by [(p-OCF3)–BINAP]Ru[O2CCH3]2 {H2 = 95.3 bar,
[catalyst] = 0.3 mM, [tiglic acid]0 = 0.125 M}.
c
[H2]0 = 50 bar, [tiglic acid] = 92.5 mM, P (total) = 166 bar, T =
25 ◦C, methanol = 1 ml.
temperature), in contrast with high enantioselectivity ob-
tained in methanol. Therefore, methanol was added as a
cosolvent or more properly, as a promoter, to the reaction
system. Not surprisingly, the enantioselectivity jumped
from 25 ∼ 35 to 50 ∼ 60%. Therefore, all of the reactions
in CO2 were carried out with added methanol (1 ml).
energies were 82.7 1.9 kJ mol−1 and 80.1 1.4 kJ mol−1
for reactions catalyzed by Ru–[(p-OCF3)–BINAP] and
Ru–BINAP, respectively. The former is slightly higher than
the latter, which is also reflected in the difference in the rate
constants, as shown in Section 2.6.
Since hydrogen is miscible with CO2, in order to com-
pare the results with the results of hydrogen pressure ef-
fect on enantioselectivity in methanol at the same hydro-
gen concentration in the catalyst phase, the hydrogen pres-
sures used for reactions in CO2 were selected to range from
0.7 to 50 bar. Unlike in methanol, enantioselectivity was
not affected by hydrogen pressure for reactions in CO2, as
shown in Table 2. This probably results from the fact that
the surrounding environment for the reaction intermediates
was CO2 instead of methanol, which changed the ligand
exchange step. We also investigated the hydrogen pressure
effect on reaction rate. As shown in Fig. 9, it seems that
When Ru–[(p-OCF3)–BINAP] was subjected to dense
CO2, it dissolved in CO2 and the solution appeared homo-
geneous and had a yellowish color, in contrast with a col-
orless solution, and lots of solids stacked on the reactor
surface when conventional Ru–BINAP complex was used.
Further investigations showed that Ru–[(p-OCF3)–BINAP]
had sufficient solubility for catalysis in CO2. Therefore, hy-
drogenation of tiglic acid with Ru–[(p-OCF3)–BINAP] was
also investigated in CO2.
Preliminary experiments showed that the reactions pro-
ceeded smoothly in CO2, but with a rather low enantiose-
lectivity (around 25∼35%, which depended on the reaction
0.10
0.25
8.6 o
23.5 o
35.1 oC
C
30 bar
50 bar
0.09
C
0.20
0.15
0.10
0.05
0.00
0.08
0.07
0.06
0.05
0.04
0
1
2
3
4
5
0
2
4
6
8
10
12
Time / hr
Time / hr
Fig. 8. Temperature effect on the tiglic acid hydrogenation reac-
tions in methanol catalyzed by [BINAP]Ru[O2CCH3]2 {H2 = 95.3 bar,
[catalyst] = 0.3 mM, [tiglic acid]0 = 0.125 M}.
Fig. 9. Hydrogen pressure effect on the tiglic acid hydrogenation reactions
in CO2 by [(p-OCF3)–BINAP]Ru[O2CCH3]2 {[tiglic acid]0 = 92.5 mM,
[catalyst] = 0.24 mM, P (total) = 166 bar, methanol = 1 ml, T = 25 ◦C}.

X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
79
The enantioselectivity remained almost the same even the
catalyst concentration was increased by a factor of three,
which is shown in Table 2.
0.25
0.20
0.15
0.10
0.05
0.00
37.0mM
92.5mM
185.5 mM
4. Discussion
It has been proposed that the tiglic acid hydrogenation
by Ru–BINAP proceeds through a monohydride mecha-
nism [16,18]. According to Ashby and Halpern [16], the
reaction is first order in hydrogen at sufficiently low hydro-
gen pressures when the heterolytic split of hydrogen by the
ruthenium–olefin complex is the rate determining step. At
sufficiently high hydrogen pressures, hydrogen will trap ev-
ery molecule of the complex and the reaction of ruthenium
complex with olefin becomes the rate determining step and
the reaction is zero order in hydrogen. At intermediate val-
ues of hydrogen pressure, the order will gradually change
from 1.0 to 0.0. The order of 0.6 observed indicates that
the experiments carried out in this study may fall into this
intermediate region.
A mechanism which includes an additional hydride route
to the olefin route proposed by Ashby and Halpern [16] is
also a possibility and this mechanism is depicted in Fig. 12.
These two routes were also proposed for the hydrogenation
of vinylcarboxylic acid derivatives catalyzed by the conven-
tional Ru–BINAP complex [19]. In this study, the hydride
route was dominant at high hydrogen pressures and con-
sisted of the following elementary steps:
0
2
4
6
8
10
12
14
Time / hr
Fig. 10. Tiglic acid concentration effect on the tiglic acid hydrogenation
reactions in CO2 by [(p-OCF3)–BINAP]Ru[O2CCH3]2 {H2 = 30 bar,
[catalyst] = 0.24 mM, P (total) = 166 bar, methanol = 1 ml, T = 25 ◦C}.
high hydrogen pressures speeded up the formation of active
species.
Substrate concentration effects in CO2 were investigated
by varying the initial tiglic acid concentration from 37.0 to
186.5 mM. The results, as shown in Fig. 10 and Table 2,
indicate that the substrates with 2 carboxyl groups had a
smaller effect on reactions in CO2 than those in methanol.
Interestingly, the substrate concentration was found to have
a positive effect on enantioselectivity, which is shown in
Table 2. The effect of catalyst concentration on reaction rate
was also investigated by varying the catalyst concentration
from 0.08 to 0.37 mM. As shown in Fig. 11, conversion
was found to increase with increasing catalyst concentration,
however the increase was not linear as would be expected.
1. the formation of a ruthenium monohydride intermediate,
2. the subsequent coordination of tiglic acid to the mono-
hydride intermediate (which is the rate-determining step
under a high hydrogen pressure),
3. hydride migration to the olefin,
4. oxidative addition of H2 and reductive elimination of the
product.
90
The olefin route, on the contrary, was dominant at low
hydrogen pressures. In this case, the hydrogenation in-
volved first coordination of tiglic acid to form the active
complex by ligand exchange reactions where the carboxy-
late group was replaced by the tiglic acid. Subsequently,
tiglic acid–Ru-hydride intermediate formed which is the
rate-determining step under a low hydrogen pressure. Based
on the references [16,18,20], the hydrogen migration step
might be fast and the Ru–C bond can be cleaved by H2
and/or methanol.
It is important to note that there was an induction period
in the reactions in methanol catalyzed by Ru–[(p-OCF3)–
BINAP]. The induction period is usually an indication of
expulsion of a ligand in order to provide a catalytically
active site. The occurrence of an induction period indicates
that the bond strength between the ruthenium and the car-
boxylate group increases with modification of the BINAP
ligand, which is in agreement with the analysis of electronic
properties of modified BINAP ligand and parent BINAP
0.08 mM
0.24 mM
80
70
60
50
40
0
2
4
6
8
10
12
14
Time / hr
Fig. 11. Catalyst concentration effect on the tiglic acid hydrogenation
reactions in CO2 by [(p-OCF3)–BINAP]Ru[O2CCH3]2{H2 = 50 bar,
[tiglic acid]0
T = 25 ◦C}.
=
92.5 mM, P (total)
=
166 bar, methanol
=
1 ml,

80
X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
O
O
P
H+
H+
Ru
O
P
O
fast
product
product
O
O
H
P
P
H+
COOH
Ru
O
H2
O
r.d.s. at
low P
r.d.s. at
high P
O
O
P
O
O
H
O
P
P
Ru
Ru
P
O
O
O
O
P
P
O
O
Ru
H+
O
H2
COOH
Fig. 12. Possible reaction pathway for hydrogenation of tiglic acid using ruthenium–BINAP complex.
ligand (31P NMR and H NMR spectra indicate that the
OCF3 groups withdraw the electrons from the phosphine).
Thus, more time is needed to replace the carboxylate groups
at the beginning of the reaction. Once reaction starts, the re-
placed carboxylate group will not participate in the catalytic
cycle. This is in agreement with the proposed mechanism.
Though the effect of hydrogen pressure on enantioselec-
tivity for Ru–BINAP is similar for Ru–[(p-OCF3)–BINAP],
a more detailed comparison between the two systems reveals
some interesting differences. At low hydrogen pressures,
the enantioselectivities obtained with both of the complexes
were similar. However, for Ru–BINAP, the enantioselectiv-
ity dropped sharply (from 90 to 50%) when the hydrogen
pressure increased from 3.4 to 95.3 bar, whereas the decrease
in enantioselectivity (from 89 to 77%) was much less for
Ru–[(p-OCF3)–BINAP], as shown in Fig. 4. This probably
results from the fact that Ru–[(p-OCF3)–BINAP] complex
might stabilize the monohydride species much better than
Ru–BINAP complex due to a stronger ruthenium hydride
bond. If one considers that the absolute configuration of the
product is determined only by the step associated with the
formation of five-membered ring of metal-alkyl intermedi-
ate, the stronger bond strength would provide more time for
the coordinated tiglic acid to rearrange to the right position
1
before the hydrogen migration to the olefin bond. As a re-
sult, the enantioselectivity with Ru–[(p-OCF3)–BINAP] is
higher than that with conventional Ru–BINAP at high hy-
drogen pressures.
The reaction in CO2 might follow a different path. As
stated in Section 3, with the addition of a small amount of
methanol such as 1 ml, the increase in enantioselectivity was
very sharp (from 30 to 56%). It was also noticed that the
substrate concentration had no effect on enantioselectivity
in methanol but had a positive effect in CO2. The results can
probably be attributed to the change of solvent properties.
Increasing the acid or methanol concentration increases the
polarity of the CO2 and makes it more like a protic solvent.
This helps with the dissociation of metal-alkyl intermediate
bond and changes the equilibrium distribution of the enan-
tiomer intermediates [18,20].
5. Conclusions
Fluorinated Ru–BINAP catalyst was developed with
OCF3-substitution of the aryl groups in BINAP skeleton.
The properties of the catalyst in the tiglic acid hydrogenation
reactions were investigated in the conventional solvent

X. Dong, C. Erkey / Journal of Molecular Catalysis A: Chemical 211 (2004) 73–81
81
methanol as well as in dense carbon dioxide. The effects of
fluorous groups that were incorporated to the conventional
BINAP on the catalysts were also investigated. The main
conclusions are as follows:
Professor Rich Given of the University of Kansas for very
helpful discussions.
References
1. The rate expressions for tiglic acid hydrogenation re-
actions in the presence of both the fluorinated catalyst
and the conventional catalyst in methanol were derived.
The orders for both hydrogen and tiglic acid were sim-
ilar for both catalysts. The reaction rate constant with
Ru–[(p-OCF3)–BINAP] was lower than that with con-
ventional catalyst due to the electron withdrawing prop-
erty of incorporated OCF3 groups.
2. A mechanism for reactions in methanol, which involves
two routes, was proposed. The mechanism doesn’t seem
to change with modification of the catalyst. The similar
orders obtained for hydrogen and tiglic acid for both
catalysts seems to substantiate this.
[1] S.C. Stinson, Chem. Eng. News 1 (2001) 79.
[2] P.G. Jessop, W. Leitner (Eds.), Chemical Synthesis Using Supercrit-
ical Fluids, 1999.
[3] J. Dupont, R.F. de Souza, P.A.Z. Suarez, Chem. Rev. 102 (2002)
3667.
[4] I.T. Horvath, J. Rabai, Science 266 (1994) 72.
[5] M.H. Abraham, A.H. Zissimos, J.G. Huddleston, H.D. Willauer, R.D.
Rogers, W.E. Acree, Ind. Eng. Chem. Res. 42 (2003) 413.
[6] G. Pozzi, F. Cinato, Chem. Commun. 8 (1998) 877.
[7] Y. Nakamura, S. Takeuchi, K. Okumura, Y. Ohgo, D.P. Curran,
Tetrahedron 58 (2002) 3963.
[8] M. Cavazzini, A. Manfredi, F. Montanari, S. Quici, G. Pozzi, Eur.
J. Organic Chem. 24 (2001) 4639.
[9] S. Takeuchi, Y. Nakamura, Y. Ohgo, D.P. Curran, Tetrahedron Lett.
39 (1998) 8691.
[10] G. Francio, K. Wittmann, W. Leitner, J. Organomet. Chem. 621
(2001) 130.
[11] D.R. Palo, C. Erkey, Organometallics 19 (2000) 81.
[12] S. Fujita, S. Fujisawa, B.M. Bhanage, Y. Ikushima, M. Arai, New
J. Chem. 26 (2002) 1479.
3. At low hydrogen pressures, the enantioselectivity with
Ru–[(p-OCF3)–BINAP] was almost the same as that
with Ru–BINAP. At higher hydrogen pressures, the
enantioselectivity with Ru–[(p-OCF3)–BINAP] complex
was much higher.
4. The modified fluorinated catalyst had a much higher sol-
ubility in dense CO2 than the conventional catalyst.
5. CO2 had a great influence on both the activity and enan-
tioselectivity.
6. Addition of methanol to CO2 was found to increase the
enantioselectivity.
[13] S. Haji, C. Erkey, Tetrahedron 58 (2002) 3929.
[14] J.J. Juliette, I.T. Horvath, J.A. Gladysz, J. Am. Chem. Soc. 121
(1999) 2696.
[15] S. Akutagawa, Appl. Catal. A 128 (1995) 171.
[16] M.T. Ashby, J. Halpern, J. Am. Chem. Soc. 113 (1991) 589.
[17] K. Radhakrishnan, P.A. Ramachandran, P.H. Brahme, R.V. Chaud-
hari, J. Chem. Eng. Data 28 (1983) 1.
[18] T. Ohta, H. Takaya, R. Noyori, Tetrahedron Lett. 31 (1990)
7189.
[19] A.S.C. Chan, C.C. Chen, T.K. Yang, J.H. Huang, Y.C. Lin, Inorg.
Chim. Acta 234 (1995) 95.
Acknowledgements
[20] M. Kitamura, M. Tsukamoto, Y. Bessho, M. Yoshimura, U.
Kobs, M. Widhalm, R. Noyori, J. Am. Chem. Soc. 124 (2002)
6649.
The authors thank Dr. Martha Morton of University
of Connecticut and Professor Richard L. Schowen and