Tunable ferrocenyl diphosphine ligands for the Ir-catalyzed enantioselective hydrogenation of N-aryl imines

Article abstract of DOI:10.1016/S0022-328X(00)00766-X

Ferrocenyl diphosphines R₂PF-P(R′)₂ are effective, tunable ligands for the iridium catalyzed enantioselective hydrogenation of N-aryl imines in the presence of iodide and acid promoters. Structure-activity/selectivity correlations were found for the hydrogenation of N-(2-ethyl-6-methylphenyl)-N-(1′-methoxymethyl)-ethylidene-amine (MEA-imine) and for 2,3,3-trimethylindolenine (TMI). Extremely high catalytic activity and moderate to good enantioselectivity were observed for the MEA imine using a catalyst generated in situ from [Ir(cod)Cl]₂ and (R)-(S)-PPF-P(3,5-Xyl)₂ (xyliphos). With the same type of catalysts, several other N-aryl imines can be hydrogenated with enantioselectivities between 31 and 96%, albeit with lower catalyst activities.

Full text of DOI:10.1016/S0022-328X(00)00766-X

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 621 (2001) 34–38
Tunable ferrocenyl diphosphine ligands for the Ir-catalyzed
enantioselective hydrogenation of N-aryl imines
Hans-Ulrich Blaser *1, Hans-Peter Buser 2, Robert Ha¨usel, Hans-Peter Jalett,
Felix Spindler *3
Sol6ias AG, Katalyse Forschung WRO-1055.630, Klybeckstr. 191, PO Box, CH-4002 Basel, Switzerland
Received 28 August 2000; accepted 22 September 2000
On the occasion of the 65th birthday of Professor Henri Brunner under the motto ‘No question: What’s right is right!’
Abstract
Ferrocenyl diphosphines R2PF-P(R%)2 are effective, tunable ligands for the iridium catalyzed enantioselective hydrogenation of
N-aryl imines in the presence of iodide and acid promoters. Structure–activity/selectivity correlations were found for the
hydrogenation of N-(2-ethyl-6-methylphenyl)-N-(1%-methoxymethyl)-ethylidene-amine (MEA-imine) and for 2,3,3-trimethylindo-
lenine (TMI). Extremely high catalytic activity and moderate to good enantioselectivity were observed for the MEA imine using
a catalyst generated in situ from [Ir(cod)Cl]2 and (R)-(S)-PPF–P(3,5-Xyl)2 (xyliphos). With the same type of catalysts, several
other N-aryl imines can be hydrogenated with enantioselectivities between 31 and 96%, albeit with lower catalyst activities.
© 2001 Elsevier Science B.V. All rights reserved.
Keywords: Enantioselective hydrogenation; N-Aryl imines; Ferrocenyl diphosphines; Ir-diphosphine catalysts; Structure–activity correlation
1. Introduction
tage of a new class of ferrocenyl diphosphine ligands
(see Fig. 1) developed by Togni and Spindler [4] is the
fact that the two phosphine groups can be introduced
sequentially. This means that a large number of differ-
ent ligands can be prepared with a relatively small
synthetic effort. In addition, their electronic and steric
properties can be varied easily and thereby be tuned to
the needs of a specific transformation. In this communi-
cation we describe the results of a structure–activity/se-
lectivity study for the iridium-catalyzed hydrogenation
of several N-aryl imines (see Fig. 2).
The technical application of catalytic enantioselective
reactions is slowly gaining acceptance in the fine and
specialty chemicals industry [1,2]. One major problem is
the often very pronounced product specificity which
means that the structure of the catalyst, usually a metal
complex with chiral ligands, has to be tuned or tailored
for each specific transformation [3]. Up to now the
search for the optimal catalyst is a very empirical
process where screening strategies and the intuition of
the chemist are the dominant elements. A major hurdle
in this endeavor is the difficult access to many chiral
phosphines, because they are not available commer-
cially and/or difficult to prepare. A particular advan-
1 *Corresponding author. Tel.: +41-61-6866161; fax: +41-61-
6866311; e-mail: hans-ulrich.blaser@solvias.com
2 Novartis Crop Protection.
3 *Corresponding author. Tel.: +41-61-6866161; fax: +41-61-
6866311; e-mail: felix.spindler@solvias.com
Fig. 1. Structure of the ferrocenyl diphosphine ligands and abbrevia-
tions used in this study.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00766-X

H.-U. Blaser et al. / Journal of Organometallic Chemistry 621 (2001) 34–38
35
Fig. 2. Structure of imine substrates used in this study.
Table 1
Hydrogenation of MEA imine 1 catalyzed by Ir–R2PF–PR2% catalysts: aryl versus alkyl substituents and effect of configuration (for R and R% see
Fig. 1) a
b
Entry
Ligand
R
Time
(h)
Conv.
(%)
Tof
ee
Comments
(h−1
)
(%)
R%
1.1
1.2
1.3
1.4
1.5
1.6
1.7
Ph
Ph
Ph
4-CF3–Ph
4-CF3–Ph
Ph
tBu
Cy
Cy
Cy
Ph
3,5-Xyl
3,5-Xyl
18
18
18
18
18
2
6
low
65
80
100
99
3
n.d.
29
18
44
n.d.
47
61
21
81
c
Me5-cp derivative
396
41
79
Xyliphos
(R)-(R)-configuration
Ph
18
92
39 d
a Reaction conditions: 50 ml autoclave, 24 mmol MEA-imine, 10% v/v acetic acid, 150 mg TBAI, s/c 800. 25 bar, 25°C, catalyst:
[Ir(cod)Cl]2/ligand.
b Excess (S)-amine.
c Solvent: MeOH/toluene, no acetic acid.
d Excess (R)-amine.
Table 2
Hydrogenation of MEA imine 1 catalyzed by Ir-(R)-(S)-R2PF–PR2% catalysts: effect of the nature of R and R% on ee (%, bold) and maximum rate
(mmol min−1) (definition of R and R%, see Fig. 1) a
R%/R
3,5-Xyl
3-tol
Ph
4-MeO–Ph
76/0.6
3-MeO–Ph
3,5-(CF3)2–Ph
38/0.01
4-CF3–Ph
4-CF3–Ph
Ph
3,5-(CF3)2–Ph
2-Naphth
3,5-Xyl
82/0.4
80/0.8
79/0.02
80/0.7
76/0.8
81/0.1
78/0.7
81/0.01
77/0.6
72/0.7
69/0.01
d
79/0.6
80/0.6
74/0.5
n.d./B0.01
71/n.d.
76/ b
49/0.02
Additional results: R=Ph, R%=4-tBu–Ph: 82 c/0.7; R=Ph, R%=3,5-(SiMe3)2–Ph: 64/0.6
a Reaction conditions: 50 ml autoclave, 24 mmol MEA-imine, 10% v/v acetic acid, 150 mg TBAI, s/c 800. 25 bar, 25°C, catalyst:
[Ir(cod)Cl]2/ligand.
b Conditions as in Table 3.
c 87% ee at −15°C.
d Solvent MeOH/toluene instead of acetic acid.
2. Hydrogenation of MEA imine 1
activity data only partially meaningful. Therefore, ex-
periments for the final ligand optimization were carried
out in a well mixed 300 ml autoclave (Table 3).
The effect of the structure of the diphosphine ligand
on the hydrogenation of MEA imine 1 was investigated
in great detail using catalysts generated in situ from
[Ir(cod)Cl]2, the respective ligand and in presence of
acid and iodide [5]. The first test series were carried out
in a 50 ml autoclave stirred with a stirring bar at an s/c
ratio of 800 at 25°C and 25 bar hydrogen pressure
(Tables 1 and 2). Later, we found that the faster
reactions were probably diffusion limited, making the
The results of a preliminary study as shown in Table
1 can be summarized as follows: (i) Only ferrocenyl
diphosphine ligands gave catalysts with medium to
good ee values and satisfactory catalyst stability (results
with other phosphine classes not shown). (ii) The two
elements of chirality must be matched for a good
catalyst performance and the absolute configuration of
the planar chirality element determines the sense of

36
H.-U. Blaser et al. / Journal of Organometallic Chemistry 621 (2001) 34–38
Table 3
Hydrogenation of MEA imine 1 catalyzed by Ir-ferrocenyldiphosphine catalysts: Effect of the nature of the 4-substituent in 3,5 dialkyl-phenyl
groups (for R and R% see Fig. 2) a
Entry
Ligand
R
Time
(h)
Conv.
(%)
Tof
ee
(%)
(h−1
)
R%
b
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
4-nPr2N-3,5-Xyl
4-nPe2N-3,5-Xyl
4-Me2N-3,5-Xyl
3,5-Xyl
4-Bn2N-3,5-Xyl
4-Me2N–Ph
4-(N-Pyr)-3,5-Xyl
4-Bn2N-3,5-iPr2–Ph
4-Me2N-3,5-iPr2–Ph
3.5
26
1
0.6
21
79
3
100
85
100
100
92
23
100
99
28 000
3300
100 000
167 000
4400
3400
33 000
4500
83
83
80
76
76
74
69
64
61
22
22
98
4500
a Reaction conditions: 300 ml autoclave, 0.51 mol MEA-imine, 10% v/v acetic acid, 70 mg TBAI, s/c 100,000, 80 bar, 50°C, catalyst:
[Ir(cod)Cl]2/ligand.
b 87% at −15 °C.
induction. With (R)-(S)-xyliphos, the (S)-enantiomer
was obtained with high activity and with 79% ee. (entry
1.6). The diastereomeric ligand (R)-(R)-xyliphos pro-
duced 39% ee of the (R)-enantiomer with a much lower
activity (entry 1.7). This effect might be due to different
conformations of the two diastereomeric Ir-xyliphos
complexes. (iii) Aryl groups at the two phosphorus
were essential, when the R% group was alkyl both ee and
tof dropped very strongly (entries 1.1–1.4, compared
with entries 1.5 and 1.6). With these results in hand, we
concentrated our efforts on the variation the P-aryl
R% we chose a number of 4-amino substituted 3,5-xylyl
derivatives. This choice was based on further indica-
tions from process development work that the 3,5-xylyl
group had a significant accelerating effect and we inter-
preted the results with R=4-tBu–Ph on an electronic
basis. These tests were carried out with an s/c ratio of
100 000 at 50°C and 80 bar under conditions were rates
were not diffusion controlled and the results are com-
pared in Table 3. Indeed, we found combinations that
led to increased enantioselectivities compared to
xyliphos (entries 3.1–3.3, compared to 3.4 and 3.6),
thereby confirming the positive effect of the 3,5-methyl
groups and of an electron donating 4-substituent. The
nature of the amine substituents had a significant but
unsystematic effect, with N-propyl and N-pentyl as
optimum. The effect was negative for a 4-(N-Pyr) (en-
try 3.7) and for bulkier groups in the 3,5 position
(entries 3.8, 3.9). Unfortunately, the catalyst activities
of all catalysts were significantly lower than for Ir/
xyliphos. For this as well as for preparative reasons,
xyliphos was chosen as ligand for the commercial pro-
cess [5].
groups in the (R)-(S)-R2PF–PR% series.
2
The results of a systematic variation of the R and R%
groups are summarized in Table 2. Except for electron
deficient R% groups, the enantioselectivities of the differ-
ent combinations varied within narrow limits between
74 and 82% ee. A comparison of the various combina-
tions indicated some relatively weak trends. For phenyl
substituents in the R position at the cyclopentadienyl
ring, the presence of CF3 groups slightly increased the
enantioselectivity, methyl groups had a negative effect
on ee, a 2-naphthyl group gave the same ee as the
phenyl group. Stronger effects were observed for the R%
position at the stereogenic center where CF3 groups
significantly decreased the enantioselectivity while
methyl groups in meta position slightly increased the ee
values. The highest ee values in this screening series
were observed with R=4-tBu–Ph (87% at −15°C).
The variations in initial rate were much stronger and
especially the presence of trifluoromethyl groups both
in the R and R% position led to a significant rate
decrease. Unfortunately, the diffusion problems men-
tioned above precluded the assessment of the effect of
strongly activating substituents (limit for initial rate ca.
0.8 mmol min−1).
3. Hydrogenation of other imines
A rather broad ligand screening was also carried out
with trimethylindolenine (TMI) 2 used as model for
cyclic N-aryl imines with fixed CꢀN geometry. In Table
4 results are sorted according to ee values. Compared
to MEA imine 1, enantioselectivities for 2 were signifi-
cantly higher, whereas tofs were much lower (no diffu-
sion problems here). Both values varied less for
different substitution patterns. As for MEA imine 1,
ligands with basic substituents showed almost no activ-
ity, so that the ee was not determined. In contrast to
the results for MEA imine 1, both 3,5-CF3 groups as
For the final ligand optimization we decided to keep
the unsubstituted phenyl group in the R position. For

H.-U. Blaser et al. / Journal of Organometallic Chemistry 621 (2001) 34–38
37
Table 4
Enantioselective hydrogenation of TMI 2 catalyzed by Ir-ferrocenyldiphosphine catalysts: effect of ligand structure (for R and R% see Fig. 1) a
Ligand
R
p
T
(°C)
Time
(h)
Conv.
(%)
Tof
ee
(%)
Comments
(bar)
(h−1
)
R%
3,5-Xyl
3,5-(CF3)2–Ph
4-MeO–Ph
Ph
Ph
Ph
Ph
Ph
3,5-Xyl
3,5-Xyl
3,5-Xyl
4-Me2N-3,5-Xyl
3,5-Xyl
40
80
40
80
40
80
80
80
80
40
40
40
40
30
15
30
50
30
50
50
50
50
30
30
30
30
47
18
17
17
23
18
17
20
17
68
18
24
33
100
99
100
96
100
99
100
100
98
29
6
5
14
15
56
11
14
15
13
14
93
89
88
88
87
86
85
80
94% ee at 5 and 15°C
91% ee at 15°C
4-tBu–Ph
4-Pr2N-3,5-iPr2–Ph
4-Pr2N-3,5-Xyl
Ph
3,5-(CF3)2–Ph
76
Ph
Ph
Ph
Cy
2-MeO–Ph
Cy
1
n.d.
n.d.
n.d.
n.d.
B1
B1
B1
tBu
Cy
5
2
a Reaction conditions: 50 ml autoclave, 3.1 mmol TMI, 10 ml toluene, 0.05 ml CF3COOH, 12 mg TBAI, s/c 250, catalyst: [Ir(cod)Cl]2/ligand.
Table 5
Enantioselective hydrogenation of various N-aryl imines catalyzed by Ir-(R)-(S)-R2PF–PR2% catalysts: effect of ligand structure on ee (best results
a
bold, for R and R% see Fig. 2)
R/R% imine
Ph/3,5-Xyl (xyliphos)
Ph/4-CF3–Ph
Ph/Ph
78
3,5-Xyl/3,5-Xyl
3,5-(CF3)2–Ph/Ph
MEA imine 1
TMI 2
Imine 3
80
87
76
71
76
93
74
81
76
78
Imine 4
Imine 5
31 (R)
67
24 (S)
96
a Reaction conditions: MEA imine 1: see Tables 1–3. TMI 2: see Table 4. Imine 3: 5 mmol imine; 5 ml toluene, 2 ml CH3COOH, 30 mg TBAI,
s/c: 200, T: 25°C, p(H2): 30 bar. Imine 4: 2.47 mmol imine; 8 ml toluene, 2 ml CH3COOH, 30 mg TBAI, s/c: 100, T: 25°C, p(H2): 40 bar. Imine
5: 2.23 mmol imine; 8 ml toluene; 0.03 ml CF3COOH, 12 mg TBAI, s/c 200, T: 25°C, p(H2): 80 bar.
4. Comparison and conclusions
well as electron donating substituents in the 4-position
led to lower ee values for 2. However, the positive
3,5-xylyl effect was more pronounced here as reported
for several other catalytic reactions with aryl phosphine
ligands [6].
The enantioselectivities with selected Ir–ligand com-
plexes for the various imines are compared in Table 5.
With the exception of imine 3, the Ir–ferrocenyldiphos-
phine catalysts with aryl-P substituents gave reasonable
to very high enantioselectivities for all N-aryl imines
tested. However, the catalyst activities varied drastically
and the optimal combination of R and R% groups on
the ferrocenyl backbone was different for each sub-
strate. At the moment, we are unable to give an explica-
tion for these findings. However, it is obvious that
especially the optimal electronic needs seem to be quite
different for each N-aryl imine.
In conclusion, this new class of ferrocenyl diphos-
phine ligands is very well suited for the Ir catalyzed
hydrogenation of N-aryl imines with moderate to very
high enantioselectivities and low to very high activities
and turnover numbers. Clearly, the modular character
of the ligands was a prime success factor for finding the
optimal catalyst with a reasonable synthetic effort.
Only very few experiments were carried with imines
3–5. Imine 3, a potential intermediate for a herbicide
originally developed by Sandoz [7] is closely related to
MEA imine 1. Imine 4 is an intermediate for the
fungicide (S)-metalaxyl [8], while imine 5 is a model
substrate. Whereas 3 could be hydrogenated with ee
values up to 78% (best ligand R, R%=Ph), imine 4 is
clearly not a good substrate for the Ir–ferrocenyl
diphosphine catalyst with best ee values of 31% for
xyliphos (with a different absolute configuration for the
two best ligands!). Somewhat unexpectedly, the highest
enantioselectivities were reached for imine 5 where the
PPF–P(4-CF3–Ph)2 gave 96% ee. In all cases, signifi-
cantly lower activities were observed (results not
shown).

38
H.-U. Blaser et al. / Journal of Organometallic Chemistry 621 (2001) 34–38
5. Experimental
5.1. Materials
5.2.1. Procedures A (with sol6ents)
Two solutions containing the substrate and the cata-
lyst (generated from [Ir(cod)Cl]2, the diphosphine, the
tetrabutylammonium iodide and (if stated) acetic acid),
respectively, were prepared under an argon atmosphere,
for each using half of the total solvent volume. The two
solutions were transferred via steel capillary into the
autoclave, which was set under argon before. After
sealing the autoclave the reaction was started immedi-
ately as described above.
[Ir(COD)Cl]2 was purchased from Johnson-Matthey,
2,3,3-trimethylindoleneine from Aldrich. The imines 1,
3, 4 and 5 were prepared by reacting the 2,6-dialkylani-
line with the corresponding ketone according to stan-
dard procedures.
The ferrocenyl diphosphine ligands were synthesized
according to [4]: 1-ferrocenylethyl-N,N-dimethylamine
was reacted with the appropriate P-chlorophosphine
ClPR2 to yield the corresponding 1-(2-phosphinoferro-
cenyl)ethyl-N,N-dimethylamine which after treatment
5.2.2. Procedures B (without sol6ents)
The appropriate amount of imine 1 was placed in a
50 ml stainless steel autoclave. Then, [Ir(cod)Cl]2, the
diphosphine, the tetrabutylammonium iodide and acetic
acid were added, and the autoclave was sealed and set
under argon. The reaction was started immediately as
described above.
with one equivalent of the sec.-phosphine HPR% in
acetic acid was converted to the 1-(2-phosphinoferro-
cenyl)-ethylphosphine (R2PF–PR%). ClPR2 and HPR%
2
2
2
were obtained either from commercial suppliers or pre-
pared in analogy to reported procedures [9].4
5.3. Analytical procedures
5.2. Typical hydrogenation procedure
The conversion of the crude reaction product was
determined by glc (column: DB 17/30W, 15 m (JCW
Scientific Inc.), temperature program: 60°C/1 min to
220°C, DT: 10° min−1). The enantiomeric excess of the
reaction product was determined by HPLC as summa-
rized in Table 6.
Hydrogenation reactions were performed in a well
stirred stainless steel autoclave (50 or 300 ml capacity).
Imine, the catalyst and the other reaction components
were placed in the autoclave (details see below); the
autoclave was sealed, flushed three times with hydrogen
pressurizing typically to 20 bars at ambient temperature
and venting the gas back, and finally charged with
hydrogen to the specified reaction pressure. After the
specified time at the indicated temperature, the auto-
clave was opened and after removal of the solvent the
conversion and ee were determined as described below.
Acknowledgements
We would like to thank Beat Eng, Markus Fischer,
Heidi Landert, Ulrich Pittelkow, Genevie`ve Thoma,
Nadia Vostenka and Andrea Holderer for their very
careful and skillful experimental work.
Table 6
References
Details for determination of enantiomeric purity of hydrogenation
products by HPLC
[1] H.U. Blaser, B. Pugin, F. Spindler, in: B. Cornils, W.A. Her-
rmann (Eds.), Applied Homogeneous Catalysis by Organometallic
Complexes, VCH, Weinheim, 1996, p. 992.
Imine
Column
Conditions/retention times
[2] H.U. Blaser, F. Spindler, M. Studer, Enantioselective catalysis in
fine chemicals production, Applied Catalysis (accepted for
publication).
[3] E.N. Jacobsen, H. Yamamoto, A. Pfaltz, Comprehensive Asym-
metric Catalysis, Springer, Berlin, 1999.
[4] A. Togni, C. Breutel, A. Schnyder, F. Spindler, H. Landert, A.
Tijani, J. Am. Chem. Soc. 116 (1994) 4061.
[5] For a case history of the discovery and the development of the
technical process see: H.U. Blaser, H.P. Buser, K. Coers, R.
Hanreich, H.P. Jalett, E. Jelsch, B. Pugin, H.D. Schneider, F.
Spindler, A. Wegmann, Chimia 53 (1999) 275.
1
2
3
4
5
Chiralcel OD
T: 25°C, hexane/2-propanol:
99.7:0.3/9.9 min (R), 10.9 min (S)
Chiralcel OD-H T: 25°C, hexane/iPrOH: 97:3/11.4 min
(−), 12.6 min (+)
Chiralcel OD
T: 25°C, hexane/iPrOH: 99.9:0.1/82.2
min (R), 97.0 min (S)
Chiralcel OD-H T: 25°C, hexane/iPrOH: 99.8:0.2/19.9
min (R), 27.7 min (S)
Chiralcel AD
T: 20°C,
hexane/(hexane/iPrOH/HNEt2
975:25:1) 70:30/8.7 min (1st
enantiomer), 9.2 min (2nd
enantiomer)
[6] G. Trabesinger, A. Albinati, N. Feiken, R.W. Kunz, P.S.
Pregosin, M. Tschoerner, J. Am. Chem. Soc. 119 (1997) 6315.
[7] K. Seckinger, R. Chollet, S. Blarer, T. Vettiger, US 5457085
(1995) assigned to Sandoz Ltd.
[8] H.U. Blaser, F. Spindler, Topics Catal. 4 (1997) 275.
[9] S. Kainz, A. Brinkmann, W. Leitner, A. Pfaltz, J. Am. Chem.
Soc. 121 (1999) 6421.
4 Reduction of the P-chlorphosphines by LiAlH4 afforded the
corresponding sec.-phosphines.