More than 100,000 Turnovers with Immobilized Ir-Diphosphine Catalysts in an Enantioselective Imine Hydrogenation

Article abstract of DOI:10.1002/1615-4169(200210)344:9<974::AID-ADSC974>3.0.CO;2-Z

A modular concept to prepare immobilized enantioselective catalysts is described, consisting of a functionalized xyliphos ligand covalently attached to a support via a linker. Immobilized xyliphos bound to silica and to polystyrene as well as soluble dime

Full text of DOI:10.1002/1615-4169(200210)344:9<974::AID-ADSC974>3.0.CO;2-Z

FULL PAPERS
More than 100,000 Turnovers with Immobilized Ir-Diphosphine
Catalysts in an Enantioselective Imine Hydrogenation
Benoit Pugin,* Heidi Landert, Felix Spindler, Hans-Ulrich Blaser
Solvias AG, R 1055.6, P.O. Box, 4002 Basel, Switzerland
Fax: ( ‡ 41)-61-6866311, e-mail: benoit.pugin@solvias.com
Received: May 15, 2002; Accepted: July 18, 2002
Dedicated to Roger Sheldon, a chemist with an amazing E(nthusiasm) factor, on the occasion of his 60th
birthday.
Abstract: A modular concept to prepare immobilized enantioselectivities but lower activities and higher
enantioselective catalysts is described, consisting of a deactivation rates than the homogeneous analogues.
functionalized xyliphos ligand covalently attached to These negative effects were tentatively explained by
a support via a linker. Immobilized xyliphos bound to the higher local catalyst concentration on the support
silica and to polystyrene as well as soluble dimeric surface leading to an increased tendency to deactiva-
xyliphos and an extractable analogue were prepared tion by irreversible dimer formation. Separation of
and tested in the Ir-catalyzed hydrogenation of a these catalysts by filtration and extraction is easy and
hindered N-arylimine used for the production of (S)- efficient.
metolachlor. The best heterogeneous catalyst 9b
exhibited TON×s >100,000 and TOF×s up to Keywords: enantioselective imine hydrogenation;
20,000 hÀ1, the best values so far for immobilized extractable Ir-xyliphos; immobilized Ir-xyliphos; mod-
catalysts. The immobilized catalysts gave similar ular immobilization concept; silica-bound xyliphos
Introduction
catalysts bound to both organic moieties and inorganic
solids.[2] Here we report on the immobilization of the Ir-
One obvious drawback of homogeneous catalysts is the xyliphos complex, a very active catalyst for the
problem of their separation from the reaction mixture. enantioselective hydrogenation of the sterically hin-
Attachment of the active metal complexes to insoluble dered MEA-imine (see Scheme 1). MEA-amine is an
or water soluble carriers is an attractive strategy intermediate for (S)-metolachlor, the most important
allowing separation by filtration or extraction.[1] How- herbicide of Syngenta, which is presently produced on a
ever, immobilization usually affects the catalytic per- >10,000 tons/year scale with the homogeneous com-
formance of the catalyst. Sometimes this effect is plex.[3] The investigations were carried out parallel to
positive, e.g., when catalyst deactivation via dimeriza- the development of the homogeneous process in order
tion is prevented by site isolation.[2a] Unfortunately, to have a further option for catalyst separation.
practical experience has shown that more often catalytic
activity and/or selectivity decrease with immobilization.
For technical applications, several requirements should
be met in order to make immobilized enantioselective
catalysts practically useful: Besides sufficient selectivity,
easy preparation or commercial availability, relatively
high activities (turnover frequency, TOF) and produc-
tivities (turnover number, TON) are required because
immobilization raises catalyst costs by a factor of 2 5.
The development of extremely efficient immobilized
catalysts is therefore of very high importance.
Different immobilization approaches are possible and
all have their advantages and disadvantages. We have
chosen the covalent attachment of the metal complex at
the chiral ligand via a suitable bifunctional linker and
have found this to be a very versatile strategy to obtain ligand.
Scheme 1. Structure of starting imine, products and xyliphos
974
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/02/34409-974 979 $ 17.50-.50/0
Adv. Synth. Catal. 2002, 344, No. 9

More than 100,000 Turnovers with Immobilized Ir-Diphosphine Catalysts
FULL PAPERS
Results
adequate mass transport. Earlier work with Rh com-
plexes immobilized on this support gave similar ee×s but
only about 1/10of the activity of their homogeneous
analogues.[2c] To immobilize 1 on polystyrene, the amino
groups of the aminomethylated polystyrene were first
reacted with an excess of 2,4-diisocyanato-toluene 4 to
Preparation of the Functionalized and Immobilized
Ligands
Over the last few years we have developed a modular give the reactive polymer 5, then the functionalized
system combining functionalized ligands with different xyliphos 1 was added. The remaining isocyanate groups
supports and isocyanate linkers in order to have a were reacted with an excess of ethanol to give the
systematic and quick access to a variety of immobilized immobilized ligand 6 with a loading of 0.116 mmol/g.
chiral catalysts.[2] We have applied this modular system
The di-xyliphos compound 8 was prepared by reaction
to immobilize xyliphos on silica gel and various organic of 1 with 1,6-diisocyanatohexane 7 in order to simulate a
moieties as depicted in Scheme 2.
high local catalyst concentration.[2a] Awater extractable
For the preparation of the ligand immobilized on xyliphos 12 was made by reacting xyliphos ligand 10
silica, we started with the functionalized xyliphos 1 carrying a chloroalkyl chain with diethyl malonate
carrying an amino function[4] which was reacted first followed by hydrolysis (see Scheme 3). Under basic
with 3-isocyanato-propyltrimethoxysilane 2 to give 3 conditions, the functionalized ligand 12 can be extracted
and then with silica to give 9 as described before.[2b] As into the water phase.
support we used silica Grace 332 that gave the best
results for Rh-catalyzed hydrogenations.[2b] Grace 332
has arelatively homogeneous pore size (average 19 nm), Hydrogenations
a pore volume of 1.55 mL/g and a high specific surface of
320m 2/g. Three immobilized xyliphos ligands 9a c with A standard protocol originally developed for the
loadings of 0.10, 0.042 and 0.013 mmol ligand/g support, homogeneous Ir-xyliphos was used where the hydro-
respectively, were synthesized.
genation reaction is carried out in presence of acetic acid
To prepare the polymer-bound ligand, we used and iodide[5] (see experimental section). Table 1 sum-
commercial aminomethylated polystyrene cross-linked marizes a number of control experiments with the
with 1% divinylbenzene. This support has to be swollen homogeneous Ir-xyliphos systems (xyliphos, function-
with an appropriate solvent (e.g., toluene) to allow alized xyliphos 3, dimeric xyliphos 8, water extractable
xyliphos 12). Table 2 shows the results with the silica and
polystyrene bound Ir catalysts with xyliphos-SiO2 9a c
and xyliphos-PS 6, respectively. In all cases, the TOF
values are calculated for the whole reaction time.
Discussion
The main objective of this work was to determine the
limits of the performance of separable catalysts and to
assess their potential in comparison with the homoge-
neous Ir-xyliphos system which is actually used in the
production of (S)-metolachlor. All catalysts were tested
with various substrate/catalyst ratios (s/c) between
20,000 and 120,000. In addition, some experiments
were carried out to identify factors that may influence
the catalytic performance.
Under our test conditions the homogeneous catalysts
can be used at s/c ratios up to 120,000 with TOF×s
Scheme 2. Functionalized xyliphos, linkers and immobilized
ligands.
Scheme 3. Water-soluble xyliphos ligand.
Adv. Synth. Catal. 2002, 344, 974 979
975

Benoit Pugin et al.
FULL PAPERS
Table 1. Control experiments with the homogeneous Ir-xyliphos systems.
Entry
Ligand
s/c[a]
TOF [hÀ1][b]
ee [%]
Comment
1.1
1.2
1.3
1.4
1.5
1.6
xyliphos
xyliphos
xyliphos
funct. xyliphos 3
extract. xyliphos 12
dimeric xyliphos 8
50,000
60,000
42,857
55,385
28,571
36,000
4,167
79
78
80
80
79
re-use[c]: ‡ 25,000 TON in 6.5 h
50,000
120,000
50,000
120,000
50,000
100 mg SiO2 added (Grace 332)
n.d.
re-use[c]: ‡ 3000 TON in 3.5 h
Reaction conditions: 19.5 g MEA-imine, 2 mL acetic acid, 10mg tetrabutylammonium iodide (TBAI), 80bar, 25 30 8C,
100% conversion.
[a]
Mol substrate/mol Ir.
Overall TOF.
[b]
[c]
Another portion of substrate was added to the hydrogenation mixture without separation of the catalyst.
Table 2. Hydrogenation with immobilized Ir-xyliphos systems.
Entry
Ligand
s/c[a]
TOF [hÀ1][b]
ee [%]
Comment
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
xyliphos-SiO2 9b
xyliphos-SiO2 9b
xyliphos-SiO2 9b
xyliphos-SiO2 9b
xyliphos-SiO2 9a
xyliphos-SiO2 9b
xyliphos-SiO2 9c
xyliphos-PS 6
20,000
20,000
120,000
250,000
50,000
50,000
50,000
50,000
50,000
10,000
20,000
12,000
9,750
2,146
6,000
1,732
1,140
1,500
77
76
78
75
77
78
77
73
70
ligand/Irˆ 2
ligand/Irˆ 2
ligand/Irˆ 2, 50 8C, 78% conv.
93% conversion
re-use[c]: no reaction in 2 h
97% conversion
95% conversion
50 8C
xyliphos-PS 6
Reaction conditions: 19.5 g MEA-imine, 2 mL acetic acid (5 mL toluene for xyliphos-PS), 10mg TBAI, 80bar, 25 30 8C,
100% conversion unless otherwise noted.
[a,b,c]
See Table 1.
between 29,000 and 60,000 hÀ1 and ee×s between 78 and observed for other catalytic systems, the polystyrene-
80%. While the TON and TOF values were significantly bound complexes (entries 2.8, 2.9) are much less active,
lower than those of the actual production process,[5] maybe because mass transport is slower in the polymer
these results serve well as standards for comparison. matrix than in the large pores of the silica gel. In
Some comments: The Ir-xyliphos catalyst was re-used contrast to the homogeneous reactions several hydro-
but with a significant loss in activity, furthermore the genations with the immobilized systems practically
addition of silica barely affected the catalyst activity stopped before complete conversion and re-use seems
(entries 1.1 and 1.2). While in the catalytic reactions the not to be possible (compare entries 1.1 and 2.6). These
two substituted xyliphos 3 and 12 gave similar activity as results indicated that the immobilized catalysts were
the parent ligand (entries 1.4 and 1.5), this was no longer more susceptible to deactivation than free Ir-xyliphos
the case for the dimeric analogue 8 (entry 1.6) which was catalyst and that they have a certain lifetime and
much less active both on first and on second use.
The most important result of Table 2 is the unprece-
gradually deactivate.
As already mentioned, we designed a number of
dented activity (TOF up to 20,000 hÀ1) and productivity experiments in order to find possible reasons for the
(TON up to 195,000) of the silica-bound catalysts lower activity (TOF×s <20,000 compared to 60,000 hÀ1
(entries 2.1 2.4). The highest TON×s/TOF×s reported for xyliphos) as well as the faster deactivation of the
in the literature for covalently attached diphosphine immobilized catalysts. The following changes go along
[1,6]
catalysts are in the range of 10,000/2000 hÀ1 for Rh
with the covalent binding of a homogeneous catalyst on
and of 33,000/400 hÀ1 for Ru.[7] Nevertheless, compared a support:
to the very active soluble catalysts, the TOF×s for the
i) Functionalization of the ligand. The additional
heterogenized analogs are lower by a factor of 2 5 functional groups which are in close proximity to the
while the enantioselectivities are only slightly affected. catalytic center may interact in a detrimental way. A
The influence of immobilization on activity and pro- comparison of the performance of xyliphos and the
ductivity becomes more pronounced with increasing s/c functionalized ligand 3 indicates that if such interactions
ratio as illustrated by entries 2.1 and 2.6. As already are present they are not significant enough to account
976
Adv. Synth. Catal. 2002, 344, 974 979

More than 100,000 Turnovers with Immobilized Ir-Diphosphine Catalysts
FULL PAPERS
for the lower performance of the immobilized catalysts with a loading of 0.1 mmol/g would have to exchange the
(compare entries 1.1 and 1.4). complete pore volume in about 2 3 seconds. As basis
ii) New chemical environment. Binding a metal we assumed a TOF of 120,000 hÀ1 as observed in the
complex to a support can lead to interactions with initial reaction phase for the homogeneous Ir-xyliphos, a
surface functional groups (e.g., free OH). The fact that pore volume of 1.55 mL/g and an imine concentration of
addition of silica gel to the homogeneous catalyst does 4.3 mmol/L. Withaloadingof0.042 mmol/gtheexchange
not significantly change its performance (entry 1.2) time would be about 6 s. If this exchange rate cannot be
shows that silica gel is not a strong catalyst poison. attained, this would mean that part of the immobilized
However, this result does not exclude the possibility that catalystsin the pores is not working. While this canexplain
the close proximity of metal center and surface groups the lower activity of the immobilized catalyst, it does not
could enhance such interactions.
iii) High local catalyst concentration. The phenom-
account for the observed enhanced deactivation.
At this time we favor the explanation that catalyst
enon of catalyst deactivation by irreversible formation deactivation is enhanced by the proximity of the Ir
of inactive dimers (or oligomers) is well known for Ir- complexes but it has to be stressed that deactivation is
diphosphine catalysts.[8 10] In a previous study with Ir- much less pronounced than for Ir catalysts with other
ppm and Ir-diop catalysts albeit under different reaction diphosphines.[9] We stopped our investigations at this
conditions (lower s/c, lower MEA-imine concentration, stage because it became clear that due to the much
no acid, in presence of solvent), we have observed higher activity and productivity, the catalyst of choice
significantly higher productivities of the immobilized for the production process would be the unfunctional-
catalysts. This result was attributed to site isolation of ized xyliphos and that catalyst separation would be
the surface bound catalytic species because among other accomplished by distillation of (S)-metolachlor.[5]
effects, the activity and productivity increased with
lower catalyst loading[2a, 9]. Even though the Ir-xyliphos
are known to have little tendency for deactivation, we Separation of Immobilized Catalysts
wanted to test this hypothesis as well. For this purpose,
we prepared the dimeric ligand 8 and immobilized The silica gel- as well as the polystyrene-bound catalysts
ligands 9a c with varying metal loadings of 0.10, 0.042 can be separated by simple filtration. Analysis of the Ir
and 0.0013 mmol/g in order to investigate the influence content revealed that in all cases >95% of the catalyst
of the local catalyst concentration for the present system was removed. The Ir catalyst with the water-soluble
(entries 2.5 2.7). Even though the results are not quite ligand 12 could be separated with >90% efficiency by
conclusive because the system with 0.042 mmol/g extraction into 1 N NaOH and we are confident that this
exhibited the highest activity, there is little doubt that can be improved if the number of carboxylic acid
the local concentration does have a strong effect on functions in the side chain is increased.
catalyst activity. This was confirmed with the dimeric
xyliphos ligand 8 that simulates a metal loading similar
to the one of the silica bound catalysts (assuming regular Conclusions
distribution). Ligand 8 has a maximum length of approx.
3.6 nm. Assuming two catalyst molecules in a spherical The modular concept allows us to prepare a large variety
ball with a diameter of 3.6 nm, the local catalyst of immobilized and extractable enantioselective cata-
concentration is about 0.15 molar. The concentration lysts. The preferred support is wide pore silica whereas
of the free Ir-xyliphos catalyst varied between polystyrene is not a suitable carrier. The silica-bound
2.5.10À4 molar for s/c 20,000 and 5.10À5 molar for s/c and the extractable catalysts are feasible for industrial
100,000. Thus, with ligand 8 a catalyst concentration is applications (TON >100,000, TOF up to 20,000 hÀ1).
simulated that is more than 500 times higher than the The lower activity of the immobilized catalysts is
free Ir-xyliphos catalyst and probably comparable to the tentatively explained by the higher local catalyst
situation on the surface of the silica. Indeed, the concentration on the support surface leading to an
hydrogenation results show that catalysts with the increased tendency to deactivation by irreversible dimer
dimeric xyliphos 8 have TON×s and TOF×s in the same formation. Separation of these catalysts by filtration and
range as xyliphos immobilized on the silica (compare extraction is easy and efficient. Although these immo-
entries 1.6 and 2.5 2.7). The fact that lowering the bilized catalysts exhibit unprecedented catalyst activity
catalyst loading by increasing the L/M ratio from 1.2 to 2 and productivity, the results demonstrate that hetero-
(compare entries 2.1 and 2.2), gives significantly more genization leads to highly complex systems that are
active catalysts is another indication that catalyst difficult to control.
deactivation is indeed enhanced by the proximity of
the Ir complexes.
iv) Limited mass transport in pores. A rough estima-
tion shows that the catalyst immobilized on Grace 332
Adv. Synth. Catal. 2002, 344, 974 979
977

Benoit Pugin et al.
FULL PAPERS
H); 31P NMR (CDCl3): d ˆÀ 25.2 (d, PPh2), 6.7 (d, PXyl2), JPP
Experimental Section
21 Hz.
General
General Procedure for the Preparation of the Silica
Gel-Bound Ligands 9a c
All experiments were carried out under argon. The function-
alized ligands 1 and 10 were prepared as previously describ-
ed.[4] Silica Grace 332 with a particle size of 35 70 mm was
obtained from Grace. Aminomethylated polystyrene cross-
linked with 1% divinylbenzene with 0.56 mmol amine groups/g
was purchased from Novabiochem.
A solution of ligand 3 in toluene (4.5 mL/g silica gel) was added
to silica gel Grace 332 that was previously dried under vacuum
at 130 8C for three hours and then set under argon. This mixture
was then heated to 90 8C and gently stirred for 20h. After
cooling to room temperature and sedimentation of the silica
gel, the supernatant solution was removed with a syringe. The
silica gel was washed several times with methanol and finally
dried under reduced pressure at 40 50 8C.
9a: 3.3 g silica gel, 400 mg (0.4 mmol) ligand 3; micro-
analysis, P 0.62%. This corresponds to a ligand loading of
0.1 mmol/g.
9b: 6.3 g silica gel, 300 mg (0.3 mmol) ligand 3; Micro-
analysis, P 0.26%. This corresponds to a ligand loading of
0.042 mmol/g.
(EtO)3Si-Xyliphos (3)
0.7 mL (2.6 mmol) of 1-triethoxysilyl-3-isocyanatopropane
were slowly added to a solution of 1.5 g (1.99 mmol) of
aminated xyliphos 1 in 10mL of dichloromethane. After
stirring overnight the solvent was removed under reduced
pressure. Chromatography (silica gel Merck 60, hexane/ethyl
1
acetate, 3/2) gave the pure product in 74% yield. H NMR
(CDCl3, characteristic signals): d ˆ 0.03 (s, 3H, Si-CH3), 0.11 (s,
3H, Si-CH3), 0.48 (m, 2H, CH2-Si-cp), 0.61 [m, 2H, CH2-
Si(OEt)3], 1.2 1.8 (m, 4H, CH2CH2NHCONHCH2CH2), 1.22
(t, J ˆ 7 Hz, 9H, O-CH2-CH3), 1.47 (m, 3H, CH-CH3), 2.20and
2.28 [two s, 6H, Ph(CH3)2], 2.95 3.25 (m, 4H, CH2-NHCONH-
CH2), 3.83 (q, J ˆ 7 Hz, 6H, O-CH2), 6.7 7.8 [m, 16H,
P(C6H5)2, PC6H3(Me)2]; 31P NMR (CDCl3): d ˆÀ 25.2 (d,
PPh2), 6.7 (d, PXyl2), JPP ˆ 21 Hz.
9c: 7.5 g silica gel, 120mg (0.12 mmol) ligand 3; Micro-
analysis, P 0.08%. This corresponds to a ligand loading of
approx. 0.013 mmol/g.
Preparation of the Extractable Xyliphos Ligand 12
1 g (1.3 mmol) 10 were added to a freshly prepared solution of
38 mmol sodium ethoxide and 11.6 mL (76 mmol) diethyl
malonate in 23 mL ethanol and stirred for 20hours at 80
90 8C in presence of 100 mg potassium iodide. After evapo-
ration of the solvent the mixture was treated with 10mL 2 N
HCl and extracted in water/ethyl acetate. The organic phase
was then dried with sodium sulfate. Chromatography (silica gel
Merck 60, hexane/ethyl ester, 10/1) gave pure 11 in 92% yield.
1H NMR (CDCl3): d ˆ 0.0 (s, 3H, Si-CH3), 0.08 (s, 3H, Si-CH3),
0.5 (m, 2H, CH2-Si), 1.2 (t, 6H, OCH2CH3) 1.48 (t, 3H, CH-
CH3), 1.82 [m, 2H, CH2-CH(COOEt)2], 2.15 and 2.25 (two s,
12H, Ph-CH3), 3.2 4.4 [m, 9H, C5H3FeC5H4, CH-CH3,
CH(COOEt)2], 4.15 (q, 4H, OCH2CH3), 6.7 7.7 [m, 16H,
P(C6H5)2 and P(C6H3Me2)]; 31P NMR (CDCl3): d ˆÀ 25.2 (d,
PPh2), 6.8 (d, PXyl2), JPP ˆ 21 Hz.
Preparation of the Polystyrene-Bound Ligand 6
600 mg of aminomethylated polystyrene were dried at 50 8C in
a high vacuum for 2 hours. After addition of 6 mL of THF, the
mixture was slowly stirred, resulting in polymer swelling, and
was subsequently treated with 1,3 mmol of tolylene-2,4-
diisocyanate 4. After 2 hours, the excess of 4 was removed by
washing 5 times with 10mL THF. The polymer was reslurried
in 6 mL of THF, treated with an orange solution of 60mg
(0.096 mmol) 1 in 2 mL THF and stirred for 2 hours at 40 8C,
with the polymer becoming yellow and the solution being
decolorized. Subsequently, 5 mL of ethanol and a spatula tip of
1,4-diazabicyclo[2.2.2]octane were added and the mixture
gently stirred overnight at room temperature. The solution
was then filtered, the yellow polymer washed four times with
THF and twice with diethyl ether and subsequently dried at
40 8C in a high vacuum. Microanalysis: 0.72% P. This
corresponds to a loading of 0.116 mmol ligand/g.
12: A mixture of 700 mg (0.78 mmol) 11 in 25 mL ethanol
and 1 g KOH in 1.5 mL water was stirred at room temperature
for 1 hour. After evaporation of the ethanol, the slurry was
extracted in toluene/diethyl ether/water. The water phase with
the product was washed with ethyl acetate, then toluene was
added. The pH was set to 2 3 whereby the product went into
the toluene phase, which was dried with sodium sulfate and
then evaporated. Product 12 was obtained in 90% yield.
1H NMR (CDCl3): d ˆ 0.0 (s, 3H, Si-CH3), 0.08 (s, 3H, Si-CH3),
0.57 (m, 2H, CH2-Si), 1.55 (m, 3H, C-CH3), 1.95 [m, 2H, CH2-
CH(COOH)2], 2.15 and 2.3 (two s, 12H, Ph-CH3), 3.2 4.4 [m,
9H, C5H3FeC5H4, CH-CH3, CH(COOH)2], 6.7 7.7 [m, 16H,
P(C6H5)2 and P(C6H3Me2)]; 31P NMR (CDCl3): d ˆÀ 25.9 (d,
PPh2), 8.4 (d, PXyl2).
Bis-Xyliphos (8)
22 microliters (0.134 mmol) of hexamethylene-diisocyanate
were added to a solution of 200 mg (0.27 mmol) aminated
Xyliphos 1 in 5 mL dichloromethane. After stirring overnight
the dichloromethane was removed under reduced pressure and
the raw material was purified by chromatography (silica gel
Merck 60, ethyl acetate). Product 8 was obtained as an orange
1
solid in 65% yield. H NMR (CDCl3, characteristic signals):
Typical Hydrogenation Procedure
d ˆ 0.1 [s, 12H (Si(CH3)2], 0.5 (m, 4H, Si-CH2), 2.2 (s, 12H, Ph-
CH3), 2.35 [s, 12H, Ph(CH3)2], 2.9 3.2 (m, 8H, CH2-N), 3.2
4.8 (m, cyclopentadiene-H and CH-CH3), 6.7 7.7 (m, 32H, Ph-
Thecatalyst precursors were prepared by mixing the ligand and
[Ir(cod)Cl]2 ( unless otherwise noted in a ratio of 1.2 ligand/1
978
Adv. Synth. Catal. 2002, 344, 974 979

More than 100,000 Turnovers with Immobilized Ir-Diphosphine Catalysts
FULL PAPERS
Ir) in small amounts of THF and evaporating the solvent. For
the hydrogenation experiments, MEA-imine, acetic acid
(AcOH) and tetrabutylammonium iodide (TBAI) were then
added to these preformed Ir complexes. The resulting mixture
was then pressed with argon through a capillary into a 50-mL
steel autoclave that was flushed with argon. The autoclave was
then sealed, flushed three times with hydrogen pressurizing
typically to 20bars, finally charged with hydrogen to the
specified pressure and the reaction started by switching on the
stirrer. The course of the reaction could be followed by the
hydrogen uptake. After the hydrogenation, the solvent was
evaporated and the conversion determined by gas chromatog-
raphy [column: DB17/30W, 15 m (JWC Scientific Inc.),
temperature program: 60 8C/1min. to 220 8C, DT: 10 8C/min].
The enantiomeric excess of the MEA-amine was determined
by HPLC (column: Daicel Chiracel-OD; solvent: hexane/2-
isopropanol, 99.7:0.3).
(Ciba-Geigy AG), EP Patent 728768 A2, 1996; Chem.
Abstr. 1996, 125, 248834.
[3] G. M. Ramos Tombo, H. U. Blaser, in Pesticide Chem-
istry and Bioscience, (Eds.: G. T. Brooks, T. R. Roberts),
Royal Society of Chemistry, Cambridge, 1999, p. 33.
[4] C. Kˆllner, B. Pugin, A. Togni, J. Am. Chem. Soc. 1998,
120, 10274.
[5] 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 1999, 53, 275.
[6] U. Nagel, E. Kinzel, J. Chem. Soc. Chem. Commun. 1986,
1098.
[7] T. Ohkuma, H. Takeno, Y. Honda, R. Noyori, Adv.
Synth. Catal. 2001, 4, 343.
[8] R. Crabtree, Acc. Chem. Res. 1979, 12, 331; D. Chadosh,
R. Crabtree, H. Felkin, S. Morehouse, G. Morris, Inorg.
Chem. 1982, 21, 1307; H. H. Wang, L. H. Pignolet, Inorg.
Chem. 1980, 19, 1470.
[9] H. U. Blaser, B. Pugin, F. Spindler, A. Togni, C. R.
Chimie 2002, 5, in press.
References
[10] H. U. Blaser, F. Spindler, in Comprehensive Asymmetric
Catalysis, Vol. I, (Eds.: E. N. Jacobsen, A. Pfaltz, H.
Yamamoto),Springer, Berlin, 1999, p. 247.
[1] Chiral Catalyst Immobilization and Recycling, (Eds.:
D. E. De Vos, I. F. J. Vankelecom, P. A. Jacobs), Wiley-
VCH, Weinheim, 2000.
[2] a) B. Pugin, J. Mol. Catal. 1996, 107, 273; b) B. Pugin, M.
M¸ller, Stud. Surf. Sci. Catal. 1993, 78, 107; c) B. Pugin
Adv. Synth. Catal. 2002, 344, 974 979
979