Iridium-imine and -amine complexes relevant to the (S)-metolachlor process: Structures, exchange kinetics, and C-H activation by IrI causing racemization

Article abstract of DOI:10.1002/chem.200306008

Iridium complexes of DMA-imine [2,6-dimethylphenyl-1′-methyl- 2′-methoxyethylimine, 1a) and (R)-DMA-amine [(1′R)-2,6- dimethylphenyl-1′-methyl-2′-methoxyethylamine, 2a] that are relevant to the catalytic imine hydrogenation step of the Syngenta (S)-Metola

Full text of DOI:10.1002/chem.200306008

FULL PAPER
Iridium–Imine and –Amine Complexes Relevant to the (S)-Metolachlor
I
ꢀ
Process: Structures, Exchange Kinetics, and C H Activation by Ir Causing
Racemization**
[
a]
[b]
[c]
[c]
Romano Dorta,* Diego Broggini, Reinhard Kissner, and Antonio Togni
Abstract: Iridium complexes of DMA-
imine [2,6-dimethylphenyl-1’-methyl-2’-
methoxyethylimine, 1a) and (R)-
DMA-amine [(1’R)-2,6-dimethylphen-
yl-1’-methyl-2’-methoxyethylamine, 2a]
that are relevant to the catalytic imine
hydrogenation step of the Syngenta
to form the imine complex 11, thus
modeling the product/substrate ex-
change step in the catalytic cycle. The
to liberate quantitatively the DMA-
iminium salt 14. On the other hand,
the chiral amine complex (R)-19
6
data suggest
a
two-step associative
formed the optically inactive h -bound
6
mechanism characterized by k =(2.6ꢁ
compound
[Ir(cod)(h -rac-2a)]BF4
1
2
ꢀ1 ꢀ1
0.3)”10 m
s
and k =(4.3ꢁ0.6)”
(rac-18) upon dissolution in THF at
room temperature, presumably via in-
tramolecular CꢀH activation. This ra-
2
ꢀ
2
ꢀ1
10
s
with the respective activation
ꢀ
1
(
S)-Metolachlor process were synthe-
sized: metathetical exchange of
Ir Cl (cod) ] (cod=1,5-cyclooctadiene)
energies EA1 =(7.5ꢁ0.6) kJmol
EA2 =(37ꢁ3) kJmol . Furthermore,
complex 11 reacted with H O to afford
the hydrolysis product [Ir(cod)(h -2,6-
dimethylaniline)]BF (12), and with I
and
ꢀ
1
cemization was found to be a two-step
ꢀ
4
ꢀ1
[
event with k’ =9.0”10
s
and k =
2
2
2
2
2
1
6
-
ꢀ5 ꢀ1
with [Ag(1a) ]BF
and [Ag((R)-
2.89”10 s , featuring an optically
2
4
2
-
3
2
a) ]BF4 afforded [Ir(cod)(k -1a)]BF
active intermediate prior to sp CꢀH
2
4
4
2
2
(
11) and [Ir(cod)(k -(R)-2a)]BF ((R)-
activation. Compounds 11, 12, rac-18,
and (R)-19 were structurally character-
ized by single-crystal X-ray analyses.
4
19)), respectively. These complexes
Keywords: CꢀH activation
· ex-
were then used in stopped-flow experi-
ments to study the displacement of
amine 2a from complex 19 by imine 1a
change kinetics · iridium · metola-
chlor · N ligands
Introduction
from hindered rotation around the C ꢀN axis) of the (1S)-3
Ar
enantiomer are responsible for most of the biological activi-
[5]
To date, Syngenta AG produces the chiral herbicide (S)-Me-
tolachlor [N-(1’-methyl-2’-methoxyethyl)-N-chloroacetyl-2-
ty. The key step in the Metolachlor synthesis is iridium-cat-
[6–8]
alyzed enantioselective imine hydrogenation
(Scheme 1).
[1]
ethyl-6-methylaniline, (3)]
in amounts of more than
The soluble catalyst system, which we studied in some
[2–4]
[9]
1
0000 t per annum in about 80% optical purity.
The
detail, is a combination of [Ir Cl (cod) ] (cod=1,5-cyclooc-
2
2
2
“chiral switch” from the racemate to an enantiomerically
enriched form (trademark Dual-Magnum) took place in
997 after it was found that the two atropisomers (resulting
1
[
a] Prof. Dr. R. Dorta
Departamento de Química, Universidad Simón Bolívar
Caracas 1080A (Venezuela)
Fax : (+58)212-9063961
E-mail: rdorta@usb.ve
[
b] Dr. D. Broggini
Department of Chemistry and Biochemistry, University of California
Santa Barbara, CA 93106–9510 (USA)
[
c] Dr. R. Kissner, Prof. Dr. A. Togni
Laboratory of Inorganic Chemistry, Swiss Federal Institute of Tech-
nology
ETH Hönggerberg, 8093 Zürich (Switzerland)
[
**] Part II. Part I: see ref. [9]. The experimental part of this work was
carried out at ETH Zürich.
Scheme 1. Industrial synthesis of Metolachlor (Syngenta).
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim DOI: 10.1002/chem.200306008 Chem. Eur. J. 2004, 10, 4546– 4555
4546

4
546– 4555
tadiene), the chiral ferrocenyldiphosphine Xyliphos (4),
tetrabutylammonium iodide (TBAI), and sulfuric acid.
MEA-imine 1b is hydrogenated under 80 bar hydrogen
thermore, the three potential metal coordinating modes of
imines 1a,b and amines 2a,b (Figure 1) were demonstrated
in reactivity studies, and all were assessed by X-ray crystal-
pressure at 323 K and at a substrate/catalyst ratio exceeding
6
1
0 to yield MEA-amine 2b in 79% ee and with an initial
6
ꢀ1 [3]
turnover frequency that is said to exceed 1.8”10 h . Cur-
rently, this hydrogenation not only represents the largest
scale enantioselective catalytic process in use in industry,
but it is also one of the fastest homogeneous systems
known, second only to certain homogeneous Ziegler–Natta
[
10]
polymerization catalysts and Noyoriꢁs ruthenium hydroge-
nation catalysts.
Figure 1. Possible coordinating functions of 1 and 2. R=Me, Et.
[11]
A possible hydrogenation cycle starts
with the coordination of imine 1 to an Ir hydride complex to
form imine adduct 5 (Scheme 2). Migratory insertion of the
3
lography. Finally, an unprecedented sp CꢀH activation by
I
the Ir (cod) fragment was inferred from the racemization of
coordinated optically pure amine 2a.
Results and Discussion
Synthesis and reactivity of DMA-imine complexes: Imine
1
a is accessible through the condensation of methoxyace-
tone and 2,6-dimethylaniline in good yields. Imine 1a was
isolated as a cis–trans equilibrium mixture in a 16:84 ratio at
room temperature [Eq. (1)]. First attempts to synthesize
iridium-imine adducts by reacting [Ir Cl (coe) ] (coe=cyclo-
2
2
4
octene) with a large excess of 1a in the presence of two
equivalents of AgBF4 failed. It showed instead that 1a
cleanly complexed the silver cation leaving the starting
[
Ir Cl (coe) ] unaltered. In a separate synthesis, the silver–
2 2 4
imine complex 10 was obtained as a white powder on a
Scheme 2. Proposed cycle of the iridium-catalyzed imine hydrogenation.
gram scale in 90% yield by reacting a toluene solution of
AgBF with two equivalents of imine 1a according to Equa-
4
[9]
tion (2). Complex 10 showed good solubility in CH Cl2
2
C=N bond into the IrꢀH bond leads to an iridium–amide
function (6) which was shown by Fryzuk et al. to add dihy-
and, to a lesser extent, in toluene and benzene. CH Cl solu-
2
1
2
tions decomposed within a day to form AgCl. The H NMR
spectrum of 10 indicated that only one of the cis–trans iso-
mers of 1a coordinated to the silver cation and that the co-
ordination of 1a did not significantly change its chemical
shifts. This silver–imine complex transmetalated quantita-
tively with [Ir Cl (cod) ] to yield the cationic iridium-imine
[12]
drogen by heterolytic activation.
The resulting iridium–
hydrido–amino species 7 liberates amine 2 upon coordina-
tion of an incoming substrate molecule (1). The catalytic
III
cycle is thought to be Ir -based and Ng Cheong Chan and
Osborn originally proposed a dissociative amine/imine ex-
2
2
2
[6]
1
change sequence.
adduct 11 [Eq. (3)]. The H NMR spectrum of 11 supported
a bidentate coordination mode of 1a, and again only one
isomer was observed. The resonance of the methoxy protons
of coordinated 1a was shifted downfield by about 0.5 ppm
to d=3.95 ppm. The imine-methyl protons resonated at d=
1.89 ppm and the methylene protons at d=5.40 ppm as com-
pared to d=1.66 and 4.20 ppm, respectively, in free 1a.
Iridium xyliphos complexes corresponding to intermediate
have recently been isolated and fully characterized, where-
5
as iridium–xyliphos–amido (6) and –amino (7) complexes
[9]
have eluded isolation in pure form so far. Therefore, we
opted to study simpler, phosphine-free model compounds
I
for intermediates 5 and 7 based on the Ir (cod) fragment.
DMA-imine 1a and DMA-amine 2a were used as close
model compounds for MEA-imine and MEA-amine (1b
and 2b), respectively, to avoid complications caused by atro-
pisomerism of iridium-bound 1b and 2b (imine 1a behaves
in much the same way as 1b under the aforementioned hy-
[13]
drogenation conditions ). Herein, we present the synthesis
I
and full characterization of Ir (cod) DMA-amine and DMA-
imine complexes which then allowed us to model the upper
part of the cycle depicted in Scheme 2. The analysis of the
kinetic data of the amine/imine substitution on the iridium
center allowed us to propose an associative mechanism. Fur-
toluene
AgBF4 þ 2 1 aƒƒƒ!½Agð1 aÞ ꢂBF ð10Þ
ð2Þ
2
4
Chem. Eur. J. 2004, 10, 4546– 4555
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4547

R. Dorta et al.
FULL PAPER
Crystals suitable for X-ray crystallography were grown by
slowly cooling a saturated THF solution of 11 from 333 K to
room temperature. We noticed a reproducible tendency of
11 to form hexagonal tubuli with diameters ranging from 1
to 2 mm. Noteworthy are the rare trigonal space group and
the large, “pizza box”-shaped unit cell (a=31.040(12), c=
Figure 2. ORTEP view of one of the three independent cationic mole-
cules of 11 (30% thermal ellipsoids). Selected bond lengths[Š] and
angles[8]: IrꢀO 2.102(19), IrꢀN 2.148(16), IrꢀC(21) 2.11(3), IrꢀC(24)
1
3.273(4) Š) containing six units. The asymmetric unit con-
tained three crystallographically independent molecules of
1. Owing to the physical properties of the crystal, only the
2
.00(2), IrꢀC(25) 2.07(3), IrꢀC(28) 2.17(2), OꢀC(12) 1.41(3), C(11)ꢀ
1
C(12) 1.55(4), NꢀC(11) 1.27(3), O-Ir-N 77.3(7), Ir-O-C(12) 119.9(13), Ir-
iridium and oxygen atoms were refined anisotropically.
Figure 2 shows one of the three
N-C(11) 116.2(16), O-C(12)-C(11) 104(2), N-C(11)-C(12) 123(2).
independent molecules that dis-
plays square-planar coordina- Table 1. Distances[Š] from the plane defined by O, Ir, and N in complexes 11 and 19. In the ORTEP plots,
the cod ligands and o-methyl groups of the aromatic rings are omitted for clarity.
tion geometry about the iridium
center, thus confirming the bi-
dentate coordination mode of
[
14]
1
a
with
a bite angle of
77.3(7)8. The greatest deviation
from the plane fitted through
the Ir, O, and N atoms and the
midpoints of the coordinated C(11)
11
19
C(1)
0.008
ꢀ0.108
ꢀ0.147
ꢀ0.222
0.092
ꢀ0.723
ꢀ0.183
0.452
0.374
0.462
C(12)
C(13)
C(14)
C=C double bonds is 0.09 Š for
the oxygen atom. The 2,6-di-
methylarene is twisted out of
the N-Ir-O plane by 918. The
IrꢀO bond length (2.102(19) Š)
is quite short and comparable
to the length of a single IrꢀO
[9,15,16]
bond.
The DMA-iridium
chelate ring is slightly puckered
a comparison with the DMA-
(
amine analogue 19 is presented
in Table 1), whereas the C(11)
atom is perfectly coplanar with
the C(12), C(13), and N atoms,
thus precluding any significant
charge delocalization from the
*
Scheme 3. Reactivity of DMA-imine complex 11.
metal into the p orbital of the
C=N double bond.
An attempt to substitute the cod ligand by treating 11
with a large excess of imine 1a to form a homoleptic com-
plex under 50 bar hydrogen pressure resulted in the libera-
tion of imine and the precipitation of metallic iridium. To
produce an amide that would serve as a model for the
postulated intermediate 6 (Scheme 2), 11 was reacted with
molar amount of TBAI again led to free imine and
[17]
[Ir I (cod) ]. Complex 11 reacted with one equivalent of
2
2
2
I in CH Cl to form quantitatively (NMR spectroscopy) the
2
2
2
iminium salt 14, which was identical to a sample prepared
[9]
by adding HBF to 1a, along with other sparingly soluble,
4
unidentified products. We speculate that iminium 14 formed
via intermediate 13 in analogy to a published enantioselec-
Na[HBEt ] (“superhydride”) in toluene solution. This result-
3
ed in de-coordination of the imine and formation of a mix-
ture of unidentified iridium hydrides. Diphosphine 4 dis-
placed the imine ligand 1a from complex 11 upon mixing to
tive CꢀH activation/iodination of the cod ligand that was
I
achieved by treating [Ir (4)(cod)]BF with iodine in
4
[18]
CH Cl . In that study we were able to isolate and fully
2
2
I
yield the cationic diphosphine iridium complex [Ir (4)-
characterize a complex (including an X-ray crystal structure)
corresponding to the postulated intermediate 13, the only
difference being the supporting diphosphine ligand 4 instead
(
cod)]BF , which was characterized elsewhere, including an
4
[9]
X-ray crystal structure (Scheme 3). Addition of an equi-
4548
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4546– 4555

Iridium–Imine and –Amine Complexes
4546– 4555
of imine 1a. We propose isomer 13 based on trans-influence
arguments that justify CꢀH activation/iodination taking
III
place trans to the N donor. Ir hydrides, such as the postu-
lated intermediate 13, are acidic and readily deprotonated
I
intramolecularly by imine 1a to leave a reactive Ir frag-
[
19]
ment. Two equivalents of H O reacted with 11 over three
2
Scheme 4. Synthesis of enantiomerically pure (R)-DMA-amine: i) p-Nos-
Cl, ii) 2,6-dimethylaniline, iii) LiAlH /THF, iv) KH, 273 K, v) CH I/THF.
4 3
6
days in CH Cl to afford quantitatively the h -dimethylani-
line Ir complex 12 and methoxyacetone. Complex 12 was
isolated as a white, air- and water-stable solid and its H
2
2
I
1
NMR spectrum showed the aromatic proton signals of the
dimethylaniline ligand shifted to higher fields. Crystals suit-
able for X-ray crystallography were grown from a CH Cl /
tion with LiAlH in THF gave enantiomerically pure 16 in
4
[24]
excellent yield, which was easily methylated to afford (R)-
[25]
2a in 75% yield and ꢃ99% ee. The specific optical rota-
2
2
6
20
D
pentane mixture. Figure 3 reveals the h bonding mode of
tion ([a] =ꢀ22.3, c=3.05 in hexane) is in accordance with
the value measured on an optically pure sample that was ob-
[26]
tained by a long series of recrystallizations.
Two equivalents of (R)-2a reacted on a gram-scale with
AgBF in toluene solution to afford the white microcrystal-
4
line silver complex [Eq. (4)] in a high yield in analogy to the
synthesis of complex 10 [Eq. (2)].
toluene
AgBF4 þ 2 2 a ƒƒƒ!½Agð2 aÞ ꢂBF ð17Þ
ð4Þ
2
4
(
R)-17 showed good solubility in toluene and particularly
in CH Cl (although with limited stability as in 10). In the
2
2
1
H NMR spectrum, a large downfield shift of the amine
proton was observed, resonating at d=4.82 ppm (cf. d=
3.4 ppm in free 2a), whereas the signals of the methoxy pro-
tons of the coordinated amine were shifted upfield. The spe-
4ꢀ
Figure 3. ORTEP view of the cationic part of 12 (BF omitted, 30%
thermal ellipsoids). Selected bond lengths[Š] and angles[8]: IrꢀC(1)
2
2
.421(7), IrꢀC(2) 2.259(7), IrꢀC(3), 2.270(7), IrꢀC(4) 2.316(9), IrꢀC(5)
.236(8), IrꢀC(6) 2.306(7), NꢀC(1) 1.363(10), C(1)ꢀC(2) 1.421(10), C(2)ꢀ
cific optical rotation is larger and of opposite sign as com-
20
D
pared to that of the free amine 2a: [a] = ++ 48 (c=0.622,
C(3) 1.406(12), C(3)ꢀC(4) 1.416(14), C(4)ꢀC(5) 1.392(14), C(5)ꢀC(6)
CH Cl ). Complex 17 cleanly transmetalated with [Ir Cl -
2
2
2
2
1
.431(11), N-C(1)-C(2) 119.1(7), N-C(1)-C(6) 121.6(7).
(
cod) ] in THF at room temperature to afford a white micro-
2
crystalline solid in >80% isolated yield. Surprisingly, this
1
compound displayed no optical activity, and the H NMR
2
,6-dimethylaniline and its puckered ring, with C(1) being
spectrum showed an upfield shift of the signals of the aro-
matic protons of up to 0.8 ppm with respect to that of 2a.
The methoxy protons resonated at higher field (d=3.34) rel-
ative to those of the DMA-imine complex 11 and approached
the value of the free amine 2a (d=3.37 ppm), thus suggest-
ing an uncoordinated methoxy ether function. These obser-
vations and the fact that the compound displayed low reac-
tivity pointed to the formation of an 18-valence electron
the atom that deviates most from planarity (0.10 Š). The re-
sulting envelope conformation is characterized by an angle
of 158 between the bisecting planes defined by C(1)-C(2)-
C(6) and by the best fit through C(2)-C(3)-C(4)-C(5)-C(6).
This best plane is 1.79 Š from the iridium center. Two N···F
contacts between the N atom and two BF4 ions, two sp -
C···F contacts between C(8) and C(18) and one BF , an
sp -C···F contact between C(3) and a fourth BF , all fall in
ꢀ
3
ꢀ
4
2
ꢀ
6
complex with the dimethylaniline ring coordinated in h
4
the range of 3.2–3.4 Š and may
indicate weak hydrogen bond-
[
20]
ing. Crystallographically char-
6
acterized Ir–h -arene complexes
[
21,22]
are rare
and the structure
III
6
of a Rh h -aniline complex
has been reported.
[23]
Synthesis and reactivity of
chiral DMA-amine complexes:
Enantiomerically pure (R)-2a
was prepared starting from
(
S)-methyllactate (Scheme 4).
Amine 15 was obtained in ac-
ceptable yield by analogy to a
[5]
published procedure. Reduc- Scheme 5. Syntheses of iridium DMA-amine complexes.
Chem. Eur. J. 2004, 10, 4546– 4555
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4549

R. Dorta et al.
FULL PAPER
Figure 5. ORTEP view of the cation of (R)-19 (30% thermal ellipsoids).
The H atoms on N and C(11) are in calculated positions. Selected bond
lengths[Š] and angles[8]: IrꢀO 2.116(10), IrꢀN 2.135(11), IrꢀC(21)
2
1
1
.10(2), IrꢀC(22) 2.08(2), IrꢀC(25) 2.06(2), IrꢀC(26) 2.084(14), OꢀC(12)
.44(2), OꢀC(14) 1.44(2), NꢀC(1) 1.40(2), NꢀC(11) 1.51(2), C(11)ꢀC(12)
.47(2), C(11)ꢀC(13) 1.53(2), C(21)ꢀC(22) 1.41(3), C(25)ꢀC(26) 1.37(3),
O-Ir-N 79.6(4), O-Ir-C(21) 97.6(6), O-Ir-C(22) 95.4(7), O-Ir-C(25)
1
61.8(6), O-Ir-C(26) 158.5(7), N-Ir-C(21) 156.4(6), N-Ir-C(22) 163.5(7),
N-Ir-C(25) 97.4(6), N-Ir-C(26) 93.1(5), Ir-N-C(11) 109.3(8), Ir-N-C(1)
1
1
19.0(8), Ir-O-C(12) 111.5(8), Ir-O-C(14) 125.0(10), N-C(11)-C(12)
10.8(13), N-C(11)-C(13) 110.6(12), Ir-N-C(1)-C(2) 59.1, C(11)-N-C(1)-
C(2) ꢀ72.1.
2
bonding mode (Figure 5). Reactivity studies showed k -
bound 2a to be very labile and (R)-19 was soluble and mod-
erately stable only in CH Cl . Complex rac-18 was inde-
2
2
pendently synthesized in good yield by refluxing a THF sol-
ution of (R)-19 (Scheme 5). The kinetics of this reaction are
discussed below. Addition of chiral diphosphine 4 to CD Cl2
2
solutions of 18 and 19 displaced amine 2a over a period of
Figure 4. ORTEP view of the two cationic conformers of rac-18 (30%
thermal ellipsoids). The H atoms on N and C(11) are in calculated posi-
tions and of arbitrary size. Selected bond lengths[Š] and angles[8] (fig-
ures in square brackets indicate the corresponding bonding parameters
for the disordered sidearm of conformer B): IrꢀC1 2.448(11), IrꢀC(2)
two days and upon mixing, respectively, to form complex
I
[Ir (4)(cod)]BF . DMA-imine 1a did not displace amine 2a
4
in complex 18, whereas the reaction with 19 quantitatively
produced DMA-imine complex 11 and free DMA-amine
during the time of mixing. The kinetics of this amine/imine
exchange are presented below and served as a model reac-
tion for the final step in the proposed catalytic cycle depict-
ed in Scheme 2 (vide supra).
2
.314(11), IrꢀC(3) 2.217(10), IrꢀC(4) 2.289(11), IrꢀC(5) 2.270(11), Irꢀ
C(6) 2.250(12), IrꢀC(21) 2.134(11), IrꢀC(22) 2.136(12), IrꢀC(25)
.134(12), IrꢀC(26) 2.156(12), NꢀC(1) 1.358(13), NꢀC(11) 1.52(2)
139(3)], C(21)ꢀC(22) 1.44(2), C(25)ꢀC(26) 1.40(2), C(1)ꢀC(2) 1.45(2),
2
[
C(2)ꢀC(3) 1.41(2), C(3)ꢀC(4) 1.39(2), C(4)ꢀC(5) 1.39(2), C(5)ꢀC(6)
1
1
.430(15), C(1)ꢀC(6) 1.45(2), C(11)ꢀC(12) 1.52(4)[144 (4)], C(11)ꢀC(13)
Crystals of rac-18 were grown from a saturated THF solu-
tion. Its X-ray structure confirmed the total racemization of
the coordinated DMA-amine (Figure 4). Complex rac-18
.49(4) [158(5)], C(12)ꢀO 1.46(3) [143(4)], OꢀC(14) 1.45(4) [145(6)],
C(1)-N-C(11) 125.6(11) [132.6(14)], N-C(11)-C(13) 110(2) [106(3)], N-
C(11)-C(12) 104.5(16) [116(2)], C(12)-C(11)-C(13) 112(3) [110(3)], N-
C(1)-C(2) 126.4(10), N-C(1)-C(6) 116.7(10), C(1)-C(2)-C(3) 118.8(10),
C(2)-C(3)-C(4) 123.7(10), C(3)-C(4)-C(5) 118.7(9), C(4)-C(5)-C(6)
¯
crystallizes in the space group P1 with one molecule in the
asymmetric unit. The side arm is statistically disordered
over two positions with 50% occupancy and with opposite
stereochemistries at C(11). The bond lengths along the side
chains should be interpreted with caution in view of their
disorderly behavior. The atoms C(11A) and C(11B) were re-
fined with isotropic thermal parameters. Figure 4 shows the
two enantiomers A and B with a selection of bond parame-
1
20.1(10), C(5)-C(6)-C(1) 120.2(9), C(11)-N-C(1)-C(2) ꢀ5.6[38.1].
fashion. This was confirmed by an X-ray crystal structure
analysis (Figure 4, vide infra) which allowed us to draw
structure rac-18 in Scheme 5. Furthermore, we succeeded in
2
isolating in excellent yield the corresponding k complex
6
(
(
R)-19 by reacting [Ir Cl (cod) ] with 17 in CH Cl at 213 K.
ters. Molecule A displays a somewhat puckered h -coordi-
2
2
2
2
2
R)-19 is a bright yellow solid which maintained optical ac-
nated arene moiety and, attached to it, the aliphatic chain
oriented towards the cod–Ir–arene core. As expected, there
are no intramolecular or intermolecular interactions be-
tween the free amine and ether functions with the iridium
2
D
0
1
tivity ([a] =178, c=0.950 in CH Cl ). Indeed, the H NMR
2
2
data pointed towards a bidentate NꢀO coordination of the
amine 2a: the amine proton resonated at d=5.61 ppm,
downfield by 2.2 ppm with respect to that of free 2a. Simi-
larly, the methoxy protons resonated at lower field than in
6
center. The conformation of the h -coordinated arene is best
described as an envelope defined by the C(1)-C(2)-C(6)
plane and the best plane through C(2)-C(3)-C(4)-C(5)-C(6).
The interplanar angle is 148. The iridium atom is 1.778 Š
2
a, and the aromatic protons resonated in the normal range.
Finally, an X-ray diffraction study confirmed the proposed
4550
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4546– 4555

Iridium–Imine and –Amine Complexes
4546– 4555
from the above-defined best plane. The C1–N vector
(
1.358 Š) is 0.04 Š shorter than in 19 and lies within the
C(1)-C(2)-C(6) plane. The C(1)-N-C(11) bond angle (1268)
_
is almost 128 larger than in NO-coordinated DMA-amine.
Furthermore, the C(11)ꢀN bond is approximately coplanar
with the arene, as indicated by the C(11)-N-C(1)-C(2) tor-
sion angle of ꢀ5.68. This same torsion angle is much larger
_
(
(
ꢀ72.18) when the DMA-amine is NO bound to iridium
19). Thus, it appears that a third large substituent on the N
atom (be it the iridium fragment in 19 or the chloroacetyl
[5]
unit in Metolachlor ) causes the out-of-plane tilting of the
arene ring.
Compound 19 crystallized from a CH Cl /Et O mixture.
2
2
2
Figure 5 reveals the square-planar coordination environment
in complex (R)-19 and provides a selection of bond parame-
ters. The DMA-amine bite angle of 79.6(4)8 is slightly larger
than the corresponding angle in the imine complex 11. The
largest deviation from the best plane defined by the Ir, O,
and N atoms and by the midpoints of the coordinated ole-
fins is ꢀ0.08 Š for the center of the C(21)=C(22) bond and
is comparable to that found in the imine adduct 11. The Irꢀ
O bond length is fairly short (2.116(10) Š) and the IrꢀN
2 2
Figure 6. UV/Vis spectra of 1a, 11, and 19 in CH Cl .
The traces obtained at 298 K at variable imine excesses
show that the reaction occurs in at least two significant
steps. If there were only one rate-determining step, neat ex-
ponential curves would be expected at such large imine ex-
cesses. The change in the shape of the trace with increasing
imine excess suggests the existence of additional slow steps
(Figure 7). Two steps are well-separated, even at the lowest
bond length of 2.135(11) Š lies in the expected range when
compared to the few examples of structurally characterized
I
[27–29]
Ir amine complexes.
No significant difference in the
trans influences of the O and N atoms on the coordinated
olefins is observed. The five-membered ring formed by Ir,
N, C(11), C(12), and O adopts a half-chair conformation.
Naturally, the DMA-amine ligand is more puckered than
the DMA-imine ligand, and a comparison with 11 is present-
ed in Table 1.
Kinetics of the DMA-amine/DMA-imine exchange on iridi-
um: The ligand-exchange reaction of Equation (5) was
monitored by means of the stopped-flow technique. Note
that Equation (5) models the upper part of the cycle in
Scheme 2 and relates intermediates 7 and 5.
Figure 7. Traces of the reaction of 19 with varying excesses of 1a record-
ed at l=424 nm.
First, the optical spectra of 1a, 11, and 19 in CH Cl were
imine excess, and the rate of the first step is obviously con-
centration-dependent, since the rise of the curve becomes
steeper with a larger imine excess. It follows that this first
step is bimolecular and, therefore, most likely represents the
addition of an imine molecule to the amine complex 19. The
second step cannot be assigned unambiguously with a
simple two-step model. If the second step were independent
of the imine concentration, the second phase of the traces
would be equal in all cases. If it were dependent on the
imine concentration, the total reaction time would shorten
with increasing imine concentration although the shapes of
the traces would remain geometrically similar. This is almost
the case for ratios of 1a:19 of 20:1, 40:1, and 80:1, but not
2
2
recorded to find the optimum wavelength for the stopped-
flow experiments (Figure 6). At l=360 nm and 424 nm, the
difference in absorbance without interference from 1a was
largest, but measurements at 424 nm offered the additional
advantage of a slightly superior instrument performance
(
lower noise level). The amine complex (initial concentra-
ꢀ4
tion of 3.59”10 m throughout) was allowed to react at
98 K with 10-, 20-, 40-, and 80-fold excesses of imine to
2
warrant pseudo-first-order conditions for the initial reaction
step for easier interpretation. The temperature dependence
of the reaction rate was investigated between 276K and
3
08 K with an 80-fold excess of imine.
Chem. Eur. J. 2004, 10, 4546– 4555
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4551

R. Dorta et al.
FULL PAPER
exactly. The final absorbance would also remain constant for
a given concentration of complex 19 because this would
limit the amount of product. The observed diminished final
absorbance suggests the formation of additional species at
large imine excesses. The traces were first coarsely analyzed
with the assumption of a two-step forward reaction (de-
scribed by k and k ) using Monte Carlo and Newton–Gauss
1
2
algorithms to fit the integrated differential equations. A
more detailed analysis was then carried out by numerical in-
tegration of more complex kinetic models in the form of
coupled differential equations. The best approximation to
the set of traces was found with a final equilibrium in which
the product complex 11 acquires an additional imine mole-
cule in a two-step process (characterized by k /k and k4/
3
ꢀ3
k , Scheme 6). The diminished absorbance at the end of the
ꢀ
4
Figure 8. Arrhenius plots of the first (a) and second step (b) of the reac-
tion of 19 with 1a.
1 2
Scheme 6. Kinetic model used to determine k and k .
traces is mainly caused by k /k . The traces were simulated
be deduced in the case of large differences. Frequency fac-
4
ꢀ4
4
ꢀ1
5
ꢀ1
with the known extinction coefficients from the spectra in
Figure 6. If more than two forward steps are allowed in the
model, the rate of the second step is found to be indepen-
dent of the initial imine concentration, and most probably
represents the elimination of the amine. The standard devia-
tion (2s) of the second rate constant becomes smaller if it is
assumed that the first step is reversible (introducing k ).
tors of 7.2”10 s and 1.7”10 s were estimated for the
first and second steps, respectively, and are of comparable
magnitude. The difference in the relative velocities of the
two steps (addition and elimination) is thus mainly deter-
mined by the respective activation energies.
Racemization via CꢀH activation: Dissolution of (R)-19 in
THF led to the clean formation of rac-18 over two days at
room temperature (Scheme 5), while solutions of (R)-19 in
CH Cl were stable. The disappearance of (R)-19 and the
ꢀ
1
This is reasonable for an addition reaction to a square-
planar complex. The values found for the bimolecular imine
addition and the amine elimination rate constants are (2.6ꢁ
2
2
2
ꢀ1 ꢀ1
ꢀ2 ꢀ1
0.3)”10 m
s
and (4.3ꢁ0.6)”10 s , respectively. Given
appearance of rac-18 went hand-in-hand with a diminishing
optical activity of the THF solution. This decay was moni-
tored by optical rotation spectrometry and NMR spectrosco-
py.
the method of approximation and the uncertainties about
the properties of the minor products, these values are relia-
ble within their decades. The equilibrium constant of the
imine addition is estimated to be 1.5”10 m . The precise
rate constants of the forward and backward reactions of this
equilibrium cannot be determined because the equilibrium
settles faster than the preceding reaction proceeds.
4
ꢀ1
Figure 9 shows the evolution of the [a]D of a THF solu-
tion of (R)-19 over 700 minutes. A good fit was obtained
with the reaction sequence shown in Scheme 7. The first
step is probably pseudofirst order, namely, k ’=k [THF]
1
1
ꢀ
4
ꢀ1
The activation energies for the first two distinct reaction
steps were determined by varying the temperature from
with k ’=9.0”10 s , whereas the second step is first order
1
with k =2.89”10 s . From the data at hand, we found the
ꢀ
5
ꢀ1
2
2
76K to 308 K (Figure 8). Step 1 requires a low activation
intermediate to be optically active with an [a] =163ꢁ9. No
D
ꢀ
1
energy of 7.5ꢁ0.6kJmol , as expected for an addition re-
hydride signals were detected in the NMR spectra during
the reaction. We postulate an intermediate such as 20
(Scheme 7), in which the ether function of the amine is dis-
placed by an incoming THF molecule. This should render
the bH atom on the stereogenic carbon atom more accessi-
ble to the iridium center. Thus, the racemization of amine
action to a free coordination site. Step 2 has a considerably
ꢀ
1
higher activation energy of 37ꢁ3 kJmol , which points to
an intramolecular rearrangement. The frequency factors ob-
tained from a narrow temperature range must be interpret-
ed with caution, although some qualitative information may
4552
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4546– 4555

Iridium–Imine and –Amine Complexes
4546– 4555
portant part of the hydrogenation cycle (i.e. product–sub-
strate exchange) was modeled, and the kinetic data suggest
that imine 1a displaces amine 2a from complex 19 by an as-
sociative mechanism. The substrate molecule (imine 1a) is
I
indeed the much better ligand for Ir than the product mole-
cule (amine 2a). In fact, the industrial solvent-free MEA-
imine hydrogenation proceeds rapidly and to completion,
despite the very large MEA-amine concentration toward
[
32]
3
the end of the reaction. Indirect evidence for sp -CꢀH ac-
tivation was found in two cases: In the I oxidation of com-
2
plex 11, leading to one equivalent of iminium salt 14 (via
the proposed CꢀH activated intermediate 13), and in the
racemization of (R)-19, leading to rac-18 (via b-CꢀH activa-
Figure 9. [a]
D
as a function of time. Measured 5 min after dissolving (R)-
1
9 in THF at 298 K.
I
tion of the amine by the Ir (cod) fragment). The latter trans-
2
6
formation of k -coordinated amine (R)-2a into racemic, h -
coordinated rac-2a bears some relevance to the industrial
process. It demonstrates that an iridium catalyst is, in princi-
ple, capable of lowering the optical purity of the product
3
[33]
amine by sp -CꢀH bond activation, and it shows a possi-
6
ble catalyst deactivation pathway by h coordination of sub-
strate or product molecules that irreversibly block coordina-
tion sites on the iridium catalyst. Hydrolysis of the iridium-
6
coordinated imine substrate molecule to afford an h -
bonded dialkylaniline iridium complex, as observed in the
formation of complex 12, is a slow reaction and unlikely to
occur during the fast imine hydrogenation reaction. In both
Scheme 7. Proposed intermediate DMA-amine complex prior to CꢀH ac-
tivation.
I
cases, the Ir (cod) fragment showed a strong tendency to co-
6
ordinate the aniline fragment in an h fashion.
3
2
a probably occurs by reversible sp CꢀH activation (bH
elimination to afford an intermediate iminium species 14,
followed by re-insertion into the IrꢀH bond). It should be
noted that CꢀH activation by oxidative addition is usually
Experimental Section
accomplished by electron-rich late transition metal centers
bearing electron-donating ligands, such as phosphines or cy-
[
34]
[9]
[
Ir
2
Cl
2
(cod)
2
],
1a, and 10 were synthesized according to published
procedures. All reactions were carried out under anaerobic and anhy-
drous conditions and reactions involving Ag salts were carried out in the
dark. Elemental analyses were performed at the Mikroelementaranaly-
tisches Laboratorium, ETH Zürich. Optical rotation spectra were record-
ed on a Perkin-Elmer341 polarimeter, and optical spectra were recorded
on a Uvikon820 spectrophotometer. The stopped-flow experiments were
performed on a Applied PhotophysicsSX17MV instrument under anaero-
bic conditions. The thermostat fluid was continuously purged with nitro-
gen (99.999%), and the rear side of the array of drive syringe pistons
was covered with a nitrogen-purged cap. The reactant solutions (19 and
clopentadienyl anions. The coordinated amine function of
I
2
a augments the electron density on the Ir center and thus
seems to render CꢀH bond activation feasible. Pre-coordi-
I
nation of an amine function to an Ir species triggering
[
30]
[31]
CꢀH and NꢀH activation is documented, and the acti-
I
vation of a CꢀH bond a to a amine function by a Ir (cod)
fragment has been demonstrated by deuterium labeling ex-
[29]
periments.
This racemization reaction may explain why
1
a in CH
2 2
Cl ) were supplied from tonometer reservoirs (Schlenk tube-
the stereoselectivity of the hydrogenation of 1b is limited to
ꢄ80%, and it points to a potential mechanism of catalyst
deactivation. In fact, rac-18 turned out to be a very stable
and inert 18 valence electron complex. The elucidation of
the precise mechanism needs further studies.
like vessels) which were filled in a glove box. The concentration of 19
was 3.5”10 m throughout, and that of 1a was 3.5”10 m, 7”10 m,
ꢀ
4
ꢀ3
ꢀ3
ꢀ
2
ꢀ2
1.4”10 m, and 2.8”10 m. In the temperature-dependence experiments,
ꢀ
2
the concentration of 1a was 2.8”10 m throughout. Recording of valid
traces was started after the stopped-flow circuit had been flushed with re-
actants until the trace shape became stable. The traces were analyzed
with the built-in fit capabilities of the Applied Photophysics control soft-
ware and subsequently with the program proFit (Quantum Soft, Zürich,
Switzerland). The more complex analyses were carried out by numerical
integration in a Microsoft Excel spreadsheet. Integration was effected by
To summarize, we succeeded in synthesizing and charac-
I
terizing imine and optically pure amine complexes of Ir
that are relevant to the Syngenta MEA-imine hydrogenation
process. We showed that the DMA-imine and DMA-amine
molecules (as very close models to MEA-imine and MEA-
th
a VBA macro which uses a 4 order Runge–Kutta algorithm. The fit to
the experimental data was calculated with the built-in Solver macro in
amine) indeed exhibited all coordination modes conceivable
based on their structure (Figure 1). Bidentate NO coordina-
the tangent mode.
_
2
[Ir(cod)(k -1a)]BF (11): THF (250 mL) was added to [Ir Cl (cod) ]
4
2
2
2
tion to iridium was demonstrated for the DMA-imine (com-
plex 11) and the DMA-amine molecule (complex 19). In
2 4
(1.007 g, 1.499 mmol) and [Ag(1a) ]BF (10, 1.730 g, 2.997 mmol). The re-
sulting green-yellow mixture was stirred for 1 h and then evaporated to
dryness. The residue was extracted with CH Cl (4”20 mL). The volatiles
were evaporated, and the residue was washed in Et O (3”80 mL) and
dried in vacuo to afford a greenish-yellow powder. Yield: 1.680 g (97%);
2
2
both complexes, switching of the coordination mode from
_
2
2
6
k -NO to h -arene were observed upon hydrolysis of 11 and
racemization of 19 leading to 12 and 18, respectively. An im-
1
2 2
H NMR (250.13 MHz, CD Cl ): d=1.43 (br, 2H), 1.70 (br, 2H), 1.90 (s,
Chem. Eur. J. 2004, 10, 4546– 4555
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4553

R. Dorta et al.
FULL PAPER
3
2
H), 2.20 (br, 4H), 2.22 (s, 6H), 2.80 (br, 2H), 3.95 (s, 3H), 4.34 (br,
H), 5.39 (s, 2H), 7.13–7.18 ppm (m, 3H); elemental analysis calcd (%)
with a syringe. The resulting mixture was stirred at room temperature for
130 h, and the volatiles were removed with a vacuum pump to afford an
for C20
H
29NBOF
4
Ir: C 41.53, H 5.05, N 2.42; found: C 41.47, H 5.23, N
olive green solid that was extracted with CH
was evaporated to yield a red oily solid that was recrystallized from THF
4 mL) and Et O (12 mL) to afford a white microcrystalline solid. Yield:
55 mg (81%). Crystals suitable for an X-ray analysis were obtained by
dissolving 95 mg of the product in THF (1.2 mL) at 333 K and then leav-
2 2
Cl (3”3 mL). The solution
2
.36.
(
2
2
Methyl-(2S)-O-(p-nitrobenzenesulfonyl)lactate: A stirred solution of (S)-
methyllactate (113.4 g, 1.089 mol) and p-nitrobenzenesulfonyl chloride
(
3
100.0 g) in ethyl acetate (0.50 L) was treated with NEt (69 mL). The
2
0
ing the solution at room temperature for two days. [a] =0 (c=0.546,
temperature was kept below 300 K. The resulting white slurry was stirred
overnight and then heated to 333 K for 1 h. After filtration, the organic
phase was washed with HCl (2m), H
ganic phase was dried over Na SO , and the volatiles were evaporated.
The residue was dissolved in CH OH (80 mL), and the solution was
D
CH
.38, N 2.41; found: C 41.20, H 5.26, N 2.34.
Method B: A THF solution (5 mL) of [Ir(cod)((R)-2a)]BF
99 mg) was refluxed for 20 h, and the mixture was then concentrated to
about 1 mL. Addition of Et O (7 mL) precipitated an off-white solid.
The supernatant solution was filtered off and the solid dried in vacuo.
2 2 4
Cl ); elemental analysis calcd (%) for C20H31NBOF Ir: C 41.38, H
5
2
O, NaOH (2m), and brine. The or-
4
((R)-19,
2
4
3
cooled to afford an off-white crystalline solid that was collected by filtra-
tion and dried in vacuo (63.6 g, 20%).
2
1
Yield: 82 mg (83%); H NMR (250.13 MHz, CD
2
Cl
2
): d=1.32 (d, 3H,
(
R)-N-(1’-Methyl-2’-methylcarboxyethyl)-2,6-dimethylaniline (15): To a
3
J=6.6 Hz), 2.00–2.30 (m, 8H), 2.13 (s, 3H), 2.17 (s, 3H), 3.35 (s, 3H),
stirred yellow solution of 2,6-dimethylaniline (111 mL, 0.900 mmol) in
chlorobenzene (400 mL) was added dropwise a solution of methyl-(2S)-
O-(p-nitrobenzenesulfonyl)lactate (160 g, 0.552 mol) in chlorobenzene
3
.44 (m, 2H), 3.95–4.10 (m, 5H), 4.38 (brd, 1H, J=9.9 Hz), 6.15–6.19
(
m, 2H), 6.36 ppm (m, 1H); elemental analysis calcd (%) for
C
20
H
31NBOF
4
Ir: C 41.38, H 5.38, N 2.41; found: C 41.13, H 5.41, N 2.47.
((R)-19): colorless solution of [Ag((R)-
(17, 628.7 mg, 1.061 mmol) in CH Cl (20 mL) was
added dropwise to an orange solution of [Ir Cl (cod) (356.2 mg,
0.5300 mmol) in CH Cl (20 mL) at 213 K over 20 min. The resulting
(
4
160 mL) at reflux temperature. The mixture was refluxed for a further
h to afford a red solution. After filtration of the solution and evapora-
tion of the volatiles, the oily residue was washed with HCl (1m) and ex-
tracted with Et O. Drying over Na SO and distillation afforded a yellow
oil. Yield: 48.7 g (45%); [a] =37.8 (c=1.25, CH OH).
R)-N-(1’-Methyl-2’-hydroxyethyl)-2,6-dimethylaniline (16): To
2
[Ir(cod)(k -(R)-2a)]BF
2a)
4
A
1
2
]BF
4
· =
8
C
7
H
8
2
2
2
2
2
]
2
2
4
2
0
3
2
2
D
lemon-yellow mixture was slowly allowed to reach 273 K over 2.5 h. The
mixture was then evaporated to afford a yellow solid that was extracted
(
a pale
yellow solution of (R)-N-(1’-methyl-2’-methylcarboxyethyl)-2,6-dimethyl-
aniline (15, 4.45 g, 21.4 mmol)in THF (100 mL) was added a ꢄ1m solu-
with CH
2
Cl
2
(3”30 mL). Concentration to ꢄ5 mL and addition of Et
2
O
(
10 mL) caused the precipitation of lemon-yellow microcrystals. Yield:
4
tion of LiAlH (21.5 mL, 21.5 mmol) in THF dropwise over 15 min by a
1
5
2 2
65 mg (92%); H NMR (250.13 MHz, CD Cl ): d=1.02 (br, 1H), 1.11
syringe. The solution was stirred for 6h, and it changed color from dark
yellow to pale yellow. The mixture was quenched by adding i) H
0.8 mL), ii) aq. NaOH (0.8 mL, 15%), iii) H
O (1.6mL), and iv) ꢄ1 g of
Na SO to the vigorously stirred solution. The salts were filtered off and
washed with Et O. The volatiles of the combined organics were removed
in vacuo. The resulting oil was redissolved in CH Cl and washed with
O (2”50 mL). The organic phase was dried with Na SO and then dis-
tilled (2”10 mbar, Tvap ꢄ359 K) to afford a pale yellow, highly viscous
oil. Yield: 3.29 g (86%); [a] =+7.24 (c=1.974, CH OH); H NMR
): d=1.07 (d, 6.3 Hz, 3H), 2.30 (s, 6H), 2.49 (br,
H), 2.94 (br, 1H), 3.35–3.55 (m, 2H), 3.65–3.75 (m, 1H), 6.80–6.85 (m,
H), 7.00–7.05 ppm (m, 2H); elemental analysis calcd (%) for
17NO: C 73.7, H 9.56, N 7.81; found: C 73.6, H 9.51, N 7.87.
R)-N-(1’-methyl-2’-methoxyethyl)-2,6-dimethylaniline ((R)-2a): A solu-
tion of (R)-N-(1’-methyl-2’-hydroxyethyl)-2,6-dimethylaniline (16,
.983 g, 11.06mmol) in THF (35 mL) was added dropwise over 10 min to
a stirred suspension of KH (0.475 g, 11.8 mmol) in THF (15 mL). H evo-
lution was observed. The mixture was stirred overnight at room temper-
ature to afford a turbid yellow solution. A solution of CH I in THF
20 mL) was added dropwise over 80 min under vigorous stirring. Large
amounts of a white precipitate formed. The mixture was concentrated to
3
(
d, 3H, J(H,H)=6.5 Hz), 1.18 (br, 1H), 1.63 (br, 1H), 1.84 (br, 2H),
2
O
2
3
5
.1–2.6(br, 4H), 2.43 (s, 3H), 3.09 (s, 3H), 3.45–3. 65 (m, 2H), 3.81 (s,
(
2
3
3
H), 3.84 (dd, 1H, J=4.9 Hz), 4.43 (dd, 1H, Jvic =11.6Hz, J=9.9 Hz),
2
4
3
.61 (brd, 1H, J=10.7 Hz), 7.05–7.09 (m, 2H), 7.14–7.19 ppm (m, 1H);
2
2
D
0
[
a] =+178 (c=0.950, CH
2
Cl
IrNO: C 41.38, H 5.38, N 2.41; found: C 41.44, H 5.52, N 2.40.
Crystals suitable for an X-ray analysis were grown by careful layering of
a solution of the compound (108 mg) in CH Cl (1 mL) with Et
5 mL).
2
); elemental analysis calcd (%) for
2
2
20 4
C H31BF
H
2
2
4
ꢀ2
2
D
0
1
2
2
2
O
3
(
(
250.13 MHz, CDCl
3
CCDC-223547, CCDC-134959, CCDC-134962, and CCDC-134960 con-
tain the supplementary crystallographic data for compounds 11, 12, 18,
and 19, respectively. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crys-
tallographic Data Centre, 12 Union Road, Cambridge CB21EZ, UK;
fax: (+44)1223-336033; or deposit@ccdc.cam.uk).
1
1
11
C H
(
1
2
3
(
Acknowledgement
3
Et
0 mL, and brine (100 mL) was added. Extraction of this mixture with
ꢀ2
We thank Dr. Volker Gramlich (ETH) for solving the crystal structure of
compound 11. R.D. thanks Novartis AG (Switzerland) and FONACIT
2
O and distillation (10 mbar, Tvap ꢄ318 K) yielded a yellowish viscous
2
0
2
0
1
oil. Yield: 1.58 g (75%); [a] =ꢀ22.3, [a] =ꢀ129 (c=3.05, hexane); H
D
365
(
Venezuela) for financial support.
NMR (250.13 MHz, CDCl
3
): d=1.19 (d, J=6.4 Hz, 3H), 2.28 (s, 6H),
3
2
7
.30–3.45 (m, 4H), 3.37 (s, 3H), 6.75–6.85 (m, 1H), 6.95–7.00 ppm (m,
H); elemental analysis calcd (%) for C12H18NO: C 74.96, H 9.44, N
.28; found: C 74.9, H 9.80, N 7.15.
[
1] C. Vogel, R. Aebi, 1972 DP 2328340 to Ciba-Geigy AG.
1
[2] H.-U. Blaser, F. Spindler, Chimia 1997, 51, 297.
[
Ag((R)-2a)
2
]BF
4
· =
8
C
7
H
8
(17): (R)-N-(1’-Methyl-2’-methoxyethyl)-2,6-di-
[
3] H.-U. Blaser, H.-P. Buser, K. Coers, R. Hanreich, H.-P. Jalett, E.
methylaniline [(R)-2a, 2.060 g, 10.66 mmol] was added dropwise to a
pinkish solution of AgBF (1.002 g, 5.147 mmol) in toluene (40 mL). The
mixture was stirred for 1 h at room temperature, and then the volatiles
were evaporated. The residue was washed with pentane (3”20 mL) and
Jelsch, B. Pugin, H.-D. Schneider, F. Spindler, A. Wegmann, Chimia
4
1
999, 53, 275.
[
[
[
4] H.-U. Blaser, Adv. Synth. Catal. 2002, 344, 17.
5] H. Moser, G. Rhys, Z. Naturforsch. B 1982, 37, 451.
6] Y. Ng Cheong Chan, J. A. Osborn, J. Am. Chem. Soc. 1990, 112,
dried in vacuo to afford a white microcrystalline powder. Yield: 2.79 g
1
(
91%); H NMR (CD
2
Cl
2
, 250.13 MHz): d=0.93 (d, 6H, J=6.1 Hz), 2.34
9
400.
(
brs, 12H), 2.66 (s, 6H), 3.27–3.50 (m, 6H), 4.82 (brd, 2H), 6.95–
[
[
[
7] F. Spindler, B. Pugin, H.-U. Blaser, Angew. Chem. 1990, 102, 56 1;
1
7
.08 ppm (m, 6H). The NMR spectrum indicated
=
8
equivalents of tolu-
Angew. Chem. Int. Ed. Eng. 1990, 29, 558.
2
0
ene. [a] =+48 (c=0.622, CH
C
6
2
Cl
2
); elemental analysis calcd (%) for
: C 50.40, H 6.63, N 4.73; found: C 50.71, H
D
8] D. Xiao, X. Zhang, Angew. Chem. 2001, 113, 3533; Angew. Chem.
Int. Ed. 2001, 40, 3425.
9] R. Dorta, D. Broggini, R. Stoop, H. Rüegger, F. Spindler, A. Togni,
Chem. Eur. J. 2004, 10, 267.
1
24
H
38AgBF
4
N
2
O
2
· =
8
C
7
H
8
.57, N 4.85.
6
[
Ir(cod)(h -rac-2a)]BF
4
(18):
Method A: THF (12 mL) was added to [Ir
2
Cl
2
(cod)
2
]
(182.2 mg,
[10] G. Fink, R. Mühlhaupt, H. H. Brintzinger, Ziegler Catalysts, Spring-
er, Berlin Heidelberg 1995.
1
0
.271 mmol) and [Ag((R)-2a) ]BF · = C H (17, 321.4 mg, 0.542 mmol)
2 4 8 7 8
4554
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4546– 4555

Iridium–Imine and –Amine Complexes
4546– 4555
[
11] R. Noyori, T. Ohkuma, Angew. Chem. 2001, 113, 40; Angew. Chem.
Int. Ed. 2001, 40, 40.
[25] Determined by chiral HPLC using Daicelꢁs Chiracel ODH column
ꢀ
1
and eluting with 2-propanol (0.35%) in hexane; 1.0 mLmin ; T=
298 K; R =7.1 min for (R)-2a; R =8.2 min for (S)-2a.
[12] D. M. Fryzuk, W. E. Piers, Organometallics 1990, 9, 986.
[13] F. Spindler (Solvias AG), personal communication.
[14] A. G. Becalski, W. R. Cullen, M. D. Fryzuk, B. R. James, G.-J. Kang,
S. J. Rettig, Inorg. Chem. 1991, 30, 5002.
f
f
[26] K. Bernauer (University of Neuchꢃtel, Switzerland), personal com-
munication.
[27] D. M. Roundhill, R. A. Bechtold, S. G. N. Roundhill, Inorg. Chem.
1980, 19, 284.
[28] C. Bianchini, E. Farnetti, M. Graziani, G. Nardin, A. Vacca, F. Za-
nobini, J. Am. Chem. Soc. 1990, 112, 9190.
[29] A. A. H. Van der Zeijden, G. van Koten, R. Luijk, R. A. Norde-
mann, A. L. Spek, Organometallics 1988, 7, 1549.
[30] R. Dorta, A. Togni, Organometallics 1998, 17, 3423.
[31] A. L. Casalnuovo, J. C. Calabrese, D. Milstein, Inorg. Chem. 1987,
26, 971.
[
[
[
15] C. A. Miller, T. S. Janik, C. H. Lake, L. M. Toomey, M. R. Churchill,
J. D. Atwood, Organometallics 1994, 13, 5080.
16] M. A. Esteruelas, A. M. López, L. A. Oro, A. Pꢂres, M. Schulz, H.
Werner, Organometallics 1993, 12, 1823.
17] S. M. Socol, C. Yang, D. W. Meek, R. Glaser, Can. J. Chem. 1992,
7
0, 2424.
[
[
18] R. Dorta, A. Togni, Organometallics 1998, 17, 5441.
19] In the presence of diphosphine ligand 4, the deprotonated Ir com-
plex underwent an SNi-type reaction to form the allyl complex
I
[32] The coordination behavior of the substrate and product molecules
3
2
I
[
IrI(4)(h ,h -C
8
H
11)]BF
4
(see ref. [18]). In the present case, the re-
may be assumed to be similar in the Ir (cod) fragment and in the
III
sulting iridium fragment is poorly supported and thus may form
sparingly soluble precipitates.
actual catalyst system (based on the Ir (xyliphos) fragment, see
ref. [9]).
[
[
[
[
[
20] G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford Uni-
[33] A comprehensive ligand screening revealed that it is difficult to
reach 90% ee. F. Spindler (Solvias AG), personal communication.
[34] R. H. Crabtree, Q. J. M. , H. Felkin, T. Fillebeen-Kahn, Synth.
React. Inorg. Met.-Org. Chem. 1982, 12, 407.
versity Press, Oxford, New York 1997.
21] R. Usón, L. A. Oro, D. Carmona, M. A. Esteruelas, C. Foces-Foces,
F. H. Cano, S. Garcia-Blanco, J. Organomet. Chem. 1983, 254, 249.
22] J. W. Steed, R. K. Juneja, R. S. Burkhalter, J. L. Atwood, J. Chem.
Soc. Chem. Commun. 1994, 2205.
23] S. Ogo, H. Chen, M. M. Olmstead, R. H. Fish, Organometallics 1996,
Received: December 10, 2003
Revised: April 14, 2004
Published online: August 2, 2004
1
5, 2009.
24] A. I. Meyers, G. S. Poindexter, Z. Birch, J. Org. Chem. 1987, 52, 892.
Chem. Eur. J. 2004, 10, 4546– 4555
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4555