Oxidative addition of dihydrogen to [Ir(bisphosphine)(1,5- cyclooctadiene)]BF4 complexes: Kinetic and thermodynamic selectivity

Article abstract of DOI:10.1021/ja981536b

The activation of dihydrogen by cationic diphosphine complexes of Rh is the rate-determining and enantiodetermining step in the catalytic asymmetric hydrogenation of prochiral enamides. The addition of H₂ to [Ir(bisphosphine)(COD)]⁺ (COD = 1,5-cyclooctadiene) complexes is examined herein as a model for stereocontrol and dynamic processes related to catalytic hydrogenation. The diastereoselectivity of H₂ oxidative addition to form diastereomeric [Ir(chiral bisphosphine)(H)₂(COD)]⁺ complexes at - 80 °C is kinetically controlled and varies substantially with the structure of the diphosphine. In one instance (chiral bisphosphine = CHIRAPHOS), the kinetic and thermodynamic selectivities of H₂ addition are inverted; i.e., the dominant kinetic product is thermodynamically less stable than the minor kinetic product. For the [IrH₂(Me-DuPhos)- (COD)]⁺ system, quantitative analysis of the 2D NOE data using two-dimensional conformer population analysis (2DCPA) establishes the absolute stereochemistry and the three- dimensional structures of the predominant conformers. Kinetic analysis of the interconversion of the [IrH₂(MeDuPhos)(COD)]⁺ diastereomers, and of their exchange with D₂, reveals a complex pathway for isotope scrambling and diastereomer interconversion that does not involve reductive elimination of H₂ to form [Ir(Me-DuPhos)(COD)]⁺. During the isotope scrambling, exquisite selectivity is seen for exchange in only two ligand sites, one on the Me- DUPHOS ligand and one on the COD ligand.

Full text of DOI:10.1021/ja981536b

J. Am. Chem. Soc. 1998, 120, 10115-10125
10115
Oxidative Addition of Dihydrogen to
[Ir(bisphosphine)(1,5-cyclooctadiene)]BF4 Complexes: Kinetic and
Thermodynamic Selectivity
Barbara F. M. Kimmich, Ekasith Somsook, and Clark R. Landis*
Contribution from the Department of Chemistry, UniVersity of WisconsinsMadison,
1101 UniVersity AVenue, Madison, Wisconsin 53706
ReceiVed May 4, 1998
Abstract: The activation of dihydrogen by cationic diphosphine complexes of Rh is the rate-determining and
enantiodetermining step in the catalytic asymmetric hydrogenation of prochiral enamides. The addition of H2
to [Ir(bisphosphine)(COD)]+ (COD ) 1,5-cyclooctadiene) complexes is examined herein as a model for
stereocontrol and dynamic processes related to catalytic hydrogenation. The diastereoselectivity of H2 oxidative
addition to form diastereomeric [Ir(chiral bisphosphine)(H)2(COD)]+ complexes at -80 °C is kinetically
controlled and varies substantially with the structure of the diphosphine. In one instance (chiral bisphosphine
) CHIRAPHOS), the kinetic and thermodynamic selectivities of H2 addition are inverted; i.e., the dominant
kinetic product is thermodynamically less stable than the minor kinetic product. For the [IrH2(Me-DuPhos)-
(COD)]+ system, quantitative analysis of the 2D NOE data using two-dimensional conformer population analysis
(2DCPA) establishes the absolute stereochemistry and the three-dimensional structures of the predominant
conformers. Kinetic analysis of the interconversion of the [IrH2(Me-DuPhos)(COD)]+ diastereomers, and of
their exchange with D2, reveals a complex pathway for isotope scrambling and diastereomer interconversion
that does not involve reductive elimination of H2 to form [Ir(Me-DuPhos)(COD)]+. During the isotope
scrambling, exquisite selectivity is seen for exchange in only two ligand sites, one on the Me-DuPHOS ligand
and one on the COD ligand.
Introduction
dihydrides under catalytic hydrogenation conditions. Whether
these complexes are intermediates or the products of side
equilibria in the catalytic cycle currently is unknown.
The catalytic asymmetric hydrogenation of prochiral sub-
strates continues to be an active area of industrial and academic
research. For example, Novartis recently announced a com-
merical process for the catalytic asymmetric hydrogenation of
imines using homogeneous iridium catalysts.1 The first suc-
cessful catalytic asymmetric hydrogenation reactions involved
the reduction of prochiral enamides.2-4 Other notable recent
developments primarily involve the synthesis of new rigid
diphosphine ligands that exhibit high enantiospecificity over a
broad range of enamide and related substrates.2,4
In contrast with the olefinic rhodium complexes of cis-
chelating diphosphines, dihydrides of analogous iridium com-
plexes can be observed directly by NMR spectroscopy. As a
result, stable dihydrides are readily observed for iridium
complexes of cyclooctadiene at -80 °C.9 This provides the
opportunity to explore factors influencing the selectivity of
authentic oxidative addition reactions and their subsequent
rearrangement dynamics.10 Although iridium complexes may
exhibit different catalytic activities and mechanisms than
analogous rhodium complexes, we note that diastereomeric
enamide complexes of the two metals with chiral diphosphines
exhibit similar stereoselectivities upon reaction with H2, with
the more stable diasteromer being less reactive toward H2.11
These results suggest that stereodifferentiating interactions in
the transition states for dihydrogen activation by rhodium
complexes may be modeled well by the corresponding iridium
complexes.
A critical step in hydrogenation reactions catalyzed by
transition metal complexes is the activation of dihydrogen. For
example, although the phenomenological steps of catalytic
enamide hydrogenation are well understood,5-7 good models
of the turnover-limiting and enantiodetermining transition states
have not yet emerged. Most current interpretations5-7 invoke
a hydrogen activation that proceeds via H2 oxidative addition
followed by insertion of the alkene into a M-H bond. The
recent characterization of a dihydride intermediate by Bargon
and co-workers8 via para-H2 enhanced H NMR spectroscopy
has provided the first direct evidence for the existence of such
1
This paper contains a detailed description of the oxidative
addition of dihydrogen to a series of [Ir(bisphosphine)(COD)]-
BF4 complexes (COD ) 1,5-cyclooctadiene). First, we present
data on the intrinsic diastereoselectivity of the oxidative addition
(1) Blaser, H. U.; Spindler, F. Top. Catal. 1997, 4, 275-282.
(2) Applied Homogeneous Catalysis with Organometallic Compounds;
Cornils, B., Herrmann, W. A., Eds.; VCH: Weinheim, 1996.
(3) Ojima, L.; Clos, N.; Bastos, C. Tetrahedron 1989, 45, 6901.
(4) Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH Publishers: New
York, 1993.
(5) Landis, C. R.; Halpern, J. J. Am. Chem. Soc. 1987, 109, 1746.
(6) Brown, J. M.; Chaloner, P. A.; Morris, G. A. J. Chem. Soc., Perkins
Trans. 2 1987, 1583.
(8) Harthun, A.; Kadyrov, R.; Selke, R.; Bargon, J. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1103.
(9) Crabtree, R. H.; Felkin, H.; Morris, G. E. J. Chem. Soc., Chem.
Commun. 1976, 716.
(10) Brauch, T. W.; Landis, C. R. Inorg. Chim. Acta 1998, 270, 285-
297.
(7) Brown, J. M.; Chaloner, P. A.; Morris, G. A. J. Chem. Soc., Chem.
Commun. 1983, 664.
(11) Armstrong, S. K.; Brown, J. M.; Burk, M. J. Tetrahedron Lett. 1993,
34, 879-882.
S0002-7863(98)01536-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/22/1998

10116 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Kimmich et al.
Table 1. 1H and 31P NMR Spectral Data for [IrH2(P-P)(COD)]BF4 in CD2Cl2 at -80 °C
1H NMRa hydride region, δ (ppm)
31P NMRb {1H},b δ (ppm)
[IrH2(COD)(P-P)]BF4
isomer, P-P )
6, (S,S)-DIOPc
IrHcis (JP-H, Hz)
IrHtrans (JP-H, Hz)
Ptrans (JP-P, Hz)
Pcis (JP-P, Hz)
maj (a)
min (b)
maj (b)
min (a)
(b)
-14.18, dd (24, 16)
-15.25, t (15)
-9.70, dd (90, 10)
-8.9, dd (92, 14)
25.2 br s
-31.6 br s
-7.2 br s
-8.9 br s
32.5 d (12)
38.4 d (18)
13.9 d (18)
11.8 d (21)
54.4 d (9)
40.1 d (9)
10.3 d (18)
8.2 d (18)
-7.7 d(18)
1.2 d (18)
1.6 d (18)
1.4 d (15)
4.3 d (18)
2.8 br s
7, (S,S)-CHIRAPHOSc
-13.45, t (16)
-9.89, dd (74, 12)
-10.48, dd (88, 12)
-11.81, dd (88, 10)
-11.70, dd (94, 11)
-10.38, dd (88, 16)
-9.85, dd (110, 19)
-11.08, dd (92, 19)
-11.70, dd (96, 16)
-10.13, dd (90, 17)
-10.46, dd (92, 17)
-11.36, dd (93, 32)
-10.03, dd (91, 28)
-11.4, m overlap
28.2 d (12)
29.5 d (15)
-19.8 br d (18)
-21.3 br d (18)
22.1 br d (12)
38.1 br d (12)
-8.2 br s
-12.73, t (15)
8, (R)-BINAPc
-13.41, dd (28, 11)
-13.25, dd (23, 12)
-15.66, dd (23, 17)
-14.04, t (21)
9, (R)-Tol-BINAPc
10, (R,R)-Me-DuPHOS
(b)
maj (a)
min (b)
(a)
(c)
(d)
(b)
i
ii
iii
14, (R,R)-NORPHOSc
15, (S,R)-BPPFAd
16, (S,R)-BPPFAcd
-12.42, dd (20, 14)
-13.54, dd (24, 12)
-14.00, dd (20, 13)
-13.12, t (16)
-12.96, dd (26, 18)
-13.29, t (21)
-12.81, dd (32, 16)
-12.9, m overlap
-12.73, dd (31, 16)
-13.96, dd (28, 17)
-16.5 br s
-5.6 br s
-4.9 br s
-5.6 d (18)
-10.6 d (18)
-19.4 d (18)
-17.47 br s
-18.8 d (20)
-6.3 d (20)
iv
maj
min
-10.89 dd (98, 26)
-11.22 dd (96, 31)
-11.82 dd (96, 28)
4.1 d (20)
1.9 d (20)
a 500 MHz. b 202 MHz. c Isomer designation based on literature assignments for [IrH2(bisphosphine)(COD)](CF3SO3) at -30 °C.9,15,16 d The
absolute configurations of the metal centers were not determined; the ordering of isomers corresponds to the ordering of the ratios presented in
Table 2.
previously have been observed to react with dihydrogen at low
temperatures to give colorless or pale colored [IrH2(phosphine)2-
(COD)]+ complexes.9,15,16 When [Ir(bisphosphine)(COD)]+
complexes containing C2-symmetric chiral bisphosphines such
as (S,S)-CHIRAPHOS, (S,S)-DIOP, (R)-BINAP, (R)-Tol-BI-
NAP, or (R,R)-Me-DuPHOS are reacted at -80 °C with
dihydrogen, two diastereomeric dihydrides can be formed
(Figure 1). The iridium dihydrides exhibit a characteristic P-H
1
coupling pattern in the hydride region of the H NMR (Table
1). As shown in Table 2, the diastereomer ratios depend on
the diphosphine used and the reaction temperatures.
Figure 1. Oxidative addition of H2 to [Ir(bisphosphine)(COD)]+
complexes containing C2-symmetric bisphosphines.
The [Ir(bisphosphine)(COD)]BF4 complexes of C1-symmetric
ligands such as (R,R)-NORPHOS, (S,R)-BPPFA, or (S,R)-
BPPFAc react to yield up to four dihydrides (Figure 2). For
example, [IrH2((S,R)-BPPFA)(COD)]BF4 (15) forms as a mix-
ture of four isomers in an approximate ratio of 25:18:28:1
(Tables 1 and 2). Previously, Ha¨lg and Ru¨egger15,16 reported
similar NMR results for some of the diphosphines listed here
(Tables 1 and 2). The absolute configurations given in Table
2, except for the Me-DuPHOS ligand, are based on the
assignments of Ha¨lg et al.15,16
Kinetic vs Thermodynamic Control of the Oxidative
Addition Reactions. Oxidative addition selectivity varies with
temperature. To probe whether the selectivity was under
thermodynamic or kinetic control, a series of warming experi-
ments was performed. In a typical experiment, the dihydride
was prepared at -80 °C and then warmed in a -45 °C bath,
and NMR spectra were taken at -80 °C (Figure 3). The
dihydride product ratios did not change substantially when the
samples were stored at -80 °C. Results of the warming
experiments are given in Table 2. For example, when [IrH2-
(Me-DuPHOS)(COD)]BF4 was warmed at -45 °C, the initial
47:1.0 ratio of diastereomers was found to equilibrate to 5.8:
1.0 (Table 2 and Figure 3). The equilibration occurred without
any noticeable decomposition of the dihydride complexes.
Oxidative additions of hydrogen to [Ir((S,S)-CHIRAPHOS)-
(COD)]BF4 and [Ir((R,R)-NORPHOS)(COD)]BF4 (Table 2) are
Figure 2. Oxidative addition of H2 to [Ir(bisphosphine)(COD)]+, where
the bisphosphine is C1-symmetric.
of hydrogen to a series of [Ir(bisphosphine)(COD)]BF4 com-
plexes containing chiral C2- and C1-symmetric bisphosphines
(Figures 1 and 2). This is followed by an in-depth study of the
oxidative addition of H2 to [Ir(R,R)-Me-DuPHOS)(COD)]BF4.
Included are the determination of the solution structure and abso-
lute configuration of the major dihydride and the determination
of the mechanisms of diastereomer and isotopic label exchange.
Results and Discussion
Formation of IrH2(bisphosphine)(COD)]BF4 Complexes.
Complexes of the type [Ir(phosphine)2(COD)]+ have been
known for several decades.12-14 These highly colored species
(14) Ha¨lg, W. J.; O¨ hrstro¨m, L. R.; Ru¨egger, H.; Venanzi, L. M. HelV.
Chim. Acta 1993, 76, 788.
(12) Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1971, 93, 3089.
(13) Haines, L. M.; Singleton, E. J. Chem. Soc., Dalton Trans. 1972,
1891.
(15) Ha¨lg, W. J.; O¨ hrstro¨m, L. R.; Ru¨egger, H.; Venanzi, L. M. Magn.
Reson. Chem. 1993, 31, 677.
(16) Ha¨lg, W. J. Ph.D. Dissertation, ETH-Zu¨rich, 1994.

Addition of H2 to [Ir(bisphosphine)(COD)]+ Complexes
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10117
Table 2. Selectivity of Oxidative Addition of H2 to [Ir(bisphosphine)(COD)]BF4
kinetic ratio
at -80 °Ca
(configuration at Ir)
thermodynamic ratio after
equiliibration at -45 °C
(configuration at Ir)
Ha¨lg product
ratio at -30 °Cf
(configuration at Ir)
twist distortion
of square planar
diene complex
[IrH2(P-P)(COD)]BF4,
P-P )
complex
6
7
8
(S,S)-DIOP
(S,S)-CHIRAPHOS
(R)-BINAP
27:1.0
1.0:1.8
>50:1
>50:1
47:1.0 (Λ, OC-6-23)
26:23:3:1
25:18:28:1b
1.3:1.0
ca. 18:1d
6.1:1.0
13:1 (Λ)
5:1 (Λ)
+9.2g
ca. >50:1.0e
ca. >50:1.0e
5.8:1.0 (Λ)
15:6:1.0:1.2
ca. 7.1:1.6:1.0:1.0d
ca. 1:19d
99:1 (∆)
-14.9h
9
(R)-Tol-BINAP
(R,R)-Me-DuPHOS
(R,R)-NORPHOSc
(S,R)-BPPFAc
10
14
15
16
+15.6
-4.0i
7:1 (∆)
(S,R)-BPPFAcc
a Ratio determined by 1H NMR spectra taken at -80 °C unless otherwise indicated. b Ratio determined by 31P NMR spectra taken at -80 °C due
to overlap in the 1H NMR. c C1-symmetric diphosphines yield four dihydride isomers. d Dihydride ratio changes, but decomposition occurs;
thermodynamic equilibrium is estimated. e Decomposition of the major dihydride occurs, and a second dihydride isomer cannot be characterized;
f
thermodynamic equilibrium is estimated. Results as determined by Ha¨lg and co-workers.15,16 g Cambridge refcode: OCPBRH. h Cambridge refcode:
BNAPRH10. i Cambridge refcode: HATJOB.
Figure 3. Warming experiment for [IrH2(Me-DuPHOS)(COD)]BF4.
also under kinetic control. On warming of a solution of [IrH2-
((R,R)-NORPHOS)(COD)]BF4 (13) to -45 °C, the initial kinetic
ratio of 26:23:3:1 gave a thermodynamic ratio of 15:6:1.0:1.2.
The CHIRAPHOS analogue displayed an even greater dif-
ferentiation between kinetic and thermodynamic control. The
kinetically favored CHIRAPHOS diastereomer 7maj (product
ratio 1.8:1.0) was found to be the thermodynamically disfavored
product (1:6.1 ratio) at -45 °C. Under similar conditions,
Ha¨lg16 reports substantial decomposition of [IrH2((S,S)-CHIRA-
PHOS)(COD)]+; the origin of this discrepancy with our results
is unclear.
The remaining dihydrides (6, 8, 9, 15, and 16) decompose to
varying degrees on warming to -45 °C. The BINAP-based
dihydrides 8 and 9 decomposed slowly to give complex mixtures
which could not be fully characterized. The DIOP analogue
decomposed somewhat more rapidly, yielding a single species
containing one hydride resonance, which has not been fully
characterized. The decomposition of the dihydride is ac-
companied by a moderate decrease of the product ratio from
27:1 to 18:1. Similarly, the [IrH2(BPPFA)(COD)]BF4 15
reacted to give a single apparent monohydride in solution,
accompanied by a variation of the dihydride ratio (product ratio
decreased from 25:18:28:1 to 7.1:1.6:1.0:1.0). Because of the
decomposition reactions at -45 °C, we are only able to estimate
the thermodynamic selectivities.
At -80 °C, the oxidative additions of H2 to [Ir(bisphosphine)-
(COD)]BF4 appear to be under kinetic control. This result is
in agreement with Eisenberg’s studies of the related dihydride
systems IrBr(CO)(CHIRAPHOS) and IrX(CO)(DIPHOS) (X )
Cl, Br, I, CN, PPh3, H).17-20 Since the reactions are under
Figure 4. Crystal structure of [Ir((R,R)-Me-DuPHOS)(COD)]BF4 (5).
(a) Top view and (b) side view showing the planarity of the phenylene
-
unit. Hydrogens and BF4 counterion were omitted for clarity.
kinetic control, molecular modeling calculations, which are
based on the stabilities of the dihydride compounds and their
intermediates, will not necessarily model the reaction selectiv-
ity.21
Crystallographic Structure of [Ir((R,R)-Me-DuPHOS)-
(COD)]BF4. [Ir(Me-DuPHOS)(COD)]BF4 (5) was isolated as
dark brown needles from CH2Cl2/Et2O. The crystal structure
of 5 is shown in Figure 4, and relevant bond distances and angles
are given in Table 3. This structure, which is very similar to
that of [Rh(COD)((S,S)-Me-DuPHOS)]SbF6,22 provides impor-
tant information about the asymmetric environment at the Ir
(19) Johnson, C. E.; Eisenberg, R. J. Am. Chem. Soc. 1985, 107, 3148.
(20) Deutsch, P. P.; Eisenberg, R. Chem. ReV. 1988, 88, 1147.
(21) Giovannetti, J. S.; Kelly, C. M.; Landis, C. R. J. Am. Chem. Soc.
1993, 115, 4040-4057.
(22) Burk, M. J.; Feaster, J. E.; Nugent, W. A.; Harlow, R. L. J. Am.
Chem. Soc. 1993, 115, 10125.
(17) Kunin, A. J.; Farid, R.; Johnson, C. E.; Eisenberg, R. J. Am. Chem.
Soc. 1985, 107, 5315.
(18) Johnson, C. E.; Fisher, B. F.; Eisenberg, R. J. Am. Chem. Soc. 1983,
105, 7772.

10118 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Kimmich et al.
Table 3. Selected Interatomic Distances and Intramolecular
Angles for [Ir((R,R)-Me-DuPHOS)(COD)]BF4
Table 4. Comparison of the Selectivity of the
[Rh(P-P)(NBD)]+-Catalyzed Hydrogenation of Prochiral Enamides
with the Kinetic Selectivity of H2 Addition to [Ir(P-P)(COD)]+
Complexes
Interatomic Distances (Å)
Ir(1)-P(1)
Ir(1)-P(2)
Ir(1)-C(1)
Ir(1)-C(2)
2.2839(8) Ir(1)-C(5)
2.214(4)
2.188(4)
1.373(6)
1.377(7)
ligand (P-P) and
kinetic selectivity of H2
addition to [Ir(P-P)(COD)]+
enatiomeric
ratio of
the product ref
2.2796(8) Ir(1)-C(6)
2.233(3)
2.197(3)
C(1)-C(2)
C(5)-C(6)
substratea
(S,S)-CHIRAPHOS, 1.8:1 (Λ) +
R ) Me; R′ ) Pr 10.1:1 (R:S) 23
Intramolecular Nonbonding Distances (Å)
R ) Me; R′ ) H
R ) Me; R′ ) iPr
R ) Me; R′ ) Ph
R ) Me; R′ ) H
R ) Ph; R′ ) Ph
8.5:1
5.6:1
24
23
C(1)-C(14)
C(5)-C(19)
C(13)-C(22)
3.539
3.589
3.151
C(20)-C(25)
C(4)-C(8)
3.497
3.658
(S,S)-DIOP, 27:1 (Λ)
5.4:1 (S:R) 25
4.0:1
99:1 (S:R)
26
27
Intramolecular Angles (deg)
85.33(3) C(2)-Ir(1)-C(5)
C(2)-Ir(1)-C(6)
(R)-BINAP, >50:1 (∆) -
P(1)-Ir(1)-P(2)
80.0(1)
89.1(2)
(R,R)-Me-DuPHOS, 47:1 (Λ) +
R ) Me; R′) H 199:1 (R:S)
22
C(1)-Ir(1)-C(5) 93.3(2)
C(1)-Ir(1)-C(6) 80.1(2)
(R,R)-NORPHOS, 26:23:3:1 (∆) - R ) Me; R′ ) Pr
8.5:1 (S:R) 23
1.5:1 (S:R) 24
1.2:1 (S:R) 28
(S,R)-BPPFA, 25:18:28:1
(S,R)-BPPFAc, 1.3:1
R ) Me, R′ ) Ph
R ) Me, R′ ) Ph
Ir(1)-P(2)-C(26) 109.7(1)
a
atom. Both the Ir-P bond lengths and the P-Ir-P bond angle
(85°) are similar to those of the related [Ir(COD)((-)-NOR-
PHOS)]PF6 species.15 As in [Rh(COD)((S,S)-Me-DuPHOS)]-
SbF6, the 1,2-phenylene ring unit is essentially planar, and only
a 6.3° dihedral twist out of the P-Ir-P plane is observed. The
rigidity of this unit positions the phospholane methyl groups
close to the Ir coordination sphere, allowing for efficient
stereochemical communication. The large dihedral angle of
15.6° observed between the P-Ir-P plane and the plane defined
by the COD olefin midpoints is again similar to the 18° dihedral
angle seen in the rhodium analogue.
Correlation of Oxidative Addition Selectivity with the
Crystallographic Structures of [Ir(bisphosphine)(diolefin)]+
Complexes. In previous investigations of the enamide hydro-
genation using complexes of the Me-DuPHOS ligand, Brown
and Burk11 noted an interesting correlation of the direction of
torsional twist between planes defined by P-Rh-P and
centroidCdC-Rh-centroidCdC positions with the sense of chiral-
ity in enamide hydrogenation. Using the definition of “twist
angle” diagrammed below, we have examined the correlation
the hydrogenation of prochiral enamides (Table 4). High
selectivities are observed in both oxidative addition and enamide
hydrogenation for BINAP, Tol-BINAP, and Me-DuPHOS
complexes. In contrast, the CHIRAPHOS-based complex 2
exhibits poor selectivity in oxidative addition of dihydrogen
(diastereomeric product ratio 1.8:1.0) but significant selectivity
as an enamide hydrogenation catalyst (enantiomeric product ratio
) 10.1:1).
Stereochemistry of [IrH2(Me-DuPHOS)(COD)]BF4 (10maj).
The absolute configuration of the major diastereomer of [IrH2-
((R,R)-Me-DuPHOS)(COD)]BF4 (10maj) was determined using
a combination of 1H-1H DQF COSY and 1H-1H NOESY 2D
NMR experiments. The structure of 10maj is shown in Figure
5, and full 1H assignments are given in Table 5 (31P NMR data
are given in Table 1). In addition to the COSY data, the
configuration of the dihydride could be assigned on the basis
of a number of important NOE interactions, in particular, those
between protons H17-H28, H20-H16, H8-H1, H24-H4,
H23-H18, H27-25A, H27-26A, H27-H19, and H27-H5
(Figure 5). These NOE interactions include those between the
hydrides and methyl H’s as well as NOEs between the methyl
or methine H’s and the phenylene moiety. The analysis of the
2D-NMR data allowed for the assignment of almost all of the
individual protons of 10maj (34 total inequivalent protons). The
methylene protons on the phospholane rings were the most
difficult to assign, and, in several cases, the shifts of the two
protons on a single carbon were essentially identical. While
these protons could not be assigned for the complex in CD2-
Cl2, full assignments could be made for 10maj in acetone-d6 (see
Supporting Information). In the process of analysis, a number
of long-range couplings between H24-H27 and H25B-H27
were found (see Figure 5). Similar long-range couplings were
observed for [IrH2((R,R)-NORPHOS)(COD)]CF3SO3 and [IrH2-
((S,S)-CHIRAPHOS)(COD)]CF3SO3, and these were attributed
to (homo)allylic coupling pathways in which the metal orbitals
play a mediating role.15
between the magnitude and sign of the twist angle with the sign
and extent of diastereoselectivity in the oxidative addition of
dihydrogen (see Table 2, columns 3 and 6) to a variety of [Ir-
(bisphosphine)(COD)]+ complexes. For the DuPHOS, NOR-
PHOS, and BINAP ligands, the kinetically preferred dihydride
diastereomer correlates with the sign of twist angle according
to (+)twistTΛ and (-)twistT∆. However, for CHIRAPHOS,
the kinetically preferred dihydride diastereomer slightly prefers
the ∆ configuration at the metal but has a (+)twist. In all four
cases, the sense of the thermodynamically preferred configu-
ration in the dihydride correlates with the sense of the twist
angle in the solid-state structure of the [Ir(bisphosphine)(COD)]+
complex.
Correlation of Oxidative Addition Selectivity with Enam-
ide Catalytic Hydrogenation Selectivity. Because the sense
of diene twist angle correlates with the sense of product chirality
in enamide hydrogenation,11 as discussed above, we were also
interested to see if the sense and magnitude of kinetically
controlled diastereoselective H2 addition to [Ir(bisphosphine)-
(COD)]+ complexes correlates with enamide enantioselectivity.
The selectivity of oxidative addition of H2 to [Ir(bisphosphine)-
(COD)]BF4 (Table 2) partially correlates to the selectivity of
(23) Scott, J. W.; Keith, D. D.; Nix, C.; Parish, D. R.; Remington, S.;
Roth, G. P.; Townsend, J. M.; Vanlentine, D.; Yang, R. J. Org. Chem.
1981, 46, 5086.
(24) Hayashi, T.; Mise, T.; Mitachi, S.; Yamamoto, K.; Kumada, M.
Tetrahedron Lett. 1976, 1133.
(25) Glaser, R.; Vainas, B. J. Organomet. Chem. 1976, 121, 249.
(26) Brown, J. M.; Murrer, B. A. Tetrahedron Lett. 1980, 581.
(27) Miyashita, A.; Yasuda, A.; Takaya, H.; Toriymi, K.; Ito, T.; Souchi,
T.; Nayora, R. J. Am. Chem. Soc. 1980, 102, 7932.
(28) Kimmich, B. F. M.; Landis, C. R. Organometallics 1996, 15, 4141.

Addition of H2 to [Ir(bisphosphine)(COD)]+ Complexes
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10119
Table 5. 1H and 13C Data for the Major Diastereomer of [IrH2((R,R)-MeDuPHOS)(COD)]BF4 (10maj) in CD2Cl2 at -80 °C
C no.
δ(13C)
11.5
14.4
16.2
21.9
26.0
JC-P (Hz)
H no.a
JP-H (Hz)
δ(1H)
assignment
17
6
18
5
17
6
18
5
0.67
0.33
1.41
1.26
2.22
1.56
1.62
2.52
2.02
2.35
2.33
2.12
1.98
1.52
1.46
2.73
1.82
2.44
2.81
3.39
2.50
3.55
2.68
3.03
4.67
4.07
4.90
4.30
7.63
7.55
7.59
7.83
Me
Me
Me
Me
22
22B
22A
26B
26A
4
15A/B
2A/B
3A/B
15A/B
3A/B
2A/B
1
14A/B
14A/B
25A
25B
21B
21A
16
13
23
20
19
24
8
9
10
COD CH2
COD CH2
COD CH2
COD CH2
26
27.3
34.3
CHCH3(6)
Me-DuPHOS CH2
Me-DuPHOS CH2
Me-DuPHOS CH2
Me-DuPHOS CH2
Me-DuPHOS CH2
Me-DuPHOS CH2
CHCH3(5)
Me-DuPHOS CH2
Me-DuPHOS CH2
COD CH2
COD CH2
COD CH2
COD CH2
CHCH3(17)
CHCH3(18)
COD vinyl
COD vinyl
COD vinyl
COD vinyl
phenyl
34.3-34.7b
1
14
35.7
36.6
41
25
21
37.9
38.8
16
13
23
20
19
24
8
9
10
39.5
44.9
72.7
79.2
82.2
34
20
84.5
130.4c
130.4c
131.3
132.3
140.1
143.3
phenyl
phenyl
phenyl
ipso
ipso
IrH trans
IrH cis
11
7/12
7/12
14
49, 31
44, 29
11
27
28
88, 16
23, 17
-10.48d
-15.74d
a Differentiation of H’s (A/B) on a single C for the phospholane CH2 groups was not possible in CD2Cl2. b Peaks are not resolved in the carbon
dimension of the HMQC. c Assignment of the ipso carbons is ambiguous. d The 1H NMR shifts above were derived from the DQF COSY. These
1
differ slightly from those derived from the 1D H NMR.
dispersed, and only the region from 34.3 to 34.7 ppm contained
1
significant overlap. Correlation of the H and 13C resonances
was achieved with a 1H-13C HMQC (heteronuclear correlation
through multiple quantum coherence) experiment. The 13C
assignments are given in Table 5. Overlap in the 13C dimension
did not allow for the assignment of all individual CH2 carbons
in the phospholane rings and CH carbons in the phenyl ring.
Assigning the structure of 10min has been complicated by overlap
of its methylene and methyl regions with the major species
10maj
.
Solution Structure Determination of 10maj by Conformer
Population Analysis. The solution structure of the major
diastereomer 10maj was determined by analyzing the time course
1
of the quantitative H 2D NOESY data with the multiconfor-
mational analysis technique, two-dimensional conformation
population analysis (2DCPA).21,29-31 This method requires the
quantitation of NOESY time courses. Initial experiments in
CD2Cl2 at -80 °C were not useable, because both positive and
negative NOE enhancements were observed, indicating that the
isotropic rotational correlation times corresponded to the null
region of NOE enhancement. The NOESY data collected in
acetone-d6 at -90 °C (mixing times of 70, 100, 130, 160, and
200 ms) resulted in enhanced, negative NOEs.
Figure 5. Dominant solution conformation of [IrH2(Me-DuPHOS)-
(COD)]BF4, 10maj, in acetone-d6 as determined by 2DCPA. The
absolute configuration of 10maj is shown, along with labels for all carbon
atoms. Selected H’s are also labeled, particularly those involved in
the NOE interactions which allowed for the structure determination.
(29) Landis, C.; Allured, V. S. J. Am. Chem. Soc. 1991, 113, 9493-
9499.
(30) Landis, C. R.; Luck, L. L.; Wright, J. M. J. Magn. Reson. Ser. B
1995, 109, 44-59.
(31) Casey, C. P.; Hallenbeck, S. L.; Wright, J. M.; Landis, C. R. J.
Am. Chem. Soc. 1997, 119, 9680.
The assignment of the 1H NMR was aided by our elucidation
.
of the 13C NMR of 10maj The 13C NMR was quite well

10120 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Kimmich et al.
Figure 6. Molecular mechanics energies of all conformers used in
the 2DCPA analysis. The two conformers obtained by quantitative
analysis of the NOE data are illustrated.
The 2DCPA analysis correlates the observed NOESY intensi-
ties or the experimental cross-relaxation rates at each mixing
time to the calculated NOE data from a set of energetically
reasonable trial models created by molecular mechanics.21,29-31
Unlike the single, average solution approach, the best repre-
sentation of 2DCPA solution structures is given by the smallest
combination of statistically significant conformers and their
populations. The goal of this 2DCPA analysis is to investigate
the detailed structure in space of the major diastereomer of the
dihydride complexes 10maj. Our analysis involved three steps:
(1) collection of NOESY time courses, (2) creation of an
ensemble of trial solution structures using molecular mechanics,
and (3) conformer population analysis of the observed NOESY
time courses to find the smallest combination of the trial
structures and their populations. The details of this analysis
are provided in the Experimental Section.
Quantitative analysis of the NOE data confirms the assign-
ment of absolute stereochemistry of 10maj in solution. The
structure of 10maj is highly constrained by the multiple ring
elements. The primary conformational degree of freedom
corresponds to the orientation of the bridging phenylene ring
with respect to the hydride ligand which is cis to both
phosphorus atoms, which we will call the axial hydride. Our
analysis of the NOE time courses finds that the most likely
conformation has the plane of the phenylene ring essentially
coplanar with the plane defined by the two P and the Ir atoms.
This structure is shown in Figures 5 and 6. Statistically, the
data are fit significantly better when two conformations are used;
the secondary conformer is shown in Figure 6. The secondary
conformer has the phenylene ring rotated toward the axial
hydride. According to molecular mechanics computations, both
conformers have energies within 9 kcal/mol of the lowest energy
conformation (see Figure 6).
Figure 7. Pathways for the mechanism of diastereomer interconversion
for [IrH2(Me-DuPHOS)(COD)]BF4.
simplify, we will refer to this approach to equilibrium as
“diastereomer exchange”.
For pathway A, the rate of diastereomer interconversion
should be first order in [Ir]tot and independent of the H2
concentration. Pathway B is kinetically distinguishable from
all other pathways because it leads to a rate of diastereomer
interconversion that is second order in [Ir]tot. Based on the high
thermodynamic stability of the dihydride complexes, 10maj and
10min, pathway B should exhibit an inverse dependence on the
[H2] because the concentration of the four-coordinate catalyst
concentration decreases with increasing [H2]. Pathway C
requires a first-order dependence of the diastereomer intercon-
version on [H2]. Kinetically, pathway D is indistinguishable
from pathway A. However, isotopic scrambling results dis-
criminate between pathways A and D (vide infra).
The diastereomer exchange of [IrH2(Me-DuPHOS)(COD)]-
1
BF4 at -45 °C was monitored by H NMR or 31P NMR as a
function of [Ir]tot and PH (Table 6). Within the precision of
2
the data, the rate is invariant of the [Ir]tot and PH . Thus,
2
pathways B and C are inconsistent with the observed kinetics.
Isotope exchange experiments (Figure 8) of H2 with d2-10maj
and d2-10min permit discrimination between pathways A and
D. If pathway A is operatiVe, the exchange of H2 with d2-10maj
must occur more rapidly than the equilibration of 10maj with
10min. Each time D2 is eliminated from d2-10maj under an
atmosphere of H2, isotope exchange will occur by the subsequent
addition of H2. The reaction of H2 with [Ir(COD)(Me-
DuPHOS)]+ at -45 °C is kinetically biased to produce 10maj
at least 10 times faster than 10min. Therefore, for 10 occurrences
of D2 elimination from d2-10maj, at most only one will lead to
diastereomer interconversion (i.e., produce d2-10min), but all 10
will lead to H2 incorporation. The observation of near
equivalent rates for diastereomer interconversion and incorpora-
tion of H2 into a solution initially consisting of >97% d2-10maj
(Table 6) precludes significant flux along pathway A.
Kinetic Studies of Isotope Exchange and Diastereomer
Interconversion of [IrH2(COD)(Me-DuPHOS)]BF4. The
dynamics of dihydride diastereomer interconversion are poten-
tially important in determining the selectivities of catalytic
asymmetric hydrogenation reactions. Several possible mech-
anisms for dihydride diastereomer interconversion are shown
in Figure 7. In the [IrH2(COD)(Me-DuPHOS)]BF4 system, the
diastereomer interconversion is an approach to equilibrium. To

Addition of H2 to [Ir(bisphosphine)(COD)]+ Complexes
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10121
Table 6. Kinetic Studies of Diastereomer Interconversion and Isotope Exchange for [IrH2(Me-DuPHOS)(COD)]BF4 (10)
experiment
solvent
observed nucleus
[Ir] (mM)
PH2 (atm)
T (°C)
t1/2 (min)
10maj h 10min
10maj h 10min
10maj h 10min
10maj h 10min
10maj h 10min
d2-10maj + H2
d2-10maj + H2
CD2Cl2
CD2Cl2
CD2Cl2
CD2Cl2
acetone-d6
CD2Cl2
CD2Cl2
CD2Cl2
acetone-d6
CD2Cl2
1H
1H
31P
31P
1H
1H
1H
1H
1H
2H
27
55
55
27
35
55
27
27
35
52
0
0
1
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
47 ( 10
47 ( 10
43 ( 10
62 ( 12
64 ( 12
47 ( 10
51 ( 10
52 ( 10
55 ( 12
56 ( 12
0.083
0
1
0.083
d2-10maj (H incorporation into M-D)
d2-10maj (H incorporation into M-D)
d2-10maj (D incorporation into H1 and H25A)
0
0
0
Figure 9. The 126-MHz 13C NMR spectra of [IrD2(Me-DuPHOS)-
(COD)]BF4 at -80 °C (a) right after formation of the dihydride at -80
°C and (b) after deuterium incorporation into sites on C25 and C1 after
equilibration at -45 °C. Deuterium incorporation sites are indicated
Figure 8. Isotope exchange experiment for [IrD2(Me-DuPHOS)(COD)]-
BF4 (10).
By elimination, only pathway D of the four pathways
illustrated in Figure 7 is consistent with the observed data (we
note that other, more complicated pathways are possible). These
results contrast Deutsch and Eisenberg’s identification of
diastereomer exchange pathways similar to A and B for
Ir(diphosphine)(CO)X (X ) halide) complexes.20
with large arrows, and asterisks indicate peaks which are due to 10min
.
and H1. Therefore, exchange of M-D with ligand C-H bonds
is exquisitely selective.
Given the constraints of the rate data for diastereomer and
isotope exchange, it is difficult to construct a simple mechanism.
The primary issues are (1) why are the isotope exchange rates
and diastereomer exchange rates equal (within experimental
resolution), and (2) why does isotope exchange proceed
exclusively into two ligand C-H bonds? We favor variant ii
of pathway D, in which the primary kinetic event corresponds
to formation of a five-coordinate intermediate. Three reasonable
five-coordinate intermediates are (1) a formally Ir(I) molecular
hydrogen complex, (2) a formally Ir(III) complex with a
dangling diphosphine ligand, and (3) a formally Ir(III) complex
with a dangling COD ligand. The latter two are coordinatively
unsaturated, 16-electron complexes which might be expected
to activate ligand C-H bonds to isotope exchange via oxidative
addition, yielding 18-electron Ir(V) complexes. One possible
scheme is detailed in Figure 10 below.
A surprising result of our investigations was the observed
growth of M-H resonances from solutions of d2-10maj and d2-
10min, even in the absence of added H2. Because there was no
net decomposition of the dihydride complexes and the solvent
was perdeuterated, this result suggests that H and D are
exchanged among M-H and ligand C-H sites. Confirmation
of such exchange is provided by 2H NMR and 13C NMR
experiments. A solution of d2-10maj at -45 °C under N2 yields
growth of two broad resonances at δ 2.5 and 2.8 ppm in the 2H
NMR spectrum and decay of the M-D resonances at δ -9.8
to -15.7 ppm. After approximately 4 h, the resonance
intensities cease to change, indicating that equilibrium has been
reached. At equilibrium, the ratio of D resonance intensities
for total M-D versus total C-D is 1.3 ( 0.4. Apparently,
there is no substantial isotope effect. The half-life for this
approach to equilibrium is similar to those for isotope label
exchange under an H2 atmosphere and for diastereomer inter-
conversion. There is no apparent difference in the rates of
deuterium label scrambling for the two diastereomers, as
determined from comparison of diastereomer ratios measured
by 31P NMR with the ratio of hydride resonances (major:minor)
For simplicity of discussion, it is easiest to focus on the
formation of 10maj and 10min from essentially pure d2-10maj
under a H2 atmosphere. Conversion of 10maj to 17maj via
decoordination of the CdC bond located cis to the Ir-D bond
is slow. From coordinatively unsaturated 17maj, oxidative
addition of the allylic C-H bond yields an η3-allyl trihydride
(18maj). Subsequent elimination of HD forms the η3-allyl
monohydride (19) and provides an entry for scrambling with
external H2 or D2. Conversion to the minor diastereomer may
occur in two ways: by rotation about the Ir-allyl centroid of
10maj to generate 10min followed by insertion of the allyl into
Ir-D bond to yield 17min, or by readdition of HD to 19 (after
rotation about the Ir-allyl centroid). A similar mechanism (not
shown) scrambles Ir-D with the methine C-H of the diphos-
1
determined by H NMR.
Due to the broadness of the 2H NMR resonances, 13C NMR
was employed to determine which ligand sites were exchange
active. The decrease in intensity and broadening of peaks at δ
35.76 (d, JP-H ) 41 Hz) and 37.89 (s) ppm in the 13C NMR
(Figure 9) demonstrates exclusive (within sensitivity limits)
incorporation of deuterium into the COD methylene (C25) and
the Me-DuPHOS methine (C1) positions. These changes are
consistent with approximately 50% exchange of D with H25A
phine ligand. From the coordinatively unsaturated 17maj
oxidative addition of the methine C-H bond to form an
,

10122 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Kimmich et al.
addition of H2 to [Ir(chiral diphosphine)(COD)]+ complexes
does not correlate well with enamide enantioselectivities.
The diastereomeric dihydrides exhibit surprisingly facile
interconversion and deuterium scrambling pathways. These
pathways do not involve reductive elimination of dihydrogen
to regenerate the square planar, four-coordinate starting complex.
The significance of these mechanistic results is that they
demonstrate the potential complexity of the dynamics of
dihydride intermediates.
Experimental Section
General Procedures. All procedures were carried out under a dry
atmosphere of nitrogen by using standard Schlenk techniques and an
inert atmosphere glovebox. Glassware was oven-dried before use.
Nitrogen (Liquid Carbonic), hydrogen (Liquid Carbonic), 8.3% hy-
drogen (balance N2, Liquid Carbonic), and deuterium (Cambridge
Isotope) were used without further purification. NMR solvents, CD2-
Cl2 and acetone-d6, were used as obtained from Cambridge Isotope
and Isotech Inc., respectively. Methylene chloride was distilled from
P2O5, and acetone was dried over molecular sieves and distilled. Diethyl
ether was freshly distilled from sodium/benzophenone ketyl.
Figure 10. Proposed mechanism of diastereomer interconversion and
isotopomer exchange for [IrD2(Me-DuPHOS)(COD)]BF4 (10).
iridaphosphacyclopropane trihydride (20) permits the requisite
D-scrambling and diastereomer interconversion processes.
32
IrCl3 was obtained from Johnson-Mathey. [Ir(COD)Cl]2 and Ir-
33
(COD)2]BF4 were prepared according to literature procedures. (R)-
BINAP, (R,R)-NORPHOS, (R)-Tol-BINAP, (S,S)-CHIRAPHOS, and
(R,R)-Me-DuPHOS were used as received from Strem Chemicals. (S,R)-
BPPFAc and (S,R)-BPPFA were prepared according to literature
procedures.34
1
2
Low-temperature H NMR (500 MHz), H{1H} NMR (76 MHz),
31P{1H} NMR (202 MHz), and 13C{1H} NMR (126 MHz) were
performed on a Bruker AM-500 spectrometer. Room-temperature 1H
NMR (300 MHz) and 31P{1H} NMR (121 MHz) were taken on a Bruker
AC-300 spectrometer. 1H DQF COSY and NOESY NMR were
The virtues of this mechanistic proposal are that it (1) yields
similar rates for approach to equilibrium of both diastereomer
interconversion and isotope exchange, (2) accounts for inde-
pendence of these rates on H2 pressure and [Ir]tot, (3) yields
similar H-D isotope ratios for both the major and minor
diastereomers throughout the diastereomer interconversion
process, and (4) accounts for highly selective incorporation of
D into just two ligand sites. Evidence for this type of a
mechanism could be found by trapping the intermediate 17 with
a ligand such as P(OMe)3. Unfortunately, the phosphite was
not a good trap for this reaction and gave rise to complexes
which contained neither hydrides nor cyclooctadiene.
performed on a Varian Unity-500 spectrometer with a 5-mm H/19F
1
probe. 1H/13C HMQC NMR was performed on a Varian Unity-500
spectrometer with a 5-mm inverse broadband probe. Probe tempera-
tures were calibrated using a methanol standard.
[Ir(P-P)(COD)]BF4 Preparation. [Ir(P-P)(COD)]BF4 complexes
were prepared in situ from [Ir(COD)2]BF4, a deep purple solid, and
the corresponding bisphosphine. [Ir(COD)2]BF4 was found to decom-
pose as a solid, even when stored at 0 °C in the absence of light, and
should be used soon after preparation.
Conclusion
[Ir(COD)((S,S)-DIOP)]BF4 (1). A solution of (S,S)-DIOP (10.0
mg, 2.0 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was added dropwise
to a solution of [Ir(COD)2]BF4 (9.9 mg, 2.0 × 10-5 mol, 1 equiv in
0.5 g of CD2Cl2), yielding a bright red solution. Data for [Ir((S,S)-
DIOP)(COD)]BF4 (1) at room temperature: 1H NMR (CD2Cl2, 300
MHz) δ 1.09 (m, 6H, CH3), 1.87-2.35 (m, 17H, free and bound COD
CH2’s), 2.70 (m, DIOP CH2’s), 3.13 (m, DIOP CH2’s), 3.80 (m, 2H,
DIOP CH’s), 4.00 (m, 2H, bound COD vinyl), 4.23 (m, 2H, bound
vinyl), 5.55 (m, 4H, free COD vinyls), 7.27-7.67 (m, 20H, phenyls)
ppm; 31P NMR (CD2Cl2, 121 MHz) δ 4.8 (s) ppm.
[Ir((S,S)-CHIRAPHOS)(COD)]BF4 (2). A solution of (S,S)-
CHIRAPHOS (9.6 mg, 2.3 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2)
was added dropwise to a solution of [Ir(COD)2]BF4 (11.2 mg, 2.3 ×
10-5 mol, 1 equiv in 0.5 g of CD2Cl2), yielding a deep green solution.
1H and 31P{1H} NMR for [Ir((S,S)-CHIRAPHOS)(COD)]BF4 (2) at
room temperature were consistent with previously published results:15
1H NMR (CD2Cl2) δ 1.06 (m, 6H, CH3), 1.75-1.98 (m, 4H, COD CH2),
2.15-2.55 (m, 14H, 2 CH’s and free and bound COD CH2), 4.02 (m,
2H, bound COD vinyls), 4.82 (m, 2H, bound COD vinyl), 5.55 (m,
4H, free COD), 7.38-7.86 (m, 20H, phenyl) ppm; 31P{1H} NMR δ
46.4 ppm.
The examination of the formation and dynamics of iridium
dihydrides provides useful insights into the selectivity of
authentic H2 oxidative addition reactions. Our studies conclu-
sively demonstrate that the kinetic and thermodynamic selectivi-
ties of oxidative addition are not the same. Indeed, for at least
one complex, [Ir(CHIRAPHOS)(COD)]+, the thermodynamic
and kinetic selectivities of dihydrogen oxidative addition have
the opposite sense. As a result, one cannot expect models of
dihydride stabilities derived from, for example, molecular
mechanics, to necessarily mirror the diastereoselectivity of the
oxidative addition transition states. Thus, it is not surprising
that previous molecular modeling studies of putative dihydride
intermediates in catalytic asymmetric hydrogenation of enamides
failed to find significant differences in stabilities between the
diastereomeric dihydrides.21 For the [IrH2(Me-DuPHOS)-
(COD)]+ complexes studied here, we have been able to establish
the solution structures, including the absolute configuration,
using NOESY methods and the 2DCPA program. The kineti-
cally preferred dihydride diastereomer is that in which the COD
ligand is twisted in the same direction as is seen in the distorted
square planar geometry of the reactant [Ir(Me-DuPHOS)-
(COD)]+. However, one must be cautious in extending this
explanation to control of the enantioselectivity of enamide
hydrogenation: the extent of diastereoselectivity observed for
(32) Crabtree, R. H.; Quirk, J. M.; Felkin, H.; Fillebeen-Khan, T. Synth.
React. Inorg. Met.-Org. Chem. 1982, 12, 407.
(33) Schenck, T. G.; Downes, J. M.; Milne, C. R. C.; MacKenzie, P. B.;
Boucher, H.; Whelan, J.; Bosnich, B. Inorg. Chem. 1985, 24, 2334.
(34) Hayashi, T.; Mise, T.; Fukushima, M.; Kagotani, M.; Nagashima,
N.; Hamada, Y.; Matsumoto, A.; Kawakami, S.; Konishi, M.; Yamamoto,
K.; Kumada, M. Bull. Chem. Soc. Jpn. 1980, 53, 1138.

Addition of H2 to [Ir(bisphosphine)(COD)]+ Complexes
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10123
Table 7. Crystallographic Data for [Ir((R,R)-Me-DuPHOS)(COD)]BF4 (5)
formula
fw
cryst color, habit
cryst size, mm
cryst syst
space group
a, Å
[C26H40IrP2]+[BF4]-
693.53
brown opaque needle
0.50 × 0.10 × 0.10
orthorhombic
P212121
F
calc, g cm-3
T, K
1.709
133(2)
1376
Mo KR; 0.710 73
Siemens P4/CCD
58.4
12 902
6419
307
0.018
F(000)
radiation; λ, Å
diffractometer
2θ(max), deg
no. reflctns collected
no. reflctns used
no. params refined
R(F), observed data
wR(F2), all data
goodness of fit on F2
9.1569(2)
b, Å
c, Å
16.0222(2)
18.3723(3)
90°
2695.47(8)
4
R ) â ) γ
V, Å3
0.045
1.038
Z
[Ir((R)-BINAP)(COD)]BF4 (3). A solution of (R)-BINAP (12.6
mg, 2.0 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was added dropwise
to a solution of [Ir(COD)2]BF4 (10.0 mg, 2.0 × 10-5 mol, 1 equiv in
0.5 g of CD2Cl2), yielding a very dark red solution. Data for [Ir((R)-
BINAP)(COD)]BF4 (3) at room temperature: 1H NMR (300 MHz, CD2-
Cl2) δ 1.80-1.96 (m, 3H, bound COD CH2’s), 2.07-2.41 (m, 13H
free COD and bound CH2’s), 4.24 (m, 2H, bound vinyl), 4.47 (m, 2H,
bound vinyl), 5.55 (m, 4H free COD vinyl), 6.46-7.90 (m, 32H,
phenyls) ppm; 31P NMR (CD2Cl2) δ 17.8 (s) ppm.
[Ir(COD)((R)-Tol-BINAP)]BF4 (4). A solution of (R)-Tol-BINAP
(10.9 mg, 1.6 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was added
dropwise to a solution of [Ir(COD)2]BF4 (8.0 mg, 1.6 × 10-5 mol, 1
equiv in 0.5 g of CD2Cl2), yielding a brown-green solution. Data for
[Ir((R)-Tol-BINAP)(COD)]BF4 (4) at room temperature: 1H NMR (300
MHz, CD2Cl2) δ 1.76 (br s, 2H, bound COD CH2’s), 1.90 (s, 6H, 2
CH3’s), 2.28 (br s, 2H, bound COD CH2’s), 2.16-2.42 (m, 20H, free
and bound COD CH2’s, 2 CH3’s), 4.16 (br s, 4H, bound COD vinyls),
5.49 (br s, 4H, free COD vinyl), 6.06-7.83 (m, 28H, phenyls) ppm;
31P NMR (CD2Cl2) δ 16.5 (s) ppm.
[Ir((R,R)-Me-DuPHOS)(COD)]BF4 (5). A solution of (R,R)-Me-
DuPHOS (6.2 mg, 2.0 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was
added dropwise to a solution of [Ir(COD)2]BF4 (10.0 mg, 2.0 × 10-5
mol, 1 equiv in 0.5 g of CD2Cl2), yielding a deep brown solution.
Assignments for [Ir((R,R)-Me-DuPHOS)(COD)]BF4 (5) are listed below
for the isolated complex. Procedures for the isolation of [Ir((R,R)-
Me-DuPHOS)(COD)]BF4 are also given below.
[Ir((R,R)-NORPHOS)(COD)]BF4 (11). A solution of (R,R)-
NORPHOS (15.0 mg, 3.2 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2)
was added dropwise to a solution of [Ir(COD)2]BF4 (16.0 mg, 3.2 ×
10-5 mol, 1 equiv in 0.5 g of CD2Cl2), yielding a deep brown-green
solution (11). 1H and 31P{1H} NMR were consistent with previously
published results:15 1H NMR (500 MHz, CD2Cl2, -80 °C) δ 0.78 (d,
1H, J ) 8 Hz, NORPHOS bridging CH2), 1.53 (m, 2H, bound CH2
COD), 1.78 (m, 3H, bound CH2 COD and NORPHOS bridging CH2),
2.30-2.60 (m, 13H, free COD, 4 bound COD CH2, and 1 CHPPh3),
2.69 (m, 1H, NORPHOS bridgehead), 3.00 (m, 1H, NORPHOS
bridgehead), 3.06 (m, 1H, CHPPh3), 3.76 (m, 2H, bound COD vinyl),
4.85 (m, 2H, bound COD vinyl), 5.35 (m, 1H, NORPHOS vinyl), 5.53
(m, 4H, free COD vinyl), 6.10 (m, 1H, NORPHOS vinyl), 7.32-7.75
(m, 20H, phenyl) ppm; 31P NMR (CD2Cl2, 202 MHz, -80 °C) δ 11.1
(d, J ) 15 Hz), 12.2 (d, J ) 15 Hz) ppm.
MHz, CD2Cl2) δ 1.50 (d, 3H, J ) 6.0 Hz, CHCH3), 1.75-1.94 (m,
4H, bound COD CH2’s), 2.00 (s, 3H, OAc), 2.11-2.55 (m, 12H, free
and bound COD), 3.98-4.51 (m, 10H, Cp’s and bound COD vinyls),
4.82 (m, 1H, 1 Cp), 5.55 (br s, 4H, free COD vinyl), 6.45 (q, J ) 6.0
Hz, CHCH3), 7.32-8.11 (m, 20H, phenyls); 31P NMR (CD2Cl2) δ 14.5
(d, JP-H ) 16 Hz), 17.7 (d, JP-H ) 16 Hz) ppm.
[Ir(COD)((R,R)-Me-DuPHOS)]BF4 (5) Isolation. [Ir(COD)2]BF4
(300.0 mg, 0.605 mmol) was placed into a Schlenk flask and dissolved
in 30 mL of freshly distilled CH2Cl2. A solution of (R,R)-Me-DuPHOS
(185.5 mg, 0.605 mmol, 1 equiv) in CH2Cl2 (30 mL) was added
dropwise over 15 min. The solution volume was reduced to ap-
proximately 5 mL, and the solution was layered with 10 mL of freshly
distilled diethyl ether. Brown needles formed overnight in the absence
of light (254 mg, 0.366 mmol, 61% yield): 1H NMR (300 MHz, CD2-
Cl2) δ 1.01 (dd, 6H, J ) 15.0, 6.9 Hz, CH3), 1.36 (dd, 6H, J ) 18.0,
7.0 Hz, CH3), 1.64 (m, 2H, J ≈ 13.2, 4.8, 2.4 Hz, CH2), 1.93 (m, 2H,
J ≈ 12.9, 4.8 Hz, CH2), 1.75-2.53 (m, 14H, COD CH2’s), 2.92 (m,
2H, J ≈ 6.8 Hz, CH), 4.60 (m, 2H, COD vinyl), 5.29 (m, 2H, vinyl),
7.67-7.78 (m, 4H, phenyl) ppm; 31P NMR (121 MHz, CD2Cl2) δ 68.6
ppm. Assignments were made using a COSY-45, which was collected
using standard pulse sequences.35,36
Crytallographic Structure Determination of [Ir(COD)((R,R)-Me-
DuPHOS)]BF4 (5). A single crystal was grown from CH2Cl2/diethyl
ether. Crystal data are summarized in Table 7. A total of 12 902
reflections with 2θ < 58.4° were collected on a Siemens P4/CCD
diffractometer using graphite-monochromated Mo KR radiation. The
structure was solved by direct methods and was refined by full-matrix
least squares on F2 values with hydrogens riding.37,38 Non-hydrogen
atoms were refined with anisotropic thermal parameters. The final R(F)
and wR(F2) factors were 0.018 and 0.045, respectively. The crystal
structure is shown in Figure 4, along with a partial numbering scheme.
Selected bond lengths and bond angles are given in Table 7. Fractional
coordinates and additional crystallographic data can be found in the
Supporting Information.
In Situ Generation of [IrH2(P-P)(COD)]BF4. [IrH2(P-P)(COD)]-
BF4 complexes were formed by bubbling cold dihydrogen (-80 °C)
through a solution of [Ir(P-P)(COD)]BF4 at -80 °C. This reaction
was almost immediate and was accompanied by a color change. The
reactions were followed by 31P NMR and the hydride region of the 1H
NMR (Tables 1 and 5). Product ratios were determined by 31P and 1H
NMR (see Table 2). The full NMR spectra are available in the
Supporting Information.
[Ir((S,R)-BPPFA)(COD)]BF4 (12). A solution of (R,R)-BPPFA
(12.9 mg, 2.1 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was added
dropwise to a solution of [Ir(COD)2]BF4 (10.2 mg, 2.1 × 10-5 mol, 1
equiv in 0.5 g of CD2Cl2), yielding a bright red solution. Data for
[Ir((S,R)-BPPFA)(COD)]BF4 (12) at room temperature: 1H NMR (300
MHz, CD2Cl2) δ 1.22 (d, 3H, J ) 6.3 Hz, CHCH3), 1.76-1.95 (m,
4H, bound COD CH2’s), 2.13-2.43 (m, 18H, free and bound COD
CH2’s and N(CH3)2), 3.79 (br s, 1H, 1 Cp), 4.04-4.59 (m, 12H, bound
COD vinyls, Cp’s and CHCH3), 5.55 (br s, 4H, free COD vinyl), 7.31-
7.99 (m, 20H, phenyls) ppm; 31P NMR (CD2Cl2) δ 15.7 (d, JP-H ) 16
Hz), 21.3 (br s) ppm.
[Ir(COD)((S,R)-BPPFAc)]BF4 (13). A solution of (R,R)-BPPFAc
(13.1 mg, 2.1 × 10-5 mol, 1 equiv in 0.5 g of CD2Cl2) was added
dropwise to a solution of [Ir(COD)2]BF4 (10.2 mg, 2.1 × 10-5 mol, 1
equiv in 0.5 g of CD2Cl2), yielding a bright red solution. Data for
[Ir(COD)((S,R)-BPPFAc)]BF4 (13) at room temperature: 1H NMR (300
[IrH2(P-P)(COD)]BF4 Warming Experiments. Solutions of
[IrH2(P-P)(COD)]BF4 at -80 °C were warmed to -45 °C (acetonitrile/
N2 bath) for periods of 2-5 h. These solutions were cooled to -80
1
°C, and their 31P and H NMR were recorded. Results are given in
Table 2. Further experimental details have been published in the thesis
of B. F. M. Kimmich.39
(35) Freeman, R.; Morris, G. A. Bull. Magn. Reson. 1979, 1, 1.
(36) Bax, A.; Freeman, R. J. Magn. Reson. 1981, 44, 542.
(37) SAINT Version 4 Software Reference Manual; Siemens Analytical
X-ray Instruments: 6300 Enterprise Dr., Madison, WI 53719-1173; 1995.
(38) Sheldrick, G. M. SHELXTL VERSION 5 Software Manual; Siemens
Analytical X-ray Instruments: 6300 Enterprise Dr., Madison, WI 53719-
1173; 1995.
(39) Kimmich, B. F. M. Ph.D. Dissertation, University of Wisconsins
Madison, 1998.

10124 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Kimmich et al.
NMR Characterization of [IrH2((R,R)-Me-DuPHOS)(COD)]BF4
(10maj). A 50 mM solution of [Ir((R,R)-Me-DuPHOS)(COD)]BF4 (5)
in CD2Cl2 was cooled to -80 °C (dry ice/acetone). A steady stream
of cold (-80 °C) H2 was bubbled through the deep brown solution for
about 5 min. A pale yellow solution was formed, >98% by NMR as
above. 1H, 13C, and 31P NMR data for [IrH2(COD)((R,R)-Me-
DuPHOS)]BF4 are given in Tables 1 (31P NMR) and 5 (1H and 13C
NMR). All data were acquired at -80 °C. A combination of DQF
COSY and NOESY spectra was used to assign the 1H NMR. The DQF
COSY and NOESY spectra are provided in the Supporting Information.
The 13C NMR, DEPT 135, and HMQC NMRs were used to assign the
carbon atoms for 10maj. Due to overlap in the 13C dimension of the
HMQC, two sets of carbon peaks could not be completely assigned.
Four carbons are overlapped at 34.3 ppm, corresponding to three CH2
and one CH carbons on the phospholane rings. Two phenyl carbons
at 130.4 ppm also could not be differentiated.
1H NOESY NMR Experiments. Spectra were obtained at -90 °C in
acetone-d6. Spectra were collected at 70, 100, 130, 160, and 200 ms
mixing times using a phase-sensitive NOESY pulse sequence43 with
an initial homospoil pulse and States-Haberkorn-Ruben phase
cycling.42 A total of 1024 increments of eight scans was collected with
2048 real points in t2. The relaxation delay was set to 1.5 s. The
Felix NMR software package was used to transform and process data.
t1 was zero-filled to 2048 points, and t2 was zero-filled to 4096 points.
A linear prediction scheme was used to correct the first point and
generate points at the tails of the FID, extending data by one-third. A
Kaiser window function and exponential linebroadening were applied
to the FID to reduce apodization artifacts. Low-order polynomial
baseline corrections were applied to each dimension. The baselines
were corrected again using FACELIFT from the National Magnetic
Resonance Facility at Madison and deconvolution facility in the Felix
program. Diagonal peak volumes at zero mixing times were estimated
by fitting the observed diagonal peaks to an exponential function. For
off-diagonal peaks, the intensities of the two symmetries related peaks
were averaged prior to normalization. Random numbers were generated
by using random function on Microsoft Excel and added for the absent
peak data point. Due to the wide range of the spectra, a noisy peak is
redefined as having peak intensity less than or equal to 2 times the
estimated local volume noise level.
Generation of Static Model Conformers. A combination of two
subsets of trial structures totaling 205 conformations was used for the
structure analysis. The first subset was generated by annealed dynamics
simulation using CERIUS2 and the Universal force field. The
electrostatic term was turned off for all calculations using the Universal
force field. Harmonic angle constraints (1000 kcal/(mol Å) for the
force constant) between both hydrides and between hydride and
phosphorus atoms were applied to prevent the collapse of two ligands
on top of one another during dynamics simulation. Simulated annealing
at constant volume and constant energy was run for 100 cycles in 10-K
increment from 900 to 0 K and from 0 to 900 K with 50 steps at each
increment, yielding 100 trial conformations. Each conformation was
then mimized without constraints until an rms gradient less that 0.001
kcal/Å was reached. The models generated by simulated annealing
differ primarily in the orientation of the bridging phenylene ring relative
to the coordination plane defined by the two phosphorus atoms and
the Ir. The type of the conformation is defined by the torsion angle
C7-P1-Ir1-P2 (torsion 1) and the torsion angle C12-P2-Ir1-P1
(torsion 2). The conformation with the torsion 1 and 2 ranging from
-10° to 10° is hereafter called the “planar” conformation. When the
phenyl ring moves up to the same side as the axial hydride with torsion
1 less than -10° and torsion 2 greater than 10°, the model is hereafter
called the “up” conformation; when the phenyl ring moves down to
the opposite side with the torsion 1 greater than 10° and the torsion 2
less than 10°, the model is hereafter called the “down” conformation.
The first subset of conformers comprises 30 “up” structures, 10 “planar”
structures, and 60 “down” structures.
1
A 500-MHz H-1H DQF COSY was obtained at -80 °C in CD2-
Cl2. The spectrum was gathered using a phase-sensitive DQF COSY
pulse sequence40,41 with an initial homospoil pulse and States-
Haberkorn-Ruben phase cycling.42 A data set with t1 and t2 dimensions
of 512 and 2048 real points, respectively, was collected using relaxation
delays of 3.5 s for eight scans of each FID. The Felix NMR software
package from Molecular Simulations, Inc. was used to transform and
process the data. t1 was zero-filled to 1024 points, and t2 was zero-
1
filled to 4096 points. The complete H-1H DQF COSY spectrum is
available in the Supporting Information.
A 500-MHz 1H NOESY was obtained at -80 °C in CD2Cl2 with a
mixing time of 300 ms. The spectrum was gathered using a phase-
sensitive 2D-NOESY pulse sequence43 with an initial homospoil pulse
and States-Haberkorn-Ruben phase cycling.42 A data set with t1 and
t2 dimensions of 512 and 2048 real points, respectively, was collected
using relaxation delays of 3.5 s for eight scans of each FID. The Felix
NMR software package from Molecular Simulations, Inc. was used to
transform and process the data. t1 was zero-filled to 1024 points, and
t2 was zero-filled to 4096 points. The complete 1H-1H NOESY
spectrum is available in the Supporting Information.
1
A 500-MHz H-13C HMQC was obtained at -80 °C in CD2Cl2.
The spectrum was gathered using a HMQC pulse sequence.44,45 A data
set with t1 and t2 dimensions of 512 and 2048 real points, respectively,
was collected using relaxation delays of 1.5 s for 24 scans of each
FID. The VNMR software package from Varian Nuclear Magnetic
Resonance Instruments was used to transform and process the data. t1
was zero-filled to 1024 points, and t2 was zero-filled to 4096 points.
The complete 1H-13C HMQC spectrum is available in the Supporting
Information.
Determination of the Solution Structure of 10maj
. A 36 mM
solution of [Ir((R,R)-Me-DuPHOS)(COD)]BF4 was prepared in acetone-
d6. The solution was reacted with H2 at -80 °C and warmed to -45
°C, yielding a 5.8:1.0 mixture of dihydrides 10maj and 10min. All 2D
experiments were acquired at -90 °C.
1H DQF COSY NMR Experiments. The spectrum was measured
using a phase-sensitive DQF COSY pulse sequence40,41 with an initial
homospoil pulse and States-Haberkorn-Ruben phase cycling at -90
°C in acetone-d6.42 A total of 1024 increments of eight scans was
collected with 2048 real points in t2. The relaxation delay was set to
1.5 s. The Felix NMR software package was used to transform and
process data. t1 was zero-filled to 2048 points, and t2 was zero-filled
to 4096 points. A linear prediction scheme was used to correct the
first point and generate points at the tails of the FID, extending data
by one-third. Gaussian linebroadening and a screwed sine-squared bell
were applied to the FID.
The second subset of conformers was generated by applying
constraints on torsions 1 and 2 during simulated annealing to provide
a fuller range of conformers. Harmonic torsion constraints ranged from
-50° to 50° in 5° increments with a 1000 kcal/(mol Å) force constant.
Angle constraints for both hydrides were used to prevent ligand collapse.
Five annealing cycles at constant volume and constant energy were
then used for each torsion constraint. The annealing cycles involved
cooling from 900 to 600 K followed by heating to 900 K with 10-K
increment and 50 steps at each increment. Each annealed structure
was then minimized, and its energy was computed with all torsion
constraints turned off. The complete set of static conformers was
generated by combining 100 structures from the first subset and 105
structures from the second subset. The correlation between the torsion
angle (torsion 2) and energy for all conformers used in the trial set is
shown in Figure 6. The lowest energy “up” conformation is ap-
proximately 4 kcal/mol higher than the lowest energy “down”
conformation.
(40) Piantini, U.; Sørensen, O. W.; Ernst, R. R. J. Am. Chem. Soc. 1982,
104, 6800.
(41) Rance, M.; Sørensen, O. W.; Bodenhausen, G.; Wagner, G.; Ernst,
R. R.; Wu¨thrich, K. Biochem. Biophys. Res. Commun. 1983, 117, 479.
(42) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson. 1982,
48, 286.
(43) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95.
(44) Bax, A.; Griffey, R. H.; Hawkins, B. L. J. Magn. Reson. 1983, 55,
301.
(45) Kessler, H.; Gehrke, M.; Griesinger, C. Angew. Chem., Int. Ed. Engl.
1988, 27, 490.
Conformer Population Analysis of the Static Ensemble. Conformer
population analysis proceeds by computing the full NOE time course
for each conformer in the trial set and then finding the most
parsimonious set of conformers that best reproduce the observed data.

Addition of H2 to [Ir(bisphosphine)(COD)]+ Complexes
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10125
Known distances between pairs of protons (21A-21B, 25A-25B,
26A-26B for τ1 optimization; 1-5, 4-6, 16-17, 13-18 for τ2
optimization) were used for calibration of the relaxation parameters.
The resulting parameters (τ1 ) 0.560 ns; τ2 ) 167 ps) were used for
running 2DCPA. The static ensemble was exhaustively fitted with each
possible single, double, and triple conformer solution in the fast-
exchange mode. Significance testing was done using Hamilton’s F-test
on every possible combination of the 10 best fits of the singles, doubles,
and triples.
addition, the appearance of the dihydride resonances for 10maj/10min
was monitored via H NMR. These experiments were repeated on a
1
27 mM solution of [IrD2((R,R)-Me-DuPHOS)(COD)]BF4 using 1 atm
8.3% H2 (balance N2). As above, the rates for diastereomer intercon-
version were determined by plotting ln([10maj]t - [10maj]∞) vs time,
giving t1/2 ) 43 ( 10 and 62 ( 12 min, respectively. Rates for isotope
exchange were determined by plotting the ln([10maj
time. This reaction was found to follow first-order kinetics, and t1/2
47 ( 10 and 51 ( 10 min, respectively.
]
∞
- [10maj]t) vs
)
The populations obtained by 2DCPA depend on the relative
weightings of the NOE data. In particular, the extent to which
reproduction of absent NOE peaks is included in the data analysis
dramatically affects the absolute value of the R-factor (and may
influence relative conformer populations). Although the dihydride
complex 10maj has 34 inequivalent protons, 458 data points of 595 total
points are assigned as absent peaks. The experimental intensities of
absent cross-peaks in our analysis were represented as intensities that
range from zero to the estimated integration noise level. The contribu-
tion of noisy and absent peaks to the error function was given by (Icalc
- Iobs)2/10. For all other peaks, the contribution to the error function
was (Icalc - Iobs)2/(Iobs)2.
A distinctive feature of the 2DCPA is that multiple conformations
may be used to fit the NOE data. The inclusion of more than one
conformation in NOE analysis always leads to better agreement between
computed and observed data. Application of the F-test to the calculation
will determine whether it is significant to use more than one conforma-
tion or even two. The single-conformer fitting shows the lowest
R-factor (R ) 0.856) with the “planar” conformation, while the two-
conformer fitting leads to a slight decrease of the R-factor (R ) 0.830)
with fractional populations of 0.79 and 0.21 of the “planar” and the
“up” conformations, respectively. The best fitting of three conformers
yields an R-factor of 0.828 and corresponds to population of the “planar”
conformer and two similar “up” conformers with fractional populations
of 0.77, 0.19, and 0.04, respectively. Application of the F-test at the
95% confidence level excluding the absent peak data points indicates
that it is statistically significant to use two rather than one conformation
in fitting the data but that three conformations do not lead to a
significantly better fit relative to two conformations.
A 27 mM solution of [IrD2((R,R)-Me-DuPHOS)(COD)]BF4 was
prepared at -80 °C as before and was purged with cold N2 for 10
min. The appearance of hydride signals was followed by 1H NMR at
-45 °C. The rate for isotope exchange was determined by plotting
the ln([10maj ∞ - [10maj]t) vs time. This reaction was found to follow
]
first-order kinetics, and the t1/2 was 52 ( 10 min. The same experiment
was repeated with a 35 mM solution of [IrD2((R,R)-Me-DuPHOS)-
(COD)]BF4 in acetone. A similar half-life was found (55 ( 11 min).
Deuterium Incorporation into C-H Bonds of 10maj. A 52 mM
solution of [IrD2((R,R)-Me-DuPHOS)(COD)]BF4 was prepared in CH2-
Cl2. This solution was purged with N2 at -80 °C. The sample tube
was placed into a NMR probe cooled to -45 °C, and new peaks were
observed growing in the 2H NMR at δ 2.5 and 2.8 ppm. These broad
peaks were overlapped, and the upfield peak appeared as a shoulder
on the peak at δ 2.5 ppm. The peaks appeared to be growing in at
similar rates. The rate for isotope exchange was determined by plotting
the ln([C-D]∞ - [C-D]t) vs time. The integrations of the C-D and
M-D sites approached 50% occupancy at equilibrium. This reaction
was found to follow first-order kinetics, and the t1/2 was 56 ( 12 min.
Reactions performed in acetone-d6 gave similar results.
1
Due to overlap in the H NMR, 13C NMR was used to assign the
sites of deuterium incorporation. A 50 mM solution of [IrD2((R,R)-
Me-DuPHOS)(COD)]BF4 was prepared in CD2Cl2 as above. 13C NMR
and DEPT were taken at -80 °C. The solution was warmed for 3 h
at -45 °C in an acetonitrile/N2 bath. 13C NMR and 135 DEPT were
taken at -80 °C. The 13C NMR differed from the initial spectrum in
two main points: (1) the minor diastereomer 10min had grown in and
(2) two peaks at δ 35.76 (d, JP-H ) 41 Hz) and 37.89 (s) ppm,
corresponding to C1 and C25A, had significantly decreased in size while
broadening due to additional coupling (Figure 2.11). This clearly
indicated that D was being incorporated at two sites in the molecule.
These were found to correspond to positions H1 and H25A. Additional
warming did not substantially change the 13C NMR of the sample.
Assignment of the proton sites involved in the exchange was determined
using HMQC data in combination with previous assignments.
Kinetics Measurements of Diastereomer Interconversion and
Isotope Exchange. NMR kinetics experiments were conducted on a
500-MHz Bruker NMR at -45 or -80 °C. We estimate that the error
in NMR-derived rate constants for all experiments is approximately
20%. The data are summarized in Table 6. The diastereomer
interconversion is an approach to equilibrium; thus, the measured rate
constants are the sum of the forward and reverse reactions.
Acknowledgment. Support for this work was provided by
the Department of Energy. We thank Dr. Randy Hayashi for
solving the crystallograpic structure of [Ir(R,R Me-DuPhos)-
(COD)](BF4) and Dr. Monty Wright for assistance with the
2DCPA analysis of the solution NOE data. We gratefully
acknowledge helpful discussions with Prof. Chuck Casey and
Dr. Charlie Fry.
A 27 mM solution of [IrH2(COD)((R,R)-Me-DuPHOS)]BF4 was
formed at -80 °C, and the solution was degassed with cold N2 for 10
min. The NMR tube was placed in a NMR probe cooled to -45 °C,
and the disappearance of the 10maj dihydrides was followed by 1H NMR.
The above kinetics run was repeated for a 55 mM solution of
[IrH2(COD)((R,R)-Me-DuPHOS)]BF4 in CD2Cl2 and for a 35 mM
solution of 10 in acetone-d6. All three kinetics experiments followed
first-order kinetics for a reaction approaching equilibrium. Plots of
the ln([10maj]t - [10maj]∞) vs time were found to be linear. The half-
lives of the interconversions were 47 ( 10, 47 ( 10, and 64 ( 12 min
(these results are indistinguishable within the error of the experiments).
A 55 mM solution of [IrD2((R,R)-Me-DuPHOS)(COD)]BF4 was
prepared by bubbling cold D2 through a solution of [Ir((R,R)-Me-
DuPHOS)(COD)]BF4 to yield an approximately 50:1 mixture of d2-
10maj to d2-10min at -80 °C. The solution was warmed at -45 °C
(acetonitrile/N2 bath) for 5-min intervals under a stream of cold H2.
The 1H and 31P NMRs of the sample were taken at -80 °C. The
disappearance of the 10maj/d2-10maj was followed by 31P NMR. In
Supporting Information Available: X-ray crystallographic
1
data for [Ir((R,R)-Me-DuPhos)(COD)](BF4), H NMR spectra
for [Ir(bisphosphine)(COD)](BF4) and [IrH2(bisphosphine)-
(COD)](BF4) complexes, and NOESY data and 2DCPA analysis
for [IrH2((R,R)-Me-DuPhos)(COD)](BF4) (126 pages, print/
PDF). See any current masthead page for ordering information
and Web access instructions.
JA981536B