Palladium-catalyzed asymmetric phosphination. Scope, mechanism, and origin of enantioselectivity

Article abstract of DOI:10.1021/ja070225a

Asymmetric cross-coupling of aryl iodides (ArI) with secondary arylphosphines (PHMe(Ar′), Ar′ = (2,4,6)-R₃C ₆H₂; R = i-Pr (Is), Me (Mes), Ph (Phes)) in the presence of the base NaOSiMe₃ and a chiral Pd catalyst precursor, such as Pd((R,R)-Me-Duphos)(trans-stilbene), gave the tertiary phosphines PMe-(Ar′)(Ar) in enantioenriched form. Sterically demanding secondary phosphine substituents (Ar′) and aryl iodides with electron-donating para substituents resulted in the highest enantiomeric excess, up to 88%. Phosphination of ortho-substituted aryl iodides required a Pd(Et-FerroTANE) catalyst but gave low enantioselectivity. Observations during catalysis and stoichiometric studies of the individual steps suggested a mechanism for the cross-coupling of PhI and PHMe(Is) (1) initiated by oxidative addition to Pd(0) yielding Pd((R,R)-Me-Duphos)(Ph)(I) (3). Reversible displacement of iodide by PHMe(Is) gave the cation [Pd((R,R)-Me-Duphos)(Ph)(PHMe(Is))][I] (4), which was isolated as the triflate salt and crystallographically characterized. Deprotonation of 4-OTf with NaOSiMe₃ gave the phosphido complex Pd((R,R)-Me-Duphos)(Ph)(PMeIs) (5); an equilibrium between its diastereomers was observed by low-temperature NMR spectroscopy. Reductive elimination of 5 yielded different products depending on the conditions. In the absence of a trap, the unstable three-coordinate phosphine complex Pd((R,R)-Me-Duphos) (PMeIs(Ph)) (6) was formed. Decomposition of 5 in the presence of PhI gave PMeIs(Ph) (2) and regenerated 3, while trapping with phosphine 1 during catalysis gave Pd((R,R)-Me-Duphos)(PHMe(Is))₂ (7), which reacted with PhI to give 3. Deprotonation of 1:1 or 1.4:1 mixtures of cations 4-OTf gave the same 6:1 ratio of enantiomers of PMeIs-(Ph) (2), suggesting that the rate of P inversion in 5 was greater than or equal to the rate of reductive elimination. Kinetic studies of the first-order reductive elimination of 5 were consistent with a Curtin-Hammett-Winstein-Holness (CHWH) scheme, in which pyramidal inversion at the phosphido ligand was much faster than P-C bond formation. The absolute configuration of the phosphine (S_P)-PMeIs(p-MeOC ₆H₄) was determined crystallographically; NMR studies and comparison to the stable complex 5-Pt were consistent with an R ^P-phosphido ligand in the major diastereomer of the intermediate Pd((R,R)-Me-Duphos)-(Ph)(PMeIs) (5). Therefore, the favored enantiomer of phosphine 2 appeared to be formed from the major diastereomer of phosphido intermediate 5, although the minor intermediate diastereomer underwent P-C bond formation about three times more rapidly. The effects of the diphosphine ligand, the phosphido substituents, and the aryl group on the ratio of diastereomers of the phosphido intermediates Pd(diphos*)-(Ar)(PMeAr′), their rates of reductive elimination, and the formation of three-coordinate complexes were probed by low-temperature ³¹P NMR spectroscopy; the results were also consistent with the CHWH scheme.

Full text of DOI:10.1021/ja070225a

Published on Web 05/03/2007
Palladium-Catalyzed Asymmetric Phosphination. Scope,
Mechanism, and Origin of Enantioselectivity
Natalia F. Blank,† Jillian R. Moncarz,† Tim J. Brunker,† Corina Scriban,†
Brian J. Anderson,† Omar Amir,† David S. Glueck,*,† Lev N. Zakharov,‡
James A. Golen,‡ Christopher D. Incarvito,§ and Arnold L. Rheingold‡,§
Contribution from the 6128 Burke Laboratory, Department of Chemistry, Dartmouth College,
HanoVer, New Hampshire, 03755, Department of Chemistry, UniVersity of California, San
Diego, 9500 Gilman DriVe, La Jolla, California, 92093, and Department of Chemistry,
UniVersity of Delaware, Newark, Delaware, 19716
Received January 17, 2007; E-mail: Glueck@Dartmouth.Edu
Abstract: Asymmetric cross-coupling of aryl iodides (ArI) with secondary arylphosphines (PHMe(Ar′), Ar′
) (2,4,6)-R3C6H2; R ) i-Pr (Is), Me (Mes), Ph (Phes)) in the presence of the base NaOSiMe3 and a chiral
Pd catalyst precursor, such as Pd((R,R)-Me-Duphos)(trans-stilbene), gave the tertiary phosphines PMe-
(Ar′)(Ar) in enantioenriched form. Sterically demanding secondary phosphine substituents (Ar′) and aryl
iodides with electron-donating para substituents resulted in the highest enantiomeric excess, up to 88%.
Phosphination of ortho-substituted aryl iodides required a Pd(Et-FerroTANE) catalyst but gave low
enantioselectivity. Observations during catalysis and stoichiometric studies of the individual steps suggested
a mechanism for the cross-coupling of PhI and PHMe(Is) (1) initiated by oxidative addition to Pd(0) yielding
Pd((R,R)-Me-Duphos)(Ph)(I) (3). Reversible displacement of iodide by PHMe(Is) gave the cation [Pd((R,R)-
Me-Duphos)(Ph)(PHMe(Is))][I] (4), which was isolated as the triflate salt and crystallographically character-
ized. Deprotonation of 4-OTf with NaOSiMe3 gave the phosphido complex Pd((R,R)-Me-Duphos)(Ph)(PMeIs)
(5); an equilibrium between its diastereomers was observed by low-temperature NMR spectroscopy.
Reductive elimination of 5 yielded different products depending on the conditions. In the absence of a trap,
the unstable three-coordinate phosphine complex Pd((R,R)-Me-Duphos)(PMeIs(Ph)) (6) was formed.
Decomposition of 5 in the presence of PhI gave PMeIs(Ph) (2) and regenerated 3, while trapping with
phosphine 1 during catalysis gave Pd((R,R)-Me-Duphos)(PHMe(Is))2 (7), which reacted with PhI to give 3.
Deprotonation of 1:1 or 1.4:1 mixtures of cations 4-OTf gave the same 6:1 ratio of enantiomers of PMeIs-
(Ph) (2), suggesting that the rate of P inversion in 5 was greater than or equal to the rate of reductive
elimination. Kinetic studies of the first-order reductive elimination of 5 were consistent with a Curtin-
Hammett-Winstein-Holness (CHWH) scheme, in which pyramidal inversion at the phosphido ligand was
much faster than P-C bond formation. The absolute configuration of the phosphine (SP)-PMeIs(p-MeOC6H4)
was determined crystallographically; NMR studies and comparison to the stable complex 5-Pt were
consistent with an RP-phosphido ligand in the major diastereomer of the intermediate Pd((R,R)-Me-Duphos)-
(Ph)(PMeIs) (5). Therefore, the favored enantiomer of phosphine 2 appeared to be formed from the major
diastereomer of phosphido intermediate 5, although the minor intermediate diastereomer underwent P-C
bond formation about three times more rapidly. The effects of the diphosphine ligand, the phosphido
substituents, and the aryl group on the ratio of diastereomers of the phosphido intermediates Pd(diphos*)-
(Ar)(PMeAr′), their rates of reductive elimination, and the formation of three-coordinate complexes were
probed by low-temperature 31P NMR spectroscopy; the results were also consistent with the CHWH scheme.
Introduction
asymmetric catalysis to the preparation of other high-value chiral
compounds, metal-catalyzed asymmetric syntheses of these
Metal-catalyzed asymmetric synthesis depends on chiral
ligands, especially phosphines.1 Chiral phosphines are usually
prepared either by resolution or by using a stoichiometric amount
of a chiral auxiliary.2 Despite the successful application of
ligands are rare.3 However, we and others have recently
described new approaches to P-stereogenic chiral phosphines
via metal-catalyzed hydrophosphination of alkenes,4 Pt- or Ru-
catalyzed alkylation of secondary phosphines,5 or Pd-catalyzed
phosphination of aryl halides.6 Better understanding of the scope
† Dartmouth College.
‡ University of California, San Diego.
§ University of Delaware.
(2) (a) Pietrusiewicz, K. M.; Zablocka, M. Chem. ReV. 1994, 94, 1375-1411.
(b) Kagan, H. B.; Sasaki, M. In The Chemistry of Organophosphorus
Compounds; Hartley, F. R., Ed.; John Wiley and Sons: Chichester, U.K.,
1990; Vol. 1, pp 51-102. (c) Tang, W.; Zhang, X. Chem. ReV. 2003, 103,
3029-3070.
(1) (a) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley-
Interscience: New York, 1994. (b) Blaser, H.-U., Schmidt, E., Eds.
Asymmetric Catalysis on Industrial Scale. Challenges, Approaches, and
Solutions; Wiley-VCH: Weinheim, 2004.
9
10.1021/ja070225a CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 6847-6858
6847

A R T I C L E S
Scheme 1
Blank et al.
Table 1. Pd((R,R)-Me-Duphos)-Catalyzed Asymmetric
Phosphination of PhI with PHMe(Is): Effect of Solvents,
Temperature, and Base on Yield and Enantioselectivitya
catalyst
equiv of
PhI
loading
(mol%)
equiv of
base
entry
solvent
T (
°
C)
yield (%)
eeb (%)
1
2
3
4
5
6
7
8
2
5
3
2.5
5
7
7
5
5
5
1
1
1
1
1
1
1
1
1
1
2
3
10
1
toluene
toluene
toluene
pentane
THF
MeCN
DMSO
toluene
toluene
toluene
toluene
toluene
toluene
toluene
22
22
22
22
22
22
22
4
50
22
22
22
22
22
71c
93d
88d
90d
69c
60c
78c
82c
60d
91d
90d
89d
96d
83d
73
72
73
78
66
58
55
82
42
70
73
75
74
75
1.05
1.07
1.1
2
2
1.1
1
9
2
10e
11f
12f
13f
14f,g
1.05
1.1
1.1
1.1
1.1
5
2.5
2.5
2.5
2.5
and limitations of these metal-catalyzed processes, and knowl-
edge of their mechanisms and the origin of enantioselectivity,
should help to make asymmetric catalysis a more efficient and
practical route to valuable chiral phosphines. With this goal in
mind, we report here details of our studies of the Pd-catalyzed
asymmetric phosphination of aryl iodides.6a
a The catalyst precursor was Pd((R,R)-Me-Duphos)(Ph)(I), and the base
was NaOSiMe3 (1.0 M in THF). bThe phosphine was complexed with
Pd((S)-Me2NCH(Me)C6H4)(µ-Cl))2 in C6D6; integration of CHMe signals
in the 1H NMR spectrum gave the de, which corresponds to the phosphine
ee (see the Supporting Information for details). cNMR yield, from integration
of the 1H NMR spectrum after addition of Pd((S)-Me2NCH(Me)C6H4)
(µ-Cl))2 to the phosphine product isolated after column chromatography
on silica gel. dIsolated yield after column chromatography. eScale: 500 mg
Results and Discussion
Selection of Catalyst and Substrates for a Test Reaction.
In any catalytic asymmetric phosphine synthesis, the chiral
ancillary ligand must resist displacement from the metal by the
excess phosphine substrate and products. Therefore, we used
the rigid bidentate chiral alkylphosphine (R,R)-Me-Duphos to
ensure tight binding to the metal.7 The secondary phosphine
PHMe(Is) (1, Is ) 2,4,6-(i-Pr)3C6H2), with a large Is and a
small Me substituent, was selected in hopes of maximizing
steric differentiation at P in the diastereomeric intermediates
and transition states.8 The base NaOSiMe3 was chosen to
minimize the concentration of the free phosphido anion
[PMeIs]-, which might reduce the Pd catalyst or undergo other
side reactions. These design elements were successful; the
reaction in Scheme 1 occurred quickly at room temperature
(about 1 h with 5 mol % catalyst loading) in high yield and
∼70% ee.6a
f
of PHMe(Is). Catalyst precursor: Pd((R,R)-Me-Duphos)(trans-stilbene).
gBase ) NaN(SiMe3)2.
Effect of Reaction Conditions. Table 1 shows the effect of
reaction conditions on the rate and selectivity of this test cross-
coupling. An excess of PhI was not required, which simplified
the workup (entries 1 and 2). The ee decreased slightly with
increased solvent polarity (entries 3-7), so the preferred solvent
was toluene or pentane. The catalyst loading could be decreased
to 2.5 mol % without any loss of enantioselectivity (entry 3).
Enantioselectivity increased at low temperature and decreased
on heating the reaction mixture (compare entries 1, 8, and 9).
Entries 10-14 showed that only 1 equiv of base was required;
using excess NaOSiMe3 or changing the base to NaN(SiMe3)2,
however, had little effect on yield and ee. In general, the reaction
appeared to be quantitative, and isolated yields were limited
by losses during chromatographic workup.
Screening of reaction conditions was done on a small scale
(20-25 mg of secondary phosphine 1). On a larger scale,
isolation of air-sensitive 2 by chromatography under nitrogen
was inconvenient, but a modified workup (entry 10) avoided
this problem. Coupling of 500 mg of PHMe(Is) (2 mmol) with
PhI using 5 mol % of Pd((R,R)-Me-Duphos)(Ph)(I) (3) was
complete in 2 h. After solvent was removed, extraction with
petroleum ether and filtration through a short pad of silica
followed by washing the silica with petroleum ether/THF (9:1)
eluent reproducibly gave the product phosphine in 90% yield
and 70% ee.
(3) (a) Burk, M. J.; Feaster, J. E.; Nugent, W. A.; Harlow, R. L. J. Am. Chem.
Soc. 1993, 115, 10125-10138. (b) Hoge, G. J. Am. Chem. Soc. 2003, 125,
10219-10227. (c) Shimizu, H.; Saito, T.; Kumobayashi, H. AdV. Synth.
Catal. 2003, 345, 185-189. (d) Marinetti, A.; Genet, J.-P. C. R. Chimie
2003, 6, 507-514.
(4) (a) Kovacik, I.; Wicht, D. K.; Grewal, N. S.; Glueck, D. S.; Incarvito, C.
D.; Guzei, I. A.; Rheingold, A. L. Organometallics 2000, 19, 950-953.
(b) Douglass, M. R.; Ogasawara, M.; Hong, S.; Metz, M. V.; Marks, T. J.
Organometallics 2002, 21, 283-292. (c) Sadow, A. D.; Haller, I.; Fadini,
L.; Togni, A. J. Am. Chem. Soc. 2004, 126, 14704-14705. (d) Sadow, A.
D.; Togni, A. J. Am. Chem. Soc. 2005, 127, 17012-17024. (e) Scriban,
C.; Kovacik, I.; Glueck, D. S. Organometallics 2005, 24, 4871-4874. (f)
Kovacik, I.; Scriban, C.; Glueck, D. S. Organometallics 2006, 25, 536-
539. (g) Scriban, C.; Glueck, D. S.; Zakharov, L. N.; Kassel, W. S.;
DiPasquale, A. G.; Golen, J. A.; Rheingold, A. L. Organometallics 2006,
25, 5757-5767. For related reactions involving metal-catalyzed enanti-
oselective addition of P-H bonds to unsaturated substrates, see: (h)
Shulyupin, M. O.; Franciu, G.; Beletskaya, I. P.; Leitner, W. AdV. Synth.
Catal. 2005, 347, 667-672. (i) Xu, Q.; Han, L.-B. Org. Lett. 2006, 8,
2099-2101. (j) Join, B.; Mimeau, D.; Delacroix, O.; Gaumont, A.-C. Chem.
Commun. 2006, 3249-3251.
Catalyst Screening. Several chiral Pd(diphos*)(Ph)(I) and
Pd(diphos*)(trans-stilbene) complexes were screened as catalyst
precursors in the reaction of Scheme 1 (Scheme 2, Table 2).9
The yields and ee of the product phosphine depended strongly
on the nature of the chiral diphos* ligand. We first investigated
bis(phospholanes) structurally analogous to Me-Duphos. Either
the original catalyst Pd((R,R)-Me-Duphos)(Ph)(I) (3) or the Pd-
(5) (a) Scriban, C.; Glueck, D. S. J. Am. Chem. Soc. 2006, 128, 2788-2789.
(b) Chan, V. S.; Stewart, I. C.; Bergman, R. G.; Toste, F. D. J. Am. Chem.
Soc. 2006, 128, 2786-2787.
(6) (a) Moncarz, J. R.; Laritcheva, N. F.; Glueck, D. S. J. Am. Chem. Soc.
2002, 124, 13356-13357. (b) Moncarz, J. R.; Brunker, T. J.; Jewett, J. C.;
Orchowski, M.; Glueck, D. S.; Sommer, R. D.; Lam, K.-C.; Incarvito, C.
D.; Concolino, T. E.; Ceccarelli, C.; Zakharov, L. N.; Rheingold, A. L.
Organometallics 2003, 22, 3205-3221. (c) Korff, C.; Helmchen, G. Chem.
Commun. 2004, 530-531. (d) Pican, S.; Gaumont, A.-C. Chem. Commun.
2005, 2393-2395.
(7) Drago, D.; Pregosin, P. S. Organometallics 2002, 21, 1208-1215.
(8) Brauer, D. J.; Bitterer, F.; Dorrenbach, F.; Hessler, G.; Stelzer, O.; Kruger,
C.; Lutz, F. Z. Naturforsch. B 1996, 51, 1183-1196.
(9) Brunker, T. J.; Blank, N. F.; Moncarz, J. R.; Scriban, C.; Anderson, B. J.;
Glueck, D. S.; Zakharov, L. N.; Golen, J. A.; Sommer, R. D.; Incarvito, C.
D.; Rheingold, A. L. Organometallics 2005, 24, 2730-2746.
9
6848 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007

Palladium-Catalyzed Asymmetric Phosphination
A R T I C L E S
Scheme 2. Chiral Diphosphine Ligands Screened in Pd-Catalyzed
Synthesis of 2 from 1
Table 3. Pd((R,R)-Me-Duphos)-Catalyzed Coupling of Secondary
Methylarylphosphines with Aryl Halidesa
entry
phosphine
ArX
time (h) yieldb (%)
eec (%)
σpd
1e
2f
3
4
5
6
7
8
9
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PHMe(Is)
PhBr
PhOTf
PhI
1
1
1
1
6
1
22
2
2
1
2
6
456
6
5
53
70
97
89
96
88
98
96
68
73
91
75
93
98
98
59
90
75
91
81
84
68
72
38
50
76
0
p-PhOC6H4I
p-MeOC6H4I
p-NH2C6H4I
p-MeC6H4I
p-CF3C6H4I
p-IC6H4I
p-NO2C6H4I
p-IC6H4I
8-QuinolylI
3,5-Me2C6H3I
o-C5H4NI
1-NaphI
88g
82
-0.32
-0.27
-0.66
-0.17
0.54
79g
78
45
40
0.18
0.78
10h
11i
12
13
14
15
16
17
18
19
20
21
22n
12
33j
27k
15
12
ndl
56
PHMe(Mes) PhI
4
0
PHMe(Mes) p-MeOC6H4I
PHMe(Mes) p-CF3C6H4I
PHMe(Phes) PhI
PHMe(Phes) p-MeOC6H4I
PHMe(Phes) p-PhOC6H4I
PHMe(Men) PhI
48
48
24
1
24
48
48
48g
7g
-0.27
0.54
0
-59m
-65g,m -0.27
-82m -0.32
Table 2. Pd(diphos*)-Catalyzed Asymmetric Synthesis of
PPhMe(Is) (2) from PHMe(Is) (1)a
73
91
23n,o PHMe(Men) PhI
entry
Diphos*
precursor
solvent
t (h)
yield (%)
eeb (%)
1c
2
3
4
5
(R,R)-Me-Duphos
(R,R)-Me-Duphos
(R,R)-Et-Duphos
(R,R)-i-Pr-Duphos
(R,R)-i-Pr-Duphos
(R,R)-Me-BPE
(S,S)-Me-FerroLANE
(S,S)-Et-FerroTANE
(S,S)-Et-FerroTANE
(S,S)-Et-FerroTANE
B
A
B
B
A
B
B
B
B
B
B
A
A
A
toluene
toluene
toluene
toluene 16
toluene 16
THF
toluene
THF
THF
toluene 24
THF 72
toluene 48
toluene 48k
toluene 24k
1
1
1
71d
97e
63d
83e
98e
100e
89e
ndh
ndh
96e
ndh
94
73
76
72
a The catalyst precursor was 5 mol % of Pd((R,R)-Me-Duphos)(trans-
stilbene). Reactions used 0.1 mmol of PHMe(Ar′), 1.05-1.1 equiv of ArX,
and NaOSiMe3 (1.0 M in THF, 1 equiv) at room temperature in toluene,
unless otherwise stated. Mes ) 2,4,6-Me3C6H2, Phes ) 2,4,6-Ph3C6H2. See
73
b
Experimental Section for details. Isolated yield after column chromatog-
71
raphy. cThe phosphine was complexed with Pd(((S)-Me2NCH(Me)C6H4)(µ-
Cl))2 in C6D6; integration of CHMe signals in the 1H NMR spectrum gave
the de which corresponds to the phosphine ee. dHammett substituent
constant.12 eCatalyst precursor: Pd((R,R)-Me-Duphos)(Ph)(I), 2 equiv of
PhBr, 50 °C. fCatalyst precursor: Pd((R,R)-Me-Duphos)(Ph)(I), 2 equiv of
PhOTf. gThe phosphine was complexed with Pd((S)-Me2NCH(Me)C6H4)(µ-
Cl))2 in C6D6; integration of P signals in the 31P NMR spectrum gave the
6
7
20
4
24
48
-18f
28
8g
9i
10
-45f
-50f
-37f
-21f
3
11j (R,S)-PPF-t-Bu
12
13
14
(R,S)-PPF-t-Bu
(R,S)-CyPF-t-Bu
(R,S)-BoPhoz
h
de which corresponds to the phosphine ee. Catalyst precursor: Pd((R,R)-
72
72
rac
rac
i
Me-Duphos)(p-NO2C6H4)(I) (see below). 2 equiv of NaOSiMe3, 2 equiv
j
of PHMe(Is). de ) diastereomeric excess. kee was determined for the
[(N-C*)Pd(P-N)][OTf] adduct formed in reaction of (N-C*)Pd(Cl)(P(8-
Quinolyl)Me(Is)) with AgOTf. l1H NMR signals of the (N-C*)Pd(P(1-
Naph)Me(Is) complex were very broad, so ee was not determined (nd).
mThe negative sign indicates that the 31P and 1H NMR chemical shift trends
for Pd adducts of these phosphine products were the opposite of those
observed for PArMe(Is) and PArMe(Mes) (Supporting Information). nFrom
ref 14. Men ) (-)-menthyl; here, the de is reported. oCatalyst precursor )
Pd((R,R)-i-Pr-Duphos)(trans-stilbene).
a The standard procedure used 5 mol % of the catalyst precursors
Pd(diphos*)(trans-stilbene) (A) or Pd(diphos*)(Ph)(I) (B), PHMe(Is), PhI
(1.05-1.1 equiv), and NaOSiMe3 (1.0 M in THF, 1 equiv) at room
temperature, except where noted. bThe phosphine was complexed with
Pd(((S)-Me2NCH(Me)C6H4)(µ-Cl))2 in C6D6; integration of CHMe signals
in the 1H NMR spectrum gave the de which corresponds to the phosphine
c
d
1
ee. 2 equiv of PhI (Table 1, entry 1). From integration of the H NMR
spectrum after addition of Pd(((S)-Me2NCH(Me)C6H4)(µ-Cl))2 to the
phosphine product isolated after column chromatography on silica gel.
eIsolated yield after column chromatography. fThis catalyst selectively gave
the opposite enantiomer of the phosphine. g10 mol % catalyst precursor, 2
Several Ni(Me-Duphos) catalyst precursors (see the Support-
ing Information) were inferior to the Pd analogue; reaction did
not go to completion, and several byproducts were formed. With
Pd, the Me-BPE ligand, less rigid than Duphos, gave reduced
ee (entry 6), while the ferrocene-based bis(phospholane) Me-
FerroLANE (entry 7) was also inferior. Catalysis with the wide-
bite-angle DuXantphos ligand was very slow, and byproducts
formed. Despite promising results for asymmetric phosphination
of secondary diarylphosphines with (S,S)-Et-FerroTANE re-
ported recently by Korff and Helmchen,6c this ligand was less
enantioselective in the test reaction (entries 8-10). It did enable
use of PhOTf with a Pd(II) precursor (entry 9; see also entry
11 for related results with PPF-t-Bu and Table 3, entry 2 for
use of PhOTf with a Pd(Me-Duphos) catalyst). Stilbene
complexes of the Josiphos derivatives PPF-t-Bu and CyPF-t-
Bu or the related ferrocenylphosphine BoPhoz (entries 12-14)
i
equiv of PhI. hnd ) not determined. 10 mol % of the catalyst precursor
was added to AgOTf and filtered into a solution of 1, PhOTf (2 equiv),
and NEt(i-Pr)2. jThe catalyst precursor was the cation [Pd(PPF-t-
Bu)(Ph)(NCMe)][OTf] (prepared in situ from the iodo complex and AgOTf
in MeCN; see the Supporting Information), 2 equiv of PhOTf, base ) NEt(i-
k
Pr)2. At 50 °C.
(0) precursor Pd((R,R)-Me-Duphos)(trans-stilbene) could be
used, without significant differences (entries 1 and 2 and entries
4 and 5 for the i-Pr-Duphos analogue). The change from Me-
to Et-Duphos had little effect (entries 1-3), and results with
(R,R)-i-Pr-Duphos were similar (entries 4 and 5). This was
surprising; the Cahn-Ingold-Prelog priority rules mean that
(R,R)-Me- and -i-Pr-Duphos have opposite absolute configura-
tions and might be expected to preferentially yield opposite
enantiomers of the product.3a NMR and structural data for the
catalyst precursors9 indicated that these results were not a
consequence of mislabeling of the ligand stereochemistry;
possible mechanistic explanations for the reversal of enanti-
oselectivity are discussed in more detail below.10
(10) For other examples of reversal of enantioselectivity in asymmetric catalysis,
see: (a) Evans, D. A.; Johnson, J. S.; Burgey, C. S.; Campos, K. R.
Tetrahedron Lett. 1999, 40, 2879-2882. (b) Sibi, M. P.; Liu, M. Curr.
Org. Chem. 2001, 5, 719-755. (c) Zanoni, G.; Castronovo, F.; Franzini,
M.; Vidari, G.; Giannini, E. Chem. Soc. ReV. 2003, 32, 115-129.
9
J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007 6849

A R T I C L E S
Blank et al.
Scheme 3. Substrate Scope in Pd-Catalyzed Asymmetric
Phosphination (R ) i-Pr, Me, Ph; for Ar Groups See Tables 3
and 4)
Table 4. Pd((S,S)-Et-FerroTANE)-catalyzed Coupling of PHMe(Is)
with o-Substituted Aryl Iodides and Organometallic
Iodocyclopentadienyl Complexesa
entry
iodo substrate
time (h)
yieldb (%)
eec (%)
1
2
o-MeOC6H4I
o-CNC6H4I
o-MeC6H4I
FcI
1
42
9
6
6
51
83
88
100
82
racemic
3
ndd
8
3
4e
5f
Figure 1. Correlation between ee of the phosphines P(p-XC6H4)Me(Is)
and the substituent constant σp (Hammett plot).12
CymI
5
a The catalyst precursor was Pd((S,S)-Et-FerroTANE)(trans-stilbene) (5
mol %). Reactions were done with 1 equiv of ArI and 1 equiv of NaOSiMe3
(1.0 M in THF) in toluene at room temperature. See Experimental Section
for details. bIsolated yield after column chromatography on silica. cThe
phosphine was complexed with (Pd((S)-Me2NCH(Me)C6H4)(µ-Cl))2 in
C6D6; integration of the 31P NMR spectrum gave the de which corresponds
to the phosphine ee. dnd ) not determined; the 31P NMR signals overlapped,
and suitable 1H NMR signals could not be identified (see Supporting
Information). eFc ) ferrocenyl, C5H4FeC5H5, 70 °C. fCym ) cymantrenyl,
C5H4Mn(CO)3, 50 °C.
entry 13, ee depended strongly on the substrate. This was
particularly evident from results with p-substituted aryl iodides,
where steric effects are presumably minimal. More electron-
donating p-substituents resulted in higher ee (see entries 3-10
with PHMe(Is)). A Hammett plot of product ee vs substituent
constants σp, which are listed in Table 3,12 gave a reasonable
correlation (Figure 1). The same trend was observed for the
other secondary phosphine substrates (Table 3, entries 16-21).
However, an attempt to exploit these observations in the
coupling of PHMe(Is) with p-Me2NC6H4I (σp ) -0.83)12 led
to immediate catalyst decomposition and formation of Pd((R,R)-
Me-Duphos)2.6b
While varying the aryl iodide with a given secondary
phosphine substrate revealed the electronic effects of p-
substitution, measuring the ee for one aryl iodide with different
secondary phosphines showed the effect of phosphine structure
on enantioselectivity (see Table 3, entries 3, 16, and 19 (PhI);
5, 17, and 20 (p-MeOC6H4I); 8 and 18 (p-CF3C6H4I); and 4
and 21 (p-PhOC6H4I)). In particular, enantiomeric excesses were
similar for PHMe(Is) and PHMe(Phes) but lower for PHMe-
(Mes), perhaps because Is and Phes are comparable in size and
both larger than Mes. Catalyst turnover did not occur with the
even bulkier PHMe(Mes*) (Mes* ) 2,4,6-(t-Bu)3C6H2).8,13
We have reported elsewhere the diastereoselective coupling
of the secondary dialkylphosphine PHMe(Men) (Men ) (-)-
menthyl) with PhI (entries 22 and 23), but catalyst decomposi-
tion occurred during the reaction and Pd((R,R)-Me-Duphos)2
formed.14
The enantiomeric excess of the new phosphines was deter-
mined by integration of the NMR spectra of their diastereomeric
adducts with the chiral Pd fragment Pd((S)-Me2NCH(Me)C6H4)-
(Cl).15 In this assay, the adducts of PMe(Ar′)(Ar) showed similar
features for Ar′ ) Is and Mes; the 31P NMR signals for the
minor diastereomer usually appeared at lower field than those
for the major one, and in the 1H NMR spectra they appeared in
the opposite order (Supporting Information Table S2). These
trends suggested that the Pd((R,R)-Me-Duphos) catalyst con-
sistently favored one hand of the phosphine product. The order
of NMR shifts was reversed for PArMe(Phes). To reflect this
difference, ee values in entries 19-21 of Table 3 are arbitrarily
gave slow turnover and no enantioselectivity; for the latter two,
ligand displacement during catalysis was observed.
Substrate Scope. Since screening of the test reaction of
Schemes 1 and 2 revealed that the original Me-Duphos ligand
gave the highest enantioselectivity, we used it to explore the
scope of the reaction. Coupling of three secondary phosphines
PHMe(Ar′) (Ar′ ) 2,4,6-R3C6H2, R ) i-Pr (Is), Me (Mes), or
Ph (Phes)) with aryl iodides ArI gave enantioenriched methyl-
diarylphosphines PMe(Ar′)(Ar) in high yields (Scheme 3, Tables
3 and 4). The racemic phosphines were prepared for comparison
using the catalyst precursor Pd(P(o-Tol)3)2Cl2 and characterized
spectroscopically, by elemental analyses and mass spectroscopy,
and, in several cases, by X-ray crystallography (Supporting
Information).
Entries 1-3 in Table 3 showed that phenyl bromide (at 50
°C) and phenyl triflate could also be used, but selectivity was
lower than that with PhI. Therefore, we studied exclusively aryl
iodides.11 The disubstituted p-diiodobenzene could be selectively
coupled with 1 or 2 equiv of PHMe(Is) (entries 9 and 11). The
Pd(Me-Duphos) catalyst tolerated other aryl iodides with p- or
m-substituents (Table 3). Double m- substitution appeared to
slow the reaction for 3,5-Me2C6H3I (entry 13), and no turnover
was observed for the analogous substrate 3,5-(CF3)2C6H3I.
Although coupling of o-iodopyridine was fast (entry 14) and
o-MeC6H4I also reacted smoothly, catalytic turnover was very
slow with other o-substituted substrates (o-MeOC6H4I, o-ClC6H4I)
or did not occur (MesI, Mes ) 2,4,6-Me3C6H2).
A Pd(Et-FerroTANE) catalyst promoted the coupling of o-
substituted aryl iodides, as well as iron and manganese iodocy-
clopentadienyl complexes, but in disappointingly low ee (Table
4). This poor selectivity did not appear to result exclusively
from catalyst decomposition, since some Pd intermediates were
observed during catalysis (Supporting Information).
For the Pd(Me-Duphos) catalyst, while yields were generally
high and reactions were fast, with a few exceptions, such as
(12) (a) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195. (b)
McDaniel, D. H.; Brown, H. C. J. Org. Chem. 1958, 23, 420-427.
(13) Yoshifuji, M.; Shibayama, K.; Inamoto, N. Chem. Lett. 1984, 115-118.
(14) Blank, N. F.; McBroom, K. C.; Glueck, D. S.; Kassel, W. S.; Rheingold,
A. L. Organometallics 2006, 25, 1742-1748.
(11) For Pd-catalyzed phosphination using aryl bromides and chlorides, see:
Murata, M.; Buchwald, S. L. Tetrahedron 2004, 60, 7397-7403.
(15) Kyba, E. P.; Rines, S. P. J. Org. Chem. 1982, 47, 4800-4801.
9
6850 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007

Palladium-Catalyzed Asymmetric Phosphination
A R T I C L E S
Table 5. Selected Bond Lengths (Å) and Angles (deg) for the
Cations [M((R,R)-Me-Duphos)(Ph)(PHMeIs)]+ (M ) Pd or Pt)a
bond/angle
M
)
Pd
M ) Pt
M-P1
2.3204(19)
2.2775(19)
2.3347(18)
2.054(7)
89.82(18)
173.4(2)
85.53(7)
87.7(2)
174.48(7)
96.53(6)
104.3(3)
118.2(2)
118.6(2)
2.3096(11)
2.2772(13)
2.3126(13)
2.082(4)
90.21(13)
172.02(16)
85.52(4)
87.26(13)
177.11(4)
96.82(4)
M-P2
M-P3
M-C
C-M-P2
C-M-P1
P2-M-P1
C-M-P3
P2-M-P3
P1-M-P3
C(Me)-P3-C(Is)
C(Me)-P3-M
C(Is)-P3-M
103.9(3)
118.19(19)
120.38(15)
a Pd complex 4 (anion ) OTf) crystallized with 2 equiv of THF. The Pt
complex (anion ) BF4) crystallized with 1 equiv of CH2Cl2.16
slightly longer than the Pt-P ones, the Pd-C(Ph) distance was
shorter than the Pt-C one. Similar observations in simpler
compounds have been rationalized on the basis of hard-soft
considerations and relativistic effects.17
Figure 2. ORTEP diagram of [Pd((R,R)-Me-Duphos)(Ph)(PHMeIs)][OTf]‚
2THF (the anion and solvent molecules were omitted, and only the P-H
hydrogen is shown).
Treatment of 3 with PHMe(Is) and NaOSiMe3 at low
temperature gave the phosphido complex Pd((R,R)-Me-Duphos)-
(Ph)(PMeIs)) (5, Scheme 4). We considered several possible
mechanisms for this reaction. In analogous Pt chemistry, reaction
of Pt((R,R)-Me-Duphos)(Ph)(X) (X ) Cl or Br) with NaOSiMe3
gave the silanolate complex Pt((R,R)-Me-Duphos)(Ph)(OSiMe3),
which reacted quickly with PHMe(Is) to give 5-Pt.5a,18 However,
the palladium iodide 3 did not react with NaOSiMe3. Alterna-
tively, the anion [PMeIs]-, formed by deprotonation of PHMe-
(Is) with NaOSiMe3, might displace iodide from 3 to give 5,
but 1 and NaOSiMe3 did not react. We cannot rule out the
possibility that such pathways operate via unfavorable equilibria,
but cation 4-OTf reacted quickly with NaOSiMe3 to give 5,
which suggests that Pd-phosphido formation occurred by
coordination followed by deprotonation (Scheme 4).
Scheme 4. Proposed Mechanism for Pd-Catalyzed Synthesis of 2
from 1 ([Pd] ) Pd((R,R)-Me-Duphos, L ) PHMe(Is) (1), L′ )
(R,R)-Me-Duphos)
listed with a minus sign; note that we have determined the
absolute configuration of only one phosphine product (see
below).
Phosphido complex 5 was characterized by low-temperature
multinuclear NMR spectroscopy (see Table 6 for 31P NMR data
and the Supporting Information). Two diastereomers which
differ in the absolute configuration of the pyramidal phosphido
ligand would be expected,19 but we observed four diastereomers
of the Pt analogue 5-Pt.16 The “extra” isomers appeared to be
conformational isomers resulting from restricted rotation in the
bulky Pt-PMeIs group. Similarly, three diastereomers of 5 were
observed (-50 °C, 92:2:0.4 ratio, Table 6). Since, unlike 5-Pt,
thermally sensitive 5 could not be purified (see below), it was
difficult to rule out the possibility that minor signals in the
spectra of 5 were due to impurities. Indeed, independent
generation of the iodo complex Pd((R,R)-Me-Duphos)(I)(PMeIs)
showed that it was present in some cases, presumably arising
Mechanism of Pd-Catalyzed Asymmetric Phosphination
(Scheme 4). Study of the individual steps in a potential
mechanism for the Pd((R,R)-Me-Duphos)-catalyzed coupling of
PHMe(Is) with PhI (Scheme 4) provided information on the
mechanism of P-C bond formation. Treatment of Pd((R,R)-
Me-Duphos)(Ph)(I) (3) with PHMe(Is) (1) led to broadening of
the 31P NMR peaks of the phosphine at room temperature. Low-
temperature 31P NMR spectroscopy showed that the phosphine
reversibly displaced iodide to form the cation [Pd((R,R)-Me-
Duphos)(Ph)(PHMe(Is))][I] (4) as a mixture of diastereomers;
the equilibrium favored neutral 3. Cation 4 was independently
synthesized as the isolable triflate salt (4-OTf) from 3, phosphine
1, and AgOTf.16 Using a stoichiometric amount of PHMe(Is)
gave a 1:1 mixture of diastereomers of 4-OTf. Heating this
material in THF, or use of excess phosphine in the original
synthesis, gave 4-OTf in 1:1.4 ratio.
(17) (a) Wisner, J. M.; Bartczak, T. J.; Ibers, J. A.; Low, J. J.; Goddard, W. A.,
III. J. Am. Chem. Soc. 1986, 108, 347-348. (b) Wisner, J. M.; Bartczak,
T. J.; Ibers, J. A. Organometallics 1986, 5, 2044-2050. (c) Oberhauser,
W.; Stampfl, T.; Bachmann, C.; Haid, R.; Langes, C.; Kopacka, H.;
Ongania, K.-H.; Bruggeller, P. Polyhedron 2000, 19, 913-923.
(18) Scriban, C.; Glueck, D. S.; Golen, J. A.; Rheingold, A. L. Organometallics
2007, 26, 1788-1800.
(19) (a) Wicht, D. K.; Glueck, D. S.; Liable-Sands, L. M.; Rheingold, A. L.
Organometallics 1999, 18, 5130-5140. (b) Wicht, D. K.; Kovacik, I.;
Glueck, D. S.; Liable-Sands, L. M.; Incarvito, C. D.; Rheingold, A. L.
Organometallics 1999, 18, 5141-5151. (c) Zhuravel, M. A.; Glueck, D.
S.; Zakharov, L. N.; Rheingold, A. L. Organometallics 2002, 21, 3208-
3214.
As expected, the crystal structure of 4-OTf (Figure 2, Table
5 and the Supporting Information) was very similar to that of
the Pt analogue; in both cases the crystal examined contained
the RP-diastereomer.16 While two of the three Pd-P bonds were
(16) For the analogous Pt complexes, see: (a) ref 5a (b) Scriban, C.; Glueck,
D. S.; DiPasquale, A. G.; Rheingold, A. L. Organometallics 2006, 25,
5435-5448.
9
J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007 6851

A R T I C L E S
Blank et al.
Table 6. 31P NMR Data for the Phosphido Complexes
M((R,R)-Me-Duphos)(Ph)(PMeIs) (M ) Pd, Pt)a
Scheme 5. [Pd] ) Pd((R,R)-Me-Duphos), L ) PHMe(Is)
T
δ
(JPt
(P1)
δ
(JPt
(P2)
δ
(JPt
(P3)
complex
(
°
C)
)
)
)
J12
J13
J23
drb
-P
-P
-P
Table S7 (Supporting Information) summarizes 31P NMR data
for 6 and analogous three-coordinate Pd(Duphos) complexes
prepared in this study (see below).21
Pd(Me-Duphos)(Ph)- -50 49.0
(PMeIs) (5a)
55.0 -34.0 27 127 13 92
5b
5c
50.1
-22.2 27 137
2
0.4
98
-29.8
130
136
On decomposition of 5 in the presence of a trap (conditions
more relevant to catalysis), three-coordinate complexes 6 were
not observed. Warming a mixture of 5 and PhI gave phosphine
2 and oxidative addition product 3 (Scheme 4). These observa-
tions, along with kinetic studies of reductive elimination (see
below), suggest that rate-determining formation of 6 from 5 was
followed by rapid reaction with PhI to yield 3. The Pd(Me-
Duphos) fragment resulting from reductive elimination could
also be trapped by the secondary phosphine substrate. During
catalytic coupling of PHMe(Is) with PhI at room temperature,
31P NMR monitoring initially showed the presence of phosphido
complex 5 and Pd((R,R)-Me-Duphos)(PHMe(Is))2 (7), which
underwent rapid exchange on the NMR time scale with PHMe-
(Is). Complex 7 was generated free of 5 by running catalysis
with a substoichiometric amount of PhI, or by addition of excess
PHMe(Is) to Pd((R,R)-Me-Duphos)(trans-stilbene) (Scheme 5).
At -40 °C, phosphine exchange in 7 was slow on the NMR
time scale and the expected mixture of four diastereomers was
observed.
Pt(Me-Duphos)(Ph)- -75 56.5
57.3 -55.9
(PMeIs) (5a-Pt)
5b-Pt
(1653) (1868) (889)
-47.3
(∼940)
-49.5
-51.6
136
2.2
5c-Pt
5d-Pt
142
134
1.2
1
a All Duphos ligands have (R,R) configuration. Chemical shifts in
ppm (85% H3PO4 external standard). Coupling constants in Hz.
Solvent ) THF-d8. For atom labeling, see the figure above. In several cases
(especially for minor diastereomers), it was not possible to assign
Duphos P resonances, so they are omitted. Data for the Pt complex is from
b
ref 16. dr ) diastereomer ratio, from integration of the 31P NMR spectra
at -50 °C.
As catalysis proceeded, complex 3 was also observed; it
became the dominant Pd species present near the end of the
reaction. In catalysis with PhOTf, the same Pd complexes were
observed,22 but the ratio of 7 to 5 was greater, consistent with
faster oxidative addition of PhI.23 These observations suggest
that the rates of oxidative addition and reductive elimination
during catalysis were similar and that Pd-P bond formation
could be faster than both steps under appropriate conditions.
Scheme 6 shows how the relative rates of interconversion
between invertomers SP- and RP-5 and their reductive elimination
might control enantioselection in catalysis. Note that this
idealized scheme neglects possible contributions to product
formation from diastereomer 5c (or other rotamers), which were
observed only at low temperature. Under catalytic conditions
at room temperature, only one set of phosphido signals was
observed; we cannot tell if this was an average resulting from
coalescence of the 5a/5b signals or if coalescence had not
occurred, and peaks due to 5b were too broadened to observe
under these conditions.24
Figure 3. Temperature dependence of the equilibrium between diastere-
omers 5a and 5b of Pd((R,R)-Me-Duphos)(Ph)(PMeIs) in THF/THF-d8. Keq
values are the average of two separate measurements (Supporting Informa-
tion).
from Pd((R,R)-Me-Duphos)I2 as an impurity in 3.20 On warming,
peaks due to minor isomer 5c disappeared and those of 5b
broadened considerably; this behavior may correspond to a
coalescence of rotamers as in analogous Pt complexes.16 We
could not observe the expected coalescence of the 5a/5b signals
because of rapid decomposition on warming (see below).
The temperature dependence of the equilibrium between
diastereomers 5a and 5b from -70 to -35 °C is shown in Figure
3. As desired (see the Introduction), the equilibrium lay far to
one side (Keq ) 70 at -70 °C in THF/THF-d8). Extrapolation
to catalytic conditions at room temperature (298 K) gave Keq
) 18 and ∆G°298 ) 1.7 kcal/mol.
Decomposition of 5 via reductive elimination on warming
gave, in the absence of a trap, a mixture of diastereomers of
the three-coordinate phosphine complex Pd((R,R)-Me-Duphos)-
(PPhMe(Is)) (6), which decomposed even at low temperature
to yield Pd((R,R)-Me-Duphos)2 and PPhMe(Is) (Scheme 4).6b
If 5a and 5b both undergo reductive elimination faster than
P inversion, then the enantiomeric ratio (er) of product 2 would
(21) The simpler three-coordinate complexes Pd((R,R)-Me-Duphos)(PR3) (R )
t-Bu or Cy) were prepared by reduction of Pd((R,R)-Me-Duphos)Cl2 with
NaBH(OMe)3 in the presence of the ligand (see the Supporting Information
for details). Although these complexes of bulky trialkylphosphines were
more stable than 6, they also decomposed to Pd((R,R)-Me-Duphos)2 at room
temperature.
(20) Data for minor diastereomer 5c are included in Table 6 since these were
reproducibly observed. In one sample of 5, under conditions designed to
maximize signal/noise, we also saw a minor (5a/5d ratio ca. 300:1) PMeIs
peak due to another phosphido complex, 5d (THF/THF-d8, -70 °C, δ -24.7
(dd, J ) 117, 35 Hz)). We could not detect the analogous Me-Duphos
resonances or tell if this material was a diastereomer of 5a-c or an impurity;
this signal disappeared on warming. For more details on independent
generation of the other potential impurities Pd((R,R)-Me-Duphos)(Ph)(PHIs)
and Pd((R,R)-Me-Duphos)(Me)(PPhIs), see the Supporting Information.
(22) Reaction mixtures with PhOTf contained 1 equiv of NaI per Pd, derived
from precursor 3. Independent generation of Pd((R,R)-Me-Duphos)(Ph)-
(OTf) (from Pd((R,R)-Me-Duphos)(trans-stilbene) and PhOTf) in the
presence of NaI gave 3, consistent with the observations in the catalytic
system.
(23) Alcazar-Roman, L. M.; Hartwig, J. F. Organometallics 2002, 21, 491-
502.
9
6852 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007

Palladium-Catalyzed Asymmetric Phosphination
A R T I C L E S
Scheme 6. Probing Relative Rates of P Inversion and Reductive
Elimination in Phosphido Complex 5: “er vs dr” Experiments ([Pd]
) Pd((R,R)-Me-Duphos)
Scheme 7 . Proposed Curtin-Hammett Kinetic Scheme for
Inversion and Reductive Elimination in Phosphido Complex 5a
a [Pd] ) Pd((R,R)-Me-Duphos). The labels for the compounds (A1 and
A4 are phosphine 2; A2 and A3 are phosphido complexes 5) and the rate
constants follow Seeman’s review.26 Rate constants k23 and k32 describe
the composite process of P inversion and rotation about the Pd-P bond
which interconverts A2 and A3, and rate constants k21 and k34 characterize
reductive elimination of the phosphido diastereomers Pd((R,R)-Me-Du-
phos)(Ph)(PMeIs) (A3 and A2, or 5a-b). The RP diastereomer A3 is
arbitrarily assumed to be the major one.
reflect the thermodynamic ratio of their precursors 4a and 4b
(er ) dr, Scheme 6). Alternatively, if interconversion of 5a and
5b is faster than reductive elimination, their relative abundance
and reductive elimination rates might result in an er of 2 different
from the original dr of cations 4 (k1 * k2; er * dr, Scheme
6).25-27 Deprotonation of 1:1 or 1.4:1 mixtures of cations 4-OTf
with NaOSiMe3 in the presence of PhI gave phosphine 2 (er )
6:1).28 Therefore, the rate of inversion was greater than or equal
to that of reductive elimination under these conditions. Generat-
ing 5 by treatment of 3 with phosphine 1, PhI, and NaOSiMe3
at room temperature led to analogous results (er ) 6:1).28 The
similar product ratio observed in the catalytic reactions provides
a rationale for the observed enantioselection.29
Mechanistic Hypothesis: Curtin-Hammett Kinetics. We
hypothesized that inversion was much faster than reductive
elimination in Scheme 6. Then, in an example of Curtin-
Hammett kinetics, the product ratio [2a]/[2b] ) P ) Keq(k34/
k21) (Scheme 7; Keq is the equilibrium constant for diastereomers
5a and 5b, while the rate constants describe their reductive
elimination rates).26 Note that Schemes 6 and 7 simplify the
reductive elimination process by omitting the three-coordinate
intermediate 6 (Scheme 4); this is justified because rate-
determining formation of 6, which fixes the P stereochemistry
in the product 2, is followed by fast release of the product. We
also assume that P-C reductive elimination proceeds with
retention of configuration at P, as we showed earlier in reductive
elimination of the analogous phosphido-borane complexes Pd-
((S,S)-Chiraphos)(o-MeOC6H4)(PMePh(BH3)).30
Following Seeman’s review, the kinetics of such Curtin-
Hammett systems can be described in terms of the empirical
Winstein-Holness rate constant kWH, which describes the total
rate of product formation.26
d
dt
d
[A1] + [A4] ) kWH([A2] + [A3])
dt
(1)
After experimental measurement of kWH, the reductive elimina-
tion rate constants k21 and k34 are given by
Keq + 1
P
P + 1
k34 ) kWH
(2)
(3)
(
)(
)
Keq
(Keq + 1)
(P + 1)
k21 ) kWH
where Keq and P are defined as in Scheme 7.
(24) Coalescence of the Pd-PMeIs 31P NMR signals near room temperature is
plausible. Interconversion of 5a and 5b requires both P inversion and
rotation about the Pd-P bond. An approximation for the barrier to this
composite process at the coalescence temperature is ∆GqTc ) 4.58Tc(10.32
+ log Tc/kc) cal/mol, where kc ) 2.22∆ν (Friebolin, H. Basic One- and
Two-Dimensional NMR Spectroscopy, 2nd ed.; VCH: Weinheim, Germany,
1993). Using the ∆ν value from Table 6, measured at 202 MHz, coalescence
at 298 K would correspond to a barrier of ca. 12 kcal/mol, similar to those
for related Pt and Pd complexes.16,19
Kinetic Studies of Reductive Elimination from Phosphido
Complex 5. Phosphido complex 5 was generated at -78 °C in
THF/THF-d8 by deprotonation of cation 4-OTf or by treatment
of Pd((R,R)-Me-Duphos)(Ph)(I) (3) with PHMe(Is) and NaO-
SiMe3. In both cases, reductive elimination in the presence of
PhI at -10 °C gave PPhMe(Is) (2) and 3 (see Scheme 4); this
process was observed by 31P NMR spectroscopy.
Since the two enantiomers of product 2 gave identical 31P
NMR spectra, the Winstein-Holness rate constant kWH could
be measured directly by monitoring the growth of the phosphine
over time by NMR spectroscopy. Alternatively, kWH could be
obtained by observing formation of Pd complex 3 or disap-
pearance of the major phosphido diastereomer 5a (we could
not obtain reliable NMR integration for the broadened peaks
of minor isomer 5b). Both decomposition of 5a and formation
of the products fit first-order kinetics.31 Thus, concentration vs
(25) Halpern, J. Science 1982, 217, 401-407.
(26) Seeman, J. R. Chem. ReV. 1983, 83, 83-134. For an application in
organometallic chemistry, see: Gately, D. A.; Norton, J. R. J. Am. Chem.
Soc. 1996, 118, 3479-3489.
(27) For a related experiment involving alkylation of Fe-phosphido complexes,
see: Crisp, G. T.; Salem, G.; Wild, S. B.; Stephens, F. S. Organometallics
1989, 8, 2360-2367.
(28) Small variations in er were observed; see the Supporting Information for
results and estimation of the error in measurement of er.
(29) Cations 4a and 4b might interconvert by dissociation of PHMe(Is),
racemization of the phosphine by reversible deprotonation/protonation, and
rebinding (Bader, A.; Nullmeyers, T.; Pabel, M.; Salem, G.; Willis, A. C.;
Wild, S. B. Inorg. Chem. 1995, 34, 384-389). If this process occurred
very quickly, and one of the diastereomeric cations was deprotonated more
rapidly than the other, then it is possible that er * dr in Scheme 6, even if
reductive elimination was faster than inversion. However, the lack of
dependence of the ee in the catalytic reactions on the concentration or the
nature of the base (Table 1, entries 11-14) suggests that the base, and,
specifically, the deprotonation step, does not affect enantioselectivity.
(30) (a) Moncarz, J. R.; Brunker, T. J.; Glueck, D. S.; Sommer, R. D.; Rheingold,
A. L. J. Am. Chem. Soc. 2003, 125, 1180-1181. (b) Reference 6b.
9
J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007 6853

A R T I C L E S
Blank et al.
complete. After isolation of the product phosphine, the ee
(product ratio P ) 8:1) was determined. Peak broadening and
competing reductive elimination precluded similar direct mea-
surement of Keq at -10 °C, but extrapolation of the data from
Figure 3 gave Keq ) 27.
These results provided two kinetically indistinguishable sets
of values for the reductive elimination rate constants k21 and
k34 of Scheme 7. If the major product phosphine was formed
from the major starting material (case 1, Keq ) 27, P ) 8), k34
) 1.4 × 10-4 s-1 and k21 ) 4.7 × 10-4 s-1 34
. Alternatively, if
the minor starting material yielded the major product (case 2,
Keq ) 27, P ) 1/8), then k34 ) 1.7 × 10-5 s-1 and k21 ) 3.7
Figure 4. Concentration vs time plots for reductive elimination of 5 in the
presence of PhI to yield 2 and 3 (THF/THF-d8, -10 °C). Blue squares:
[5a]. Red circles: [2]. Black triangles: [3]. The smooth lines are fits to the
equations A ) A0e-kt and B ) A0(1 - e-kt) for decay of 5a (A) and growth
of 2 and 3 (B) with k ) 1.13 × 10-4 s-1 (r2 ) 0.996 for decay of 5a). The
same value of k was used to fit the growth of products 2 and 3. Directly
fitting concentration vs time data for 2 and 3 yielded slightly different rate
constants; see Table 7, entry 1 and the Supporting Information for more
details.
× 10-3 s-1 34
. Thus, in both cases, reductive elimination of the
minor phosphido intermediate 5b occurred more quickly than
that for the major one. If 5b decayed about 3 times faster than
5a (case 1), this would simply reduce the ee, which would still
be controlled by the thermodynamic preference for 5a. In case
2, however, where reductive elimination of 5b is about 200 times
faster than that of 5a, the effect of Keq on the product ratio
would be outweighed by the relative rates of reductive elimina-
tion.
time plots yielded three independent determinations of kWH for
each run (Figure 4; the smooth curves are fits to first-order decay
of 5a and growth of 2 (red circles) and 3 (black triangles)). As
shown in Table 7, values for kWH obtained by observing the
disappearance of 5a were consistent with those calculated from
the appearance of the products 2 and 3.
These results enabled us to confirm the hypothesis that P
inversion in phosphido intermediates 5 was much faster than
reductive elimination (Schemes 6 and 7). We have shown
previously that the barrier to P inversion in analogous Pt and
Pd phosphido complexes was the same within experimental
error.19c It was not possible to measure the rate of P inversion
in 5. However, the rate constants for the process of inversion
and rotation about the Pt-P bond which interconverts the
diastereomers of the Pt analogues Pt((R,R)-i-Pr-Duphos)(Ph)-
(PMeIs) (8-Pt) and Pt((R,R)-Me-Duphos)(I)(PMeIs) were ca. 5
× 102 s-1 and 8 × 102 s-1 at the coalescence temperature of
-10 °C,24 respectively, at least 105 times greater than the rate
constants for reductive elimination from 5 at the same
temperature.16b
Decomposition of 5a generally gave the most reliable fits;
data for formation of products 2 and 3 are included in Table 7
as an indication of the error in determination of kWH. The results
did not differ significantly with the two different precursors
(compare entries 1-4 and 5-6)32 or depend on initial concen-
tration of 5. Using higher concentrations of PhI gave similar
results (Supporting Information), consistent with first-order
decomposition.33 Without PhI (entry 4), decomposition of 5 gave
a mixture of diastereomers of three-coordinate intermediate 6,
which decomposed to yield 2, Pd((R,R)-Me-Duphos)2, and,
presumably, Pd(0) (the 31P NMR signal of 2 broadened over
time, perhaps due to an exchange process with a zerovalent Pd
complex of the phosphine). The first-order rate constant for
decay of 5a under these conditions was comparable to that in
the presence of PhI. An approximate value for kWH is then 1.5-
Absolute Configurations of the Products and Intermedi-
ates. Determination of the absolute configurations of the product
phosphine 2 and the intermediate phosphido complexes 5 would
distinguish the two possibilities described above and thereby
establish the origin of enantioselectivity in the catalytic asym-
metric synthesis of 2 from 1.
((0.5) × 10-4 s-1
.
A suitable phosphine, P(p-MeOC6H4)Me(Is) (2-An), was
prepared catalytically in 83% ee. As discussed above, the
consistent trends in the 31P and 1H NMR spectra of the adducts
of the phosphines in Table 3 with (Pd((S)-Me2NCH(Me)C6H4)-
(µ-Cl))2 suggested that the major enantiomer of this phosphine
and its Ph analogue 2 had the same absolute configuration (see
the Supporting Information). Six recrystallizations of 2-An from
i-PrOH gave a white crystalline solid with ee >99% by HPLC
on a chiral stationary phase (ChiralPak AD; see the Supporting
Information for details). The crystal structure of this highly
enriched sample showed the presence of two independent
molecules per asymmetric unit. Both had SP-configuration
(Figure 5). Chiral HPLC analysis of the single crystal used for
structure determination confirmed that it was the major enan-
Extracting the reductive elimination rate constants k21 and
34 from eqs 2 and 3 required kWH, the product ratio P, and Keq.
k
The kinetics studies yielded both kWH and P. The latter was
measured under conditions (see Table 7, entry 1 and the
Supporting Information) where complete decomposition of 5
occurred at -10 °C. Thus, the reaction was monitored until ca.
80% conversion. Samples were then transferred to a cold bath
(-10 °C, ice/brine/NaCl), which was placed in a freezer
(-10 °C) and stored for at least 24 h until the reaction was
(31) (a) See the Supporting Information for demonstration that kWH can be
measured by observing decay of either diastereomer of starting material 5,
as a complement to the more usual method, observing product formation.
(b) Concentration vs time data were fit using TableCurve2D software
(SYSTAT Software Inc., http://www.systat.com/products/TableCurve2D/
help/?sec)1106, accessed July 2006), using A ) A0e-kt and [B] )
[A0](1 - e-kt) respectively for the first-order formation of B from A. See
the Supporting Information for more details.
(34) The error in determining kWH is relatively large (Table 7), while there are
also errors associated with the measurement of Keq and P and, hence, the
derived rate constants k21 and k34. For our purposes of comparing the
magnitude of k21/k34 with k23/k32 and determining which diastereomer of 5
leads to the major product, however, it is not necessary to determine these
numbers with great precision.
(32) The different precursors yielded 5 along with the byproducts NaOTf and
NaI, respectively. While it is possible that the presence of these salts might
lead to different rates of reductive elimination, our data are indistinguishable
within experimental error.
(33) Hartwig, J. F. Acc. Chem. Res. 1998, 31, 852-860.
9
6854 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007

Palladium-Catalyzed Asymmetric Phosphination
A R T I C L E S
Table 7. Kinetics Data for Reductive Elimination from Pd((R,R)-Me-Duphos)(Ph)(PMeIs) (5) in the Presence of PhI at -10 °C in
a
THF/THF-d8
1
1
1
1
equiv of
PhI
k
WH (s-
,
k
WH (s-
,
k
WH (s-
,
k ,
WH (s-
entry
precursor
decay of 5a)
formation of 2)
formation of 3)
average)
1b
2
3
4
5
A (0.20 M)
A (0.187 M)
A (0.11 M)
A (0.15 M)
B (0.145 M)
B (0.20 M)
2
1.13 × 10-4
1.31 × 10-4
1.34 × 10-4
1.92 × 10-4
1.28 × 10-4
1.49 × 10-4
9.06 × 10-5
8.60 × 10-5
2.12 × 10-4
d
1.07 × 10-4
1.04 × 10-4
1.09 × 10-4
2.03 × 10-4
1.92 × 10-4
1.13 × 10-4
1.92 × 10-4
1.5
1.9
0
2
2
c
2.63 × 10-4
d
1.19 × 10-4
9.19 × 10-5
6
2.33 × 10-4
1.94 × 10-4
a Phosphido complex 5 was generated by deprotonation of precursor A (4-OTf) with NaOSiMe3 or by treatment of precursor B (3) with PHMe(Is) and
NaOSiMe3. As in Figure 4, concentration vs time data for 5a, 2, and 3 were fit separately to first-order exponential functions. See Experimental Section and
b
c
d
Supporting Information for more details. Product ee ) 79%. Concentration vs time data for 3 were too noisy to give an acceptable fit. Without the PhI
trap, 3 was not formed, and formation and decomposition of the intermediate 6 prevented monitoring [2] over time.
Figure 5. ORTEP diagram of one of the two independent molecules of
SP-P(p-MeOC6H4)Me(Is) (2-An).
tiomer of phosphine 2-An. Racemic 2-An crystallized in a
different space group (Supporting Information).
Figure 6. Partial ORTEP diagram of 5-Pt,16b used as a model to show
NOEs observed between Pd-Ph hydrogens H2-H6 and the PMeIs
hydrogens (C40 ) P-Me; C38 and C39 ) i-Pr Me; H37 ) i-Pr CH) in
the major diastereomer 5a (THF/THF-d8, -50 °C). Selected distances in
5-Pt (in Å) were estimated using the calculated hydrogen atom positions:
H37-H6 2.393, H37-H5 4.495, H37-H2 4.895, H38-H6 3.860 (av),
H39-H6 3.790 (av), H40-H2 4.099 (av), H40-H3 4.851 (av), H40-H4
5.053 (av), H40-H5 4.400 (av), H40-H6 3.532 (av).
We could not crystallize thermally unstable intermediate 5,
so its absolute configuration was investigated by low-temper-
ature NMR spectroscopy, with comparison to NMR and
crystallographic results for RP-5-Pt.16b,18 The similarity between
the structures of cations 4 and 4-Pt (Table 5) suggested that
5-Pt would be a good structural model for 5. Unfortunately,
direct confirmation that the solution structure of the major
diastereomer of 5-Pt was the same as that in the solid state was
not possible because rapid P inversion would interconvert the
diastereomers even at low temperature. However, low-temper-
ature NOESY studies of 5-Pt showed that the solution structure
of the major diastereomer was consistent with the solid-state
one (RP).18 In the absence of similar NMR data for the minor
isomer, not available because of its low concentration, we could
not rule out the possibility that the major solution isomer is
actually SP. However, comparison to the structures of Pt((R,R)-
i-Pr-Duphos)(Ph)(PMeIs) (8-Pt) and related complexes de-
scribed elsewhere suggested that the major diastereomer of 5-Pt
contained an RP phosphido group.18
Scheme 8 . Approximate Rate and Equilibrium Constants for
Inversion and Reductive Elimination of 5a (A3) and 5b (A2) at
-10 °C in THF ([Pd] ) Pd((R,R)-Me-Duphos))
Multinuclear NMR and NOESY data for the major diaster-
eomers of 5 and 5-Pt were in excellent agreement (Supporting
Information), so it is likely that the absolute configuration of
major diastereomer 5a was also RP. Figure 6 shows selected
NOEs between PMeIs and Pd-Ph hydrogens in 5a, which are
consistent with the solid-state structure of 5-Pt (see the footnote
in Figure 6 for selected H-H distances, estimated using the
calculated hydrogen atom positions, which correspond to the
observed NOEs).
Since P-C reductive elimination from Pd(II) proceeded with
retention of configuration, SP-2 must be formed from RP-5.30
(The apparent inversion of configuration is a consequence of
the sequence rules as a P-Pd bond is converted to a P-Ph
bond.) Thus, it appears that the major enantiomer of product
phosphine 2 was formed from the major diastereomer of
intermediate phosphido complex 5. Scheme 8 summarizes these
conclusions and includes the approximate reductive elimination
and inversion rate constants.
In contrast to the classic major-product-from-minor-in-
termediate result in asymmetric hydrogenation,25,35 these results
suggest that enantioselectivity was determined mainly by the
thermodynamic preference for 5a, although faster reductive
elimination of minor diastereomer 5b reduced the product ratio.
We reached a similar conclusion in Pt-catalyzed asymmetric
9
J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007 6855

A R T I C L E S
Blank et al.
Table 8. Substituent Effects on Diastereomer Ratio in the
Scheme 9. Structure-Reactivity Relationships in Pd(Duphos) Aryl
Phosphido Complexesa
Phosphido Complexes Pd((R,R)-Duphos)(Ph)(PMeAr′) at -60 °C
in THFa
Duphos (number)
Ar
′
dr
(R,R)-i-Pr-Duphos (8)
(R,R)-Me-Duphos (5)
(R,R)-Me-Duphos (9)
(R,R)-Me-Duphos (10)
Is
Is
Phes
Mes
31:1:4:3
137:2.4:1
>100:1b
7.3:1
a Diastereomer ratios were determined by 31P NMR integration. Four
diastereomers were observed for 8, three for 5,20 only one for 9, and two
b
for 10. Only one diastereomer of 9 was observed.
Table 9. Substituent Effects on Diastereomer Ratio (Keq) in the
Phosphido Complexes Pd((R,R)-Me-Duphos)(p-XC6H4)(PMeIs) at
-60 °C in THFa
12
X
Keq
σp
NH2
MeO
H
34
48
57
33
56
-0.32
-0.27
0
0.18
0.78
I
NO2
a Phosphido complexes: [Pd] ) Pd((R,R)-i-Pr-Duphos) (R ) i-Pr), X
) H, R′ ) i-Pr (8); [Pd] ) Pd((R,R)-Me-Duphos) (R ) Me), X ) H, R′ )
i-Pr (5), R′ ) Ph (9), R′ ) Me (10); R′ ) i-Pr, X ) NH2 (5-NH2), X ) I
(5-I), X ) NO2 (5-NO2).
a The equilibrium constants (diastereomer ratios between 5a and 5b and
their analogues) were determined by 31P NMR integration.
reomer of Pd((R,R)-Me-Duphos)(Ph)(PMePhes) (9) was ob-
served; the dr at -60 °C was lower for PMeIs analogue 5 (three
diastereomers) and lower still for the two diastereomers of
PMeMes complex 10, as expected if the equilibrium constant
depended strongly on steric effects. Reductive elimination from
PMePhes complex 9, which occurred at -60 °C, was faster than
that in 5. So was decomposition of PMeMes complex 10, which
took about 3 h at -20 °C (compare 5, in which reductive
elimination was not complete after 3 h at -10 °C). Decomposi-
tion of Pd((R,R)-Me-Duphos)(Ph)(PMePhes) (9) in the presence
of PhI gave a mixture of the three-coordinate Pd((R,R)-Me-
Duphos)(Ph)(PPhMe(Phes)) (6-Phes) and 3, which suggests that,
in contrast to 5, reductive elimination and trapping of Pd(0)
with PhI were similar in rate. Thus, modifying the aryl
substituent in the phosphido ligand resulted in significant
changes to both thermodynamics and kinetics of these analogues
of 5. However, generation and reductive elimination of 9 and
10, at least in our initial experiments, was less clean than the
chemistry of 5, precluding analogous kinetic studies of reductive
elimination.
Pd-Ar Group. We generated a series of Pd((R,R)-Me-
Duphos)(p-XC6H4)(PMeIs) intermediates (X ) NH2, MeO, H,
I, NO2) to assess the effect of the para substituent on the
equilibrium between the diastereomers and their rates of
reductive elimination. Table 9 shows the dependence of Keq
([5a]/[5b], as in Figure 3) at -60 °C on the aryl substituent.
As expected, given the evidence above for steric effects on
diastereomer ratio, the values were large and similar in
magnitude. The variations observed at this temperature did not
correlate with the electronic properties of the p-substituents as
judged by their Hammett constants or with enantioselectivity
in catalysis (Table 3, Figure 1).
alkylation involving the closely related intermediate 5-Pt, where
P-C bond formation involved nucleophilic attack of the
phosphido group on benzyl bromide.18
Structure-Reactivity Relationships. Assuming that this
mechanistic model is a general one, we generated a series of
derivatives Pd(diphos*)(Ar)(PMeAr′) to investigate the depen-
dence of dr (Keq), reductive elimination rate, and ee on the Pd-
bound chiral diphosphine ligand, the secondary phosphine
substrate, and the aryl iodide (Scheme 9). The intermediates
were formed either on direct treatment of Pd(diphos*)(Ar)(I)
with PHMe(Ar′) and NaOSiMe3 or on deprotonation of the
cations [Pd(diphos*)(Ar)(PHMe(Ar′))][OTf]. 31P NMR data for
these phosphido complexes (Supporting Information) were
similar to those shown in Table 6 for 5.
Diphos*. Pd((R,R)-i-Pr-Duphos)(Ph)(PMeIs) (8), as observed
for 8-Pt, existed as a mixture of four diastereomers at low
temperature.16 We presume that each of the expected “inver-
tomers,” which differ in their phosphido P configuration, was
a mixture of two rotamers in these sterically hindered complexes.
Changing from Me-Duphos in 5 to i-Pr-Duphos in 8 (and in
the Pt analogues) reduced the diastereomeric ratio (Table 8) and
also affected the kinetics of reductive elimination. P-C bond
formation in 8 occurred even at -60 °C, under conditions where
5 decomposed very slowly. Apparently, a combination of these
factors resulted in the unusual observation that these ligands,
of opposite absolute configuration, gave the same enantiomer
of 2 with very similar ee in catalysis (Table 2).10 These results
were consistent with the models of Schemes 7 and 8, but a
similar quantitative analysis was not possible because of the
presence of appreciable quantities of three diastereomers of 8
at low temperature.
Reductive elimination rates were qualitatively similar in this
series, except for the p-NO2 derivative, which underwent
anomalously fast reductive elimination, even at -70 °C. In the
crystal structure of Pd((R,R)-Me-Duphos)(p-NO2C6H4)(I) (Sup-
porting Information), the Pd-C bond length of 2.060(4) Å
(average value, two molecules in the unit cell) was slightly
shorter than that in the Pd-Ph derivative (2.084(3) Å).9 Thus,
Phosphido Ligand. The dr was also strongly affected by
modifying the P-aryl substituent (Table 8). Only one diaste-
(35) (a) Landis, C. R.; Halpern, J. J. Am. Chem. Soc. 1987, 109, 1746-1754.
(b) Drexler, H.-J.; Baumann, W.; Schmidt, T.; Zhang, S.; Sun, A.;
Spannenberg, A.; Fischer, C.; Buschmann, H.; Heller, D. Angew. Chem.,
Int. Ed. 2005, 44, 1184-1188. (c) Schmidt, T.; Baumann, W.; Drexler,
H.-J.; Arrieta, A.; Heller, D.; Buschmann, H. Organometallics 2005, 24,
3842-3848.
9
6856 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007

Palladium-Catalyzed Asymmetric Phosphination
A R T I C L E S
1H or 13C NMR chemical shifts are reported vs Me4Si and were
it seems unlikely that a ground-state effect based on large
changes in the Pd-C bond strength could explain the relative
rates of reductive elimination. Instead, a transition-state effect
based on the increased electrophilicity of the Ar group may
explain this observation and, more generally, the fast reductive
elimination for p-nitrophenyl groups in Pd-mediated formation
of C-X bonds (X ) N, O, S, P).33,36
1
determined by reference to the residual H or 13C solvent peaks. 31P
NMR chemical shifts are reported vs H3PO4 (85%) used as an external
reference. Coupling constants are reported in Hz. Unless otherwise
noted, peaks in NMR spectra are singlets, and absolute values are
reported for coupling constants. Elemental analyses were provided by
the Schwarzkopf Microanalytical Laboratory. Mass spectra were
obtained at the University of Illinois Urbana-Champaign.
Thus, changes in the structure of the key intermediate Pd-
(Duphos)(Ar)(PMeAr′) had significant effects on both the
equilibrium between the diastereomers and the rate of reductive
elimination. The lack of correlation between dr, reductive
elimination rate, and enantioselectivity in catalysis (Table 3)
is consistent with the CHWH model for enantioselection
(Scheme 7).
Unless otherwise noted, reagents were from commercial suppliers.
The following compounds were prepared by literature procedures:
8-iodoquinoline,38 (N,N)-dimethyl-4-iodo-aniline,39 FcI (Fc ) C5H5-
FeC5H4),40 CymI (Cym ) C5H4Mn(CO)3),41 PHMe(Is) and PHMe-
(Mes*),8 (Pd(S)-Me2NCH(Me)C6H4)(µ-Cl))2,42 Pd(diphos*)(trans-
stilbene), Pd((R,R)-Me-Duphos)(Ph)(I) and Pd(Me-FerroLANE)(Ph)(I),9
PhesBr,43 Pd(P(o-Tol)3)2Cl2,44 and Pd(TMEDA)(p-NO2C6H4)(I).45 Syn-
theses of the secondary phosphines PHMe(Phes) and PHMe(Mes) are
described in the Supporting Information.
Conclusions
General Procedure for Catalysis. To a solution of Pd((R,R)-Me-
Duphos)(Ph)(I) (3 mg, 0.004 mmol, 0.05 equiv) in toluene (0.5 mL) in
a 20 mL vial was added PhI (11 µL, 0.092 mmol), followed by PHMe-
(Is) (23 mg, 0.092 mmol) and then NaOSiMe3 (92 µL of a 1.0 M THF
solution, 0.092 mmol). The pale yellow solution turned dark yellow
and then orange upon addition of the base. The solution was transferred
to an NMR tube, and the vial was rinsed with additional toluene (0.5
mL), which was then transferred to the NMR tube. After ca. 10 min a
white solid precipitated. The progress of the reaction was monitored
by 31P{1H} NMR. Once the reaction was complete (ca. 1 h), the solvent
was removed in Vacuo. The yellow residue was dissolved in THF (0.5
mL), and the product phosphine was purified by column chromatog-
raphy on silica (9:1 petroleum ether/THF eluent, 4 in. column, 1 cm
diameter) and isolated as a clear colorless oil which became a white
solid upon storage at -25 °C for 4 d to give 16 mg (53% yield) of
isolated product. See the Supporting Information for ee and NMR yield
determination.
[Pd((R,R)-Me-Duphos)(Ph)(PHMe(Is))][OTf] (4a-b). To a stirring
solution of Pd((R,R)-Me-Duphos)(Ph)(I) (75 mg, 0.12 mmol) in THF
(2 mL) was added PHMe(Is) (31 mg, 0.12 mmol) followed by dropwise
addition of AgOTf (32 mg, 0.12 mmol), giving a brown precipitate.
The reaction mixture was stirred for 30 min. The solution was filtered
though Celite. The precipitate was rinsed with THF (2 × 2 mL). The
yellow-greenish solution was concentrated in Vacuo. The 31P{1H} NMR
spectrum showed a 1:1 ratio of diastereomers. The solution was layered
with petroleum ether and cooled overnight at -25 °C to yield a yellow
oil. Trituration with petroleum ether gave a pale-yellow solid, which
was dried in Vacuo to yield 81 mg (75%) of product as a 1:1 mixture
of diastereomers.
We have described enantioselective synthesis of P-stereogenic
phosphines by Pd-catalyzed asymmetric phosphination. The
reaction proceeded in high yields and up to 88% ee. Several
chiral diphosphines were suitable ligands; Me-Duphos resulted
in the highest selectivity for the test cross-coupling of PHMe-
(Is) and PhI. A variety of aryl iodides and secondary phosphines
could be used; enantioselectivity depended strongly on substrate
structure. Sterically demanding secondary phosphine Ar′ sub-
stituents and aryl iodides with electron-donating para substituents
gave the highest ee values.
Kinetic studies of reductive elimination suggested that the
rate of phosphorus inversion in the diastereomeric Pd-phosphido
intermediates Pd((R,R)-Me-Duphos)(Ph)(PMeIs) (5) was much
faster than the rate of subsequent reductive elimination. As a
result of this Curtin-Hammett behavior, summarized in Schemes
7 and 8, the product ratio depended on both the equilibrium
ratio of diastereomers 5 and their relative rates of reductive
elimination (P ) Keq(k34/k21)).
The substituent effects observed highlight the problems and
opportunities this mode of enantioselection poses for rational
catalyst design. From both this work and analogous Pt chemistry,
it appears relatively easy to engineer phosphido intermediates
with a large Keq by using phosphido substituents of markedly
different size.16 The hard part, which remains a general challenge
in asymmetric catalysis, is to control the relative rates of reaction
of major and minor diastereomers (k34/k21). In the case of 5,
studied in detail here, the faster reductive elimination of the
minor diastereomer (k34/k21 ≈ 1/3) worked against Keq to result
in a reduced ee. Further studies in these and related systems
will seek to understand the factors which control this reactivity,
in order to design systems where the major diastereomer reacts
more quickly than the minor one.
Anal. Calcd for C41H60F3O3P3PdS: C, 55.38; H, 6.80. Found: C,
55.44; H, 6.75. 31P{1H} NMR (THF-d8), a and b are the two
diastereomers: δ 73.5 (dd, J ) 24, 26, a), 72.0 (dd, J ) 369, 24, b),
70.4 (dd, J ) 29, 24, b), 69.8 (dd, J ) 375, 26, a), -54.3 (br dd, J )
375, 24, a), -59.0 (dd, J ) 369, 29, b). 1H NMR (THF-d8): δ 8.17-
8.14 (m, 1H, Ar), 8.08-8.00 (m, 3H, Ar), 7.86-7.78 (m, 4H, Ar),
7.72-7.70 (m, 1H, Ar), 7.42-7.40 (m, 1H, Ar), 7.38-7.35 (m, 1H,
Ar), 7.27-7.20 (m, 5H, Ar), 7.16-7.11 (m, 2H, Ar), 7.08-7.05 (m,
1H, Ar), 6.98-6.97 (m, 1H, Ar), 6.94-6.91 (m, 1H, Ar), 6.75-6.72
(m, 1H, Ar), 6.63 (dm, JPH ) 355, 1H, PH), 6.32 (dm, JPH ) 373, 1H,
Experimental Section
General Details. Unless otherwise noted, all reactions and manipu-
lations were performed in dry glassware under a nitrogen atmosphere
at 20 °C in a dry box or using standard Schlenk techniques. Petroleum
ether (bp 38-53 °C), ether, THF, CH2Cl2, and toluene were dried using
activated alumina columns similar to those described by Grubbs.37 NMR
spectra were recorded by using Varian 300 or 500 MHz spectrometers.
(38) Kondo, Y.; Manabu, S.; Uchiyama, M.; Sakamoto, T. J. Am. Chem. Soc.
1999, 121, 3539-3540.
(39) Barluenga, J.; Gonzalez, J. M.; Garcia-Martin, M. A.; Campos, P. J.;
Asensio, G. J. Org. Chem. 1993, 58, 2058-2060.
(40) Guillaneux, D.; Kagan, H. B. J. Org. Chem. 1995, 60, 2502-2505.
(41) Lo Sterzo, C.; Miller, M. M.; Stille, J. K. Organometallics 1989, 8, 2331-
2337.
(42) Tani, K.; Brown, L. D.; Ahmed, J.; Ibers, J. A.; Nakamura, A.; Otsuka, S.;
Yokota, M. J. Am. Chem. Soc. 1977, 99, 7876-7886.
(43) Olmstead, M. M.; Power, P. P. J. Organomet. Chem. 1991, 408, 1-6.
(44) Paul, F.; Patt, J.; Hartwig, J. F. Organometallics 1995, 14, 3030-3039.
(45) Kruis, D.; Markies, B. A.; Canty, A. J.; Boersma, J.; van Koten, G. J.
Organomet. Chem. 1997, 532, 235-242.
(36) Widenhoefer, R. A.; Buchwald, S. L. J. Am. Chem. Soc. 1998, 120, 6504-
6511.
(37) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518-1520.
9
J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007 6857

A R T I C L E S
Blank et al.
PH), 3.78-3.71 (m, 2H), 3.28-2.80 (m, 10H), 2.62-2.12 (m, 6H),
2.03-1.79 (m, 6H), 1.71-1.56 (m, 3H), 1.71-1.67 (m, 3H, P-Me,
overlapping with previous signals), 1.53 (dd, JPH ) 7, JHH ) 4, 3H,
Me), 1.49 (dd, JPH ) 7, JHH ) 3, 3H, Me), 1.38 (d, JHH ) 6, 6H, i-Pr),
Deprotonation of Cations 4a-b (er vs dr Experiments). To a
solution of Pd((R,R)-Me-Duphos)(Ph)(I) (3, 29 mg, 0.047 mmol) in
THF (1 mL) were added PHMe(Is) (12 mg, 0.047 mmol) and PhI (10
µL, 0.089 mmol, 1.9 equiv). This reaction mixture was added to solid
AgOTf (12 mg, 0.047 mmol). A white solid precipitated, and the
mixture was filtered through Celite twice and then transferred into an
NMR tube fitted with a rubber septum. 31P{1H} NMR (THF) of this
mixture showed a 1:1 ratio of cations 4a and 4b. NaOSiMe3 (48 µL of
a 1.0 M THF solution, 0.048 mmol) was added via microliter syringe,
and the mixture immediately turned orange. The reaction was complete
(by 31P{1H} NMR) after 1 h, and black solid was observed in the NMR
tube. The reaction mixture was filtered through Celite, and the solvent
was removed in Vacuo. Isolation, purification, and ee determination of
2 were performed as described above. Similar experiments were also
done on isolated samples of 4-OTf having different diastereomer ratios
(1:1 and 1.4:1). See the Supporting Information for details.
Sample Procedure for Kinetics Experiments: Reductive Elimi-
nation from Pd-Phosphido Intermediates in the Presence of PhI.
Method I. To a solution of 4-OTf (53.4 mg, 0.06 mmol) in 1 mL of
THF/THF-d8 was added PhI (7 µL, 0.06 mmol, 1 equiv) via syringe.
The reaction mixture was transferred to the inner part of a coaxial
NMR tube, which contained PPh2Me (∼5 mg) in 0.2 mL of THF in
the outer part of the tube as an external standard. The tube was capped
with a septum and cooled to -78 °C. NaOSiMe3 (60 µL, 1.0 M solution
in THF, 0.06 mmol) was injected via syringe, and the tube was
transferred to the precooled NMR probe. The reaction was monitored
by 31P{1H} NMR spectroscopy until ca. 80% reaction completion. Then
the NMR tube was transferred to a -10 °C cold bath (ice/brine/NaCl)
and kept in a -10 °C refrigerator for 24 h. Isolation of the product
and determination of the ee were done according to the standard
procedure.
1.33 (d, JHH ) 6, 6H, i-Pr), 1.29 (d, JHH ) 7, 6H, i-Pr), 1.27 (d, JHH
)
6, 6H, i-Pr), 1.26 (d, JHH ) 7, 6H, i-Pr), 1.39-1.25 (m, 2H, overlapping
with previous signals), 1.17 (dd, JPH ) 15, JHH ) 7, 3H, Me), 1.21-
1.13 (m, 7H, overlapping with previous signals), 1.06-1.02 (m, 3H,
P-Me), 0.93-0.82 (m, 9H), 0.83 (dd, JPH ) 14, JHH ) 7, 3H, Me,
overlapping with previous peak), 0.75 (dd, JPH ) 15, JHH ) 7, 3H,
Me). IR (KBr): 2960, 2918, 2863, 1650, 1634, 1562, 1458, 1419, 1381,
1271, 1221, 1145, 1117, 1023, 897, 765, 734, 699, 638, 528, 451.
Changing the Diastereomeric Ratio of Cations 4. Method I. A
solution of 4-OTf (30 mg, 0.034 mmol) in THF with initial diastere-
omeric ratio 1:1 was heated at 50 °C for 46 h. The diastereomeric ratio
changed to 1:1.4 (a/b). Method II. To a stirred solution of Pd((R,R)-
Me-Duphos)(Ph)(I) (50 mg, 0.081 mmol) in THF (2 mL) was added
PHMe(Is) (40.6 mg, 0.162 mmol, 2 equiv). After 5 min, when the
reaction mixture changed color from pale yellow to orange-brown,
AgOTf (22.5 mg, 0.087 mmol, 1.1 equiv) was added as a suspension
in THF (2 mL); a gray-brown precipitate formed. The reaction mixture
was filtered and concentrated to 0.5 mL. Petroleum ether was layered
on top of the solution, and slow diffusion overnight in a refrigerator
yielded a brown oil that solidified upon trituration with petroleum ether.
The diastereomeric ratio of the resulting compound was 1:1.4 (a/b).
Generation of Pd((R,R)-Me-Duphos)(Ph)(PMeIs) (5a-b). Method
1. To a cold (-78 °C) solution of Pd((R,R)-Me-Duphos)(Ph)(I) (3, 45
mg, 0.07 mmol), PHMe(Is) (18 mg, 0.07 mmol), and PhI (9 µL, 0.08
mmol, 1.1 equiv) in THF/THF-d8 (1.5 mL, 1:1) in an NMR tube was
added NaOSiMe3 (73 µL of a 1.0 M solution in THF, 0.073 mmol).
The tube was placed in the precooled (-50 °C) NMR probe. Method
2. To a cold (-78 °C) solution of 4-OTf (generated in situ by reaction
of Pd((R,R)-Me-Duphos)(Ph)(I) (44 mg, 0.07 mmol), PHMe(Is) (18
mg, 0.07 mmol), and AgOTf (18 mg, 0.07 mmol) in the presence of
PhI followed by filtration to remove AgI; alternatively, isolated 4-OTf
was used) was added NaOSiMe3 (71 µL of a 1.0 M solution in THF,
0.071 mmol). The tube was placed in the precooled (-50 °C) NMR
probe. In both cases, a color change to orange was observed on
formation of 5.
Method II. To a solution of 3 (49.3 mg, 0.08 mmol) was added PhI
(18 µL, 33 mg, 0.16 mmol, 2 equiv) via syringe followed by PHMe-
(Is) (20 mg, 0.08 mmol) as a solution in 0.5 mL of THF with several
drops of THF-d8. The reaction mixture was transferred to the inner
part of a coaxial NMR tube as described above. The tube was capped
with a septum and cooled to -78 °C. NaOSiMe3 (80 µL, 1.0 M solution
in THF, 0.08 mmol) was injected via syringe, and the tube was
transferred to the precooled (-10 °C) NMR probe. Reaction monitoring
and workup were as those in Method I above.
31P NMR Study of Catalysis at Room Temperature. An NMR
tube equipped with a rubber septum was charged with Pd((R,R)-Me-
Duphos)(Ph)(I) (21 mg, 0.034 mmol), PHMe(Is) (85 mg, 0.34 mmol,
10 equiv), and PhI (71 mg, 0.35 mmol, 10.2 equiv) in ca. 1 mL of
THF with a few drops of THF-d8 added. The tube was cooled to
-78 °C, and NaOSiMe3 (350 µL of 1.0 M solution in THF, 0.35 mmol,
10.3 equiv) was added via cannula. The tube was briefly inverted in
the cold bath to mix it and then placed in the NMR probe at 21 °C.
The reaction was monitored by 31P NMR over ca. 1 h, when it was
complete. A related experiment was done on the same scale, but using
PhOTf (79 mg, 0.35 mmol) instead of PhI.
Acknowledgment. We thank the National Science Founda-
tion, Union Carbide (Innovation Recognition Program), and the
Beckman Foundation (Scholarship for O.A.) for support, and
Solvias and Eastman for gifts of chiral phosphines.
Supporting Information Available: Experimental procedures
and characterization data. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA070225A
9
6858 J. AM. CHEM. SOC. VOL. 129, NO. 21, 2007