Enantioselective homogeneous catalysis and the '3,5-dialkyl meta- effect'. MeO-BIPHEP complexes related to heck, allylic alkylation, and hydrogenation chemistry

Article abstract of DOI:10.1021/ja964406g

The enantioselectivities arising from a Pd-catalyzed Heck reaction (>98% ee) and an allylic alkylation (>90% ee) using a 3,5-di-tert-butyl-MeO-BIPHEP chiral auxiliary (1) are reported. Higher ee's are observed with the 3,5- dialkyl substituents than with

Full text of DOI:10.1021/ja964406g

J. Am. Chem. Soc. 1997, 119, 6315-6323
6315
Enantioselective Homogeneous Catalysis and the “3,5-Dialkyl
Meta-Effect”. MeO-BIPHEP Complexes Related to Heck,
Allylic Alkylation, and Hydrogenation Chemistry
Gerald Trabesinger,† Alberto Albinati,‡ Nantko Feiken,† Roland W. Kunz,§
Paul S. Pregosin,*,† and Matthias Tschoerner†
Contribution from the Laboratorium fu¨r Anorganische Chemie, ETH Zentrum,
8092 Zu¨rich, Switzerland, Organische Chemie, UniVersita¨t Zu¨rich, 8057 Zu¨rich, Switzerland,
and Chemical Pharmacy, UniVersity of Milan, I-20131 Milan, Italy
ReceiVed December 23, 1996. ReVised Manuscript ReceiVed April 24, 1997X
Abstract: The enantioselectivities arising from a Pd-catalyzed Heck reaction (>98% ee) and an allylic alkylation
(>90% ee) using a 3,5-di-tert-butyl-MeO-BIPHEP chiral auxiliary (1) are reported. Higher ee’s are observed with
the 3,5-dialkyl substituents than with the unsubstituted parent MeO-BIPHEP. It is proposed that the observed
dialkyl “meta-effect”, on enantioselectivity, is the combined result of a more rigid and slightly larger chiral pocket
and that this effect will have some generality in homogeneous catalysis. Detailed NMR studies on the allyl complex
[Pd(PhCHCHCHPh)(1)]PF6 (5), and the model hydrogenation catalyst [RuH(cymene)(1)]BF4 (6), reveal restricted
rotation about several of the P-C(ipso) bonds of the phosphorus substituents containing the 3,5-di-tert-butyl groups.
The X-ray structure of 6 reveals that the cymene ligand is not symmetrically bound to the Ru atom. This observation
is interpreted as an expression of the chiral pocket of 1. MM3* calculations on 6 support the NMR findings and
reproduce the X-ray results.
Introduction
terms of the relative ee, the 3,5-di-tert-butyl analog is at least
7% better than the parent unsubstituted derivative.3
Enantioselective homogeneous catalysis with late transition
metals is enjoying increasing application in organic synthesis,
with hydrogenation, allylic alkylation, and Heck chemistry
among the most successful reactions.1 Although the necessary
difference in activation energy, ∆∆Gq, to achieve an enantio-
meric excess (ee) of >90% is modest, <3 kcal/mol, an
increasing number of chiral auxiliaries are capable of ac-
complishing this task in a variety of transformations.2
The Ru(II) chemistry of 1 has proven to be both fruitful and
novel. Starting from Ru(OAc)2(1), one can isolate a stable
“stripped-down” cationic five-coordinate hydrido-bissolvento
complex, [RuH(i-PrOH)2(1)]BF4 (2), from the hydrogenation
reaction mixture.4 This is the first well-characterized chiral five-
coordinate bis(phosphine)ruthenium hydride complex stable as
a solvento complex. Compound 2 is also an effective hydro-
genation precursor. Moreover, ligand 1 is capable of serving
as a six-electron donor5 in that one of the biaryl double bonds
can coordinate to Ru(II) as shown in the η5-C8H11-pentadienyl
complex 3.
We have recently become involved with the bis(3,5-di-tert-
butyl)-substituted chiral bidentate phosphine MeO-BIPHEP 1,
originally prepared by Schmid and co-workers.3 Ru complexes
of 1 find practical application in that they afford >95% ee’s in
the enantioselective hydrogenation of a prochiral pyrone.3 In
† ETH Zentrum.
‡ University of Milan.
§ Universita¨t Zu¨rich.
X Abstract published in AdVance ACS Abstracts, June 1, 1997.
(1) Trost, B. M. ; van Vranken, D. L. Chem. ReV. 1996, 96, 395. Genet,
J. P.; Acros. Org. Acta 1995, 1, 4. Tanner, D. Angew. Chem., Int. Ed. Engl.
1994, 106, 625. Brown, J. M. Chem. Soc. ReV. 1993, 25. Kitamura, M.;
Noyori, R. In Modern Synthetic Methods; Scheffold, R., Ed.; Springer-
Verlag: New York, 1989; 5, p 116.
(2) Togni. A.; Venanzi, L. M. Angew. Chem. 1994, 33, 497. Hayashi,
T. In Catalytic Asymmetry Synthesis; Ojima, I., Ed.; VCH Publishers, Inc.:
New York, 1993; p 325.
We report here (a) new catalytic results using 1 in both the
Pd-catalyzed enantioselective Heck and allylic alkylation reac-
tions and (b) new solid-state and solution structural chemistry
for Ru and Pd complexes based on 1. Using the results from
these studies, together with new MM3* calculations, allows a
(4) Currao, A.; Feiken, N.; Macchioni, A.; Nesper, R.; Pregosin P. S.;
Trabesinger, G. HelV. Chim. Acta 1996, 79, 1587.
(5) Feiken, N.; Pregosin, P. S.; Trabesinger, G. Organometallics 1997,
16, 537-543.
(3) Schmid, R.; Broger, E. A.; Cereghetti, M.; Crameri, Y.; Foricher, J.;
Lalonde M.; Mueller, R. K.; Scalone, M.; Schoettel, G.; Zutter, U. Pure
Appl. Chem. 1996, 68, 131.
S0002-7863(96)04406-X CCC: $14.00 © 1997 American Chemical Society

6316 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Trabesinger et al.
rationalization of the higher ee’s observed with 1 and with a
number of different chiral bis(3,5-dialkylaryl)phosphine ligands.
of Pd(II).14a-17 For JOSIPHOS, CHIRAPHOS, BINAP, and
BIPHEP type complexes one can define the shape of the phenyl
chiral array, using NOESY methods. The steric interactions
between these phenyls and, e.g., a coordinated 1,3-diphenylallyl
substrate, are pin-pointed via intramolecular NOEs. As a
consequence of these short contacts this structural π-allyl
chemistry is accompanied by specific η3-η1-η3-isomerization
processes.16 Given the detail available from these NMR studies
we have undertaken a similar set of measurements on 5.
The 31P NMR spectrum of the diphenylallyl complex 5
reveals two components, 5a and 5b, in the ratio ca. 1.7:1, which,
Results
Catalytic Results. The enantioselective Heck reaction,6,8
involving dihydrofuran and phenyl triflate, described by Ha-
yashi,6 proceeds slowly with 1 as auxiliary, but in very high
ee, as shown in eq 1. An analogous experiment with the
unsubstituted parent MeO-BIPHEP (i.e, no meta-substituent)
afforded 4a in 84% ee (71% chemical yield). As in the Ru
chemistry, noted above, the 3,5-di-tert-butyl-MeO-BIPHEP
analog is superior in terms of enantioselectivity. For the same
reaction, Pfaltz and co-workers8 have achieved the same ee using
an oxazoline-based chiral auxiliary. Interestingly, Hayashi finds
that 4b and not 4a is the major product using (R)-BINAP as
auxiliary.
We find both the chemical yields, and to a lesser extent the
ee’s for this Heck chemistry, to be sensitive to the nature of
the base and solvent. The observed ee using 1 plus Pd(OAc)2
with Ag2CO3 in DMF was 90%; however, the chemical yield
of 4a was only 2.1%. Using THF as solvent, with NEt(i-Pr)2
as base, at 70 °C for 7 d, also gives ca. 98% ee, but again low
yields (34% and 3% for 4a and 4b, respectively). Use of the
standard Pd(0) reagent Pd2(dibenzylidene acetone)3‚C6H6 at 70
°C for 7 d in benzene affords yields of ca. 44% and 2% for 4a
and 4b, respectively, with the ee’s remaining ca. 98%. Hayashi
and co-workers6 have previously reported a base dependency.6
The enantioselective allylic alkylation of the classical 1,3-
diphenyl acetate substrate13 (eq 2) was carried out using the
with the help of 1H and 13C NMR measurements, can be
assigned to syn/syn and syn/anti isomers. Obviously, the
observed ee for the allylation in eq 2 does not correspond to
the measured ground-state populations of the two diastereomers.
The same can be said for the ee and populations involving the
allylic alkylation chemistry of the parent MeO-BIPHEP
auxiliary.14a
The room temperature 1H spectrum for 5 shows that several
of the aromatic and tert-butyl proton signals are broad. A series
of variable temperature and 2-D NMR experiments lead to the
assignment of the pertinent resonances (see Table 1 for details)
and reveals restricted rotation for the R, â, and γ-phenyl rings
of 5a, at room temperature (see Figure 1a). At 193 K all four
PhCH(OAc)CHdCHPh + CHR2- f
PhC*H(CHR2)CHdCHPh + OAc- (R ) CO2Me) (2)
1,3-diphenylallyl complex [Pd(PhCHCHCHPh)(1)]PF6 (5) as
catalyst precursor. We find a 70% yield of product with an ee
of ca. 91%. The analogous control experiment with the parent
MeO-BIPHEP afforded a ca. 87% ee14 so that again 1 is
superior, if only slightly.
NMR Spectroscopy of 5. We have recently determined the
3-D solution structures for a series of chiral π-allyl complexes
rings are frozen, and Figure 1b shows the 16 nonequivalent tert-
butyl signals (8 each for 5a and 5b). The 20 nonequivalent
ortho-ring protons are readily detected via a 2-D 31P,1H-
correlation (see Figure 2). The variable temperature studies
reveal that 5b has two rings frozen at ambient temperature (the
four unlabeled sharp singlets in the top half of the figure). There
is no observable exchange between 5a and 5b at ambient
temperature in CD2Cl2. In uncoordinated 1 there is relatively
fast rotation around these P-C(ipso) bonds at ambient temper-
ature.
(6) Ozawa, F.; Kubo. A.; Matsumoto, Y.; Hayashi, T.; Nishioka, I.;
Yanagi, K.; Moriguchi, K Organometallics 1993, 12, 4188. In this paper,
4b and not 4a is the major product.
(7) Cabri, W.; Candiani, I. Acc. Chem. Res. 1995, 28, 2. Brown, J. M.;
Perez-Torrente, J.; Alcock, N.; Clase, H. J. Organometallics 1995, 14, 207.
(8) Pfaltz, A. Acta Chem. Scand. 1996, 50, 189.
The 13C chemical shifts of the allyl termini in 5a (102.2 and
83.8 ppm) and 5b (98.1 and 90.7 ppm) fall in the expected
range.14a,15,16 However, the intramolecular differences, ∆13C,
(9) Reiser, O. Angew. Chem. 1993, 105, 576.
(10) von Matt. P.; Lloyd-Jones, G. C.; Minidis, A. B. E.; Pfaltz, A.;
Macko, L.; Neuburger, M.; Zehnder, M.; Ru¨egger, H, Pregosin, P. S. HelV.
Chim. Acta 1995, 78, 265.
(11) Rieck, H.; Helmchen, G. Angew. Chem. 1995, 107, 2881. Knu¨hl,
G.; Sennhenn, P.; Helmchen, G. J. Chem. Soc., Chem. Commun. 1995, 1845.
Knu¨hl, G.; Sennhenn, P.; Helmchen, G. J. Chem. Soc., Chem. Commun. .
1995, 1845.
(12) (a) Mackenzie, P. B.; Whelan, J.; Bosnich, B. J. Am. Chem. Soc.
1985, 107, 2046. (b) Auburn, P. R.; Mackenzie, P. B.; Bosnich, B. J. Am.
Chem. Soc. 1985, 107, 2033. (c) Andersson, P. G.; Harden, A.; Tanner, D.;
Norrby, P. O. Chem. Eur. J. 1995, 1, 12.
(15) (a) Barbaro, P.; Currao, A.; Herrmann, J.; Nesper, R.; Pregosin, P.
S.; Salzmann. Organometallics 1996, 15, 1879. (b) Albinati, A.; Pregosin,
P. S.; Wick, K. Organometallics 1996, 15, 2419. (c) Herrmann, J.; Pregosin,
P. S.; Salzmann, R.; Albinati, A. Organometallics 1995, 14, 3311. (d)
Pregosin, P. S.; Ru¨egger, H.; Salzmann, R.; Albinati, A.; Lianza, F.; Kunz,
R. W. Organometallics 1994, 13, 83. (e) Pregosin, P. S.; Ru¨egger, H.;
Salzmann, R.; Albinati, A.; Lianza, F.; Kunz, R. W. Organometallics 1994,
13, 5040.
(16) Pregosin, P. S.; Salzmann, R. Coord. Chem. ReV. 1996, 155, 35.
(17) (a) Togni, A.; Burckhardt, U.; Gramlich, V.; Pregosin, P.; Salzmann,
R. J. Am. Chem. Soc. 1996, 118, 1031. (b) Pregosin, P. S.; Salzmann, R.;
Togni, A. Organometallics 1995, 14, 842. (c) Breutel, C.; Pregosin, P. S.;
Salzmann, R.; Togni, A. J. Am. Chem. Soc. 1994, 116, 4067.
(13) Trost, B. M. Acc. Chem. Res. 1980, 13, 385.
(14) (a) Barbaro, P. Pregosin, P. S.; Salzmann, R.; Albinati, A.; Kunz,
R., W. Organometallics 1995, 14, 5160. (b) Bolm, C.; Kaufmann, D.;
Gessler, S.; Harms, K. J. Organomet. Chem. 1995, 502, 47.

EnantioselectiVe Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6317
Table 1. 1H-, 13C-, and 31P-NMR Dataa for 5a and 5b
5a (syn/syn)
5b (syn/anti)
1H
13C
1H
13C
3
4
5
7
3′
4′
5′
7′
6.22
6.79
6.07
3.17
6.36
7.17
7.03
3.23
7.38
6.44
1.43
0.98
7.33
7.48
6.93
1.33
1.20
7.53
6.72
6.55
1.06
1.09
7.26
6.41
7.76
1.06
1.09
7.61
5.85
6.32
3.93
114.0
6.25
6.88
6.53
3.15
6.37
7.06
7.17
3.19
7.47
6.45
1.38
0.97
7.52
6.70
6.90
1.06
1.26
7.33
6.92
6.38
1.16
123.9
57.1
56.8
57.1
57.0
o-R
o-R
t-R
t-R
p-R
o-â
o-â
t-â
t-â
p-â
o-γ
o-γ
t-γ
t-γ
p-γ
o-δ
o-δ
t-δ
t-δ
p-δ
8
7.13
7.79
1.15
1.45
7.52
1
Figure 1. (a) tert-Butyl region of the H-spectrum for 5 at ambient
102.1
109.6
83.8
temperature (top). (b) Spectrum at 193 K (bottom). The low-temperature
spectrum reveals eight nonequivalent tert-butyl resonances in both
isomers due to hindered rotation around the P-C(ipso) axes. The
assignment can be made via a series of variable temperature measure-
ments together with NOE results (CD2Cl2, 500 MHz).
9
10
PA
PB
29.7
25.7
75.5
24.5
20.9
71.0
2J(PA-PB)
a Chemical shifts in parts per million, J values in hertz; 193 K,
CD2Cl2. b o ) phosphine ortho protons, t ) t-Bu methyl groups, p )
phosphine para protons. See Chart 1.
X-ray Structure of [RuH(p-cymene)(1)]BF4 (6). Given the
effectiveness of the ruthenium hydrido complex 1 in the
enantioselective hydrogenation,3 we prepared the model cymene
complex 6 as shown in eq 3. Given the dearth of structures for
between the allyl termini in 5a (14.4 ppm) and 5b (7.4 ppm)
are suggestive. Assuming (a) ∆13C is due to steric effects (and
specifically, the D,F-ring-stacking indicated below,
2HBF4
Ru(OAc)2(1) + cymene
8
H2 (1 atm), CH2Cl2
[RuH(cymene)(1)]BF4 (6) + 2 HOAc + HBF4 (3)
chiral hydridoruthenium compounds of chelating phosphines,
the solid-state structure of complex 6 was determined by X-ray
diffraction methods.
and described previously14a) and (b) that the highest frequency
carbon, shown by the arrow, will be attacked preferentially as
it is not as strongly bound and thus more electrophilic, then
one can successfully predict the chirality of the product. This
rationale requires that the two isomers be in equilibrium; i.e.,
5a reacts preferentially and then 5b is drawn off via the
equilibrium. This is reasonable as the observed kinetics are
quite slow.12a,b
Unfortunately, substantial signal overlap together with de-
veloping slow rotation around the allyl C(allyl)-C(phenyl)
carbon-carbon bonds hindered a quantitative analysis of the
rotational barriers; however, the spectroscopy was somewhat
simpler for the arene Ru complex 6.
An ORTEP plot for the cation is given in Figure 3. The
immediate coordination sphere consists of the six π-complexed
arene carbons and the two P atoms of the chiral bidentate MeO-
BIPHEP. There is a THF solvent molecule in the unit cell.
The hydride was located by minimizing the ligand repulsion
(see the Experimental Section). The space in which the hydride
sits is relatively large, and the associated bond lengths and bond
angles fall in the expected ranges. As in the structure of 2, we
note that coordinated 1 takes up a substantial amount of space
around the metal. Table 2 contains a list of selected bond
lengths and bond angles for 6 and Table 3, experimental
parameters.

6318 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Trabesinger et al.
Table 2. Selected Bond Lengths (Å) and Bond Angles (deg) for 6
Ru-P2
2.291(2)
2.298(7)
2.333(7)
2.260(7)
1.66(9)
Ru-P1
Ru- C9
Ru-C11
Ru-C13
2.326(2)
2.306(7)
2.313(7)
2.251(7)
Ru-C8
Ru-C10
Ru-C12
Ru-Hyd
C8-C9
1.442(11)
1.527(11)
1.402(10)
1.421(10)
C8-C13
C9-C10
C11-C12
C11-C17
1.390(11)
1.402(10)
1.420(10)
1.532(10)
C8-C14
C10-C11
C12-C13
P1-Ru-P2
P2-Ru-C10
P2-Ru-C8
P2-Ru-C12
P1-Ru-C10
P1-Ru-C8
P1-Ru-C12
P1-Ru-Hyd
91.07(6)
116.3(2)
162.8(2)
99.3(2)
112.6(2)
104.4(2)
169.5(2)
86(3)
P2-Ru-C11
P2-Ru-C9
P2-Ru-C13
P1-Ru-C11
P1-Ru-C9
P1-Ru-C13
P2-Ru-Hyd
94.9(2)
150.5(2)
127.3(2)
144.7(2)
95.9(2)
133.9(2)
82(3)
Table 3. Experimental Data for the X-ray Diffraction Study of
[RuH(p-Cymene)(1)]BF4‚C4H8O (6)
formula
mol wt
C84H118BF4O3P2Ru
1425.70
crystal dim, mm
data collectn T, K
cryst syst
space group
a, Å
0.50 × 0.40 × 0.30
180(1)
orthorombic
P212121
15.856(3)
20.643(2)
24.396(3)
7985(2)
4
Figure 2. 31P,1H-correlation spectrum of 5a and 5b showing the
aromatic region. Five correlation peaks are visible for each 31P-spin
(20 for the two diastereomers): one each due to the biaryl protons, 5
and 5′, and four from the nonequivalent orthoprotons. Once these are
assigned, NOE results can be used for detailed structural work (CD2-
Cl2, 500 MHz).
b, Å
c, Å
V, Å3
Z
F(calcd), g cm-3
1.186
2.834
µ, cm-1
radiation
Mo KR (graphite monochromated,
λ ) 0.710 69 Å)
no. of indep data collected
no. of obsd reflns (no)
(|Fo| > 3.0σ(|F|))
transmission coeff.
no. of params refined (nv)
Ra
7605
6193
0.9982-0.9371
650
0.0547
a
Rw
0.0691
GOF
3.092
2
a R ) ∑(|Fo - (1/k)Fc|)/∑|Fo|. Rw ) [∑w(Fo - (1/k)Fc)2/∑w|Fo| ]1/2
,
2
2
where w ) [σ2(Fo)]-1. σ(Fo) ) [σ2(Fo ) + f4(Fo )]1/2/2Fo. f ) 0.03.
1 is a relatively large ligand, there is no unusual lengthening of
the metal-phosphorus bond in 6. The P-Ru-P angle, 91.07-
(6)°, is normal.
On the other hand, the observed Ru-C(cymene) distances
in 6, ca. 2.25-2.33 Å, are all relatively long. These separations
are in agreement with the analogous distances in 7, but in
contrast to the ca. 2.18-2.24 Å range of Ru-C(cymene) values
found in complexes of pyrazolylmethane and pyrazolyl borate,20
among others.21-24 Close inspection of the Ru-C(cymene)
bond lengths in 6 reveals that the distances Ru-C12 and Ru-
C13 are shorter than Ru-C9 and Ru-C10, suggesting a “tilted”
mode of coordination. The aromatic C-C bond lengths within
the arene, ca. 1.39-1.44 Å are as expected.19-28
Figure 3. ORTEP plot for the cation of 6. The shaded ellipsoids show
the coordinated arene (C8-C17), plus the Ru, P1, P2, and hydride
atoms.
The Ru-P distances in 6 are modest in length, 2.291(2) and
2.326(2) Å. Cymene complexes of ruthenium are well-known
catalyst precursors, and the structures of several phosphine-
containing complexes have been determined. Zanetti et al.18
reported structural details for [RuCl(cymene)(JOSIPHOS)](PF6)
(7), which contains a large ferrocene-based chiral bidentate
phosphine. They find longer Ru-P separations of 2.368(2) and
2.390(2) Å. In the cationic mono-phosphine-hydride complex
[RuH(styrene)(cymene)(PPh3)]+ (8) Faller and Chase19 report
the Ru-P distance to be 2.297(7) Å. Consequently, although
(20) Bhambri, S.; Tocher, D. A. Polyhedron 1996, 15, 2763.
(21) Yamamoto, Y.; Sato, R.; Matsuo, F.; Sudoh, C.; Igoshi, T. Inorg.
Chem. 1996, 35, 2329.
(22) Tocher, D. A.; Gould, R. O.; Stephenson, T. A.; Bennett, M. A.;
Ennett, J. P.; Matheson, T. W.; Sawer, L.; Shah, V. K. J. Chem. Soc., Dalton
Trans. 1983, 1571.
(23) Porter, L. C.; Polam, J. R.; Bodige, S. Inorg. Chem. 1995, 34, 998.
(24) ; Bennett, M. A.; Goh, L. Y.; Willis, A. C. J. Am. Chem. Soc. 1996,
118, 4984. Bennett, M. A.; Matheson, T. W.; Robertson, G. B.; Smith, A.
K.; Tucker, P. A. Inorg. Chem. 1981, 20, 2353.
(25) Hashiguchi, S.; Fujii, A.; Takehara, J.; Ikariya, T.; Noyori, R. J.
Am. Chem. Soc. 1995, 117, 7562.
(18) Zanetti, N. C.; Spindler, F.; Spencer, J.; Togni, A.; Rihs, G.
Organometallics 1996, 15, 860.
(19) Faller, J. W. Chase, K. J. Organometallics 1995, 14, 1592.

EnantioselectiVe Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6319
Figure 4. Section of the 2-D NOESY for 6 at 193 K showing that
two of the ortho protons of the δ ring are in slow exchange (CD2Cl2,
500 MHz).
One of the most interesting features of the structure concerns
the ca. 45° twist made by the arene C11-C17 vector, relative
to a symmetric position defined by the metal and the two P
atoms. Obviously, the chiral array is enforcing a steric
preference upon this complexed arene.
Figure 5. 31P,1H-correlation spectrum of 6 showing the aromatic
region. Ten correlation peaks are visible: one each due to the biaryl
protons, 5 and 5′, and eight from the nonequivalent ortho protons. This
is an unusually well-resolved example and also demonstrates the
four nonequivalent para protons as singlets (193 K, CD2Cl2, 500
MHz).
Chart 1
Solution Structure for RuH(p-cymene)(1) (6). We have
1
previously reported5 a set of variable temperature H-spectra
demonstrating that 6 is dynamic on the NMR time scale.
Analysis of these spectra reveals the same type of restricted
rotation about the P-C(ipso) bonds, noted above in 5. All of
the P-aryl rotations are restricted at 193 K; however, the δ-ring
is rotating just fast enough so that exchange cross-peaks between
its two ortho protons are visible (see Figure 4). The eight
nonequivalent ortho protons of 1 (plus the o-binaphthyl proton)
can be located and eventually assigned starting with a 31P,1H-
correlation, as shown in Figure 5. The four relatively sharp
singlets at ca. 7.35 ppm, in Figure 5, represent the four
nonequivalent para protons of the rings. At ambient temperature
only one ring, â, is rotating slowly (see Chart 1) in contrast to
5. The activation barriers (kJ/mol), determined by line shape
analyses of the tert-butyl signals, for rotation about the three
relatively unhindered P-C bonds are 50.0 (γ), 42.4 (R), and
36.2 (δ) with the two smallest values associated with the two
pseudo-equatorial substituents (see the Experimental Section for
details). In the absence of 2-(and/or 6) substitution, P-phenyl
substituents in complexes of bidentate chiral phosphines nor-
mally rotate freely.29 Therefore, our data for 5 and 6 represent
the first report of such restricted rotation, and indicate a well-
defined, relatively rigid chiral pocket.
The remaining proton signals can be assigned by classical
2-D measurements, and details of these can be found in Table
4. Figure 6 shows that there is a selective NOE from the Ru-H
resonance to one of the two cymene aromatic protons ortho to
the methyl group (and also a strong NOE to one ortho to the
i-Pr). This confirms that the Me-C vector of the cymene is
rotated relative to the P-Ru-P plane, as found in the solid-
state structure. The rotation brings one of these two protons
(26) Mashima, K.; Kusano, K.; Ohta, T.; Noyori, R. Takaya, H. J. Chem.
Soc., Chem. Commun. 1989, 1208.
(27) Bank, J.; Gevert, O.; Wolfsberger, W.; Werner, H. Organometallics
1995, 14, 4972.
(28) Bhambri, S.; Tocher, D. A. J. Organomet. Chem. 1996, 507, 291.
(29) There are indications that Binap-allyl complexes show the same
restricted rotation, although the activation energy is somewhat lower; i.e.,
the rotation is not yet slow at ambient temperature. See: Ru¨egger, H.; Kunz,
R. W.; Ammann, C. J.; Pregosin, P. S. Magn. Reson. Chem. 1992, 29, 197.

6320 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Trabesinger et al.
Table 4. NMR Data for 6a
1H (13C)
3
5
3′
5′
6.33
7.61
6.23
7.18
7.90
1.48
7.35
7.95
1.24
7.31
7.61
1.37
7.31
7.95
1.45
7.39
4
7
4′
7′
o-R
t-R
7.08
3.12
7.09
2.97
6.05
0.91
o-R
t-R
p-R
o-â
t-â
p-â
o-γ
t-γ
p-γ
o-δ
t-δ
p-δ
9
o-â
t-â
6.61
0.96
o-γ
t-γ
6.82
1.09
o-δ
t-δ
6.71
1.04
3.35 (107.3)
2.76 (94.7)
1.56
0.71
-9.80
10
13
15
17
6.06 (92.7)
5.98 (89.4)
1.07
12
14
Figure 7. Superimposed calculated (MM3*) and X-ray experimental
results for the backbone structure of 6. The tert-butyl groups have been
omitted for clarity. The two structures are seen to be closely related.
16
1.61
hydrideb
a 193 K, CD2Cl2, o ) ortho, t ) t-Bu methyl group, p ) para. See
with the observed solution NOEs, given the uncertainty in the
hydride position. The qualitative solution structure as deter-
mined by NOE methods is in agreement with the structure in
the solid state.
2
Chart 1. b PA ) 52.6, PB ) 54.0, J(PA-PB) ) 46 Hz.
The coordinated cymene of 6 reveals several interesting
chemical shifts: (a) The 13C resonances for the two cymene
carbons ortho to the i-Pr group are separated by 17.8 ppm (107.2
and 89.4 ppm), with the former at unusually high frequency,
on the basis of literature expectations.24-28 We assume that
this difference stems from the asymmetric bonding of the arene
to the metal, as suggested by the X-ray data. The two carbons
ortho to the methyl group are only 1.7 ppm apart. (b) The four
aromatic cymene protons are quite different, with two of these
at very low frequency, as shown:
The protons at 2.76 and 3.35 ppm are shielded as a consequence
of the local anisotropic effects of the equatorial phenyl substi-
tutents30 situated on the P atoms.
MM3* Calculations. Given the interesting structural features
of 6, we carried out MM3* calculations on three cationic
complexes: [RuH(1)(cymene)]+ (6), the unsubstituted MeO-
BIPHEP parent analog, and the 3,5-dimethyl derivative, i.e.,
for 1 with R ) H, CH3, and C(CH3)3.
Figure 6. Section of the ROESY spectrum showing the hydride
contacts to the aromatic cymene protons. The moderate H12 contact
together with the strong H13 contact proves that the cymene is rotated
as described in the text. Note that there are no contacts to H9 and H10
(CD2Cl2, 193 K, 500 MHz, carrier frequency set on -5.0 ppm).
much closer to the hydride, i.e.,
The immediate coordination sphere X-ray data for 6 were used
as input parameters. Figure 7 shows part of the calculated
structure of 6 superimposed on the solid-state result. Clearly
the calculated structure is in good agreement (R ) 0.065 Å)
with the experimentally observed result, suggesting, as in
previous studies, that the method is producing realistic results.
From the X-ray data one finds separations of 2.6(1) Å for Ru-H
to H13 and 3.2(1) Å for Ru-H to H12, in reasonable agreement
(30) Pretsch E.; Seibl. J.; Simon, W. Strukturaufkla¨rung organischer
Verbindungen Springer-Verlag: Berlin, 1986.

EnantioselectiVe Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6321
Table 5. Approximate Rotational Barriers (kJ/mol) about the
P-C(ipso) Phenyl Substituents As Calculated by MM3* a
ring
H
Me
t-Bu (6)
R
â
δ
γ
30
50
30
50
35
50
30
55
45
110
40
125
a â and γ are pseudoaxial and R and δ pseudoequatorial rings. The
three complexes are [RuH(1)(cymene)]+ (6), the unsubstituted MeO-
BIPHEP parent analog, and the 3,5-dimethyl derivative; i.e., for 1 R
) H, CH3, and C(CH3)3.
Figure 8. Superimposed calculated (MM3*) results for the BIPHEP
backbone structures of the H-, Me-, and t-BuRu cations. The meta
substitutents have been omitted for clarity. The three structures are seen
to be closely related.
between pseudoaxial and pseudoequatorial rings will decrease
on going from R ) tert-butyl to R ) Me.
The unexpectedly high barriers to rotation in 6 arise from
the inability of the axial rings to rotate past the biaryl moiety.
Since there is free rotation in uncoordinated 1 at ambient
temperature, complexation to the transition metal compresses
1 somewhat. It is important to note that the closest contacts
from the chiral auxiliary to the substrate (cymene in 6, but 1,3-
diphenylallyl in 5) stem from the pseudoequatorial rings. For
these rings, the calculations suggest a small but noticeable
increase of ca. 5-10 kJ/mol in the rotational barrier, on going
from R ) H to R ) t-Bu. Therefore, the entire chiral array
becomes increasingly rigid, relative to the parent complex, with
increasing size of the meta-substituents.
Figure 8 shows the calculated BIPHEP backbone structures
for the three cations, again superimposed upon one another, to
highlight potential differences. Inspection of this figure, plus
analysis of various angular data, reveals that the structures of
these three complexes are closely related; i.e., introduction of
the meta substitutents does not change the basic structure.
The effect of arene rotation on structure was considered.
Unexpectedly, the calculations suggest four relatively low energy
possibilities for each of the three cations. For 6, two of these
have almost identical energies and might well be populated,
whereas two are found somewhat higher in energy (ca. 16 and
27 kJ/mol above the minimum). If one defines an angle between
the Ru-H and Me-C(ipso) vectors as φ, then the two lowest
Discussion
From the catalytic results in both the Heck reaction and the
allylic alkylation, it is clear that the 3,5-di-tert-butyl substituents
improve the enantioselectivity. A similar observation has been
made by Schmid and co-workers3 in the enantioselective
hydrogenation of a prochiral pyrone. RajanBabu and co-
workers31 have recently reported that 3,5-dimethyl substitution
can improve the ee’s in the Rh-catalyzed hydrogenation of
dehydroamino acids using the chelating ligands shown below.
energy solutions for 6 have φ ) ca. -120° (this is the lowest
energy structure and corresponds to the X-ray and NMR result)
and + 120° (ca. 1 kJ higher). The remaining structures have φ
) -30° and +150°, with the latter at highest energy. Conse-
quently, the observed positioning of the arene may not be
general and moreover need not have much significance. Indeed,
we have observed a second isomer in solution, at the ca. 5%
level at low temperature, which might well be the second low-
energy rotational isomer. The calculations also suggest that the
two arene protons, H13 and H12, should be separated by ca.
2.3 and 3.0 Å, respectively, from the hydride, in good agreement
with the NOE measurement shown in Figure 6.
For four different substrates, they found that the observed ee
was 5%, 6%,12%, and in one case 35% better than with the
parent chelate. Recently, 3,5-dimethyl substitution has been
shown to afford a good ee in the Ir-catalyzed hydrogenation
of prochiral imines using the chiral ferrocene phosphine
9.32 Empirically, it seems that there is a “3,5-dialkyl meta-
For the dimethyl and parent complexes, the picture is
qualitatively the same. The lowest energy rotamer always has
φ ) -120°, and the highest, with φ ) +150°, is at least ca. 17
kJ/mol above the minimum energy; however, the relative
energies of the remaining isomers vary.
The calculated rotational barriers about the four different
P-phenyl substituents, for the three cations, are shown in Table
5. These are not (and are not expected to be) in agreement
with the solution data. Solvent, ion-pairing, and counterion
effects are likely to be considerable, so we take these numbers
to have relative significance. Nevertheless, Table 5 shows, for
6, that there are two groups of two, with the pseudoaxial rings
having the highest barriers, in rough agreement with our solution
observations. However, assigning the highest barrier to the
γ-ring contradicts the solution observation for 6. Further, the
calculations predict that the difference in rotational barriers
effect” on the observed enantioselectivity. This effect can be
sizeable, ca. 14%, as in the Heck reaction noted above; however,
it is more often modest in magnitude, perhaps only ca. 5-10%.
In terms of energy, the observed ∆ee’s correspond to only ca.
2-4 kJ/mol in ∆∆Gq.
(31) RajanBabu, T. V.; Ayers, T. A.; Casalnuovo, A. L. J. Am. Chem.
Soc. 1994, 116, 4101.
(32) Cited in Togni, A. Chimia 1996, 50, 87. Also cited in Chem. Eng.
News 1996, July, 22, 38.

6322 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Trabesinger et al.
The standard deviations on intensities were calculated in terms of
statistics alone, while those on Fo were calculated as shown in Table
3.
The Pd complex 5 and the Ru complex 6 show restricted
rotation about several P-C bonds. Moreover, there are indica-
tions that a similar restricted rotation is observable in an Ir(III)
complex of 9.33 Our MM3* calculations on the 3,5-di-tert-
butyl-MeO-BIPHEP complex 6 suggest that the restricted
rotations are to be expected and that the entire chiral array is
increasingly rigid, relative to the parent complex. Freezing one
or more rings results in a more selective chiral pocket, in
agreement with our earlier calculations.14a It is only necessary
for the ∆∆Gq to “remember” a fraction of the calculated energy
differences to rationalize the observed improved ee’s, relative
to the parent complex.
On the basis of these various results with the Pd, Ru, Rh,
and Ir complexes, we suggest that the observed dialkyl meta-
effect on enantioselectivity is the combined result of a more
rigid and slightly larger chiral pocket. This effect will have
some generality in enantioselective homogeneous catalysis and
may be useful in chiral ligand design.
It is worth remembering that the chiral pocket for any given
auxiliary, be it a MeO-BIPHEP, a ferrocene type,17,18 or a
sugar-based diarylphosphinite,31,34 will interact differently with
coordinated ligands of varying size. Consequently, the extent
of this 3,5-dialkyl meta-effect is expected to be substrate
dependent. Further, enantioselective catalysis in which elec-
tronic effects combine with steric effects, e.g., hydrocyanation,34
could well lead to an insignificant dialkyl meta-effect, so that
each reaction will require individual consideration.
The structure was solved by a combination of direct and Fourier
methods and refined by full-matrix least squares. During the refine-
ment, a Fourier difference map revealed the presence of a clathrated
solvent molecule (THF), which was included in the refinement.
Anisotropic displacement parameters were used for the heavy atoms,
the cymene moiety, the ring substituents, and the solvent molecule; all
other atoms were refined isotropically. Increasing the number of
parameters, while giving an unfavorable observation to parameters ratio,
did not yield a significantly better model.37 As can be judged from
the large displacement parameters of some atoms, two of the tert-butyl
groups are disordered even though no model for it could be constructed.
It proved impossible to locate the hydride ligand on the difference
Fourier maps; however, its position was unambiguously located by using
the energy minimization program HYDEX38 and successfully refined
using an isotropic temperature factor. The function minimized was
[∑w(|Fo| - 1/k|Fc|)2]) with w ) [σ2(Fo)]-1. No extinction correction
was deemed necessary. The scattering factors used, corrected for the
real and imaginary parts of the anomalous dispersion, were taken from
the literature.39 The handedness of the structure was tested by refining
both enantiomorphs; the coordinates giving the significantly37 lower
Rw factor were used. Upon convergence the final Fourier difference
map showed no significant peaks. All calculations were carried out
by using the Enraf-Nonius MOLEN crystallographic programs.35
Calculations. The MacroModel system as described by Still and
co-workers40 was used for the calculations on the MeO-BIPHEP
complexes. The force field used was MM3*. Missing constants were
added and optimized by hand to fit the experimental structure. The
values used may be found as part of the Supporting Information together
with one sample structure in MacroModel format. Monte Carlo
techniques were used for the conformational analysis.
Experimental Section
Barriers to the rotation of the phenyl rings were carried out with the
angle driver technique supplied by MacroModel. All coordinates,
except the driven dihedral angle, were optimized at each step.
Dynamics Calculations. The activation energies for the solution
dynamics were determined using the method of Bain et al.41 The
calculated values of 50.0 ( 0.9 (γ), 42.4 ( 2.7 (R), and 36.2 ( 2.0 (δ)
are associated with correlation coefficients of 0.994, 0.997, and 0.987,
respectively. These activation energies were obtained with difficulty
due to the complexity of the proton spectra, i.e., eight different
environments, with similar chemical shifts, often overlapping.
Preparation of [Pd(η3-PhCHCHCHPh)(1)]PF6 (5). [Pd(η3-Ph-
CHCHCHPh)(µ-Cl)]2 (33.5 mg, 50 µmol) was dissolved in 7 mL of
acetone by gentle warming. TlPF6 (35 mg, 0.1 mmol) was added
General Procedures. All reactions were performed in an atmo-
sphere of Ar using standard Schlenk techniques. Dry and oxygen free
solvents were used. Ru(OAc)2(1a) (2) was provided by F. Hoffman-
La Roche, Basel. Routine 1H (300.13 MHz) and 31P (121.5 MHz) NMR
spectra were recorded with a Bruker DPX-300 spectrometer. Chemical
shifts are given in parts per million, and coupling constants (J) are
given in hertz. The two-dimensional studies were carried out at 500
MHz for 1H. NOESY measurements were carried out as reported
previously.14a,15 The 31P,1H-correlation experiments were carried out
at low temperature. The ROESY spectrum for 6 was measured twice
at 193 K (500 MHz, CD2Cl2), with a 0.4 s spinlock. The carrier
frequency was set in the middle of the proton spectrum in the first
measurement and at ca. -5 ppm in the second. IR spectra were
recorded with a Perkin-Elmer 882 infrared spectrophotometer. El-
emental analyses and mass spectroscopic studies were performed at
ETHZ.
Crystallography. A suitable crystal, mounted on a glass fiber, was
cooled to 180 K by using an Enraf-Nonius FR558SH nitrogen gas-
stream cryostat installed on a CAD4 diffractometer, which was used
for the space group determination and for the data collection. Unit
cell dimensions were obtained by least squares fit of the 2θ values of
25 high-order reflections (9.55 e θ e 18.09°). Selected crystal-
lographic and other relevant data are listed in Table 3 and Supporting
Information S1.
and the precipitated TlCl filtered through Celite after 10 min.
A
solution of (S)-1 (104 mg, 1 mmol) in CH2Cl2 was added. Evaporation
to dryness and recrystallization of the residue from CH2Cl2/pentane
gave the product as a yellow crystalline solid. Yield: 114 mg, 73
µmol, 73% of 5‚CH2Cl2. Anal. Calcd for C85H109O2F6P3Pd‚CH2Cl2
(1561.1): C, 66.2; H, 7.17. Found: C, 66.5; H, 7.05. FAB-MS: m/e
1329.4 (M+ - PF6), 1136.3 [Pd(1)].
Allylic Alkylation. The procedure applied for the allylic alkylation
was the same as described by Pfaltz et al.8,10 Complex 5 was used as
the catalyst precursor. The product was purified by column chroma-
tography (silicagel F60, hexane-ethyl acetate, 65:35) The yield of the
(+)-(R)-product was 70%. Anal. Calcd for C20H20O4 (324.4): C,
74.06; H, 6.21. Found: C, 74.26; H, 6.07. EI-MS: m/e 324.2 (M+).
1H-NMR (CDCl3): δ 3.53 (s, 3H), 3.71 (s, 3H), 3.96 (d, J ) 10.9 Hz,
1H), 4.27 (dd, J ) 8.4, 10.9 Hz, 1H), 6.33 (dd, J ) 8.4, 15.8 Hz, 1H)
6.49 (d, J ) 15.8 Hz, 1H), 7.20-7.35 (m, 10Har). HPLC: OD-H
(Chiralcel, Daicel), UV-detector 254 nm, hexane-i-PrOH, 99:1, flow
0.5 mL/min; 90.5% ee, (+)-(R)-product; tR ) 24.5 min, tS ) 26.9 min.
Data were measured with variable scan speed to ensure constant
statistical precision on the collected intensities. Three standard
reflections were used to check the stability of the crystal and of the
experimental conditions and measured every 90 min; no significant
variation was detected. Data were corrected for Lorentz and polariza-
tion factors using the data reduction programs of the MOLEN
crystallographic package.35 An empirical absorption correction was
also applied (azimuthal (Ψ) scans of two reflections having ø > 88°).36
(37) Hamilton , W. C. Acta Crystallogr. 1965, 17, 502.
(38) Orpen, A. G. J. Chem. Soc., Dalton Trans. 1980, 2509.
(39) International Tables for X-ray Crystallography; Kynoch: Birming-
ham, England, 1974; Vol. IV.
(33) H. Ru¨egger, ETH Zentrum, Zu¨rich, unpublished results.
(34) RajanBabu, T. V.; Casalnuovo, A. L. J. Am. Chem. Soc. 1996, 118,
6325.
(35) MOLEN, Enraf-Nonius Structure Determination Package; Enraf-
Nonius: Delft, The Netherlands, 1990.
(36) North, A. C. T.; Phillips, D. C.; Mathews, F. S. Acta Crystallogr.,
Sect. A 1968, 24, 351.
(40) F. Mohamadi, F.; Richards, N. J. F.; Guida, W. C.; Liskamp, R.
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440.
(41) Bain, A. D.; Duns G. J. J. Magn. Reson. 1995, 112A, 258.

EnantioselectiVe Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6323
Heck Reaction (Arylation of 2,3-Dihydrofuran with Phenyl
Iodide or Phenyl Triflate). Procedure a. PdCl2(1) (18.1 mg, 15
µmol, 3 mol %) and Ag2CO3 (138 mg, 0.5 mmol, 2 equiv) were
suspended in 5 mL of DMF in an ampule with a magnetic stirring bar.
C6H5I (102 mg, 0.5 mmol) and 2,3-dihydrofuran (105 mg, 1.5 mmol)
were added and the ampule was sealed. The mixture was stirred over
7 d at 50 °C. The reaction mixture was added to 250 mL of pentane
with stirring and the resulting suspension filtered. The solution was
washed with water, dried over MgSO4, and concentrated in vacuo. The
crude products were separated and purified via column chromatography
(silica gel F60, hexane-ethyl acetate, 98:2). (+)-(S)-2-Phenyl-2,3-
dihydrofuran: Rf ) 0.4, 4a. (-)-(S)-2-Phenyl-2,5-dihydrofuran: Rf
) 0.15, 4b. The NMR data for 4a and 4b were consistent with those
of the the literature. Results for the catalytic runs are given in the
table below.
a 8.2 M solution in water, 0.11 mmol) was followed by 1.2 equiv (10
µL, 0.066 mmol) of p-cymene. The reaction mixture was then heated
until refluxing started. After cooling to room temperature, the argon
atmosphere was replaced by an atmosphere of H2 (ca. 1.2 atm). The
solution was heated again until it started to reflux. The pale yellow
solution then stood overnight at room temperature under an atmosphere
of H2. After being transferred to a Schlenk tube the reaction mixture
was taken to dryness in vacuo, and the residue washed twice with 3
mL of hexane and then redissolved in 4 mL of CH2Cl2. The resulting
solution was extracted twice with 4 mL of H2O, to remove inorganics,
and the CH2Cl2 dried over MgSO4. After filtration and drying in vacuo,
removal of solvent affords the analytically pure product as a yellow
powder in 81% yield. MS (FAB): m/e (found) 1267.8 (M+). IR
(KBr): 2030 cm-1 (Ru-H; w), 1125-1045 cm-1 (BF4-; s, br).
Selected NMR data: 31P{1H}-NMR (CD2Cl2, 193 K): δ 54.0, 52.6,
2J(P,P) ) 46. 1H-NMR (500 MHz, CD2Cl2): -9.80 (hydride); 7.95,
(3H), 7.61, 6.82, 6.71, 6.61, 6.06 (the eight ortho protons in Figure
2b); 3.12, 2.97 (2 OMe). Anal. Calcd for C, 70.94; H, 8.26. Found:
C, 70.19; H, 8.26.
Alternative Preparation of 6. It is determined that hydrogen gas
is not necessary for the preparation of 6. The solvent is a sufficient
source of the hydride ligand. To a solution of [Ru((R)-1)(OAc)2] (14.0
mg, 0.0112 mmol) in 1 mL of 2-propanol in a NMR tube was added
5.7 equiv (10 µL, 0.064 mmol) of p-cymene followed by the addition
of 2.0 equiv (2.7 µL of a 8.2 M solution in water, 0.022 mmol) of
HBF4‚H2O. After being transferred to a Schlenk vessel, the reaction
mixture was taken to dryness in vacuo. The residue was washed twice
with 2 mL of hexane and redissolved in 4 mL of CH2Cl2. Subsequently
the solution was extracted twice with 2 mL of H2O and dried over
MgSO4. After filtration and drying in vacuo, the product was obtained
as a yellow powder in 78% yield.
Procedure6,8 b1. (S)-(1) (61.9 mg, 60 µmol, 6 mol %) and Pd2-
(dba)3 (14.9 mg, 30 µmol {Pd}, 3 mol %) were solved in 5 mL of
THF, placed in an ampule with a magnetic stirring bar, and stirred at
60 °C for 20 min. The resulting homogeneous orange solution was
cooled to room temperature and then treated with 2,3-dihydrofuran
(377.8 µL, 5 mmol, 5 equiv), phenyl triflate (157 µL, 1 mmol), and
N,N-diisopropylethylamine (510 µL, 3 mmol, 3 equiv). The ampule
was closed and the mixture stirred at 70 °C for 7 d. The reaction
mixture was worked up as described above. A second experiment,
b2, was run with benzene as solvent.
Procedure6 c. (S)-(1) (61.9 mg, 60 µmol, 6 mol %) and Pd(OAc)2
(6.7 mg, 30 µmol, 3 mol %) were placed in an ampule with a magnetic
stirring bar. This mixture was treated with 3 mL of benzene and 2,3-
dihydrofuran (377.8 µL, 5 mmol, 5 equiv), and then stirred for 20 min.
Phenyl triflate (157 µL, 1 mmol) and N,N-diisopropylethylamine (510
µL, 3 mmol, 3 equiv) were added, and the ampule was sealed. The
reaction mixture was stirred at 40 °C for 7 d. The products were
isolated as described above.
The absolute configuration was determined by the optical rotation
of the products in comparison with the literature. Optical yields were
measured by both 1H-NMR (CDCl3) with the use of the optically active
chemical shift reagent tris[3-((heptafluoropropyl)hydroxymethylene)-
camphorato]europium(III) [Eu(hfc)3] and HPLC (OD-H (Chiralcel,
Daicel), UV-detector 258 and 225 nm, hexane-i-PrOH, 99.5:0.5, flow
0.5 mL/min; 2-phenyl-2,3-dihydrofuran; tR ) 11.7 min, tS ) 12.7 min).
Acknowledgment. It is our pleasure to dedicate this paper
to Professor Dieter Seebach on the occasion of his 60th birthday.
P.S.P. thanks the Swiss National Science Foundation, the ETH
Zurich, and F. Hoffmann-La Roche Ltd. for financial support.
A.A. thanks MURST for partial support. We also thank F.
Hoffmann-La Roche Ltd. for a gift of BIPHEP ligands and the
complex Ru(OAc)2(1), as well as Johnson Matthey for the loan
of precious metals. Special thanks go to Dr. Renzo Salzmann
for measuring several of the early NMR spectra on the 1,3-
diphenylallyl complex.
(S)-2-phenyl-2,3-dihydrofuran (S)-2-phenyl-2,5-dihydrofuran
procedure
a
yield (optical purity)
yield (optical purity)
17.4%
13.8%
(73% ee)
34%
(>98% ee)
44%
(>98% ee)
65-70%
(>98% ee)
(22% ee)
3%
(>98% ee)
2%
(>98% ee)
3%
(>98% ee)
Supporting Information Available: Tables of bond lengths
and angles, complete atomic coordinates, anisotropic displace-
ment coefficients, and isotropic displacement coefficients for
hydrogen atoms, ORTEP plot with a full numbering scheme,
and force field parameters added to mm3.fld and sample
structure in MacroModel format for MM3* calculations (23
pages). See any current masthead page for ordering and Internet
access instructions.
b1
b2
c
Preparation of [RuH(p-cymene)(1)]BF4 (6). This complex was
originally prepared starting from Ru(OAc)2(1) (68.4 mg, 0.0547 mmol),
in 1 mL of 2-propanol. Addition of 2.1 equiv of HBF4‚H2O (14 µL of
JA964406G