Ruthenium(II) complexes with chiral tetradentate P2N2 ligands catalyze the asymmetric epoxidation of olefins with H2O2

Article abstract of DOI:10.1021/om000481v

The five-coordinate complexes of the type [RuCl(PNNP)]PF₆ (PNNP = tetradentate ligand with a P₂N₂ donor set) are prepared by chloride abstraction from [RuCl₂(PNNP)]. A mixture of Δ-cis-β- and Λ-cis-β-[RuCl₂ (1a-κ⁴P,N,N,P)] (2a; 1a = N, N′-bis[o-(diphenylphosphino)-benzylidene]-2, 2′-diimino-1,1′-(S)-binaphthylene), prepared by reaction of 1a with [RuCl₂-(PPh₃)₃], reacts with Tl[PF₆], giving the five-coordinate [RuCl (1a-κ⁴P,N,N,P)]PF₆ (3a). The related trans-[RuCl₂(1b-κ⁴P,N,N,P)] (2b; 1b = N,N′-bis[o-(diphenylphosphino) benzylidene]-(1S,2S)-diiminocyclohexane) reacts with Tl[PF₆] to give [RuCl(1b-κ⁴P,N,N,P)]PF₆ (3b). With the amino ligand N,N′-bis[o-(diphenylphosphino) benzylidene]-(1S,2S)-diaminocyclohexane (1c), the aqua complex [RuCl(OH₂)(1c-κ⁴P,N,N,P)]PF₆ (5c) is obtained by reaction of Tl-[PF₆] with [RuCl₂(PPh₃) (1c-κ³P,N,N)] (4), which has been isolated and structurally characterized. The reactivity of the five-coordinate 2b with CO and oxygen donors such as water, Et₂O, THF, and methanol is reported. Both 3 and 5 catalyze the asymmetric epoxidation of olefins with hydrogen peroxide as oxidant. Enantiomeric excesses up to 42% were obtained in the enantioselective epoxidation of styrene and of other unfunctionalized olefins. The reaction is highly stereospecific, as the epoxidation of (Z)-2-methylstyrene gives a cis:trans ratio of 99:1.

Full text of DOI:10.1021/om000481v

Organometallics 2000, 19, 4117-4126
4117
Ru th en iu m (II) Com p lexes w ith Ch ir a l Tetr a d en ta te P 2N2
Liga n d s Ca ta lyze th e Asym m etr ic Ep oxid a tion of Olefin s
w ith H2O2
Robert M. Stoop, Stephan Bachmann, Massimiliano Valentini,† and
Antonio Mezzetti*
Laboratory of Inorganic Chemistry, Swiss Federal Institute of Technology, ETH Zentrum,
CH-8092 Zu¨rich, Switzerland
Received J une 7, 2000
The five-coordinate complexes of the type [RuCl(PNNP)]PF6 (PNNP ) tetradentate ligand
with a P2N2 donor set) are prepared by chloride abstraction from [RuCl2(PNNP)]. A mixture
of ∆-cis-â- and Λ-cis-â-[RuCl2(1a -κ4P,N,N,P)] (2a ; 1a ) N,N′-bis[o-(diphenylphosphino)-
benzylidene]-2,2′-diimino-1,1′-(S)-binaphthylene), prepared by reaction of 1a with [RuCl2-
(PPh3)3], reacts with Tl[PF6], giving the five-coordinate [RuCl(1a -κ4P,N,N,P)]PF6 (3a ). The
related trans-[RuCl2(1b-κ4P,N,N,P)] (2b; 1b ) N,N′-bis[o-(diphenylphosphino)benzylidene]-
(1S,2S)-diiminocyclohexane) reacts with Tl[PF6] to give [RuCl(1b-κ4P,N,N,P)]PF6 (3b). With
the amino ligand N,N′-bis[o-(diphenylphosphino)benzylidene]-(1S,2S)-diaminocyclohexane
(1c), the aqua complex [RuCl(OH2)(1c-κ4P,N,N,P)]PF6 (5c) is obtained by reaction of Tl-
[PF6] with [RuCl2(PPh3)(1c-κ3P,N,N)] (4), which has been isolated and structurally character-
ized. The reactivity of the five-coordinate 2b with CO and oxygen donors such as water,
Et2O, THF, and methanol is reported. Both 3 and 5 catalyze the asymmetric epoxidation of
olefins with hydrogen peroxide as oxidant. Enantiomeric excesses up to 42% were obtained
in the enantioselective epoxidation of styrene and of other unfunctionalized olefins. The
reaction is highly stereospecific, as the epoxidation of (Z)-2-methylstyrene gives a cis:trans
ratio of 99:1.
In tr od u ction
intrinsic danger related with concentrated H2O2 solu-
tions, there is the decomposition reaction catalyzed by
most transition metals and the handling of a biphasic
system.3 Also, water generally deactivates most oxida-
tion catalysts due to strong coordination to the metal
center. Finally, much lower enantioselectivities are
generally obtained with hydrogen peroxide than with
hypochlorite or PhIO.4-7 Thus, there is considerable
interest in developing enantioselective catalytic systems
as an alternative to the manganese-based ones.
Ruthenium complexes containing nitrogen-donor
ligands have been largely investigated in asymmetric
oxidation catalysis. Thus, ruthenium complexes con-
taining porphyrin,8 oxazoline,9 or salen10 chiral ligands
The enantioselective catalytic epoxidation of olefins
is an essential synthetic method.1 Despite the fact that
recent breakthroughs have been achieved in this field,
several problems are still unresolved. These concern, in
particular, the nature of the oxidant and the product
selectivity. Thus, Mn-salen catalysts give excellent
enantioselectivity for tri- and cis-disubstituted olefins,
but partial loss of configuration occurs.1 Also, the
enantioselectivity with mono- or trans-disubstituted
olefins is much lower.1 As for the oxidant, the best
results are obtained with expensive or environmentally
nonbenign oxidants, such as iodosyl benzene and hypo-
chlorite. Clearly, the ideal oxidant is molecular oxygen
or, even better, air, but this goal remains far away.2
After dioxygen, hydrogen peroxide is considered the
most environmentally friendly and economic of oxi-
dants,3 but a number of problems hamper its use in
(asymmetric) oxidation reactions. In addition to the
(3) Strukul, G., Ed. Catalytic Oxidation with Hydrogen Peroxide as
Oxidant; Kluwer: Dordrecht, The Netherlands, 1992.
(4) See, for instance: Katsuki, T. J . Mol. Catal. A: Chem. 1996,
113, 87.
(5) (a) Berkessel, A.; Frauenkron, M.; Schwenkreis, T.; Steinmetz,
A. J . Mol. Catal. A: Chem. 1997, 117, 339. (b) Berkessel, A.;
Frauenkron, M.; Schwenkreis, T.; Steinmetz, A.; Baum, G.; Fenske,
D. J . Mol. Catal. A: Chem. 1996, 113, 321. (c) Schwenkreis, T.;
Berkessel, A. Tetrahedron Lett. 1993, 34, 4785.
(6) Bolm, C.; Kadereit, D.; Valacchi, M. Synlett 1997, 687.
(7) (a) Pietika¨inen, P. Tetrahedron 1998, 54, 4319. (b) Pietika¨inen,
P. Tetrahedron Lett. 1994, 35, 941.
(8) (a) Groves, J . T.; Viski, P. J . Org. Chem. 1990, 55, 3628 and
references therein. (b) O’Malley, S.; Kodadek, T. J . Am. Chem. Soc.
1989, 111, 9116. (c) Mansuy, D.; Battioni, P.; Renaud, J . P.; Guerin,
P. J . Chem. Soc., Chem. Commun. 1985, 155. (d) Zhang, R.; Yu, W. Y.;
Lai, T. S.; Che, C. M. Chem. Commun. 1999, 409. (e) Gross, Z.; Ini, S.
J . Org. Chem. 1997, 62, 5514. (f) Berkessel, A.; Frauenkron, M. J .
Chem. Soc., Perkin Trans. 1 1997, 2265.
* To whom correspondence should be addressed. E-mail: mezzetti@
inorg.chem.ethz.ch.
† Spin-diffusion measurements.
(1) For recent reviews, see: (a) Ito, Y. N.; Katsuki, T. Bull. Chem.
Soc. J pn. 1999, 72, 603. (b) J acobsen, E. N. In Catalytic Asymmetric
Synthesis; Ojima, I., Ed.; VCH: New York, 1993; Chapter 4.2.
(2) (a) The Activation of Dioxygen and Homogeneous Catalytic
Oxidation; Barton, D. H. R., Martell, A. E., Sawyer, D. T., Eds.;
Plenum: New York, 1993. (b) J ames, B. R. Dioxygen Activation and
Homogeneous Catalytic Oxidation; Sima`ndi, L. I., Ed.; Elsevier:
Amsterdam, 1991; p 195. Selected examples: (c) Groves, J . T.; Quinn,
R. J . Am. Chem. Soc. 1985, 107, 5790. (d) Neumann, R.; Dahan, M.
Nature 1997, 388, 353. (e) Neumann, R.; Dahan, M. J . Am. Chem. Soc.
1998, 120, 11969.
(9) (a) End, N.; Pfaltz, A. Chem. Commun. 1998, 589. (b) Nishiyama,
H.; Shimada, T.; Itoh, H.; Sugiyama, H.; Motoyama, Y. Chem. Com-
mun. 1997, 1863.
10.1021/om000481v CCC: $19.00 © 2000 American Chemical Society
Publication on Web 08/25/2000

4118 Organometallics, Vol. 19, No. 20, 2000
Stoop et al.
Sch em e 1
Ch a r t 2
Ch a r t 1
(S)-binaphthylene (1a ) and N,N′-bis[o-(diphenylphos-
phino)benzylidene]-(1S,2S)-diiminocyclohexane (1b) are
best prepared by condensation of Ph2P(o-CHO-C6H4) (2
equiv) with the corresponding enantiomerically pure
diamine. This procedure avoids the racemate resolution
process that has been previously reported for 1b.18 The
amino analogue 1c was prepared according to the
reported procedure, which involves NaBH4 reduction of
1b in ethanol solution.18 All ligands are crystalline
solids and are stable with respect to oxidation and
hydrolysis. The ligands 1b,c have been used in the
asymmetric ruthenium-catalyzed transfer hydrogena-
tion of ketones,19 whereas the ligand (S)-1a is new.20
[Ru Cl(1a -K4P ,N,N,P )]P F 6. The reaction of [RuCl2-
(PPh3)3] with the binaphthyl ligand 1a in boiling
benzene or toluene gives trans-[RuCl2(1a -κ4P,N,N,P)]
(trans-2a ) as the only product, as indicated by the
singlet at δ 46.3 in the 31P NMR spectrum. However,
trans-2a failed to react with Tl[PF6] in CH2Cl2, CHCl3,
or boiling benzene. Thus, we prepared 2a as a mixture
of the trans and cis isomers by reacting [RuCl2(PPh3)3]
and 1a in CH2Cl2 at room temperature. This procedure
affords 2a as a mixture of the ∆-cis-â, Λ-cis-â, and trans
isomers in a 5:4:1 ratio, as determined by 31P NMR
spectroscopy. The minor cis isomer cis-â-2a ′′ (40%),
have been successfully used in asymmetric epoxidation
of unfunctionalized olefins, in particular trans-disub-
stituted ones. However, the use of H2O2 as oxidant
together with ruthenium catalysts is particularly
rare11-13 and is unprecedented in asymmetric catalysis,
whereas phosphine complexes of platinum have been
successfully used in asymmetric (ep)oxidation reactions
with H2O2 as the oxidant.14
Thus, following early reports of the use of phosphine
complexes in catalytic oxidation,15 we investigated the
use of five-coordinate chiral complexes of the type [RuCl-
(P-P)2]+ as epoxidation catalysts (P-P* ) chiral diphos-
phine).16 Recently, we have also found that the cationic
complexes [RuCl(PNNP)]PF6 (where PNNP is a tetra-
dentate chiral ligand containing two imines and two
phosphines) catalyze the asymmetric epoxidation of un-
functionalized olefins with hydrogen peroxide as the
primary oxidant (Scheme 1).17 We report here in full
on the ruthenium coordination chemistry of the PNNP
ligands depicted in Chart 1, as well as on their applica-
tion in the enantioselective epoxidation of a number of
unfunctionalized olefins with aqueous hydrogen perox-
ide as the oxidant.
which exhibits two doublets at δ 49.4 and 41.4 (J PP′
31.9 Hz) was separated from the other two isomers
cis-â-2a ′ (50%, two doublets at δ 52.1 and 43.3, J PP′
)
)
33.5 Hz) and trans-2a (10%) by column chromatography.
Resu lts a n d Discu ssion
To attribute, at least tentatively, the absolute con-
figuration at the metal in 2a -c, we calculated the
relative energies of the diastereomeric complexes [RuCl2-
(PNNP)] by means of molecular modeling (Cerius2)21
(Charts 2-4). For all three ligands 1a -c, the trans
isomer is the least strained one and should be the
thermodynamic product as far as steric factors are
Syn th esis of th e P NNP Liga n d s. The ligands N,N′-
bis[o-(diphenylphosphino)benzylidene]-2,2′-diimino-1,1′-
(10) (a) Takeda, T.; Irie, R.; Shinoda, Y.; Katsuki, T. Synlett 1999,
1157. (b) Kureshi, R. I.; Khan, N. H.; Abdi, S. H. R.; Iyer, P. J . Mol.
Catal. A: Chem. 1997, 124, 91.
(11) Sheldon, R. A.; Barf, G. A. J . Mol. Catal. A: Chem. 1995, 102,
23.
(12) Fisher, J . M.; Fulford, A.; Bennett, P. S. J . Mol. Catal. 1992,
77, 229.
(13) Behr, A.; Eusterwiemann, K. J . Organomet. Chem. 1991, 403,
215.
(14) (a) Baccin, C.; Gusso, A.; Pinna, F.; Strukul, G. Organometallics
1995, 14, 1161. (b) Gusso, A.; Baccin, C.; Pinna, F.; Strukul, G.
Organometallics 1994, 13, 3442. (c) Sinigalia, R.; Michelin, R. A.; Pinna,
F.; Strukul, G. Organometallics 1987, 6, 728.
(15) (a) Bressan, M.; Morvillo, A. Inorg. Chem. 1989, 28, 950. (b)
Morvillo, A.; Bressan, M. J . Mol. Catal. A: Chem. 1997, 125, 119. (c)
Maran, F.; Morvillo, A.; d’Alessandro, N.; Bressan, M. Inorg. Chim.
Acta 1999, 288, 122.
(18) Wong, W. K.; Chik, T. W.; Hui, K. N.; Williams, I.; Feng, X.;
Mak, T. C. W.; Che, C. M. Polyhedron 1996, 15, 4447.
(19) Gao, J . X.; Ikariya, T.; Noyori, R. Organometallics 1996, 15,
1087.
(20) The synthesis of the racemic ligand 1a has been previously
reported: (a) Wong, W. K.; Gao, J . X.; Wong, W. T.; Che, C. M.
Polyhedron 1993, 12, 2063. (b) Wong, W. K.; Gao, J . X.; Wong, W. T.;
Cheng, W. C.; Che, C. M. J . Organomet. Chem. 1994, 277.
(21) Cerius2 is a molecular modeling program based on the universal
force field (UFF): Rappe´, A. K.; Casewit, C. J .; Colwell, K. S.; Goddard,
W. A., III; Skiff, W. M. J . Am. Chem. Soc. 1992, 114, 10024. Rappe´, A.
K.; Colwell, K. S.; Casewit, C. J . Inorg. Chem. 1993, 32, 3438. PDB
files of the minimized structures are available from the authors on
request.
(16) Stoop, R. M.; Bauer, C.; Setz, P.; Wo¨rle, M.; Wong, T. Y. H.;
Mezzetti, A. Organometallics 1999, 18, 5691.
(17) Stoop, R. M.; Mezzetti, A. Green Chem. 1999, 39.

Ru(II) Complex Catalysis of Olefin Epoxidation
Organometallics, Vol. 19, No. 20, 2000 4119
Ch a r t 3
spectrum of 3a shows an AX system with δ 79.8 and
44.5 and a J P,P′ coupling constant of 27.6 Hz, indicating
mutually cis P atoms. The high-frequency resonance at
δ 79.8 is diagnostic of five-coordinate Ru(II) complexes
containing a five-membered P-Ru-N chelate ring.22,23
Addition of 1 equiv of NEt4Cl to 3a in CDCl3 quantita-
tively yields the minor cis-â isomer 2a ′′ as expected on
the basis of microscopic reversibility.
[Ru Cl(1b-K4P ,N,N,P )]P F 6. The ligand 1b reacts with
[RuCl2(PPh3)3] in CH2Cl2 at room temperature, giving
the corresponding six-coordinate dichloro complex [RuCl2-
(1b-κ4P,N,N,P)] (2b). Under these conditions, 2b is
obtained as a 1:4 mixture of the cis-â and the trans
isomer, respectively (Chart 3). The isomer distribution
is based on the 31P NMR spectrum, where the singlet
at δ 47.7 is attributed to the trans isomer, as supported
by the X-ray structure of trans-2b.19 The latter has been
obtained by Noyori as the major product of the reaction
of 1b and [RuCl2(dmso)4] in boiling toluene.19 We
tentatively attribute the configuration at ruthenium in
the cis-â isomer as Λ-cis-â on the basis of molecular
modeling, which indicates that the Λ-cis-â isomer is 1.5
kcal mol-1 more stable than the ∆-cis-â one (Chart 3),
but caution is required in view of the very small energy
difference.
Ch a r t 4
In contrast with 2a , both trans-2b and cis-â-2b react
with Tl[PF6] in dry CH2Cl2, giving the red-brown
monochloro complex [RuCl(1b-κ4P,N,N,P)]PF6 (3b), as
indicated by the 31P NMR spectra of the reaction
between 2b (4:1 trans/cis mixture) and Tl[PF6]. Thus,
the isomeric mixture of 2b was used in the preparation
of 3b as such. Complex 3b tenaciously retains variable
amounts of the solvents used for crystallization, includ-
ing hexane, which results in inaccurate elemental
analyses. However, its formulation is supported by mass
spectrometry (FAB+), showing an [M + 1]+ peak at m/z
796. The cationic nature of 3b is confirmed by its molar
concerned. However, a perusal of the energy data
reveals that electronic factors, which are not taken into
account in the UFF calculation, clearly influence the
relative stability of the isomers. Thus, Cerius2 calcula-
tions suggest that one cis-R isomer of 2a ,b is more stable
than both cis-â isomers, but no cis-R isomer has ever
been observed. Indeed, a trans arrangement of the
phosphines is disfavored due to the competition for the
same π-electrons of ruthenium. Assuming that the
equilibrium composition (i.e., 100% trans isomer) is
reached through intermediates of descending energy,
then the last cis-â isomer observed (either ∆ or Λ)
should be the most stable one. However, in the case of
2a , the nearly equal distribution of the cis-â isomer 2a ′
(50%) and 2a ′′ (40%) precludes even a tentative attribu-
tion of the absolute (∆ or Λ) configuration at the metal.
The mixture of the three isomers of 2a was used as
such in the reaction with Tl[PF6] to yield the red-brown
five-coordinate [RuCl(1a -κ4P,N,N,P)]+ (3a ) as a single
isomer. The insoluble trans isomer does not react and
is filtered off upon crystallization. The isomer mixture
consisting of cis-â-2a ′′ and trans-2a (3:1, obtained by
column chromatography) was treated with Tl[PF6] in
an NMR tube (CDCl3). Monitoring of the reaction by
means of 31P NMR spectroscopy showed that the (major)
isomer cis-â-2a ′ is converted into the (minor) one
cis-â-2a ′′, which reacts in turn to give [RuCl(1a -
κ4P,N,N,P)]+ (3a), whereas trans-2a remains unchanged.
The stable five-coordinate species [RuCl(1a -κ4P,N,N,P)]-
PF6 (3a ) was characterized by elemental analysis, FAB+
conductivity (ΛM ) 40 Ω-1 cm2 mol-1 in 10-3 mol dm-3
in CH2Cl2 solution. Together with the unexceptional 31
)
P
NMR signal observed for the [PF6]- anion, this rules
out the coordination of the anion to ruthenium. Complex
3b is extremely reactive toward water and other oxygen
donors in CH2Cl2 or CDCl3 solution, which restricts the
scope of solvents that can be used without forming six-
coordinate adducts (see below).
In rigorously dry CD2Cl2 or CDCl3 solutions, the
31P NMR spectrum of 3b consists of an AX system (δ
49.7 and 59.0) with a J P,P′ coupling constant of 28.2 Hz.
The presence of inequivalent, mutually cis P atoms is
consistent with the bent arrangement of the tetra-
dentate ligand of the cis-â isomers in Chart 3. How-
ever, the chemical shifts are lower than expected for a
five-coordinate complex. In fact, the related five-
coordinate complex [RuCl(Ph2PCH2CH2NMe2)2]+, whose
mononuclear nature is supported by a X-ray study,
shows a singlet at δ 77.5 in the 31P NMR spectrum.22
Similarly, [RuCl(ppye)2]+ (ppye ) 1-(diphenylphosphino)-
2-(2-pyridyl)ethane) displays a 31P NMR singlet at δ
70.5.23 Moreover, the latter complex dimerizes in solu-
tion and in the solid state, forming the binuclear
dication [(ppye)2Ru(µ-Cl)2Ru(ppye)2]2+, which has been
(22) Shen, J . Y.; Slugovc, C.; Wiede, P.; Mereiter, K.; Schmid, R.;
Kirchner, K. Inorg. Chim. Acta 1998, 268, 69.
(23) Costella, L.; Del Zotto, A.; Mezzetti, A.; Zangrando, E.; Rigo,
P. J . Chem. Soc., Dalton Trans. 1993, 3001.
1
MS, and 31P and H NMR spectroscopy. The 31P NMR

4120 Organometallics, Vol. 19, No. 20, 2000
Stoop et al.
F igu r e 1. Stimulated echo pulse sequence used for diffu-
sion measurements. δ is the length of the gradient pulse,
G its strength, and ∆ the time allowed for diffusion.
characterized by X-ray diffraction.23 Thus, we suspected
that 3b is present, at least in solution, as the doubly
chloro-bridged binuclear species [(1b-κ4P,N,N,P)Ru(µ-
Cl)2Ru(1b-κ4P,N,N,P)]2+, containing two six-coordinate
ruthenium centers. To assess the nuclearity of 3b, we
performed spin-diffusion measurements (see below).
The diffusion coefficients of 3a and [RuCl(CO)(1b-
κ4P,N,N,P)]PF6 (6) were measured by 1H NMR spec-
troscopy in CDCl3 solution. Complex 6, which is known
to be a mononuclear complex, was chosen as reference.
The diffusion coefficient (D) was obtained from pulsed
field gradient spin-echo (PGSE) experiments24 with the
pulse sequence in Figure 1 systematically changing
the strength of the gradient G.25 Plots of the relative
intensity of the NMR signals, ln(I/I0), vs the square of
the gradient amplitude are linear (Figure S1; Support-
ing Information). Their slopes are directly proportional
to the diffusion coefficients D according to eq 1, which
F igu r e 2. ORTEP view of [RuCl2(PPh3)(1c-κ3P,N,N)] (4)
(30% probability ellipsoids).
Ta ble 1. Selected Bon d Dista n ces (Å) a n d An gles
(d eg) for [Ru Cl2(P P h 3)(1c-K4P ,N,N,P )] (4)
Ru-N(1)
Ru-P(1)
Ru-Cl(1)
2.157(9)
2.280(3)
2.462(3)
Ru-N(2)
Ru-P(3)
Ru-Cl(2)
2.231(8)
2.346(3)
2.403(3)
N(1)-Ru-N(2)
N(2)-Ru-P(1)
N(2)-Ru-P(3)
N(1)-Ru-Cl(2)
P(1)-Ru-Cl(2)
N(1)-Ru-Cl(1)
P(1)-Ru-Cl(1)
Cl(2)-Ru-Cl(1)
78.5(3)
N(1)-Ru-P(1)
N(1)-Ru-P(3)
P(1)-Ru-P(3)
N(2)-Ru-Cl(2)
P(3)-Ru-Cl(2)
N(2)-Ru-Cl(1)
P(3)-Ru-Cl(1)
90.6(3)
170.0(3)
98.8(1)
88.9(2)
89.2(1)
88.2(2)
104.2(1)
169.0(2)
92.2(2)
87.0(2)
92.2(1)
79.4(2)
88.3(1)
166.4(1)
I
δ
3
ln
) -(γδ)2G2 ∆ -
D
(1)
C6D6. On the basis of the large energy difference
between ∆-cis-â-2c and Λ-cis-â-2c (25.7 kcal mol-1), the
configuration at ruthenium in cis-â-2c is most probably
∆ (Chart 4).
( )
(
)
I0
was used to fit the experimental data. As D is correlated
to the radii of the molecules,26 and the difference for
the slopes is less than 1%, 3a and 6 have almost the
same hydrodynamic radii; i.e., both 6 and 3a are
mononuclear species.
The PPh3 adduct [RuCl2(PPh3)(1c-κ3P,N,N)] (4) was
isolated from the reaction mixture by column chroma-
tography and fully characterized, including an X-ray
investigation. The 31P NMR spectrum of 4 shows two
doublets for the coordinated P atom of the PNNP ligand
and for PPh3 at δ 46.5 and 38.8. The dangling P atom
of the PNNP ligand appears as a singlet at δ -14.7,
close to the value of the free ligand (δ -15.9). Single
crystals of 4 were obtained by slow diffusion of CH2Cl2
into PriOH. The results of the X-ray study are sum-
marized in Figure 2 and Table 1. Along with the
complex, the unit cell contains one molecule of CH2Cl2
as solvent of crystallization. The complex has a distorted-
octahedral coordination with trans chlorides and a
meridional arrangement of the coordinated P,N,N moi-
ety. The Ru-Cl, Ru-P, and Ru-N distances are in the
range expected for similar complexes of general formula
[RuCl2N2P2].19,22,27 Interestingly, N(1) is more tightly
bound to ruthenium than N(2) (2.157(9) vs 2.231(8) Å).
This is also reflected by the Ru-P distances, as the PPh3
ligand is less tightly bound than the PPh2 group of the
PNNP ligand (2.346(3) vs 2.280(3) Å). There are two
possible reasons for the relatively long Ru-N(2) and
Ru-P(3) distances as compared to Ru-N(1) and Ru-
P(1), namely the weaker (or absent) chelate effect or the
steric interaction between PPh3 and the -o-C6H4-PPh2
group. Both effects also explain the lability of the PPh3
Rea ction of [Ru Cl2(P P h 3)n (1c)] (n ) 0, 1) w ith Tl-
[P F 6]. As the reaction of trans-[RuCl2(1c-κ4P,N,N,P)]
(2c) with Tl[PF6] gave no identifiable products, we
attempted to prepare cis-2c as described above. In
contrast with 1a ,b, the room-temperature reaction of
[RuCl2(PPh3)3] with 1c in CH2Cl2 yielded a mixture of
different products rather than of diastereomers. In the
typical reaction outcome, the products formed are the
reaction intermediate [RuCl2(PPh3)(1c-κ3P,N,N)] (4)
(60%) (see below), the already reported19 trans-2c (10%),
and a further species which we tentatively formulate
as cis-â-2c (30%). Accordingly, the FAB+ mass spectrum
of the product mixture of 2c and 4 shows the fragmen-
tation pattern expected for [RuCl2(1c-κ4P,N,N,P)] along
with the molecular peak of 4. As a further support for
the formulation of cis-â-2c, trans-2c is obtained quan-
titatively by refluxing the mixture of trans-2c and cis-
â-2c (separated by column chromatography from 4) in
(24) Stilbs, P. Prog. Nucl. Magn. Reson. 1987, 19, 1 and references
therein.
(25) Examples of this method: (a) Valentini, M.; Pregosin, P. S.;
Ru¨egger, H. Organometallics 2000, 19, 2551. (b) Pichota, A.; Pregosin,
P. S.; Valentini, M.; Wo¨rle, M.; Seebach, D. Angew. Chem., Int. Ed.
Engl. 2000, 39, 153. (c) J iang, Q.; Ru¨egger, H.; Venanzi, L. M. Inorg.
Chim. Acta 1999, 290, 64.
(26) This relationship is expressed from the Stokes-Einstein
equation: D ) kT/6πηrH, where η is the viscosity of the solution and
rH the hydrodynamic radius.
(27) Song, J . H.; Cho, D. J .; J eon, S. J .; Kim, Y. H.; Kim, T. J .; J eong,
J . H. Inorg. Chem. 1999, 38, 893.

Ru(II) Complex Catalysis of Olefin Epoxidation
Organometallics, Vol. 19, No. 20, 2000 4121
ligand. The dangling P atom is 5.58 Å away from
ruthenium. The closing of the Cl(1)-Ru-Cl(2) angle
(166.4(1)°) reflects the steric crowding and chelate ring
strain in the complex, as the chloro ligands are pushed
away from the bulky PPh3 toward the amino moiety
trans to it. The bite angles for the P,N and N,N chelate
rings are 90.6(3) and 78.5(3)°, respectively. The cyclo-
hexyl ring has a regular chair conformation.
[Ru Cl(L)(P NNP )]P F 6. We investigated the reactiv-
ity of the 16-electron complex 3b toward small molecules
such as CO, H2O, and other N- or O-donors (L) to
give the octahedral adducts [RuCl(L)(PNNP)]PF6 (see
Experimental Section).
(30%). The isomer distribution is independent of the
excess of water used in the synthesis and does not
change upon heating the mixture of 5b′ and 5b′′ for
several hours (CDCl3 at reflux). Complex 5b does not
dissociate the aqua ligand upon dissolution in dry
solvents (CH2Cl2 or CHCl3), even in the presence of
molecular sieves. The coordinated water molecule is
displaced by reaction of 5b with HPF6 in Et2O, which
quantitatively affords the ether adduct [RuCl(OEt2)(1b-
κ4P,N,N,P)]+. The latter can be independently prepared
by addition of Et2O (10 equiv) to 3b.
Finally, the aqua derivative [RuCl(OH2)(1c-
κ4P,N,N,P)]PF6 (5c) is apparently the last product
of chloride abstraction from the dichloro complexes
of ligand 1c with Tl[PF6]. Thus, complex 4 reacts
with Tl[PF6], giving 5c and 1 equiv of PPh3 (eq 2). In
The cationic species [RuCl(1b-κ4P,N,N,P)]+ (2b) reacts
instantaneously with CO in CH2Cl2 at room tempera-
ture, giving [RuCl(CO)(1b-κ4P,N,N,P)]PF6 (6). In this
product, the CO ligand occupies the position trans to a
phosphine, as indicated by the 31P and 13C NMR data
(J P,C ) 110.0 Hz, J P′,C ) 15.1 Hz). The ν(CO) stretching
vibration of 6 (2016 cm-1, KBr) is slightly higher than
in similar cationic complexes featuring a trans P-Ru-
[RuCl2(PPh3)(1c-κ3P,N,N)] + Tl[PF6] + H2O f
[RuCl(OH2)(1c-κ4P,N,N,P)]PF6 + PPh3 + TlCl (2)
CO moiety, such as [RuCl(CO)(ppye-P,N)2]+ (1996 cm-1 23
)
addition, the reaction of the mixture of cis-â-2c, trans-
2c, and 4 with Tl[PF6] in CH2Cl2 gives 5c instead of
the putative five-coordinate [RuCl(1c-κ4P,N,N,P)]+. The
characterization of 5c was complicated by its limited
stability in solution, as it decomposed during recrystal-
lization. However, besides the analogy with 5b, the
formulation of 5c as an aqua complex is supported by
the molecular peak at m/z 817 in its FAB+ mass
spectrum and by the fact that it does not react with an
excess of water (in CDCl3). The AX system in the 31P
NMR spectrum indicates that 5c is formed as a single
isomer.
and [RuCl(CO)(Ph2PCH2CH2NMe2)2]+ (1989 cm-1).22
The comparison between 6 and the latter species sug-
gests that the [RuCl(1b-κ4P,N,N,P)]+ fragment is more
electron poor than [RuCl(Ph2PCH2CH2NMe2)2]+ and,
thus, a stronger Lewis acid, as expected from the ligand
1 being a weaker donor than Ph2PCH2CH2NMe. Upon
refluxing in 2-methoxyethanol, 6 isomerizes to the
species 6′ in which the carbonyl (ν(CO) 1981 cm-1, KBr)
is cis to both P atoms (J P,C ) 15.8 and 15.5 Hz). In the
latter isomer, the ligand trans to CO is either an imine
or Cl, but this cannot be assessed by IR data alone.28
We never observed the putative reaction intermediate,
the five-coordinate [RuCl(1c-κ4P,N,N,P)]PF6 (3c), even
when the reaction was carried out in rigorously dried
solvents. As water can be present, at worst, only in
traces under the reaction conditions employed (Schlenk
techniques under argon, solvent distilled over CaH2),
we conclude that “[RuCl(1c-κ4P,N,N,P)]+” (3c) is the
most oxophilic species of the 3a -c series. As 5c can be
conveniently prepared by reacting the product mixture
of 2c and 4 with Tl[PF6], we used this synthetic path
(instead of the reaction with pure 4) in the catalytic
assays described below.
[Ru Cl(OH2)(P NNP )]P F 6. The five-coordinate com-
plexes 3a ,b react with most O-donors (L), such as H2O,
THF, CH3OH, and diethyl ether, to give the correspond-
ing adducts [RuCl(L)(PNNP)]+. As discussed above, the
oxophilicity of the putative amino analogue 3c is so
high that only the aqua complex 5c was ever detected.
Thus, in view of the activation of hydrogen peroxide,
the reactivity of 3a and 3b with water was investigated
in detail. The binaphthyl analogue 3a is the least
reactive of the 3a -c series. In the presence of water (1
equiv), it forms a species that is tentatively formulated
as an aqua complex. The isolation of [RuCl(OH2)(1a -
κ4P,N,N,P)]PF6 was not attempted, as the reaction is
reversible and 3a is reformed upon addition of molecular
sieves to the solution.
Aqua complexes of ruthenium(II) are common in
combination with hard donors, such as in [Ru(OH2)2-
(bipy)2]n+ 29 and in [Ru(OH2)6]2+ 30
In contrast, soft
.
coligands, such as phosphines, disfavor water coordina-
tion at Ru(II), unless intramolecular H-bonding is
possible, as in complexes with coordinated O-containing
anions,31 or when a strong π-acceptor enhances the
Lewis acidity, as in [RuCl2(CO)(OH2)(PEt3)2].32 Aqua
complexes stabilized by the combination of soft P donors
and hard nitrogen ligands have already been reported,
e.g. [Ru(OH2)(PEt3)2(terpy)2]2+ and [Ru(OH2)(dppe)-
(TpiPr)] (terpy ) 2,2′:6′,2′′-terpyridine; TpiPr ) hydrido-
In contrast, the reaction of 3b with an excess of
H2O gives the stable aqua complex [RuCl(OH2)(1b-
κ4P,N,N,P)] (5b), which was isolated in the solid state
and fully characterized. Bands at 3485 and 1602 cm-1
in the IR spectrum are attributed to the O-H stretching
and H-O-H bending, respectively. The molar conduc-
tivity (ΛM ) 39 Ω-1 cm2 mol-1) and the 31P NMR signal
of the uncoordinated [PF6]- anion indicates that the
complex is an electrolyte in CH2Cl2 solution. In the 31P
NMR spectrum, two AX systems featuring cis coupling
constants (J P,P′ ) 31.8 and 26.9 Hz) reveal that two
different cis isomers are present, 5b′ (70%) and 5b′′
(29) Durham, B.; Wilson, S. R.; Hodgson, D. J .; Meyer, T. J . J . Am.
Chem. Soc. 1980, 102, 600.
(30) Bernhard, P.; Bu¨rgi, H. B.; Hauser, J .; Lehmann, H.; Ludi, A.
Inorg. Chem. 1982, 21, 3936.
(31) (a) Mahon, M. F.; Whittlesey, M. K.; Wood, P. T. Organo-
metallics 1999, 18, 4068. (b) Herold, S.; Mezzetti, A.; Venanzi, L. M.;
Albinati, A.; Lianza, F.; Gerfin, T.; Gramlich, V. Inorg. Chim. Acta
1995, 235, 215.
(28) See, for instance: (a) Ziessel, R.; Toupet, L.; Chardon-Noblat,
S.; Deronzier, A.; Matt, D. J . Chem. Soc., Dalton Trans. 1997, 3777.
(b) Abbenhuis, R. A. T. M.; del R´ıo, I.; Bergshoef, M. M.; Boersma, J .;
Veldman, N.; Spek, A. L.; van Koten, G. Inorg. Chem. 1998, 37, 1749.
(32) Sun, Y.; Taylor, N. J .; Carty, A. J . Inorg. Chem. 1993, 32, 4457.

4122 Organometallics, Vol. 19, No. 20, 2000
Stoop et al.
Ta ble 2. Asym m etr ic Ep oxid a tion of Styr en esa
Ch a r t 5
tris(3,5-diisopropylpyrazolyl)borato).33,34 The aqua
complexes [Ru(OH2)L5]n+ are of particular interest, as
their 2e oxidation can give oxo complexes of the type
[RuIV(O)L5], which have been used in a number of
stoichiometric and catalytic oxygen-transfer reactions
with alkanes,35 alkenes,2c,36,37 alcohols,38 phosphines,39
and sulfides.40 The electrochemistry of aqua complexes
of ruthenium(II) has been investigated in great detail
by Meyer,37 Che,41 and Takeuchi.42 Thus, the aqua
complex [RuCl(OH2)(1a -κ4P,N,N,P)]+ (3a ) is possibly
related to the hypothetical ruthenium(IV) oxo species
[RuCl(O)(1a -κ4P,N,N,P)]+, the putative intermediate in
the catalytic epoxidation described below. To this point,
it is worth mentioning that we recently reported a
related oxo complex of osmium(IV), [OsCl(O)(dcpe)2]+
(dcpe ) 1,2-bis(dicyclohexylphosphino)ethane).43
Asym m etr ic Ep oxid a tion . The cationic complexes
3a ,b and 5b were used as catalyst precursors for the
enantioselective epoxidation of the unfunctionalized
olefins in Chart 5 with hydrogen peroxide as oxidant.
An alternative protocol exploits the generation of the
catalyst in situ by reaction of the dichloro complexes
2a -c with Tl[PF6]. In the case of the amino ligand 1c,
this procedure yields the corresponding aqua derivative
5c. In general, the dichloro complexes 2a -c are cata-
lytically inactive, and no reaction occurs either without
the catalysts (3a ,b and 5b,c) or in the presence of the
free ligand.44 The reaction conditions were optimized
using 3b as catalyst and styrene as the substrate. The
reaction products were quantified by gas chromatogra-
phy, and enantiomeric excesses were determined by GC
analysis on a chiral column. In a preliminary com-
munication,17 we have reported the effect of a series of
parameters, such as oxidant, solvent, temperature, and
additives, on the reaction catalyzed by 3b as precatalyst
and styrene as substrate.
selectivity (%)b
epoxide PhCHO confignc (%)
t
conversn
(%)
eed
run cat. substr (h)
1
2
3
4
5
6
7
3b
7
7
8
7
8
7
7
6
2
4
6
4
2
6
35
35
25
39
18
20
18
81
70
60
68
53
12
7
9
9
5
9
6
S
S
nd
S
nd
S
37
42
25
30
12
40
<5
2be
3b
5b
5b
5ce
3a e
17
27
S
a
Reaction conditions: olefin (0.96 mmol), catalyst (9.6 µmol),
CH2Cl2 (5 mL), H2O2 (30%, 9.8 M, 6.68 mmol, 0.7 mL) added
b
in one portion, 22 °C. By GC analysis with decane as internal
standard. c By comparison with an authentic sample (chiral
d
GC analysis). Determined by chiral GC (Supelco R-DEX 120).
e Catalyst prepared in situ (see Experimental Section).
In the optimized procedure, aqueous hydrogen per-
oxide (30%, 7 equiv vs the olefin) is added in one portion
to a CH2Cl2 solution containing styrene, the catalyst
precursor (1 mol %), and decane as internal standard.
Gas is evolved from the biphasic reaction system at the
beginning of the reaction, apparently due to partial
H2O2 decomposition. Gas evolution generally stops after
45 min of reaction time. For most olefins, oxidative
cleavage of the CdC bond also occurs along with epoxide
formation. Benzaldehyde is detected in the reaction
mixture of acyclic olefins.
The epoxidation of styrene (7) in the presence of (S,S)-
3b (1 mol %) and H2O2 (7 equiv vs 7) gives (S)-styrene
oxide with 81% selectivity and 37% ee after 6 h with
35% olefin conversion (Table 2, run 1). The oxidative
cleavage of the CdC double bond is a minor side
reaction, as only 9% of the initial styrene is converted
to benzaldehyde after 6 h. The mass balance indicates
that 10% of styrene forms products other than styrene
oxide or benzaldehyde. The recovery of (atactic) poly-
styrene from the reaction mixture according to the mass
balance shows that the loss of mass is due to polymer-
ization of the olefin. The catalytic system can be also
formed in situ by reacting the dichloro complex 2b and
Tl[PF6] in CH2Cl2 for 12 h. After filtration, the solution
of the catalyst is added to a CH2Cl2 solution of the
olefin and decane. This method gives 35% conversion
after 2 h with 42% ee for styrene oxide, the highest value
in this study (run 2).45 Longer reaction times do not
improve the conversion. In contrast to other systems,
p-chlorostyrene (8) does not perform better than styrene
(run 3).8e
(33) Lawson, H. J .; J anik, T. S.; Churchill, M. R.; Takeuchi, K. J .
Inorg. Chim. Acta 1990 74, 197.
(34) Takahashi, Y.; Hikichi, S.; Akita, M.; Moro-oka, Y. Organo-
metallics 1999, 18, 2571.
(35) Representative papers: (a) Groves, J . T.; Nemo, T. E. J . Am.
Chem. Soc. 1983, 105, 6243. (b) Goldstein, A. S.; Drago, R. S. J . Chem.
Soc., Chem. Commun. 1991, 21. (c) Bailey, A. J .; Griffith, W. P.; Savage,
P. D. J . Chem. Soc., Dalton Trans. 1995, 3537.
(36) Representative papers: (a) Dobson, J . C.; Seok, W. K.; Meyer,
T. J . Inorg. Chem. 1986, 25, 1513. (b) Che, C. M.; Li., C. K.; Tang, W.
T.; Yu, W. Y. J . Chem. Soc., Dalton Trans. 1992, 3153. (c) Goldstein,
A. S.; Beer, R. H.; Drago, R. S. J . Am. Chem. Soc. 1994, 116, 2424.
(37) Stultz, L. K.; Binstead, R. A.; Reynolds, M. S.; Meyer, J . T. J .
Am. Chem. Soc. 1995, 117, 2520 and references therein.
(38) Recent papers: (a) Catalano, V. J .; Heck, R. A.; Immoos, C. E.;
O¨ hman, A.; Hill, M. G. Inorg. Chem. 1998, 37, 2150, and refs 1-7
therein. (b) Lebeau, E. L.; Meyer, T. J . Inorg. Chem. 1999, 38, 2174.
(39) Le Maux P.; Bahri, H.; Simonneaux, G.; Toupet, L. Inorg. Chem.
1995, 34, 4691.
The aqua derivative 5b is generally less efficient than
3b (runs 4 and 5). This is rather surprising, as the use
of aqueous H2O2 implies that the aqua complex 5b must
be present when 3b is used as catalyst precursor.
However, this effect could be related to the large number
of diastereomers that these PNNP complexes can form
(40) Hua, X.; Shang, M.; Lappin, A. G. Inorg. Chem. 1997, 36, 3735.
(41) Che, C. M.; Yam, V. W. W. In Advances in Transition Metal
Coordination Chemistry; Che, C. M., Ed.; J AI Press: London, 1996;
Vol. 1, p 209.
(42) Marmion, M. E.; Takeuchi, K. J . J . Am. Chem. Soc. 1988, 110,
1472.
(43) Barthazy, P.; Wo¨rle, M.; Mezzetti, A. J . Am. Chem. Soc. 1999,
121, 480.
(44) Iminium salts have been found to catalyze the epoxidation of
olefins: Ahmed, G.; Garnett, I.; Goacolou, K.; Wailes, J . S. Tetrahedron
1999, 55, 2341.
(45) To exclude the possibility that traces of TlCl affect the catalytic
results, we have studied the effect of added TlCl. When TlCl is not
filtered off after the reaction of Tl[PF6] with [RuCl2(1b-κ4P,N,N,P)],
the reaction gives results analogous to those of run 1 (25% conversion,
28% selectivity for the epoxide after 6 h). No oxidation occurs in the
presence of TlCl (1 mol %) and without the metal complex.

Ru(II) Complex Catalysis of Olefin Epoxidation
Organometallics, Vol. 19, No. 20, 2000 4123
Ta ble 3. Asym m etr ic Ep oxid a tion of cis-Olefin sa
Ta ble 4. Asym m etr ic Ep oxid a tion of tr a n s-Olefin sa
selectivity (%)
t
conversn
(%)
eeb
selectivity (%)
eeb
run cat. substr (h)
epoxide other confign (%)
t
conversn
(%)
run cat. substr (h)
epoxide RCHO confign (%)
1
2
3
3b
5b
3b
9
9
2
2
2
2
2
1
2
22
6
100
100
100
78
72
70
55
53
50
55
4
c
c
f
f
f
f
f
nd
nd
25d
22e
1
2
3
4
3b
5b
3b
5b
11
11
12
12
2
2
4
4
26
14
45
14
62
52
69
65
17
4
0
1R,2Rc
4
1S,2Rg 41
1S,2Rg 39
1S,2Rg 25
1S,2Rg 29
1S,2Rg <5
1R,2Rc 10
10
10
10
10
10
4h 3b
nd
nd
12
12
5
6
7
5b
5ci
3a
0
a
See Table 2 (footnote a) for reaction and analytical conditions.
15
Determined by chiral GC (Supelco R-DEX 120). c By comparison
with an authentic sample (chiral GC analysis).
b
a
See Table 2 (footnote a) for reaction and analytical conditions.
Determined by chiral GC (Supelco R-DEX 120). c Traces of trans
b
d
with conversion and selectivity comparable to those of
styrene (Table 4, run 1). The aqua complex 5b is less
active, but some enantioselectivity is observed (run 2).
Similarly, trans-stilbene gives moderate conversion and
chemoselectivity, together with low ee values (runs 3
and 4).
epoxide are detected (cis:trans ratio is 99:1). ee of trans isomer
is 38%. e ee of trans isomer is 23%. f Overoxidation products have
g
been detected by GC MS (see text). By the sign of the optical
rotation. Substrate:catalyst ratio is 200. Catalyst prepared in
situ (see Experimental Section).
h
i
(see below). The amino ligand 1c, tested with the in situ
procedure described above, gives much lower epoxide
selectivity (run 6) than the complexes with the imino
ligand 1b. The mass balance suggests that this is due
to olefin polymerization. Finally, the binaphthylene
ligand 1a gives the lowest enantioselectivity and epoxide
selectivity (run 7). Again, the missing mass can be
explained with the competing polymerization of the
olefin.
The epoxidation of (Z)-2-methylstyrene (9) catalyzed
by complex 3b gives the cis-epoxide as the main product
and only traces of trans-epoxide (the cis:trans ratio is
99:1) (Table 3, run 1). The enantioselectivities of the cis
and trans epoxides are 25% and 38% ee, respectively.
The aqua complex 5b gives similar selectivity, but much
lower conversion (run 2). The stereospecificity of the
epoxidation reaction catalyzed by 3b and 5b is much
higher than that observed in Mn-catalyzed epoxidations
of cis-disubstituted olefins.4 Within the assumption that
the catalytic cycle involves an oxo species, this stereo-
specificity suggests concerted oxene transfer to the
double bond.1,6
With 3b as catalyst precursor, the cis-constrained
olefin 1,2-dihydronaphthalene (10) is quantitatively
converted after 2 h, forming the (-)-(1S,2R)-epoxide
with 55% selectivity and 41% ee (Table 3, run 3). Minor
amounts of 1,2-dihydroxy-3,4-dihydronaphthalene and
1-oxo-2,3,4-trihydronaphthalene are detected by GC MS
among the reaction products. Quantitative conversion
of 10 is obtained also with a substrate to catalyst ratio
of 200:1 (run 4). The aqua complex 5b shows similar
activity and chemoselectivity, but the ee is lower (25%,
run 5), and the amino analogue 5c gives both lower ee
(29%) and lower conversion (78%) than 3b (run 6). With
5c, the epoxide selectivity drops during the reaction:
monitoring by GC shows that the epoxidation reaction
stops at 78% conversion of 10 after 1 h with 55% epoxide
selectivity. After 2 h of reaction time, the selectivity for
the epoxide drops to 4%. GC-MS analysis indicates that
the epoxide decomposes to the dialdehyde. Eventually,
no epoxide is observed after 6 h of reaction time.
We tested trans-â-methylstyrene (11) and trans-
stilbene as models of trans-disubstituted olefins. Com-
plex 3b gives nearly racemic epoxide with 11 (4% ee),
Finally, we investigated the fate of the catalyst 3b
upon addition of H2O2 to check the stability of the coor-
dinated phosphine ligands toward oxidation. Aqueous
H2O2 (30%, 1 equiv) was layered over a CDCl3 solution
of 3b, and the reactions occurring upon diffusion into
the organic phase were monitored by 31P and H NMR
1
spectroscopy. After 10 min, the solution contains equal
amounts of starting 3b and of aqua complex 5b (by
integration of the 31P NMR signals). Interestingly, only
the isomer 5b′ is observed. After 90 min, the signals of
3b have disappeared, and the signal of aqua complex
5b′ is present along with those of unidentified products
at δ 36 (m), 31.9 (s), and 30.0 (s). Integration of the 31P
spectrum indicates that 5b′ is about 50% of total. After
48 h, 5b′ has apparently reacted to give an unidentified
species having 31P NMR signals in the δ range 30-40.
On the basis of chemical shifts and multiplicity, the
latter compounds could be metal complexes containing
coordinated phosphine oxides,46 whose nature will be
the subject of future investigations. In conclusion, the
formation of aqua derivative 5b′ is the fastest reaction
in solution and is followed by a reaction that probably
involves oxidation of the phosphine arms.
This result gives some clue regarding the catalytic
reaction. As we generally observe vigorous gas evolution
immediately after addition of H2O2, we believe that the
complex itself (rather than decomposition products
thereof) decomposes H2O2. Decomposition of H2O2 by
transition-metal compounds is a very general reaction.
It can occur with a catalase mechanism if a RudO
species is involved. Alternatively, traces of Ru(III) can
start a Fenton process.3 Also, the observation of one
single isomer of the aqua complex 5b as the kinetic
product of H2O addition to 3b offers an explanation for
the different reactivities observed with 3b and 5b. In
fact, the catalytic experiments were performed with the
isolated aqua complex containing a mixture of diastereo-
mers 5b′ and 5b′′, which might have different catalytic
behavior.
(46) See, for insance: (a) J ia, G.; Ng, W. S.; Chu, H. S.; Wong, W.
T.; Yu, N. T.; Williams, I. D. Organometallics 1999, 18, 3602. (b) Maj,
A. M.; Pietrusiewicz, K. M.; Suisse, I.; Agbossou, F.; Mortreux, A.
Tetrahedron: Asymmetry 1999, 10, 831.

4124 Organometallics, Vol. 19, No. 20, 2000
Stoop et al.
with a microprocessor-controlled gradient unit and a multi-
nuclear probe with an actively shielded Z-gradient coil. The
complexes were dissolved in distilled CDCl3 and measured at
296 K without spinning. The shape for the gradients was
rectangular; their length (δ) was 5 ms, the strength (G) was
varied in the course of the experiments, from 0.56 to 24.08 G
cm-1, every 0.56 G cm-1, and the diffusion time (∆) was equal
to 66 ms.
To the best of our knowledge, this is the first asym-
metric epoxidation catalyzed by a ruthenium complex
that exploits hydrogen peroxide as oxidant.47 With
the exception of 1,2-dihydronaphthalene, the olefin
conversion is moderate (up to 45% for 12), and the
decomposition of hydrogen peroxide occurs parallel to
epoxidation. However, the relatively small excess of
H2O2 used (7 equiv vs olefin) is worthy of note, as
transition-metal complexes efficiently decompose H2O2.3
Most interestingly, the system activity is up to a TOF
of 18 h-1 (Table 2, run 2), which is an order of
magnitude larger than for ruthenium systems contain-
ing N- or N,O-donor ligands.9 The high activity of the
complexes [RuCl(PNNP)]+ explains why the catalytic
epoxidation reaction effectively competes with the de-
composition of the oxidant. A general feature of the
above reactions is the high chemoselectivity, with
selectivity for epoxide formation as high as 81%. This
is better than that observed with other ruthenium-based
catalytic systems, which generally afford substantial
amounts of cleavage products.11,15,16,48
Mass spectra were measured by the MS service of the
Laboratorium fu¨r Organische Chemie (ETH Zu¨rich). A 3-NO-
BA (3-nitrobenzyl alcohol) matrix and a Xe atom beam with a
translational energy of 8 keV were used for FAB+ MS. Optical
rotations were measured using a Perkin-Elmer 341 polarim-
eter with a 1 dm cell. Elemental analyses were carried out by
the Laboratory of Microelemental Analysis (ETH Zu¨rich).
N,N′-Bis[o-(d ip h en ylp h osp h in o)b en zylid en e]-2,2′-d i-
im in o-1,1′-(S)-bin a p h th ylen e (1a ). (S)-(-)-2,2′-Diamino-1,1′-
binaphthalene (257 mg, 0.905 mmol) and 2-(diphenylphosphino)-
benzaldehyde (522 mg, 1.810 mmol) were stirred in toluene
(15 mL) at room temperature for 14 h. The resulting yellow
solution was refluxed in a Dean-Stark apparatus for 2 h.
Evaporation of the solvent under vacuum and recrystallization
of the resulting yellow oil from MeOH/toluene (4:1) gave a
yellow solid. Yield: 700 mg (94%). [R]20D: -126.4 ( 2° (c ) 1).
1H NMR (250 MHz, CDCl3): δ 8.9 (d, 2H, H-CdN, J P,H ) 6
Hz), 7.85 (m, 2H, arom), 7.55-6.8 (m, 36H, arom), 6.6 (m, 2H,
arom). 31P NMR (101 MHz, CDCl3): δ -15.2 (s) (δ -16.5
was reported for (rac)-1c).18 MS (EI): m/z 829 (M+, 98), 751
([M - Ph]+, 40), 540 ([M - NdC-PPh2]+, 100). Other analytic
and spectral data are as in ref 20.
Con clu sion
We have shown that new cationic ruthenium com-
plexes containing tetradentate chiral ligands with a
P2N2 donor set are effective catalysts for the asymmetric
epoxidation of alkenes with hydrogen peroxide. Clearly,
chloride abstraction is necessary to catalytic activity,
as the dichloro species [RuCl2(PNNP)] are catalytically
inactive under the same conditions. In particular, H2O2
is activated with high efficiency, which points to the role
played by the high oxophilicity of the five-coordinate
complexes. Indeed, best results are obtained with 3b,
which forms the stable aqua complex 5b. Either reduced
oxophilicity (as in 3a ) or reduced stability of the aqua
complex (as in 5c) decreases the efficiency of the system.
The stereospecificity of the epoxidation reaction sug-
gests that the intermediates involved have little radical
character.
N,N′-Bis[o-(d ip h en ylp h osp h in o)ben zylid en e]-(1S,2S)-
d iim in ocycloh exa n e ((S,S)-1b ). (1S,2S)-(+)-1,2-Diamino-
cyclohexane (63 mg, 0.54 mmol) and 2-(diphenylphosphino)-
benzaldehyde (317 mg, 1.08 mmol) were dissolved in toluene
(10 mL), and the resulting orange-yellow solution was stirred
for 10 h at room temperature. Then, the reaction solution was
heated to reflux in a Dean-Stark apparatus, after which
toluene was evaporated. The resulting yellow oil was recrys-
tallized from toluene/MeOH to give a white solid. Yield: 320
1
mg (90%). [R]23 ) -74.9 ( 0.2° (CHCl3, c ) 1). H NMR (250
D
MHz, CDCl3): δ 8.7 (d, 2H, NdCH), 7.7 (m, 2H, arom), 7.3-
7.15 (m, 24H, arom), 6.8 (m, 2H, arom), 3.1 (m, 2H, N-CH),
1.2-1.7 (m, 8 H, CH2). 31P NMR (101 MHz, CDCl3): δ -13.9
(s, 2P). IR (KBr, cm-1): 1635 (s, νCN). MS (EI): m/z 658 (M+,
2), 581 ([M - Ph]+, 42), 288 ([M - 2PPh2]+, 100). Anal. Calcd
for C44H40N2P2: C, 80.22; H, 6.12; N, 4.25. Found: C, 80.16;
H, 6.38; N, 4.29.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. Reactions with air- or moisture-
sensitive materials were carried out under an argon atmo-
sphere using Schlenk techniques. Styrene, 1,2-dihydronaph-
thalene, rac-styrene oxide, (R)-styrene oxide, and (1S,2S)-
(+)-1,2-diaminocyclohexane were obtained from Fluka AG.
trans-â-Methylstyrene, (1R,2R)-trans-â-methylstyrene oxide,
and 2-(diphenylphosphino)benzaldehyde were purchased
from Aldrich; Tl[PF6] was obtained from Strem Chemicals.
[RuCl2(PPh3)3], cis-â-methylstyrene, rac-1,2-dihydronaph-
thalene oxide, rac-trans-â-methylstyrene oxide, and rac-cis-â-
methylstyrene oxide were prepared according to literature
procedures (see the Supporting Information). NMR spectra
were recorded on Bruker AVANCE spectrometers. 1H and
13C positive chemical shifts in ppm are downfield from
tetramethylsilane. 31P NMR spectra were referenced against
external 85% H3PO4. The diffusion measurements were
performed on a Bruker AVANCE 400 spectrometer equipped
N,N′-Bis[o-(d ip h en ylp h osp h in o)ben zylid en e]-(1S,2S)-
d ia m in ocycloh exa n e ((S,S)-1c). Ligand 1b (1.00 g, 1.52
mmol) and NaBH4 (608 mg, 16 mmol) were refluxed for 24 h
in ethanol (100 mL). The solution was cooled, and the solvent
was removed under vacuum. The resulting solid was sus-
pended in CH2Cl2 (70 mL) and filtered over Celite. The solvent
was removed under vacuum, and the resulting creamy solid
was recrystallized from CH2Cl2/hexane. Yield: 674 mg (67%).
1H NMR (250 MHz, CDCl3): δ 7.58-7.53 (m, 4 H, arom), 7.34-
7.13 (m, 22 H, arom), 6.87-6.82 (m, 2 H, arom), 4.15 (d, 2 H,
PhCHH′N, J H,H′ ) 13.4 Hz), 3.91 (d, 2 H, PhCHH′N, J H,H′
)
13.4), 2.4 (br, 1 H, N-CH), 2.05 (br, 1 H, N-CH), 1.67 (br,
2H, NH), 1.47-1.04 (m, 8 H, Cy H). 31P NMR (101 MHz,
CDCl3): -15.9 (s, 2P). Other analytic and spectral data are as
in ref 18.
[Ru Cl2(1a -K4P ,N,N,P )] (2a ). Ligand 1a (511 mg, 0.78
mmol) and [RuCl2(PPh3)3] (743.2 mg, 0.78 mmol) were stirred
in CH2Cl2 (10 mL) for 14 h, after which hexane (50 mL) was
added. When the solution was concentrated, red microcrystals
precipitated, which were filtered off, washed with hexane, and
dried under vacuum. Recrystallization was from CH2Cl2/
hexane. Yield: 668 mg (86%). This procedure gives a mixture
of two different cis-â isomers (50 and 40%) and a trans isomer
(10%). 1H NMR (250 MHz, CDCl3): cis-â-2a ′ (50%), δ 8.4 (d, 1
(47) We did not attempt to develop a mechanistic interpretation of
enantioselection, since the (hypothetical) oxo intermediate [RuCl(O)-
(PNNP)] can be formed in seven diastereomers (one trans, four cis-â,
and two cis-R isomers). The analysis of the 14 transition states is a
daunting task, also in view of the low energy differences between them,
as indicated by the moderate to low enantioselectivity.
(48) Augier, C.; Malara, L.; Lazzeri, V.; Waegell, B. Tetrahedron Lett.
1995, 36, 8775.

Ru(II) Complex Catalysis of Olefin Epoxidation
Organometallics, Vol. 19, No. 20, 2000 4125
H, J P,H ) 8.3 Hz, NdCH), 8.3 (s, 1 H, arom), 8.1-6.4 (m, 37
H, arom), 5.63 (d, 1 H, J P,H ) 8.6 Hz, NdCH), 5.24 (m, 2 H,
arom); cis-â-2a ′′ (40%), δ 8.4 (d, 1 H, J P,H ) 8.3 Hz, NdCH);
trans-2a (10%), δ 8.55 (d, 2 H, J P,H ) 6.1 Hz, NdCH). 31P NMR
(101 MHz, CDCl3): cis-â-2a ′, δ 49.4 (d, 1P, J P,P′ ) 31.9 Hz),
yielded pure 4 and a mixture of trans-2c and cis-â-2c. Data
for orange-red 4 are as follows. 31P NMR (250 MHz, CDCl3):
δ 45.8 (d, 1 P, PNNP coord), 38.8 (d, 1 P, PPh3), -14.7 (s, 1 P,
dangling PNNP). MS (FAB+): m/z 1096 (M+, 60), 834 ([M -
PPh3]+, 100), 799 ([M - PPh3 - Cl]+, 27), 764 ([M - PPh3
-
41.4 (d, 1 P, J P,P′ ) 31.9 Hz); cis-â-2a ′′, δ 52.3 (d, 1P, J P,P′
)
2Cl]+, 15). Anal. Calcd for C62H59Cl2N2P3Ru‚CH2Cl2: C, 64.02,
H, 5.20, N, 2.37. Found: C, 63.57, H, 5.41, N, 2.45.
33.5 Hz), 43.5 (d, 1P, J P,P′ ) 33.5 Hz); trans-2a , δ 46.4 (s, 2P).
MS (FAB+): m/z 1001 ([M + 1]+, 10), 1000 (M+, 13), 966
([M + 1 - Cl]+, 75), 965 ([M - Cl]+, 100). Anal. Calcd for
X-r a y An a lysis of [Ru Cl2(P P h 3)(1c-K3P ,N,N)] (4). Red
crystals of 4 were grown from CH2Cl2/PriOH. A needle
(0.60 × 0.06 × 0.02 mm) was mounted on a glass capillary.
Crystal data for C63H61Cl4N2P3Ru: orthorhombic, P212121, cell
dimensions (298 K) a ) 10.3532(2) Å, b ) 17.6566(3) Å, c )
31.6328(3) Å, and V ) 5782.6(2) Å3 with Z ) 4 and Dc ) 1.358
Mg/m3, µ ) 0.580 mm-1 (Mo KR, graphite monochromated),
λ ) 0.710 73 Å, F(000) ) 2440. The data were collected at
room temperature on a Siemens SMART platform in the θ
range 1.29-20.82°. The structure was solved with SHELXTL
using direct methods. Of the 25 499 measured reflections with
index ranges -10 e h e 10, -17 e k e 15, and -31 e l e 31,
and 6061 independent reflections, 4677 reflections with I >
2σ(I) were used in the refinement (full-matrix least squares
on F2 with anisotropic displacement parameters for all non-H
atoms). Hydrogen atoms were introduced at calculated posi-
tions and refined with the riding model and individual isotropic
thermal parameters for each group. Final residuals were R1
) 0.0657 (I > 2σ(I)), R1 ) 0.0993 (all 6061 reflections), wR2
) 0.1404 (all data), GOF ) 1.133. Maximum and minimum
difference peaks were +0.55 and -0.44 e Å-3; largest and mean
∆/σ ) -0.004 and 0.000. Atomic coordinates, anisotropic
displacement coefficients, and an extended list of interatomic
distances and angles are available as Supporting Information.
Selected interatomic distances and angles are reported in
Table 1.
C
58H42N2P2Cl2Ru‚0.5CH2Cl2: C, 67.34; H, 4.15; N, 2.68.
Found: C, 67.53; H, 4.85; N, 2.72. The minor cis isomer cis-
â-2a ′′ was separated from cis-â-2a ′ and trans-2a by column
chromatography over alumina with CH2Cl2/hexane (1:1) as
eluent.
[Ru Cl(1a -K4P ,N,N,P )]P F 6 (3a ). [RuCl2(1a -κ4P,N,N,P)] (373
mg, 0.37 mmol) and Tl[PF6] (131 mg, 0.37 mmol) were stirred
for 16 h in CH2Cl2 (20 mL). After filtration of TlCl over Celite,
pentane (50 mL) was added. Partial evaporation of the solvent
yielded a purple precipitate, which gave red-brown microcrys-
tals after recrystallization from CH2Cl2/pentane. Yield:
282 mg (69%). 1H NMR (300 MHz, CDCl3): δ 8.84 (d, 1 H,
NdC-H, J P,H ) 9.6 Hz), 8.60 (s, 1 H, arom), 8.25 (d, 1 H, arom,
J H,H′ ) 8.6 Hz), 8.08 (m, 2 H, arom), 7.9-6.4 (m, 36 H, arom),
4.94 (d, 1 H, NdC-H, J P,H ) 8.6 Hz). 31P NMR (250 MHz,
CDCl3): δ 79.8 (d, 1 P, J P,P′ ) 27.6 Hz), 44.5 (d, 1 P, J P,P′
)
27.6 Hz). MS (FAB+): m/z 965 (M+, 100), 930 ([M - Cl]+, 10).
Anal. Calcd for C58H42N2F6P3ClRu‚0.25CH2Cl2: C, 61.82, H,
3.79, N, 2.48. Found: C, 61.61, H, 4.07, N, 2.69.
[Ru Cl2(1b-K4P ,N,N,P )] (2b). [RuCl2(PPh3)3] (282 mg, 0.294
mmol) and 1b (194 mg, 0.294 mmol) were dissolved in CH2Cl2
(30 mL), and the resulting brown solution was stirred at room
temperature for 4 h. Addition of hexane and partial evapora-
tion of the solvent gave a bordeaux red solid. Yield: 223 mg
1
(91%). H NMR (250 MHz, CDCl3): cis-â-2b (20%), δ 8.96 (d,
[Ru Cl(CO)(1b-K4P ,N,N,P )]P F 6 (6). Complex 2b (100 mg,
0.104 mmol) and Tl[PF6] (44 mg, 0.125 mmol) were dissolved
in CH2Cl2 (10 mL). The red solution was stirred for 3 h at room
temperature and then filtered over Celite and saturated with
CO gas. Addition of hexane to the resulting orange solution
and partial evaporation of the solvent yielded a yellow
1H, J P,H ) 10.2 Hz, NdCH); trans-2b (80%), δ 8.90 (d, 2H,
J P,H ) 8.7 Hz, NdCH), 7.68-6.51 (m, 28H, arom), 4.18 (d, 2H,
N-CH, J H,H′ ) 8.2 Hz), 2.72 (d, 2H, cyclohexyl-CH2, J H,H′
11.6 Hz), 2.06 (m, 2H, cyclohexyl-CH2), 1.42 (m, 4H, cyclohexyl-
CH2). 31P NMR (101 MHz, CDCl3): cis-â-3, 88.3 (d, 1 P, J P,P′
)
)
1
31.7 Hz), 36.1 (d, 1 P, J P,P′ ) 31.7 Hz); trans-2b, 47.7 (s, 2 P).
MS (FAB+): m/z 830 (M+, 100), 795 ([M - Cl]+, 54), 759
([M - 2Cl]+, 10). Anal. Calcd for C44H40Cl2N2P2Ru: C, 63.62;
H, 4.85; N, 3.37. Found: C, 63.46; H, 4.90; N, 3.13.
precipitate. Yield: 88 mg (87%). H NMR (250 MHz, CDCl3):
δ 8.92 (s, 1H, NdCH), 8.64 (d, 1H, NdCH, J P,H ) 9.4 Hz),
8.24-6.41 (m, 28H, arom), 2.79 (m, 1H, N-CH), 2.52 (m, 1H,
N-CH), 1.94 (m, 2H, CH2), 1.36-1.21 (m, 6H, CH2). 31P NMR
(101 MHz, CDCl3): δ 35.0 (d, 1P, J P,P′ ) 28.8 Hz), 22.6 (d, 1P,
J P,P′ ) 28.2 Hz), -144.4 (septet, 1P, J P,F′ ) 714 Hz, PF6). 13C
NMR (75 MHz, CDCl3): δ 194.8 (dd, 1C, J P,C ) 110.0 Hz, J P′,C
) 15.1 Hz, CO). IR (KBr, cm-1): 2016 (s, νCO), 1640 (m, νCN).
MS (FAB+): m/z 824 (M+, 28), 795 ([M - CO]+, 100). Anal.
Calcd for C45H40N2OF6P3ClRu: C, 55.82; H, 4.16; N, 2.89.
Found: C, 55.56; H, 4.57; N, 2.84.
[Ru Cl(1b-K4P ,N,N,P )]P F 6 (3b). Complex 2b (244 mg, 0.294
mmol) and Tl[PF6] (124 mg, 0.352 mmol) were stirred at
room temperature in CH2Cl2 (30 mL) for 3 h. After filtration
over Celite and addition of hexane, evaporation of CH2Cl2
yielded a red-brown solid. Yield: 204 mg (74%). 1H NMR
(250 MHz, CD2Cl2, over molecular sieves): δ 8.85 (d, 1 H,
J P,H ) 10.0 Hz, NdCH), 8.6 (br s, 1 H, NdCH). 31P NMR (101
MHz, CD2Cl2, over molecular sieves): δ 59.0 (d, 1P, J P,P′ ) 28.2
[Ru Cl(H2O)(1b-K4P ,N,N,P )]P F 6 (5b). Complex 2b (302
mg, 0.363 mmol) and Tl[PF6] (165 mg, 0.471 mmol) were
dissolved in CH2Cl2 (40 mL). The red solution was stirred for
3 h at room temperature and then filtered over Celite into a
PriOH-H2O (1:1) solution. Partial evaporation of the solvent
Hz), 49.7 (d, J P,P′ ) 28.2 Hz, 1 P), -144.4 (septet, 1 P, J PF
)
714 Hz, PF6). ΛM ) 40 Ω-1 cm2 mol-1 (10-3 mol dm-3
CH2Cl2 solution). MS (FAB+): m/z: 796 ([M + H]+, 100), 760
([M - Cl]+, 17).
1
At t em p t ed Syn t h esis of cis-â-[R u Cl2(1c-K4P ,N,N,P )]
(2c). Ligand 1c (971 mg, 1.467 mmol) and [RuCl2(PPh3)3]
(1.467 g, 1.467 mmol) were stirred in CH2Cl2 (70 mL) for 14
h. After it was refluxed for an additional 24 h, the solution
was cooled, and hexane was added. Partial evaporation of the
solvent under vacuum gave a brown solid, which was filtered
off, washed with hexane, and dried under vacuum. Recrystal-
lization from CH2Cl2/hexane yielded a 6:3:1 mixture of 4, cis-
â-2c, and trans-2c. Yield: 942 mg (ca. 65%). 31P NMR (101
MHz, CDCl3): 4 (60%), δ 45.8 (d, 1 P, J P,P′ ) 29.4 Hz), 38.0 (d,
1 P, J P,P′ ) 29.4 Hz); cis-â-2c (30%), δ 54.0 (d, 1 P, J P,P′ ) 28.7
Hz), 37.3 (d, 1 P, J P,P′ ) 28.7 Hz); trans-2c (10%), δ 42.5 (s, 2
P). See below for data of 4. MS (FAB+): m/z 1096 (M+, 9, 4),
834 (M+, 100, 2c), 799 ([M - Cl]+, 11, 2c), 761 ([M - 2Cl -
3H]+, 93, 2c), 263 ([Ph3PH]+, 42, from 4). Column chromatog-
raphy of the reaction mixture (alumina, CH2Cl2/hexane (1:1))
yielded a yellow-orange solid. Yield: 288 mg (83%). H NMR
(250 MHz, CDCl3): isomer 5b′ (70%), δ 8.80 (d, J P,H ) 10.2
Hz, 1 H, NdCH), 8.68 (br s, 1 H, NdCH); isomer 5b′′ (30%), δ
9.21 (d, J P,H ) 9.0 Hz, 1 H, NdCH), 8.88 (d, J P,H ) 9.0 Hz, 1
H, NdCH). 31P NMR (101 MHz, CDCl3): isomer 5b′, δ 65.0
(d, J P,P′ ) 31.8 Hz, 1 P), 45.5 (d, J P,P′ ) 31.8 Hz, 1 P); isomer
5b′′, 42.9 (d, J P,P′ ) 26.9 Hz, 1 P), 50.9 (d, J P,P′ ) 26.9 Hz, 1 P),
-144.4 (septet, J PF ) 714 Hz, 1 P, PF6). IR (CHCl3): νOH
)
3685, 3604, 3485 (br) cm-1; νCN ) 1630, 1602 cm-1. ΛM ) 39
Ω-1 cm2 mol-1 (10-3 mol dm-3 CD2Cl2 solution). MS (FAB+):
m/z 813 (M+, 10), 795 ([M - H2O]+, 100), 759 (M - Cl - H2O]+,
24). Anal. Calcd for C44H42ClF6N2OP3Ru: C, 55.15; H, 4.42;
N, 2.92. Found: C, 55.83; H, 4.94; N, 2.37.
[Ru Cl(H2O)(1c-K4P ,N,N,P )]P F 6 (5c). A mixture of 2c and
4 (227 mg, ca. 0.23 mmol) and Tl[PF6] (94 mg, 0.27 mmol) were
dissolved in CH2Cl2 (10 mL). The brown solution was stirred

4126 Organometallics, Vol. 19, No. 20, 2000
Stoop et al.
for 12 h at room temperature and then filtered over Celite into
a PriOH-H2O (1:1) solution. Partial evaporation of the solvent
yielded a yellow-orange solid. Yield: 125 mg (ca. 67%). The
complex decomposes to unidentified products upon recrystal-
lization from CH2Cl2/hexane. 31P NMR (101 MHz, CDCl3): δ
65.8 (d, J P,P′ ) 31.9 Hz, 1P), 45.8 (d, J P,P′ ) 31.9 Hz, 1 P),
-144.4 (septet, J PF ) 714 Hz, 1 P, PF6). MS (FAB+): m/z 817
(M+, 16), 800 ([M - H2O]+, 79), 763 (M - Cl - H2O - H]+,
100).
in CH2Cl2 (5 mL). The suspension was stirred overnight and
filtered over Celite (to remove TlCl) into a CH2Cl2 solution (5
mL) of the olefin (1.9 mmol) and decane (0.34 mmol, internal
standard). The aqua derivative 5c was prepared in situ from
the 6:3:1 mixture of 4, cis-â-2c, and trans-2c described above
(20 mg, ca. 20 µmol) and Tl[PF6] (7 mg, 20 µmol) by stirring
overnight in CH2Cl2 (5 mL). The resulting solution was filtered
(over Celite, to remove TlCl) into a CH2Cl2 solution (5 mL) of
the olefin (1.92 mmol) and decane (0.34 mmol), and H2O2 (1.4
mL, 3.2 mmol) was added in one portion.
All reactions were monitored by GC for at least 6 h.
Unreacted olefin and products (epoxide and aldehyde) were
quantified by achiral GC (cross-linked silicon column) using
decane as internal standard. The enantiomeric excess of the
epoxide was determined by chiral GC (Alpha-Dex120 Supelco).
Details are given in the Supporting Information.
[Ru Cl(L)(1b-K4P ,N,N,P )]P F6.The [RuCl(L)(1b-κ4P,N,N,P)]-
PF6 adducts with L ) Et2O, THF, MeOH, and CH3CN were
prepared as follows: [RuCl(1b-κ4P,N,N,P)]PF6 (15 mg, 16
mmol) was dissolved under argon in CD2Cl2 (0.5 mL) or CDCl3,
and L (10 equiv) was added with a microsyringe. 31P NMR (101
MHz, CDCl3): Et2O (CD2Cl2), δ 64.0 (d, 1 P, J P,P′ ) 30.1 Hz),
48.2 (d, 1 P, J P,P′ ) 30.1 Hz); THF (CD2Cl2), δ 61.3 (d, 1 P, J P,P′
) 32.5), 43.5 (d, 1 P, J P,P′ ) 32.5); MeOH (CD2Cl2), δ 61.5 (d,
1 P, J P,P′ ) 32.5), 43.5 (d, 1 P, J P,P′ ) 32.5); CH3CN (CDCl3), δ
48.5 (d, 1 P, J P,P′ ) 26.6), 42.2 (d, 1 P, J P,P′ ) 26.6).
Ca ta lytic Ep oxid a tion . In a typical catalytic run, the
olefin (0.96 mmol), decane (0.17 mmol), and the precatalyst
(9.6 µmol, 1 mol %) were dissolved in dry, freshly distilled
CH2Cl2 (5 mL) under argon. Aqueous hydrogen peroxide
(Perhydrol Merck, 30%, 9.8 M solution; 0.7 mL, 6.86 mmol)
was added in one portion to the brown solution with vigorous
stirring. In the alternative in situ procedure, 3a ,b were
prepared from 2a ,b (19 µmol) and Tl[PF6] (6.6 mg, 19 µmol)
Ack n ow led gm en t. S.B. and R.M.S. gratefully ac-
knowledge financial support from the Swiss National
Science Foundation.
Su p p or tin g In for m a tion Ava ila ble: Details of the X-ray
study of 4 and details of catalytic reactions and analytic
procedures. This material is available free of charge via the
Internet at http://pubs.acs.org.
OM000481V