The first complete identification of a diastereomeric catalyst-substrate (alkoxide) species in an enantioselective ketone hydrogenation. Mechanistic investigations

Article abstract of DOI:10.1021/ja0102991

The enantioselective hydrogenations of the dialkyl 3,3-dimethyloxaloacetate ketone substrates (2, 3, and 4; alkyl = Me, ⁱPr, and ^tBu, respectively) were catalyzed by [Ru((R)-BINAP)(H)(MeCN)_n(sol)_3-n]-(BF₄) (1, n = 0-3, sol = THF or MeOH, (R)-BINAP = (R)-2,2′-bis(diphenylphosphino)-1,1′-binaphthyl) in up to 82% ee (R). Reaction of the active catalyst 1 with 1 equiv of substrate (2, 3, or 4) in THF or MeOH solution formed the diastereomeric catalyst-alkoxide complexes [Ru((R)-BINAP)(MeCN)(OCH(CO₂R)-(C(CH₃)₂ CO₂R))](BF₄) (5/6 R = Me, 8/9 R = ⁱPr, and 10 R = ^tBu, respectively) via hydride addition to the ketone carbonyl carbon and ruthenium addition to oxygen. The absolute configurations at the alkoxide groups ((R)- for the major diastereomers 5, 8, and 10) were determined via cleavage of the ruthenium-alkoxide bond with 1 equiv of HBF₄·OEt₂. The solution structures of the major diastereomer catalyst-alkoxide complexes (5, 8, and 10) were unambiguously determined by variable-temperature NMR spectroscopy. The major diastereomers (5, 8, and 10) had the same absolute configuration as the major product enantiomers from the catalytic hydrogenation of 2, 3, and 4 with 1 as catalyst. The ratio of major to minor alkoxide diastereomers was similar to the ee of the catalytic hydrogenation. The catalyst-alkoxide complexes are formed at temperatures as low as -30 °C with no other precursors or intermediates observed by NMR showing that ketone-hydride insertion is likely not the turnover limiting step of the catalytic hydrogenation. Results from the stoichiometric hydrogenolysis of 5/6, 8/9, or 10 indicate that their formation is rapid and only partially reversible prior to the irreversible hydrogenolysis of the ruthenium-oxygen bond. The stereoselectivities of the formation and hydrogenolysis of 5/6, 8/9, and 10 sum up to equal the stereoselectivities of the respective catalytic hydrogenations of 2, 3, and 4. The rates of the hydrogenolysis were consistent with these diastereomers being true catalytic intermediates.

Full text of DOI:10.1021/ja0102991

Published on Web 03/16/2002
The First Complete Identification of a Diastereomeric
Catalyst-Substrate (Alkoxide) Species in an Enantioselective
Ketone Hydrogenation. Mechanistic Investigations
†
Christopher J. A. Daley and Steven H. Bergens*
Contribution from the Department of Chemistry, UniVersity of Alberta, Edmonton,
Alberta T6G 2G2, Canada
Received February 2, 2001
Abstract: The enantioselective hydrogenations of the dialkyl 3,3-dimethyloxaloacetate ketone substrates
i
t
(
2, 3, and 4; alkyl ) Me, Pr, and Bu, respectively) were catalyzed by [Ru((R)-BINAP)(H)(MeCN)
n
(sol)3-n]-
(
BF ) (1, n ) 0-3, sol ) THF or MeOH, (R)-BINAP ) (R)-2,2′-bis(diphenylphosphino)-1,1′-binaphthyl) in
4
up to 82% ee (R). Reaction of the active catalyst 1 with 1 equiv of substrate (2, 3, or 4) in THF or MeOH
solution formed the diastereomeric catalyst-alkoxide complexes [Ru((R)-BINAP)(MeCN)(OCH(CO R)-
) (5/6 R ) Me, 8/9 R ) Pr, and 10 R ) Bu, respectively) via hydride addition to the
2
i
t
3 2 2 4
(C(CH ) CO R))](BF
ketone carbonyl carbon and ruthenium addition to oxygen. The absolute configurations at the alkoxide
groups ((R)- for the major diastereomers 5, 8, and 10) were determined via cleavage of the ruthenium-
alkoxide bond with 1 equiv of HBF
4
‚OEt . The solution structures of the major diastereomer catalyst-
2
alkoxide complexes (5, 8, and 10) were unambiguously determined by variable-temperature NMR
spectroscopy. The major diastereomers (5, 8, and 10) had the same absolute configuration as the major
product enantiomers from the catalytic hydrogenation of 2, 3, and 4 with 1 as catalyst. The ratio of major
to minor alkoxide diastereomers was similar to the ee of the catalytic hydrogenation. The catalyst-alkoxide
complexes are formed at temperatures as low as -30 °C with no other precursors or intermediates observed
by NMR showing that ketone-hydride insertion is likely not the turnover limiting step of the catalytic
hydrogenation. Results from the stoichiometric hydrogenolysis of 5/6, 8/9, or 10 indicate that their formation
is rapid and only partially reversible prior to the irreversible hydrogenolysis of the ruthenium-oxygen bond.
The stereoselectivities of the formation and hydrogenolysis of 5/6, 8/9, and 10 sum up to equal the
stereoselectivities of the respective catalytic hydrogenations of 2, 3, and 4. The rates of the hydrogenolysis
were consistent with these diastereomers being true catalytic intermediates.
Introduction
the many published studies of enantioselective ketone hydro-
genations, we know of no report describing the complete or
unambiguous structural characterization of a diastereomeric
We report the first complete and unambiguous structure
determination of a diastereomeric catalyst-substrate complex
alkoxide) in an enantioselective ketone hydrogenation. The
ketone-catalyst compound, whether it is a putatiVe catalytic
(
1
a,3
intermediate or not.
As a result, the nature of the steric
enantioselective catalytic hydrogenation of ketones is an
1
extensively studied and utilized reaction. Remarkable advances,
(
2) Selected examples of advances in enantioselective ketone hydrogenations.
Ruthenium catalysts: (a) Ohkuma, T.; Ishii, D.; Takeno, H.; Noyori, R. J.
Am. Chem. Soc. 2000, 122, 6510. (b) Mikami, K.; Korenaga, T.; Ohkuma,
T.; Noyori, R. Angew. Chem., Int. Ed. 2000, 39, 3707. (c) Ohkuma, T.;
Koizumi, M.; Ikehira, H.; Yokozawa, T.; Noyori, R. Org. Lett. 2000, 2,
especially by Noyori and co-workers, have recently been
achieved in enantioselective ketone hydrogenations, including
development of quite reactive catalyst systems1a,2 that effect the
659. (d) Noyori, R.; Ohkuma, T. Pure Appl. Chem. 1999, 71, 1493. (e)
1
a,2l
highly enantioselective hydrogenations of keto esters,
of
and even
Despite these advances, and despite
Mikami, K.; Korenaga, T.; Terada, M.; Ohkuma, T.; Pham, T.; Noyori, R.
Angew. Chem., Int. Ed. Engl. 1999, 38, 495. (f) Cao, P.; Zhang, X. J. Org.
Chem. 1999, 64, 2127. (g) Fehring, V.; Selke, R. Angew. Chem., Int. Ed.
Engl. 1998, 37, 1827. (h) Ohkuma, T.; Koizumi, M.; Doucet, H.; Pham,
T.; Kozawa, M.; Murata, K.; Katayama, E.; Yokozawa, T.; Ikariya, T.;
Noyori, R. J. Am. Chem. Soc. 1998, 120, 13529. (i) Ohkuma, T.; Doucet,
H.; Pham, T.; Mikami, K.; Korenaga, T.; Terada, M.; Noyori, R. J. Am.
Chem. Soc. 1998, 120, 1086. (j) Nagel, U.; Roller, C. Z. Naturforsch. 1998,
1a,2h,i,p
1a,2e,f,o-q
olefinic ketones,
of alkyl-aryl ketones,
of alkyl ketones.2
g,j,k,o-q
*
To whom correspondence should be addressed. E-mail: steve.bergens@
ualberta.ca.
†
Current address: Harvard University, 12 Oxford Street, Cambridge,
53b, 267. (k) Doucet, H.; Ohkuma, T.; Murata, K.; Yokozawa, T.; Kozawa,
MA, 02138.
M.; Katayama, E.; England, A. F.; Ikariya, T.; Noyori, R. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 1703. (l) Ager, D. J.; Laneman, S. A. Tetrahe-
dron: Asymmetry 1997, 8, 3327 and references within. (m) Gen eˆ t, J. P.;
Ratovelomanana-Vidal, V.; Ca n˜ o de Andrade, M. C.; Pfister, X.; Guerreiro,
P.; Lenoir, J. Y. Tetrahedron Lett. 1995, 36, 4801. (n) Burk, M. J.; Harper,
T. G. P.; Kalberg, C. S. J. Am. Chem. Soc. 1995, 117, 4423. (o) Ohkuma,
T.; Ooka, H.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 10417.
(p) Ohkuma, T.; Ooka, H.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1995, 117, 2675. Rhodium catalyst: (q) Jiang, Q.; Jiang, Y.;
Xiao, D.; Cao, P.; Zhang, X. Angew. Chem., Int. Ed. Engl. 1998, 37, 1100.
(
1) Examples of use in pharmaceuticals, industry, and academia: for a thorough,
recent review, see: (a) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed.
2
001, 40, 40-73. Other examples include: (b) Chirality in Industry:
Collins, A. N., Sheldrake, G. N., Crosby, J., Eds.; John Wiley & Sons:
New York, 1997. (c) Reductions in Organic Chemistry: Abdel-Magid, A.
F., Ed.; Adv. Chem. Ser. No. 641; American Chemical Society: Wash-
ington, DC, 1996. (d) Noyori, R. Asymmetric Catalysis in Organic Synthesis,
Wiley: New York, 1994. (e) Catalytic Asymmetric Synthesis: Ojima, I.,
Ed.; VCH Publishers: New York, 1993.
3680
9
J. AM. CHEM. SOC. 2002, 124, 3680-3691
10.1021/ja0102991 CCC: $22.00 © 2002 American Chemical Society

Identification of a Catalyst−Substrate (Alkoxide) Species
A R T I C L E S
interactions within this scientifically and economically important
enantioselective catalytic reaction can only be studied by using
indirect experimental evidence.1a,3,4
Prior studies from this laboratory have demonstrated the utility
of our catalyst system [Ru((R)-BINAP)(H)(MeCN)n(sol)3-n]-
the same absolute configuration as the major product enantiomer.
Observation was made at low temperatures of the diastereomeric
olefin-hydride insertion by the olefin-catalyst adduct as a
putative step in a catalytic olefin hydrogenation. The success
obtained from use of 1 to study olefin hydrogenations has
prompted us to extend its use to study the mechanism of
enantioselective ketone hydrogenations.
In this study, we report our mechanistic investigations of the
catalytic hydrogenations of the ketone substrates dimethyl,
diisopropyl, and di-tert-butyl 3,3-dimethyloxaloacetate (2, 3, and
6
c
(BF4) (1, BINAP is 2,2′-bis(diphenylphosphino)-1,1′-binaphthyl,
n ) 0-3, sol ) MeOH, acetone, or THF depending on reaction
5
medium) for the enantioselective hydrogenation of olefins. This
active catalyst is generated in high yields, in pure form, and it
contains both a hydride and labile solvento ligands (eq 1). This
4
) using 1 as catalyst (eq 2). We also report the solution-state
structures of the major diastereomeric ketone-hydride insertion
products between these substrates and 1, the protonolysis of
these alkoxides with HBF4‚OEt2, and evidence for the mech-
anism of these catalytic hydrogenations.
rare combination of features makes this system well-suited to
study the mechanism of enantioselective hydrogenations. For
Results and Discussion:
6
example, use of this system allowed the first isolation and X-ray
structure determination of a diastereomeric catalyst-alkyl
complex (formed by olefin-hydride insertion) of the same
absolute configuration as the major product enantiomer of a
We chose ketones 2, 3, and 4 as substrates because they are
nonenolizable and because they contain a distribution of
functional groups that often results in high enantiomeric excess
(ee) in catalytic hydrogenation: a 1,4-dicarbonyl unit with the
6a
8
catalytic olefin hydrogenation. Use of this system also allowed
prochiral functionality at the 2-position. The catalytic hydro-
7
the first structure determination of a diastereomeric olefin-
genations of 2, 3, and 4 with 1 as catalyst (50 °C, 50 atm of
H2, 50 h) proceeded with complete conversion and with ee’s
ranging from moderate to good in MeOH (2, ee 59% (R); 3, ee
catalyst adduct that contained a hydride ligand and that was of
(
3) Some partial characterization or identifications have been made. Two
examples for rhodium systems: (a) Hatat, C.; Karim, A.; Kokel, N.;
Mortreux, A.; Petit, F. New J. Chem. 1990, 14, 141. (b) Tani, K.; Tanigawa,
E.; Tatsuno, Y.; Otsuka, S. J. Organomet. Chem. 1985, 279, 87. Three
examples in ruthenium systems: (c) King, S. A.; DiMichele, L. In Catalysis
of Organic Reactions; Scaros, M. G., Prunier, M. L., Eds.; Marcel-
Dekker: New York, 1995; Vol. 62, p 157. (d) Mezzetti, A.; Tschumper,
A.; Consiglio, G. J. Chem. Soc., Dalton. Trans. 1995, 49. (e) Geraty, S.
M.; Harkin, P.; Vos, J. G. Inorg. Chim. Acta 1987, 131, 217. Achiral ligand
or substrate complexes: (f) Agbossou, F.; Carpentier, J.-F.; Hatat, C.; Kokel,
N.; Mortreux, A. Organometallics 1995, 14, 2480. (g) Hayashi, Y.; Komiya,
S.; Yamamoto, T.; Yamamoto, A. Chem. Lett. 1984, 1363. (h) Hiraki, K.;
Katayama, R.; Yamaguchi, K.; Honda, S. Inorg. Chim. Acta 1982, 59, 11.
For structures of two catalytic intermediates (each not containing substrate)
in Ru(II)-catalyzed transfer hydrogenations of ketones see: (i) Haack, K.-
J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R. Angew. Chem., Int.
Ed. Engl. 1997, 36, 285.
6
8% (R), and 4, ee 82% (R)) and in THF (2, ee 59% (R); 3, ee
66% (R), and 4, ee 80% (R)) (eq 2). The absolute configuration
of all the major product enantiomers was the same (R), and the
product ee increased as the steric bulk of the ester groups
increased.
(7) For examples of solution-state studies and structure determinations utilized
in the investigation of other possible catalytic intermediates see: (a) van
der Slot, S. C.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Iggo, J. A.;
Heaton, B. T. Organometallics 2001, 20, 430. (b) Hiraki, K.; Ishimoto, T.;
Kawano, H. Bull. Chem. Soc. Jpn. 2000, 73, 2099. (c) Del Rio, I.; Pamies,
O.; van Leeuwan, P. W. N. M.; Claver, C. J. Organomet. Chem. 2000,
608, 115. (d) Uriz, P.; Fernandez, E.; Ruiz, N.; Claver, C. Inorg. Chem.
Commun. 2000, 3, 515. (e) Gridnev, I. D.; Higashi, N.; Asakura, K.;
Imamoto, T. J. Am. Chem. Soc. 2000, 122, 7183. (f) Yi, C. S.; Lee, D. W.
Organometallics 1999, 18, 5152. (g) van Rooy, A.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M. J. Organomet. Chem. 1997, 535, 201. (h)
Giovannetti, J. S.; Kelly, C. M.; Landis, C. R. J. Am. Chem. Soc. 1993,
115, 4040. (i) McCulloch, B.; Halpern, J.; Thompson, M. R.; Landis, C.
R. Organometallics 1990, 9, 1392. (j) Brown, J. M.; Maddox, P. J. Chem.
Commun. 1987, 1276. (k) Landis, C. R.; Halpern, J. J. Am. Chem. Soc.
1987, 109, 1746. (l) Brown, J. M.; Chaloner, P. A.; Morris, G. A. Chem.
Commun. 1983, 664. (m) Brown, J. M.; Murrer, B. A. J. Chem. Soc., Perkin
Trans. 2 1982, 489. (n) Brown, J. M.; Chaloner, P. A. J. Chem. Soc., Perkin
Trans. 2 1982, 711. (o) Brown, J. M.; Chaloner, P. A.; Glaser, R.; Geresh,
S. Tetrahedron 1980, 36, 815. (p) Brown, J. M.; Chaloner, P. A. Chem.
Commun. 1980, 344. (q) Brown, J. M.; Chaloner, P. A. J. Am. Chem. Soc.
1980, 102, 3040. (r) Brown, J. M.; Chaloner, P. A. Chem. Commun. 1979,
613.
(
4) Other mechanistic investigations of rhodium- and ruthenium-catalyzed
enantioselective hydrogenation of ketones. Rhodium: (a) Agbossou, F.;
Carpentier, J.-F.; Mortreux, A.; Surpateanu, G.; Welch, A. J. New J. Chem.
1
996, 20, 1047. (b) Mezzetti, A.; Tschumper, A.; Consiglio, G. J. Chem.
Soc., Dalton Trans. 1995, 49. (c) Pasternak, H.; Pruchnik, F. P. Polish J.
Chem. 1992, 66, 865. (d) Chiba, M.; Takahashi, H.; Morimoto, T.; Achiwa,
K. Tetrahedron Lett. 1987, 28, 3675. (e) Tor o¨ s, S.; Heil, B.; Kollar, L.;
Marko, L. Acta Chim. Hung. 1985, 119, 135. (f) Ojima, I.; Kogure, T. J.
Organomet. Chem. 1980, 195, 239. Ruthenium: (g) Abdur-Rashid, K.;
Lough, A. J.; Morris, R. H. Organometallics 2000, 19, 2655. (h) S a´ nchez-
Delgado, R. A.; Valencia, N.; M a´ rquez-Silva, R.-L.; Andriollo, A.; Medina,
M. Inorg. Chem. 1986, 25, 1106.
(
5) (a) Wiles, J. A.; Lee, C. E.; McDonald, R.; Bergens, S. H. Organometallics
1
996, 15, 3782. (b) Daley, C. J. A.; Wiles, J. A.; Bergens, S. H. Can. J.
Chem. 1998, 76, 1447-1456.
(
6) (a) Wiles, J. A.; Bergens, S. H. J. Am. Chem. Soc. 1997, 119, 2940. (b)
Wiles, J. A.; Bergens, S. H. Organometallics 1998, 17, 2228. (c) Wiles, J.
A.; Bergens, S. H. Organometallics 1999, 18, 3709.
(8) This arrangement is found in commonly utilized R,â-unsaturated acid
substrates and their derivatives (e.g. itaconic acid/esters and R-acetamido-
cinnamic acid/esters) for highly enantioselective catalytic hydrogenations.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002 3681

A R T I C L E S
Daley and Bergens
Scheme 1. Test for Ruthenium-Alkoxide Formationa
We carried out stoichiometric reactions between 2 and the
active catalyst 1 at room temperature in THF and MeOH to
investigate the mechanism of these ketone hydrogenations.
9
These reactions resulted in the immediate loss of hydride signal
1
in the H NMR spectrum with concomitant formation of two
species (5 and 6, in a 79.5:20.5 ratio) of similar NMR
representation, suggesting they were diastereomeric catalyst-
alkoxides (eq 3). The corresponding reaction between the
a
As expected, the 13C APT signal of C1 is inverted with respect to the
reaction 1 + 2 (C1-H forms) versus 1-d1 + 2 (C1-D forms).
diisopropyl ester (3) and the active catalyst 1 was also complete
on mixing at room temperature to form the major and minor
product diastereomers (8 and 9) in a 90:10 ratio. The same
reaction with the bulky di-tert-butyl ester 4 formed exclusively
the product diastereomer 10.
Figure 1. The three possible structures (I, II, and III) of the major
diastereomer formed from reaction of 1 with 2, 3, or 4.
It was imperative to determine the structures of 5, 6, 8, 9,
and 10 to confirm that ketone-hydride insertion occurred, and
to investigate the steric interactions within these possible
catalytic intermediates. Attempts to obtain crystals suitable for
structure determination by X-ray diffraction of 5, 6, or any of
the catalyst-substrate complexes reported in this manuscript
were unsuccessful. NMR and isotope labeling experiments were
used instead to obtain the solution-state structures of these
species.
reaction between 2 and 1 as ruthenium-alkoxides resulting from
rapid, exergonic ketone-hydride insertion.
The remaining assignments of the broad structural features
were completed as follows. Both ester groups were bonded to
13
31
13
1
ruthenium as shown by C- P coupling in the C{ H} NMR
spectra. The coupling constants were similar for both ester
groups and ranged from 2 to 3 Hz, making it impossible to
13
31
determine with C- P coupling which of the ester groups were
Characterization of Catalyst-Substrate Species. (a) Broad
Structural Features. NMR experiments showed that 5 and 6
were ruthenium-alkoxide diastereomers resulting from ketone-
hydride insertion with hydride addition to carbon and ruthenium
addition to oxygen (eq 3). Specifically, the signals for the alkoxy
protons (RuOCH-) were identified as the overlapping peaks
trans to a phosphorus center. The presence of a single
coordinated MeCN ligand was confirmed by H NMR and by
1
15
15
31
N NMR of the N-labeled isotopomer. The signals in the
P
15
NMR spectra of the N-labeled MeCN isotopomers of 5 and 6
had only small P- N coupling constant (5: JP-N ) 2.8 Hz;
31
15
2
2
6
:
JP-N ) 2.9 Hz), showing that the MeCN ligand occupied
1
at δ 3.84 (br s) in the H NMR spectrum as follows. HMQC
10
a coordination site mutually cis to both phosphorus centers.
These data prove that diastereomers 5 and 6 contain one MeCN
ligand, and they contain 2 as a tridentate alkoxide ligand bonded
to ruthenium in a fac arrangement through the ester carbonyl
groups and the alkoxide group (Figure 1: shown for one
diastereomer). Such fac-tridentate bonding by the alkoxide of
experiments showed these signals correlated to the overlapping
signals at δ 87 in the 13C NMR spectrum. HMBC experiments
indicated the peaks at δ 87 arose from the alkoxy carbon centers.
We synthesized the deuterium-labeled diastereomers 5(RuOCD-)
and 6(RuOCD-) by reaction of 2 with 1-d1 (1-d1 is obtained by
reaction of [Ru((R)-BINAP)(1-3;5-6-η-C8H11)(MeCN)](BF4) (7)
with deuterium gas).6b This substitution resulted in loss of the
2
is expected as molecular models suggest the alkoxide cannot
readily adopt a mer-tridentate arrangement about ruthenium.
1
alkoxide signal at δ 3.84 in the H NMR spectrum of the
The magnitudes of the 31P NMR chemical shifts and coupling
constants for the major diastereomers (5, 8, and 10) formed by
the ketone-hydride insertion were similar. Further, the mag-
product, and inversion of the signal at δ 87 in 13C APT NMR
experiments (Scheme 1), proving that these signals correspond
to the proton and the carbon atom of the alkoxide, respectively.
These experiments unambiguously identify the products of the
3
nitudes of JP-C (2 to 3 Hz) for the ester carbonyls and the
(
9) King et al. (ref 3c) reported similar observations for reaction between â-keto
(10) We have shown previously (ref 6c) in the related complex [Ru((R)-BINAP)-
15
+
esters and the partially characterized hydrides obtained from hydrogenation
(H)( N-MeCN) ] that only when MeCN is coordinated trans to phosphorus
31 15 2
3
of Ru(BINAP)(Cl)
2
/NEt
3
in CH
2
Cl
2
. The products reported by King were
in these complexes does it show significant P- N coupling ( JP-N g 8
Hz).
the corresponding â-hydroxy ester and unidentified ruthenium complexes.
3
682 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002
9

Identification of a Catalyst−Substrate (Alkoxide) Species
A R T I C L E S
magnitudes of JP-N (∼ 3 Hz) in the 15N-labeled isotopomers
of 8, 9, and 10 all showed that the alkoxide ligands in these
complexes bind to ruthenium in a tridentate manner with a fac
arrangement of the ester carbonyl and alkoxide groups. Thus
the major diastereomers formed by the ketone-hydride insertion
have the same broad structural features (Figure 1).
2
(b) Absolute Configurations. The absolute configurations
at the alkoxide carbon centers were determined by protolytic
cleavage of the ruthenium-alkoxide bond in 5(RuOCD-) and
6(RuOCD-) with HBF4‚OEt2 at room temperature in CH2Cl2. The
protonolysis is complete upon mixing to generate the dicationic
adducts [Ru((R)-BINAP)(MeCN)((R and S)-MeCO2CD(OH)-
C(Me)2CO2Me)](BF4)2. Addition of excess MeCN liberated
MeCO2CD(OH)C(Me)2CO2Me (11-d1) and formed the known
dicationic ruthenium compound [Ru((R)-BINAP)(MeCN)4]-
11
(
BF4)2 (14, eq 4). No H-D exchange occurred at the alkoxide
Figure 2. Assignments of substrate H and 13C NMR signals of 8 along
1
with key HMBC correlations (shown by arrows to structure).
They differ by which coordination sites are occupied by the
alkoxide and ester groups. In diastereomer I the alkoxy group
is coordinated trans to P2, the phosphorus center attached to
the naphthyl ring on the same side of the P1-Ru-P2 plane as
the MeCN ligand; in II the alkoxy group is coordinated trans
to P1, the phosphorus center attached to the naphthyl ring on
the opposite side of the P1-Ru-P2 plane as the MeCN ligand;
and in III the alkoxy group is coordinated cis to both phosphorus
centers. Overlap of signals in the H and C NMR spectra of
the diastereomers 5 and 6 prevented us from obtaining unam-
biguous correlation data to distinguish between I, II, or III.
Such overlap was minimal for diastereomers 8 and 9 because
the corresponding minor diastereomers were present in low
concentrations. The completed characterization of the major
diisopropyl diastereomer 8 is described below.
carbon of the product, showing that epimerization did not occur
during the protonolysis. Furthermore, the ee of the liberated
alcohol 11-d1 was 60% (R), which corresponds to the ratio of
1
13
5(RuOCD-) to 6(RuOCD-) used for this experiment (CH2Cl2 solution;
8
0:20). Combined, the results from protonolysis of 5(RuOCD-)
and 6(RuOCD-) unambiguously assign the absolute configuration
of the major alkoxide diastereomer 5 as R, and that of the minor
diastereomer 6 as S. Protonolysis of a 90:10 mixture of 8(RuOCD-)
and 9(RuOCD-) generated (after liberation by MeCN) the alcohol
(
d) Substrate Signals. Figure 2 shows these assignments.
1
1
i
i
Analysis of the COSY and selective H{ H} decoupling data
assigned by correlation the signals from the protons in each
isopropyl group (isopropyl A: δ 0.76 and 1.15smethyl groups,
and δ 4.76smethyne proton; isopropyl B: δ 1.09 and 1.23s
methyl groups, and δ 5.42smethyne proton). Analysis of the
PrCO2CD(OH)C(Me)2CO2 Pr (12-d1) in 79% ee (R) without
observable H-D exchange. Finally, protonolysis of the exclu-
sively formed diastereomer 10(RuOCD-) generated (after liberation
t
t
by MeCN) (R)- BuCO2CD(OH)C(Me)2CO2 Bu (13-d1) exclu-
sively, without observable H-D exchange. In each case, the
observed ee corresponds to the observed diastereomeric excess
1
3
HMQC data identified the C signals that corresponded to the
isopropyl methyne carbons, the alkoxy carbon, and the methyl
groups on the backbone of the coordinated substrate (Figure 2;
carbons 3 and 10; 5; and 7 and 8, respectively). With these
assignments, the analysis of the HMBC data showed that the
alkoxide methyne proton coupled to both ester carbonyl carbon
signals at δ 182 and 190. Further, the backbone methyl groups
coupled with only the ester carbonyl carbon signal at δ 182,
unambiguously assigning it to the ester carbonyl group â to
the alkoxide unit (carbon 9), and assigning the remaining
carbonyl carbon signal (δ 190) to the R ester carbonyl group
(de) of the respective alkoxide intermediates. The major
diastereomer of the catalyst-alkoxide complex formed by
reaction of 2, 3, or 4 with the active catalyst 1 therefore contains
the (R)-alkoxide in all cases. This absolute configuration matches
that of the major product enantiomer of the catalytic hydrogena-
tion of these substrates.
(
c) Fine Structural Features. Figure 1 shows the only three
possible isomers of the major ruthenium-alkoxide diastereomers
, 8, and 10 that are consistent with the absolute configuration
5
and the bulk structural features determined thus far (vide supra).
They are all of R absolute configuration at the alkoxy carbon,
the MeCN ligand is in a coordination site cis to both phosphorus
centers, and the substrate is bonded to ruthenium as a tridentate
ligand through the alkoxy and ester carbonyl oxygen atoms.
(carbon 4). HMBC data also showed that the methyne proton
signal from the isopropyl unit A (δ 4.76) coupled to the R ester
1
2
carbonyl carbon (δ 190). Therefore isopropyl unit B is that
(
12) The methyne proton signal of isopropyl group B (δ 5.42) did not show
any correlation to either ester carbonyl carbon signal (by HMBC experi-
ments), that the methyne proton signal of isopropyl group A (δ 4.76) did
is sufficient proof of assignments.
(
11) Mashima, K.; Hino, T.; Takaya, H. J. Chem. Soc., Dalton Trans. 1992,
2
099.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002 3683

A R T I C L E S
Daley and Bergens
Figure 3. Assignment of ortho protons on the phenyl and naphthyl rings
bonded to P1 and P2 from 31P- H HETCOR and variable-temperature
ROESY data. The quadrant of space about the ruthenium center was
identified.
1
Figure 4. Important ROE contacts observed between the substrate and
(
R)-BINAP ligand at ambient-temperature to -40 °C (a, b, and c; all strong)
and key ROE contacts at -80 °C (d, e, and f; all strong) that conclusively
identified III as the structure of 8 (as well as for 5 and 10).
of the â ester group. Similar assignments were made for the
substrate unit in both the dimethyl major diastereomer 5 and
the di-tert-butyl diastereomer 10, except that for 10 the HMBC
be on the naphthalene ring bonded to P2 and on the same side
of the ruthenium-phosphine plane as CH3CN (Figure 3, P2-
1
experiments could not distinguish the tert-butyl group H NMR
1
o-Naph). This P2-o-Naph H signal also showed a ROE contact
13
signals to the R or â ester groups due to the limits of detection.
e) BINAP Signals and Correlations. Figure 3 shows the
1
at ambient-temperature to the H signal (δ 7.64) from two
(
o-phenyl protons associated with P2. This signal is thereby
unambiguously assigned to the equatorial P2 o-phenyl protons
as they are the only ortho protons close to the P2-o-Naph
conformation of coordinated (R)-BINAP and the spatial distri-
bution of its components relative to the cis-coordinated MeCN
ligand. The relative tilt of the naphthyl rings forces one phenyl
ring of each phosphorus axial and the other equatorial. Placing
MeCN in the cis-coordination site shown in Figure 3 orients
the P2 axial phenyl (P2-Phax) ring on the opposite side of the
P1-Ru-P2 plane as MeCN, and the naphthyl group (P2-Naph)
on the same side of the plane as MeCN. Consequently, P1-
Phax is on the same side of the P1-Ru-P2 plane as MeCN and
P1-Naph is opposite. As (R)-BINAP is C2-symmetric, it is of
no consequence which side of the P1-Ru-P2 plane the MeCN
ligand is initially placed for this assignment of structure. The
choice shown in Figure 3 was arbitrary.
1
proton. Finally, the H NMR signal for the P1-o-Naph proton
was found at δ 7.02. The signals of both P1-o-Phax (observed
at <-40 °C) and P1-o-Pheq (observed at <-80 °C) were
31
1
identified by P- H HETCOR studies and ROESY studies at
low temperatures that are described further below.
Having identified the absolute configuration at the alkoxide
group, and having assigned the NMR signals to key substrate
and BINAP elements, it was possible to use low-temperature
ROESY experiments to unambiguously assign the structure of
the major diastereomer. At -40 °C, the backbone methyl group
proton signals (δ 0.8, Figure 4, correlations b and c) showed
ROE contacts to the P2-o-Pheq (δ 7.64) and the P2-o-Phax (δ
31
1
Variable-temperature P- H HETCOR experiments assigned
1
the H NMR signals to the ortho protons on the phenyl and
8
.76) proton signals. Thus one backbone methyl group is close
3
1
naphthyl rings bonded to P1 and P2 (Figure 3). Each P signal
to both the P2 axial and equatorial phenyl rings. Further, the
methyne proton and one of the methyl group protons of
isopropyl unit B showed ROE contacts to P2-o-Pheq (Figure 4,
correlation a), showing this group is proximal to P2 as well.
At -80 °C the alkoxy methyne proton (δ 3.85) had ROE
contacts with the P2-o-Phax (Figure 4, correlation d) as well as
will show from three (with rapid rotation about the P-phenyl
1
bond) to five (with slow rotation about the P-phenyl bond) H
correlations with the two axial o-phenyl protons, with the two
equatorial o-phenyl protons, and with the one o-naphthyl proton.
1
Ambient-temperature studies identified only two H correlations
1
to P2 and one H correlation to P1, showing that rotation about
14
with a P1-o-Pheq proton (δ 9.15, Figure 4, correlation e).
most of the phosphorus-phenyl bonds was rapid at room
Inspection of molecular models shows that the only alkoxide
diastereomer (of I, II, or III (Figure 1)) with the alkoxy methyne
proton in proximity to P2-Phax and P1-Pheq, with the isopropyl
unit B in proximity to P2-Pheq, and with a backbone methyl
group in proximity to P2-Pheq and P2-Phax is III (Figure 4).
This assignment is unambiguous. The major diastereomers of
1
temperature, and that some o-phenyl H NMR signals were in
the baseline of the spectrum. Indeed, variable low-temperature
investigations (-40 to -100 °C) were required to identify all
the ortho protons. Some of these signals only grew out of the
baseline at -80 °C, suggesting that rotation about the P-phenyl
bond can be quite facile in BINAP complexes.
Figure 3 shows assignments of the BINAP-MeCN and
BINAP-BINAP correlations. Ambient-temperature ROESY
experiments showed a ROE contact between CH3CN and only
one o-naphthyl signal (δ 8.02). This ortho proton must therefore
(14) The signal at δ 9.15 was assigned to a P -o-Ph proton as follows. 31P-
1
eq
1
H HETCOR experiments carried out at -100 °C showed correlation to
P
1
. ROESY experiments carried out at -80 °C showed a ROE contact
with P -o-Ph (Figure 4, correlation f). The only o-phenyl protons attached
2
ax
to P
protons are on the opposite side of the P
signal showed exchange (-80 °C) with the H NMR signal at δ 6.2
indicating that the latter arises from the other P -o-Pheq proton exchanged
by rotation about the P -Pheq bond. The remaining P -o-phenyl protons
were then assigned to P -o-Phax (δ 7.15 and 7.95 at - 40 °C) as they
were the only ones left unassigned.
1
that are in proximity to P
2
-o-Phax are P
1
-o-Pheq as the P
1
-o-Phax
1
-Ru-P
2
plane (Figure 4). This
1
(
13) The experimental limits of detection of the HMBC NMR experiment are
generally 3-bond correlations. The tert-butyl methyl protons are four bonds
away from their respective ester carbonyl carbon atoms and thus correlations
were not observed.
1
1
1
1
3684 J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002

Identification of a Catalyst−Substrate (Alkoxide) Species
A R T I C L E S
the hydride insertion reactions between 1 and the ketones 2, 3,
calculations substantiate,17 that protonation or hydrogen bonding
to the carbonyl oxygen of unfunctionalized and certain func-
tionalized ketones is required for, or enhances the rate of,
ketone-hydride insertion. That there is little kinetic impediment
toward ketone-hydride insertion in this system shows that
functionalized ketone-hydride insertion can be quite rapid in
the absence of acid for hydride-catalyst systems with a vacant
coordination site cis to the hydride.
1
5
and 4 all therefore have the solution-state structure III. These
species are of the same absolute configuration as the major
product enantiomer of the catalytic hydrogenation. Further
experiments were carried out to investigate whether these species
were true catalytic intermediates.
Low-Temperature Studies. Stoichiometric reactions between
1
and each of the ketone substrates 2, 3, and 4 in THF-d8 were
Isolated Ruthenium-Alkoxide Complexes as Probable
Catalytic Intermediates. The rates of stoichiometric hydro-
genolysis of the ruthenium-alkoxide bonds in 5(RuOCD-),
carried out at low temperatures and monitored by NMR
spectroscopy to investigate the ketone-hydride insertion (eq
5
). No evidence of reaction was observed below -30 °C. The
6(RuOCD-), 8(RuOCD-), 9(RuOCD-), and 10(RuOCD-) were much lower
than the rates of ketone-hydride insertion. The ketone-hydride
insertions were complete on mixing at room temperature, and
they even occurred at -30 °C. The stoichiometric hydrogenoly-
ses only occurred at significant rates under the conditions of
the catalytic hydrogenation (50 °C and 50 atm of H2). The
stoichiometric hydrogenolyses were complete after 1 h under
these conditions. This rate is comparable to the turnover
frequencies of the catalytic hydrogenations (∼1 turnover/h in
18
THF or in MeOH). Within the accuracy of these experiments,
the rate of the stoichiometric hydrogenolysis is consistent with
the turnover frequency of the catalytic hydrogenation. These
rate experiments thus cannot rule out the diastereomeric
alkoxides as catalytic intermediates.
The rate and structural data suggest that the diastereomeric
catalyst-alkoxides are catalytic intermediates. If this suggestion
is correct, the mechanism of the catalytic hydrogenation involves
a rapid, exergonic ketone-hydride insertion followed by a
significantly slower, turnover-limiting hydrogenolysis of the
ruthenium-alkoxide bond. This proposed mechanism imposes
two preliminary requirements on the system that should be
fulfilled before the mechanism can be considered further. The
first requirement is that the diastereomeric catalyst-alkoxides
actually form under catalytic conditions. The second require-
ment, imposed by the relative rates of the exergonic insertion
insertion begins at -30 °C to form the ruthenium-alkoxide
diastereomers (5/6, 8/9, and 10) without other observable
intermediates. Assuming ketone-hydride insertion in these
systems is analogous to olefin-hydride insertion, in that it
2
requires prior coordination of the ketone as an η -π-ligand cis
1
6
to the hydride, these results show that ketone coordination is
slower than ketone-hydride insertion in this system. An
intriguing alternative is that this “insertion” reaction proceeds
via nucleophilic attack of the hydride ligand on the carbonyl
carbon. Such a process would not require coordination of the
ketone to ruthenium and would be facilitated by coordination
of the ester groups. Regardless of the mechanistic details, this
exergonic ketone-hydride insertion is too facile to be the
turnover-limiting step of catalytic hydrogenations involving 5/6,
(fast) and the hydrogenolysis (slow), is that the alkoxides are
the predominant ruthenium-containing species in solution during
catalysis. To investigate these issues, the catalytic hydrogenation
of the di-tert-butyl ester 4 (50 °C and 50 atm of H2) was
interrupted by depressurization and cooling to room temperature
after ∼4 turnovers. As predicted by the proposed mechanism,
the catalyst-alkoxide diastereomer 10 was the sole detectable
ruthenium-containing species present in the reaction mixture
after the hydrogenation was interrupted. Assuming that 10
formed during the catalytic reaction, and not as some conse-
quence of depressurization, this observation indicates the
diastereomeric alkoxides are the major ruthenium species in
solution during the catalytic hydrogenation.
8
/9, and 10 as intermediates. Further, the ratios of the product
alkoxide diastereomers formed at low temperatures equaled
within experimental error the ratios formed at room temperature,
showing the diastereoselectivity of the insertion is only mildly
sensitive to temperature changes.
Although the ketone groups in 2, 3, and 4 are activated toward
reduction by the R-ester groups, they are rather sterically
crowded by the gem-methyl groups and (for 4) to some extent
The stoichiometric hydrogenolysis of 5(RuOCD-), 6(RuOCD-),
8(RuOCD-), 9(RuOCD-), and 10(RuOCD-) were carried out in MeOH
and THF under the conditions of the catalytic hydrogenation
(under 50 atm of H2 at 50 °C) to determine if ketone-hydride
insertion was reversible on the time scale of the hydrogenolysis.
If the reverse of ketone-deuteride insertion (a net â-deuteride
1
a
by the tert-butyl ester substituents. Noyori has proposed, and
(
15) The analysis of the variable-temperature ROESY NMR data for the tert-
butyl complex 10 also corroborated the structure. While the complex 5
was complicated by overlap with signals from 6, the resulting patterns also
correspond to those observed with the complexes of 8 and 10.
16) For references on ketone/aldehyde metal hydride insertions see: (a)
Nietlispach, D.; Veghini, D.; Berke, H. HelV. Chim. Acta 1994, 77, 2197.
(
(
b) Portnoy, M.; Milstein, D. Organometallics 1994, 13, 600. (c)
(17) (a) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122, 1466.
(b) Alonso, D. A.; Brandt, P.; Nordin, S. J. M.; Andersson, P. G. J. Am.
Chem. Soc. 1999, 121, 9580.
(18) The turnover frequencies were determined by interrupting catalytic
hydrogenations of 2, 3, and 4 at time t and dividing the number of turnovers
by t.
Vanderzeijden, A. A. H.; Bosch, H. W.; Berke, H. Organometallics 1992,
1
7
1, 2051. (d) Vanderzeijden, A. A. H.; Berke, H. HelV. Chim. Acta 1992,
5, 513. (e) Esteruelas, M. A.; Valero, C.; Oro, L. A.; Meyer, U.; Werner,
H. Inorg. Chem. 1991, 30, 1159. (f) Rakowski, M. C.; Muetterties, E. L.
J. Am. Chem. Soc. 1977, 99, 739 and references sited therein.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002 3685

A R T I C L E S
Daley and Bergens
Scheme 2. Sequence of Steps for H-D Exchange at Alkoxide
Carbon of Diastereomeric Ruthenium-Alkoxides
indicate that the rate of hydrogenolysis is faster than the rate
the alkoxides revert to free catalyst and ketone under conditions
of the catalytic hydrogenation. If this mechanistic proposal is
correct, rapid preequilibrium between the diastereomeric alkox-
ides will not occur during the catalytic hydrogenation, and the
ratio of the diastereomeric alkoxides will reflect the ratio of
alcohol product enantiomers. In other words, the enantioselec-
tivity of such a mechanism is weighted more toward formation
of the diastereomeric alkoxides and less toward hydrogenolysis
of the ruthenium-alkoxide bond. We found that the ratio of
alcohol product enantiomers obtained from the catalytic hydro-
genation of 2 in THF (79.5 (R):20.5 (S) (the same ratio was
obtained in MeOH)) was similar to the ratios of the catalyst-
alkoxide diastereomers formed by the stoichiometric ketone-
hydride insertions between 2 and 1 carried out at room
temperature under ∼1 atm of dihydrogen (79 (R):21 (S) in THF
(75 (R):25 (S) in MeOH)). Such similarities between the alcohol
catalysis product enantiomer ratio and the catalyst-alkoxide
diastereomer ratio were also obtained by using the diisopropyl
ketone substrate 3 (catalysis product enantiomer ratio, 83 (R):
17 (S); stoichiometric diastereomer ratio, 90 (R):10 (S)) and the
di-tert-butyl ketone substrate 4 (catalysis product enantiomer
ratio, 90 (R):10 (S); stoichiometric diastereomer ratio, ∼100
1
9
(
R): ∼0 (S)) in THF. These similarities in product ratios
elmination) occurs at a rate faster than hydrogenolysis of the
ruthenium-alkoxide bond, H-D exchange will occur at the
alkoxide carbon via the following sequence of steps. â-Deuteride
elmination within the diastereomeric catalyst-alkoxides fol-
lowed by substrate dissociation will form the ketone (2, 3, or
between the stoichiometric insertions and the catalytic hydro-
genations imply that the stoichiometric insertions carried out
under ∼1 atm of dihydrogen represent the insertions that proceed
under catalytic conditions (50 atm of H2 and 50 °C). This
implication is supported by two observations. First, the de of
the diastereomeric catalyst-alkoxide 10 isolated from the
interrupted catalytic hydrogenation of 4 (∼100%) was the same
as the de’s of 10 obtained from the room temperature and low-
temperature stoichiometric insertions (∼100%). Second, the de
of the insertion reaction is only mildly sensitive to temperature
changes (vide supra). These stoichiometric insertion reactions
therefore support the prediction by the proposed mechanism that
the enantioselection of the catalytic hydrogenation is weighted
toward the ketone-hydride insertion step. We point out,
however, that the de of the diastereomeric alkoxide 10 was
4
, respectively) and the deuterium-labeled catalyst (1-d1)
(Scheme 2). We have shown previously that the active catalyst
1
rapidly undergoes H-D exchange under ∼1 atm of D2 even
6b
at low temperatures. Any 1-d1 generated by â-deuteride
elmination within the diastereomeric catalyst-alkoxides will
undergo H-D exchange immediately upon formation under
catalytic conditions (50 atm of H2, 50 °C) to generate the
unlabeled catalyst 1. Ketone-hydride insertion between 1 and
the substrate followed by hydrogenolysis will then form the
unlabeled product alcohol. The amount of H-D exchange at
the alkoxide carbon in the product ((MeO2C)CD(OH)C(Me)2-
∼
100%, while the ee’s of the catalytic hydrogenations were
(
6
∼
CO2Me)) of stoichiometric hydrogenolysis of 5(RuOCD-) and
(RuOCD-) under catalytic conditions (50 atm of H2, 50 °C) was
30% in MeOH and ∼40% in THF. Only 15% H-D exchange
from 80 to 84%, showing the hydrogenolysis also contributes
to the enantioselection by such a catalytic cycle, but to a lesser
extent than the ketone-hydride insertion.
was observed in MeOH and 20% in THF during the stoichio-
metric hydrogenolysis of 8(RuOCD-) and 9(RuOCD-). The stoichio-
metric hydrogenolysis of 10(RuOCD-) resulted in only 10% H-D
exchange in MeOH and 20% exchange in THF. These experi-
ments prove that ketone-hydride insertion is only partially
reversible (in one case only 10%) on the time scale of
hydrogenolysis of the ruthenium-alkoxide bond under the
conditions of catalytic hydrogenation. In other words, the rate
of hydrogenolysis of the ruthenium-alkoxide bond is faster than
the rate the catalyst alkoxides revert to catalyst and free ketone
under catalysis conditions. The hydrogenolysis does not proceed
via a rapid (fully established) preequilibrium between the
diastereomeric alkoxides (traditionally called Curtin-Hammett
conditions) under catalytic conditions.
Finally, and as expected if the diastereomeric alkoxides were
true catalytic intermediates, the ee’s of the alcohols obtained
from the stoichiometric hydrogenolyses carried out under
catalytic conditions (50 atm of H2, 50 °C) were quite close to
the ee’s obtained from the catalytic hydrogenation (% ee from
hydrogenolysis (ee from catalytic hydrogenation): 5(RuOCD-)/
6
9
(RuOCD-), 60 (59) in MeOH, 58 (59) in THF; 8(RuOCD-)/
(RuOCD-), 70 (68) in MeOH, 69 (66) in THF; 10(RuOCD-), 84
(82) in MeOH, 84 (80) in THF). The slight differences between
the stoichiometric and catalytic ee’s likely arise from experi-
mental factors such as reaction occurring during warm-up of
the pressure reactor during the stoichiometric hydrogenolysis
and from isotope effects (the ee for the stoichiometric hydro-
genolysis of 10 in THF was 82% versus 84% for 10(RuOCD-)).
These experiments show that the product ee’s from stoichio-
metric hydrogenolysis of the mixture of diastereomeric alkoxides
The structure, rate, interruption, and deuterium exchange data
support the proposal that the catalytic hydrogenation proceeds
via a rapid ketone-hydride insertion followed by a slow
hydrogenolysis of the ruthenium alkoxide bond. The data further
(19) These stoichiometric reactions were not carried out in MeOH.
3686 J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002

Identification of a Catalyst−Substrate (Alkoxide) Species
A R T I C L E S
Figure 5. Proposed catalytic cycle. The active catalyst 1 proceeds through a major and minor pathway where the major pathway also corresponds to the
major diastereomeric catalyst-alkoxide species observed in solution (5, 8, and 10).
formed at room temperature under ∼1 atm of dihydrogen is
quite close to the product ee’s of the catalytic hydrogenation.
This evidence, along with the proof from the deuterium
exchange experiments that the rate of hydrogenolysis is faster
than the rate the diastereomeric alkoxides revert to free catalyst
and ketone (vide supra), is strong indication that the diastereo-
meric alkoxides are actual catalytic intermediates. In other
words, the mixture of diastereomeric alkoxides formed at room
temperature reacts with dihydrogen to form alcohol product in
nearly the same ee as the catalytic hydrogenation at a rate faster
than they could invert absolute configuration by reversion to
free catalyst and ketone. We point out, however, that high-
pressure kinetics and NMR studies are required to rigorously
prove this conclusion.
1. The ketone substrate (2, 3, or 4) then undergoes fast,
exergonic insertion into the Ru-H bond to yield the major and
minor diastereomeric catalyst-alkoxide intermediates (5 and
6; 8 and 9; and 10 from ketones 2, 3, and 4, respectively). This
insertion is only partially reversible on the time scale of
hydrogenation, and it contributes greatly to the enantioselection
by the catalytic cycle. In the subsequent steps, the ruthenium-
alkoxide bond is irreversibly cleaved, likely by the addition of
dihydrogen, releasing the chiral alcohol product and reforming
the active catalyst species 1 to complete the catalytic cycle. As
the catalyst-alkoxides are 18-electron species, it is proposed
that either desolvation of the MeCN ligand or dissociation of a
coordinated ester occurs prior to addition of H2 and subsequent
reductive elimination of the alcohol products.
Proposed Mechanistic Pathway and Origins of Enantio-
selection. The rate of hydrogenolysis of the ruthenium-alkoxide
bonds in the isolated diastereomeric alkoxides was consistent
with the catalytic turnover frequency under the same reaction
conditions. The ratio of diastereomeric alkoxides formed by
stoichiometric ketone-hydride insertion is similar to the ratio
of enantiomers formed by the catalytic hydrogenation. The
stoichiometric hydrogenolysis does not proceed by a rapid (fully
established) preequilibrium between the diastereomeric alkox-
ides, and the stoichiometric ee of the product alcohols is quite
close to the catalytic ee. Taken together, these results are strong
evidence that the diastereomeric catalyst-alkoxides formed by
ketone-hydride insertion reactions between the active catalyst
The weight of evidence that the diastereomeric alkoxides are
catalytic intermediates warrants consideration of possible links
between their structures and the enantioselection by the catalytic
hydrogenation. As stated previously, that the rate of hydro-
genolysis is faster than the rate of reversion to free catalyst and
ketone, and that the ratio of diastereomeric alkoxides is similar
to the ratio of product alcohol enantiomers is strong evidence
that the catalytic enantioselection is weighted toward the steps
prior to the hydrogenolysis. Both the ratio of product enantio-
mers from the catalytic hydrogenation and the ratio of diaster-
eomeric alkoxides increase when the steric bulk of the ester
groups is increased. This is evidence that there is a relationship
between the relative energies of the diastereomeric alkoxides
and the enantioselection by the catalytic cycle. By assuming
the minor diastereomer has the same broad structural features
(alkoxide group and MeCN are both cis to both phosphorus
centers, and both ester groups coordinated) the steric origins of
1
and the ketone substrates 2, 3, and 4 are catalytic intermedi-
ates. Figure 5 shows the mechanism for the catalytic hydrogena-
tion we propose based on the mechanistic and structural
evidence presented in this paper. Hydrogenation of the catalyst
precursor (7) forms cyclooctane and the active catalyst species
J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002 3687

A R T I C L E S
Daley and Bergens
Figure 6. Possible steric influence on enantioselection. Increasing ester group size increases the severity of steric interactions of the minor diastereomer,
and increases the overall enantioselection, the tert-butyl ester giving the greatest enantioselection. Note rotation about the phosphorus-phenyl bonds is rapid
at room temperature.
the energy difference between the diastereomers becomes
apparent. Figure 6 illustrates this steric interaction.
reversible ketone-hydride insertion and hydrogenolysis of the
ruthenium-alkoxide bond, but it is weighted toward the
ketone-hydride insertion. As is the case with any mechanism
study, caution must be exercised when extending these results
to other systems.
Inspection of molecular models reveals no significant steric
repulsions in the major diastereomer. The minor diastereomer,
however, contains a strongly repulsive steric interaction between
the ester alkoxy group A (R-ester to ketone carbonyl) and P2-
Pheq. The steric crowding between these groups increases as
does the size of the ester alkoxy groups, and we propose that it
is this interaction that is responsible for the enantioselection by
the catalytic cycle.
Experimental Section
Materials and Methods. All operations were performed under an
argon atmosphere with standard Schlenk techniques. The solvents were
dried and distilled under an argon atmosphere by standard methods
before use.21 The argon gas (Praxair, 99.998%) was passed through a
Conclusions
drying train containing 3 Å molecular sieves and P
Trace quantities of oxygen were removed from the dihydrogen gas
Praxair, 99.99%) by passage through an Alltech Oxy-Trap. All
4
O10 before use.
This work presents the first complete and unambiguous
structure determination of a diastereomeric substrate-catalyst
adduct in an enantioselective ketone hydrogenation. It also
provides the first direct experimental information on the reaction
dynamics of how such diastereomers form, on how they react,
and on the steric forces within these probable catalytic inter-
mediates. In these systems, ketone-hydride insertion is rapid
and exergonic, and it may proceed without prior formation of
(
commercial reagents (Aldrich or Fluka) were recrystallized or distilled
under an argon atmosphere before use. [Ru((R)-BINAP)(MeCN)(1-3:
5
,6-η-C
8
H
11)](BF
4
) (7),5a dimethyl oxaloacetate, and di-tert-butyl
22
22
oxaloacetate were prepared using established procedures.
Unless stated otherwise, all H, 13C, and 31P NMR spectra were
measured with Bruker AM-400 or AM-500 spectrometers. H and
1
1
13
C
NMR chemical shifts are reported in parts per million (δ) relative to
2
an η -π-ketone-catalyst complex. Further, the rate of hydro-
31
TMS with the solvent as an internal reference. P NMR chemical shifts
genolysis is higher than the rate of reversion to catalyst and
ketone. The structural and mechanistic evidence we obtained
strongly supports that these catalytic hydrogenations operate by
the mechanism shown in Figure 5. In this mechanism, and unlike
hydrogenations of MAC and related olefin substrates catalyzed
are reported in parts per million (δ) relative to an 85% H PO external
3
4
reference. All 13C and P NMR are 1H decoupled unless stated
otherwise. Mass spectra were measured with a Kratos MS50 spec-
trometer. Microanalyses were performed at the University of Alberta
Microanalysis Laboratory. Optical rotations were measured with a
Perkin-Elmer 241 polarimeter at 589 nm with 1.0 dm cells. Specific
31
+
by [Rh(chiral bisphosphine)(sol)2] , the diastereomeric catalyst-
D
rotations, [R] , are reported in degrees per decimeter at 25 °C, and the
concentration (c) is given in grams per 100 mL.
substrate intermediate present in the highest concentration leads
to the major product enantiomer.20 The net enantioselectivity
of the catalytic cycle is partitioned between the partially
Diisopropyl Acetylenedicarboxylate. To a 100 mL flask was
-1
transferred acetylene dicarboxylic acid (20.0242 g, 1.756 × 10 mol),
(
20) As stated previously, a high-pressure kinetics investigation should support
this mechanistic scheme. We note that the alkoxides cannot be mere catalytic
sinks because they react with little reversible formation at approximately
the same rate as the turnover frequency.
(21) Casey, M.; Leonard, J.; Lygo, B.; Procter, G. AdVanced Practical Organic
Chemistry; Chapman & Hall: London, 1990; p 28.
(22) Sucrow, W.; Grosz, K. P. Synth. Commun. 1979, 9, 603.
3688 J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002

Identification of a Catalyst−Substrate (Alkoxide) Species
concentrated H SO (2.5 mL), and 2-propanol (150 mL). The solution
A R T I C L E S
2
4
mg, 86.8% yield). A similar procedure was utilized for the reduction
1
13
23
was stirred and heated to reflux (80 °C) for 5 h, after which it was
of 12 (85.5% yield). H and C NMR correspond to literature data.
cooled to room temperature and then concentrated under reduced
[Ru((R)-BINAP)(MeCN)(OC(H)(CO CH )(C(CH ) (CO CH )))]-
2
3
3 2
2
3
pressure. The remaining clear-colorless liquid was taken up in Et
2
O
4
(BF ) (5/6). To a 20 mL solvent Schlenk was transferred 7 (100.5 mg,
-
4
(100 mL) and sequentially washed with a 2 N solution of NaOH until
1.05 × 10 mol). The flask was placed under vacuum and refilled
with argon gas (×3) to remove all traces of oxygen. To a sidearm flask
the aqueous layer was basic (2 × 20 mL) and with distilled H O (4 ×
2
-
4
2
5 mL). The organic layer was dried over MgSO
4
for 24 h, filtered,
was transferred 2 (21.7 mg, 1.15 × 10 mol) in the glovebox. Next,
solvent (1.5 mL; acetone, THF, or methanol) was added to the flask
via gastight syringe and the solution was transferred to the solvent
Schlenk via cannula. The flask was rinsed with solvent (8.5 mL) and
transferred to the solvent Schlenk. The tube was flushed with dihy-
drogen gas for 2 min and then pressurized to 20 psig. The tube was
shaken vigorously until all of the catalyst precursor was in solution,
and for a further 5 min to ensure complete reaction. The tube was then
depressurized to atmospheric pressure under argon gas, the solution
transferred to a sidearm flask, and the solvent removed under reduced
pressure. The solid residue was redissolved in THF (2.0 mL) and
precipitated with pentane (25 mL). The solution was filtered and the
solid washed with pentane (3 × 5 mL) and dried under high vacuum
and washed with Et
under reduced pressure yielding the clear-colorless liquid product (30.2
g, 86.8%). The product was used without further purification. H NMR
2
O (3 × 100 mL), and the solvent was removed
1
3
(
400.1 MHz, CD
CR, R ) CO C(CH
R ) CO C(CH
4C, RCtCR, R ) CO
2
Cl
2
, 25 °C): δ 1.28 (d, 12H, JH-H ) 6.5 Hz, RCt
3
2
3
)
2
(H)), 5.12 (quint, 2H, JH-H ) 6.5 Hz, RCtCR,
13
2
3
)
2
(H)). C NMR (100.6 MHz, CD
C(CH (H)), 71.34 (2C, RCtCR, R )
(H)), 74.68 (2C, RCtCR, R ) CO C(CH (H)), 151.45
C(CH (H)).
Dimethyl 3,3-Dimethyloxaloacetate (2). To a dry 100 mL flask
2 2
Cl , 25 °C): δ 21.58
(
2
3 2
)
CO
2
C(CH
3
)
2
2
3 2
)
(2C, RCtCR, R ) CO
2
3 2
)
was transferred potassium carbonate (1.76 g, 12.8 mmol). The flask
was flushed with dry argon gas for 20 min and then charged with
dimethyl oxaloacetate (1.02 g, 6.38 mmol) and acetone (30 mL,
distilled). A condenser was added to the flask and the system was kept
under an argon atmosphere. The solution was stirred for 5 min then
methyl iodide (2.4 mL, 38.3 mmol) was added down the condenser,
and the mixture was heated and stirred under reflux for 20 h. The
solution was then cooled and the solvent removed under reduced
3
1
for 2 h (96.3 mg, 88.7% yield). The ratio of products observed by
P
NMR analysis when synthesized was 79:21 in THF and 75:25 in
methanol. To obtain 15N-enriched complexes, the compounds were
15
dissolved in acetone and excess NCMe was added. The solution was
stirred for 2 h and 31P NMR analysis indicated >95% exchange of
1
4
15
NCMe with NCMe had occurred. FT-IR (CH
2
Cl
2
, 22 °C): 1608
cm- (w, νCdO), 1636 cm (s, νCdO). ESI-MS (pos) m/z 954.1 (M) ,
1
-1
+
pressure. The solid residue was washed with CH
passed through a column of alumina (10 cm long, 1 cm diameter). The
column was further washed with CH Cl (20 mL) and the solvent was
removed under reduced pressure to yield a clear pale yellow liquid
2 2
Cl (20 mL) and
+
exact mass calcd for C54
NP RuBF
.37. 5 (absolute configuration (R)): H NMR (400 MHz, acetone-d
25 °C): δ 0.94 (s, 3H, C(CH ), 1.03 (s, 3H, C(CH ), 1.93 (s, 3H,
CH CN), 3.41 (s, 3H, R-C(O)OCH ), 3.62 (s, 3H, â-C(O)OCH ), 3.84
br s, 1H, Ru-O-CH, overlapped with minor), 6.45-7.95 (m, 32H,
H
48
O
5
NP
2
Ru , 954.2. Anal. Calcd for C54
48 5
H O -
2
4
: C, 62.32; H, 4.65; N, 1.35. Found: C, 61.97; H, 4.44; N,
2
2
1
1
6
,
1
3
)
2
3 2
)
(
6
(
0.97 g, 80.8% yield). H NMR (400 MHz, CD
H, C(CH ), 3.67 (s, 3H, CO CH ), 3.82 (s, 3H, CO
400 MHz, CD Cl , 25 °C): δ 21.94 (2C, C(CH
CO CH ), 52.92 (1C, C(CH ), 53.33 (1C, CO CH ), 160.94 (1C, CO
CH ), 173.24 (1C, CO ), 191.77 (1C, C(O)CO CH ). CI-MS (pos)
m/z 206.1 ((M + NH
2 2
Cl , 25 °C): δ 1.39 (s,
13
3
3
3
)
3 2
2
3
2
3
CH ). C NMR
(
2
2
3 2
) ), 52.76 (1C,
13
BINAP, overlapped with minor). C NMR (400 MHz, acetone-d
C): δ 4.09 (1C, CH CN), 21.04 (1C, C(CH ), 25.35 (1C, C(CH
6.98 (1C, C(CH ), 55.0 (1C, CO CH ), 55.1 (1C, CO CH ), 86.7
1C, Ru-O-CH, overlapped with minor), 126.8 (1C, CH CN), 127-
42.5 (BINAP, overlapped with minor), 183.5 (1C, â-CO CH ), 190.2
, overlapped with minor). P NMR (400 MHz, acetone-
6
, 25
2
3
3
)
2
2
3
2
-
°
4
(
1
(
d
3
)
3 2
3
)
2
),
3
2
CH
+
3
2
3
)
3 2
2
3
2
3
+
4
), exact mass calcd for C
H
8 12
O
5
+ NH
4
)
3
2
06.2). Anal. Calcd for C
8 12 5
H O : C, 51.06; H, 6.43. Found: C, 50.89;
2
3
H, 6.58.
31
1C, R-CO
6
2
CH
3
Diisopropyl 3,3-Dimethyloxaloacetate (3). The synthesis was
performed with a similar procedure as in the synthesis of dimethyl-
2
2
, 25 °C): δ 58.9 (d, 1P, JP-P ) 45.5 Hz), 62.3 (d, 1P, JP-P ) 45.5
31
15
Hz). P NMR (400 MHz, acetone-d
compound: δ 58.9 (dd, 1P, JP-P ) 45.0 Hz, 2JP-N ) 2.8 Hz), 62.3
(
6
, 25 °C) of NCMe enriched
1
3
(
,3-dimethyloxaloacetate. H NMR (400 MHz, CDCl
3
, 25 °C): δ 1.19
2
3
(H)), 1.33 (d, 6H, 3JH-H ) 6.5
d, 6H, JH-H ) 6.5 Hz, CO
2
C(CH
3
)
2
2
2
dd, 1P, JP-P ) 45.0 Hz, JP-N ) 2.8 Hz).
(absolute configuration (S)): 1H NMR (400 MHz, acetone-d
C): δ 1.12 (br s, 3H, C(CH ), 1.19 (br s, 3H, C(CH ), 1.89 (s, 3H,
CN), 3.37 (br s, 3H, R-C(O)OCH ), 3.52 (br s, 3H, â-C(O)OCH ),
3
Hz, CO
6
(
2
C(CH
3
)
2
(H)), 1.42 (s, 6H, C(CH
3
)
2
), 5.03 (quint, 1H, JH-H
3
)
C-
6
6
, 25
.5 Hz, CO
CH (H)).
Di-tert-butyl 3,3-Dimethyloxaloacetate (4). The synthesis was
performed with a similar procedure as in the synthesis of dimethyl 3,3-
2
C(CH
3
)
2
(H)), 5.11 (quint, 1H, JH-H ) 6.5 Hz, CO
2
°
CH
3
)
2
3 2
)
3 2
)
3
3
3
3
3
d
4
.84 (br s, 1H, Ru-O-CH, overlapped with major), 6.45-7.95 (m,
2H, BINAP, overlapped with major). C NMR (400 MHz, acetone-
13
1
dimethyloxaloacetate. H NMR (400 MHz, CDCl
H, C(CH ), 1.43 (s, 9H, CO C(CH ), 1.55 (s, 9H, CO
FT-IR (CH Cl
, 22 °C): 1744 cm- (shoulder, νCdO), 1725 cm (s,
3
, 25 °C): δ 1.39 (s,
6
, 25 °C): δ 4.3 (1C, CH
3
CN), 22.0 (1C, C(CH
3
)
2
), 23.2 (1C, C(CH
CH ), 86.7 (1C,
CN, overlapped
with major), 127-142.5 (BINAP, overlapped with major), 190.5 (1C,
3 2
) ),
6
3
)
2
2
3
)
3
2
C(CH ).
3 3
)
9.2 (1C, C(CH ), 53.1 (1C, CO
)
3 2
2
CH ), 55.0 (1C, CO
3
2
3
1
-1
2
2
Ru-O-CH, overlapped with major), 126.8 (1C, CH
3
ν
CdO).
-
3
31
2
,2-Dimethylbutan-1,3,4-triol. LiAlH
4
(0.122 g, 3.21 × 10 mol)
â-CO
(400 MHz, acetone-d
63.7 (br d, 1P, JP-P ) 45.0 Hz). P NMR (400 MHz, acetone-d
2
CH
3
2 3
), 191.2 (1C, R-CO CH , overlapped with major). P NMR
2
was transferred to a three-necked flask that had an addition funnel and
septa. To the addition funnel were transferred the hydrogenated product
6
, 25 °C): δ 57.4 (br d, 1P, JP-P ) 45.0 Hz),
2
31
6
, 25
-
4
15
2
1
3 (90.0 mg, 3.28 × 10 mol) and Et
transferred Et O (15 mL) via cannula. The system was flushed with
argon gas for 10 min and then left under an argon atmosphere. The
flask was cooled to -5 °C and, after 5 min, the alcohol Et O solution
was added dropwise over 10 min. The addition funnel was rinsed with
Et O and the solution was allowed to react for 1.5 h at ambient-
2
O (5 mL). To the flask was
°C) of NCMe enriched compound: δ 57.4 (dd, 1P, JP-P ) 45.5 Hz,
2
2
2
J
P-N ) 2.9 Hz), 63.7 (dd, 1P, JP-P ) 45.5 Hz, JP-N ) 2.9 Hz).
[Ru((R)-BINAP)(MeCN)(OC(H)(CO CH(CH )(C(CH (CO CH-
(CH ))](BF ) (8/9). (8/9) was synthesized in a similar manner as (5/
6) in THF. FT-IR (CH
2
2
3
)
2
3
)
2
2
2
3
)
2
4
-
1
-1
2
Cl
2
, 22 °C): 1610 cm (w, νCdO), 1638 cm
+
2
(s, νCdO). ESI-MS (pos) m/z 1010.2 (M) , exact mass calcd for
C
+
temperature. The flask was then cooled back to -5 °C and distilled
water (125 µL) was added slowly over 30 s. Next, 0.1 N HCl (125
µL) and distilled water (375 µL) were added dropwise in succession.
The solution was allowed to react for 30 min, then it was filtered and
58
H
56
O
5
NP
2
Ru , 1010.268. Anal. Calcd for C58
H
56
O
5
NP
2
RuBF
4
: C,
63.51; H, 5.15; N, 1.28. Found: C, 63.15; H, 5.32; N, 1.41. 8 (absolute
1
configuration (R)): H NMR (400 MHz, THF-d
8
, 25 °C): δ 0.76 (d,
3
3H, JH-H ) 5.5 Hz, R-CO
2
(CH(CH
3
)
2
)), 0.80 (s, 3H, C(CH ), 0.98
3 2
)
the solid washed with Et
under reduced pressure yielding the clear-colorless liquid product (38.2
2
O (4 × 10 mL). The solvent was then removed
(23) Matsuo, T.; Mori, K.; Matsui, M. Tetrahedron Lett. 1976, 23, 1979.
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3689
9

A R T I C L E S
Daley and Bergens
2
(
1
s, 3H, C(CH
3
3
)
2
), 1.09 (d, 3H, 3JH-H ) 5.5 Hz, â-CO
.15 (d, 3H, JH-H ) 5.5 Hz, R-CO (CH(CH )), 1.23 (d, 3H, JH-H
5.5 Hz, â-CO (CH(CH )), 1.89 (s, 3H, NCCH ), 3.82 (s, 1H, Ru-
(CH(CH )), 5.42
2
(CH(CH
3
3
)
2
)),
THF-d
J
8
, -30 °C): δ 59.8 (d, 1P, JP-P ) 45.5 Hz, 5), 63.0 (d, 1P,
2
2
2
3
)
2
P-P ) 45.5 Hz, 5), 56.4 (br d, 1P, JP-P ∼ 45 Hz, 6), 63.7 (br d, 1P,
)
2
3
)
2
3
2JP-P ∼ 45 Hz, 6). Similar results were observed upon low-temperature
investigation of both 3 and 4. The only observable species upon
warming were the alkoxide intermediates 8 and 9 for 3 (in the same
ratio observed in stoichiometric reaction of 1 with the ketone at room
temperature) and only 10 for 4.
3
O-CH), 4.76 (br quint, 1H, JH-H ) 5.5 Hz, R-CO
2
3 2
)
3
(
br quint, 1H, JH-H ) 5.5 Hz, â-CO
2
(CH(CH
, 25 °C): δ 3.9 (1C, CH
, 2 from R-ester group and 1 from â-ester group),
, from â-ester group), 46.0 (1C, C(CH ), 72.0
)), 73.0 (1C, R-CO (CH(CH )), 87.0 (1C, Ru-
O-CH), 124-142 (46C, BINAP carbons), 182 (1C, â-CO (CH(CH )),
)). P NMR (400 MHz, THF-d , 25 °C): δ
6.0 (d, 1P, JP-P ) 47.0 Hz, P of minor complex), 60.4 (d, 1P, JP-P
3 2
) )), 6.40-8.02 (m, 32H,
1
3
BINAP). C NMR (400 MHz, THF-d
2
2
8
3
CN),
0-21 (3C, CH(CH
5.0 (1C, CH(CH
(CH(CH
3 2
)
)
3 2
3
)
2
Typical Procedure for Hydrogenation. To a 25 mL sidearm flask
was transferred 7 (10.0 mg, 1.04 × 10 mol) and 50 equiv of the
-
5
(1C, â-CO
2
)
3 2
2
)
3 2
-4
2
)
3 2
corresponding ketone (2, 3, or 4; 5.21 × 10 mol) in a glovebox.
Solvent (methanol or THF, 5.0 mL) was added to the sealed flask via
gastight syringe. The flask was transferred to a pre-flushed, argon gas,
stainless steel bomb. The bomb was flushed for a further 15 min with
argon gas, then with dihydrogen gas for 10 min, and finally it was
pressurized to 50 atm. Once stabilized, the bomb was placed in a 50
°C oil bath and reacted for 50 h. The bomb was then cooled to ambient
temperature and depressurized, and the flask was placed under reduced
pressure to remove the solvent. The residue was passed through a
3
1
1
5
90 (1C, R-CO
2
(CH(CH
)
3 2
8
2
2
1
2
)
45.0 Hz, P
1
of major complex), 64.1 (d, 1P, JP-P ) 45.0 Hz, P
2
of
2
major complex), 66.4 (d, 1P, JP-P ) 47.0 Hz, P
2
of minor complex).
15
3
1
6
P NMR (400 MHz, acetone-d , 25 °C) of NCMe enriched
2
2
compound: δ 56.0 (dd, 1P, JP-P ) 47.0 Hz, JP-N ) 3.0 Hz, P
minor complex), 60.4 (dd, 1P, JP-P ) 45.0 Hz, JP-N ) 2.7 Hz, P
major complex), 64.1 (dd, 1P, JP-P ) 45.0 Hz, JP-N ) 2.7 Hz, P
major complex), 66.4 (dd, 1P, JP-P ) 47.0 Hz, JP-N ) 3.0 Hz, P
1
of
of
of
of
2
2
1
2
2
2
2
2
2
Florisil plug with Et
2
O (∼10 mL) to remove the catalyst. The Et
2
O
minor complex).
was removed under reduced pressure and the clear-colorless liquid
1
products were analyzed by H NMR. The enantiomeric excesses were
[
Ru((R)-BINAP)(MeCN)(OC(H)(CO
CH ))](BF
in THF. FT-IR (CH
2
C(CH
3
)
3
)(C(CH
3
)
2
(CO
2
C-
determined by either 1H NMR, with added shift reagent (tris[3-
(heptafluoropropylhydroxy-methylene)-(+)-camphorato]europium(I-
II)) and its comparison to that of the racemic alcohol and added shift
reagent (used for 11 and 12),24 or by chiral GC analysis (13). The
absolute configuration of the major enantiomeric product 11 was
determined by comparison to the reported optical rotation of (S)-11
(
3
)
3
4
) (10). 10 was synthesized in the same manner as (5/6)
-
1
-1
2
Cl
2
, 22 °C): 1608 cm (w, νCdO), 1644 cm (s,
+
ν
NP
CdO). ESI-MS (pos) m/z 1038.3 (M) , exact mass calcd for C60
RuBF , 1038.298. Anal. Calcd for C60H O NP RuBF : C, 64.06;
2 4 60 5 2 4
H
60
O
5
-
25
H, 5.38; N, 1.25. Found: C, 63.32; H, 5.44; N, 1.54. 10 (absolute
1
configuration (R)): H NMR (400 MHz, THF-d
H, C(CH ), 1.07 (s, 3H, C(CH ), 1.17 (s, 9H, R-CO
s, 9H, â-CO C(CH ), 1.92 (s, 3H, CH CN), 3.98 (s, 1H, Ru-O-
CH), 6.5-8.04 (m, 32H, BINAP). C NMR (400 MHz, THF-d , 25
C): δ 3.78 (1C, CH CN), 22.24 (1C, C(CH ), 24.97 (1C, C(CH
), 28.78 (3C, OC(CH ), 48.20 (1C, C(CH
), 88.23 (1C, OC(CH
8
, 25 °C): δ 0.86 (s,
26
(
[R]
enantiomeric product for 12 and 13 was determined by reduction of
the products with LiAlH to the corresponding triol, 2,2-dimethylbutan-
,3,4-triol, and comparison to the reported optical rotation of (3R)-
D
+33°, c 1.42, CHCl
3
). The absolute configuration of the major
3
(
)
3 2
)
3 2
2
C(CH ), 1.50
3 3
)
2
3
)
3
3
1
3
4
8
1
(
°
2
8
3
3
)
)
3 3
2
3
)
)
2
),
),
23 1
-)-2,2-dimethylbutan-1,3,4-triol ([R]
D
-16°, c 1.06, EtOH). H NMR
8.30 (3C, OC(CH
7.32 (1C, OC(CH
3
)
3
3
2
of 11 (400.1 MHz, CDCl
3
3 2
, 25 °C): δ 1.12 (s, 3H, C(CH ) ), 1.21 (s,
3
)
3
3 3
)
), 89.48 (1C, Ru-O-
C-
2 3 3 8
), 190.20 (R-C CO C(CH ) ). P NMR (400 MHz, THF-d , 25
3
H, C(CH
)
3 2
), 3.23 (br d, 1H, 3JH-H ∼ 6 Hz, C(H)(OH)), 3.66 (s, 3H,
CH), 126.5-142.6 (42C, aromatics, BINAP), 183.41 (1C, â-CO
(
°
Hz). P NMR (400 MHz, acetone-d
compound: δ 59.5 (dd, 1P, P
(
2
3
3
1
OCH
3
), 3.72 (s, 3H, OCH
3
), 4.30 (br d, 1H, JH-H ∼ 6 Hz, C(H)(OH)).
CH
3
)
3
1
2
, 2JP-P ) 45.5
, 25 °C) of NCMe enriched
, JP-P ) 45.0 Hz, JP-N ) 2.8 Hz), 63.5
H NMR of 12 (400.1 MHz, CDCl
3 3 2
, 25 °C): δ 1.12 (s, 3H, C(CH ) ),
C): δ 59.5 (d, 1P, P
1
, JP-P ) 45.5 Hz), 63.5 (d, 1P, P
2
3
3
1
15
1.22 (s, 3H, C(CH ), 1.23 (d, 6H, JH-H ) 6.5 Hz, CO
(H))), 1.24 (d, 6H, JH-H ) 6.5 Hz, CO
H-H ) 6.5 Hz, C(H)(OH)), 4.30 (d, 1H, C(H)(OH)), 4.99 (quint, 1H,
(H))), 5.08 (quint, 1H, JH-H ) 6.5 Hz,
(H))). H NMR of 13 (400.1 MHz, CDCl , 25 °C): δ
.14 (s, 3H, C(CH ), 1.68 (s, 3H, C(CH ), 1.44 (s, 9H, CO C(CH ),
.48 (s, 9H, CO C(CH ), 3.19 (d, 1H, C(H)(OH)), 4.17 (d, C(H)-
3
3
)
2
2
C((CH
(H))), 3.19 (d, 1H,
3 2
) -
6
2
2
2
C((CH
)
3 2
1
3
2
2
J
J
dd, 1P, P
2
, JP-P ) 45.0 Hz, JP-N ) 2.8 Hz). For details on the
3
3
H-H ) 6.5 Hz, CO
2
1
C((CH
3
)
2
assignment of BINAP signals refer to the Results and Discussion
section.
CO
2
C((CH
)
3 2
3
1
1
(
)
3 2
)
3 2
2
3 3
)
Low-Temperature NMR Investigation of Catalyst and Substrate
2
3 3
)
-
5
Interactions. Compound 7 (19.5 mg, 2.03 × 10 mol) was partially
OH)).
Stoichiometric Hydrogenolysis of [Ru((R)-BINAP)(MeCN)(OC-
D)(CO R)(C(CH (CO R))](BF ) (5-d /6-d , 8-d /9-d , and 10-d ).
and 6-d (105.0 mg,
.01 × 10 mol) was transferred to a 100 mL sidearm. The flask was
dissolved in THF-d (0.6 mL) in an NMR tube under an argon
8
atmosphere. At room temperature, the tube was flushed with dihydrogen
gas, pressurized (1-2 atm), and shaken until a golden orange solution
was generated (∼5 min). The dihydrogen atmosphere was replaced by
(
2
3
)
2
2
4
1
1
1
1
1
In the glovebox, the mixture of complexes 5-d
1
1
-
4
1
1
31
argon and the resulting solution was analyzed by H and P NMR
spectroscopy at -80 °C. NMR spectroscopic analysis indicated a
mixture of two ruthenium-hydrido species ([Ru((R)-BINAP)(H)(THF-
sealed and the solvent (MeOH or THF, 48.6 mL) added via gastight
syringe. The sidearm was placed in a preflushed stainless steel bomb
and the bomb was further flushed for 15 min with argon gas and with
dihydrogen gas for 10 min, and finally pressurized to 50 atm. The bomb
was allowed to react for 1 h (the time, on average, for a single turnover
in the operating catalytic reaction) in a 50 °C oil bath. Once the bomb
was depressurized, the flask was removed and bubbled with oxygen to
destroy the catalyst. The flask was then placed under reduced pressure
to remove the solvent. The residue was passed through a Florisil plug
8 n 4
d ) (MeCN)3-n](BF ) (n ) 0-3, with n ) 2 as major species (75%)))
-
4
and cyclooctane were present. At -80 °C, 2 (4.0 mg, 1.15 × 10
mol) was injected into the NMR tube via gastight syringe. The tube
was removed from the cooling bath, shaken for ∼15 s, and then
31
immediately placed in a precooled (-80 °C) NMR probe. The P NMR
1
spectrum at -80 °C remained unchanged, as did the H NMR spectrum
except for the introduction of 2. Upon warming the NMR probe, the
with Et
2
O (5 mL) to remove the catalyst, the solvent removed under
reduced pressure, and the product analyzed by NMR. H NMR of THF
31
1
P and H NMR remained unchanged until the temperature reached
1
3
1
-
30 °C. At -30 °C, the P NMR slowly showed the signs of 5 and
6
peaks growing in and no other peaks were observed. Once warmed
(
24) 11: The ratio of the methoxy signals (ca. δ 4.0) was used to determine the
ee. The ratio of these peaks was 1:1 for racemic alcohol. 12: The ratio of
the backbone methyl signals (ca. δ 1.2 and 1.1) was used to determine the
ee. The ratio of these peaks was 1:1 for racemic alcohol.
3
1
to ambient temperature, the P NMR showed only the presence of 5
and 6 in the same ratio observed with the stoichiometric reaction at
room temperature. P (162.0 MHz, THF-d , -80 to -40 °C): δ 72.2
3
1
8
(25) Performed on a Beta Dex 120 column in acetone solution. Initial oven
temperature was 120 °C for 45 min then it was increased at a rate of 10
2
2
2
(d, JP-P ) 42.5 Hz, A), 72.4 (d, JP-P ) 49.5 Hz, B), 77.1 (d, JP-P
)
°
(
C/min to 200 °C. The enantiomers were eluted at ∼ 45 min for
S)-enantiomer and at ∼46 min for (R)-enantiomer.
(26) Seebach, D.; Wasmuth, D. HelV. Chim. Acta 1980, 63, 197.
2
4
3.5 Hz, A), 81.4 (d, JP-P ) 49.5 Hz, B). Approximate percentages
3
1
of hydrido species present: A (25%) and B (75%). P (162.0 MHz,
3
690 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002
9

Identification of a Catalyst−Substrate (Alkoxide) Species
A R T I C L E S
reaction (400.1 MHz, CDCl
3
, 25 °C): δ 4.3 (0.4 H as compared to 3H
solution was stirred under argon atmosphere for 2 min. To this solution
methoxy signals at ca. δ 3.8). The ee was determined to be 58% (R).
was added 1 equiv of HBF
immediate color change from a deep orange to a pale yellow. The
solution was stirred for 5 min then the flask was placed under reduced
pressure to remove the solvent. The residue was analyzed by H, H,
and 31P NMR and determined to have produced the desired alcohol
product, with complete retention of deuterium at the alkoxy position,
4
2
‚OEt via gastight syringe, resulting in an
1
H NMR of MeOH reaction (400.1 MHz, CDCl
3
, 25 °C): δ 4.3 (0.25
H as compared to 3H methoxy signals at ca. δ 3.8). The ee was
determined to be 60% (R). The stoichiometric hydrogenations of the
1
2
complexes 8-d
1
/9-d
H NMR of THF reaction of 8-d
.3 (0.2 H as compared to 3H methoxy signals at ca. δ 3.8). The ee
1
and 10-d
1
were performed in a similar manner.
1
1
/9-d (400.1 MHz, CDCl , 25 °C): δ
1
3
4
and the [Ru((R)-BINAP)(MeCN)
passed through a Florisil plug with Et
4
](BF
4 2
) byproduct. The sample was
1
was determined to be 69% (R). H NMR of MeOH reaction of 8-d
-d (400.1 MHz, CDCl , 25 °C): δ 4.3 (0.15 H as compared to 3H
methoxy signals at ca. δ 3.8). The ee was determined to be 70% (R).
1
/
2
O and the solvent removed, and
9
1
3
the product residue 11-d was analyzed for the enantiomeric excess
1
(58% ee (R)). Similar reactions were performed with the ruthenium-
alkoxides 8-d /9-d and with the ruthenium-alkoxide 10. Reaction with
1
H NMR of THF reaction of 10-d
1 3
(400.1 MHz, CDCl , 25 °C): δ 4.2
1
1
(
0.2 H as compared to 3H methoxy signals at ca. δ 3.8). The ee was
8-d /9-d yielded 79% ee (R), and reaction with 10-d yielded >99%
1
1
1
1
determined to be 84% (R). H NMR of MeOH reaction of 10-d
1
(400.1
ee (R).
MHz, CDCl , 25 °C): δ 4.3 (0.1 H as compared to 3 H methoxy signals
3
at ca. δ 3.8). The ee was determined to be 84% (R).
Reaction of [Ru((R)-BINAP)(MeCN)(OC(D)(CO
Acknowledgment. We sincerely appreciate the expert as-
sistance of Mr. G. Bigam and Dr. T. Nakashima of the
University of Alberta High Field NMR Laboratory. We also
thank the Natural Sciences and Engineering Research Council
of Canada and the University of Alberta for financial support.
2
R)(C(CH
3 2
) -
(
CO
2
R)))](BF
4
) (5-d
1
/6-d
1
, 8-d
1
/9-d
1
, and 10-d
/6-d
1
) with HBF
) (150.4 mg, 1.44 × 10
4
‚OEt . The
2
-
4
ruthenium-alkoxide complexes (5-d
mol) were transferred to a 50 mL sidearm. The flask was sealed, placed
1
1
under reduced pressure, and refilled (×3) to remove all traces of oxygen.
To the flask was added CH
2
Cl
2
(5.0 mL) via gastight syringe. The
JA0102991
J. AM. CHEM. SOC.
9
VOL. 124, NO. 14, 2002 3691