Hydrogen transfer to carbonyls and imines from a hydroxycyclopentadienyl ruthenium hydride: Evidence for concerted hydride and proton transfer

Article abstract of DOI:10.1021/ja002177z

Reaction of { [2,5-Ph₂-3,4-Tol₂(η⁵-C ₄CO)]₂H}Ru₂(CO)₄(μ-H) (6) with H₂ formed [2,5-Ph₂-3,4-Tol₂(η⁵-C ₄COH)Ru(CO)₂H] (8), the active species in catalytic carbonyl reductions developed by Shvo. Kinetic studies of the reduction of PhCHO by 8 in THF at - 10 °C showed second-order kinetics with ΔH? = 12.0 kcal mol^-1 and ΔS? = -28 eu. The rate of reduction was not accelerated by CF₃CO₂H, and was not inhibited by CO. Selective deuteration of the RuH and OH positions in 8 gave individual kinetic isotope effects k_RuH/k_RuD = 1.5 ± 0.2 and k_OH/k_OD = 2.2 ± 0.1 for PhCHO reduction at 0 °C. Simultaneous deuteration of both positions in 8 gave a combined kinetic isotope effect of k_ORuH/k_ODRuD = 3.6 ± 0.3. [2,5-Ph₂-3,4-Tol₂(η⁵-C ₄COSiEt₃) Ru(CO)₂H] (12) and NEt₄⁺[2,5-Ph₂-3,4-Tol ₂(η⁴-C₄CO) Ru(CO)₂H]^- (13) were unreactive toward PhCHO under conditions where facile PhCHO reduction by 8 occurred. PhCOMe was reduced by 8 30 times slower than PhCHO; MeN=CHPh was reduced by 8 26 times faster than PhCHO. Cyclohexene was reduced to cyclohexane by 8 at 80 °C only in the presence of H₂. Concerted transfer of a proton from OH and hydride from Ru of 8 to carbonyls and imines is proposed.

Full text of DOI:10.1021/ja002177z

1090
J. Am. Chem. Soc. 2001, 123, 1090-1100
Hydrogen Transfer to Carbonyls and Imines from a
Hydroxycyclopentadienyl Ruthenium Hydride: Evidence for
Concerted Hydride and Proton Transfer
Charles P. Casey,* Steven W. Singer, Douglas R. Powell, Randy K. Hayashi, and
Michael Kavana
Contribution from the Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706
ReceiVed June 19, 2000. ReVised Manuscript ReceiVed NoVember 29, 2000
Abstract: Reaction of {[2,5-Ph2-3,4-Tol2(η5-C4CO)]2H}Ru2(CO)4(µ-H) (6) with H2 formed [2,5-Ph2-3,4-Tol2(η5-
C4COH)Ru(CO)2H] (8), the active species in catalytic carbonyl reductions developed by Shvo. Kinetic studies
of the reduction of PhCHO by 8 in THF at -10 °C showed second-order kinetics with ∆H‡ ) 12.0 kcal mol-1
and ∆S‡ ) -28 eu. The rate of reduction was not accelerated by CF3CO2H, and was not inhibited by CO.
Selective deuteration of the RuH and OH positions in 8 gave individual kinetic isotope effects kRuH/kRuD ) 1.5
( 0.2 and kOH/kOD ) 2.2 ( 0.1 for PhCHO reduction at 0 °C. Simultaneous deuteration of both positions in
8 gave a combined kinetic isotope effect of kOHRuH/kODRuD ) 3.6 ( 0.3. [2,5-Ph2-3,4-Tol2(η5-C4COSiEt3)Ru-
(CO)2H] (12) and NEt4+[2,5-Ph2-3,4-Tol2(η4-C4CO)Ru(CO)2H]- (13) were unreactive toward PhCHO under
conditions where facile PhCHO reduction by 8 occurred. PhCOMe was reduced by 8 30 times slower than
PhCHO; MeNdCHPh was reduced by 8 26 times faster than PhCHO. Cyclohexene was reduced to cyclohexane
by 8 at 80 °C only in the presence of H2. Concerted transfer of a proton from OH and hydride from Ru of 8
to carbonyls and imines is proposed.
Introduction
Scheme 1
Catalytic reduction of polar functional groups mediated by
transition metal complexes has emerged as a complement to
stoichiometric reduction by metal hydrides such as LiAlH4 and
NaBH4.1 Interest has been spurred by the development of
complexes that catalyze the asymmetric reduction of ketones
and imines. Chiral arene-Ru(II)-TsDPEN (N-(p-toluenesulfonyl)-
1,2-diphenylethylenediamine) (1) catalysts reported by Noyori
have achieved high levels of enantioselectivity in asymmetric
transfer hydrogenations with 2-propanol as the reductant
(Scheme 1).2 Recent work by Mashima and Ikariya has extended
Noyori’s design to isoelectronic Cp*Rh(III) and Cp*Ir(III)-
TsDPEN and TsCYDN (N-p-toluenesulfonyl-1,2-cyclohexadi-
amine) catalysts.3 In these systems, the active catalysts are
proposed to be metal hydrides containing an acidic -NH proton
that promotes enantioselective hydrogen transfer. In the cymene-
Ru and Cp*Ir cases, these active species have been isolated
and characterized. In addition, a number of groups have reported
enantioselective reduction catalysts that are enhanced by an
“-NH effect”, the presence of acidic amine hydrogen that is
thought to enable the active hydride catalyst to achieve higher
levels of selectivity and efficiency.4
that may function as powerful catalysts for the reduction of
unsaturated organic compounds.5 We became intrigued by
Shvo’s ruthenium reduction catalyst system in which a hydroxy-
cyclopentadienyl hydride 2, which contains electronically
coupled acidic and hydridic hydrogens, has been proposed as
the active hydrogen transfer species.6 Shvo’s catalyst system
employs the diruthenium complex 3 and either H2 or HCO2H
for the reduction of aldehydes and ketones at high temperature.
3 also catalyzed the hydrogenation of alkenes and alkynes by
H2 but not by HCO2H. Shvo has proposed that 3 dissociates
into 2 and coordinatively unsaturated dienone dicarbonyl A.
Unsaturated intermediate A is proposed to react with H2 or
HCO2H to form 2. Shvo converted 3 to 2 at 105 °C under 35
atm of H2 or at 65 °C in the presence of excess HCO2H and
showed that 2 reduces cyclohexanone to cyclohexanol at room
temperature.6,7
We have had a long-standing interest in studying organome-
tallic complexes that have acidic and hydridic hydrogens and
(1) Fehring, V.; Selke, R. Angew. Chem., Int. Ed. Engl. 1998, 37, 1827.
(2) (a) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97. (b)
Haack, K.-J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 285. (c) Matsumura, K.; Hashiguchi, S.; Ikariya,
T.; Noyori, R J. Am. Chem. Soc. 1997, 119, 8738.
(3) (a) Mashima, K.; Abe, T.; Tani, K. Chem. Lett. 1998, 1199. (b)
Mashima, K.; Abe, T.; Tani, K. Chem. Lett. 1998, 1201. (c) Murata, K.;
Ikariya, T.; Noyori, R. J. Org. Chem. 1999, 64, 2186.
In recent work with the Shvo catalyst, Ba¨ckvall has used 3
for the dynamic kinetic resolution of secondary alcohols. The
(5) (a) Casey, C. P.; Wang, Y.; Tanke, R. S.; Hazin, P. A.; Rutter, E.
W., Jr. New J. Chem. 1994, 18, 43. (b) Casey, C. P. J. Organomet. Chem.
1990, 400, 205.
(6) Shvo, Y.; Czarkie, D.; Rahamim, Y. J. Am. Chem. Soc. 1986, 108,
7400.
(7) (a) Menashe, N.; Shvo, Y. Organometallics 1991, 10, 3885. (b)
Menashe, N.; Salant, E.; Shvo, Y. J. Organomet. Chem. 1996, 514, 97.
(4) For other references to transfer hydrogenation catalysts, see ref 3 in:
Alonso, D. A.; Brandt, P.; Nordin, S. J. M.; Andersson, P. G. J. Am. Chem.
Soc. 1999, 121, 9580.
10.1021/ja002177z CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/20/2001

EVidence for Concerted Hydride and Proton Transfer
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1091
Scheme 2
Scheme 5
Scheme 6
Scheme 3
proton transfer, CO recoordination, and alcohol dissociation
complete the sequence.
A third mechanism involves η5-η3 ring slippage of 2
concurrent with aldehyde complexation and subsequent in-
tramolecular hydride addition (Scheme 5).10 Shvo has proposed
that the hydrogenation of alkenes and alkynes proceeds through
a ring-slip mechanism.11
Scheme 4
Here we describe detailed kinetic studies of the reduction of
aldehydes, ketones, and imines by a tolyl-substituted analogue
of 2. These studies provide persuasive evidence that these
reductions proceed by concerted hydrogen transfer promoted
by the moderately acidic CpOH proton.
Results
Synthesis of Tolyl-Substituted Ruthenium Complexes for
Mechanistic Investigations. A series of ruthenium complexes
were prepared from 2,5-diphenyl-3,4-(di-p-tolyl)cyclopentadi-
enone (4) for use in detailed mechanistic studies. The 1H NMR
tolyl methyl singlets provide a convenient signature to distin-
guish the various ruthenium complexes.
resolution exploits the ability of 3 to reversibly dehydrogenate
alcohols in the presence of a lipase.8 The lipase selectively
acylates one alcohol enantiomer while 3 racemizes the other
enantiomer.
The broad scope of the Shvo catalyst and its similarity to the
growing family of H+/H- reduction catalysts prompted us to
initiate a detailed mechanistic study of the reactivity of 2 with
carbonyls. At the beginning of our work, we considered three
possible mechanisms for hydrogen transfer. First, simultaneous
transfer of H+ from OH and H- from RuH might occur outside
the coordination sphere of the metal (Scheme 3). Noyori has
proposed a similar mechanism for reduction of ketones by the
Ru(II)-TsDPEN system.3a
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru(CO)3 (5) was synthesized
in 80% yield by heating Ru3(CO)12 and 2,5-diphenyl-3,4-(di-
p-tolyl)cyclopentadienone (4) in refluxing toluene (Scheme 6).
Dienone tricarbonyl complex 5 had spectroscopic features very
similar to those reported for the tetraphenyl analogue.12 The
1H NMR resonance for the tolyl methyl group was observed at
δ 1.82.
{[2,5-Ph2-3,4-Tol2(η5-C4CO)]2H}Ru2(CO)4(µ-H) (6). By
using a procedure similar to that reported by Shvo for 3,
treatment of cyclopentadienone ruthenium complex 5 with
(9) Gladysz has studied related σ- and π-aldehyde rhenium complexes.
(a) Garner, C. M.; Me´ndez, N. Q.; Kowalczyk, J. J.; Ferna´ndez, J. M.;
Emerson, K.; Larsen, R. D.; Gladysz, J. A. J. Am. Chem. Soc. 1990, 112,
5146. (b) Me´ndez, N. Q.; Mayne, C. L.; Gladysz, J. A. Angew. Chem., Int.
Ed. Engl. 1990, 29, 1475. (c) Me´ndez, N. Q.; Arif, A. M.; Gladysz, J. A.
Angew. Chem., Int. Ed. Engl. 1990, 29, 1473.
(10) The hydroxycyclopentadienyl ligand can ring slip in three directions,
but only the symmetric ring slip is shown.
(11) Shvo, Y.; Goldberg, I.; Czerkie, D.; Reshef, D.; Stein, Z. Organo-
metallics 1997, 16, 133.
A second possible mechanism involves CO dissociation from
2 to form the unsaturated Ru monocarbonyl hydride B (Scheme
4). Coordination of an aldehyde followed by intramolecular
hydride addition forms an alkoxide complex.9 Intramolecular
(8) (a) Larsson, A. L. E.; Persson, B. A.; Ba¨ckvall, J. E. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1211. (b) Persson, B. A.; Larsson, A. L. E.; Le
Ray, M.; Ba¨ckvall, J. E. J. Am. Chem. Soc. 1999, 121, 1645. (c) Persson,
B. A.; Huerta, F. F.; Ba¨ckvall, J. E. J. Org. Chem. 1999, 64, 5237. (d)
Huerta, F. F.; Laxmi, Y. R. S.; Ba¨ckvall, J. E. Org. Lett. 2000, 2, 1037. (e)
Laxmi, Y. R. S.; Ba¨ckvall, J. E. Chem. Commun. 2000, 611.
(12) (a) Bruce, M. I.; Knight, J. R. J. Organomet. Chem. 1968, 12, 411.
(b) Bailey, N. A.; Jassal, V. S.; Vefghi, R.; White, C. J. Chem. Soc., Dalton
Trans. 1987, 2815.

1092 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
Casey et al.
Scheme 7
Scheme 8
Figure 1. X-ray Crystal Structure of {[2,5-Ph2-3,4-Tol2(η5-C4CO)]2H}-
Ru2(CO)4(µ-H) (6). Thermal ellipsoids are drawn at 50% probability.
Na2CO3 in acetone gave a 50% yield of diruthenium complex
6. We developed a more convenient and more efficient one-
step procedure for the synthesis of 6. Refluxing Ru3(CO)12 and
cyclopentadienone 4 in CH3OH for 40 h produced a 59% yield
1
of 6 as a pure yellow precipitate (Scheme 6). In the H NMR
1
The H and 13C NMR spectra of 8 were consistent with its
formulation as a hydroxycyclopentadienyl ruthenium hydride.
In the 1H NMR spectrum, a hydride resonance was observed at
δ -9.76 and a broad singlet assigned to the hydroxyl proton
was observed at δ 8.60. A single tolyl methyl resonance was
observed at δ 2.17, indicating 8 had formed without any
byproducts. In the 13C NMR spectrum of 8, the cyclopentadienyl
C-O resonance appeared at δ 137 and the RuCO resonance
appeared δ 203.16
spectrum of 6 in C6D6, the bridging hydride resonance appeared
at δ -17.68 and the bridging hydroxyl resonance appeared as
a broad singlet at δ 12.90.6 Addition of D2O led to a greatly
diminished intensity of the δ 12.90 resonance due to OH/OD
exchange.
In the X-ray crystal structure of 6 (Figure 1), the Ru(1)-
C(1) bond length of 2.40 Å and the C(1)-O(1) bond length of
1.29 Å provide evidence for η5-coordination of the cyclopen-
tadienyl ligands.13 The two Ru(CO)2 units are twisted ap-
proximately 45 °C relative to one another and the eight aryl
groups are arranged in a propeller orientation.
The equilibrium constant for formation of 8 from H2 was
much less favorable in hydrocarbon solvents than in THF.
Essentially complete conversion to 8 was achieved under only
1 atm of H2 in THF-d8. However, when 6 was heated under 1
atm of H2 at 80 °C in C6D6, incomplete conversion to 8 was
seen and a 75:25 equilibrium mixture of 6:8 was formed after
72 h (Keq ) [8]2/[6][H2] ≈ 0.02 atm-1).17 Solutions of 8 in
toluene-d8, CD3CN, and DMSO-d6 were conveniently prepared
from THF solutions of 8 by removing THF under high vacuum
and dissolving the resulting oily solid in the solvent of choice.18
The reaction of diruthenium hydride 6 with HCO2H produced
solutions of 8 containing excess formic acid. When a 10-fold
excess of 96% HCO2H was added to a THF-d8 solution of 6,
no reaction was observed at room temperature. However, upon
heating at 80 °C, vigorous evolution of gas occurred. After 10
min, the evolution of gas had ceased and 1H NMR showed that
6 had been completely converted to 8.
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru2(CO)2]2 (7). Prolonged heat-
ing of Ru3(CO)12 and cyclopentadienone 4 in heptane for 48 h
led to the precipitation of ruthenium dimer 7 as a mustard yellow
powder in 74% yield (Scheme 6).14 The tolyl methyl 1H NMR
resonances made it easy to distinguish dimer 7 (δ 1.69) from
diruthenium hydride complex 6 (δ 1.79). The X-ray crystal
structure of 7 (see Supporting Information) confirmed its dimeric
structure.15
[2,5-Ph2-3,4-Tol2(η5-C4COH)Ru(CO)2H] (8). Reaction of
either H2 or formic acid with diruthenium complex 6 in THF-
d8 produced the monomeric hydroxycyclopentadienyl ruthenium
hydride 8 (Scheme 7). A resealable NMR tube containing a
solution of 6 in THF-d8 under 1 atm of H2 at -78 °C was sealed
and heated at 80 °C. After 2 h, the hydride resonance for 6 at
δ -18.23 had decreased in intensity and a new hydride
resonance assigned to 8 appeared at δ -9.76. The 1:2 ratio of
these resonances indicated 50% of 6 had been converted to 8.
After 8 h, equilibrium was reached and a 98:2 mixture of 8:6
was observed (Keq ) [8]2/[6][H2] ≈ 1 atm-1).
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru(CO)2PPh3 (9) and [2,5-
Ph2-3,4-Tol2(η4-C4CO)]Ru(CO)2NH(CH3)(CH2Ph)(10) were
synthesized by ligand addition to ruthenium dimer 7 at room
temperature (Scheme 8). Addition of PPh3 to a CH2Cl2 solution
of 7 led to the isolation of phosphine adduct 9 as a pale yel-
low powder in 74% yield.19 Similarly, addition of NH-
(CH3)(CH2Ph) to a CH2Cl2 solution of 7 gave a 79% yield of
amine adduct 10 as a dark green solid.20
(13) In contrast, (η4-C4Ph4CO)Ru(CO)3 exhibits η4-coordination with a
Ru(1)-C(1) bond distance of 2.53 Å and a C(1)-O(1) bond length of 1.22
Å. Blum, Y.; Shvo, Y.; Chodosh, D. F. Inorg. Chim. Acta 1985, 97, L25.
(14) 7 was identified by comparing its spectroscopic data to that of
analogous complexes. Mays, M. J.; Morris, M. J.; Raithby, P. R.; Shvo,
Y.; Czarkie, D. Organometallics 1989, 8, 1162.
[2,5-Ph2-3,4-Tol2(η5-C4COH)]Ru(CO)2(O2CCF3) (11). Ad-
dition of excess CF3CO2H to a CH2Cl2 solution of ruthenium
(15) The structure of 7 is similar to that obtained for the tetraphenyl
analogue by Shvo and Mays.14 They described the dimer as having η4-
cyclopentadienyl ligands with ketonic carbonyls bound to the second Ru.
However, we believe that both compounds are better described as having
the η5-cyclopentadienyl rings with alkoxy oxygens bound to the second
ruthenium. The 2.40 Å Ru(1)-C(1) and the 1.28 Å C(1)-O(1) bond lengths
of 7 are very similar to the related bond lengths in diruthenium hydroxy-
cyclopentadienyl ruthenium hydride 6 and support the formulation of 7 as
an η5-alkoxycyclopentadienyl complex.
(16) The spectroscopic data closely match the reported data for the
tetraphenyl analogue.6
(17) Shvo and Mays attributed the thermodynamic stability of 2 in THF
to intermolecular hydrogen bonding between the hydroxy group and THF,
which prevents loss of H2 and reversion to 3.14
(18) Solutions prepared this way contained 0.5-1.0 equiv of THF/Ru.
(19) A crystal structure of the tetraphenyl analogue was recently
published. Yamazaki, S.; Taira, Z. J. Organomet. Chem. 1999, 578, 61.

EVidence for Concerted Hydride and Proton Transfer
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1093
Scheme 9
Figure 2. X-ray Crystal Structure of NEt4+[2,5-Ph2-3,4-Tol2(η4-C4-
CO)Ru(CO)2H]- (13). Thermal ellipsoids are drawn at 50% probability.
The NEt4+ cation is omitted for clarity.
dimer 7 followed by addition of pentane led to the isolation of
the hydroxycyclopentadienyl trifluoroacetate complex 11 as a
light green powder in 85% yield (Scheme 8). The structure of
11 was confirmed by X-ray crystallography (see Supporting
Information).
Scheme 10
[2,5-Ph2-3,4-Tol2(η5-C4COSiEt3)Ru(CO)2H] (12) was pre-
pared by addition of SiEt3H to a CH2Cl2 solution of ruthenium
dimer 7 (Scheme 8). The formation of 12 can be rationalized
as an oxidative addition of Et3SiH to a monomeric cyclopen-
Scheme 11
1
tadienone ruthenium dicarbonyl intermediate. The H NMR
spectrum (C6D6) showed a single tolyl CH3 resonance at δ 1.81
and a hydride resonance at δ -9.20.
NEt4+[2,5-Ph2-3,4-Tol2(η4-C4CO)Ru(CO)2H]- (13). The
anionic alkoxycyclopentadienyl hydride 13 was synthesized to
compare its reactivity toward aldehydes with that of the
hydroxycylopentadienyl hydride 8. The most convenient pre-
parative route to pure 13 was addition of a THF solution of
KBEt3H to a suspension of NEt4Br and 5 in THF at -78 °C
(Scheme 9). After the mixture was warmed to room temperature
for 20 min and aqueous NaOH workup to destroy excess
borohydride, 13 was isolated in 77% yield as an off-white
powder. The 1H NMR spectrum of 13 exhibited a hydride
resonance at δ -9.68 and a single tolyl methyl resonance at δ
2.15.
When a solution of 8 in THF-d8 was heated at 80 °C under
1.7 atm CO, an equilibrium between 8 and dienone ruthenium
tricarbonyl complex 5 was slowly established (Scheme 10).
After 72 h, an equilibrium 87:13 mixture of 5(δ 2.21):22 8(δ
2.16) was observed (Keq ) [5 ) 0.0226 mmol][H2 ) 0.0226
mmol]/[8 ) 0.0034 mmol][CO ) 0.1 mmol] ) 1.5). The
equilibrium was also approached from the opposite direction.
When a solution of 5 in THF-d8 was heated at 80 °C under 1.7
atm of H2 for 72 h, a 30:70 equilibrium mixture of 5 and 8 was
established (Keq ) [5 ) 0.0057 mmol][H2 ) 0.1 mmol]/[8 )
0.0133 mmol][CO ) 0.0133 mmol] ) 3.2). In both experiments
the formation of diruthenium complex 6 was not detected.
The intermediacy of formyl complex 14 in the synthesis of
1
13 was demonstrated by H NMR spectroscopy (Scheme 9).
After addition of KBEt3H to a THF-d8 solution of tricarbonyl
5 at -78 °C and warming to -30 °C, the 1H NMR spectrum of
the golden yellow solution exhibited a single tolyl methyl
resonance at δ 2.17 and a singlet at δ 12.70 characteristic of a
metal formyl group.21 Formyl complex 14 underwent decarbo-
nylation to K+[2,5-Ph2-3,4-Tol2(η5-C4CO)Ru(CO)2H]- with t1/2
) 15 min at 23 °C.
The X-ray crystal structure of 13 (Figure 2) supports
η4-coordination of the cyclopentadienone ligand to the Ru center.
The long Ru(1)-C(1) distance of 2.45 Å, the short C(1)-O(1)
bond length of 1.244 Å, and the 12.7° tilt of the C(1)-O(1)
bond out of the plane defined by C(2)-C(5) provide evidence
for η4-coordination.
Stability of Hydroxycyclopentadienyl Ruthenium Hydride
8. Solutions of hydroxycyclopentadienyl ruthenium hydride 8
were stable under H2, but decomposed under N2 or in the
presence of air. In THF-d8 under N2 at room temperature, the
rate of reversion to the diruthenium complex 6 and H2 was slow
(t1/2 ≈ 1-2 days).
Exchange of 13CO with hydroxycyclopentadienyl ruthenium
hydride 8 was studied to determine whether reversible CO
dissociation occurs. Fluorescent room light was found to
promote 13CO exchange. A THF-d8 solution of 8 under 1.5 atm
1
of 13CO exposed to room light was monitored by H NMR
spectroscopy. After 15 min, a three-line pattern was observed
at δ -9.76 consisting of a center line for unlabeled 8 flanked
by a doublet (2JCH ) 9.5 Hz) for monolabeled 8-13CO (4:1 )
8:8-13CO) (Scheme 11). After 1.5 h in room light, a five-line
pattern was observed due to the superposition of a central line
for unlabeled 8, a doublet for 8-13CO, and a triplet (2JCH ) 9.5
Hz) for double-labeled 8-(13CO)2 (1:1:0.2 ) 8:8-13CO:
8-(13CO)2). The 13CO exchange rate was slowed dramatically
when the experiment was repeated in an NMR tube kept in the
dark. After 1 h, less than 5% 13CO incorporation was seen. The
observation of facile CO exchange at room temperature requires
(21) Gladysz, J. A. AdV. Organomet. Chem. 1982, 20, 1.
(22) The reaction was stopped at that point, and the THF-d8 was removed
in vacuo and replaced with C6D6. The 1H NMR spectrum in C6D6 confirmed
that the resonance at δ 2.21 in THF-d8 corresponded to a resonance at δ
1.82 in C6D6, the tolyl resonance for tricarbonyl 5.
(20) Similar amine complexes have been synthesized. (a) Abed, M.;
Goldberg, I.; Stein, Z.; Shvo, Y. Organometallics 1988, 7, 2054. (b) Shvo,
Y.; Abed, M.; Laine, Blum, Y.; Laine, R. M. Isr. J. Chem. 1986, 27, 267.

1094 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
Casey et al.
serious consideration of photochemically induced CO dissocia-
tion as a possible step in reactions of 8.
Scheme 12
Acidity of Hydroxyl Proton in Hydroxycyclopentadienyl
Hydride 8. Knowledge of the acidity23 of the hydroxy proton
of 8 is crucial in evaluating the role of proton transfer in the
reduction of PhCHO by 8. CH3CN was chosen as the solvent
for determination of the pKa of 8 because it allows compari-
son with a broad spectrum of organic acids.24 Norton’s IR
method for determination of the pKa of metal carbonyl hydrides
in CH3CN was readily adapted to the study of 8.25
When NEt3 (0.072 M, pKa ) 18.5 in CH3CN) was added to
a 0.070 M CH3CN solution of 8, IR bands due to 8 at 2014 and
1954 cm-1 decreased in intensity and new bands at 1985 and
1920 cm-1 assigned to NEt3H+[2,5-Ph2-3,4-Tol2(η5-C4CO)Ru-
(CO)2H]- (15) grew in. A 1:3.6 ratio of 8:15 was determined
by comparing the absorbances with independently obtained
molar extinction coefficients. The pKa of 8 was calculated to
be 17.5.26 When 1 equiv of pyridine (pKa ) 12.4 in CH3CN)
was added to a CH3CN solution of 8, no deprotonation was
observed. When 1 equiv of 1,1,3,3-tetramethylguanidine (pKa
) 23.3 in CH3CN) was added to a CH3CN solution of 8,
complete deprotonation to the guanidinium salt was observed.
The hydroxyl proton of 8 in CH3CN is 9 pKa units more acidic
than phenol (pKa ) 26.6 in CH3CN) and 3 pKa units more acidic
than benzoic acid (pKa ) 20.7 in CH3CN).27 The moderately
strong acidity of the hydroxyl proton must be considered as an
important factor in the reaction of 8 with PhCHO.
Equation 1 shows the second rate law for PhCHO reduction.
The second-order rate constant at -10 °C is k2 ) (3.0 ( 0.2)
× 10-4 M-1 s-1 28
.
d[PhCH2OH]
d[8]
dt
d[PhCHO]
-
) -2
) 2
)
dt
dt
d[6]
dt
2
) 2k2[8][PhCHO] (1)
The temperature dependence of the rate of reduction of
PhCHO by 8 was measured between 0 and -15 °C: k2 ) (7.7
( 0.5) × 10-4 M-1 s-1 at 0 °C, (5.9 ( 0.4) × 10-4 M-1 s-1 at
-5 °C, and (2.0 ( 0.3) × 10-4 M-1 s-1 at -15 °C. An Eyring
plot gave ∆H‡ ) 12.0 ( 1.5 kcal mol-1 and ∆S‡ ) -28 ( 5
eu. The high negative entropy of activation is consistent with
an associative reaction in which the transition state for hydrogen
transfer is highly ordered.
Kinetics of PhCHO Reduction by Hydroxycyclopentadi-
enyl Ruthenium Hydride 8. When an excess of PhCHO was
added to a THF-d8 solution of 8 at -78 °C and then warmed to
room temperature, complete disappearance of the hydride
resonance at δ -9.76 and the hydroxyl resonance at δ 8.60
1
was observed in the H NMR spectrum within 10 min. New
The rate of PhCHO reduction was not inhibited by CO
pressure. Essentially the same rate of hydride disappearance was
observed under N2 as under 1 atm of CO at 0 °C (kobs ) 1.13
× 10-3 s-1).29 This is consistent with the fact that PhCHO
reduction is faster than 13CO exchange in the dark.
resonances were observed at δ 4.57 corresponding to the CH2
group of PhCH2OH, and at δ -18.23 for the diruthenium
hydride 6 (Scheme 12). The NMR yield of PhCH2OH was
determined with ferrocene as an internal standard. The formation
of PhCH2OH was independently confirmed by gas chromatog-
raphy.
Trapping of Dienone Dicarbonyl Ruthenium Intermediate
C. The stoichiometry and reaction products suggest that
hydrogen transfer from 8 to PhCHO initially produces a
coordinatively unsaturated dienone ruthenium dicarbonyl inter-
mediate C that is trapped by an additional equivalent of 8 to
give the diruthenium hydride 6. In an effort to intercept
intermediate C, the reaction of 8 with PhCHO (1.0 M) was
carried out at 0 °C in the presence of PPh3 (0.4 M).30 In addition
to the disappearance of 8 and the formation of PhCH2OH, a
The kinetics of PhCHO reduction by 8 were measured by
adding an excess (typically >10-fold excess) of aldehyde to a
THF-d8 solution of 8 at -78 °C and monitoring the disappear-
1
ance of 8 by H NMR spectroscopy in an NMR probe cooled
to -10 °C. The disappearance of 8 followed pseudo-first-order
kinetics to over 80% conversion, indicating a first-order
dependence on 8. Four kinetic runs were carried out at -10 °C
with aldehyde concentrations of 0.8 (kobs ) 5.0 × 10-4 s-1),
1.0 (kobs ) 6.2 × 10-4 s-1), 1.2 (kobs ) 8.0 × 10-4 s-1), and
1.6 M (kobs ) 1.0 × 10-3 s-1). A linear plot of kobs vs PhCHO
concentration established a first-order dependence on PhCHO.
1
new tolyl resonance at δ 2.17 in the H NMR spectrum and a
new phosphorus resonance at δ 38.4 in the 31P NMR spectrum
provided evidence for the formation of [2,5-Ph2-3,4-Tol2(η4-
C4CO)]Ru(CO)2PPh3 (9). The absence of a bridging hydride
(23) The few pKa values determined for hydroxycyclopentadienyl ligands
in water are in the same range as organic acids. [C5H4OH]Fe(CO)2Cl was
calculated to have a pKa of ∼4 in H2O. Weiss, E.; Mere´nyi, R.; Hu¨bel, W.
Chem. Ber. 1962, 95, 1170.
1
resonance in the H NMR spectrum showed that none of the
diruthenium hydride 6 was formed in the reaction. Thus, 0.4 M
(24) pKa measurements are available for a much narrower range of
compounds in THF. Coeztee, J. F. Prog. Phys. Org. Chem. 1967, 4, 45.
(25) (a) Jordan, R. F.; Norton, J. R. J. Am. Chem. Soc. 1982, 104, 1255.
(b) Moore, E. J.; Sullivan, J. M.; Norton, J. R. J. Am. Chem. Soc. 1986,
108, 2257.
(28) Under catalytic conditions, carbonyl reduction by 8 is reversible.
Reduction of PhCDO (1.0 M) by 8 at 0 °C resulted in no D incorporation
into the product 6. This result illustrates that under the stoichiometric
conditions used in these kinetic studies, no reversible alcohol dehydroge-
nation occurs.
(29) The slightly slower observed rate for hydride disappearance is
attributable to the trapping of approximately 20% of dienone dicarbonyl
intermediate C by CO to form tricarbonyl 5.
(30) A control experiment showed that 8 does not react with PPh3 at
room temperature. This is somewhat surprising since 13CO exchange occurs
into 8 under these conditions. This result may be accounted for by the steric
bulk of the tetraarylcyclopentadienyl ligand. For a similar example, see:
Adams, H.; Bailey, N. A.; Browning, A. F.; Ramsden, J. A.; White, C. J.
Organomet. Chem. 1990, 387, 305.
(26) Keq ) [NEt3H+][15]/[NEt3][8]; pKa(8-OH)) pKa(NEt3H+) + pKeq.
(27) In general, the pKa values of organic -OH acids in CH3CN are
higher relative to ammonium ions than in water because CH3CN solvates
ionic species poorly, and deprotonation of the cationic ammonium salts to
neutral amines is thermodynamically favored in CH3CN. (a) Chantooni,
M. K., Jr.; Kolthoff, I. M. J. Am. Chem. Soc. 1970, 92, 7025. (b) Kolthoff,
I. M.; Chantooni, M. K., Jr.; Bhowik, S. J. Am. Chem. Soc. 1966, 88, 5430.
For a recent illustrative example see: Shafiq, F.; Szalda, D. J.; Creutz, C.;
Bullock R. M. Organometallics 2000, 19, 824.

EVidence for Concerted Hydride and Proton Transfer
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1095
Table 1. Kinetic Deuterium Isotope Effects on the Reduction of
PhCHO by 8
Scheme 13
kOHRuH/kOHRuD
1.5 ( 0.2
2.2 ( 0.1
3.6 ( 0.3
1.6 ( 0.2
2.3 ( 0.4
k
k
OHRuH/kODRuH
OHRuH/kODRuD
kODRuH/kODRuD
OHRuD/kODRuD
k
and a new tolyl resonance at δ 2.18 assigned to hydroxy-
trifluoroacetate complex 11 appeared. No resonances corre-
sponding to the diruthenium hydride 6 were observed. Appar-
ently the initially formed intermediate C is trapped by oxidative
addition of trifluoroacetic acid.
PPh3 is a much more efficient trap for the intermediate C than
is the ruthenium hydride 8.
The rate of conversion of PhCHO to PhCH2OH was the same
as in the absence of PPh3. However, because only 1 equiv of 8
was consumed in the presence of PPh3 compared with 2 equiv
in the absence of PPh3, the rate of disappearance of 8 is only
half as rapid in the presence of PPh3 (eq 2). The second-order
rate constant in the presence of PPh3 (k2 ) (9.1 ( 0.5) × 10-4
M-1 s-1 at 0 °C) was approximately the same as in the absence
of PPh3 (k2 ) (7.7 ( 0.6) × 10-4 M-1 s-1 at 0 °C).
Even though CF3CO2H is a much stronger acid than the
hydroxy group of 8, the rate of PhCHO reduction by 8 in the
presence of 0.4 M CF3CO2H was less than twice as fast as in
the absence of acid. This indicates that the hydroxyl group of
8 is an unusually effective intramolecular general acid for
promotion of hydride transfer from ruthenium to PhCHO.
Reduction of PhCHO by the Anionic Alkoxycyclopenta-
dienyl Ruthenium Hydride 13. The relative importance of the
hydride donor strength of the ruthenium hydride and the acidity
of the hydroxycyclopentadienyl in PhCHO reduction was
investigated by studying the reaction of the anionic alkoxycy-
clopentadienyl ruthenium hydride NEt4+[2,5-Ph2-3,4-Tol2(η5-
C4CO)Ru(CO)2H]- (13) with PhCHO. This complex should be
a better hydride donor than 8 because of the negative charge
on the complex, but it does not possess an acidic proton. No
reaction between PhCHO (1.0 M) and hydride anion 13 in CD3-
CN was observed after 80 min at room temperature. The purity
of the PhCHO used in the reduction altered the rate of hydride
disappearance markedly. While no reduction of freshly distilled
PhCHO was observed, relatively fast reductions were observed
with undistilled PhCHO stored in air (t1/2 ≈ 15 min at room
temperature for a 1.0 M solution).34 Addition of the aldehyde
oxidation product PhCO2H speeded up reduction rates dramati-
cally.35 The accelerating effect of PhCO2H in the reductions
with anionic hydride 13 underscores the critical importance of
the general acid catalysis in the reductions with hydroxy-hy-
dride 8.
No PhCHO Reduction by Triethylsiloxycyclopentadienyl
Ruthenium Hydride 12. To further explore the role of the
hydroxycylopentadienyl proton in mediating the reduction of
PhCHO, the reaction of PhCHO with [2,5-Ph2-3,4-Tol2(η5-
C4COSiEt3)Ru(CO)2H] (12), which has no acidic hydrogen, was
studied. No reaction between PhCHO and the triethylsiloxy-
hydride 12 in C6D6 was observed after 24 h at room tempera-
ture.36 This result confirms the importance of the hydroxyl in
promoting the carbonyl reduction. To test for the possibility
that hydride transfer might occur reversibly from 12 to PhCHO
to generate a benzyl alkoxide intermediate, a solution of 12 and
PhCDO in C6D6 was monitored after 24 h at room temperature.
No decrease in the hydride resonance of 12 at δ -9.20 was
seen, indicating that no reversible hydride addition to 12 was
occurring (Scheme 13).
d[PhCH2OH]
d[8]
dt
d[PhCHO]
-
) -
)
)
dt
dt
d[9]
dt
) k2[8][PhCHO] (2)
Deuterium Isotope Effects. Isotope effects on PhCHO
reduction were investigated by selective deuteration of the
hydroxy and hydride positions of 8. To ensure that adventitious
proton sources (small amounts of PhCOOH or H2O) did not
attenuate the kinetic isotope effect on the -OH proton, D2O
was added to the reactions.31 As a control, the rate of PhCHO
(1.0 M) reduction at 0 °C was measured in the presence of H2O
(0.2 M). The reaction proceeded with kobs ) (1.60 ( 0.02) ×
10-3 s-1, a small increase from the rate in the absence of water.
The rate of PhCHO (1.0 M) reduction at 0 °C in the presence
of D2O (0.2 M) proceeded with kobs ) (7.3 ( 0.1) × 10-4 s-1
.
Dideuterated 8-d2 (>90% isotopic purity at the hydride posi-
tion)32 was prepared by heating a THF-d8 solution of 7 under
D2 (1 atm at -196 °C) at 80 °C for 8 h. PhCHO (1.0 M)
reduction by 8-d2 at 0 °C in the presence of H2O (0.2 M)
proceeded with kobs ) (1.04 ( 0.10) × 10-3 s-1. PhCHO (1.0
M) reduction by 8-d2 in the presence of D2O (0.2 M) proceeded
with kobs ) (4.5 ( 0.3) × 10-4 s-1. The deuterium isotope
effects determined from these rates are summarized in Table 1.
PhCHO Reduction in the Presence of CF3CO2H. The
mechanism shown in Scheme 12 involves simultaneous transfer
of hydride from ruthenium and protonation of the PhCHO
oxygen by the hydroxy group. This proton transfer takes
advantage of the high acidity of the hydroxycyclopentadienyl
group. To test whether other Bronsted acids could act as general
acids to transfer a proton to the benzaldehyde oxygen and
accelerate hydride transfer from 8, the rate of reaction of 8 with
PhCHO (1.0 M) in the presence of CF3CO2H (0.4 M, pKa )
12.6 in CH3CN)33 at 0 °C was measured by 1H NMR
spectroscopy (k2 ) (1.09 ( 0.03) × 10-3 M-1 s-1). The hydride
resonance of 8 disappeared and new resonances for PhCH2OH
Reduction of Other Polar Double Bonds by 8. Reduction
of PhCOMe by 8 occurred more slowly than reduction of
PhCHO. Reaction of 8 with PhCOMe (1.2 M) in THF-d8 at 22
°C gave 1-phenylethanol and diruthenium complex 6. The rate
(31) In the absence of added D2O, nonreproducible and internally
inconsistent results were obtained for kinetic isotope effects on the hydroxyl
proton.
(32) At room temperature the OH proton of 8 readily exchanged with
external D2O in THF-d8, but no exchange with Ru-H was observed. When
a THF-d8 solution of 8 and D2O was heated at 80 °C exchange with Ru-H
was observed. This exchange was faster than exchange of 8 with D2 in
THF-d8. These results suggest that 8 is in equilibrium with a dihydrogen
complex that dissociates H2, leading to exchange with D2. Casey, C. P.;
Singer, S. W. Unpublished results.
(34) In contrast, the rate of PhCHO reduction by 8 was the same with
pure or PhCOOH-contaminated PhCHO.
(35) The requirement of an acid to promote carbonyl reduction by anionic
transition metal hydrides has been previously observed. Gaus, P. L.; Kao,
S. C.; Youngdahl, K.; Darensbourg, M. Y. J. Am. Chem. Soc. 1985, 107,
2428.
(33) Jasinski, T.; El-Harakany, A. A.; Halaka, F. G.; Sadek, H. Croat.
Chem. Acta 1978, 51, 1.
(36) Reactions with 12 were done in C6D6 because the complex was
very sensitive to hydrolysis in THF-d8.

1096 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
Casey et al.
constant for reduction was k2 ) (2.0 ( 0.2) × 10-4 M-1 s-1 at
22 °C. Extrapolating the rate of reduction of PhCHO to 22 °C
(k ) 6.0 × 10-3 M-1 s-1) indicates that PhCOMe reduction is
30 times slower than PhCHO reduction.37
hydrogen transfer, the aldehyde is outside the coordination
sphere of the metal. The second-order rate law requires a
transition state involving both 8 and the aldehyde; the large
negative ∆S‡ ) -28 eu is consistent with a concerted mech-
anism.
The facile hydrogen transfer to carbonyl groups led us to
examine reduction of imines. From Shvo’s work, we anticipated
that the resulting amine after hydrogen transfer would bind to
the dienone dicarbonyl intermediate C to form a stable amine
adduct 10. Measurement of the rate of hydrogen transfer to
MeNdCHPh was attempted at -10 °C to compare directly with
the rate of PhCHO reduction, but the reaction proceeded too
quickly to accurately measure a rate. At -25 °C, the rate
constant for imine reduction was k2 ) (2.63 ( 0.05) × 10-3
M-1 s-1. Extrapolating the rate of PhCHO reduction to -25
°C (k ) 1.0 × 10-4 M-1 s-1) indicates that imine reduction is
26 times faster than PhCHO reduction. The product of the
reaction is amine adduct 10, which was confirmed by indepen-
dent synthesis.
Solvent Effects on Hydrogen Transfer. The rate of hydro-
gen transfer from hydroxy-hydride 8 to PhCHO was signifi-
cantly increased by changing the solvent from THF-d8 to
toluene-d8. The kinetics of the reaction of 8 with PhCHO (1.0
M) in toluene-d8 at -45 °C were monitored by following the
disappearance of the hydride resonance of 8 (kobs ) 1.2 × 10-3
s-1, k2 ) (6.0 ( 0.2) × 10-4 M-1 s-1, ∆G‡ ) 16.6 kcal mol-1).
The rate of hydrogen transfer to PhCHO was approximately 50
times faster in toluene than THF (∆∆G‡ ) 1.8 kcal mol-1).
Hydrogen transfer to PhCOCH3 (k2 ) (2.3 ( 0.2) × 10-4 M-1
s-1 at -10 °C) and MeNdCHPh (k2 ) (4.50 ( 0.25) × 10-3
M-1 s-1 at -60 °C) was also faster in toluene than THF.38
Alkene Reactivity. While hydrogen transfer from 8 to polar
double bonds occurs rapidly below room temperature, cyclo-
hexene was unreactive. No reaction was observed at room
temperature when an excess of cyclohexene was added to a
THF-d8 solution of 8. At 80 °C, loss of H2 from 8 to form
diruthenium complex 6 occurred without observable hydrogena-
tion of cyclohexene. However, when 8 and cyclohexene were
heated at 80 °C under H2 (1 atm), slow catalytic hydrogenation
of cyclohexene occurred (∼5 turnovers in 24 h).39 Similar
turnover rates were observed when a toluene-d8 solution of
triethylsiloxy-hydride 12 and excess cyclohexene was heated
at 80 °C under H2. Under the conditions of this catalytic
hydrogenation, hydrides 8 and 12 were the only observable Ru
species.40
The observation of primary deuterium isotope effects for both
Ru-D and O-D provides the strongest evidence for a concerted
hydrogen transfer mechanism involving both Ru-H and O-H
bond breaking in the transition state. Ru-H transfer from one
molecule of 8 concerted with O-H transfer from a second
molecule of 8 can be excluded since such a mechanism would
require a rate dependence on [8]2. The kinetic isotope effect of
1.5 ( 0.2 for Ru-D is similar to previously measured isotope
effects for rate-determining hydride additions.41 The isotope
effect of 2.2 ( 0.1 for O-D is consistent with rate-determining
proton transfer, though it is smaller than the isotope effect
reported for proton transfer in many acid-catalyzed reactions.42
As expected for a concerted reaction, the product of the two
individual isotope effects (1.5 × 2.2 ) 3.3) is within experi-
mental error of the combined D isotope effect measured for
RuD-OD (3.6 ( 0.3).
Unfortunately, there are no experimental or theoretical studies
of kinetic isotope effects on concerted proton and hydride
transfer in organometallic complexes to compare to the isotope
effects observed here.43 However, a concerted hydrogen transfer
mechanism has been proposed for certain dehydrogenase
enzymes.44 In particular, theoretical studies on models of lactate
dehydrogenase, which catalyzes the redox interconversion of
lactate and pyruvate, have estimated isotope effects for a
concerted proton and hydride transfer.45 Calculated primary
kinetic isotope effects for a dihydropyridine/formaldehyde/
imidazolium system were 1.6-2.0 for hydride transfer and 2.3-
2.4 for proton transfer.46 Isotope effects of similar magnitude
were observed experimentally for the reduction of p-benzo-
quinone by 10-methylacridan in the presence of substituted
acetic acids.47
Two different two-step mechanisms can also be excluded.
Rapid and reversible hydride donation from ruthenium to the
aldehyde carbonyl followed by proton transfer to the carbonyl
oxygen can be excluded by the absence of deuterium exchange
between PhCDO and siloxycyclopentadienyl ruthenium hydride
12. Similarly, rapid and reversible protonation of the aldehyde
carbonyl followed by hydride transfer from ruthenium appears
unlikely because specific acid catalysis is normally accompanied
by an inverse isotope effect.42
Discussion
Though concerted hydrogen transfer is the most economical
mechanistic explanation for observation of isotope effects at
Concerted Hydrogen Transfer Mechanism. We have
obtained persuasive evidence that reduction of aldehydes by 8
occurs by simultaneous transfer of hydride from ruthenium and
of a proton from the CpOH group (Scheme 12). During the
(41) Bullock, R. M. In Transition Metal Hydrides; Dedieu, A., Ed.;
VCH: New York, 1992; Chapter 8, pp 263-307.
(42) Jencks, W. P. Catalysis in Chemistry and Enzymology; Dover:
Mineola, NY, 1987; Chapter 4, pp 243-282.
(37) Using activation parameters obtained from PhCHO reductions
corresponds to a difference in activation barriers of 2.0 kcal mol-1 (22.3
kcal mol-1 vs 20.3 kcal mol-1).
(38) PhCHO reductions were also run in CD3CN (k ) (9.7 ( 0.6) ×
10-3 M-1 s-1 at 0 °C) and in DMSO-d6 (k ) (1.00 ( 0.04) × 10-3 M-1
s-1 at 21 °C). These measured rates indicate that the strength of hydrogen
bonding interactions between the hydroxyl proton of 8 and the solvent
dramatically changes the rate of aldehyde reduction.
(39) No catalytic hydrogenation was observed when 8 and excess HCO2H
were heated to 80 °C in the presence of cyclohexene in THF-d8.
(40) At room temperature, 8 catalyzes the isomerization of 1-pentene to
a thermodynamic mixture of cis- and trans-2-pentene. Addition of methyl
acrylate to 8 forms an alkyl complex by cis hydride addition across the
CdC double bond. These alkene reactions are promoted by light, so
photochemical CO dissociation appears to be the first step in these reactions.
It is unclear whether photochemical CO dissociation is involved in catalytic
hydrogenation of alkenes at elevated temperature. Casey, C. P.; Singer, S.
W. Unpublished results.
(43) Noyori has reported an isotope effect of 1.5 ( 0.1 for dehydroge-
nation of 2-propanol by the 16 e- precursor to 1.2b
(44) For background on dehydrogenase enzymes, see: Abeles, R. H.;
Frey, P. A.; Jencks, W. P. Biochemistry; Jones and Bartlett: Boston, 1992;
Chapter 16, pp 447-477.
(45) (a) Wilkie, J.; Williams, I. H. J. Am. Chem. Soc. 1992, 114, 5423.
(b) Wilkie, J.; Williams, I. H. J. Chem. Soc., Perkin Trans. 2 1995, 1559.
(46) Coleman, C. A.; Rose, J. G.; Murray, C. J. J. Am. Chem. Soc. 1992,
114, 9755.
(47) The small isotope effects were rationalized by the coupling of heavy
atom motion to proton transfer.45,46 Though the concerted transfers from 8
differ from these model system because the proton and hydride are
electronically coupled, the small observed isotope effects may also be due
to extensive heavy atom motion (lengthening of the carbonyl C-O bond
and slippage of the η5-hydroxy-cyclopentadienyl ligand to an η4-cyclopen-
tadienone ligand). Analysis of primary kinetic isotopes on hydrogen transfer
in complexes related to 8 and analysis of heavy atom isotope effects will
afford a more detailed picture of the transition state for concerted transfer.

EVidence for Concerted Hydride and Proton Transfer
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1097
Scheme 14
barrier and an isotope effect will be seen at only one site. For
concerted mechanisms, isotope effects should be seen for both
steps even if the substrates are varied over a wide range.
The acidity of the Cp-OH (pKa ) 17.5 in CH3CN) is crucial
for successful hydride transfer from Ru-H. This hydroxyl group
is more acidic than either phenol (pKa ) 26.6 in CH3CN) or
benzoic acid (pKa ) 20.7 in CH3CN). When the acidic hydrogen
was replaced by -SiEt3 in compound 12, no hydride transfer
to PhCHO occurred. The anionic compound 13 with CpO- is
expected to be a better hydride donor than neutral 8, but it
proved to be unreactive toward PhCHO. The ability of 8 to
transfer hydride from Ru clearly requires simultaneous proton
transfer from CpO-H to the aldehyde oxygen. The small
observed acceleration of PhCHO reduction in the presence of
excess CF3CO2H reinforces the importance of having the proton
and hydride electronically coupled to achieve efficient hydrogen
transfer.
The proposed concerted hydrogen transfer mechanism gener-
ates the coordinatively unsaturated dienone dicarbonyl inter-
mediate C and free alcohol. Intermediate C was successfully
trapped by PPh3 and CF3CO2H. Though an alcohol complex
may form transiently, it was not detected spectroscopically and
probably does not play a role in hydrogen transfer. The absence
of a detectable alcohol complex contrasts sharply with aldimine
hydrogen transfer, in which a stable amine complex is formed
after the transfer occurs.
Direct reductions by 8 require a polar multiple bond. The
observations that cyclohexene does not react with 8 in the
absence of H2 and that reduction only occurs under H2 suggests
a very different mechanism for alkene reduction. We will
address the mechanism of alkene reduction in detail in a future
paper.
An alternative mechanism involves coordination of aldehyde
to Ru followed by hydride transfer and requires prior CO
dissociation (Scheme 4). Since the rate of reduction of PhCHO
by 8 was not retarded under CO atmosphere, reversible CO
dissociation can be excluded as a reaction step. Since 13CO
exchange into 8 was much slower than the reaction of 8 with
PhCHO, rate-determining CO loss can also be excluded as a
reaction step. Rate-determining CO dissociation is also incon-
sistent with the observation of isotope effects and the second-
order kinetics.
A ring-slip mechanism (Scheme 5) provides a route to
aldehyde coordination without CO loss. If this mechanism were
operative, then the triethylsiloxy ruthenium hydride 12 should
be equally susceptible to ring slip, aldehyde coordination, and
hydride transfer to give a Ru-O-CHDPh intermediate (Scheme
13). However, in the absence of an acidic proton the reduction
would not be completed. The fact that PhCDO does not ex-
change with 12 is taken as evidence against the ring-slip mecha-
nism. The observed isotope effects for reduction of PhCHO by
8 are not readily interpreted in terms of a ring-slip mechanism.
Comparison of Substrates. Hydrogen transfer from 8 to
PhCHO occurs approximately 30 times faster than transfer to
PhCOMe, consistent with the enhanced electrophilicity of the
aldehyde. However, the MeNdCHPh undergoes a more rapid
hydrogen transfer than PhCHO, even though it has a more
weakly polarized carbon-heteroatom bond. Since proton trans-
fer of the acidic hydroxyl proton of 8 is crucial in promoting
hydride transfer from Ru, the faster relative rate of MeNdCHPh
reduction may be attributed to the stronger basicity of the
nitrogen heteroatom. The structure of the transition state for
hydrogen transfer to MeNdCHPh probably involves a larger
degree of substrate protonation than in transfer to PhCHO.
both centers, a stepwise mechanism in which the barriers for
proton transfer followed by hydride transfer are approximately
equal (k-1 ≈ k2 in Scheme 14) cannot be excluded. If k-1 and
k2 differ by a factor of 10 or greater, then an isotope effect
would be seen only for the slower step. In this mechanism,
endothermic proton transfer from oxygen is followed by fast
hydride transfer from ruthenium. This hydride transfer must be
extremely fast to have the same barrier as the exothermic back
proton transfer between the aldehyde oxygen and the oxygen
on the Cp ring. Such exothermic proton transfers between
oxygen centers occur at near diffusion controlled rates.48
For a concerted mechanism, the product of the individual
isotope effects {(kOHRuH/kODRuH) × (kOHRuH/kOHRuD) ) 2.2 (
0.1 × 1.5 ( 0.2 ) 3.3 ( 0.5} should be equal to the isotope
effect observed for doubly labeled material (kOHRuH/kODRuD
)
3.6 ( 0.3). Within experimental error, this is the result observed.
For the two-step mechanism shown in Scheme 14, the isotope
effect observed for doubly labeled material should be smaller
than the product of the individual isotope effects. For example,
assuming no equilibrium isotope on formation of the protonated
carbonyl intermediate and equal barriers for OH and RuH
transfer, one must employ isotope effects of 3.4 on OH transfer
and 2.0 on RuH transfer to obtain the observed isotope effects
of kOHRuH/kODRuH ) 2.2 ( 0.1 and kOHRuH/kOHRuD )1.5 ( 0.2
(see Supporting Information, pp 70-74). The combination of
the two isotope effects predicts an isotope effect for double-
labeled material of kOHRuH/kODRuD ) 2.7, which is smaller than
the observed value of 3.6 ( 0.3, but not far outside of error
estimates. If we assume equal barriers for OH and RuH transfer
and an equilibrium isotope effect of 1.5 favoring aldehyde
protonation by 8 compared with aldehyde deuteration by 8-OD,
one must employ kinetic isotope effects of 3.0 on OH transfer
and 2.0 on RuH transfer to obtain the observed isotope effects
(see Supporting Information, pp 75-79).49 This scheme then
predicts an isotope effect for double-labeled material of kOHRuH
/
kODRuD ) 3.0. Again, this number is smaller than the observed
3.6 ( 0.3, but not far outside of error estimates.
To distinguish between our favored concerted mechanism and
a two-step mechanism with nearly equal barriers, we plan to
investigate isotope effects on different substrates. For a two-
step mechanism with a significantly different substrate, it is
likely that one of the two steps will have a substantially different
(48) The diffusion-controlled rate is derived from Jencks’ discussion of
stepwise and concerted acid-base catalysis in aqueous solution and is
intended as an estimate of the rate in organic solvents. (a) Jencks, W. P.
Acc. Chem. Res. 1976, 9, 425. (b) Jencks, W. P. Acc. Chem. Res. 1980, 13,
161.
(49) The equilibrium between 8 + PhHCdOD+ and 8-OD + PhHCd
OH+ should favor deuterium at the site with the strongest bond, 8-OD.
Wiberg, K. B. Physical Organic Chemistry; John Wiley and Sons: New
York, 1964; Part II, Chapter 7, pp 273-278.

1098 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
Casey et al.
Comparison with Other Metal Hydride Catalysts. As noted
in the introduction, a number of metal complexes that contain
both a metal hydride and an acidic proton on a coordinated
ligand have recently been shown to be catalysts for transfer
hydrogenation of ketones. These catalysts may well be mecha-
nistically related to 8. The reactivity of Noyori’s Ru(II)-TsDPEN
catalyst 1 (Scheme 1) resembles that of 8. Both catalysts transfer
hydrogen reversibly to carbonyl substrates, and the presence of
an acidic proton on a coordinated ligand is crucial for efficient
hydrogen transfer.2a Though no detailed kinetic studies of the
reduction of carbonyls by 1 have been reported, extensive
computational studies have established that the concerted
hydrogen transfer from 1 to ketones has a lower activation
barrier than routes involving substrate coordination to the metal
center followed by hydride transfer.50
Hydrogen transfer to imines provides a striking reactivity
difference between 8 and 1. Hydrogen transfer from 8 is faster
to imines than to ketones, while 1 reduces ketones but is
unreactive toward imines.51 Calculations show that the activation
barrier for hydrogen transfer from 1 to an imine is 5-11 kcal
mol-1 higher than the barrier for hydride transfer to carbonyls.
This difference in relative reactivities may be accounted for by
the more acidic pKa of the OH group of 8 compared with that
of the NH2 group of 1.
Comparison to Neutral Hydrides in the Presence of Strong
Acids. Hydroxy-hydride 8, which incorporates both acidic and
hydridic hydrogens in the same molecule, may be compared to
the reduction systems in which hydride and proton are delivered
from separate molecules.52 Bullock and Norton found that the
neutral transition metal hydrides, CpW(CO)3H and CpMo-
(CO)3H, reduced carbonyls to alcohols in the presence of strong
Bronsted acids (CF3CO2H, CF3SO2H). In the reduction of
acetone by CpW(CO)3H/CF3SO2H, the alcohol complex [Cp-
(CO)3W(OHCH(CH3)2)][OTf] (16) was isolated and character-
ized. Kinetic studies on the CpMo(CO)3H/CF3CO2H reduction
of acetone showed the rate-determining step in the reduction
was hydride transfer to a protonated ketone.
The Bullock-Norton ionic carbonyl hydrogenations and the
Shvo system both utilize an acidic proton to promote hydride
transfer. In both cases, ketone reduction is kinetically favored
over dihydrogen evolution. 8 and related transfer hydrogenation
catalysts differ from the ionic hydrogenation systems in that
the active catalyst is formed from relatively unreactive precur-
sors (alcohols, H2, HCO2H) and transfers proton and hydride
in concert, not stepwise. Presumably, having the proton and
hydride electronically coupled allows the reduction to proceed
with the less acidic -OH proton in 8. The concerted mechanism
has significant benefits in providing a mild method to reduce
carbonyls with excellent chemoselectivity. Another difference
between the Bullock-Norton ionic carbonyl hydrogenations and
reductions by 8 is that alcohol complexes such as 16 are isolated
from the former system but no evidence for alcohol complexes
was seen in reductions by 8. The neutral intermediate C derived
from hydrogen transfer from 8 binds alcohols poorly; this allows
the catalyst to turn over without product alcohol inhibition.
Relevance to Shvo Catalytic System. The kinetic studies
of aldehyde reduction by 8 confirm Shvo’s hypothesis that 8 is
the active species in the catalytic reactions. Since the rate of
stoichiometric reduction of PhCHO by 8 is so much faster than
catalytic reduction, it is clear that regeneration of 8 after
hydrogen transfer is the rate-determining step in the catalysis.
In accord with this proposal, Shvo has shown by IR spectroscopy
that the resting state in the catalytic reduction of cyclohexanone
at 85 °C and 12 atm of H2 is the diruthenium complex 3.53
Since the rate of reduction of PhCHO by 8 is so rapid, it
should be possible to devise much more reactive catalysts by
structural modifications that would prevent formation of the
unreactive diruthenium species 6. For example, increasing the
steric bulk of the cyclopentadienone ligand might prevent the
formation of diruthenium species but not greatly interfere with
either the rate of regeneration of the hydroxy-hydride from H2
or the rate of aldehyde reduction. We are exploring the synthesis
of such complexes in the hope of developing highly active
catalysts.
While the stoichiometric reduction of PhCHO by 8 is
immeasurably faster than reduction of cyclohexene, Shvo has
reported that cyclohexanone is catalytically reduced only eight
times faster than cyclohexene. We are initiating detailed
mechanistic investigations of alkene reduction by 8 to under-
stand the underlying reasons for chemoselective reductions in
both stoichiometric and catalytic systems.
Experimental Section
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru(CO)3 (5). Ru3(CO)12 (400 mg, 0.62
mmol) and 2,5-diphenyl-3,4-(di-p-tolyl)cyclopenta-2,4-dien-1-one (4)
(772 mg, 1.86 mmol) were dissolved in toluene (25 mL) and heated at
reflux until the solution became light red (2-3 h).54 The solution was
concentrated to ∼2 mL under vacuum and pentane (10 mL) was added
to afford 5 (890 mg, 80%) as a pale yellow precipitate, which was
collected by filtration, mp 164-166 °C dec. IR (CH2Cl2) 2083 (s), 2023
1
(br) cm-1. H NMR (C6D6, 300 MHz) δ 1.82 (s, 6H, tolyl CH3), 6.62
(d, 3J ) 8.0 Hz, 4 H, tolyl), 7.02 (m, 10 H, phenyl and tolyl), 7.94 (m,
4H, phenyl). 13C{1H} NMR (C6D6, 75 MHz) δ 20.9 (CH3), 81.5 (C
3,4 of Cp), 108.4 (C 2,5 of Cp), 127.2-138.5 (8 resonances, aromatic),
174.8 (C1 of Cp), 195.5 (CO). HRMS (EI) calcd (found) for C34H24O4-
Ru: 598.0711 (598.0699).
[2,5-Ph2-3,4-Tol2(η5-C4CO)]2H}Ru2(CO)4(µ-H) (6). Ru3(CO)12 (300
mg, 0.46 mmol) and 4 (578 mg, 1.40 mmol) were suspended in
methanol (60 mL) and heated at reflux for 40 h. A bright yellow
precipitate 6 (465 mg, 59%) was collected by filtration, mp 200-204
1
°C dec. IR (CH2Cl2) 2036 (vs), 2004 (m), 1977 (s) cm-1. H NMR
(C6D6, 300 MHz) δ -17.68 (s, RuHRu), 1.79 (s, 12H, tolyl CH3), 6.58
(d, 3J ) 8.4 Hz, 8 H, tolyl), 6.95 (m, 12 H, phenyl), 7.09 (d, 3J ) 8.4
Hz, 8 H, tolyl), 7.50 (m, 8 H, phenyl), 12.90 (br s, OH). 13C{1H} NMR
(C6D6, 75 MHz) δ 20.9 (CH3), 88.2 (C 3,4 of Cp), 104.2 (C 2,5 of
Cp), 127.0-137.8 (8 resonances, aromatic), 155.3 (C1 of Cp), 202.0
(CO). Anal. Calcd for C66H50O6Ru2: C, 69.46; H, 4.41. Found: C,
69.56; H, 4.31.
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru2(CO)2]2 (7). Ru3(CO)12 (531 mg,
0.83 mmol) and 4 (957 mg, 2.49 mmol) were suspended in heptane
(175 mL) and heated at reflux for 48 h. The mustard yellow precipitate
was collected by filtration to give 7 (1.04 g, 74%), mp 195-200 °C
dec. IR (CH2Cl2) 2019 (s), 1967 (s) cm-1. 1H NMR (C6D6, 300 MHz)
δ 1.69 (s, 12H, tolyl CH3), 6.49 (d, 3J ) 8.1 Hz, 8 H), 6.88 (m, 20 H,
tolyl and phenyl), 7.50 (m, 8 H, phenyl). 13C{1H} NMR (C6D6, 75
MHz) δ 20.8 (CH3), 88.1 (C 3,4 of Cp), 99.2 (C 2,5 of Cp), 126.4-
137.4 (8 resonances, aromatic), 171.5 (C1 of Cp), 200.8 (CO). Anal.
Calcd for C66H48O6Ru2: C, 69.58; H, 4.24. Found: C, 69.82; H, 4.12.
(50) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122,
1466.
(51) Under catalytic conditions, Noyori has reported that the Ru(II)-
TsDPEN catalyst efficiently reduces imines to optically active amines with
HCO2H:NEt3 (5:2) as the hydrogen source. The selectivity for imines over
ketones is dramatic enough that the imine reductions may be run in acetone
solvent. However, the observation that 1 is unreactive toward imines
suggests that, under catalytic conditions, it is the protonated imminium ion
that is reduced. Uematsu, N.; Fujii, A.; Hashiguchi, S.; Ikariya, T., Noyori,
R. J. Am. Chem. Soc. 1996, 118, 4916.
(53) Blum, Y.; Czarkie, D.; Rahamim, Y.; Shvo, Y. Organometallics
1985, 4, 1459.
(52) Song, J. S.; Szalda, D. J.; Bullock, R. M.: Lawrie, C. J. C.; Rodkin,
M. A.; Norton, J. R. Angew. Chem., Int. Ed. Engl. 1992, 31, 1233.
(54) It is crucial during the reflux to keep the oil bath temperature below
130 °C to prevent product decomposition.

EVidence for Concerted Hydride and Proton Transfer
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1099
[2,5-Ph2-3,4-Tol2(η5-C4COH)]Ru(CO)2H (8). A solution of 6 (20
mg, 0.02 mmol) in THF-d8 (0.45 mL) in a resealable NMR tube was
degassed by three successive freeze-pump-thaw cycles and placed
under 1 atm of H2 at -78 °C. The tube was sealed at -78 °C and
new satellites were observed for the hydride resonance at δ -9.76 after
2 h.
pKa Determination of -OH Proton of 8. Solvent was removed in
vacuo from a THF solution of 8 prepared from 6 and H2. The yellow
oily residue was redissolved in CH3CN (0.070 M solution) and NEt3
(3.0 µL, 0.072 M) was added via gastight syringe. An aliquot of the
solution was syringed into a 0.099 mm CaF2 IR cell. Two resonances
at 2014 and 1954 cm-1 are due to 8 and two at 1985 and 1920 cm-1
are due to NEt3H+[2,5-Ph2-3,4-Tol2(η5-C4CO)Ru(CO)2H]- (15). The
absolute concentrations of the two species were determined by obtaining
1
heated at 80 °C in a constant-temperature bath for 8 h. The H NMR
spectrum of the yellow-orange solution showed >98% conversion to
1
8. No isolation was attempted. IR (THF) 2012 (s), 1951 (s) cm-1. H
NMR (THF-d8, 500 MHz) δ -9.76 (s, RuH), 2.17 (s, 6H, tolyl CH3),
6.80 (d, 3J ) 7.9 Hz, 4 H, tolyl), 6.96 (d, 3J ) 7.9 Hz, 4 H, tolyl), 7.23
(m, 6 H, phenyl), 7.47 (m, 4 H, phenyl), 8.60 (br s, OH). 13C{1H}
NMR (THF-d8, 126 MHz) δ 21.1 (CH3), 92.2 (C 3,4 of Cp), 104.7 (C
2,5 of Cp), 127.9-137.5 (8 resonances, aromatic), 137.6 (C1 of Cp),
203.0 (CO).
8 was also synthesized from HCO2H and 6. HCO2H (7 µL, 0.20
mmol) was added via syringe under a stream of N2 to a solution of 6
in THF-d8 (0.45 mL, 0.04 M solution). The 1H NMR spectrum showed
no reaction at room temperature. After the solution was heated at 80
°C until gas evolution ceased (10 min), the 1H NMR spectrum showed
complete conversion to 8.
the molar absorptivities of 8 and 15. For 8: 2014 (2895 M-1 cm-1
)
and 1954 cm-1 (2650 M-1 cm-1). For 15: 1985 (2108 M-1 cm-1) and
1920 cm-1 (2108 M-1 cm-1).55 From the molar absorptivites, the ratio
of 8 to 15 was determined to be 1:3.6. By using the known pKa of
NEt3 (pKa ) 18.5) and the observed equilibrium constant (Keq ) 12.2),
a pKa of 17.5 was determined for the -OH proton of 8.
Kinetics of PhCHO Reduction by 8. General procedure: PhCHO
was added via gastight syringe under a stream of N2 to a resealable
NMR tube containing a THF-d8 solution of 8 (0.05-0.08 M) and
ferrocene (internal NMR integration standard) at -78 °C. The tube
was evacuated and nitrogen readded. The tube was inserted into an
NMR spectrometer precooled to -10 ( 0.5 °C. The time between
removal of the tube from the cold bath and acquisition of the first
spectrum was ∼3 min. The reactions were followed through 2-3 half-
lives and the disappearance of 8 was measured by integrating its hydride
resonance at δ -9.76 relative to ferrocene and plotting its disappearance
with respect to time. All the kinetic runs were done with at least a
10-fold excess of PhCHO and followed pseudo-first-order kinetics. The
products formed were PhCH2OH and 6 as confirmed by 1H NMR and
GC.
[2,5-Ph2-3,4-Tol2(η5-C4COH)]Ru(CO)2(O2CCF3) (11). CF3CO2H
(55 µl, 0.90 mmol) was added via syringe to a CH2Cl2 (10 mL) solution
of 7 (110 mg, 0.10 mmol). After 1 h, solvent was evaporated under
vacuum and the residue was recrystallized from CH2Cl2/pentane to give
11 (112 mg, 85%) as an air-stable light green powder, mp 160-162
1
°C dec. IR (CH2Cl2) 2045 (s), 1995 (s) cm-1. H NMR (CD2Cl2, 300
MHz) δ 2.25 (s, 6H, tolyl CH3), 6.10 (br s, OH), 6.89 (AA′BB′, 8 H,
tolyl), 7.40 (m, 10 H, phenyl). 13C NMR{1H} (CD2Cl2, 125 MHz) δ
1
21.3 (CH3), 86.9 (C 3,4 of Cp), 101.0 (C 2,5 of Cp), 114.9 (q, JCF
)
189 Hz, CF3), 126.0-139.3 (8 resonances, aromatic), 142.2 (C1 of Cp),
163.7 (q, 2JCF ) 36 Hz, CF3CO2), 198.0 (CO). Anal. Calcd for C35H25O3-
RuF3: C, 61.49; H, 3.68. Found: C, 61.15; H, 3.52.
[2,5-Ph2-3,4-Tol2(η5-C4COSiEt3)Ru(CO)2H] (12). SiEt3H (175 µL,
1.09 mmol) was added via syringe to a CH2Cl2 (8 mL) solution of 7
(200 mg, 0.17 mmol). After 4 h, solvent was evaporated under vacuum
and the resulting green oil was dissolved in hexane (3 mL), filtered,
and cooled to -35 °C to give 12 as a moisture-sensitive green solid
(93 mg, 40%), mp 125-127 °C dec. IR (CH2Cl2) 2015 (s), 1954 (s)
[2,5-Ph2-3,4-Tol2(η5-C4COD)]Ru(CO)2D (8-d2). After a suspension
of 7 (15 mg, 0.013 mmol) in THF-d8 (0.37 mL) under 1 atm of D2
1
(added at -196 C) was heated at 80 °C for 8 h, H NMR indicated
>98% conversion to 8-d2 (0.07 M solution). The extent of deuteration
at both the hydroxy and hydride positions was determined to be >90%
1
by H NMR comparison with the tolyl methyl resonance.
Deuterium Isotope Effect Measurements. A THF-d8 solution of 8
(0.07 M), prepared from 6 and H2, was degassed and PhCHO (47 µL,
1.0 M) and D2O (1.6 µL, 0.2 M) were added at -78 °C. The tube was
inserted into an NMR spectrometer precooled to 0 ( 0.5 °C. The
1
3
cm-1. H NMR (C6D6, 300 MHz) δ -9.20 (s, RuH), 0.50 (q, J ) 7.3
Hz, SiCH2), 0.74 (t, 3J ) 7.3 Hz, SiCH2CH3), 1.81 (s, 6H, tolyl CH3),
6.62 (d, 3J ) 8.0 Hz, 4 H, tolyl), 7.00 (m, 6 H, phenyl), 7.29 (d, 3J )
8.0 Hz, 4 H, tolyl), 7.70 (m, 4 H, phenyl). 13C{1H} NMR (C6D6, 125
MHz) δ 5.3 (SiCH2), 7.1 (SiCH2CH3), 21.2 (CH3), 95.8 (C 3,4 of Cp),
105.3 (C 2,5 of Cp), 128.6-137.6 (8 resonances, aromatic), 134.1 (C1
of Cp), 203.1 (CO). HRMS (EI) calcd (found) for C39H40O3RuSi:
686.1783 (686.1815).
1
disappearance of the RuH H NMR resonance was monitored over 3
half-lives: kobs ) 7.4 × 10-4 s-1; t1/2 ) 15.6 min. In a second kinetic
run, kobs ) 7.1 × 10-4 s-1 and t1/2 ) 16.3 min. The products of the
reaction were 6-OD and PhCH2OD.
Hydrogenation of Cyclohexene Catalyzed by 8 in THF. Cyclo-
hexene (39 µL, 0.38 mmol) was added via syringe to a THF-d8 solution
of 8 (0.09 M) prepared from 6 and H2 at -78 °C. The solution was
degassed and placed under H2 (1 atm) at -78 °C. The tube was warmed
NEt4+[2,5-Ph2-3,4-Tol2(η4-C4CO)Ru(CO)2H]- (13). KBEt3H (550
µL, 1.0 M in THF) was added via syringe to a solution of 5 (200 mg,
0.34 mmol) and NEt4Br (70 mg, 0.34 mmol) in THF (8 mL) at -78
°C. The solution was stirred for 20 min at room temperature and 1 N
NaOH (5 mL) solution was added to destroy excess borohydride. THF
was evaporated under vacuum leaving an aqueous solution with a
yellow-green precipitate. The precipitate was washed with water (3 ×
5 mL) and hexane (10 mL) to give 13 (180 mg, 77%) as an off-white
air-sensitive powder. IR (THF) 1974 (s), 1910 (s) cm-1. 1H NMR (THF-
1
to room temperature. No new resonances were observed by H NMR
after 30 min. After the solution was heated at 80 °C for 20 min, a new
resonance was observed at δ 1.44, indicating cyclohexane formation.
After 24 h, ∼5 turnovers of cyclohexene had occurred.
X-ray Crystal Structure Determination for {[2,5-Ph2-3,4-Tol2-
(η5-C4CO)]2H}Ru2(CO)4(µ-H) (6). X-ray quality crystals were grown
by cooling a CH2Cl2/hexane solution of 6 to -10 °C for one week. A
yellow transparent plate-shaped crystal of dimensions 0.42 × 0.36 ×
0.02 mm was selected for structural analysis. Intensity data for this
compound were collected with a Bruker SMART CCD area detector
mounted on a Bruker P4 goniometer using graphite-monochromated
Mo Ka radiation (λ ) 0.71073 Å). The sample was cooled to 133(2)
K. The intensity data, which nominally covered one and a half
hemispheres of reciprocal space, were measured as a series of φ
oscillation frames each of 0.4 ° for 60 s/frame. The detector was
operated in 512 × 512 mode and was positioned 5.00 cm from the
sample. Coverage of unique data was 97.4% complete to 25.00° in θ.
Cell parameters were determined from a nonlinear least-squares fit of
8192 peaks in the range 3.0 < θ < 25.0°. The first 50 frames were
repeated at the end of data collection and yielded 224 peaks showing
3
d8, 300 MHz) δ -9.68 (s, RuH), 1.09 (br t, J ) 6.5 Hz, NCH2CH3),
3
3
2.15 (s, 6H, tolyl CH3), 3.15 (q, J ) 6.5 Hz, NCH2CH3), 6.71 (d, J
) 7.4 Hz, 4 H, tolyl), 7.04 (m, 10 H, tolyl and phenyl), 7.58 (m, 4 H,
phenyl). 13C{1H} NMR (THF-d8, 125 MHz) δ 7.8 (NCH2CH3), 21.4
(s, tolyl CH3), 52.9 (NCH2CH3), 84.2 (C 3,4 of Cp), 102.8 (C 2,5 of
Cp), 124.6-139.2 (8 resonances, aromatic), 171.0 (C1 of Cp), 207.5
(CO). MS (MALDI-TOF) calcd (found) for C33H25O3Ru- 571.1 (571.1).
Exchange of 13CO with 8. A solution of 8 (0.08 M) in THF-d8
prepared from 6 and H2 was degassed, cooled to -78 °C, and placed
under 13CO (1 atm). The tube was warmed to room temperature and
shaken frequently. After 15 min, the hydride resonance of 8 at δ -9.76
was flanked by a doublet satellite (2JCH ) 9.5 Hz); the integrated ratio
of the central peak to the doublet was 4:1. After 1.5 h, two additional
outer lines were observed due to a triplet; the integrated ratio of the
central singlet:doublet:triplet was 1:1:0.2. The reaction was discontinued
after 3 h (t1/2 for disappearance of unlabeled 8 was ∼1.5 h). When the
13CO exchange experiment was repeated in the absence of light, no
(55) The molar absorptivities of the 1,1,3,3-tetramethylguanidinium salt
17 were used to determine the concentration of 15. The CO stretches for
17 were observed at 1980 and 1915 cm-1
.

1100 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
Casey et al.
a variation of -0.02% during the data collection. A total of 29452
data were measured in the range 2.48 < θ < 25.00°. The data were
corrected for absorption by the empirical method56 giving minimum
and maximum transmission factors of 0.8129 and 0.9898. The data
were merged to form a set of 11117 independent data with R(int) )
0.0484.
data collection and yielded 139 peaks showing a variation of 0.02%
during the data collection. A total of 20519 data were measured in the
range 1.99 < θ < 25.00°. The data were corrected for absorption by
the empirical method55 giving minimum and maximum transmission
factors of 0.662 and 0.942. The data were merged to form a set of
6772 independent data with R(int) ) 0.0560.
The triclinic space group P1h was determined by statistical tests and
verified by subsequent refinement. The structure was solved by direct
methods and refined by full-matrix least-squares methods on F2.57
Hydrogen atom positions were initially determined by geometry and
were refined by using a riding model. Non-hydrogen atoms were refined
with anisotropic displacement parameters. A total of 773 parameters
were refined against 47 restraints and 11117 data to give wR(F2) )
0.1536 and S ) 1.080 for weights of w ) 1/[σ2(F2) + (0.0868P)2 +
The triclinic space group P1h was determined by statistical tests and
verified by subsequent refinement. The structure was solved by direct
methods and refined by full-matrix least-squares methods on F2.56
Hydrogen atom positions were initially determined by geometry and
were refined with use of a riding model. Non-hydrogen atoms of the
cation and anion were refined with anisotropic displacement parameters.
A total of 481 parameters were refined against 21 restraints and 6772
data to give wR(F2) ) 0.1895 and S ) 1.036 for weights of w ) 1/[σ2-
2
2.9570P], where P ) [Fo + 2Fc2]/3. The final R(F) was 0.0525 for
(F2) + (0.1240P)2], where P ) [Fo + 2Fc2]/3. The final R(F) was
2
the 8174 observed, [F > 4σ(F)], data. The largest shift/s.u. was 0.049
in the final refinement cycle. The final difference map had maxima
and minima of 1.207 and -0.800 e/Å3, respectively.
0.0650 for the 5106 observed, [F > 4σ(F)], data. The largest shift/s.u.
was 0.004 in the final refinement cycle. The final difference map had
maxima and minima of 1.348 and -0.920 e/Å3, respectively.
X-ray Crystal Structure Determination for NEt4+[2,5-Ph2-3,4-
Tol2(η5-C4CO)Ru(CO)2H]- (13). X-ray quality crystals were grown
by slow diffusion of hexane into a saturated THF solution of 13. A
yellow prism-shaped crystal of dimensions 0.38 × 0.26 × 0.06 mm
was selected for structural analysis. Intensity data for this compound
were collected with a Bruker SMART CCD area detector mounted on
a Bruker P4 goniometer using graphite-monochromated Mo KR
radiation (λ ) 0.71073 Å). The sample was cooled to 133(2) K. The
intensity data, which nominally covered one and a half hemispheres
of reciprocal space, were measured as a series of φ oscillation frames
each of 0.4° for 20 s/frame. The detector was operated in 512 × 512
mode and was positioned 5.00 cm from the sample. Coverage of unique
data was 91.2% complete to 25.00° in θ. Cell parameters were
determined from a nonlinear least-squares fit of 4799 peaks in the range
2.54 < θ < 26.89°. The first 50 frames were repeated at the end of
Acknowledgment. Financial support from the Department
of Energy, Office of Basic Energy Sciences, is gratefully
acknowledged. Steven Singer thanks the NIH (5T32 GM 08505)
for support under a Chemistry Biology Interface Training Grant.
Grants from NSF (CHE-9629688) and NIH (I S10 RR04981-
01) for the purchase of NMR spectrometers and from NSF
(CHE-9105497) for the purchase of X-ray instruments are
acknowledged.
Supporting Information Available: General experimental
methods, experimental details, and characterization for 9, 10,
and 14, kinetics of reduction by 8, and calculations of isotope
effects and X-ray crystallographic data for 6, 7, 11, and 13
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
(56) Sheldrick, G. M. SADABS; Program for Empirical Absorption of
Area Detector Data; University of Go¨ttingen, Germany, 1996.
(57) Sheldrick, G. M. SHELXTL Version 5 Reference Manual. Bruker-
AXS, 6300 Enterprise Dr., Madison, WI, 1994.
JA002177Z