Intramolecular trapping of an intermediate in the reduction of imines by a hydroxycyclopentadienyl ruthenium hydride: Support for a concerted outer sphere mechanism

Article abstract of DOI:10.1021/ja073370x

Reduction of imines by [2,5-Ph₂-3,4-Tol₂(η ⁵-C₄COH)]Ru(CO)₂H (1) produces kinetically stable ruthenium amine complexes. Reduction of an imine possessing an intramolecular amine was studied to distinguish between inner sphere and outer sphere mechanisms. 1,4-Bn¹⁵NH(c-C₆H₁₀)=NBn (12) was reduced by 1 in toluene-d₈ to give 85% of [2,5-Ph ₂-3,4-Tol₂(η⁴-C₄CO)](CO) ₂RuNHBn(c-C₆H₁₀)¹⁵NHBn (16-RuN, ¹⁵N), resulting from coordination of the newly formed amine to the ruthenium center, and 15% of trapping product [2,5-Ph₂-3,4-Tol ₂(η⁴-C₄CO)](CO)₂Ru ¹⁵NHBn(c-C₆H₁₀)NHBn (16-Ru¹⁵N,N), resulting from coordination of the intramolecular trapping amine. These results provide support for an outer sphere transfer of hydrogen to the imine to generate a coordinatively unsaturated intermediate, which can be trapped by the intramolecular amine. An opposing mechanism, requiring coordination of the imine nitrogen to ruthenium prior to hydrogen transfer, cannot readily explain the observation of the trapping product 16-Ru¹⁵N,N.

Full text of DOI:10.1021/ja073370x

Published on Web 08/31/2007
Intramolecular Trapping of an Intermediate in the Reduction
of Imines by a Hydroxycyclopentadienyl Ruthenium Hydride:
Support for a Concerted Outer Sphere Mechanism
Charles P. Casey,* Timothy B. Clark, and Ilia A. Guzei
Contribution from the Department of Chemistry, UniVersity of Wisconsin-Madison,
Madison, Wisconsin 53706
Received May 22, 2007; E-mail: casey@chem.wisc.edu
Abstract: Reduction of imines by [2,5-Ph2-3,4-Tol2(η5-C4COH)]Ru(CO)2H (1) produces kinetically stable
ruthenium amine complexes. Reduction of an imine possessing an intramolecular amine was studied to
distinguish between inner sphere and outer sphere mechanisms. 1,4-Bn15NH(c-C6H10)dNBn (12) was
reduced by 1 in toluene-d8 to give 85% of [2,5-Ph2-3,4-Tol2(η4-C4CO)](CO)2RuNHBn(c-C6H10)15NHBn (16-
RuN,15N), resulting from coordination of the newly formed amine to the ruthenium center, and 15% of
trapping product [2,5-Ph2-3,4-Tol2(η4-C4CO)](CO)2Ru15NHBn(c-C6H10)NHBn (16-Ru15N,N), resulting from
coordination of the intramolecular trapping amine. These results provide support for an outer sphere transfer
of hydrogen to the imine to generate a coordinatively unsaturated intermediate, which can be trapped by
the intramolecular amine. An opposing mechanism, requiring coordination of the imine nitrogen to ruthenium
prior to hydrogen transfer, cannot readily explain the observation of the trapping product 16-Ru15N,N.
Scheme 1. Concerted Outer Sphere Reduction Mechanism
Introduction
The area of ligand-metal bifunctional catalysis has led to
valuable methods for the hydrogenation of polar unsaturated
organic substrates.1-7 The first example of this type of catalyst,
reported by Shvo and co-workers in 1986,2-4 has been the
subject of detailed mechanistic investigation8 by both Ba¨ckvall’s
group9,10 and our group.11-15 For the stoichiometric reduction
(1) For reviews of ligand-metal bifunctional catalysis, see: (a) Noyori, R.;
Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40-73. (b) Noyori, R.;
Kitamura, M.; Ohkuma, T. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5356-
5362. (c) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. ReV.
2004, 248, 2201-2237. (d) Ikariya, T.; Murata, K.; Noyori, R. Org. Biomol.
Chem. 2006, 4, 393-406.
(2) Shvo, Y.; Czarkie, D.; Rahamim, Y.; Chodosh, D. F. J. Am. Chem. Soc.
1986, 108, 7400-7402.
(3) Menashe, N.; Shvo, Y. Organometallics 1991, 10, 3885-3891.
(4) Menashe, N.; Salant, E.; Shvo, Y. J. Organomet. Chem. 1996, 514, 97-
102.
of benzaldehyde by the active reducing agent, the hydroxycy-
clopentadienyl ruthenium hydride 1, we found second-order
kinetics and significant primary deuterium isotope effects for
transfer of both OH and RuH.11 We proposed an outer sphere
mechanism involving concerted addition of the acidic hydroxy
hydrogen and the hydridic ruthenium hydride to an unsaturated
substrate (Scheme 1). DFT calculations supported this concerted
outer sphere mechanism.12 In contrast, Ba¨ckvall has proposed
an alternative inner sphere mechanism involving coordination
of the heteroatom to ruthenium concurrent with η5 to η3 ring
slippage of the hydroxycyclopentadienyl ligand, which precedes
the concerted hydrogenation step (Scheme 2).16,17
(5) Persson, B. A.; Huerta, F. F.; Ba¨ckvall, J.-E. J. Org. Chem. 1999, 64, 5237-
5240.
(6) Martin-Matute, B.; Edin, M.; Boga´r, K.; Kaynak, F. B.; Ba¨ckvall, J.-E. J.
Am. Chem. Soc. 2005, 127, 8817-8825.
(7) Fransson, A.-B. L.; Xu, Y.; Leijondahl, K.; Ba¨ckvall, J.-E. J. Org. Chem.
2006, 71, 6309-6316.
(8) For mechanistic and computational work on other ligand-metal bifunctional
hydrogenation catalyst systems, see: (a) Yamakawa, M.; Ito, H.; Noyori,
R. J. Am. Chem. Soc. 2000, 122, 1466-1478. (b) Abdur-Rashid, K.;
Clapham, S. E.; Hadzovic, A.; Harvey, J. N.; Lough, A. J.; Morris, R. H.
J. Am. Chem. Soc. 2002, 124, 15104-15118. (c) Sandoval, C. A.; Ohkuma,
T.; Mun˜iz, K.; Noyori, R. J. Am. Chem. Soc. 2003, 125, 13490-13503.
(d) Casey, C. P.; Johnson, J. B. J. Org. Chem. 2003, 68, 1998-2001. (e)
Åberg, J. B.; Samec, J. S. M.; Ba¨ckvall, J.-E. Chem. Commun. 2006, 2771-
2773.
(9) Johnson, J. B.; Ba¨ckvall, J.-E. J. Org. Chem. 2003, 68, 7681-7684.
(10) Samec, J. S. M.; EÄ ll, A. H.; Åberg, J. B.; Privalov, T.; Eriksson, L.; Ba¨ckvall,
J.-E. J. Am. Chem. Soc. 2006, 128, 14293-14305.
Distinguishing between the inner (Scheme 2) and outer sphere
(Scheme 1) reduction mechanisms is challenging due to the
(11) Casey, C. P.; Singer, S. W.; Powell, D. R.; Hayashi, R. K.; Kavana, M. J.
Am. Chem. Soc. 2001, 123, 1090-1100.
(12) Casey, C. P.; Bikzhanova, G. A.; Cui, Q.; Guzei, I. A. J. Am. Chem. Soc.
2005, 127, 14062-14071.
(15) Casey, C. P.; Bikzhanova, G. A.; Guzei, I. A. J. Am. Chem. Soc. 2006,
128, 2286-2293.
(16) Csjernyik, G.; EÄ ll, A. H.; Fadini, L.; Pugin, B.; Ba¨ckvall, J.-E. J. Org.
Chem. 2002, 67, 1657-1662.
(13) Casey, C. P.; Johnson, J. B.; Singer, S. W.; Cui, Q. J. Am. Chem. Soc.
2005, 127, 3100-3109.
(14) Casey, C. P.; Johnson, J. B. J. Am. Chem. Soc. 2005, 127, 1883-1894.
9
10.1021/ja073370x CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 11821-11827
11821

A R T I C L E S
Casey et al.
Scheme 2. Ba¨ckvall’s Concerted Inner Sphere Mechanism
Scheme 3. Casey-Bikzhanova Intramolecular Trapping
Experiment
reduction of an imine by 1 in the presence of an added amine
led only to a ruthenium complex derived from the reduced imine.
This result is that predicted by the inner sphere mechanism,
but it might also be explained by formation of intermediate A
within a solvent cage that collapsed to the ruthenium amine
complex more rapidly than the amine could break its hydrogen
bond to the cyclopentadienone carbonyl and escape from the
solvent cage to be trapped by an external amine.
structural similarities between the first formed reactive inter-
mediates A and C. In the outer sphere mechanism, the first
formed intermediate upon reduction (A) is linked to the
ruthenium complex by a hydrogen bond from the carbonyl
oxygen of the cyclopentadienone ligand. In the inner sphere
mechanism, the analogous intermediate links the newly reduced
substrate to the complex by a Ru-X bond in addition to the
hydrogen bond.
To distinguish between these two explanations, we turned to
an intramolecular trapping experiment using the difunctional
amino-substituted imine 6. Reduction of imine 6 by 1 at reduced
temperatures (g-20 °C) in toluene-d8 resulted in a 50:50
mixture of kinetically stable amine complexes 7 and 8 (Scheme
3).12 Upon warming to 25 °C, the ratio of amine complexes
shifted to a thermodynamic 94:6 ratio of 8:7. Reduction of imine
6 by 1 in CD2Cl2 gave somewhat less of the intramolecular
trapping product 8 (33%).20
These mechanisms are potentially distinguishable based on
trapping experiments. The outer sphere mechanism involves the
coordinatively unsaturated intermediate A, which might be
trapped prior to alcohol coordination, while the inner sphere
mechanism leads directly to a coordinated alcohol. However,
related ruthenium alcohol complexes have never been directly
observed presumably because of rapid dissociation.18,19 Con-
sequently, recent mechanistic investigations in this area
have focused on the reduction of imines by the Shvo complex
because reduction produces kinetically stable amine complexes
(4, Scheme 1).10,12,14,15,18 The outer sphere mechanism predicts
that intermediate A (Scheme 1) might be kinetically trapped
by an added amine. In contrast, the inner sphere mechanism is
more restrictive and requires the absence of kinetic trapping
products.
These trapping experiments are consistent with an outer
sphere reduction followed by competition between coordination
of the newly formed amine and intramolecular trapping within
the solvent cage.21 For the trapping amine (shown in red,
Scheme 4) to access the coordinatively unsaturated ruthenium
center, the ligand-amine hydrogen bond of the newly formed
amine (shown in blue) must be exchanged with the trapping
amine via intermediate E (Scheme 4). The 50:50 ratio of 7:8
requires that the rate of hydrogen bond breaking (k1) be faster
than the rate of amine complexation (k2).
Because the inner sphere mechanism results from pre-
coordination of the imine to ruthenium, the inner sphere
mechanism imposes two major restrictions not required by the
outer sphere mechanism. First, in the inner sphere mechanism,
the newly formed amine is “born” in the coordination sphere
(C), which does not allow additional kinetic trapping agents to
intercept the ruthenium if the complex is kinetically stable. The
outer sphere mechanism permits (but does not require) inter-
or intramolecular trapping of coordinatively unsaturated inter-
mediate A by a trapping agent (Scheme 1). Second, stereospe-
cific trans addition to an imine is required by the inner sphere
mechanism due to pre-coordination and the required concerted
transition state for reduction.15 In contrast, the outer sphere
mechanism allows both cis and trans reduction to occur.
In 2006, Ba¨ckvall reported an analogous intramolecular
trapping experiment to probe the mechanism of imine reduction
by 1′.10 Ba¨ckvall considered the possibility that reduction of 6
occurred by an inner sphere mechanism but that the η2
unsaturated intermediate (C in Scheme 2) might have undergone
a “ruthenium migration via the aromatic system” to produce
the intramolecular trapping product 8.22 He studied the reduction
of the amine-substituted ketimine 9 in which the aromatic linker
between the two nitrogen centers was replaced by a saturated
(20) Bikzhanova, G. A. Ph.D. Thesis, University of Wisconsin-Madison, 2005.
(21) Additional control experiments were consistent with cleavage of the amine
hydrogen bond producing intermediate E and the diamine within the solvent
cage. For example, the addition of an external amine (4-isopropylaniline)
in the reduction of imine 6 only results in intramolecular and no
intermolecular trapping products. Upon warming the 50:50 mixture of 7
and 8 to 24 °C, in the presence of the external amine, p-(i-Pr)C6H4NH2, a
6:94 ratio of 7:8 was formed without the generation of the corresponding
[p-(i-Pr)C6H4NH2]-Ru complex. This result is consistent with a solvent
cage effect.12
Previous Trapping Experiments. In 2005, we reported inter-
and intramolecular trapping experiments in the reduction of
imines by ruthenium hydride complex 1. We found that the
(17) Privalov, T.; Samec, J. S. M.; Ba¨ckvall, J.-E. Organometallics 2007, 26,
2840-2848.
(18) Casey, C. P.; Bikzhanova, G. A.; Ba¨ckvall, J.-E.; Johansson, L.; Park, J.;
Kim, Y. H. Organometallics 2002, 21, 1955-1959.
(19) Casey, C. P.; Vos, T. E.; Bikzhanova, G. A. Organometallics 2003, 22,
901-903.
(22) DFT calculations by Privalov, Samec, and Ba¨ckvall17 have shown an energy
minimum for a Ru(CO)2 complex having an η2-tetraphenylcyclopentadi-
enone with one phenyl of the tetraphenyl cyclopentadienone bound η2 to
ruthenium. This suggests that ruthenium migration via the aromatic system
might be possible.
9
11822 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Trapping of Intermediate in the Reduction of Imines
A R T I C L E S
Scheme 4. Outer Sphere Mechanism Model for Reduction of Imine 12
by electron donation from the p-amino substituent24 and would
be expected to break more readily and give rise to trapping
products. To test this hypothesis, we designed imine 12, which
has an intramolecular trapping agent (Figure 1). In the outer
sphere mechanism, reaction of 12 with 1 would generate an
intermediate similar to D, having a weaker hydrogen bond than
in the Ba¨ckvall-Samec trap 9, but without the intervening
π-system of the Casey-Bikzhanova trap 6 that might cloud the
results. In addition to the weakened hydrogen bond, the newly
formed amine has increased nucleophilicity.25 As a consequence
of these two opposing effects (weakened hydrogen bond vs
increased nucleophilicity), it is very difficult to predict how
changes in the nature of the initially formed amine will affect
the amount of trapping products observed. Here, we report the
results of these new trapping experiments that provide support
for an outer sphere reduction mechanism.
Scheme 5. Ba¨ckvall-Samec Intramolecular Trapping Experiment
linker (Scheme 5).23 Reduction of 9 by ruthenium hydride 1′ in
CD2Cl2 led only to the formation of ruthenium complex 10,
resulting from coordination to newly formed amine, and no
detectable intramolecular trapping product 11 was observed.
Complex 10 was kinetically stable up to -8 °C, at which point
isomerization to the thermodynamically favored complex 11 was
observed. Ba¨ckvall concluded that this experiment provided
support for the inner sphere mechanism (Scheme 2).
New Trapping Experiments. Ba¨ckvall’s preferred interpre-
tation of these two intramolecular trapping experiments is that
both proceed by an inner sphere mechanism but the Casey-
Bikzhanova observation of a trapping product in the reduction
of 6 resulted from migration of ruthenium across the π-system.
An alternative interpretation is that both proceed by an outer
sphere mechanism but that Ba¨ckvall-Samec’s failure to observe
a trapping product in the reduction of 9 results from more rapid
amine coordination than release of the amine hydrogen bond
and formation of trapping product 11 (Scheme 6). Why might
the relative rates of coordination to nitrogen as compared to
breaking of the hydrogen bond be so much faster for intermedi-
ate D′ from reduction of 9 (Scheme 6) than for intermediate D
from reduction of 6 (Scheme 4)? We hypothesized that the
hydrogen bond to the amine in D (derived from 6) is weakened
Figure 1. Imine substrate design.
Results
Design and Synthesis of Intramolecular Trapping Agent.
The 15N-labeled pseudo-symmetric imine 12 was designed to
test this mechanistic hypothesis.26 It offers three advantages:
(1) The symmetrical diamine produced upon reduction has no
thermodynamic preference for binding either amine to ruthe-
nium. (2) 15N NMR spectroscopy provides quantitative mea-
surement of product ratios.27 (3) The absence of an aryl linker
simplifies interpretation of intramolecular trapping. Imine 12
was straightforwardly synthesized as shown in Scheme 7.28
Synthesis and Characterization of Ruthenium Diamine
Complexes. Because the reduction of imine 12 was anticipated
to produce cis and trans isomers of diamine ruthenium com-
plexes, pure samples of unenriched trans-16 and cis-16 were
independently synthesized (Scheme 8). Reductive amination of
benzaldehyde with trans-1,4-cyclohexanediamine and NaBH4
gave N,N′-dibenzyl-trans-1,4-cyclohexanediamine (trans-15).
Reaction of trans-15 with {[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru-
(23) See the Supporting Information for discussion of Ba¨ckvall’s observation10
of an intermolecular trapping of an intermediate in the reduction of an
imine by 1′ in the presence of an amine.
(24) We initially anticipated that the strength of the hydrogen bond of D′ would
be greater than that of D due to the increased electron donation from
+
p-NH2C6H4 versus Ph [p-NH2C6H4NH3 (pKa ) 6.08) is less acidic than
+
anilinium C6H5NH3 (pKa ) 4.58)]. During the review process, a referee
pointed out that the cyclohexyl substituent of D′ would weaken the hydrogen
+
bond as compared to the benzyl substituent on D [PhCH2NH3 (pKa
)
+
9.30) is more acidic than c-C6H11NH3 (pKa ) 10.64)]. These pKa values
nearly offset the anticipated difference between substrates D and D′. (a)
Brown, H. C.; McDaniel, D. H.; Hafliger, O. Dissociation Constants. In
Determination of Organic Structures by Physical Methods; Braude, E. A.,
Nachod, F. C., Eds.; Academic Press: New York, 1955; Vol. 1, p 590. (b)
Hall, H. K., Jr. J. Am. Chem. Soc. 1957, 79, 5441-5444.
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11823

A R T I C L E S
Casey et al.
Scheme 6. Outer Sphere Mechanism Model for Reduction of Imine 9
cyclohexylamine ruthenium complex; its 15N chemical shift was
Scheme 7. Synthesis of Imine 12
1
determined at natural abundance using a 2D H-15N HSQC
1
NMR experiment. The 2D H-15N HSQC NMR spectrum of
cis-16 showed a strong cross-peak for the complexed amine at
δ -311.7 and a less intense cross-peak for the free amine at δ
-334.6.30 The 2D 1H-15N HSQC NMR spectrum of trans-16
showed only a strong cross-peak for the complexed amine at δ
-312.0; no cross-peak for the free amine was observed in this
natural abundance experiment.31
Scheme 8. Syntheses of cis-16 and trans-16
Low-temperature Trapping Experiments were conduc-
ted with 15N-labeled imine 12 in toluene-d8 at -45 °C. Reac-
tion of 12 (0.11 M) with ruthenium hydride 1 occurred below
-45 °C. 15N NMR spectra at -45 °C showed complete
disappearance of 12 and broad resonances in the δ -300 to
-330 range. 15N NMR spectra obtained at 0 °C gave sharp
resonances whose ratios did not change over time at 0 °C. Four
peaks were observed in the 15N NMR spectrum at 0 °C: δ
-334.66 (21%, cis-16-RuN,15N), -324.63 (64%, trans-16-
RuN,15N), -311.29 (12%, trans-16-Ru15N,N), and -310.95
(∼3%, cis-16-Ru15N,N). Overall, this corresponds to a 76:24
ratio of trans:cis diamine complexes. About 15% of 16 has the
original 15N-labeled amine coordinated to ruthenium, while 85%
of 16 has the nitrogen derived from the newly formed amine
coordinated to ruthenium.27
The initial ratio of isomeric complexes of 16 was unchanged
up to 24 °C, demonstrating that the ratios observed at 0 °C are
kinetically determined. Upon heating at 50 °C, the ratio of
isomers of 16 changed somewhat after 2 h and reached a steady
state after 4 h (Scheme 9).32 The intensities of the four 15N NMR
resonances indicated a 40:40:10:10 ratio of trans-16-RuN,15N:
trans-16-Ru15N,N:cis-16-RuN,15N:cis-16-Ru15N,N. Thus, the15Ν-
label was evenly distributed between the free amine and the
coordinated amine in both the cis and the trans isomers. In terms
of the outer sphere mechanism, this requires breaking both the
N-Ru bond and the hydrogen bond of the amine to the dienone
carbonyl, followed by coordination of one of the equivalent
(CO)2}2 14 gave trans-16 (Scheme 8), which was characterized
by X-ray crystallography (see the Supporting Information).
Similarly, cis-16 was prepared starting with cis-1,4-cyclohex-
anediamine.
The 15N-labeled 12 (δ -327.8) provided a good model for
the 15N NMR chemical shift of a free N-benzyl cyclohexyl
amine.29 The ruthenium complex of N-benzyl isopropylamine,
[2,5-Ph2-3,4-Tol2(η4-C4CO)]Ru(CO)2(i-PrNHBn) 17 (δ -308.0),
provided a good model for the chemical shift of a N-benzyl
(25) The nucleophilicity parameters of amines (N) are known to deviate largely
from pKa values, suggesting that the rates of hydrogen-bond breakage (k1,
Scheme 4) and amine complexation (k2) are not just dependent on the pKa
value. For example, PhNH3+ (pKa ) 4.58, N ) 12.99 in water) is 0.71 pKa
units more acidic than p-CH3OC6H4NH3+ (pKa ) 5.29, N ) 16.53 in water),
but is 3.54 N units less nucleophilic. Brotzel, F.; Chu, Y. C.; Mayr, H. J.
Org. Chem. 2007, 72, 3679-3688.
(30) See the Supporting Information.
(31) Coupling of the unbound nitrogen to the NH hydrogen averages to zero
due to rapid exchange processes. This was confirmed by obtaining a non-
decoupled 15N NMR spectrum of the products of reduction of 12 that
showed no coupling between 15N and 1H for trans-16-RuN,15N.
(32) Upon heating the products of the reduction of 12 by 1 to 50 °C, three
decomposition products were observed in addition to isomers of 16. One
of the decomposition products observed upon heating was identified as
bisruthenium diamine complex {[2,5-Ph2-3,4-Tol2(η5-C4CO)]Ru(CO)2}2-
{µ-[trans-1,4-(BnNH)-c-C6H10(N HBn)]} (15N-23). See the Supporting
Information.
(26) Ba¨ckvall also used 15N NMR analysis in his trapping experiments.10
(27) An inverse gated pulse sequence and a relaxation delay of 35 s were used
to allow quantitative comparisons of 15N NMR integrals.
(28) A similar strategy was employed by Ba¨ckvall for the synthesis of
imine 9.10
(29) 15N NMR chemical shifts are reported on the δ scale with respect to
nitrobenzene as the standard.
9
11824 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Trapping of Intermediate in the Reduction of Imines
A R T I C L E S
Scheme 9. Trapping Experiment: Reduction of Imine 12 by 1
Scheme 10. Reduction of Imine 12 by 1 in the Presence of
Benzylamine
Reduction of imine 12 by complex 1 was also examined in
CD2Cl2 at -20 °C. Four peaks were observed in the 15N NMR
spectrum at -20 °C: δ -335.2 (9%, cis-16-RuN,15N), -322.7
(82%, trans-16-RuN,15N), -311.9 (7%, trans-16-Ru15N,N), and
-311.3 (∼2%, cis-16-Ru15N,N). This corresponds to an 89:11
ratio of trans:cis isomers, which is higher than that seen in
toluene-d8. In CD2Cl2, more of the nitrogen coordinated to
ruthenium was derived from the newly formed amine (91%).
The decreased amount of 16 with the 15N-labeled amine
coordinated to ruthenium (9%) in CD2Cl2 is similar to the
decreased quantity of trapping product 8 observed from imine
6 (Scheme 3, 50% in toluene-d8 and 33% in CD2Cl2).
Discussion
The possibility that the formation of trapping product 16-
Ru15N,N is an artifact of the experimental conditions and results
from a small amount of rearrangement of the 16-RuN,15N was
given serious consideration. If the reaction mixture was
inadvertently warmed during insertion into the pre-cooled NMR
probe, a small amount of rearrangement of 16-RuN,15N to 16-
Ru15N,N could have occurred upon warming. This possibility
can be excluded because the amine complex rearranges ex-
tremely slowly at room temperature and requires heating at
50 °C for 4 h to reach equilibrium (Scheme 9).
The kinetic formation of trans-16-Ru15N,N and cis-16-
Ru15N,N requires the intervention of a coordinatively unsatur-
ated intermediate and is consistent with an outer sphere reduction
mechanism (Scheme 9). These trapping products cannot result
from “ruthenium migration via the aromatic system”10 as
Ba¨ckvall suggested as a route to intramolecular trapping product
8 in an inner sphere mechanism. The observation of these
trapping products provides strong support for an outer sphere
mechanism (Scheme 1) and is inconsistent with an inner sphere
mechanism involving hydroxycyclopentadienyl ring slippage
(Scheme 2).
Because the coordinatively unsaturated intermediate is not
intermolecularly trappable, coordination of the newly formed
amine to ruthenium must be much faster than diffusion from
the solvent cage.12 The presence of a hydrogen bond from the
newly formed amine to the dieone carbonyl could help to explain
slow diffusion of the amine from the solvent cage.12 In the case
of intramolecular trapping agents, we need to explain why more
than one-half of the product is derived from coordination of
the newly formed amine. Again, hydrogen bonding of the newly
formed amine to the dienone carbonyl can explain selective
coordination of the new amine.12 For intramolecular trapping,
the hydrogen bond must be broken to give a solvent cage
containing the unsaturated intermediate and a free diamine.
nitrogens of the diamine. The rate of isomerization of dialkyl
amine complex 16 is slower than that seen earlier for the aryl,
alkyl amine complexes resulting from reduction of imines 6
and 9, which occurred at lower temperatures. This is attributed
to greater basicity of the dialkyl amine. The change in the 15N
label distribution upon heating further confirms the kinetic nature
of the initial ratio of isomers.
It was also noted that the ratio of trans:cis isomers increased
from 3:1 to 4:1 upon heating. This was independently investi-
gated by heating the unlabeled cis-16 in toluene-d8 at 50 °C for
2 h. Formation of a ∼4:1 equilibrium mixture of trans-16:cis-
1
16 was observed by H NMR spectroscopy. This cis-trans
equilibration is best explained by reversible dehydrogenation
of cis-16 back to 12 and 1. We have previously observed
reversible hydrogenation of imines by 1 as a required process
in imine isomerization.14,15,33
The reduction of imine 12 by 1 in the presence of benzy-
lamine (1:1 ratio) in toluene-d8 was studied to determine if
intermolecular kinetic trapping occurred at low temperatures
1
(Scheme 10). H NMR spectroscopy showed less than 5% of
the benzylamine complex 18 (low intensity of resonance at δ
3.1) was formed. Upon heating at 40 °C for 1 h, approximately
20% of 18 was formed concurrent with an increase from 17%
to 27% of products with the 15N label bound to ruthenium. Thus,
the rate of intramolecular scrambling of 15N-label is similar to
the rate of intermolecular amine exchange. Previously, we had
seen more rapid intramolecular amine exchange between 7 and
8 than intermolecular amine exchange with p-isopropylaniline.12
(33) For a more detailed discussion of benzylamine as an intermolecular trap
in the reduction of imine 12 by 1, see the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11825

A R T I C L E S
Casey et al.
Scheme 11. Partitioning of Unsaturated Intermediates between Hydrogen Bond Breaking (k2) and Amine Coordination (k2)
Scheme 12. Pathway to Trapping Products in Inner Sphere
Mechanism
Hydrogen bonding to either amine can then occur within the
solvent cage followed by coordination to ruthenium. Escape of
the diamine from the solvent cage therefore requires both
breaking of hydrogen-bonding interactions and diffusion
apart.
For the three different intramolecular trapping experiments,
the product ratios depend on the relative rates of coordination
of the newly formed amine (k2) and breaking of the hydrogen
bond to the newly formed amine (k1). Both of these rates should
be faster for a more electron-rich amine, which would be both
more nucleophilic and also a weaker hydrogen-bond donor.
While both k1 and k2 are expected to increase with increased
electron-rich amines, Mayr has found that the correlation
between the nucleophilicity parameter N and the pKaH value is
poor.25 Therefore, it is impossible to predict the dependence of
this rate ratio (and the amount of trapping product) on the
structure of the amine.
The extent of intramolecular trapping product formation
provides a measure of the relative rates of amine coordination
(k2) and of hydrogen bond breaking (k1) (Scheme 11). For the
reduction of 12 by 1 in toluene-d8, the 85:15 ratio 16-RuN,15N:
16-Ru15N,N is explained by 70% immediate coordination of
the newly formed amine and 30% breaking of hydrogen bond
(to form E′′) that can result in coordination of either diamine
nitrogen (k2′′/k1′′ ) 2.3) (Scheme 11). For the reduction of
Ba¨ckvall’s imine 9 by 1′ in CD2Cl2, the observation of only
the product (10) from coordination of the newly formed amine
suggests that coordination of the amine is much faster than
breaking of the relatively strong hydrogen bond to the aryl amine
(k2′/k1′ . 1). The strength of the hydrogen bond allows
coordination before hydrogen bond breaking occurs. For the
reduction of Casey and Bikzhanova’s imine 6 in toluene-d8, the
observation of a 50:50 ratio of 7:8 indicates that amine
coordination is a bit slower than hydrogen bond breaking (k2/
k1 < 1). It should also be noted that less trapping product was
observed from 6 in CD2Cl2 than in toluene-d8, suggesting that
k2/k1 is larger in CD2Cl2.34
Additional Constraints on the Inner Sphere Mechanism
To Account for Observed Trapping Products. While the outer
sphere mechanism provides a consistent explanation for all of
the data that have been reported, an inner sphere mechanism
requires a competing process for the generation of 9-15%
trapping products 16-Ru15N,N from the reduction of imine 12.
Ba¨ckvall has suggested that the initially formed η2-16 e-
ruthenium amine complex (H, Scheme 12) could dissociate the
amine, followed by re-coordination of the alternative amine
(34) The formation of complexes 16-RuN,15N and 16-Ru15N,N from intermedi-
ate E′′ could also proceed directly from E′′ to the respective amine complex
without a hydrogen-bonded intermediate [D′′ or F′′ (analogous to F′ in
Scheme 6)]. This variation in the mechanism does not change the
conclusions described in the text.
9
11826 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Trapping of Intermediate in the Reduction of Imines
A R T I C L E S
1
Scheme 13. Stereochemistry of Imine Reduction
to-bulb distillation (135-145 °C/0.08 mmHg). H NMR (C7D8, 500
MHz) δ 7.47 (d, J ) 8.0, 2H), 7.30 (d, J ) 8.0, 2H), 7.26 (m, 4H),
7.13 (m, 2H), 4.47 (d, J ) 15.7, 1H), 4.44 (d, J ) 15.4, 1H), 3.57 (d,
J ) 13.6, 1H), 3.56 (d, J ) 13.2, 1H), 2.63 (dt, J ) 11.5, 5.2, 1H),
2.48 (ddd, J ) 12.2, 8.5, 3.5, 1H), 2.36 (m, 1H), 2.15 (m, 1H), 1.73
(m, 1H), 1.3 (m, 1H), 1.57 (m, 1H), 1.32 (m, 1H), 1.12 (m, 1H), 0.72
(br s, 1H). 13C NMR (C6D6, 125 MHz) δ 172.0, 142.1, 128.93, 128.92,
128.69, 128.67, 128.5, 127.5, 127.0, 54.9 (d, J ) 4.4), 54.8, 51.6 (d, J
) 4.4), 37.3 (d, J ) 1.4), 33.7 (d, J ) 2.1), 32.9 (d, J ) 2.9), 26.0 (d,
J ) 1.4). 15N NMR (C7D8, 50.7 MHz) δ -327.8. IR (thin film) 3060,
3026, 2925, 2856, 1658, 1494, 1452 cm-1. HRMS (ESI) m/z calcd for
C20H25N15N (M + H)+ 294.1988, found 294.1979.
Scheme 14. Reversible Imine Reduction
Low-Temperature Reduction of Imine 12 by 1 in Toluene-d8. A
solution of [2,5-Ph2-3,4-Tol2(η4-C4COH)]}Ru(CO)2H (1) in toluene-
d8 was prepared as previously described11 by heating a solution of {-
[2,5-Ph2-3,4-Tol2(η5-C4CO)]2H}Ru2(CO)4(µ-H) (0.023 g, 0.020 mmol)
in THF-d8 (0.4 mL) in a medium-walled resealable NMR tube under 1
atm H2 at 90 °C for 14 h. Solvent was evaporated under vacuum, and
the residue was dissolved in toluene-d8 (0.15 mL). The solution of 1
in toluene-d8 was frozen in liquid nitrogen, and a solution of imine 12
(0.20 mL, 0.040 mmol, 0.2 M 12 in toluene-d8) was added under
nitrogen gas. The tube was sealed, warmed to -78 °C, and inserted
into a pre-cooled (-45 °C) NMR spectrometer. 15N NMR spectra at
-45 °C showed three broad resonances: δ -334.6 (cis-16-RuN,15N),
-324.2 (trans-16-RuN,15N), -310.2 (trans-16-Ru15N,N). The NMR
probe was warmed to 0 °C, and a second 15N NMR spectrum showed
four resonances: δ -334.7 (21%, cis-16-RuN,15N), -324.6 (64%,
trans-16-RuN,15N), -311.3 (12%, trans-16-Ru15N,N), -311.0 (∼3%,
cis-16-Ru15N,N). The sealed NMR tube was then heated at 50 °C for
4 h in an oil bath. An 15N NMR spectrum at 0 °C showed the four
expected resonances, δ -334.7 (10%, cis-16-RuN,15N), -324.6 (40%,
trans-16-RuN,15N), -311.3 (40%, trans-16-Ru15N,N), -311.0 (10%,
cis-16-Ru15N,N), and three new resonances, δ -310.1 due to {[2,5-
Ph2-3,4-Tol2(η5-C4CO)]Ru(CO)2}2{µ-[trans-1,4-(Bn15NH)-c-C6H10-
(NHBn)]} 15N-23 (about the same integration as the δ -324.6 and
-311.3 resonances), and two unassigned resonances at δ -324.4 (about
one-half the integration of the δ -324.6 and -311.3 resonances) and
-312.7 (about one-third the integration of the δ -324.6 and -311.3
resonances).
nitrogen (Scheme 12). We believe that this process would not
be kinetically competent because it requires extremely rapid
cleavage of the Ru-N bond of H at 0 °C and generation of a
very high-energy 14 e- ruthenium complex I. It is highly
unlikely that dissociation of the Ru-N bond of H would be
much faster than the dissociation of amine from 16-RuN,15N,
which requires heating at 50 °C.
Very High Reactivity of the Coordinatively Unsaturated
Intermediate. The coordinatively unsaturated intermediate D,
with an amine hydrogen bonded to the dienone carbonyl,
displays unusually high reactivity. The collapse of this inter-
mediate by coordination of nitrogen to ruthenium occurs faster
(for D′ and D′′) than breaking the weak hydrogen bond between
the amine and a dienone carbonyl and much faster than hydrogen
bond breaking followed by diffusion apart. Earlier we demon-
strated that hydrogenation of an imine occurred with some
selectivity for trans addition of dihydrogen (Scheme 13). This
stereospecificity requires that the coordination of the newly
formed amine to ruthenium occurs more rapidly than lone pair
inversion at the nitrogen center. In the case of the reduction of
N-alkyl imines by 1, we found isomerization and deuterium
scrambling evidence for reversible imine hydrogenation (Scheme
14). Reversible hydrogenation requires that unsaturated inter-
mediate D transfer hydride back from carbon to ruthenium faster
than amine coordination to nitrogen.
Acknowledgment. Financial support was provided by the
Department of Energy, Office of Basic Energy Sciences, the
National Science Foundation (CHE-0209476), and the National
Institutes of Health in the form of a NRSA fellowship to T.B.C.
(GM077962-01). Grants from the NSF (CHE-9629688) and NIH
(I S10 RR04981-01) for the purchase of NMR spectrometers
are acknowledged. We thank Dr. Galina A. Bikzhanova for
insightful discussions, Ms. Lara Spencer for assistance with
X-ray crystallography, and Drs. Charles Fry and Monica Ivancic
for assistance with NMR spectrometry.
Conclusion
Intramolecular trapping of an intermediate in the reduction
of imine 12 by ruthenium hydride 1 provides strong support
for an outer sphere mechanism for imine reduction. We infer
that the reduction of other polar unsaturated substrates such as
ketones and aldehydes proceeds by similar mechanisms.
Supporting Information Available: General experimental
information, syntheses of 13, trans-15, and cis-15, and ruthe-
nium complexes trans-16, cis-16, and 17, selected 15N NMR
spectra, and X-ray crystal structures of trans-16 and 23. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Experimental Section
1,4-BnNd(c-C6H10)15NHBn (12). A solution of ketone 13 (0.750
g, 3.69 mmol) and benzylamine (0.480 mL, 4.43 mmol) in 10 mL of
benzene containing 4 Å mol sieves (ca. 1 g) was heated at 50 °C for
88 h. The reaction mixture was filtered through Celite and concentrated
in vacuo. Imine 12 was obtained as a yellow oil (0.40 g, 37%) by bulb-
JA073370X
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11827