A mechanism displaying autocatalysis: The hydrogenation of acetophenone catalyzed by RuH(S-binap)(app) where app is the amido ligand derived from 2-Amino-2-(2-pyridyl)propane

Article abstract of DOI:10.1021/om700849w

The 2-(aminomethyl)pyridine (ampy) ligand is known to activate ruthenium complexes for the catalytic hydrogenation of ketones. Here we prepare well-defined catalysts using the new ligand 2-amino-2-(2pyridyl)propane (appH) in order to elucidate the role of the pyridyl group. The ligand has two methyl groups on the a-carbon to block β-hydride elimination reactions. It reacts with RuHCl(S-binap)(PPh₃) to produce the orange-yellow complex RuHCl(S-binap)(appH) (2). In the presence of a strong base (KO′Bu), complex 2 is converted into an active catalyst for the H₂- hydrogenation of acetophenone in benzene under mild conditions (20°C, 5 atm H₂). Solutions of 2 rapidly react with KO1Bu under an argon atmosphere to produce a deep red amidohydrido complex RuH(S-binap)(app) (3), which is an active catalyst. A crystal structure determination of 3 represents the first structure of a Ru - binap hydrido-amido complex. It reveals a five-coordinate Ru(II) center with a short Ru-N(amido) distance (1.962(3) A) and a trigonal planar geometry at the amido nitrogen. The kinetic experiments using 3 as a catalyst and acetophenone as a substrate in benzene show that the rate of 1-phenylethanol production is dependent on both catalyst and H₂ concentrations. These results parallel the behavior of the conventional Noyori-type Ru(II) catalysts with diamine ligands. However, unique features of catalysis with 3 are as follows: (1) the formation of a dihydride is thermodynamically unfavorable at 1 atm H₂, 20°C; (2) the rate shows a dependence on the product concentration since it increases as the product builds up during the reaction in an autocatalytic fashion. A significant increase in the initial rate is observed when a critical concentration of rac-1-phenylethanol is present at the beginning of the reaction. The addition of 2-propanol in benzene raises the rate as well, and the fastest H ₂-hydrogenation is achieved if 2-propanol is used as a solvent. This alcohol effect is favored by the pyridyl ligand app since it was not observed for the similar catalyst RuH(NHCMe₂CMe₂NH ₂)(binap). While 3 is an exceptional catalyst for H ₂-hydrogenation in 2-propanol (TOF > 6700 h^-1 at 20°C, 5 atm H₂), it has a lower activity in transfer hydrogenation from the same solvent under comparable conditions (TOF 110 h ^_1 at 20°C, 1 atm Ar). DFT calculations on the model amido complex Ru(H)(PH₃)₂(HNCH₂C₅H ₄N) (4) confirm that the splitting of H₂ to give the trans dihydride is the turnover-limiting step and lies 9 kcal/mol in free energy above the transition state for the ketone hydrogenation step. The formation of the dihydride is entropically unfavorable. The theoretical activation barrier for H2 splitting is lowered by 5 kcal/mol by an alcohol-assisted mechanism but still remains higher in energy than the ketone hydrogenation step. This latter step can also be alcohol-assisted and can result in a different ee in the product alcohol than without alcohol assistance, as observed experimentally for reactions using 2-propanol versus benzene as the solvent. With alcohol present, an alkoxohydridoruthenium(II) complex is calculated to be the catalyst resting state.

Full text of DOI:10.1021/om700849w

Organometallics 2007, 26, 5987-5999
5987
A Mechanism Displaying Autocatalysis: The Hydrogenation of
Acetophenone Catalyzed by RuH(S-binap)(app) Where app Is the
Amido Ligand Derived from 2-Amino-2-(2-pyridyl)propane
Alen Hadzovic, Datong Song, Christina M. MacLaughlin, and Robert H. Morris*
DaVenport Laboratory, Department of Chemistry, UniVersity of Toronto, 80 St. George Street,
Toronto, Ontario M5S 3H6, Canada
ReceiVed August 23, 2007
The 2-(aminomethyl)pyridine (ampy) ligand is known to activate ruthenium complexes for the catalytic
hydrogenation of ketones. Here we prepare well-defined catalysts using the new ligand 2-amino-2-(2-
pyridyl)propane (appH) in order to elucidate the role of the pyridyl group. The ligand has two methyl
groups on the R-carbon to block â-hydride elimination reactions. It reacts with RuHCl(S-binap)(PPh3) to
produce the orange-yellow complex RuHCl(S-binap)(appH) (2). In the presence of a strong base (KOt-
Bu), complex 2 is converted into an active catalyst for the H2-hydrogenation of acetophenone in benzene
under mild conditions (20 °C, 5 atm H2). Solutions of 2 rapidly react with KOtBu under an argon
atmosphere to produce a deep red amidohydrido complex RuH(S-binap)(app) (3), which is an active
catalyst. A crystal structure determination of 3 represents the first structure of a Ru-binap hydrido-amido
complex. It reveals a five-coordinate Ru(II) center with a short Ru-N(amido) distance (1.962(3) Å) and
a trigonal planar geometry at the amido nitrogen. The kinetic experiments using 3 as a catalyst and
acetophenone as a substrate in benzene show that the rate of 1-phenylethanol production is dependent on
both catalyst and H2 concentrations. These results parallel the behavior of the conventional Noyori-type
Ru(II) catalysts with diamine ligands. However, unique features of catalysis with 3 are as follows: (1)
the formation of a dihydride is thermodynamically unfavorable at 1 atm H2, 20 °C; (2) the rate shows a
dependence on the product concentration since it increases as the product builds up during the reaction
in an autocatalytic fashion. A significant increase in the initial rate is observed when a critical concentration
of rac-1-phenylethanol is present at the beginning of the reaction. The addition of 2-propanol in benzene
raises the rate as well, and the fastest H2-hydrogenation is achieved if 2-propanol is used as a solvent.
This “alcohol effect” is favored by the pyridyl ligand app since it was not observed for the similar catalyst
RuH(NHCMe2CMe2NH2)(binap). While 3 is an exceptional catalyst for H2-hydrogenation in 2-propanol
(TOF > 6700 h-1 at 20 °C, 5 atm H2), it has a lower activity in transfer hydrogenation from the same
solvent under comparable conditions (TOF 110 h-1 at 20 °C, 1 atm Ar). DFT calculations on the model
amido complex Ru(H)(PH3)2(HNCH2C5H4N) (4) confirm that the splitting of H2 to give the trans dihydride
is the turnover-limiting step and lies 9 kcal/mol in free energy above the transition state for the ketone
hydrogenation step. The formation of the dihydride is entropically unfavorable. The theoretical activation
barrier for H2 splitting is lowered by 5 kcal/mol by an alcohol-assisted mechanism but still remains
higher in energy than the ketone hydrogenation step. This latter step can also be alcohol-assisted and can
result in a different ee in the product alcohol than without alcohol assistance, as observed experimentally
for reactions using 2-propanol versus benzene as the solvent. With alcohol present, an alkoxohydridoru-
thenium(II) complex is calculated to be the catalyst resting state.
the amido N-H bond.4,8,9 They have been recognized as
intermediates in amination and hydroamination of arylha-
lides.10,11 Ruthenium amido complexes have featured promi-
nently as active catalysts or catalyst intermediates in hydrogena-
tion7,12-20 and transfer hydrogenation16,21-24 of ketones and in
Introduction
Transition metal amido complexes play an important role in
both catalysis (as catalysts and/or intermediates in catalysis) and
stoichiometric reactions.1-3 For example, they function as strong
bases in stoichiometric carbon-hydrogen bond activation
reactions,4-6 dihydrogen splitting,7 and insertion reactions into
(6) Feng, Y.; Lail, M.; Foley, N. A.; Gunnoe, T. B.; Barakat, K. A.;
Cundari, T. R.; Petersen, J. L. J. Am. Chem. Soc. 2006, 128, 7982-7994.
(7) Abdur-Rashid, K.; Faatz, M.; Lough, A. J.; Morris, R. H. J. Am.
Chem. Soc. 2001, 123, 7473-7474.
(8) Rais, D.; Bergman, R. G. Chem.-Eur. J. 2004, 10, 3970 - 3978.
(9) Fox, D. J.; Bergman, R. G. J. Am. Chem. Soc. 2003, 125, 8984-
8985.
(10) Hartwig, J. F.; Richards, S.; Baranano, D.; Paul, F. J. Am. Chem.
Soc. 1996, 118, 3626-3633.
(11) Alcazar-Roman, L. M.; Hartwig, J. F.; Rheingold, A. L.; Liable-
Sands, L. M.; Guzei, I. A. J. Am. Chem. Soc. 2000, 122, 4618-4630.
(12) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40-73.
* Corresponding author. E-mail: rmorris@chem.utoronto.ca.
(1) Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. ReV. 1989, 95,
1-40.
(2) Fulton, J. R.; Holland, A. W.; Fox, D. J.; Bergman, R. G. Acc. Chem.
Res. 2002, 35, 44-56.
(3) Gunnoe, T. B. Eur. J. Inorg. Chem. 2007, 1185-1203.
(4) Fulton, J. R.; Bouwkamp, M. W.; Bergman, R. G. J. Am. Chem.
Soc. 2000, 122, 8799-8800.
(5) Eckert, N. A.; Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Inorg.
Chem. 2004, 43, 3306 -3321.
10.1021/om700849w CCC: $37.00 © 2007 American Chemical Society
Publication on Web 10/20/2007

5988 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
Scheme 1. â-Hydride Elimination from an Amido Ligand
Scheme 2. Proposed Mechanism of Action of Ruthenium
Hydridoamine Complexes in the Catalytic Hydrogenation of
Ketones
conjugate addition reactions.24-26
A major concern in the preparation of coordinatively unsatur-
ated transition metal amido complexes is their ability to undergo
â-hydride elimination from the carbon backbone of the ligand,
generating a hydrido-imine complex (Scheme 1).7,10,27,28 In order
to overcome this problem, ligands without hydrogen atoms on
carbon R to the nitrogen donor have been used. The earliest
examples of such ligands include N(SiMe3)2-, NHPh-, and
NPh2-.1 Other, more elaborate mono- and polydentate ligands
followed.29,30
Our group has reported the isolation and characterization of
Ru(II) amido complexes Ru(H)(P2)(H2NCMe2CMe2NH), P2 )
R-binap or (PPh3)2,7,13 that have an amido ligand with the
R-positions blocked with methyl groups. These complexes
proved to constitute a part of a catalytic cycle for hydrogenation
of ketones to alcohols in Noyori-type catalytic systems (Scheme
2). In this cycle, the hydridoamido complex heterolytically
cleaves the H2 molecule over the Ru-N amido bond to produce
a trans dihydride amine species. The trans dihydride complex
that is formed in this way transfers a hydride and a proton (an
equivalent of H2) to the carbonyl group of the ketone to produce
alcohol and regenerates the hydridoamido complex. This
catalytic cycle explains well the diamine or “N-H effect”
previously observed by Noyori and co-workers.22,31 Similar
hydrido-amido complexes, derived from nonblocked diamines,
proved to be unstable and could be observed only in certain
NMR experiments along with their decomposition products.32,33
There are several reports in the recent literature of an
increased catalytic activity of ketone hydrogenation catalysts
when a conventional diamine ligand is replaced with 2-(ami-
nomethyl)pyridine (ampy). Noyori and co-workers have ob-
served that RuXY(S-tolbinap)(ampy) (X, Y ) Cl; X ) H, Y )
BH4) are excellent (pre)catalysts for hydrogenation of
bulky, deactivated tert-alkyl ketones for which typical RuX2-
(diphosphine)(diamine) catalysts are less effective.34 This
increased activity of ampy over diamine catalysts has been
explained by the presence of a flat pyridine ring that lowers a
nonbonding repulsion between the incoming substrate and the
backbone of the catalyst. Baratta and co-workers developed the
catalyst precursors RuCl2(diphosphine)(ampy)35 and RuCl-
(CNN)(diphosphine), where CNN is 6-(4′-methylphenyl)-2-
pyridylmethylamine, a tridentate ligand forming orthometalated
Ru(II) complexes.36 These are remarkably active in transfer
hydrogenation of ketones from 2-propanol at elevated temper-
atures. We have described trans-RuHCl(PPh3)2(ampy) and cis,-
cis-Ru(H)2(PPh3)2(ampy), precursors to the catalyst system that
is much more active in the H2-hydrogenation of acetophenone
than those obtained from related complexes trans-RuHCl(PPh3)2-
(R,R-dach) and cis,cis-Ru(H)2(PPh3)2(R,R-dach) (R,R-dach )
1R,2R-diaminocyclohexane).32
Some of us previously designed the ligand 2-phenyl-6-(2-
aminoisopropyl)pyridine, with methyl groups on the R-carbon
of the amino group to allow the synthesis of stable Pd and Pt
amido compounds.37 Here we report the preparation of the novel
ligand 2-amino-2-(2-pyridyl)propane (appH), an analogue of
ampy with two methyl groups on the carbon R to the amine.
This ligand design enables the isolation of the Ru(II) hydrido-
amido complex RuH(S-binap)(app) and its full characterization.
The other known complex of this type, RuH(R-binap)(NHCMe2-
CMe2NH2), has not been isolated in a crystalline form.7 This
provides us with a well-defined catalytic system to study and
model with DFT calculations.
(13) 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.
(14) Sandoval, C. A.; Ohkuma, T.; Mun˜iz, K.; Noyori, R. J. Am. Chem.
Soc. 2003, 125, 13490-13503.
(15) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. ReV.
2004, 248, 2201-2237.
(16) Mun˜iz, K. Angew. Chem., Int. Ed. 2005, 44, 6622-6627.
(17) Clapham, S. E.; Morris, R. H. Organometallics 2005, 24, 479-
481.
(18) Sandoval, C. A.; Ohkuma, T.; Utsumi, N.; Tsutsumi, K.; Murata,
K.; Noyori, R. Chem. Asian J. 2006, 1, 102-110.
(19) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2006, 128,
13700-13701.
(20) Ohkuma, T.; Tsutsumi, K.; Utsumi, N.; Arai, N.; Noyori, R.; Murata,
K. Org. Lett. 2007, 9, 255-257.
(21) Haack, K. J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R. Angew.
Chem., Int. Ed. Engl. 1997, 36, 285-288.
(22) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122,
1466-1478.
Results and Discussion
(23) Boubekeur, L.; Ulmer, S.; Ricard, L.; Mezailles, N.; LeFloch, P.
Organometallics 2006, 25, 315-317.
Synthesis of the appH Ligand. The ligand was prepared
according to Scheme 3 following a route based partly on the
previously reported synthesis of 2-aminomethyl-6-phenylpyri-
dine.37 Refluxing the alcohol 1a in a mixture of CH3CN/BF3-
(24) Ikariya, T.; Murata, K.; Noyori, R. Org. Biomol. Chem. 2006, 4,
393-406.
(25) Watanabe, M.; Ikagawa, A.; Wang, H.; Murata, K.; Ikariya, T. J.
Am. Chem. Soc. 2004, 126, 11148-11149.
(26) Clapham, S. E.; Guo, R.; Zimmer-DeIuliis, M.; Rasool, N.; Lough,
A.; Morris, R. H. Organometallics 2006, 25, 5477-5486.
(27) Diamond, S. E.; Mares, F. J. Organomet. Chem. 1977, 142, C55 -
C57.
(33) Abbel, R.; Abdur-Rashid, K.; Faatz, M.; Hadzovic, A.; Lough, A.
J.; Morris, R. H. J. Am. Chem. Soc. 2005, 127, 1870-1882.
(34) Ohkuma, T.; Sandoval, C. A.; Srinivasan, R.; Lin, Q.; Wei, Y.;
Mun˜iz, K.; Noyori, R. J. Am. Chem. Soc. 2005, 127, 8288-8289.
(35) Baratta, W.; Herdtweck, E.; Siega, K.; Toniutti, M.; Rigo, P.
Organometallics 2005, 24, 1660-1669.
(28) Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7010-7011.
(29) Cooper, M. K.; Downes, J. M. Inorg. Chem. 1978, 17, 880-884.
(30) Fryzuk, M. D.; MacNeil, P. A.; Rettig, S. J.; Secco, A. S.; Trotter,
J. Organometallics 1982, 1, 918 - 930.
(31) Ohkuma, T.; Ooka, H.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J.
Am. Chem. Soc. 1995, 117, 2675-2676.
(32) Abdur-Rashid, K.; Abbel, R.; Hadzovic, A.; Lough, A. J.; Morris,
R. H. Inorg. Chem. 2005, 44, 2483-2492.
(36) Baratta, W.; Chelucci, G.; Gladiali, S.; Siega, K.; Toniutti, M.;
Zanette, M.; Zangrando, E.; Rigo, P. Angew. Chem., Int. Ed. 2005, 44,
6214-6219.
(37) Song, D.; Morris, R. H. Organometallics 2004, 23, 4406-4413.

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5989
Scheme 3. Preparation of the appH Ligand
Scheme 5. Preparation of the Complex
Ru(H)Cl(S-binap)(appH) (2) from RuCl2(S-binap)(PPh3)
the last PPh3 ligand with a desired diamine. As such, it can be
time-consuming and the isolation of the intermediates is usually
accompanied by losses of material. We looked for a new,
simpler route to prepare the Ru(H)Cl(diphosphine)(diamine)
precursors in order to shorten the preparation time and increase
the overall yields.
The chlorohydrido complex Ru(H)Cl(S-binap)(appH) (2) can
be prepared in a two-step, one-pot reaction starting from RuCl2-
(S-binap)(PPh3)41 (Scheme 5). In the first step, the starting
dichloro complex is reacted with hydrogen gas under basic
conditions to produce a deep red solution of the triphenylphos-
phine complex Ru(H)Cl(S-binap)(PPh3) and the salt [Et3NH]-
Cl as a precipitate. This reaction presumably proceeds via a
dihydrogen intermediate, RuCl2(η2-H2)(S-binap)(PPh3), in which
the dihydrogen ligand is acidic enough to be deprotonated by
the base NEt3, as proposed in the formation of Ru(H)Cl(PPh3)3
from RuCl2(PPh3)3 under analogous conditions.42 The triph-
enylphosphine complex is not isolated from this reaction, but
rather is reacted with appH to produce Ru(H)Cl(S-binap)(appH)
(2). Complex 2 can be isolated as an orange-yellow solid.
Complex 2 exists as two diastereomers in approximately a
9:1 ratio on the basis of the integration in the 31P NMR and 1H
spectra. The hydride-decoupled 31P NMR spectrum has two sets
of resonances in an AB pattern consistent with the presence of
the two isomers. The hydride region of the 1H NMR spectrum
has two doublets of doublets at -15.8 and -16.3 ppm for the
major and minor diastereomer, respectively. The magnitude of
Scheme 4. General Preparation of the Complexes
trans-RuHCl(diphosphine)(diamine)
Et2O for 72 h gives the amide 1b in about 30% yield.38 The
bromo group can be easily removed with Raney Ni/H2 under
strongly basic conditions39 to produce N-(2-(pyridin-2-yl)propan-
2-yl)acetamide (1c) in about 65% yield. An acid-catalyzed
hydrolysis of 1c yields the new compound appH in good yield
(>90%).
Any attempt to increase the yield of 1b by changing the CH3-
CN:BF3Et2O ratio or prolonging the reflux did not result in an
increase in yield. The bromo substituent in the starting alcohol
seems to be necessary because 2-(pyridin-2-yl)propan-2-ol does
not react under similar conditions. The reason for this difference
in the reactivity is not clear at the moment. The efforts to shorten
the synthetic route by converting the alcohol to azide, which
could be subsequently reduced to the pyridyl-amine, failed. The
usual conversion of tertiary alcohols to azides using NaN3 in
acidic media did not produce the intermediate azide. The lack
of reactivity could be due to the acid protonating the pyridyl
nitrogen instead of converting the -OH group to the good
leaving group H2O.
Syntheses of Ru(II) Complexes. The ruthenium(II) chloro-
hydrido compounds of the general formula RuHCl(diphosphine)-
(diamine) have proven to be valuable precursors to the
catalytically active amido and dihydride complexes.7,13,33 These
can be prepared following the general and well-established route
in Scheme 4.40
This approach requires the preparation and isolation of the
two intermediates Ru(H)Cl(PPh3)3 and Ru(H)Cl(diphosphine)-
(PPh3) before the final product is obtained by substitution of
2
the JHP coupling constants is consistent with a hydride ligand
cis to the P donor atoms of the S-binap on ruthenium. The proton
resonances of the two methyl groups of each diastereomer
appear as sharp singlets, consistent with the different environ-
ments syn to chloride and syn to hydride. Two broad doublets
at 2.7 and 3.4 (2JHH ) 9.6 Hz) are assigned to inequivalent NH2
protons; apparently these are coincident for the two diastereo-
mers.
Complex 2 rapidly reacts with KOtBu under an argon
atmosphere, eliminating an equivalent of HCl, to produce the
deep red hydridoamido complex Ru(H)(S-binap)(app) (3) (Scheme
6). A sharp doublet of doublets in the hydride region of the 1H
2
NMR spectrum at -15.5 ppm and the magnitude of JHP
coupling constants (26.1 and 40 Hz) indicate that the hydride
ligand remains approximately cis with respect to two inequiva-
lent phosphorus nuclei. Two CH3 groups appear as two sharp
singlets at 1.33 and 1.58 ppm, exhibiting a slight upfield shift
compared to their resonances in 2. A doublet at 4.3 ppm is
assigned to the NH resonance with a 3JHP ) 4.3 Hz due to the
(38) Sjo¨berg, K. Acta Chem. Scand. 1968, 22, 1787-1790.
(39) Kammerer, H.; Horner, L.; Beck, H. Chem. Ber. 1958, 91, 1376-
1379.
(40) Abdur-Rashid, K.; Lough, A. J.; Morris, R. H. Organometallics
2001, 20, 1047-1049.
(41) Joshi, A. M.; Thorburn, I. S.; Rettig, S. J.; James, B. R. Inorg. Chim.
Acta 1992, 198-200, 283-296.
(42) Jessop, P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155-
284.

5990 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
Scheme 6. Preparation of the Hydridoamido Complex
Ru(H)(S-binap)(app)
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 3
bond (Å)
angle (deg)
Ru1-H1ru
1.48(4)
P2-Ru1-H1ru
84(2)
127(2)
74(2)
106(2)
145.2(1)
168.69(9)
77.2(1)
100.23(9)
90.97(4)
94.5(1)
Ru1-N1
Ru1-N2
Ru1-P1
Ru1-P2
1.962(3)
2.122(3)
2.234(1)
2.230(1)
N1-Ru1-H1ru
N2-Ru1-H1ru
P1-Ru1-H1ru
P2-Ru1-N1
P1-Ru1-N2
N1-Ru1-N2
N2-Ru1-P2
P1-Ru1-P2
N1-Ru1-P1
Figure 1. Structure of 3 with thermal ellipsoids shown at 50%
probability. All the hydrogen atoms except the hydride and the
amido NH are omitted for clarity.
coupling to a phosphorus trans to the amido nitrogen. This
1
assignment is based on the integration and on a H 2D COSY
experiment in which this resonance did not show coupling to
any other proton in the complex. The 31P{1H} NMR spectrum
has two doublets for two inequivalent phosphorus atoms (71
2
and 82.7 ppm, JPP ) 31 Hz). This is similar to that of the
spectrum of the complex RuH(R-binap)(NHCMe2CMe2NH2).7
As mentioned in the Introduction, amidohydrido complexes
are proposed to be within the catalytic cycle for the asymmetric
hydrogenation of polar bonds catalyzed by ruthenium binap
complexes7,12-14 but have not been structurally characterized.
The X-ray crystal structure of 3 (Figure 1 and Table 1) reveals
a five-coordinate, distorted trigonal-bipyramidal geometry at Ru,
with the P1 and pyridyl nitrogen (N2) approximately trans to
each other (P1-Ru1-N2 ) 168.69(9)°) and occupying the
apical positions. The amido nitrogen (N1)-ruthenium distance
is 1.962(3) Å, which is comparable to the 1.967(1) Å bond
length in the similar amido complex Ru(H)(PPh3)2(HNCMe2-
CMe2NH2),13 but significantly shorter than the N(amino)-Ru
distance of 2.163(2) Å in Ru(H)Cl(PPh3)2(ampy).32 The other
hydrido amido complex, RuH(PPh2C6H4CH2NCMe2CMe2-
NHCH2C6H4PPh2), has an amido nitrogen-Ru distance of
2.001(2) Å.43 The sum of angles around N1 (357.4°) is as
expected for a planar, sp2-hybridized nitrogen atom involved
in pπ(N)fdπ(Ru) dative bonding. There is a small H-Ru-P2
angle of 83.7(1)° for these σ donors across from the amido
nitrogen as observed for similar five-coordinate RuHXL3, RhH2-
XL2, and IrH2XL2 complexes.13,44
Preliminary work shows that RuCl2(app)(R-binap) can be
prepared in 78% yield from RuCl2(PPh3)(R-binap) by stirring
in toluene for 16 h (31P NMR (C6D6) 44.2 (d), 47.6 (d)) and
identified by a single-crystal structure determination of poor
quality.
Catalytic Activity. In the presence of a strong base such as
KOtBu, complex 2 acts as a precatalyst for the H2-hydrogenation
of acetophenone in benzene under mild reaction conditions. The
complete conversion of acetophenone to 1-phenylethanol is
achieved in less than 1.5 h with the substrate:catalyst:KOtBu
ratio of 300:1:25 under 10 atm of H2 and room temperature.
The alcohol is produced in 27% ee for R-1-phenylethanol. The
Figure 2. Kinetic data for hydrogenation of acetophenone catalyzed
by 3 at 20 °C in C6H6: [3] ) 0.0003 M, [acetophenone] ) 0.3 M;
constant pressure of 5 atm H2 (run 1, [); [3] ) 0.0003 M,
[acetophenone] ) 0.18 M, 5 atm H2 (run 2, 9, corresponding ee
values 0); [3] ) 0.0006 M, [acetophenone] ) 0.18 M, 5 atm H2
(run 3, *); and [3] ) 0.0003 M, [acetophenone] ) 0.18 M, 10 atm
H2 (run 4, 2).
related complex Ru(H)Cl(ampy)(PPh3)2 (ampy ) 2-aminom-
ethylpyridine) gives 98% conversion after 1.3 h at room
temperature using S:C:B ratio 709:1:17 and 5 atm H2.32 The
amido complex 3, on the other hand, is an active catalyst under
base-free conditions. Using the same S:C ratio of 300:1, the
complete hydrogenation of acetophenone to 1-phenylethanol was
accomplished within the same reaction time.
The ketone 1-acetonaphthone was also tested since this more
stereoactive substrate is often hydrogenated to the alcohol in
high ee.31,45,46 The amido complex 3 efficiently hydrogenates
this substrate but only in 23% ee (R).
Kinetic Study. Several kinetic runs were performed using 3
as the catalyst and benzene as the solvent at constant hydrogen
pressure (Figure 2). The behavior of this system, at low ketone
conversions, is qualitatively similar to that previously reported
for the amido complex Ru(H)(HNCMe2CMe2NH2)(PPh3)213 and
the dihydrides Ru(H)2(R-binap)(tmen)13 and Ru(H)2(R,R-dach)-
33
(PPh3)2 (tmen ) 2,3-diamino-2,3-dimethylbutane; R,R-dach
) (1R,2R)-1,2-diaminocyclohexane). The rate of alcohol pro-
duction increases with an increase of either the catalyst
concentration (run 2 vs run 3) or the dihydrogen concentration
(run 2 vs run 4). On the other hand, the initial rate is not
(43) Li, T.; Churlaud, R.; Lough, A. J.; Abdur-Rashid, K.; Morris, R.
H. Organometallics 2004, 23, 6239-6247.
(44) Riehl, J. F.; Jean, Y.; Eisenstein, O.; Pelissier, M. Organometallics
1992, 11, 729-737.
(45) Ralph, C. K.; Akotsi, O. M.; Bergens, S. H. Organometallics 2004,
23, 1484-1486.
(46) Chai, L. T.; Wang, W. W.; Wang, Q. R.; Tao, F. G. J. Mol. Catal.
A 2007, 270, 83-88.

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5991
Scheme 7. Proposed Simple Mechanism (see below for a
more complete mechanism) for the Hydrogenation of
Acetophenone Catalyzed by 3 in Benzene with Low
Concentrations of Alcohol (Path A) and High Concentrations
of Alcohol (Path B)
Figure 3. Data for the catalytic hydrogenation of acetophenone
(0.18 M) in C6H6 at 20 °C with [3] ) 0.0003 M under 5 atm of H2
and no additive (run 2, [), rac-1-phenylethanol added (0.04 M,
initial concentration) (run 5, 9), and 2-propanol added (0.06 M)
(run 6, 2).
such as 2-propanol and ethanol are the solvents of choice for
the hydrogenation of arylketones catalyzed by some amine-
hydrido ruthenium complexes.48 In their kinetic study of the
hydrogenation of acetophenone in 2-propanol using RuH(η1-
BH4)(S-tolbinap)(1S,2S-dpen) (1S,2S-dpen ) 1S,2S-diphenyl-
ethylenediamine) as a precatalyst, Noyori and co-workers
suggested a transition state similar to that of path B for the
regeneration of the active catalyst.14 More recently, Casey and
co-workers have established that there is ethanol-assisted loss
of the H2 molecule from Shvo’s hydroxycyclopentadienyl
ruthenium catalyst.49
dependent on the substrate concentration (run 1 vs run 2). It is
not clear for this complex, nonlinear reaction system why, later
in the reaction, the higher ketone concentration (run 1) leads to
a lower rate. However, overall these observations indicate that
the heterolytic splitting of dihydrogen is the rate-determining
step.
The data of Figure 2 show that a critical concentration of the
product has to build up in order for the rate to show a significant
increase. This critical concentration can be estimated from the
plots to be 9 mM for run 2 and 20 mM for runs 3 and 4 of
Figure 2. Therefore this product concentration seems to be a
function of both the catalyst concentration and H2 pressure, and
it roughly doubles by doubling either [Ru] or [H2]. This also
points to the alcohol’s involvement in the reaction between the
amido complex 3 and dissolved H2. As more alcohol is produced
during the reduction of the substrate, the participation of the
alcohol in the H2 splitting becomes more important and this
produces an autocatalytic effect.
If 0.04 M racemic 1-phenylethanol is added at the beginning
of the reaction, the resulting initial rate is significantly higher
(Figure 3, Table 2). Additionally, the production of 1-phenyl-
ethanol remains almost linear until 100% conversion. If 2-pro-
panol is added instead of rac-1-phenylethanol, the rate increases
again but not as much as with the reaction product as an additive.
The reason for the different effect on the rate of two alcohols
may be the difference in acidity of the alcohols, the latter more
acidic than the former. Since the proposed alcohol-assisted
splitting of the H2 molecule likely involves a simultaneous
protonation of the amido nitrogen and reprotonation of the
alcohol by the η2-H2 ligand, it is reasonable to expect that the
acidity of the alcohol would be important. This is supported by
the calculations described below.
These observations on experiments at low ketone conversions
and therefore low alcohol concentrations suggest the sequence
of events as shown in path A of Scheme 7. The amido complex
3 heterolytically cleaves the H2 molecule into a hydride on the
metal and a proton on the amido nitrogen. The trans dihydride
complex thus formed is responsible for the addition of H(-)
/
H(+) to the CdO bond of the substrate to give 1-phenethanol in
about 27% ee (R). Interestingly a solution of complex 3 in
benzene-d6, unlike the previous hydridoamido complexes,13,47
does not react with 1 atm dihydrogen on the NMR scale (35
mg of complex in 0.55 mL of benzene-d6 and 1 atm of H2).
Therefore the equilibrium constant for H2 addition according
to path A lies far to the left.
While the kinetic experiments with the dihydride catalysts
Ru(H)2(R-binap)(tmen) and Ru(H)2(R,R-dach)(PPh3)2 show a
linear production of alcohol over time, the systems with the
RuH(S-binap)(app) catalyst show a significant acceleration in
the rate as the reaction progresses. This suggests that the actual
mechanism is more complicated than the one shown by path A
of Scheme 7. The likely explanation for this observation is a
lowering of the energy barrier for the dihydrogen splitting due
to the assistance of the alcohol product. The alcohol might assist
in the dihydrogen splitting via a six-membered transition state
shown in path B of Scheme 7. In this scenario the alcohol
protonates an amido nitrogen while the H2 molecule is hetero-
lytically cleaved between a Ru(II) center and an oxygen atom
of the alkoxide. It is likely that this protonation-cleavage
process occurs simultaneously through a hydrogen-bonded
network.
The highest rate for the H2-hydrogenation of acetophenone
using 3 as a catalyst is observed in 2-propanol (run 7, Table 2).
If benzene as a solvent is replaced with 2-propanol and other
conditions are kept the same as in run 2, 95% conversion of
the substrate to alcohol is achieved in only 5 min. This converts
to a TOF of over 6750 h-1 in 2-propanol (run 7) compared to
the final TOF of about 550 h-1 in benzene (run 2), calculated
Several other groups have suggested a similar alcohol-assisted
mechanism in metal-ligand bifunctional catalysis. Ikariya and
co-workers proposed, and also confirmed by use of isotope
labeling experiments, that an alcohol-assisted dihydrogen split-
ting transition state accounts for the finding that the alcohols
(48) Ito, M.; Hirakawa, M.; Murata, K.; Ikariya, T. Organometallics 2001,
20, 379-381.
(49) Casey, C. P.; Johnson, J. B.; Singer, S. W.; Cui, Q. J. Am. Chem.
Soc. 2005, 127, 3100-3109.
(47) Rautenstrauch, V.; Hoang-Cong, X.; Churlaud, R.; Abdur-Rashid,
K.; Morris, R. H. Chem.-Eur. J. 2003, 9, 4954-4967.

5992 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
Table 2. Turnover Frequencies, tof, for the Reduction of Acetophenone with 3 as a Catalyst (In all runs: [3] ) 0.0003 M,
[acetophenone] ) 0.18 M, 5 atm H2, 20 °C; runs 1-4 are shown in Figure 2)
run
solvent
additive
H or T?a
TOF, h-1 b
ee, % (R)
25-27
1-4
benzene
benzene
benzene
2-propanol
2-propanol
none
H
H
H
H
T
550
1520
1150
6750
115
5
6
7
8
rac-C6H5CH(OH)CH3 (0.04 M)
2-propanol (0.06 M)
29
32
37
none
none
a H ) H2 hydrogenation; T ) transfer hydrogenation. b Values calculated on the basis of the conversion for the last point of each run.
functional and the SDD54 effective core potentials (ECP) for
ruthenium is based on the previous reports showing that this
combination is superior to the B3LYP functional and other
ECP.55,56 Recently, Lynch et al. have demonstrated that the
mPW1PW91 functional is also better for the prediction of the
transition states and the energy barriers.57 All other nonmetallic
elements were treated with the 6-311++G(d, p) level of theory.
For the gas-phase DFT calculations, catalyst 3 has been
replaced with the model amido system Ru(H)(PH3)2(2-NHCH2-
(C5H4N)) (4), in which the app- ligand has been replaced with
a deprotonated ampy ligand, and S-binap, with the two PH3
ligands. The optimized geometry of 4 (Figure 4) is quite similar
to that of 3. The (HNCH2C5H4N)- ligand is almost planar, with
the bridging -CH2- group displaced from the ligand plane.
The length of the Ru-N(amido) bond in 4 is calculated to be
1.972 Å, and this compares very well with the corresponding
distance of 1.962(3) Å in 3. The geometry at the amido nitrogen
in the model is trigonal planar, with the sum of the angles around
nitrogen of 357.3° (compared with 357.4° in 3). The Ru-
N(pyridine) distance in 4 is 2.116 Å, while the corresponding
one in 3 is 2.122(3) Å. The calculated Ru-H distance is
somewhat longer than that in 3 from the X-ray study: 1.573
and 1.48(4) Å for 4 and 3, respectively. However, this difference
is likely to be the result of the fact that the X-ray diffraction
method generally underestimates the metal-hydride distance.
The bond angles are reproduced with less precision than the
bond lengths. Thus, the hydride ligand is shifted toward the
PH3 trans to the amido nitrogen, much like in 3, but the
P-Ru-H angle in 4 (81.7°) is smaller than the corresponding
one in 3 (83.7(2)°). On the other hand the P2-Ru-N1 angle is
larger in 4 than in 3 (154.6° vs 145.2(1)°). However, in the last
case the difference can be a consequence of replacing a rigid,
bidentate ligand (S-binap) with the two much smaller mono-
dentate PH3 ligands. The discrepancy between the bond angles
in a real complex and its model has been reported previously
for various DFT functionals and basis sets.56
Chart 1
for the last point of run 2. Under comparable solvent transfer
hydrogenation conditions, that is, under Ar instead of H2 gas, a
conversion of only 12% was measured after 40 min. This gives
a TOF of about 115 h-1 (Table 2, run 8).
There is a similar 10-fold increase in the catalysis rate
constant on going from benzene to 2-propanol as the solvent
for the precatalyst RuHCl(PPh2C6H4CH2NHC6H10NHCH2C6H4-
PPh2) (complex C, Chart 1) containing a tetradentate P-NH-
NH-P ligand.47 The kinetic study with this catalyst revealed
that in either benzene or 2-propanol the addition of the H2 across
the Ru-N amido bond of the complex D is the turnover-limiting
step in the catalytic cycle.
The ee of the 1-phenylethanol for runs 1-6 (Figure 2, Table
2) ranges from 25 to 27% (R alcohol). A small increase to 32%
was observed for the H2-hydrogenation in 2-propanol (run 7)
and further to 37% in transfer hydrogenation (run 8). These are
somewhat higher than the ee value of 14% observed by Noyori
for the hydrogenation of acetophenone catalyzed by the system
RuCl2(S-tolbinap)(ampy)/ base34 using 2-propanol as a solvent.
We find that RuCl2(R-binap)(app) (note the use of R-binap here)
when activated with KOtBu in 2-propanol displays a TOF of
about 50 h-1 under the conditions listed in Table 2 to produce
the S alcohol in 34% ee. Therefore the presence of the methyl
groups of the appH ligand accounts for the difference in ee
between the ampy and appH systems.
DFT Calculations. Theoretical studies have proven to be a
useful tool in elucidating the mechanistic aspects of the catalytic
hydrogenation of the polar, unsaturated bonds. Noyori and co-
workers used this approach to support their metal-ligand
bifunctional catalysis mechanism for transfer hydrogenation of
ketones.22 Some of us have used the results of DFT calculations
as a support for the mechanism of H2-hydrogenation of ketones
established on the basis of kinetic data.13 Hedberg et al. have
extended these calculations to explain the alcohol effect on
Noyori-type catalysts.50 Other substrates have also been exam-
ined. For example, the hydrogenation of CO2 using cis-Ru(H)2-
(PMe3)4 catalysts has been a subject of a theoretical investiga-
tion.51
Mechanism for Low Concentrations of Alcohol, Path A.
The theoretical reaction coordinate diagram for the H2 activation
and the subsequent acetone hydrogenation in the gas phase
without alcohol assistance is shown in Figure 5. The structures
corresponding to the steps of dihydrogen activation are shown
in Figure 4, while those of acetone hydrogenation are shown in
Figure 6. The profile of Figure 5 supports the experimental
findings corresponding to low conversions (path A of Scheme
7) that the reaction between the amido catalyst and H2 gas to
produce trans,cis-Ru(H)2(PH3)2(ampy) (5) via transition state
(52) Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664-675.
(53) Burke, K.; Perdew, J. P.; Wang, Y. In Electronic Density Functional
Theory: Recent Progress and New Directions; Dobson, J. F., Vignale, G.,
Das, M. P., Eds.; Plenum: New York, 1997; p 81.
(54) Leininger, T.; Nicklass, A.; Stoll, H.; Dolg, M.; Schwerdtfeger, P.
J. Chem. Phys. 1996, 105, 1052-1059.
(55) Gusev, D. G. J. Am. Chem. Soc. 2004, 126, 14249-14257.
(56) Zhang, Y.; Guo, Z.; You, X.-Z. J. Am. Chem. Soc. 2001, 123, 9378-
9387.
We extended our experimental results with a theoretical study
in order to gain more information on the catalytic cycles shown
in Scheme 7. Our choice of the restricted mPW1PW9152,53
(50) Hedberg, C.; Ka¨llstrom, K.; Arvidsson, P. I.; Andersson, P. G.;
Brandt, P. J. Am. Chem. Soc. 2005, 127, 15083-15090.
(51) Ohnishi, Y.; Matsunaga, T.; Nakao, Y.; Sato, H.; Sakaki, S. J. Am.
Chem. Soc. 2005, 127, 4021-4032.
(57) Lynch, B. J.; Truhlar, D. G. J. Phys. Chem. A 2001, 105, 2936-
2941.

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5993
Figure 4. Optimized gas-phase structures of the model complexes and the transition state (TS1) for the dihydrogen splitting without an
alcohol. Selected bond lengths and angles: 4 Ru-N1 2.116 Å, Ru-N2 1.971 Å, Ru-H3 1.573 Å, Ru-N2-H4 126.5°; [4·H2] Ru-N2
2.103 Å, H1-H2 0.812 Å, Ru-H1 1.845 Å, Ru-H2 1.832 Å, Ru-H3 1.601 Å, Ru-N2-H4 107.5°, TS1 Ru-N2 2.153 Å, H1-H2 1.029
Å, Ru-H1 1.809 Å, N2-H2 1.832 Å, Ru-H3 1.634 Å, Ru-N2-H4 108.2°, 5 Ru-N2 2.166 Å, H1-H2 2.6 Å, Ru-H1 1.681 Å, Ru-H3
1.703 Å, Ru-N2-H4 105.6°. Selected APT charges of [4·H2]: H1 +0.095, H2 +0.091, N2 -0.48.
If it is rotated to eclipse the Ru-N(pyridine) bond and
reoptimized, the H2 returns to eclipse the Ru-N(amido) during
the optimization steps. The higher stability of the former
structure stems from the higher electron density around the
formally double Ru-N(amido) bond that could contribute to
back-bonding to σ* (H2). The consequence of the H2 coordina-
tion is the loss of planarity at the N(amido) (sum of angles at
N is 327.4°), its rehybridization from sp2 to sp3, and the resulting
release of the nitrogen lone pair from pπ(N)fdπ(Ru) dative
bonding. Consequently, the Ru-N(amido) bond elongates from
1.972 Å in 4 to 2.103 Å in [4·H2] and the APT charge on
nitrogen atom decreases from -0.454 to -0.479. reflecting its
increased basicity.
The acidic η2-H2 ligand in [4·H2] is deprotonated by the amido
Figure 5. Calculated energies (at the mPW1PW91/SDD level for
nitrogen to produce the trans dihydride trans,cis-Ru(H)2(PH3)2-
Ru and mPW1PW91/6-311++G(d,p) level for the nonmetallic
(ampy) (5) via the transition state TS1 (Figure 4). The start of
elements) for the H2-hydrogenation catalytic cycle using 4 as a
N-H and Ru-H bond formation in TS1 is signaled, among
model catalyst and acetone as a substrate (∆H dashed line, ∆G
other changes, by an elongation of the H-H distance (to 1.029
solid line). The ∆G values (298.15 K, 1 atm) are corrected for the
Å) and a shortening of the N‚‚‚H and Ru‚‚‚H distances to 1.832
zero-point energies but uncorrected for translation and rotation.
and 1.809 Å, respectively. The imaginary frequency for this
mode is 1294i cm-1. The activation parameters for TS1 (∆Gq
TS1 is the turnover-limiting step with theoretical (gas-phase)
) 19.7 kcal mol-1, ∆Hq ) 9.93 kcal mol-1, ∆Sq ) -32.7 cal
mol-1 K-1) compare well with those experimentally determined
for Ru(H)2(R-binap)(tmen) (∆Gq ) 16.6 kcal mol-1, ∆Hq )
8.6 kcal mol-1, ∆Sq ) -27 cal mol-1 K-1) and Ru(H)-
(PPh3)2(HNCMe2CMe2NH2) (∆Gq ) 14.5 kcal mol-1, ∆Hq )
7.6 kcal mol-1, ∆Sq ) -23 cal mol-1 K-1).13
activation energies of 19.7 (∆Gq) and 9.9 kcal/mol (∆Hq),
respectively. The energies of hydrogenation of acetone are
calculated to be -4.31 (∆G) and -13.9 kcal mol-1 (∆H),
respectively.
The dihydrogen activation steps (Figures 4 and 5) start with
the 4/H2 system consisting of free 4 and free H2. When H2 is
placed above the plane defined by Ru and two N atoms (at
approximately 3.5 Å distance) and the system is optimized, the
dihydrogen complex [4·H2] is obtained (Figure 4). This complex
is the dihydrogen-hydride species Ru(η2-H2)(H)(PH3)2-
(HNCH2C5H4N). The H-H distance in [4·H2] (0.812 Å) is
longer than the H-H distance in free H2 (0.744 Å). The APT
charges on the H atoms (+0.095 and +0.091) indicate that the
coordinated H2 molecule becomes acidic.42 The H2 ligand is
oriented in such a way that it eclipses the Ru-N(amido) bond.
The geometry at Ru in the trans dihydride 5 is that of a
distorted octahedron. Overall, it closely resembles the geometry
of a related trans dihydride Ru(H)2(R-binap) (H2NCMe2CMe2-
NH2), whose X-ray crystal structure has been determined.7 The
Ru-H distances in 5 are 1.681 and 1.703 Å, while those in
Ru(H)2(R-binap) (H2NCMe2CMe2NH2) are 1.64(3) and 1.70-
(3) Å. The Ru-NH2 bond lengths are also comparable with
2.166 Å in 5 and 2.202(2)/2.193(2) Å in the full complex.
Interestingly, the P-Ru-P angle in 5 (91.4°) is almost identical
Figure 6. Optimized gas-phase structures of the model complexes and the transition state (TS2) for the H2-hydrogenation of acetone with
5 as catalyst. Selected bond lengths (Å): [5·acetone] H1-C7 2.926, O1-C7 1.216, O1-H2 1.980, H2-N2 1.019, N2-Ru 2.165, Ru-H1
1.688, TS2 H1-C7 1.817, O1-C7 1.241, O1-H2 1.725, H2-N2 1.035, N2-Ru 2.139, Ru-H1 1.724, [4·2-PrOH]-A H1-C7 1.127,
O1-C7 1.419, O1-H2 1.004, N2-H2 1.698, Ru-N2 2.052, Ru-H1 2.181, [4·2-PrOH]-B H1-C7 1.100, O1-C7 1.419, O1-H2 1.006,
N2-H2 1.715, Ru-N2 2.094, Ru-O1 2.380.

5994 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
Figure 7. Optimized gas-phase structures of the model complexes and the transition states for the 2-propanol- and 1-phenylethanol-
assisted splitting of H2. Selected bond lengths (Å): [4·H2·2-PrOH] H5-H6 0.813, Ru-H5 1.830, Ru-H6 1.849, O1-H6 2.434, H2-N2
1.708, TS3 H5-H6 0.857, Ru-H5 1.817, O1-H6 1.648, H2-N2 1.154, [5·2-PrOH] H6-H5 1.648, Ru-H5 1.706, O1-H6 0.990, H2-
N2 1.024, O1-H2 1.933, [4·H2·PhEtOH] H5-H6 0.814, Ru-H5 1.831, Ru-H6 1.842, O1-H6 2.543, H2-N2 1.660, TS4 H5-H6 0.851,
Ru-H5 1.818, O1-H6 1.684 Å, H2-N2 1.120, [5·PhEtOH] H6-H5 1.487, Ru-H5 1.708, O1-H6 0.995, H2-N2 1.022, O1-H2 2.027.
Selected APT charges of [4·H2·2-PrOH]: H5 +0.10, H6 +0.14, N2 -0.58; [4·H2·PhEtOH]: H5 +0.11, H6 +0.12, N2 -0.58.
as a model of a ketone hydrogenation reaction. There are small
barriers to this exothermic reaction. First the trans dihydride 5
interacts with acetone to produce the molecular complex [5·
acetone] (Figure 6), where the carbonyl oxygen forms a
hydrogen bond with the axial proton of the NH2 group and the
carbonyl carbon interacts weakly with the hydride on Ru. This
interaction causes a slight elongation in the CdO bond from
1.207 Å in free acetone to 1.216 Å in [5·acetone].
The hydrogenation of acetone proceeds via a transition state
TS2 that has an imaginary frequency at 230i cm-1. The transfer
of H(-) and H(+) in an outer-sphere mechanism is facile with
the activation parameters ∆Ghyd ) 8.7 kcal mol-1, ∆Hhyd
)
q
q
q
-4.7 kcal mol-1, and ∆Shyd ) -76.6 cal mol-1 K-1. In TS2
there is a significant shortening of the C‚‚‚H and O‚‚‚H distances
in the OdC‚‚‚H-Ru and CdO‚‚‚H-N interactions, respec-
tively, and an elongation of the C-O bond. Other changes are
also apparent during the vibration at the imaginary frequency.
These include a significant displacement of the carbonyl carbon
from the plane defined by the two methyl carbons and the
oxygen atom and also a movement of the hydride ligand toward
the electrophilic carbon in the CdO group.
Figure 8. Comparison of the calculated energies for the alcohol-
free H2 splitting (from the middle to the right) and the alcohol-
assisted H2 splitting (from the middle to the left). The zero value
at the middle is the starting relative energy of free 4 and H2 for the
alcohol-free path and free 4, H2, and alcohol for the alcohol-assisted
path. The ∆G values (298.15 K, 1 atm) are corrected for the zero-
point energies but uncorrected for translations and rotations. The
solid line connects ∆H, while the dashed line connects ∆G values.
The values for R ) C6H5(CH3)CH- are underlined, while those
for R ) (CH3)2CH- are in italics.
The primary product of acetone hydrogenation is the adduct
[4·2-PrOH]-A, in which 2-propanol interacts with the amido
complex through N2‚‚‚H2-O1 (1.70 Å) and C7-H1‚‚‚Ru (2.18
Å) interactions (Figure 6). Complex [4·2-PrOH]-A can rearrange
to produce [4·2-PrOH]-B, which is, according to calculations,
the most stable structure of the cycle (Figure 5). 2-Propanol is
still in close contact with the amido nitrogen (N2‚‚‚H2-O1
distance 1.71 Å); however the Ru‚‚‚H1 interaction is lost and
rather the O1 atom from 2-propanol moves above the ruthenium
atom (O1‚‚‚Ru 2.38 Å). As a result, the Ru-N(amido) bond in
[4·2-PrOH]-B is significantly longer than the same bond in free
4 (2.094 vs 1.972 Å). This indicates a decrease in the
pπ(N)fdπ(Ru) dative bonding due to the fact that the nitrogen
lone pair is involved in the interaction with the alcohol. Also,
to the P-Ru-P angle in Ru(H)2(R-binap)(H2NCMe2CMe2NH2)
(91.5°). The formation of 5 is entropically unfavorable, and 5
lies about 2 kcal/mol in free energy above free H2 and 4. This
may be why the formation of the corresponding dihydride trans-
Ru(H)2(S-binap)(appH) from 3 and dihydrogen gas (1 atm) has
not yet been observed.
Figure 5 also shows the energy profiles for the reaction of
the trans dihydride 5 with acetone via transition state TS2 to
give back the hydridoamido complex 4 and release 2-propanol

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5995
Figure 9. Optimized starting structures (gas phase) for the alcohol-assisted hydrogenation of acetone. Selected bond lengths (Å): 6 H1-
C10 3.114, C10-O2 1.218, O2-H5 1.819, O1-H2 2.009, 7 H1-C10 2.973, C10-O2 1.221, O2-H2 2.145, O2-H5 1.902.
each with one imaginary frequency at 446i and 227i cm-1. The
Scheme 8
splitting of the H2 and the simultaneous protonation of the amido
nitrogen are indicated by a decrease in HH‚‚‚O and OH‚‚‚N
distances. After the heterolytic H2 cleavage, the complexes [5·
ROH] are produced in which the alcohol is hydrogen bonded
to the trans dihydride complex 5. Hedberg et al. calculated a
very similar enthalpy profile for the RuH(NHCH2CH2NH2)-
(PH3)2/H2/MeOH system.50
The activation parameters for the alcohol-assisted transition
states TS3 for 2-PrOH (∆Gaq ) 14.7 kcal mol-1, ∆Haq ) -5.81
q
kcal mol-1, ∆Sa ) -68.7 cal mol-1 K-1) and TS4 for
1-phenylethanol (∆Gaq ) 12.2 kcal mol-1, ∆Haq ) -8.72 kcal
the geometry around the N(amido) is not trigonal planar (sum
mol-1, ∆Sa ) -70.3 cal mol-1 K-1) lie significantly below
q
of angles at N is 334.6°). Its APT charge decreases from -0.474
the ones calculated for the alcohol-free TS1 (Figure 8). The
in 4 to -0.540 in [4·2-PrOH]-B, indicative of higher localization
q
change in ∆Ga due to 2-propanol assistance is approximately
of the negative charge. In this environment, the Ru center is
-5 kcal/mol, and this should result, on the basis of gas-phase
transition state theory, in an increase in the rate constant by a
factor of 4000. In the actual experiment, a factor of about 10 is
observed on going from benzene to 2-propanol. There must be
more electron-deficient. After 2-propanol is released from [4·
2-PrOH]-B, the active catalyst 4 is regenerated.
It is interesting to note that every attempt to obtain a stable
structure [5·acetophenone], an analogue of [5·acetone], failed.
other factors in solution compared to the gas phase that have
The result of the optimization steps was the hydrogenation of
not been considered, such as those arising from differences in
the substrate and the formation of 4 and 1-phenylethanol, a
the pre-exponential factor of the rate constant equation for the
structural analogue to [4·2-PrOH]-A. The same outcome was
ternary transition state complex TS3 versus the secondary
obtained regardless of the size of basis sets used for nonmetallic
transition state complex TS1 or other interactions of the
elements (3-21G(d,p) and 6-311++G(d,p)) and the distance
catalyst or transition states with further alcohol molecules.
between 5 and acetophenone molecules in the Gaussian input
The factor of 10 increase in rate constant cannot be simply
file. This result suggests that the energy barrier for the
explained by a nonspecific polarity effect since the change in
acetophenone hydrogenation is significantly lower than that for
polarity during the hydrogenation of a 0.2 M ketone solution
acetone and in turn lies even lower compared to the H2 splitting
in benzene is negligible, yet a sizable increase in rate is
step.
observed.
Alcohol-Assisted H2 Splitting Mechanism, Path B. This
Alcohol-Assisted Ketone Hydrogenation. Since Figure 8
mechanism (path B of Scheme 7), involving either 2-propanol
indicates that the solvated dihydride complex [5·2-PrOH] would
be favored over “free” 5 in 2-propanol solutions, we investigated
how the ketone hydrogenation by this dihydride is affected by
the presence of a 2-propanol molecule. Very recently Handgraaf
and Meijer have utilized ab initio molecular dynamics methods
to evaluate the role of the alcohol solvent in the transfer
hydrogenation of formaldehyde to methanol using (η6-
C6H6)RuH(NH2CH2CH2O) as a catalyst.58 They showed that
the solvent molecules can play a specific role in the substrate
reduction process and that simple solvation models cannot
account for this.
or 1-phenylethanol, has also been investigated theoretically
(Figures 7-9). The addition of the η2-H2 ligand to the electron-
deficient and coordinatively unsaturated Ru center in [4·ROH]-
B (R ) (CH3)2CH- or C6H5(CH3)CH-) produces stable η2-
dihydrogen complexes [4·H2·ROH] (Figure 7). These consist of
[4·H2] (see above) in a hydrogen-bonded network with the
respective alcohol. One H atom of the dihydrogen ligand is
oriented toward the oxygen atom in the alcohol (as opposed to
the amido nitrogen in [4·H2]), while the N(amido)‚‚‚HO
hydrogen bond remains intact. The H-H distances in [4·H2·
ROH] are similar to that of [4·H2], but the dihydrogen atoms in
[4·H2·ROH] have a positive charge (APT) (Figure 7) that is
approximately 0.05 units greater than those of [4·H2] (Figure
4). Furthermore, the amido nitrogen is more basic, with a
negative charge of -0.58 for [4·H2·ROH] compared to -0.48
for [4·H2]. The dihydrogen is cleaved between the Ru and the
oxygen atom via the transition states TS3 and TS4 (Figure 7),
We examined two possible starting points for the alcohol-
assisted hydrogenation of acetone (H2 transfer), systems 6 and
7 (Figure 9).
(58) Handgraaf, J.-W.; Meijer, E. J. J. Am. Chem. Soc. 2007, 129, 3099-
3103.

5996 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
Figure 10. Optimized gas-phase structures of the model complexes and the transition state (TS5) for the alcohol-assisted ketone hydrogenation
starting from 6. Selected bond lengths (Å): TS5 H1-C10 1.786, C10-O1 1.241, O1-H5 1.658, O2-H2 1.869, [8·2-PrOH] Ru-O1
2.206, O1-C10 1.388, O1-H5 1.513, O2-H2 1.779, 8 Ru-O1 2.182, O1-H2 1.825, O1-C10 1.379.
Figure 11. Optimized gas-phase structures of the model complexes and the transition state (TS6) for the alcohol-assisted ketone hydrogenation
starting from 7. Selected bond lengths (Å): TS6 H1-C10 2.087, H2-O1 1.882, H5-O1 1.859, O1-C10 1.235, 9 H1-C10 1.125, O1-
C10 1.400, O1-H2 1.033, H2-N2 1.581, H5-O1 1.811.
Structure 6 (Figure 9) results from insertion of the acetone
carbonyl group into the OH‚‚‚HRu hydridic-protonic bond of
the complex [5·2-PrOH]. Thus an eight-membered ring is formed
through a hydrogen bond network. Structure 7 is obtained by
adding a 2-propanol molecule to [5·acetone] so that it forms a
hydrogen bond to the carbonyl group of the acetone. This retains
the original six-membered ring of the complex [5·acetone].
System 7 is slightly higher in energy than 6 (∆H ) 0.70 kcal/
mol, ∆G ) 1.03 kcal/mol), and thus the two could be in
equilibrium.
Starting from complex 6, we expected to complete the
hydrogenation of acetone by “shuttling” the protons through
the hydrogen bond network according to Scheme 8 to give
complex 10. However, our attempts to optimize the structure
of complex 10 failed. Instead the calculations predict the
formation of the alkoxohydridoruthenium(II) complex [8·2-
PrOH] (Figure 10). This species is the product of only a hydride
transfer to the carbonyl carbon.
These results suggest that the alcohol -OH group is not acidic
enough to transfer its proton to the substrate. The proton on
the neighboring -NH2 group is tied up in a hydrogen bond
with the alcohol molecule. The optimized structure of [8·2-
PrOH] is very similar to the crystal structure of the related Ru-
(II) phenoxide complex RuH(OPh)(tmen)(PPh3)·HOPh (tmen )
2,3-diamino-2,3-dimethylbutane),26 in which a phenol molecule
forms a six-membered ring Ru-O‚‚‚HO(Ph)‚‚‚NH with the
pseudo-octahedral complex through a hydrogen bond network.
However, the bonds in [8·2-PrOH] are much shorter than the
corresponding ones in the phenoxide complex. Thus the Ru-O
distance in the phenoxide complex is 2.364(2) Å, while the
calculated one in [8·2-PrOH] is 2.206 Å and compares well with
the Ru-O distance of 2.239(2) Å in a similar complex, trans-
RuH(OC6H4-p-CH3)(dmpe)2 (dmpe ) Me2PCH2CH2PMe2).59
The calculated hydrogen bonds in [8·2-PrOH] are also shorter
(OH‚‚‚HRu 1.51 Å and NH‚‚‚OH 1.78 Å) compared to the
phenoxide/phenol complex (OH‚‚‚HRu 1.72 Å and NH‚‚‚OH
2.04 Å). This difference in hydrogen bonds could be the
consequence of inaccuracy in hydrogen atom positions and bond
lengths as determined by X-ray crystallography. The O(alco-
hol)-O(Ru alkoxide) (2.51 vs 2.52 Å in [8·2-PrOH] and
phenoxide, respectively) and O(alcohol)-N(Ru) distances (2.80
vs 2.93 Å) are comparable. The transition state for the formation
of [8·2-PrOH], TS5, has the activation parameters ∆Gaq ) 14.4
q
q
kcal mol-1, ∆Ha ) -17.7 kcal mol-1, and ∆Sa ) -107 cal
mol-1 K-1 (relative to the free reactants: amido 4, acetone,
H2, and 2-propanol) with an imaginary frequency at 170i cm-1
.
q
It is interesting to note that the calculated values for ∆Ga for
TS3 (2-propanol-assisted H2 splitting) and TS5 are very close:
14.7 and 14.4 kcal/mol. The loss of the 2-propanol molecule
produces alkoxide complex 8. Complex 8 is characterized by a
short H2‚‚‚O1 distance of 1.82 Å accompanied by a small N2-
Ru-O1 angle of 71.6° (compared to 84.9° in [8·2-PrOH]) while
maintaining the Ru-O1 bond almost unchanged at 2.182 Å.
Bergens and co-workers19 have recently reported the observation
of RuH(2-PrO)(R-binap)(R,R-dpen) formed from [RuH(H2)(R-
binap)(R,R-dpen)]+ in 2-propanol in the presence of a strong
base (KOtBu). They have provided evidence that this alkoxide
complex does not react with H2 (2 atm) without a base present
at ambient temperatures over prolonged times (10 h). The
remarkable stability of the observed alkoxide was attributed to
(59) Burn, M. J.; Fickes, M. G.; Hollander, F. J.; Bergman, R. G.
Organometallics 1995, 14, 137-150.

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5997
Scheme 9. Expanded Mechanism to Include Catalysis in
Benzene and 2-PrOH
Figure 12. Part of the energy profile for the alcohol-assisted
hydrogenation of acetone, starting from 6 or 7 (Figure 9). The solid
lines connect the ∆H, while the dashed lines connect the ∆G values.
The energy values for the cycle starting with 6 are in normal font,
while those for the cycle starting from 7 are in italics. The energies
are relative to free 4, acetone, 2-propanol, and H2 for easier
comparison with the energy profiles of Figure 8. The reaction
sequence ends with two free 2-propanol and the regenerated free
amido 4.
the intra- and/or intermolecular hydrogen bonding between an
NH group and the 2-PrO- ligand and/or solvent molecules.19
Our calculations support this observed stability. Given that the
alkoxohydridoruthenium complexes are not active in the H2-
hydrogenation of ketones26 and have high stability,19 they
possibly represent the catalyst resting state in 2-propanol
solutions. The addition of a strong base presumably shifts the
equilibrium toward the formation of free 4 and thus decreasing
the concentration of the resting state species.19 Experimental
observations have shown that the addition of base has a
beneficial effect on the rate of hydrogenation.14,33 However,
there is a small difference in free energy between 8 and complex
[4·2-PrOH]-B (0.6 kcal/mol), suggesting that the two might be
in a rapid equilibrium.
catalysts for H2 transfer to ketones, but the alcohol assistance
of H2 transfer could also account for these variations.
An alternative pathway starts from the hydrogen-bonded
system 7 (Figure 9). In this case a structure 9 is found (Figure
11), in which the proton from the amino group and the hydride
from the RuH moiety have been fully transferred to the ketone
substrate. In 9 the hydrogen bond between 2-propanol and
hydrogenated substrate (O(2)H(5)‚‚‚O(1)) is retained and de-
scribed as [4·2-PrOH]-A with one 2-propanol molecule hydrogen
On the basis of the experimental and theoretical results, we
propose an expanded mechanism for catalysis shown in Scheme
9. It includes both an alcohol-free pathway (path A, the domain
within the dotted lines) that operates in the benzene solutions
at low alcohol concentrations and an alcohol-assisted pathway
(path B) for high alcohol concentrations and/or solutions in
alcohol. Path A has been described previously apart from the
intermediate [4·2-PrOH]-B, an alcohol adduct of the amidohy-
dridoruthenium complex 4. Path B circles around outside of
the dotted line domain, where additional 2-PrOH is present. It
includes the alcohol-assisted H2 splitting step TS3, two possible
alcohol-assisted hydrogenations of the ketone via TS5 or TS6,
and the formation of alkoxide intermediates 8 and 9.
bonded to the newly produced alcohol. The transition state TS6
has the activation parameters ∆Gaq ) 13.0 kcal mol-1, ∆Ha
)
q
-19.0 kcal mol-1, and ∆Sa ) -107 cal mol-1 K-1 (relative
to the free reactants: amido 4, acetone, H2, and 2-propanol)
with an imaginary frequency at 75i cm-1. The energy profile
of the two pathways is shown in Figure 12.
q
The transition states for alcohol-assisted hydrogen splitting
(TS3) and alcohol-assisted H2 transfer (TS5 and TS6) are very
close in free energy (14.7, 14.4, and 13.0 kcal/mol for TS3,
TS5, and TS6, respectively), but in any case the H2 splitting
has a higher energy barrier. Furthermore, the orientation of the
substrate molecule in TS5 and TS6 is different from that in
TS2 (alcohol-free hydrogenation). Since this is the step that
determines enantioselectivity, this difference in orientation can
explain the difference in obtained ee values between the
reactions in benzene and 2-propanol (27 vs 32%; see Table 2).
It might also explain the dramatic changes in ee with solvent
reported by Ohkuma et al. with the very similar complex RuCl2-
((S)-tolbinap)(pica), where pica and ampy refer to the same
ligand, 2-aminomethylpyridine.34 Ohkuma et al. attributed the
changes to variations in the structures of the alkoxides cis-RuH-
(OR)((S)-tolbinap)(pica), which they propose to be the active
Conclusions
The design and synthesis of a novel appH ligand, lacking
R-hydrogens, which is an analogue of 2-(aminomethyl)pyridine,
permits the isolation of a stable amidohydrido complex, RuH-
(S-binap)(app) (3). It is prepared from a parent chloro-hydrido
complex, RuHCl(S-binap)(appH) (2), by dehydrochlorination
by use of the strong base KOtBu. The crystal structure of 3
reveals the presence of a short Ru-N(amido) bond and a
trigonal-planar geometry around the amido nitrogen. Like similar
amido complexes previously reported,13,47 amido 3 is an active
catalyst for hydrogenation of ketones without a base added. The
substrate-reducing complex is likely to be trans-RuH2(S-binap)-
(appH) according to our calculations and by analogy with other

5998 Organometallics, Vol. 26, No. 24, 2007
HadzoVic et al.
similar systems,13,33 but the formation of this dihydride is
unfavorable entropically and has not yet been observed. A
kinetic study in benzene shows that systems using 3 as a catalyst
behave differently from other diamine systems such as trans-
Ru(H)2(R-binap)(tmen), Ru(H)2(PPh3)2(R,R-dach), RuH(PPh3)2-
(H2NCMe2CMe2NH), and RuHCl(PPh2C6H4CH2NHC6H10-
NHCH2C6H4PPh2).13,33,47 While the measured rates in all cases
show a dependence on both Ru and H2 concentration, the rate
of catalysis with 3 is also a function of the product concentration
and therefore is autocatalytic.
purchased from Sigma-Aldrich and used without further purification.
1
The H and 13C NMR spectra were recorded on a Mercury 400
1
spectrometer operating at 400 MHz for H and 100 MHz for 13C.
The 31P NMR spectra were obtained on a Gemini 300 spectrometer
1
(121.5 MHz). The H and 13C NMR were measured relative to
partially deuterated solvent peaks but are reported relative to
tetramethylsilane. The 31P chemical shifts are reported relative to
85% H3PO4 as an external reference. The elemental analysis was
performed at the Department of Chemistry, University of Toronto,
on a Perkin-Elmer 2400 CHN elemental analyzer.
The catalytic hydrogenations and the kinetic measurements were
performed under constant pressures of H2 gas in a 50 mL Parr high-
pressure reactor. A constant temperature for these experiments was
maintained using a Fisher Scientific IsoTemp 1016D water bath.
The samples were analyzed by chiral GC on a Perkin-Elmer
Autosystem XL with a Chrompack capillary column (ChirasilDEX
CB 25 m × 0.25 mm). Hydrogen was used as a carrier gas at a
column pressure and temperature of 5 psi and 130 °C, an injector
temperature of 250 °C, and an FID temperature of 275 °C. The
retention times were as follows: acetophenone, 5.20 min; (R)-1-
phenylethanol, 8.95 min; and (S)-1-phenylethanol, 9.50 min.
That indeed the product affects the hydrogenation rate was
confirmed by adding rac-1-phenylethanol to the reaction mixture
prior to the hydrogenation. The addition of 2-propanol also
increases the rate, but not as much as PhCH(OH)Me. Moreover,
the highest rate and TOF were achieved when 2-propanol was
used as solvent. These findings have been explained by a second
pathway, the alcohol-assisted H2 splitting mechanism in which
added alcohol lowers the energy of this rate-limiting step. This
alcohol effect has been observed previously in other metal-
ligand bifunctional catalysis. It seems to be more pronounced
for the aminomethylpyridine type of ligands than simple
diamines. The reason for this different behavior is not clear.
One possible explanation is the decrease in steric bulk on one
side of amine ligand due to the planarity of the pyridine ring.
This might also explain why the use of this type of pyridine
ligand is so effective in the hydrogenation of bulky ketones.34
The gas-phase DFT calculations support the experimental
results. We find that the rate-limiting step is the heterolytic
cleavage of H2 by the Ru amido complex with the following
Synthesis of N-(2-(Pyridin-2-yl)propan-2-yl)acetamide, 1c.
Potassium hydroxide (4 g, 0.07 mol) and 1b (0.7 g, 2.7 mmol)
were dissolved in 40 mL of dry methanol under an argon
atmosphere. This solution was then added to 1.4 g of Raney nickel
in 10 mL of dry methanol. This was placed under 1 atm of H2 and
stirred for 4 h. The mixture was then filtered through a pad of Celite
and the solvent from the filtrate evaporated in Vacuo. Distilled water
(10 mL) was added to the yellowish residue and the mixture
extracted with CHCl3 (3 × 10 mL). The combined organic layers
were dried over MgSO4. Removal of the solvent gave a pale yellow
solid. Yield: 0.32 g (65%). 1H NMR (400 MHz, CDCl3, δ): 1.72
(s, 6H, -CH3), 2.01 (s, 3H, -CH3), 7.15 (ddd, JHH ) 7.6, 4.8, 1.2
Hz, 1H, py), 7.36 (m, 1H, py), ca. 7.66 (s, br, 1H, NH), 7.67 (dt,
JHH ) 7.2, 1.6 Hz, 1H, py), 8.47 (m, 1H, py). 13C NMR (100 MHz,
CDCl3, δ): 24.85, 27.66, 56.66, 119.67, 122.02, 137.25, 147.77,
164.76, 169.56. EI-MS (m/z): 178 (M+, 25%), 163 (35%), 135
(65%), 121 (100%).
parameters: ∆Gq ) 19.7 kcal mol-1, ∆Hq ) 9.93 kcal mol-1
,
∆Sq ) -32.7 cal mol-1 K-1. The presence of alcohol
significantly lowers the activation barrier for this process by
about 5 kcal/mol in free energy (∆Gaq ) 14.7 kcal mol-1, ∆Haq
) -5.81 kcal mol-1, ∆Saq ) -68.7 cal mol-1 K-1 in the case
of 2-PrOH) but not enough to bring it below the hydrogenation
step (∆Ghyd ) 8.7 kcal mol-1, ∆Hhyd ) -4.7 kcal mol-1
,
q
q
∆Shydq ) -76.6 cal mol-1 K-1). We have also investigated two
alcohol-assisted mechanisms of ketone H2-hydrogenation in
2-propanol solvent. These can explain the differences in
enantioselectivity observed between the hydrogenations in
benzene and 2-propanol because of the different orientations
of the substrate molecule in the transition states TS2 versus
TS5 and TS6. They also indicate that alkoxide complexes are
the catalyst resting states. These pathways may be broadly
applicable to other Noyori-type catalyst systems.
Synthesis of 2-Amino-2-(2-pyridyl)propane (appH). A yellow
solution of 1c (0.3 g, 1.7 mmol) in 6 M HCl (40 mL) was refluxed
in air overnight. The cooled reaction mixture was made strongly
basic (pH 10 against universal pH paper) with 30% NaOH. This
was extracted with CHCl3 (3 × 10 mL). The organic extracts were
dried over MgSO4, and the solvent was removed in Vacuo, leaving
the product as a pale brown oil. Yield: 0.21 g (91%). The ligand
was used without further purification. 1H NMR (400 MHz, CDCl3,
δ): 1.45 (s, 6H, -CH3), 1.88 (s, br, 2H, -NH2), 7.06 (m, 1H, py),
7.4 (d, br, J ) 7.6 Hz, 1H, py), 7.58 (dt, J ) 7.6, 1.4 Hz, 1H, py),
8.50 (s, br, 1H, py). 13C NMR (100 MHz, CDCl3, δ): 31.37, 54.14,
118.53, 121.40, 136.53, 148.76, 168.32. EI-MS (m/z): 137 (M+,
25%), 121 (100%), 104, (35%), 80 (20%), 58 (95%).
This report has also introduced the direct synthesis of RuHCl-
(binap)(appH) from RuCl2(binap)(PPh3) according to Scheme
5. This is proving to be a valuable route for the synthesis of a
variety of other ruthenium precatalysts.
Synthesis of Ru(H)Cl(S-binap)(appH), 2. Triethylamine (0.11
g, 1 mmol, 0.15 mL) was added to a suspension of RuCl2(S-binap)-
(PPh3) (0.58 g, 0.55 mmol) in 8 mL of dry THF in a 50 mL Schlenk
flask under an H2 atmosphere. The reaction was stirred for 4 h at
room temperature, resulting in a deep red solution. After purging
the flask with argon, neat appH (0.09 g, 0.6 mmol) was added and
stirring continued for 1 h. The solvent from the resulting intense
orange-yellow solution was removed in Vacuo. The solids were
extracted with a minimum amount of dry THF and filtered to
remove the insoluble [Et3NH]Cl salt. The addition of the 1:1
hexanes/ether mixture precipitated the orange-yellow product. The
solids were filtered off, washed with a hexanes/ether mixture, and
vacuum-dried. Yield: 0.35 g (70%). Anal. Calcd: C, 69.30; H,
Experimental Section
General Considerations. Unless otherwise stated, all the opera-
tions were carried out under an inert atmosphere using standard
Schlenk and glovebox techniques. 1a, 1b,37 RuCl2(PPh3)2,60 and
RuCl2(S-binap)(PPh3)41 were prepared according to the literature
procedures. THF, diethyl ether, hexanes, and benzene were dried
and distilled from sodium benzophenone ketyl under an argon
atmosphere. 2-Propanol and methanol were dried and distilled from
Mg/I2 also under Ar. Acetophenone was dried over molecular sieves
and distilled under an argon atmosphere. Deuterated solvents were
purchased from Cambridge Isotope Laboratories and degassed and
dried over molecular sieves prior to use. All other reagents were
1
5.71; N, 2.89. Found: C, 69.02; H, 5.75; N, 2.97. H NMR (thf-
2
d8, δ), major isomer: -15.79 (dd, JHP ) 28 and 24.5 Hz, 1H,
(60) Hollan, P. A.; Stephenson, T. A.; Wilkinson, G. Inorg. Synth. 1970,
12, 237-240.
RuH), 1.46 (s, 3H, -CH3), 1.84 (s, 3H, -CH3), 2.7 (bd, 1H, NH,

A Mechanism Displaying Autocatalysis
Organometallics, Vol. 26, No. 24, 2007 5999
2
2JHH ) 9.6 Hz), 3.4 (bd, 1H, NH, JHH ) 9.6 Hz), 6.2-8.4 (m,
Table 3. Summary of X-ray Parameters for Complex 3
36H); minor isomer: -16.28 (dd, 2JHP ) 29.8 and 21.7 Hz, RuH),
1.22 (s, -CH3), 1.52 (s, -CH3). 31P{1H} NMR (thf-d8, δ), major
isomer: 66.2 (d, 2JPP ) 39.4 Hz), 72.1 (d); minor isomer: 62.3 (d,
2JPP ) 41.4 Hz), 68.7 (d). 13C NMR (thf-d8, δ): 28.66, 31.8, 62.22,
119.36, 122.02, 124.91, 125.6, 125.76, 126.13, 126.57 (m), 127.67
(m), 127.92, 128.25, 128.4, 128.5, 128.9, 129.2, 129.4, 129.5,
133.87, 134.6, 134.73, 135.45, 136.69, 136.79, 138.76, 138.82. IR
(Nujol, cm-1): 1995 (νRuH), 3329, 3402 (νNH).
empirical formula
fw
temperature, K
wavelength, Å
cryst syst
space group
a, Å
C52H44N2P2Ru‚0.5C6H14
902.99
150(2)
0.71073
monoclinic
P21
11.139(2)
12.076(2)
16.788(3)
90
95.22(3)
90
2248.7(8)
2
1.334
0.459
938
b, Å
c, Å
R, deg
Synthesis of RuH(S-binap)(app), 3. Ru(H)Cl(S-binap)(appH)
(0.260 g, 0.28 mmol) was placed in a 25 mL Schlenk flask and
dissolved in 4 mL of dry THF under an argon atmosphere. The
solid KOtBu (0.08 g, 0.7 mmol) was added to this orange-yellow
solution, immediately turning its color deep red. The reaction was
stirred for 40 min, and THF was removed in Vacuo. Solids were
extracted with approximately 1 mL of dry THF and filtered through
a pad of Celite. Hexanes (7 mL) were added to the filtrate,
precipitating a red product, which was filtered, washed with
hexanes, and vacuum-dried. Yield: 0.20 g (78%) 1H NMR (C6D6,
â, deg
γ, deg
volume, Å3
Z
density (calcd), g cm-3
absorp coeff, mm-1
F(000)
cryst size, mm
range, θ ,deg
no. of reflns collected
no. of indep reflns
completeness to θ ) 27.49°
refinement method
goodness-of-fit on F2
final R indices [I > 2σ(I)]
R indices (all data)
largest diff peak and hole, e Å-3
0.44 × 0.40 × 0.38
2.91-27.49
23 147
9593 [R(int) ) 0.0453]
99.1%
2
δ): -15.46 (dd, JHP ) 26.1 and 40 Hz, 1H, RuH), 1.34 (s, 3H,
full-matrix least-squares on F2
1.070
3
-CH3), 1.57 (s, -CH3), 4.3 (d, JHP ) 4.3 Hz, 1H, NH), 5.7 (m,
2
1H), 6.6 - 8.8 (m, 35 H). 31P{1H} NMR (C6D6, δ): 71.0 (d, JPP
R1 ) 0.0431, wR2 ) 0.1114
R1 ) 0.0482, wR2 ) 0.1156
1.212 and -0.943
) 31 Hz), 82.7 (d). 13C NMR (C6D6, δ): 33.87, 35.42, 72.67,
120.03, 120.35, 125.30, 125.60 (m), 125.7-129.3 (peaks covered
due to the NMR solvent), 133.3, 133.54, 134.5 (m), 137.8, 139.3,
139.6, 140.07, 140.42, 142.62, 143.04, 144.2, 144.68, 155.74,
174.24. IR (Nujol, cm-1): 1903 (νRuH), 3234 (νNH). Several attempts
to obtain a satisfactory elemental analysis on the crystalline,
spectroscopically pure product failed due to the high sensitivity of
this complex to air and nitrogen and difficulties with combustion.
The typical result: Anal. Calcd: C, 73.07; H, 5.69; N, 3.10.
Found: C, 69.35; H 5.42; N, 2.82.
and refine the structure by direct methods. The hydride ligand was
located and refined with isotropic thermal parameters. The crystal-
lographic data are summarized in Table 3. Refinement of data
revealed the presence of solvent molecules in the unit cell that
refined the best as 2-methylpentane with an occupancy of 1/2 for
all the solvent atoms.
Catalytic Tests with 2 and 3. A solution containing acetophe-
none (0.250 g, 2.1 mmol) and precatalyst 2 (7 mg, 7.5 × 10-6
mol) was dissolved in 5 mL of benzene. This solution was injected
in a high-pressure reactor under the H2 gas previously thermostated
at 20 °C. This was followed by the addition of a KOtBu (0.021 g,
0.19 mmol) suspension in 0.2 mL of benzene. After the addition
of the base, the H2 pressure was set to 10 atm. The same procedure
was followed for the catalytic reaction with 3 except the base was
not added.
Computational Details. All DFT calculations were performed
using Gaussian 03 (Revision D.01)61 using a restricted hybrid
mPW1PW91 functional52,53 with SDD54 ECP (GenECP) for ruthe-
nium and 6-311++G(d,p) level for nonmetallic elements using a
default grid. The synchronous transit-guided quasi-Newton (STQN)
method (QST2) was used for the optimization of transition states.62
All transition states had one imaginary frequency, consistent with
the reaction course. The reported energies are for gas phase at
298.15 K and 1 atm and are corrected for zero-point energies but
uncorrected for the translation and rotation.
Kinetic Measurements. The standard solutions of 3 (1.5 mM)
and acetophenone (0.833 M) were prepared by dissolving a required
amount of a compound in 5 mL of benzene or 2-propanol. The
fresh solution of 3 was prepared before each run. The reaction
mixtures were prepared by pipetting the required volumes of the
standard solutions and benzene or 2-propanol. For the runs in which
they were used as additives, the alcohols (racemic 1-phenylethanol
and 2-propanol) were added in required amounts to the acetophe-
none standard solution so that, after the dilution, the concentration
of rac-1-phenylethanol was 0.04 M and for 2-propanol 0.06 M.
For all the runs the solution of acetophenone, followed by the
solution of 3, was injected into the already thermostated reactor
under the desired pressure of H2 to give the final reaction volume
of 5 mL. The reaction time was measured from the addition of
complex 3 solution. The samples were withdrawn from the reactor
at regular time intervals of 5 min. The sampling took about 10 s,
during which time the pressure in the reactor fell to almost
atmospheric. The samples taken were analyzed by chiral GC. The
transfer hydrogenation experiment was carried in an argon glovebox
in well-sealed and stirred 10 mL round-bottom flask.
Acknowledgment. This work was supported by grants to
R.H.M. from NSERC Canada and PRF as administered by
American Chemical Society. Digital Specialty is thanked for a
gift of S-binap. We thank Prof. Dmitri Goussev for information
on DFT calculations.
Supporting Information Available: The rate law for the
hydrogenation of acetophenone in benzene with 3 as a catalyst
obtained by numerical integration of the data from the Figures 2
and 3. X-ray structural data and crystallographic file (CIF) for
complex 3. XYZ coordinates for all the optimized structures and
the calculated energies. This material is available free of charge at
http://pubs.acs.org.
OM700849W
X-ray Structure Determination of 3. Single crystals of Ru-
(H)(S-binap)(app)‚0.5C6H14 were grown by a slow diffusion of
hexanes into a concentrated solution of the complex in benzene.
The data were collected using a Nonius Kappa-CCD diffractometer
at 150 K with Mo KR (λ ) 0.71073 Å) and integrated and scaled
using the Denzo-SMN package. SHELEX V6.10 was used to solve
(61) Frisch, M. J.; et al. Gaussian 03, Revision C.02 ed.; Gaussian Inc.:
Wallingford, CT, 2004. See the Supporting Information for the complete
author list.
(62) Peng, C.; Ayala, P. Y.; Schlegel, B. H.; Frisch, M. J. J. Comput.
Chem. 1996, 17, 49-56.