Mechanistic investigation of the hydrogenation of ketones catalyzed by a ruthenium(II) complex featuring an N-heterocyclic carbene with a tethered primary amine donor: Evidence for an inner sphere mechanism

Article abstract of DOI:10.1021/om101152m

The complex [Ru(p-cymene)(m-CH₂NH₂)Cl]PF₆ (1) catalyzes the H₂-hydrogenation of ketones in basic THF under 25 bar of H₂ at 50 °C with a turnover frequency (TOF) of up to 461 h^-1 and a maximum conversion of 99%. When the substrate is acetophenone, the TOF decreases significantly as the catalyst to substrate ratio is increased. The rate law was then determined to be rate = k _H[Ru]_tot[H₂]/(1 + K_eq[ketone]), and [1] is equal to [Ru]_tot if catalyst decomposition does not occur. This is consistent with the heterolytic splitting of dihydrogen at the active ruthenium species as the rate-determining step. In competition with this reaction is the reversible addition of acetophenone to the active species to give an enolate complex. The transfer to the ketone of a hydride and proton equivalent that are produced in the heterolytic splitting reaction yields the product in a fast, low activation barrier step. The kinetic isotope effect was measured using D₂ gas and acetophenone-d₃, and this gave values (k_H/k_D) of 1.33 ± 0.15 and 1.29 ± 0.15, respectively. The ruthenium hydride complex [Ru(p-cymene)(m-CH ₂NH₂)H]PF₆ (2) was prepared, as this was postulated to be a crucial intermediate in the outer-sphere bifunctional mechanism. This is inactive under catalytic conditions unless it is activated by a base. DFT computations suggest that the energy barriers for the addition of dihydrogen, heterolytic splitting of dihydrogen, and concerted transfer of H⁺/H^- to the ketone for the outer-sphere mechanism would be respectively 18.0, 0.2, and 33.5 kcal/mol uphill at 298 K and 1 atm. On the other hand, the energy barriers for an inner-sphere mechanism involving the decoordination of the amine group of the NHC ligand, the heterolytic splitting of dihydrogen across a Ru-O(alkoxide) bond, and hydride migration to the coordinated ketone, are respectively 15.5, 17.5, and 15.6 kcal/mol uphill at 298 K and 1 atm. This is more consistent with the experimental observation that the heterolytic splitting of dihydrogen is the turnover-limiting step. This was confirmed by showing that an analogous complex with a tethered teritiary amine group has comparable activity for the H₂-hydrogenation of acetophenone. The related complex [Os(p-cymene)(m-CH₂NH ₂)Cl]PF₆ (6) was synthesized by a transmetalation reaction with [Ni(m-CH₂NH₂)₂](PF₆) ₂ (5) and [Os(p-cymene)Cl₂]₂, and its catalytic activity toward hydrogenation of acetophenone was also investigated.

Full text of DOI:10.1021/om101152m

1236 Organometallics 2011, 30, 1236–1252
DOI: 10.1021/om101152m
Mechanistic Investigation of the Hydrogenation of Ketones Catalyzed by
a Ruthenium(II) Complex Featuring an N-Heterocyclic Carbene with a
Tethered Primary Amine Donor: Evidence for an Inner Sphere Mechanism
Wylie W. N. O, Alan J. Lough, and Robert H. Morris*
Davenport Laboratory, Department of Chemistry, University of Toronto, 80 St. George Street, Toronto,
Ontario M5S 3H6, Canada
Received December 7, 2010
The complex [Ru(p-cymene)(m-CH2NH2)Cl]PF6 (1) catalyzes the H2-hydrogenation of ketones in basic
THF under 25 bar of H2 at 50 °C with a turnover frequency (TOF) of up to 461 h-1 and a maximum
conversion of 99%. When the substrate is acetophenone, the TOF decreases significantly as the catalyst to
substrate ratio is increased. The rate law was then determined to be rate = kH[Ru]tot[H2]/(1 þ
Keq[ketone]), and [1] is equal to [Ru]tot if catalyst decomposition does not occur. This is consistent with
the heterolytic splitting of dihydrogen at the active ruthenium species as the rate-determining step. In
competition with this reaction is the reversible addition of acetophenone to the active species to give an
enolate complex. The transfer to the ketone of a hydride and proton equivalent that are produced in the
heterolytic splitting reaction yields the product in a fast, low activation barrier step. The kinetic isotope
effect was measured using D2 gas and acetophenone-d3, and this gave values (kH/kD) of 1.33 ( 0.15 and
1.29 ( 0.15, respectively. The ruthenium hydride complex [Ru(p-cymene)(m-CH2NH2)H]PF6 (2) was
prepared, as this was postulated to be a crucial intermediate in the outer-sphere bifunctional mechanism.
This is inactive under catalytic conditions unless it is activated by a base. DFT computations suggest that
the energy barriers for the addition of dihydrogen, heterolytic splitting of dihydrogen, and concerted
transfer of Hþ/H- to the ketone for the outer-sphere mechanism would be respectively 18.0, 0.2, and 33.5
kcal/mol uphill at 298 K and 1 atm. On the other hand, the energy barriers for an inner-sphere mechanism
involving the decoordination of the amine group of the NHC ligand, the heterolytic splitting of dihydrogen
across a Ru-O(alkoxide) bond, and hydride migration to the coordinated ketone, are respectively 15.5,
17.5, and 15.6 kcal/mol uphill at 298 K and 1 atm. This is more consistent with the experimental
observation that the heterolytic splitting of dihydrogen is the turnover-limiting step. This was confirmed by
showing that an analogous complex with a tethered teritiary amine group has comparable activity for the
H2-hydrogenation of acetophenone. The related complex [Os(p-cymene)(m-CH2NH2)Cl]PF6 (6) was
synthesized by a transmetalation reaction with [Ni(m-CH2NH2)2](PF6)2 (5) and [Os(p-cymene)Cl2]2, and
its catalytic activity toward hydrogenation of acetophenone was also investigated.
Introduction
hydride complexes,3 or more recently by main-group com-
pounds such as frustrated Lewis pairs.4 Bifunctional
catalysis,5,6 which was originally proposed by Noyori and
co-workers in the pioneering work in the hydrogenation of
polar bonds using the catalyst RuH(η6-p-cymene)((S,S)-
Tsdpen) (Tsdpen = N-(p-toluenesulfonyl)-1,2-diphenyl-
ethylenediamine)7 and, more generally, late-transition-metal
catalysts containing M-H and N-H groups8,9 have superior
The hydrogenation of polar bonds using molecular hydro-
gen has remained by far the most efficient process in terms of
atom economy.1,2 However, dihydrogen is inert to reaction
with most organic substrates of interest, and so it must be
activated by a transition-metal complex, forming metal
*To whom correspondence should be addressed. E-mail: rmorris@
chem.utoronto.ca.
(1) (a) Kubas, G. J. Chem. Rev. 2007, 107, 4152–4205. (b) de Vries,
J. G., Elsevier, C. J., Eds. The Handbook of Homogeneous Hydrogena-
tion; Wiley-VCH: Weinheim, Germany, 2004; Vols. 1-3.
(2) Ito, M.; Ikariya, T. Chem. Commun. 2007, 5134–5142.
(3) (a) Jessop, P. G.; Morris, R. H. Coord. Chem. Rev. 1992, 121, 155–
284. (b) Morris, R. H. Can. J. Chem. 1996, 74, 1907–1915. (c) Rosales,
M. Coord. Chem. Rev. 2000, 196, 249–280. (d) Young, M. B.; Barrow,
J. C.; Glass, K. L.; Lundell, G. F.; Newton, C. L.; Pellicore, J. M.; Rittle,
K. E.; Selnick, H. G.; Stauffer, K. J.; Vacca, J. P.; Williams, P. D.; Bohn,
D.; Clayton, F. C.; Cook, J. J.; Krueger, J. A.; Kuo, L. C.; Lewis, S. D.;
Lucas, B. J.; McMasters, D. R.; Miller-Stein, C.; Pietrak, B. L.; Wallace,
A. A.; White, R. B.; Wong, B.; Yan, Y.; Nantermet, P. G. J. Med. Chem.
2004, 47, 2995–3008. (e) DuBois, M. R.; DuBois, D. L. Chem. Soc. Rev.
2009, 38, 62–72.
(4) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46–76.
(5) (a) Muniz, K. Angew. Chem., Int. Ed. 2005, 44, 6622–6627. (b) Ito,
M.; Ikariya, T. J. Synth. Org. Chem. Jpn. 2008, 66, 1042–1048. (c)
€
Grutzmacher, H. Angew. Chem., Int. Ed. 2008, 47, 1814–1818. (d)
Kuwata, S.; Ikariya, T. Dalton Trans. 2010, 39, 2984–2992. (e) Grotjahn,
D. B. Top. Catal. 2010, 53, 1009–1014.
(6) Ikariya, T.; Murata, K.; Noyori, R. Org. Biomol. Chem. 2006, 4,
393–406.
(7) Haack, K. J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R.
Angew. Chem., Int. Ed. 1997, 36, 285–288.
(8) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org. Chem. 2001,
66, 7931–7944.
(9) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev.
2004, 248, 2201–2237.
r
pubs.acs.org/Organometallics
Published on Web 02/22/2011
2011 American Chemical Society

Article
Organometallics, Vol. 30, No. 5, 2011 1237
activities in the selective reduction of polar bonds to produce
valuable alcohols and amines. Several studies, both
theoretical10-12 and experimental,13-15 including the study of
kinetic isotope effects,16-18 have been conducted to under-
stand the mechanism of action using the “NH-effect” and the
bifunctional nature of the true form of catalytically active
species. The presence of the N-H group and its relation-
ship to the outer-sphere bifunctional19,20 and inner-sphere
mechanisms21,22 have also been studied by many research groups.
The many catalyst systems that were studied by our
research group23-28 which undergo efficient H2-hydroge-
nation of ketones and imines, including those of trans-
Ru(H)2((R)-binap)(tmen), (OC-6-22)-Ru(H)2(PPh3)2(tmen)
(tmen = 2,3-dimethylbutane-2,3-diamine), and trans-Ru-
(H)2(κ4-P2(NH)2)26,29 (P2(NH)2 = tetradentate diphosphi-
nediamine ligand), were found to have the heterolytic
splitting of the coordinated η2-H2 ligand on the active
species as the rate-determining step from various
mechanistic and computational studies.25-28 These are
active catalysts, without prior activation with base, and
catalyze efficiently the reduction of ketones by H2 under
mild conditions.24-27 The energy barrier calculated for the
model complex (OC-6-22)-Ru(H)2(PH3)2(en) (en = ethyl-
enediamine) was found to be higher in the heterolytic
splitting of H2 compared to the concerted transfer of Hþ/
H- to the ketone in a six-membered-ring transition
state.8-11 The corresponding coordinatively unsaturated
complexes containing a ruthenium-amido bond were iso-
lated, and these were also found to activate dihydrogen
to give the trans-dihydride complexes.24-27 For the sys-
tem trans-Ru(H)2(diamine)((R)-binap), Bergens and co-
workers suggested a ruthenium(II) alkoxide complex was
indeed formed prior to the formation of such amido
complexes.15,30
Not much work has been devoted to study the mechanism
of action of catalysts containing phosphine-amine
ligands (P-NH2).31-34 The ruthenium catalysts containing
these ligands effect not only the reduction of ketones32
but also the hydrogenation of a broad range of substrates,
including imines,34 esters,35 epoxides,31 and other
polar bonds.36 All these may utilize the same bifunctional
mechanism involving the action of the M-H and N-H
groups. The notion of replacing the phosphine with
an N-heterocyclic carbene (NHC) donor, in particular,
a donor-functionalized NHC, is therefore attractive to
achieve the goal of greener chemistry.37 We have previously
reported that the transfer hydrogenation catalyst [Ru(p-
cymene)(m-CH2NH2)Cl]PF6 (1) is effective for the hydro-
genation of acetophenone to 1-phenylethanol in basic
2-propanol at 75 °C. This system reached a maximum
conversion of 96% and a turnover frequency (TOF) of
880 h-1 (Figure 1).38 Here we present our study toward its
H2-hydrogenation activity in the reduction of ketones
and a detailed mechanistic investigation, including kinetic stud-
ies and theoretical computations that were performed
to study the possibility of the cooperative nature of the
ligand and the metal center in this new class of ligand
system.
(10) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122,
1466–1478.
(11) (a) Di Tommaso, D.; French, S. A.; Catlow, C. R. A. J. Mol.
Struct. (THEOCHEM) 2007, 812, 39–49. (b) Puchta, R.; Dahlenburg,
L.; Clark, T. Chem. Eur. J. 2008, 14, 8898–8903. (c) Chen, Y.; Tang,
Y. H.; Lei, M. Dalton Trans. 2009, 2359–2364. (d) Lei, M.; Zhang, W. C.;
Chen, Y.; Tang, Y. H. Organometallics 2010, 29, 543–548.
(12) (a) Zhang, H. H.; Chen, D. Z.; Zhang, Y. H.; Zhang, G. Q.; Liu,
J. B. Dalton Trans. 2010, 39, 1972–1978. (b) Chen, Z.; Chen, Y.; Tang,
Y. H.; Lei, M. Dalton Trans. 2010, 39, 2036–2043.
(13) (a) Maire, P.; Buttner, T.; Breher, F.; Le Floch, P.; Grutzmacher,
H. Angew. Chem., Int. Ed. 2005, 44, 6318–6323. (b) Friedrich, A.; Drees,
M.; auf der Gunne, J. S.; Schneider, S. J. Am. Chem. Soc. 2009, 131,
17552–17553.
(14) (a) Sandoval, C. A.; Ohkuma, T.; Utsumi, N.; Tsutsumi, K.;
Murata, K.; Noyori, R. Chem. Asian J. 2006, 1, 102–110. (b) Ohkuma,
T.; Utsumi, N.; Tsutsumi, K.; Murata, K.; Sandoval, C.; Noyori, R. J.
Am. Chem. Soc. 2006, 128, 8724–8725.
(15) Hamilton, R. J.; Leong, C. G.; Bigam, G.; Miskolzie, M.;
Bergens, S. H. J. Am. Chem. Soc. 2005, 127, 4152–4153.
(16) Sandoval, C. A.; Ohkuma, T.; Muniz, K.; Noyori, R. J. Am.
Chem. Soc. 2003, 125, 13490–13503.
(17) Kass, M.; Friedrich, A.; Drees, M.; Schneider, S. Angew. Chem.,
Int. Ed. 2009, 48, 905–907.
(18) Zimmer-De Iuliis, M.; Morris, R. H. J. Am. Chem. Soc. 2009,
131, 11263–11269.
(19) (a) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40–
73. (b) Ma, G. B.; McDonald, R.; Ferguson, M.; Cavell, R. G.; Patrick,
B. O.; James, B. R.; Hu, T. Q. Organometallics 2007, 26, 846–854. (c)
Baratta, W.; Ballico, M.; Esposito, G.; Rigo, P. Chem. Eur. J. 2008, 14,
5588–5595. (d) Sandoval, C. A.; Shi, Q. X.; Liu, S. S.; Noyori, R. Chem.
Asian J. 2009, 4, 1221–1224.
(20) Ito, M.; Hirakawa, M.; Murata, K.; Ikariya, T. Organometallics
2001, 20, 379–381.
(21) (a) Leong, C. G.; Akotsi, O. M.; Ferguson, M. J.; Bergens, S. H.
Chem. Commun. 2003, 750–751. (b) Phillips, S. D.; Fuentes, J. A.;
Clarke, M. L. Chem. Eur. J. 2010, 16, 8002–8005.
(30) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2008, 130,
11979–11987.
(31) Ito, M.; Hirakawa, M.; Osaku, A.; Ikariya, T. Organometallics
2003, 22, 4190–4192.
(32) (a) Guo, R.; Lough, A. J.; Morris, R. H.; Song, D. Organome-
tallics 2004, 23, 5524–5529. (b) Jia, W. L.; Chen, X. H.; Guo, R. W.;
Sui-Seng, C.; Amoroso, D.; Lough, A. J.; Abdur-Rashid, K. Dalton
Trans. 2009, 8301–8307.
(22) Lundgren, R. J.; Rankin, M. A.; McDonald, R.; Schatte, G.;
Stradiotto, M. Angew. Chem., Int. Ed. 2007, 46, 4732–4735.
(23) Abdur-Rashid, K.; Lough, A. J.; Morris, R. H. Organometallics
2001, 20, 1047–1049.
(24) Abdur-Rashid, K.; Faatz, M.; Lough, A. J.; Morris, R. H. J. Am.
Chem. Soc. 2001, 123, 7473–7474.
(25) (a) 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. (b) Abbel, R.; Abdur-Rashid, K.; Faatz, M.; Hadzovic, A.;
Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2005, 127, 1870–1882.
(26) Rautenstrauch, V.; Hoang-Cong, X.; Churlaud, R.; Abdur-
Rashid, K.; Morris, R. H. Chem. Eur. J. 2003, 9, 4954–4967.
(27) Clapham, S. E.; Morris, R. H. Organometallics 2005, 24, 479–
481.
(28) Hadzovic, A.; Song, D.; MacLaughlin, C. M.; Morris, R. H.
Organometallics 2007, 26, 5987–5999.
(29) (a) Li, T.; Churlaud, R.; Lough, A. J.; Abdur-Rashid, K.;
Morris, R. H. Organometallics 2004, 23, 6239–6247. (b) Li, T.; Bergner,
I.; Haque, F. N.; Zimmer-De Iuliis, M.; Song, D.; Morris, R. H.
Organometallics 2007, 26, 5940–5949.
(33) (a) Dahlenburg, L.; Gotz, R. Eur. J. Inorg. Chem. 2004, 888–905.
(b) Blaquiere, N.; Diallo-Garcia, S.; Gorelsky, S. I.; Black, D. A.;
Fagnou, K. J. Am. Chem. Soc. 2008, 130, 14034–14035.
(34) Abdur-Rashid, K.; Guo, R. W.; Lough, A. J.; Morris, R. H.;
Song, D. T. Adv. Synth. Catal. 2005, 347, 571–579.
(35) (a) Saudan, L. A.; Saudan, C. M.; Debieux, C.; Wyss, P. Angew.
Chem., Int. Ed. 2007, 46, 7473–7476. (b) Kuriyama, W.; Ino, Y.; Ogata,
O.; Sayo, N.; Saito, T. Adv. Synth. Catal. 2010, 352, 92–96.
(36) (a) Ito, M.; Sakaguchi, A.; Kobayashi, C.; Ikariya, T. J. Am.
Chem. Soc. 2007, 129, 290–291. (b) Ito, M.; Koo, L. W.; Himizu, A.;
Kobayashi, C.; Sakaguchi, A.; Ikariya, T. Angew. Chem., Int. Ed. 2009,
48, 1324–1327.
(37) (a) Lee, H. M.; Lee, C. C.; Cheng, P. Y. Curr. Org. Chem. 2007,
11, 1491–1524. (b) Kuhl, O. Chem. Soc. Rev. 2007, 36, 592–607. (c)
Normand, A. T.; Cavell, K. J. Eur. J. Inorg. Chem. 2008, 2781–2800. (d)
Corberan, R.; Mas-Marza, E.; Peris, E. Eur. J. Inorg. Chem. 2009, 1700–
1716. (d) Diez-Gonzalez, S.; Marion, N.; Nolan, S. P. Chem. Rev. 2009,
109, 3612–3676.
(38) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2009,
28, 6755–6761.

1238 Organometallics, Vol. 30, No. 5, 2011
O et al.
Table 2. H2-Hydrogenation of Ketones Catalyzed by Complex 1
Figure 1. Transfer hydrogenation of acetophenone catalyzed
by complex 1 in basic 2-propanol.
conversnc (%/h)
entrya
substrate
TOF (h-1
)
Table 1. H2-Hydrogenation of Acetophenone to 1-Phenylethanol
Catalyzed by Complex 1
1
a
b
b
c
d
e
f
g
h
i
88/1
78/1
40/1
94/1
94/1
0/1
94/1
79/1
3/1
99/2
98/2
77/3
213
208
179
461
377
2
3b
4
99/1.5
99/1.5
0/2
99/1.5
96/2.5
4/1.5
99/2.5
99/0.5
5
6
7
8
9
10
11
260
215
5
164
838
conversnc (%/h)
74/1
70/0.16
entrya C/B/Sb solvent
base
TOF (h-1 d
)
j
a Reactions were carried out in a 50 mL Parr hydrogenation reactor at
25 bar of H2 pressure at 50 °C using THF (6 mL). Potassium tert-
butoxide was used as base. The C/B/S ratio was 1/8/200. b The C/B/S
ratio was 1/8/400. c Conversions were determined by GC and are
reported as an average of two runs.
1
2
3
4
5
6
7
8e
1/8/200
1/8/200
1/8/200
1/0/200
0/8/200
1/8/600
iPrOH KOtBu
96/1
88/1
5/1
0/1
0/1
20/1
6/1
0/1
99/1.5
99/2
11/2
287
213
THF
THF
THF
THF
THF
KOtBu
NaOtBu
none
KOtBu
KOtBu
KOtBu
KOtBu
57/2.5
28/3
122
73
1/8/1200 THF
1/8/200 THF
H2-Hydrogenation of Other Ketones Catalyzed by Complex
1. Other ketones with different steric bulk were investigated
(Table 2, entries 1-6). In general, TOF values increase with a
larger group next to the polar bond. Of note, substrates
without R-C-H protons next to the carbonyl group did not
show inhibition of the catalysis by the substrate. For exam-
ple, a 2-fold increase in the substrate loading in the hydro-
genation of benzophenone gave similar initial rates and TOF
values (Table 2, entries 4 and 5, and the Supporting Informa-
tion, Figure S1). On the other hand, complex 1 did not
catalyze the hydrogenation of deoxybenzoin in basic THF
(entry 6). All these results are suggestive of the formation of
an enolate complex derived from the active ruthenium
species. This likely causes a decrease in the concentration
of the active ruthenium species within the catalytic cycle (vide
infra). An increase in the donor ability on going from chloro
to methoxy substituents at the 40-position of the aryl group
on the ketone led to a slight decrease in the TOF values
(entries 7 and 8). Placing the chloro group ortho to the
ketone results in a dramatic decrease in rate (entry 9).
Pinacolone and benzaldehyde were also effectively hydro-
genated under similar reaction conditions (Table 2, entries 10
and 11).
Kinetic Studies. The mechanism of action of the H2-
hydrogenation of ketones catalyzed by complex 1 was
probed by determining the rate law. Acetophenone was
chosen as the substrate of interest, and its concentration
along with that of 1-phenylethanol was conveniently mon-
itored by gas chromatography (GC). The catalytic condi-
tions were varied from the standard conditions (0.83 mM [1],
25 bar of H2, 0.17 M acetophenone and 7.4 mM potassium
tert-butoxide; catalyst/base/H2/ketone = 1/8/120/200) by
changing the concentration of a single component of interest,
using 0.28-0.83 mM of 1, 5-25 bar of H2, 0.17-1.0 M
a Reactions were carried out in a 50 mL Parr hydrogenation reactor at
25 bar of H2 pressure at 50 °C using THF or iPrOH (6 mL) as the solvent.
b C/B/S = catalyst/base/substrate. c Conversions were determined by
GC and are reported as an average of two runs. d TOF = turnover
frequency, measured from the slope of the linear portion of an [alcohol]
versus time plot. e Reaction conducted under argon.
Results and Discussion
H2-Hydrogenation of Acetophenone Catalyzed by Complex
1. The results pertaining to the catalytic activity of complex 1
in the H2-hydrogenation of acetophenone are given in Table
1. Full conversion to 1-phenylethanol is achieved in 1.5 h
under 25 bar of H2 in 2-propanol at 50 °C with a catalyst to
base to substrate (C/B/S) ratio of 1/8/200 on activation in the
presence of potassium tert-butoxide (KOtBu) as the base.
This took somewhat longer (2 h) when an aprotic solvent,
tetrahydrofuran (THF), was used instead (Table 1, entries 1
and 2). The ruthenium(II) complex is not a catalyst when not
activated by base. It exhibited low activity when sodium tert-
butoxide, which has a low solubility in THF, was used. In
addition, the ruthenium(II) complex is not a catalyst when
activated by base in the absence of H2 in THF (Table 1,
entries 3, 4 and 8). More importantly, an increase in substrate
loading to a catalyst to substrate (C/S) ratio of 1/1200
decreased the TOF value by 3-fold, giving a value of 73
h
-1 (Table 1, entries 6 and 7). The catalyst [Ru(η5-Cp*)(m-
CH2NH2)(py)]PF6 (Cp* = pentamethylcyclopentadienyl;
py = pyridine), which contains the same chelating primary
amine-NHC ligand, tolerates high substrate loading (C/S
ratio up to 1/11 500) when activated by base.39
(39) O, W. W. N.; Lough, A. J.; Morris, R. H. Chem. Commun. 2010,
46, 8240–8242.

Article
Organometallics, Vol. 30, No. 5, 2011 1239
Table 3. Kinetic Data for the Hydrogenation of Acetophenone
Catalyzed by Complex 1
initial rate (v0), 10-5 M s-1
[H2/D2]b,
mM
runa [1], mM
[ketone], M
exptlc
calcdd
1
2
3
4
5
6
7
8
0.28
0.55
0.83
0.83
0.83
0.83
0.83
0.83
0.83
0.83
0.83
100
100
100
60
32
20
100
100
100
33
0.17
0.17
0.17
0.17
0.17
0.17
0.33
0.50
1.0
1.7
3.4
4.9
2.9
1.5
0.76
3.4
2.8
1.7
1.3
4.9
1.6
3.2
4.8
2.9
1.5
0.97
3.6
2.8
1.7
9
10e
11f
0.17
0.16
100
a Reactions were carried out in a 50 mL Parr hydrogenation reactor at
the required conditions at 50 °C. THF was the solvent, and potassium
tert-butoxide (7.4 mM, C/B = 1/8) was used as the base. b The
solubilities of H2 and D2 were taken from ref 40. c Values obtained from
the least-squares fits of the data plotted in Figures 2 and 3. d Values
calculated from the proposed rate law given by eq 1. e Reaction
conducted using 8 bar of D2 gas. f Reaction conducted using acetophe-
none-d3 (0.16 M) and H2 gas (25 bar).
Figure 3. Linear plot showing the relationship between the
reciprocal of the initial rate (1/v0 in 103 M-1 s) and acetophe-
none concentration (M). The rate (kH) and equilibrium (Keq)
constants were derived from the slope and the y intercept
according to eq 2, respectively.
conditions and kinetic data from the initial linear portion of
the plots of [alcohol] versus time are given in Figure 2 and in
Table 3.
Under pseudo-first-order conditions, it was determined
that the rates, within experimental error, are first order in the
concentrations of the complex 1 and of hydrogen. A plot of
the initial rate (v0, in M s-1) versus [1] (in mM) yielded a
straight line with a slope of (60 ( 1) ꢀ 10-3 s-1, while that
of v0 versus [H2] in THF40 gave a straight line with a slope
of (48 ( 1) ꢀ 10-5 s-1. Each plot passed through the origin.
A plot of initial rate (in M s-1) versus [ketone] (in M)
yielded a hyperbola which does not pass through the
origin (Figure 2) and gave a reaction order of -0.6 on the
ketone. The rate law, on the other hand, is zero order in
the concentrations of 1-phenylethanol and potassium tert-
butoxide within the range of concentrations of interest
(vide infra).
Figure 2. Kinetic data showing the production of 1-phenyletha-
nol from acetophenone catalyzed by complex 1 in basic THF: (a)
dependence on the catalyst concentration (1); (b) dependence on
the hydrogen concentration; (c) dependence on the acetophe-
none concentration. The inset shows the dependence of initial
rates (v0, ꢀ10-5 M s-1) and the concentration of the analyte of
interest. Details of the reaction conditions are given in Table 3
and in the Supporting Information (Tables S1-S3).
acetophenone, 0-35 mM 1-phenylethanol, and 2.9-18 mM
potassium tert-butoxide. The reaction temperature was kept
at 50 °C for all of the runs. Some representative catalytic

1240 Organometallics, Vol. 30, No. 5, 2011
O et al.
Scheme 1. Reaction Scheme Showing the Definitions of the Rate
and Equilibrium Constants
occur. By taking the reciprocal of eq 1, the reciprocal of the
rate and the ketone concentration is related by eq 2.
1
1
Keq½ketoneꢁ
¼
þ
ð2Þ
rate
kH½Ruꢁtot½H2ꢁ kH½Ruꢁtot½H2ꢁ
This will allow the calculation of kH and Keq values, respec-
tively, using the slope and the y intercept of such a linear plot.
Indeed, a plot of the reciprocal of initial rate (in M-1 s) versus
[acetophenone] (in M) using the kinetic data is linear with a y
intercept of (12.9 ( 0.7) ꢀ 103 M-1 s-1 and a slope of (46.6 (
1.2) ꢀ 103 M-2 -1 (Figure 3). Thus the rate and equilibrium
S
constants obtained using the rate law given in eq 1 for the
range of acetophenone, complex 1, hydrogen, 1-phenyletha-
nol and base concentrations studied are
A possible general form of the course of the reaction is
proposed as shown in Scheme 1 to explain the kinetic data
and the experimental findings. First, an active species is
quickly formed by the reaction of complex 1 with a base.
This species can then either reversibly react with acetophe-
none (Keq(H/D)) to produce an enolate complex or, in the rate-
determining step (kH/D), irreversibly react with hydrogen (or
deuterium) gas to give a second reactive ruthenium complex
responsible for the reduction of the ketone to the product
alcohol (Scheme 1). This second complex is likely to be a
hydride (or deuteride) formed by the heterolytic splitting of
dihydrogen (or dideuterium).
kH ¼ 0:94 ( 0:05 M-1 s-1
Keq ¼ 3:6 ( 0:2 at 323 K
The Keq value was also used to calculate the rate constant
kH from the plots of initial rate versus [1] or [H2], and the
values obtained are within experimental error compared to
the value obtained from the plot of reciprocal of initial rate
versus [acetophenone]. The observed and calculated initial
rates given in Table 3 also are in good agreement.
Similar observations were reported in the H2-hydrogena-
tion of acetophenone using the precatalyst trans-RuHCl-
(κ4-(S,S)-cyP2(NH)2) when activated by base under 12 bar
of H2 in 2-propanol at 20 °C: a 3-fold increase in the C/S
ratio to 1/12 500 led to a decrease in the initial rate by
25%.26 For the system which might undergo the outer-
sphere bifunctional mechanism, it was postulated that
a reversible reaction between acetophenone and the
ruthenium-amido complex RuH(κ4-(S,S)-cyP2(NH)-
(NH2)) forms the ruthenium-ketone adduct, and this
equilibrium outcompetes the coordination of dihydrogen
to ruthenium. This had the effect of slowing catalysis, since
it was proposed that the heterolytic splitting of η2-H2 by the
amido complex to afford the bifunctional dihydride catalyst
trans-Ru(H)2(κ4-(S,S)-cyP2(NH2)) was rate-limiting. The
binding of the ketone to the amido complex might also lead
to the formation of a stable enolate complex: this could occur
by the deprotonation of an R-C-H proton from the coordi-
nated acetophenone by the ruthenium-amido complex
(Scheme 1).
Effect of the Base on Catalysis. The relationship between
the concentration of potassium tert-butoxide (KOtBu) and
rates were examined by means of kinetic studies. A plot of
initial rate (in M s-1) versus [KOtBu] (in M) yielded a straight
line with a slope of -(9.5 ( 1.3) ꢀ 10-4 s-1 which does not
pass through the origin (see the Supporting Information,
Figure S2). The reaction order was -0.2, as determined by
the kinetic data. The concentration of the alkoxide base,
therefore, does not greatly influence the rate of catalysis. In
addition, potassium tert-butoxide in high concentrations
is known to form stable aggregates in solvents with low
dielectric constants, such as THF. For instance, the reac-
tion of Pt(papH)Cl (papH = κ3-C,N,N-2-phenyl-6-(2-
aminoisopropyl)pyridine) with 5 equiv of KOtBu in benzene
afforded the amido complex [Pt(pap)]2(KOtBu)8(KCl).41
The chloride anion was encapsulated by seven potassium
ions, and the potassium ions were bridged by oxygens of the
tert-butoxide anions. The decrease in rate in catalysis with an
increasing concentration of the base (up to 18 mM) can
plausibly be attributed to complex ion-pairing and aggrega-
tion effects between cationic potassium and ruthenium
species with the anionic tert-butoxide, the enolate of acet-
ophenone, and the alkoxide of 1-phenylethanol. Potassium
ions are known to form solvates with THF molecules, but
these are weak in comparison to strong anion-cation ion-
pairing interactions. On the other hand, a decrease in rate
was also observed when a low concentration of base (1.2
mM) was used in catalysis (C/B/S = 1/1.4/200).
The rate law (eq 1) can be derived from this general
reaction scheme (see the Supporting Information for
derivation):
d½ketoneꢁ
dt
kH½Ruꢁtot½H2ꢁ
rate ¼ -
¼
ð1Þ
1 þ Keq½ketoneꢁ
in which [Ru]tot is the total concentration of ruthenium
species, after complex 1 is activated with base. This suggests
that [1] is equal to [Ru]tot if catalyst decomposition does not
It has been shown that for certain transfer42 and H2-
hydrogenation43 systems, particular alkali-metal cations
(40) This was modeled using the data obtained for 1,2-dimethox-
yethane at 323 K. The Henry’s Law constant for the solubility of H2 in
THF at 50 °C was therefore modeled to be 3.98 ꢀ 10-2 M/bar; see: (a)
Solubility Data Series, Hydrogen and Deuterium, 5/6; Young, C. L., Ed.;
Pergamon Press: New York, 1981. Deuterium gas is 1.046 times more
soluble than hydrogen gas in water at 50 °C; see: (b) Muccitelli, J.; Wen, W. Y.
J. Solution Chem. 1978, 7, 257–267.
(41) Song, D.; Morris, R. H. Organometallics 2004, 23, 4406–4413.
(42) (a) Vastila, P.; Zaitsev, A. B.; Wettergren, J.; Privalov, T.;
Adolfsson, H. Chem. Eur. J. 2006, 12, 3218–3225. (b) Wettergren, J.;
Buitrago, E.; Ryberg, P.; Adolfsson, H. Chem. Eur. J. 2009, 15, 5709–
5718.
(43) Hartmann, R.; Chen, P. Angew. Chem., Int. Ed. 2001, 40, 3581–
3585.

Article
Organometallics, Vol. 30, No. 5, 2011 1241
can accelerate the catalysis by acting as a Lewis acid. This
possibility was examined by our system. The addition of
[2.2.2]cryptand in equimolar amount to potassium ions (C/
B/S = 1/1.4/200) led to comparable reaction rates with
respect to the standard condition for catalysis (C/B/S = 1/
8/200). An equimolar amount of 18-crown-6 that was added
with respect to base (C/B/S = 1/8/200, [KOtBu] = 7.4 mM)
had no effect on the reaction rate. The use of a base with
sodium ions (NaOtBu, Table 1, entry 3) gave low activity in
catalysis. This, however, was attributed to the limited solu-
bility of the base in THF solution. Therefore, the cations of
the alkoxide base do not significantly affect the rate of
catalysis.26
Effect of Alcohols on Catalysis. A plot of initial rate (in
M s-1) versus [1-phenylethanol] (in M) added in excess
obtained in kinetic studies yielded a straight line with a slope
of (1.1 ( 0.5) ꢀ 10-4 s-1 which does not pass through the
origin (see the Supporting Information, Figure S3). The
reaction order was -0.1, as determined by the kinetic data.
The product alcohol and 2-propanol, therefore, have mini-
mal effect on catalysis and do not contribute to the rate law.
Of note, the rate of conversion was somewhat higher using
2-propanol (Table 1), due to the combined activities of both
H2 and transfer hydrogenation.38
Figure 4. Linear plots showing the relationship between the
reciprocal of the initial rate (1/v0 in 103 M-1 s) and acetophe-
none concentration (M), using (a) D2 gas (8 bar) and acetophe-
none, (b) H2 gas (25 bar) and acetophenone, and (c) H2 gas (25
bar) and acetophenone-d3 (0.16-0.49 M), in the production of
1-phenylethanol from acetophenone catalyzed by complex 1 in
basic THF. Detailsofthereactionconditionsaregiven inTable3
and in the Supporting Information (Tables S4-S6).
Isotope Effects and Deuterium Labeling Studies. The mea-
surement of kinetic isotope effects (KIE)44 is a powerful
technique to gain important mechanistic insight in catalysis.
The use of both hydrogen and deuterium gas in the measure-
ment of the kinetic isotope effect for hydrogenation catalysts
is common,17,45 but it is less well known for homogeneous
bifunctional catalysts for the H2-hydrogenation of polar
bonds.16,18 Noyori and co-workers have reported a KIE
value (kH/kD) of 2 for the hydrogenation of acetophenone
catalyzed by trans-RuH(η1-BH4)((S)-tolbinap)((S,S)-dpen)
in the presence of base in H2/2-propanol versus D2/2-pro-
panol-d8.16 We have recently reported a KIE value of 2.0 (
0.1 for the hydrogenation of acetophenone catalyzed by
trans-Ru(H)2((R)-binap)(tmen). This was interpreted as an
early transition state involving the heterolytic splitting of
dihydrogen (dideuterium) by the ruthenium catalyst where
there is little weakening of the H-H/D-D bond.18
The rates ofthereaction using D2 gas were examinedwithin
a given range of acetophenone concentrations (0.083-0.16
M) at low D2 gas pressure (8 bar) and 50 °C in basic THF to
obtainmeaningfulkinetic data. ThekD andKeq(D) valueswere
calculated from the slope and y intercept of a plot of the
reciprocal of the initial rate (in M-1 s) and [acetophenone] (in
M) as in eq 2. This gave a linear plot with a y intercept of (51 (
5) ꢀ 103 M-1 s-1 and a slope of (15 ( 4) ꢀ 104 M-2 s-1. The
catalysis conditions along with the results of the kinetic data
and calculated values of kD and Keq(D) are given in Figure 4,
in Tables 3 and 4, and in the Supporting Information (Tables
S4 and S5 and Figure S4). The experimentally determined
kinetic isotope effect value (kH/kD), within experimental
error, is 1.33 ( 0.15 for using H2 versus D2 gas.
Table 4. Isotope Effect for the Hydrogenation of Acetophenone
Catalyzed by Complex 1
deuterium sourcea
kD (M-1 s-1 b
)
Keq(D)
KIE (kH/kD)
b
D2 (8 bar, 33 mM)
acetophenone-d3
(0.16 - 0.49 M)
0.70 ( 0.07
0.72 ( 0.05
2.9 ( 0.8
1.3 ( 0.2
1.33 ( 0.15
1.29 ( 0.11
a Reactions were carried out in a 50 mL Parr hydrogenation reactor at
the required conditions at 50 °C. THF was the solvent, and potassium
tert-butoxide (7.4 mM) was used as base. b Values obtained from the
slopes and intercepts of the linear plots given in Figure 4 and eq 2.
by allowing enolate formation, it was important to determine
the effect of deuteration of these positions in acetophenone.
Kinetic studies were therefore performed using acetophe-
none-d3 (0.16-0.49 M) at 25 bar of H2 and 50 °C in basic
THF. The kD and Keq(D) values were computed in a fashion
similar to that described above by first plotting the reciprocal
of initial rate (in M-1 s) versus [acetophenone-d3] (in M,
Figure 4) to yield a y intercept of (17 ( 1) ꢀ 103 M-1 s-1 and a
slope of (22 ( 3) ꢀ 103 M-2 s-1. The catalysis conditions,
results of the kinetic data, and calculated values of kD and
Keq(D) are given in Tables 3 and 4 and in the Supporting
Information (Tables S4 and S6 and Figure S5).
The deuterated products were examined by NMR spec-
troscopy. For catalysis using 8 bar of D2 gas (C/B/S = 1/8/
200) at 50 °C, deuterium incorporation into the methyl
groups of the unreacted acetophenone and of the product
1-phenylethanol was observed at 50% conversion by use of
1H (in CDCl3) and 2D NMR analysis, with C6D6 serving as
an external reference. Deuteration of the phenyl ring as a
result of C-H activation of the ketone substrate was not
observed. Deuteration by D2 gas at the alcoholic oxygen and
the carbonyl carbon was observed (>81% enrichment).
Likewise, for catalysis using acetophenone-d3 and 25 bar of
H2 gas (C/B/S = 1/8/200) at 50 °C, hydrogen scrambling
with deuterium in the methyl group was observed in both
acetophenone and 1-phenylethanol at 91% conversion.
Since the presence of R-C-H groups adjacent to the
carbonyl group of a ketone was shown to inhibit catalysis
(44) Westheimer, F. H. Chem. Rev. 1961, 61, 265–273.
(45) (a) Chock, P. B.; Halpern, J. J. Am. Chem. Soc. 1966, 88, 3511–
3514. (b) Brown, J. M.; Parker, D. Organometallics 1982, 1, 950–956. (c)
Kitamura, M.; Tsukamoto, M.; Bessho, Y.; Yoshimura, M.; Kobs, U.;
Widhalm, M.; Noyori, R. J. Am. Chem. Soc. 2002, 124, 6649–6667. (d)
Imamoto, T.; Itoh, T.; Yoshida, K.; Gridnev, I. D. Chem. Asian J. 2008,
3, 1636–1641.

1242 Organometallics, Vol. 30, No. 5, 2011
O et al.
Scheme 2. Possible Outer-Sphere Mechanism That Involves Bifunctional Catalysis in the Hydrogenation of Acetophenone
Scheme 3. Synthesis of Complex 2 from Reaction of Complex 1
and Basic 2-Propanol
deuterated ketone (H2, acetophenone-d3) (Scheme 1 and
Table 4). We conclude that the small kinetic isotope effect
that is observed upon the use of D2 gas in the reduction of
acetophenone is indicative of an early transition state in the
heterolytic splitting of H2 combined with an expected
inverse equilibrium isotope effect for the H2/D2 binding to
the active ruthenium species.46
18
The Disfavored Outer-Sphere Mechanism. Two mechan-
istic proposals should be considered in light of the experi-
mental findings described above. While the outer-sphere
mechanism9 is expected, there is stronger evidence for an
inner-sphere mechanism as described below. In an outer-
sphere mechanism which would involve the bifunctional
nature of a primary amine group (NH2) and the me-
tal-hydride bond (Ru-H), the activation of complex 1 with
base should give a metal-amido complex containing a
trigonal-planar nitrogen donor atom (Scheme 2). Coordina-
tion of dihydrogen to such a complex and subsequent
heterolytic splitting of the molecule should give the bifunc-
tional metal-hydride and protic amine grouping. The hy-
dride and the proton equivalent would then transfer to the
ketone in the outer coordination sphere, giving the product
alcohol, without coordination to the metal center. In this
mechanism the observed inhibition by enolizable ketones
would be explained by the reaction of the ketone with the
metal-amido complex.26 Deprotonation of the methyl
group should afford the enolate complex with the primary
amine group restored.
In order to identify reaction intermediates in the outer-
sphere mechanism, reactions of excess or stoichiometric
amounts of base (KOtBu, KH, NaBH4, or K-Selectride)
with complex 1 in THF were tried, but these gave multiple
unidentifiable products. Attempts to isolate such products
failed and instead led to instant decomposition. Attempts to
prepare the ruthenium-hydride complex 2 (Scheme 2) by
using dihydrogen (up to 5 bar) in excess KOtBu in stirring
THF at 25 or 50 °C also failed. However, complex 2 could be
prepared by warming complex 1 and 3 equiv of sodium
Deuteration at the alcoholic oxygen and the carbonyl
carbon was not important in this case (<3% enrichment).
Deuterium scrambling into the methyl group of the aceto-
phenone and 1-phenylethanol may occur via protonation/
deuteration of an enolate complex, since a stoichiometric
reaction of acetophenone-d3 and 1-phenylethanol in basic
THF under an argon atmosphere at 50 °C also leads to
deuterium/proton scrambling in the methyl group of the
ketone but not in the alcohol (see the Supporting Informa-
tion, Figure S6). The former could also occur by Hþ/Dþ
exchange with the amine ligand or with the alcoholic proton
of coordinated 1-phenylethanol or tert-butanol. The forma-
tion of stable ruthenium(II) complexes containing strong
deuterium-heteroatom bonds in the ligands and a strong
ruthenium-deuteride bond is thermodynamically favored.
This makes the transfer of Dþ/D- to the ketone from the
metal difficult.
The use of D2 and acetophenone-d3 as deuterium sources
helps to confirm the rate law equation and the kinetic model
shown in Scheme 1. If the heterolytic splitting of dihydrogen
is the rate-determining step in the catalytic cycle, the use of
acetophenone-d3 as a substrate should have a small kinetic
isotope effect on kH but a large isotope effect on the
equilibrium between the active species and the enolate com-
plex. This is exactly what is observed. On the other hand, the
use of D2 gas gave a kinetic isotope effect but had little
influence on the equilibrium between the active species and
the enolate complex. The equilibrium constant values for
enolate formation decrease as the number of deuterium
atoms available in the system increases on going from the
use of a protic source (H2, acetophenone) to a partially
(46) Parkin, G. Acc. Chem. Res. 2009, 42, 315–325 and references
therein.

Article
Organometallics, Vol. 30, No. 5, 2011 1243
Table 5. Selected Crystal Data, Data Collection, and Refinement
Parameters for 2 and 6a
6
2
empirical formula
C21H27ClF6
N3OsP
692.08
C21H28F6N3
RuP C4H8O
640.61
3
fw
lattice type
space group
T, K
monoclinic
P21/n
150
12.7810 (15)
16.890 (2)
13.5646 (16)
90
113.406(9)
90
2687.3(6)
4
1.711
4.959
1344
0.20 ꢀ 0.08 ꢀ 0.02
2.8-25.0
9182/4338
orthorhombic
Pbca
150
16.8737(3)
14.6964(6)
22.3909(7)
90
90
90
5552.6(3)
8
1.533
0.686
2624
0.22 ꢀ 0.12 ꢀ 0.08
2.59-25.00
24 408/4879
˚
a, A
˚
b, A
˚
c, A
R, deg
β, deg
γ, deg
3
˚
V, A
Z
F
calcd, Mg m-3
μ(Mo KR), mm-1
F(000)
cryst size, mm3
θ range collected, deg
no. of collected/
unique rflns
abs cor
max, min transmission
coeff
semiempirical from equivalents
0.960, 0.828
Figure 5. ORTEP diagram of 2 ([Ru(p-cymene)(m-CH2NH2)-
H]PF6) depicted with thermal ellipsoids at the 30% probability
level. The counteranion, solvent molecule, and most of the
hydrogens have been omitted for clarity. Selected bond dis-
0.9073, 0.4371
no. of params refined
goodness of fit
R1 (I > 2σ(I))
wR2 (all data)
286
0.933
342
1.038
˚
tances (A) and bond angles (deg): Ru(1)-C(1), 2.029(7); Ru-
0.0598
0.1520
0.0568
0.1800
(1)-N(3), 2.145(5); Ru(1)-H(1ru), 1.79(7); Ru(1)-C(12),
2.274(6); Ru(1)-C(15), 2.232(6); C(1)-Ru(1)-N(3), 91.9(2);
C(1)-Ru(1)-H(1ru), 81(2); H(1ru)-Ru(1)-N(3), 84(2).
-3
˚
peak and hole, e A
1.536 and -1.777
1.556 and -0.630
P
P
P
a Definition of R indices: R1 = (Fo - Fc)/ (Fo); wR2 = [ [w(Fo
2
P
- Fc2)2]/ [w(Fo2)2]]1/2
.
2-propoxide in 2-propanol solution at 50 °C for 3 h, giving a
brown solid upon isolation in 57% yield (Scheme 3). The use
of KOtBu or a reaction that was conducted in refluxing
2-propanol gave impure products. Complex 2 is oxygen
sensitive both in solution and in the solid state. A solution
of 2 in dichloromethane readily converts to the cation of
complex 1 at room temperature in hours and slowly reacts
with solvents such as acetonitrile after prolonged standing
for days.
2-propanol) conditions. No catalytic activity was observed
in the H2 hydrogenation of acetophenone in the absence of
base when a C/S ratio of 1/200 was used. A conversion of 3%
to 1-phenylethanol and a maximum conversion of 9% were
achieved in 2 and 19 h, respectively, under transfer hydro-
genation conditions in the absence of base when the same
C/S ratio was used. However, addition of base (KOtBu)
to such a reaction mixture activates the complex to
catalyze the hydrogenation of acetophenone. Full con-
version to 1-phenylethanol is achieved in 45 min with a
C/B/S ratio of 1/8/200 under identical H2-hydrogenation
conditions as above and to 88% conversion in 2 h under
identical transfer hydrogenation conditions. Plots of
the concentration of product alcohol versus time for both
H2-hydrogenation and transfer hydrogenation were sig-
moidal with a short induction period of 6 and 30 min,
respectively (Figure 6). The induction period can be ex-
plained by the conversion of 2 by reaction with the base into a
catalytically active species. The addition of 1-phenylethanol
to the starting mixture led to a slower reaction rate and a
slightly longer induction period. Therefore, the alcohol does
not act as a proton shuttle in the heterolytic splitting of H2 in
the transition state, in contrast to other bifunctional systems
that were studied.2,20,28,39
Complex 2 was unambiguously characterized by an X-ray
diffraction study (Figure 5, Table 5). The compound adopts
a piano-stool geometry about the metal center. The Ru-
˚
Ccarbene bond (2.029(7) A) was shorter than that of complex 1
(2.092(5) A).38 The large bite angle between the carbene and
˚
primary amine donor containing the metal center
(C(1)-Ru(1)-N(3) = 91.9(2)°) is typical of L-Ru-L an-
gles for ruthenium arene complexes ([Ru(η6-arene)HL2]þ)
containing a hydride ligand.7,22,47 Note that the contact
˚
distance of 2.45 A between the ruthenium hydride and the
primary amine protons (Ru-H H-N-) is comparable to
3 3 3
that for the complex RuH(η6-p-cymene)((S,S)-Tsdpen)
˚
(contact distance 2.29 A) reported by Noyori and co-
workers.7 In solution, the Ru-Ccarbene resonance was
observed at 185.3 ppm in the 13C NMR spectrum. The
1
Ru-H resonance was observed at -7.82 ppm in the H
NMR spectrum in CD3CN. This lies in the region for
analogous ruthenium-hydride complexes of the type
[Ru(η6-arene)HL2]þ.7,22,47
The catalytic activity of the hydride complex 2 was
then tested under H2-hydrogenation (25 bar, 50 °C in THF
or 2-propanol) and transfer hydrogenation (75 °C in
Interestingly, the cationic complex 2 containing a ruthe-
nium hydride and primary amine grouping does not transfer
its hydride and proton equivalent to a ketone. This was
confirmed by the lack of reaction between complex 2 and a
stoichiometric amount of acetophenone in THF-d8 at 25 or
50 °C. It was expected that the metal hydride and a protic
amine grouping would be suitable to effect bifunctional
catalysis of polar bonds.2,6,8,9 The use of an N-heterocyclic
carbene as the ligand in this cationic complex clearly shows
(47) Chaplin, A. B.; Dyson, P. J. Organometallics 2007, 26, 4357–
4360.

1244 Organometallics, Vol. 30, No. 5, 2011
O et al.
Figure 7. Reaction profile showing the hydrogenation of trans-
4-phenyl-but-3-en-2-one catalyzed by complex 1 in basic THF at
25 bar of H2 pressure and 50 °C: (blue circles) trans-4-phenylbut-
3-en-2-ol; (red squares) 4-phenylbutan-2-ol; (green triangles)
4-phenylbutan-2-one. The C/B/S ratio was 1/8/200.
Scheme 4. Hydrogenation of trans-4-Phenyl-but-3-en-2-one
Catalyzed by Complex 1
hydrogenation of the conjugated olefin could also occur by a
1,4-addition reaction.48 A ligand in the active ruthenium
species might also rearrange or dissociate to allow the
coordination of the olefin and its reduction to occur. This
was tested using trans-4-phenyl-but-3-en-2-one as the sub-
strate (Scheme 4). Under similar catalytic conditions (C/B/S
= 1/8/200, 25 bar of H2 at 50 °C), complex 1 catalyzed the
reduction of the polar bond and the olefin, giving the
products trans-4-phenyl-but-3-en-2-ol (33% conversion),
4-phenylbutan-2-one (10% conversion), and the fully hydro-
genated product 4-phenylbutan-2-ol (24% conversion) in 3 h
(Scheme 4 and Figure 7).
The Favored Inner-Sphere Mechanism. Given the experi-
mental data that are not supportive of the outer-sphere
mechanism, an inner-sphere mechanism is proposed.9,49
The ketone must coordinate to the metal center to allow
the hydride migration from the metal to the carbonyl group.
To effect coordination of the ketone, the primary amine
group can decoordinate from the metal center, or the arene
ligand on the ruthenium center can slip to η4 coordination
(Figure 8). Of note, a seven-membered ring is formed with
the chelating primary amine-NHC ligand, including the
nitrogen and carbene donor and the metal center. Facile
coordination and recoordination of the tethered group can
occur under catalytic conditions. This has been proposed for
Figure 6. Reaction profiles showing the hydrogenation of acet-
ophenone catalyzed by complex 1 (blue squares) and complex 2
(red circles) in (a) basic THF at 25 bar of H2 pressure and 50 °C
and (b) 2-propanol at 75 °C. The C/B/S ratio was 1/8/200. The
transfer hydrogenation of acetophenone catalyzed by complex 1
in basic 2-propanol was reported in ref 38.
that there is no activity in this case. Stradiotto and co-
workers have recently reported a highly active zwitterionic
ruthenium catalyst containing a phosphine-amine ligand
that catalyzes the transfer hydrogenation of ketones in
refluxing 2-propanol. On the other hand, the hydride com-
plex of the ruthenium catalyst showed no activity toward
ketone reduction, regardless of whether an excess amount of
base was added. Such a hydride complex, however, has no
protic N-H functionality on its ligand.22 The cationic
charge seems to reduce the nucleophilicity of the hydride
ligand so that there is diminished catalytic activity.
If the outer sphere mechanism operates in the H2-hydro-
genation of ketone catalyzed by complex 1 in basic solvents,
the hydrogenation of an R,β-unsaturated ketone would most
likely afford the reduction of the carbonyl group,9 although
(48) (a) Ito, M.; Kitahara, S.; Ikariya, T. J. Am. Chem. Soc. 2005, 127,
6172–6173. Michael addition reactions catalyzed by bifunctional cata-
lysts can also occur via 1,4-addition without prior coordination of the
conjugated olefin to the metal center; see: (b) Ikariya, T.; Gridnev, I. D.
Chem. Rec. 2009, 9, 106–123.
(49) Daley, C. J. A.; Bergens, S. H. J. Am. Chem. Soc. 2002, 124,
3680–3691.

Article
Organometallics, Vol. 30, No. 5, 2011 1245
chelates to the metal, will form a seven-membered ring as
observed in complex 1. The in situ generation of the silver(I)
carbene complex derived from 3 and subsequent transmeta-
lation of the NHC ligand to ruthenium(II),55 followed by
counteranion metathesis, gave a compound formulated on
the basis of elemental analysis as [Ru(p-cymene)Cl-
(CH3NC3H2N(CH2)3N(CH3)2)]PF6 1.5 DMSO (4). The
3
use of DMSO as cosolvent for the reaction is unavoidable,
as ligand 3 has limited solubility in most organic solvents
except alcohols. Attempts to isolate the intermediate silver(I)
carbene complex containing the corresponding imidazolyli-
dene ligand failed, as this quickly decomposed.
Figure 8. Possible reaction intermediates for the inner-sphere
mechanism involving hydride migration to the coordinated
ketone substrate: (left) decoordination of the chelating amine
group from the NHC ligand; (right) ring slippage of the arene
ring.
The NMR spectra of 4 in CD2Cl2 at 253 K indicate that
compound 4 is a mixture of two complexes, one with a
chelating NHC ligand and another with a nonchelating
NHC and coordinated DMSO ligand. Infrared spectroscopy
usually provides a means of distinguishing between O- and
S-bound forms of the dimethyl sulfoxide ligand56 but did not
prove helpful because of the complex spectrum in the finger-
print region (see the Supporting Information, Figure S7). The
experimental procedures and the characterization of the
compounds are given in the Supporting Information.
Complex 4 is an effective catalyst for the H2-hydrogena-
tion of acetophenone with an activity comparable to that of
complex 1 under similar reaction conditions (C/B/S = 1/10/
240, 25 bar of H2 at 50 °C) in basic THF with KOtBu as a
base. A conversion of 76% to 1-phenylethanol was achieved
within 2 h of reaction. Like complex 1, no induction period
was observed (see the Supporting Information, Figure S8).
Therefore, the protons of the amino group in complex 1 are
not needed for catalysis. This is another example showing
that the “NH effect” is not always operational in catalysts for
polar bond hydrogenation.21,22
Theoretical Considerations. The catalytic cycles involved in
both outer- (Scheme 2) and inner-sphere mechanisms (Figure 8
and below) were investigated by using density functional
theory (DFT) methods. A MPW1PW91 functional was cho-
sen, as this gave better predictions of energy barriers and
transition states.57 For computational ease, the p-cymene
ligand was simplified to a benzene ligand. The alkoxide base,
1-phenylethanol, and acetophenone were simplified to 2-prop-
oxide, 2-propanol, and acetone, respectively. The cations of
the base and the counteranions of the catalytically active
species were omitted throughout the calculations.
the transfer hydrogenation of ketones using a complex of
type Ru(η6-arene)(κ2-N-N)Cl (N-N = 2-hydroxylphenylbis-
(pyrazol-1-yl)methane and other derivatives) under base-
free conditions. The authors proposed that a nitrogen donor
of a pyrazolyl group decoordinates from the metal center
during catalysis and acts as a base in the deprotonation of the
coordinated alcohol.50 Elsevier and co-workers recently
reported the use of N-heterocyclic carbene complexes with
a tethered tertiary-amine group of palladium(0), [Pd(NHC-
amine)(alkene)], in the transfer hydrogenation of alkynes to
(Z)-alkenes under base-free conditions. Experimental find-
ings suggest the tethered amine group dissociates from the
metal center and acts as a base in catalysis.51 Chu and co-
workers reported the use of a half-sandwich complex of
ruthenium containing a tethered tertiary amine group on
the cyclopentadienyl ligand, which acts as a base to assist the
heterolytic cleavage of dihydrogen.52 Nolan and co-workers
have recently reported the mechanism of racemization of
chiral alcohols catalyzed by ruthenium(II) systems containing
N-heterocyclic carbene ligands. This was suggested to under-
go an inner-sphere mechanism which involved the dissocia-
tion of a coordinated carbonyl ligand. An Ru-O(alkoxide)
bond was formed in the presence of alkoxide, and this is
responsible for the formation of a ruthenium hydride inter-
mediate in the racemization of chiral alcohol. This is further
supported by NMR studies and DFT computations.53 The
feasibility of the amino group dissociation has been explored
using computational methods (vide infra). Of note, it is less
likely for the primary amine group to act as a base in
catalysis, as complex 1 must be activated by an alkoxide
base to become active.
The reaction coordinate diagram for the outer-sphere
mechanism is given in Figure 9. The ruthenium-amido
complex A1 reacts with dihydrogen to give the η2-dihydrogen
complex B. The amido nitrogen of A1 is highly charged (APT
charge -0.37 ESU; Table S8 in the Supporting Information)
In order to probe the importance of the primary amine
group tethered to the N-heterocyclic carbene ligand in com-
plex 1, we attempted to make and test the N,N-dimethyla-
mine analogue. Methylation of the primary amine group of
complex 1 was not successful. We did succeed in preparing a
closely related compound by reaction of the imidazolium salt
354 with silver(I) oxide and [Ru(p-cymene)Cl2]2 in one pot
(Scheme 5). Note that the carbene ligand derived from 3, if it
˚
and has a short Ru-N bond (1.89 A, Figure 10). The sum of
the angles around nitrogen is 358° in the optimized structure.
Therefore, the nitrogen is sp2 hybridized and the Ru-N has a
double-bond character. The Ru-N bond is longer in B (2.09
˚
A) and the double-bond character is lost, which allows
dihydrogen coordination to the coordinatively unsaturated
complex. The energy barrier at 298 K and 1 atm for the
dihydrogen addition to A1 on going from structure B0, which
has the dihydrogen molecule outside the coordination
(50) Carrion, M. C.; Sepulveda, F.; Jalon, F. A.; Manzano, B. R.;
Rodriguez, A. M. Organometallics 2009, 28, 3822–3833.
(51) (a) Warsink, S.; Hauwert, P.; Siegler, M. A.; Spek, A. L.; Elsevier,
C. J. Appl. Organomet. Chem. 2009, 23, 225–228. (b) Hauwert, P.;
Boerleider, R.; Warsink, S.; Weigand, J. J.; Elsevier, C. J. J. Am. Chem.
Soc. 2010, 132, 16900–16910.
(52) Chu, H. S.; Lau, C. P.; Wong, K. Y.; Wong, W. T. Organome-
tallics 1998, 17, 2768–2777.
(53) Bosson, J.; Poater, A.; Cavallo, L.; Nolan, S. P. J. Am. Chem.
Soc. 2010, 132, 13146–13149.
(54) Jimenez, M. V.; Perez-Torrente, J. J.; Bartolome, M. I.; Gierz,
V.; Lahoz, F. J.; Oro, L. A. Organometallics 2008, 27, 224–234.
(55) Warsink, S.; de Boer, S. Y.; Jongens, L. M.; Fu, C. F.; Liu, S. T.;
Chen, J. T.; Lutz, M.; Spek, A. L.; Elsevier, C. J. Dalton Trans. 2009,
7080–7086.
(56) James, B. R.; Morris, R. H. Can. J. Chem. 1980, 58, 399–408.
(57) Lynch, B. J.; Truhlar, D. G. J. Phys. Chem. A 2001, 105, 2936–
2941.

1246 Organometallics, Vol. 30, No. 5, 2011
O et al.
Figure 9. Reaction coordinate diagram for an outer-sphere mechanism for the H2-hydrogenation of acetone starting from A1 and
moving to the right and the enolate formation starting from A1 and moving to the left. The free energies are reported relative to G in
kcal/mol. All computed structures were optimized at 1 atm and 298 K in the gas phase.
Scheme 5. Synthesis of Complex 4 by in Situ Generation of the Silver(I) Carbene Complex and Subsequent Transmetalation of the NHC
Ligand to Ruthenium(II)
0
sphere, to the transition state TSB ,B is 18.0 kcal/mol uphill.
The entropy change (ΔSq) for dihydrogen coordination
is -25.9 cal/(mol K). The dihydrogen is weakly activated
in the transition state and also in structure B, as the H-H
Ru(H)2(PH3)2(en), have charges of -0.20 and -0.15 ESU.18
From structure C, acetone is hydrogen-bonded to the complex,
forming D, and this goes þ33.5 kcal/mol uphill from C plus
acetone at 298 K and 1 atm, leading to the transition state TSD,E
,
58
˚
distances are short (0.76 and 0.83 A, respectively). Struc-
ture B leads to the transition state TSB,C, in which the H-H
distance increases to 0.88 A and the negative charge on nitro-
gen increases from -0.37 to -0.40 ESU. Note that the
energy difference between B and TSB,C is only 0.2 kcal/mol.
The heterolytic splitting of H2 by A1 leads to C, which
is -39.5 kcal/mol downhill from TSB,C. The Ru-H distance is
1.58 A, which is similar within the ESD to the value of the
Ru-H distance (1.79(7) A) in the crystal structure of complex
2. The single-bond character of Ru-N is retained (2.16 A),
and the charge on nitrogen is further reduced to -0.24
ESU. The charges on ruthenium and hydride are -0.72
and -0.042 ESU, respectively. The smaller charge on the
hydride may explain the poor activity of complex 2 toward
reaction with acetophenone. The hydrides of the model
complex of a ketone hydrogenation catalyst, (OC-6-22)-
where the concerted transfer of dihydrogen to the ketone
from Ru-H and Ru-NH2 occurs via a six-membered-ring
transition state.9-11 The Ru-H bond is elongated (to 1.84 A),
˚
˚
˚
as is the N-H bond (from 1.02 to 1.30 A). In addition, the
C-H and C-O bonds of the alcohol are formed simulta-
˚
neously, with bond distances of 1.27 and 1.33 A, respectively.
The charge of ruthenium is reduced (-0.32 ESU) and the
hydride is more negatively charged (-0.43 ESU). This leads to
the ruthenium-amido-isopropyl alcohol adduct E, held by
weak electrostatic interactions. The elimination of the alcohol
product results in the regeneration of the amido complex A1,
which is -15.6 kcal/mol downhill from TSD,E. The alcohol
could also coordinate to the amido complex, forming F.
Subsequent deprotonation of the alcohol by the amide ligand
causes the formation of the ruthenium-alkoxide complex G.
This has a basic alkoxide ligand (APT charge on O -0.82
ESU) and is more thermodynamically stable than A1 plus 2-
propanol by 2.4 kcal/mol. Of note, the Ru-C(carbene) bond
˚
˚
˚
(58) Morris, R. H. Coord. Chem. Rev. 2008, 252, 2381–2394.

Article
Organometallics, Vol. 30, No. 5, 2011 1247
Figure 10. Selected computed structures A-C and G for the outer-sphere mechanism in the H2-hydrogenation of acetone and the
transition state structures for the heterolytic splitting of H2 (TSB,C) and for the concerted transfer of a hydride/proton pair to the ketone
˚
(TSD,E). The bond lengths (A) are given in the structures. The APT charges (ESU) of these structures are given in Table S8 in the
Supporting Information.
˚
distances vary from 2.03 to 2.08 A for all the computed
structures, and they are comparable to those of complexes 1
and 2 in the crystal structures (Figure 10).38
Starting from the alkoxide complex G, the decoordination
of the amine group is an uphill process by þ15.5 kcal/mol at
298 K and 1 atm (Figure 11). This process is entropically
favored (ΔS = 16.6 cal/(Mol K)). This afterward provides
the coordinatively unsaturated complex H, which has a short
The enolate complex A3 is formed from A1 via coordina-
tion of acetone, giving A2, and this leads to the transition
state TSA2,A3, where deprotonation of the methyl group
occurs. This is þ17.9 kcal/mol uphill from A1 plus acetone.
The product A3 is thermodynamically less stable than A1 plus
acetone by 7.3 kcal/mol (Figure 9). The computations thus
accurately reflect comparable energy barriers for the H2
addition and enolate formation from the amido complex
A1 (18.0 versus 17.9 kcal/mol at 298 K and 1 atm).
The concerted transfer of Hþ/H- to acetone has a much
higher barrier of þ33.5 kcal/mol in comparison to the
aforementioned steps, which contradicts the experimental
findings. The calculation also predicts that the ruthe-
nium-hydride complex C (or complex 2) is the resting state
of the catalytic cycle. This is more stable than the ruthe-
nium-alkoxide and ruthenium-amido complexes by 15.5
and 22.4 kcal/mol.
An alternative proposal is the inner-sphere mechanism,
which involves the decoordination of the amine group from
the NHC ligand or ring slippage of the coordinated arene
ring.10 The computed structures of the latter have higher
energies (>25 kcal/mol with respect to structure G). The
ring slipped structures were also higher in energy than
those structures computed for the mechanism involving
the decoordination of the amine group (see below). Thus,
they are not considered as the possible intermediates (the
structures and their energies are given in the Supporting
Information).
˚
Ru-O distance compared to G (1.90 and 2.06 A, respec-
tively, Figures 10 and 12). The charges on the oxygens of the
alkoxide are comparable in H and G (-0.82 and -0.77,
respectively; Tables S8 and S9 in the Supporting In-
formation). The variation in the Ru-C(carbene) bond dis-
tance is similar to that of the outer-sphere mechanism for all
of the computed structures (Figure 12). In addition, the
˚
Ru-N(amine) distances are more than 5.0 A for all of the
computed structures. For all, the dihedral angle of the phenyl
and the imidazolidene rings range from 83.3° for H and 95.0°
for K. For comparison, the dihedral angles for the crystal
structures of complexes 1 and 2 are 54.6 and 54.3°.38
Coordination of molecular hydrogen in a side-on fashion
to H then occurs, giving I. Both QST3 and QST2 searches
were performed, but these failed to locate a transition state
for dihydrogen addition to H. The dihydrogen is again
weakly activated in I with a short H-H distance of
58
˚
0.85 A. This process works against entropy (ΔS = -16.0
cal/(mol K)). Heterolytic splitting of η2-H2 takes place
through transition state TSI,J. The H-H distance increases
˚
to 0.92 A, and the charge on oxygen increases from -0.68 to
-0.73 ESU. The energy barrier for the heterolytic splitting of
H2 is þ33.0 kcal/mol uphill from G plus dihydrogen or þ17.5
kcal/mol from H plus dihydrogen at 298 K and 1 atm. The
energy difference between I and TSI,J is 3.7 kcal/mol,
indicative of an early transition state. Again, the heterolytic

1248 Organometallics, Vol. 30, No. 5, 2011
O et al.
Figure 11. Reaction coordinate diagram for the inner-sphere mechanism in the H2-hydrogenation of acetone starting from H and
moving to the right. The amino group is decoordinated and is not involved in hydrogen bonding. Moving to the left from H leads to the
unstable enolate complex M. The free energies are reported relative to G in kcal/mol. All computed structures were optimized at 1 atm
and 298 K in the gas phase.
Figure 12. Selected computed structures H-K for the inner-sphere mechanism involving decoordination of the amine group of the
NHC ligand in the H2-hydrogenation of ketone and the transition state structures for the heterolytic splitting for H2 (TSI,J) and for the
˚
hydride attack on the coordinated ketone (TSK,H). The bond lengths (A) are given in the structures. The APT charges (ESU) of these
structures are given in Table S9 in the Supporting Information.

Article
Organometallics, Vol. 30, No. 5, 2011 1249
splitting of H2 is entropically demanding (ΔSq = -19.9
cal/(mol K)).
For the inner-sphere mechanism, decoordination of the
amine group of the NHC ligand takes place from the
2-propoxide complex G to allow the β-hydride elimination
to take place, forming H. This proceeds through the transi-
tion state TSK,H, leading to the hydride complex K
(Figure 11). The β-hydride elimination of the alkoxide is
uphill by 6.6 kcal/mol from H. The hydride ligand from K
can then attack the carbonyl group of the substrate; this is the
same step as described in the inner-sphere H2 hydrogenation
mechanism, with the same energy barrier of þ15.6 kcal/mol
uphill (Figure 11). Overall, this has lower energy barriers
compared with the outer-sphere mechanism and seems more
likely to occur during catalysis.
The ruthenium-hydride complex containing a coordi-
nated alcohol is then formed from TSI,J, leading to J, which
is -27.7 kcal/mol downhill. The structure has a short Ru-H
˚
bond (1.57 A). The charges on the ruthenium and the
hydride are -0.52 and -0.039 ESU. Decoordination of
2-propanol and recoordination of acetone forms K, and a
further energy of þ15.6 kcal/mol uphill is required, leading
to the transition state TSK,H. The hydride ligand on the
ruthenium complex attacks the coordinated acetone via a
four-membered-ring transition state.9,53 The Ru-H bond
˚
elongates to 1.65 A and the Ru-O bond shortens (from 2.14
˚
to 2.09 A), due to an increased charge on oxygen
Role of Complex 2 in Catalysis. Complex 2 is not likely to
be an intermediate in the H2-hydrogenation reaction, as
attempts to prepare this from complex 1 in basic THF under
H2 failed. In addition, it does not react with acetophenone
(see below). This, however, may be the resting state of the
catalyst in the transfer hydrogenation, as it can be prepared
under similar catalytic conditions. According to the calcula-
tions, the amine group of structure K can recoordinate,
which displaces the coordinated acetone to give structure
C, which is analogous to complex 2. This is -33.1 kcal/mol
downhill from the transition state TSH,K. The fact that a base
is needed to activate complex 2 in both H2-hydrogenation
and transfer hydrogenation reactions is difficult to explain.
In fact, a stoichiometric reaction of complex 2 and KOtBu in
THF-d8 at 50 °C under argon produces a reaction mixture
that lacks a hydride signal in the 1H NMR spectrum.
Subsequent addition of a stoichiometric amount of aceto-
phenone to this reaction mixture did not lead to the forma-
tion of 1-phenylethanol or 1-phenylethoxide. The species
that is formed from the reaction of complex 2 and an
alkoxide base may be responsible for the activation of H2.
Synthesis of an Osmium(II) Complex with an NHC and a
Primary Amine Donor. For comparison, the osmium(II)
complex 6 was synthesized by transmetalation38 of the
(from -0.57 to -0.60 ESU; Table S9 in the Supporting
˚
Information). Likewise, the C-H (1.58 A) and C-O (1.30
A) bonds of the alcohol are formed simultaneously. The
˚
charge of ruthenium is reduced (-0.70 ESU), and the
hydride is more highly charged (-0.10 ESU). This leads to
structure H and completes the cycle. We are aware of the
presence of the geometric isomers of these structures: in
particular, in which the amine proton forms a hydrogen
bond with the oxygen on the alkoxide or alcohol groups.
These have higher energies compared to those described
above. A QST3 search in the transition state of the hetero-
lytic splitting of dihydrogen failed to locate any imaginary
frequency. The energies and the geometries of such struc-
tures are given in the Supporting Information.
The enolate complex M is formed from H upon coordina-
tion of acetone and subsequent deprotonation of the coordi-
nated alkoxide base. No transition state is located between
structures L and M. This is þ24.8 kcal/mol uphill starting
from H. All this reflects different energy barriers for the
H2-splitting and enolate formation from H (17.5 versus 24.8
kcal/mol at 298 K and 1 atm). The hydride migration step has
a lower energy barrier of 19.6 kcal/mol, which coincides with
the experimental findings. Of note, the energy barrier for
H2-splitting starting from the alkoxide G plus dihydrogen is
even higher (þ33.0 kcal/mol), and the computations predict
that this is the resting state of the catalytic cycle. This predicts
that the heterolytic splitting of dihydrogen is truly the rate-
determining step. The formation of an enolate complex, on
the other hand, should occur via the amido complex, as the
energy barrier is 24.8 kcal/mol uphill starting from G and
moving successively from A1 to A2, to the transition state
TSA2,A3, and then to A3 (see Figure 9). This has a much lower
energy barrier than that starting from H. Of note, the amido
complex could form initially from the alkoxide complex and
deprotonation of the amine proton by the internal alkoxide
base. Bergens and co-workers proposed a similar mechan-
ism for the complex RuH(OR)(diamine)((R)-binap).15,30
Possible Mechanism for Transfer Hydrogenation. The com-
putations also gave important information on the transfer
hydrogenation of ketones.38 The transfer hydrogenation might
proceed via the outer-sphere mechanism starting from the
amido complex A1, then to E, TSD,E, and D, and finally to C,
the ruthenium-hydride complex (Figure 9). The energy barrier
starting from A1 plus 2-propanol is þ15.6 kcal/mol uphill at
298 K and 1 atm. The concerted transfer of Hþ/H- proceeds in
a manner identical with that described for the H2-hydrogena-
tion, going from C to A1, with an energy barrier of þ33.5 kcal/
mol. This high barrier argues against this mechanism, and in
fact, the ruthenium-hydride complex 2 has very low activity
under transfer hydrogenation conditions in the absence of base.
chelating ligand (C-NH2) from complex 5 to [Os(p-
59,60
cymene)Cl2]2
in refluxing acetonitrile solution (Scheme
6). An air-stable yellow solid was isolated in 93% yield.
The X-ray crystal structure of the complex showed a piano-
stool geometry about the metal center with an η6-cymene
ligand and the chloro and the chelating C-NH2 ligands
˚
(Figure 13 and Table 5). The Os-Ccarbene bond (2.07(1) A) is
comparable to those in compounds containing an NHC
ligand reported in the literature.60-62 A diagnostic 13C
NMR feature of complex 6 in solution is the Os-Ccarbene
resonance at 159.4 ppm, which lies in the expected range for
analogous complexes.60-62
The catalytic activity of complex 6 in the hydrogenation of
acetophenone was tested using reaction conditions similar to
those for the ruthenium(II) complex 1. The complex, when
activated by base (KOtBu) in THF, gave 23% conversion to
1-phenylethanol using 25 bar of H2 as the hydrogen source
(C/B/S = 1/8/200) in 3 h at 50 °C. It gave 99% conversion in
(59) Cabeza, J. A.; Maitlis, P. M. Dalton Trans. 1985, 573–578.
(60) Castarlenas, R.; Esteruelas, M. A.; Onate, E. Organometallics
2005, 24, 4343–4346.
(61) (a) Eguillor, B.; Esteruelas, M. A.; Olivan, M.; Puerta, M.
Organometallics 2008, 27, 445–450. (b) Castarlenas, R.; Esteruelas,
M. A.; Lalrempuia, R.; Olivan, M.; Onate, E. Organometallics 2008,
27, 795–798.
(62) Castarlenas, R.; Esteruelas, M. A.; Onate, E. Organometallics
2008, 27, 3240–3247.

1250 Organometallics, Vol. 30, No. 5, 2011
O et al.
Scheme 6. Synthesis of Osmium(II) Complex 6 by Transmeta-
lation of the C-NH2 Ligand from 5
20 h at 50 °C if 2-propanol was used as solvent. On the other
hand, the activated complex catalyzed the transfer hydro-
genation of acetophenone to 40% conversion to the product
alcohol in 18 h at 75 °C (C/B/S = 1/8/200). This has lower
activity in the transfer hydrogenation of acetophenone in
2-propanol solution compared to that of complex 138 and
analogous systems reported in the literature.62,63 The poor
activity of complex 6 when activated in comparison to 1
might be attributed to the oxophilic nature of the metal
center to form stable osmium(II)-alkoxide complexes. Of
note, the osmium(II) complex showed no activity when it is
not activated by potassium tert-butoxide. Similar DFT
calculations were also performed with the osmium(II) com-
plex using the inner-sphere mechanism. The osmium(II)
dihydrogen complex with a decoordinated amine group
similar to I is 30.7 kcal/mol uphill from the alkoxide com-
plex analogous to G plus dihydrogen (see the Supporting
Information). The related complex [Os(p-cymene)(NHC)-
(OH)]OTf (NHC = bis(2,6-diisopropylphenyl)imidazolidene,
IPr), was believed to undergo an inner-sphere mechanism for
the transfer hydrogenation of aldehydes in 2-propanol solution,
whereas the hydroxyl ligand acts as an internal base in the
formation of an Os-O(alkoxide) bond.62
Figure 13. ORTEP diagram of 6 ([Os(p-cymene)(m-CH2NH2)-
Cl]PF6) depicted with thermal ellipsoids at the 30% probability
level. The counteranion and most of the hydrogens have been
˚
omitted for clarity. Selected bond distances (A) and bond angles
(deg): Os(1)-C(1), 2.07(1); Os(1)-N(3), 2.137(9); Os(1)-Cl(1),
2.422(3); Os(1)-C(12), 2.163(7); Os(1)-C(15), 2.201(6); C-
(1)-Os(1)-N(3), 91.1(4); C(1)-Os(1)-Cl(1), 87.5(3); Cl-
(1)-Os(1)-N(3), 80.0(3).
to a ruthenium-alkoxide complex. The alkoxide ligand
labilizes the cis amine ligand.64 Decoordination of the amine
group of the NHC ligand provides a vacant site for the
coordination of dihydrogen (Scheme 7). Ring-opening reac-
tions are common for chelating ligands that form a seven-
membered ring with the metal. Catalysis then proceeds with
the heterolytic splitting of dihydrogen by the internal alk-
oxide base, and then hydride attacks the coordinated ketone.
In addition, the alkoxide complex can convert to the ruthe-
nium-amido complex which is responsible for the formation
of an enolate complex. The current study should assist in the
rational design of more robust and active hydrogenation
catalysts using N-heterocyclic carbenes as ligands.
Conclusion
In summary, the H2-hydrogenation of ketones catalyzed
by complex 1 was studied and the mechanism of action was
investigated by both experimental and theoretical means.
The ruthenium-hydride complex 2, which was thought to be
the crucial intermediate for bifunctional catalysis, was iso-
lated. This, however, showed no activity under catalytic
conditions unless when activated by a base. This hydride
complex is believed to be the resting state in the transfer
hydrogenation mechanism. The cationic charge on the metal
center is likely to decrease the nucleophilicity of the hydride
ligand. In fact, this is a rare example of a catalyst with an
M-H and N-H grouping that fails to undergo bifunctional
catalysis using the “NH effect”, which was originally pro-
posed by Noyori and co-workers.8 Kinetic studies including
the studies of isotope effects using D2 gas and acetophenone-
d3 support the theoretical predictions that an early transition
state in the heterolytic splitting of H2 is the rate-determining
step of the catalytic cycle. The outer-sphere mechanism
involving bifunctional catalysis of ketone reduction is dis-
favored according to experimental studies. Computational
studies also suggest a high energy barrier for the concerted
transfer of Hþ/H- to the ketone compared to dihydrogen
addition and subsequent heterolytic splitting.
Experimental Section
Synthesis. All of the preparations and manipulations, except
where otherwise stated, were carried out under a nitrogen or
argon atmosphere using standard Schlenk-line and glovebox
techniques. Dry and oxygen-free solvents were always used. The
syntheses of [1-(2-(aminomethyl)phenyl)-3-methylimidazol-2-
ylidene]chloro(η6-p-cymene)ruthenium(II) hexafluoropho-
sphate (1) and bis[1-(2-(aminomethyl)phenyl)-3-methylimida-
zol-2-ylidene]nickel(II) hexafluorophosphate (5) have been
59,60
reported previously.38 The synthesis of [Os(p-cymene)Cl2]2
was reported in the literature. All other reagents and solvents
were purchased from commercial sources and were used as
received. Deuterated solvents were purchased from Cambridge
An alternative to the outer-sphere bifunctional mechanism
is therefore proposed on the basis of experimental and
theoretical findings. First, complex 1, when activated, leads
(64) (a) Flood, T. C.; Lim, J. K.; Deming, M. A.; Keung, W.
Organometallics 2000, 19, 1166–1174. (b) Holland, A. W.; Bergman,
R. G. J. Am. Chem. Soc. 2002, 124, 14684–14695. (c) Clarkson, A. J.;
Buckingham, D. A.; Rogers, A. J.; Blackman, A. G.; Clark, C. R. Inorg.
Chem. 2000, 39, 4769–4775.
(63) Faller, J. W.; Lavoie, A. R. Org. Lett. 2001, 3, 3703–3706.

Article
Organometallics, Vol. 30, No. 5, 2011 1251
Scheme 7. Proposed Mechanism for the H2-Hydrogenation of Ketones Catalyzed by Complex 1 and Base
Isotope Laboratories and Sigma Aldrich and degassed and dried
over activated molecular sieves prior to use. NMR spectra were
recorded on a Varian 400 spectrometer operating at 400 MHz
(CD3CN, δ): -72.9 (d, JPF = 706 Hz). 13C{1H} NMR (CD3CN,
δ):185.3 (Ru-Ccarbene), 141.4 (CPh), 134.0 (CPh), 132.2 (CPh),
130.5 (CPh), 129.3 (CPh), 126.6 (CPh), 124.3 (Cimid.), 123.5
(Cimid.), 108.7 (CAr-p-cymene), 105.1 (CAr-p-cymene), 88.9 (CAr-p-cymene),
85.6 (CAr-p-cymene), 81.3 (CAr-p-cymene), 80.9 (CAr-p-cymene),
47.4 (CH2), 39.8 (CH3), 32.4 (CH of (CH3)2CH of p-cymene),
24.1 (CH3 of (CH3)2CH of p-cymene), 23.9 (CH3 of (CH3)2CH of
p-cymene), 19.1 (CH3 of p-cymene). MS (ESI, methanol/water;
m/z): 424.1 [M]þ. Attempts at elemental analyses failed to give an
acceptable carbon content, while hydrogen and nitrogen contents
are in the acceptable range. Typical results are as follows. Anal.
Calcd for C21H28F6N3PRu: C, 44.37; H, 4.96; N, 7.39. Found: C,
42.91; H, 4.80; N, 7.34.
for H, 100 MHz for 13C, and 376 MHz for 19F. The H and
13C{1H} NMR spectra were measured relative to partially
deuterated solvent peaks but are reported relative to tetra-
methylsilane (TMS). All 19F chemical shifts were measured
relative to trichlorofluoromethane as an external reference.
Elemental analyses were performed at the Department of
Chemistry, University of Toronto, on a Perkin-Elmer 2400
CHN elemental analyzer. Single-crystal X-ray diffraction data
were collected using a Nonius Kappa-CCD diffractometer with
1
1
˚
Mo KR radiation (λ = 0.710 73 A). The CCD data were
integrated and scaled using the Denzo-SMN package. The
structures were solved and refined using SHELXTL V6.1.
Refinement was by full-matrix least squares on F2 using all
data. Details are given in Table 5.
Synthesis of [1-(2-(Aminomethyl)phenyl)-3-methylimidazol-2-
ylidene]chloro(η6-p-cymene)osmium(II) Hexafluorophosphate
([Os(p-cymene)(m-CH2NH2)Cl]PF6, 6). A Schlenk flask was
charged with 5 (46 mg, 0.064 mmol) and the [Os(p-cymene)Cl2]2
dimer (50 mg, 0.063 mmol). Dry acetonitrile (8 mL) was added
to the reaction mixture, and it was refluxed under an argon
atmosphere for 2.5 h to give a green solution. The solvent was
then evacuated. The residue was extracted with THF (4 mL)
and filtered through a pad of Celite. The volume of solvent was
reduced (2 mL), and addition of diethyl ether (12 mL) to the
THF solution yielded a pale yellow precipitate which was
collected and dried in vacuo. Yield: 81 mg, 93%. Crystals
suitable for an X-ray diffraction study were obtained by slow
diffusion of diethyl ether into a saturated solution of 6 in a 1/1
acetonitrile/methanol mixture under a nitrogen atmosphere. 1H
NMR (CD2Cl2, δ): 7.70 (m, 3-CH of Ph, 1H), 7.59 (m, 4-CH and
5-CH of Ph, 2H), 7.41 (m, 6-CH of Ph, 1H), 7.24 (d, JHH = 1.95
Hz, 5-CH of imid, 1H), 7.22 (d, JHH = 1.95, 4-CH of imid, 1H),
5.88 (m, br, NH2, 2H), 5.58 (d, JHH = 5.44 Hz, 2-Ar-CH of p-
cymene, 1H), 5.51 (d, JHH = 5.51 Hz, 6-Ar-CH of p-cymene,
1H), 5.30 (d, JHH = 5.51 Hz, 5-Ar-CH of p-cymene, 1H), 4.98
(d, JHH = 5.44 Hz, 3-Ar-CH of p-cymene, 1H), 4.21 (m, CH2,
1H), 4.03 (s, CH3, 3H), 3.15 (dt, JHH = 2.54, 12.28 Hz, CH2,
1H), 2.48 (sept, JHH = 6.88 Hz, CH of (CH3)2CH of p-cymene,
1H), 1.77 (s, CH3 of p-cymene, 3H), 1.14 (d, JHH = 6.91 Hz,
CH3 of (CH3)2CH of p-cymene, 3H), 1.07 (d, JHH = 6.91 Hz,
CH3 of (CH3)2CH of p-cymene, 3H). 19F NMR (CD2Cl2,
δ): -72.0 (d, JPF = 712 Hz). 13C{1H} NMR (CD2Cl2, δ):
159.4 (Os-Ccarbene), 138.9 (CPh), 132.8 (CPh), 131.0 (CPh),
130.5 (CPh), 130.0 (CPh), 125.6 (Cimid.), 124.9 (CPh), 123.5
(Cimid.), 102.8 (CAr-p-cymene), 92.5 (CAr-p-cymene), 78.2 (CAr-p-
cymene), 76.9 (CAr-p-cymene), 75.6 (CAr-p-cymene), 73.1 (CAr-p-cymene),
Synthesis of [1-(2-(Aminomethyl)phenyl)-3-methylimidazol-2-
ylidene]hydrido(η6-p-cymene)ruthenium(II) Hexafluoropho-
sphate ([Ru(p-cymene)(m-CH2NH2)H]PF6, 2). A Schlenk flask
was charged with 1 (50 mg, 0.083 mmol) in 2-propanol solution
(14 mL). The solution was warmed to 50 °C under an argon
atmosphere. A solution of sodium 2-propoxide (20 mg, 0.24
mmol) in 2-propanol (6 mL) was added to this stirred solution
over the course of 0.5 h, whereupon the color of the reaction
mixture turned from yellow to deep red and then to brown. The
solution was stirred for a further 3 h. After the reaction had gone
to completion, the solvent was removed under vacuum. The
solid residue was extracted with tetrahydrofuran (THF, 4 mL)
and filtered through a pad of Celite. The addition of pentane (16
mL) to the THF solution yielded a brown precipitate, which was
collected and dried in vacuo. Yield: 27 mg, 57%. Crystals
suitable for an X-ray diffraction study were obtained as a
THF solvate by slow diffusion of pentane into a saturated
solution of 2 in tetrahydrofuran under a nitrogen atmo-
1
sphere. H NMR (CD3CN, δ): 7.56 (m, 3-CH and 4-CH of
Ph, 2H), 7.46 (m, 5-CH of Ph, 1H), 7.39 (d, JHH = 7.34 Hz,
6-CH of Ph, 1H), 7.33 (d, JHH = 1.72 Hz, 5-CH of imid, 1H),
7.21 (d, JHH = 1.72, 4-CH of imid, 1H), 5.06 (d, JHH = 5.65
Hz, 2-Ar-CH of p-cymene, 1H), 5.02 (d, JHH = 5.92 Hz,
6-Ar-CH of p-cymene, 1H), 4.82 (d, JHH = 5.92 Hz, 5-Ar-
CH of p-cymene, 1H), 4.40 (d, JHH = 5.65 Hz, 3-Ar-CH of
p-cymene, 1H), 3.68 (m, CH2, 1H), 3.61 (s, CH3, 3H), 3.40 (m,
br, NH2, 2H), 2.64 (dt, JHH = 3.66, 12.17 Hz, CH2, 1H), 2.40
(sept, JHH = 6.80 Hz, CH of (CH3)2CH of p-cymene, 1H),
1.52 (s, CH3 of p-cymene, 3H), 1.21 (d, JHH = 6.80 Hz, CH3
of (CH3)2CH of p-cymene, 3H), 1.14 (d, JHH = 6.80 Hz, CH3
of (CH3)2CH of p-cymene, 3H), -7.82 (s, Ru-H). 19F NMR
47.7 (CH ), 39.8 (CH3), 31.0 (CH of (CH3)2CH of p-cymene),
23.9 (CH3 of p-cymene), 21.2, (CH3 of (CH3)2CH of p-cymene),
18.7 (CH3 of (CH3)2CH of p-cymene). MS (ESI, methanol/
2

1252 Organometallics, Vol. 30, No. 5, 2011
O et al.
water; m/z): 548.2 [M]þ. HRMS (ESI, methanol/water; m/z):
calcd for C21H27N3ClOsþ [M]þ 548.1502, found 548.1463. Anal.
Calcd for C21H27ClF6N3OsP: C, 36.44; H, 3.93; N, 6.07. Found:
C, 37.04; H, 3.44; N, 5.51.
potassium tert-butoxide are given in Table 3 and in the Support-
ing Information (Tables S1-S3). All of the conversions were
reported as an average of at least two GC runs. The reported
conversions were reproducible.
Catalysis. Oxygen-free tetrahydrofuran (THF) used for all of
the catalytic runs was stirred over sodium for 2-3 days under
argon and freshly distilled from sodium benzophenone ketyl
prior to use. Acetophenone was vacuum-distilled over phos-
phorus pentoxide (P2O5) and stored under nitrogen prior to use.
All of the other substrates were vacuum-distilled, dried over
activated molecular sieves, and stored under nitrogen prior to
use. All of the hydrogenation reactions were performed at
constant pressure using a stainless steel 50 mL Parr hydrogena-
tion reactor. The temperature was maintained at 50 °C using a
constant-temperature water bath. The reactor was flushed
several times with hydrogen gas at 2-4 bar prior to the addition
of catalyst/substrate and base solutions.
In a typical run (Table 2, entry 2), the catalyst 1 (3 mg, 5.0
μmol), benzophenone (181 mg, 0.99 mmol), and potassium tert-
butoxide (5 mg, 0.044 mmol) were dissolved in THF (4 and 2
mL, respectively) under a nitrogen atmosphere. The catalyst/
substrate and base solutions were taken up by means of two
separate syringes and needles in a glovebox. The needles were
stoppered, and the syringes were taken to the reactor. The
solutions were then injected into the reactor against a flow of
hydrogen gas. The hydrogen gas was adjusted to the desired
pressure. Small aliquots of the reaction mixture were quickly
withdrawn with a syringe and needle under a flow of hydrogen at
timed intervals by venting the Parr reactor at reduced pressure.
Alternatively, small aliquots of the reaction mixture were
sampled from a stainless steel sampling dip tube attached to a
modified Parr reactor. The dip tube was 30 cm in length with an
inner diameter of 0.01 in., and a swing valve was attached to the
end of the sampling tube. Other technical details were previously
reported.18 Two small aliquots of samples were thereby with-
drawn quickly at timed intervals by opening the swing valve, and
the first two aliquots were discarded. All samples for gas
chromatography (GC) analyses were diluted to a total volume
of approximately 0.50 mL using oxygenated THF.
Kinetic Isotope Effect Studies. D2 (99.8% D) gas and acet-
ophenone-β,β,β-d3 (99% D) were purchased from Cambridge
Isotope Laboratories. Acetophenone-d3 was vacuum-distilled,
dried over activated molecular sieves, and stored under nitrogen
prior to use. D2 gas was used as received without further
purification. For experiments using deuterium gas, the modified
Parr reactor with a stainless steel sampling dip tube was used.
The reactor was degassed by evacuation for 10 min, followed by
refilling with 2-4 bar of deuterium gas. This was repeated three
times. The reaction mixtures were then injected into the reactor
against a flow of deuterium gas, and the gas pressure was
adjusted to 8 bar. The tabulated results of the initial rates with
varying concentrations of acetophenone are given in the Sup-
porting Information (Tables S4-S6). Experiments using acet-
ophenone-d3 were conducted similarly to the kinetic studies
described above.
Computational Details. All density functional theory (DFT)
calculations were performed using the Gaussian03 package65
with the restricted hybrid mPW1PW91 functional.66 Ruthenium
was treated with the SDD67 basis set to include relativistic effects
and an effective core potential, and other heteroatoms were
treated with the double-ζ basis set 6-31þþG**, which includes
diffuse functionals and additional p orbitals on hydrogen as well
as additional d orbitals on carbon, nitrogen, and oxygen. All
geometry optimizations were conducted in the gas phase; these
were followed by full vibrational analyses at 1 atm and 298 K.
Reported energies were corrected from zero-point energies but
uncorrected for translations and rotations. The QST3 method
was used to locate transition states. All transition states reported
were found to have a single imaginary frequency. The calculated
atomic polar tensor (APT) charges were used to reflect more
accurately the charge distribution on each atom under the
derivative of dipole moments with respect to an applied external
electric field.68
Acknowledgment. The NSERC of Canada is thanked
for a Discovery Grant to R.H.M. NSERC Canada and
the Ministry of Education of Ontario are thanked for
graduate scholarships to W.W.N.O.
A Perkin-Elmer Clarus 400 chromatograph equipped with a
chiral column (CP chirasil-Dex CB 25 m ꢀ 2.5 mm) with an
autosampling capability was used for GC analyses. Hydrogen
was used as a mobile phase at a column pressure of 5 psi with a
split flow rate of 50 mL/min. The injector temperature was 250
°C, and the FID temperature was 275 °C. The oven temperature
and the retention times (tR, tp, /min) for all the substrates and
alcohol products are given in the Supporting Information
(Table S7). All of the conversions were reported as an average
of two GC runs. The reported conversions were reproducible.
Kinetics. A standard solution of the catalyst 1 (1.66 mM) was
prepared by dissolving the complex (25 mg) in THF (25 mL).
Reaction mixtures were then prepared in THF by dispensing the
required amount of the catalyst and weighed amounts of
acetophenone, phenylethanol, and potassium tert-butoxide (or
a standard solution of potassium tert-butoxide (17.8 mM) in
THF for base-dependent studies) and diluted to a total volume
of 6 mL. The hydrogenation reaction and sampling techniques
were described as above in order to obtain at least five data
points including the origin (time 0 min) at below 40% conversion
to the product alcohol. The tabulated results of initial rates with
varying concentrations of catalyst, acetophenone, hydrogen, and
Supporting Information Available: CIF files giving X-ray
structural data for complexes 2 and 6, text, tables, and figures
giving details for catalysis and tabulated results for kinetic data,
experimental procedures, and characterization data of complex
4, Cartesian coordinates, energies for all of the computed
structures, and the complete citation for reference 65, and
AVI files giving animations for the computed transition states.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(65) Frisch, M., et al. Gaussian 03, Revision C.02; Gaussian Inc.,
Wallingford, CT, 2004.
(66) (a) Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664–675. (b)
Burke, K.; Perdew, J. P.; Wang, Y. Electronic Density Functional
Theory: Recent Progress and New Directions; Plenum: New York, 1997.
(67) Leininger, T.; Nicklass, A.; Stoll, H.; Dolg, M.; Schwerdtfeger,
P. J. Chem. Phys. 1996, 105, 1052–1059.
(68) Cioslowski, J. J. Am. Chem. Soc. 1989, 111, 8333–8336.