Experimental and computational investigation of C-N bond activation in ruthenium N-heterocyclic carbene complexes

Article abstract of DOI:10.1021/ja109702h

A combination of experimental studies and density functional theory calculations is used to study C-N bond activation in a series of ruthenium N-alkyl-substituted heterocyclic carbene (NHC) complexes. These show that prior C-H activation of the NHC ligand

Full text of DOI:10.1021/ja109702h

Published on Web 12/03/2010
Experimental and Computational Investigation of C-N Bond
Activation in Ruthenium N-Heterocyclic Carbene Complexes
L. Jonas L. Ha¨ller,† Michael J. Page,‡ Stefan Erhardt,†,§ Stuart A. Macgregor,*,†
Mary F. Mahon,‡ M. Abu Naser,† Andrea Ve´lez,† and Michael K. Whittlesey*,‡
School of Engineering and Physical Sciences, Heriot-Watt UniVersity, Edinburgh EH14 4AS, and
Department of Chemistry, UniVersity of Bath, ClaVerton Down, Bath BA2 7AY, United Kingdom
Received October 28, 2010; E-mail: S.A.Macgregor@hw.ac.uk; chsmkw@bath.ac.uk
Abstract: A combination of experimental studies and density functional theory calculations is used to study
C-N bond activation in a series of ruthenium N-alkyl-substituted heterocyclic carbene (NHC) complexes.
These show that prior C-H activation of the NHC ligand renders the system susceptible to irreversible
C-N activation. In the presence of a source of HCl, C-H activated Ru(IiPr2Me2)′(PPh3)2(CO)H (1, IiPr2Me2
) 1,3-diisopropyl-4,5-dimethylimidazol-2-ylidene) reacts to give Ru(IiPrHMe2)(PPh3)2(CO)HCl (2, IiPrHMe2
) 1-isopropyl-4,5-dimethylimidazol-2-ylidene) and propene. The mechanism involves (i) isomerization to a
trans-phosphine isomer, 1c, in which hydride is trans to the metalated alkyl arm, (ii) C-N cleavage to give
an intermediate propene complex with a C2-metalated imidazole ligand, and (iii) N-protonation and propene/
Cl- substitution to give 2. The overall computed activation barrier (∆Eqcalcd) corresponds to the isomerization/
C-N cleavage process and has a value of +24.4 kcal/mol. C-N activation in 1c is promoted by the relief
of electronic strain arising from the trans disposition of the high-trans-influence hydride and alkyl ligands.
Experimental studies on analogues of
1 with different C4/C5 carbene backbone substituents
(Ru(IiPr2Ph2)′(PPh3)2(CO)H, Ru(IiPr2)′(PPh3)2(CO)H) or different N-substituents (Ru(IEt2Me2)′(PPh3)2(CO)H)
reveal that Ph substituents promote C-N activation. Calculations confirm that Ru(IiPr2Ph2)′(PPh3)2(CO)H
undergoes isomerization/C-N bond cleavage with a low barrier of only +21.4 kcal/mol. Larger N-alkyl
groups also facilitate C-N bond activation (Ru(ItBu2Me2)′(PPh3)2(CO)H, ∆Eqcalcd ) +21.3 kcal/mol), and in
this case the reaction is promoted by the formation of the more highly substituted 2-methylpropene.
Introduction
that is, their susceptibility to undergo degradation via bond
activation reactions.2 Thus, Grubbs, Piers, and Dorta have all
N-heterocyclic carbenes (NHCs) are now universally recog-
nized as being a class of ligands that can impart exceptional
catalytic activity onto metal centers. Numerous reviews and,
more recently, a number of textbooks have detailed the role of
NHCs in enhancing the ability of metals to catalyze a range of
reactions, most commonly metathesis, C-C coupling, and H-X
addition across CdO and CdC bonds.1 Consideration of the
metathesis literature highlights another, less appreciated and
certainly less well understood aspect of M-NHC complexes,
shown during efforts to prepare more active ruthenium-based
metathesis catalysts the occurrence of irreversible C-H activa-
tion of an N-substituent on the carbene, resulting in the catalyst
lifetime being compromised and the catalyst efficiency being
reduced.3
Ruthenium complexes appear to be particularly prone to
M-NHC degradation reactions, with not only C-H activation,
but also C-C activation and migration/insertion reactions being
documented.4 In 2006, we showed that treatment of
Ru(PPh3)3(CO)HCl with 2 equiv of 1,3-diisopropyl-4,5-dim-
ethylimidazol-2-ylidene (IiPr2Me2) at 70 °C afforded a mixture
of products consisting not only of the C-H activated carbene
complex cis-Ru(IiPr2Me2)′(PPh3)2(CO)H (1a), but also more
unexpectedly, of 2 (Scheme 1), which results from C-N
activation of one of the N-iPr linkages in the carbene.5 Upon
† Heriot-Watt University.
‡ University of Bath.
§ Present address: Departamento de Qu´ımica, Faculdade de Cieˆncias e
Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal.
(1) (a) Bourissou, D.; Guerret, O.; Gabba¨ı, F. P.; Bertrand, G. Chem. ReV.
2000, 100, 39. (b) Herrmann, W. A.; Weskamp, T.; Bo¨hm, V. P. W.
AdV. Organomet. Chem. 2001, 48, 1. (c) Herrmann, W. A. Angew.
Chem., Int. Ed. 2002, 41, 1290. (d) N-Heterocyclic Carbenes in
Synthesis; Nolan, S. P., Ed.; Wiley-VCH: Weinheim, Germany, 2006.
(e) N-Heterocyclic Carbenes in Transition Metal Catalysis; Glorius,
F., Ed.; Springer-Verlag: Heidelberg, Germany, 2007. (f) Kantchev,
E. A. B.; O’ Brien, C. J.; Organ, M. G. Angew. Chem., Int. Ed. 2007,
46, 2768. (g) Hahn, F. E.; Jahnke, M. C. Angew. Chem., Int. Ed. 2008,
47, 3122. (h) Marion, N.; Nolan, S. P. Acc. Chem. Res. 2008, 41,
1440. (i) D´ıez-Gonza´lez, S.; Marion, N.; Nolan, S. P. Chem. ReV. 2009,
109, 3612. (j) Poyatos, M.; Mata, J. A.; Peris, E. Chem. ReV. 2009,
109, 3677. (k) Dro¨ge, T.; Glorius, F. Angew. Chem., Int. Ed. 2010,
49, 6940. (k) N-Heterocyclic Carbenes in Transition Metal Catalysis
and Organocatalysis; Cazin, C. S. J., Ed.; Springer-Verlag: Heidelberg,
Germany, 2011.
(2) (a) Crudden, C. M.; Allen, D. P. Coord. Chem. ReV. 2004, 248, 2247.
(b) Cavell, K. J. Dalton Trans. 2008, 6676. (c) Albrecht, M. Chem.
ReV. 2010, 110, 576.
(3) (a) Trnka, T. M.; Morgan, J. P.; Sanford, M. S.; Wilhelm, T. E.; Scholl,
M.; Choi, T.-L.; Ding, S.; Day, M. W.; Grubbs, R. H. J. Am. Chem.
Soc. 2003, 125, 2546. (b) Hong, S. H.; Wenzel, A. G.; Salguero, T. T.;
Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2007, 129, 7961. (c)
Leitao, E. M.; Dubberley, S. R.; Piers, W. E.; Wu, Q.; McDonald, R.
Chem.sEur. J. 2008, 14, 11565. (d) Vieille-Petit, L.; Luan, X.; Gatti,
M.; Blumentritt, S.; Linden, A.; Clavier, H.; Nolan, S. P.; Dorta, R.
Chem. Commun. 2009, 3783.
9
18408 J. AM. CHEM. SOC. 2010, 132, 18408–18416
10.1021/ja109702h  2010 American Chemical Society

C-N Bond Activation in Ruthenium Carbene Complexes
A R T I C L E S
Scheme 1
further heating, 2 underwent tautomerism to the N-bound species
3 in the presence of excess NHC.6,7
activated complex 1a in the presence of a chloride source
(either bis(triphenylphosphoranylidene)ammonium chloride
([PPN]Cl) or Ru(PPh3)3(CO)HCl) gave a mixture of 2 and 3
and (ii) C-N activation was completely shut down when
Ru(PPh3)3(CO)HCl was reacted with IiPr2Me2 under 1 atm
of H2, conditions under which C-H activation is reversible.
C-N activation has now been observed in a number of both
early (Y, Zr) and late (Rh, Ir, Pd) metal NHC complexes.9
However, these reactions either involve poorly defined metal
complexes generated in situ using imidazolium salt precursors
or take place under highly forcing or heterogeneous conditions.
Such processes do not lend themselves to detailed mechanistic
studies or the thorough characterization of decomposition
products that is so essential (as highlighted by Grubbs3b) if a
proper understanding of decomposition pathways is to be
obtained. In contrast, our ruthenium system provides a well-
defined set of complexes that allow for the first time an
understanding of the C-N activation decomposition reaction.
Herein, we report a combined experimental and computational
mechanistic study of C-N bond activation originating in 1a
that provides insight into (i) how C-N activation takes place,
(ii) the role of C-H activation, and (iii) the importance of N-
and backbone NHC-substituents and metal ancillary ligands. Of
particular note is the finding that C-N activation is intimately
linked to the presence of a hydride ligand, highlighting the fact
that NHC complexes susceptible to C-H activation may be
particularly vulnerable to a subsequent irreversible decomposi-
tion via C-N activation.
C-N activation of NHCs was first described only in 2004
by Cloke, Caddick, and co-workers.8 Exposure of a THF
solution of Ni(cod)2 and excess 1,3-di-tert-butylimidazol-2-
ylidene (ItBu) to sunlight for two weeks yielded a dinuclear
Ni(II) species containing two C-N activated ItBu ligands.
Shorter irradiation times led to the isolation of an intermediate
species containing a C-H activated N-tBu group, which upon
heating with additional ItBu at 70 °C afforded the same final
C-N activated product, along with isobutene. The importance
of C-H activation as a forerunner to C-N activation
suggested by the nickel work was reinforced by our observa-
tions on ruthenium that (i) direct heating of the C-H
(4) (a) Jazzar, R. F. R.; Macgregor, S. A.; Mahon, M. F.; Richards, S. P.;
Whittlesey, M. K. J. Am. Chem. Soc. 2002, 124, 4944. (b) Giunta,
D.; Ho¨lscher, M.; Lehmann, C. W.; Mynott, R.; Wirtz, C.; Leitner,
W. AdV. Synth. Catal. 2003, 345, 1139. (c) Burling, S.; Mahon, M. F.;
Paine, B. M.; Whittlesey, M. K.; Williams, J. M. J. Organometallics
2004, 23, 4537. (d) Abdur-Rashid, K.; Fedorkiw, T.; Lough, A. J.;
Morris, R. H. Organometallics 2004, 23, 86. (e) Cabeza, J. A.; del
R´ıo, I.; Miguel, D.; Sa´nchez-Vega, M. G. Chem. Commun. 2005, 3956.
(f) Galan, B. R.; Gembicky, M.; Dominiak, P. M.; Keister, J. B.; Diver,
S. T. J. Am. Chem. Soc. 2005, 127, 15702. (g) Becker, E.; Stingl, V.;
Dazinger, G.; Puchberger, M.; Mereiter, K.; Kirchner, K. J. Am. Chem.
Soc. 2006, 128, 6572. (h) Hong, S. H.; Chlenov, A.; Day, M. W.;
Grubbs, R. H. Angew. Chem., Int. Ed. 2007, 46, 5148. (i) Burling, S.;
Paine, B. M.; Nama, D.; Brown, V. S.; Mahon, M. F.; Prior, T. J.;
Pregosin, P. S.; Whittlesey, M. K.; Williams, J. M. J. J. Am. Chem.
Soc. 2007, 129, 1987. (j) Galan, B. R.; Pitak, M.; Gembicky, M.;
Keister, J. B.; Diver, S. T. J. Am. Chem. Soc. 2009, 131, 6822. (k)
Burling, S.; Mas-Marza´, E.; Valpuesta, J. E. V.; Mahon, M. F.;
Whittlesey, M. K. Organometallics 2009, 28, 6676. (l) Benhamou,
L.; Wolf, J.; Ce´sar, V.; Labande, A.; Poli, R.; Lugan, N.; Lavigne, G.
Organometallics 2009, 28, 6981. (m) Armstrong, R.; Ecott, C.; Mas-
Marza´, E.; Page, M. J.; Mahon, M. F.; Whittlesey, M. K. Organo-
metallics 2010, 29, 991.
Results and Discussion
Further Experimental Observations on C-N Activation. The
transformation of 1a to 2 involves a number of separate
processes, including not only cleavage of the C-N bond but
also reorientation from a cis- to a trans-phosphine arrangement,
loss of propene, and, ultimately, addition of “HCl”. As shown
in Scheme 1, reaction from Ru(PPh3)3(CO)HCl requires at least
2 equiv of IiPr2Me2, suggesting that the HCl lost upon converting
Ru(PPh3)3(CO)HCl to 1a is trapped as the imidazolium chloride
(5) Burling, S.; Mahon, M. F.; Powell, R. E.; Whittlesey, M. K.; Williams,
J. M. J. J. Am. Chem. Soc. 2006, 128, 13702.
(6) For other tautomerization processes, see: (a) Sundberg, R. J.; Bryan,
R. F.; Taylor, I. F.; Taube, H. J. Am. Chem. Soc. 1974, 96, 381. (b)
Sini, G.; Eisenstein, O.; Crabtree, R. H. Inorg. Chem. 2002, 41, 602.
(c) Wiedemann, S. H.; Lewis, J. C.; Ellman, J. A.; Bergman, R. G.
J. Am. Chem. Soc. 2006, 128, 2452. (d) Alvarez, E.; Conejero, S.;
Paneque, M.; Petronilho, A.; Poveda, M. L.; Serrano, O.; Carmona,
E. J. Am. Chem. Soc. 2006, 128, 13060. (e) Alvarez, E.; Conejero, S.;
Lara, P.; Lo´pez, J. A.; Paneque, M.; Petronilho, A.; Poveda, M. L.;
del R´ıo, D.; Serrano, O.; Carmona, E. J. Am. Chem. Soc. 2007, 129,
14130. (f) Buil, M. L.; Esteruelas, M. A.; Garce´s, K.; Oliva´n, M.;
On˜ate, E. J. Am. Chem. Soc. 2007, 129, 10998. (g) Ruiz, J.;
Perandones, B. F. J. Am. Chem. Soc. 2007, 129, 9298. (h) Esteruelas,
M. A.; Ferna´ndez-Alvarez, F. J.; On˜ate, E. Organometallics 2008, 27,
6236. (i) Gribble, M. W., Jr.; Ellman, J. A.; Bergman, R. G.
Organometallics 2008, 27, 2152. (j) Song, G.; Su, Y.; Periana, R. A.;
Crabtree, R. H.; Han, K.; Zhang, H.; Li, X. Angew. Chem., Int. Ed.
2010, 49, 912.
(9) (a) Mas-Marza´, E.; Poyatos, M.; Sanau´, M.; Peris, E. Organometallics
2004, 23, 323. (b) Mas-Marza´, E.; Poyatos, M.; Sanau´, M.; Peris, E.
Inorg. Chem. 2004, 43, 2213. (c) Huynh, H. V.; Meier, N.; Pape, T.;
Hahn, F. E. Organometallics 2006, 25, 3012. (d) Yen, S. K.; Koh,
L. L.; Huynh, H. V.; Hor, T. S. A. Dalton Trans. 2007, 3952. (e)
Wang, X.; Chen, H.; Li, X. Organometallics 2007, 26, 4684. (f)
Cabeza, J. A.; del R´ıo, I.; Miguel, D.; Sa´nchez-Vega, M. G. Angew.
Chem., Int. Ed. 2008, 47, 1920. (g) Cooke, C. E.; Jennings, M. C.;
Katz, M. J.; Pomeroy, R. K.; Clyburne, J. A. C. Organometallics 2008,
27, 5777. (h) Han, Y.; Hong, Y.-T.; Huynh, H. V. J. Organomet. Chem.
2008, 693, 3159. (i) Ye, J.; Zhang, X.; Chen, W.; Shimada, S.
Organometallics 2008, 27, 4166. (j) Liu, L.; Wang, F.; Shi, M. Eur.
J. Inorg. Chem. 2009, 1723. (k) Sabiah, S.; Lee, C.-S.; Hwang, W.-
S.; Lin, I. J. B. Organometallics 2010, 29, 290. (l) Hu, Y.-C.; Tsai,
C.-C.; Shih, W.-C.; Yap, G. P. A.; Ong, T.-G. Organometallics 2010,
29, 516.
(7) The factors controlling the relative energies of C-bound 2 and N-bound
3 have been probed using DFT calculations: Ha¨ller, L. J. L.;
Macgregor, S. A. Eur. J. Inorg. Chem. 2009, 2000.
(8) Caddick, S.; Cloke, F. G. N.; Hitchcock, P. B.; de K. Lewis, A. K.
Angew. Chem., Int. Ed. 2004, 43, 5824.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18409

A R T I C L E S
Ha¨ller et al.
Scheme 2
originating in 1a and proceeding through 1c. Experimentally, we
have previously shown that isomerization of 1a to 1c proceeds
through the five-coordinate agostic intermediate 5 (Scheme 3).10
Labeling studies suggest that 5 is formed via protonation of a Ru(0)
intermediate, 4, which itself is derived from 1a via C-H reductive
elimination. 5 can be deprotonated by strong bases to give 1c, a
step which calculations indicate occurs via a novel C-H activation
of a nonagostic C-H bond. This mechanism is consistent with
the isomerization of 1a to 1c occurring in the presence of
[IiPr2Me2H]Cl, which acts as a source of first a proton and then a
strong base. A similar scheme could account for the formation of
1b via deprotonation of an isomer of 5 where the H and CO ligands
have exchanged positions.
Calculations on 1a, 1b, and 1c showed 1a to be most stable
(relative energy 0.0 kcal/mol), followed by 1b (E ) +1.6 kcal/
mol) and 1c (E ) +4.3 kcal/mol), and this ordering is in good
qualitative agreement with the distribution of these three isomers
in benzene at equilibrium. Computed geometries for 1a and 1c
were also in good agreement with those determined experi-
mentally.
[IiPr2Me2H]Cl. We propose that it is this [IiPr2Me2H]Cl that
provides the HCl necessary to transform 1a into 2. In support
of this, when a benzene solution of 1a was heated at 70 °C in
the presence of [IiPr2Me2H]Cl, complete conversion to a mixture
of 2, 3, and propene was observed over a period of 24 h.
Alternative HCl sources, such as [ItBuH]Cl or [NH4]Cl, were
also shown to be capable of converting 1a to the same mixture
of products under similar conditions, although when a solution
of HCl itself (4 M in dioxane) was used, decomposition of all
the ruthenium species in solution was observed through elimina-
tion of imidazolium salt.
When the reaction of 1a with [IiPr2Me2H]Cl at 70 °C was
monitored over shorter time periods, isomerization to a mixture
of 1a and the previously reported complexes 1b and 1c in a
1:0.5:0.03 ratio was observed within 20 min,4i,10 with no sign
of either the C-N activated species 2 or its N-tautomer 3.
Complexes 1b and 1c exhibit a trans disposition of PPh3 ligands
and differ in the relative positions of the hydride and carbonyl
ligands (Scheme 2). Of particular note is that 1c and the C-N
activated species 2 exhibit the same stereochemical orientation
of NHC, PPh3, and CO ligands. A reasonable postulate is
therefore that 1a first isomerizes to 1c, which then undergoes
C-N activation to afford 2. In support of this, thermolysis of
a solution of 1c and [ItBuH]Cl in C6D6 for 1 h gave a mixture
of 2 and 3 (ca. 50% conversion), whereas similar treatment of
1a gave no more than 3% conversion to 2 and 3. This implies
that although 1a can undergo rapid isomerization, the low
equilibrium abundance of 1c relative to 1a results in significantly
reduced formation of 2 and 3.
Figure 1 shows computed reaction profiles for C-N activation
in 1a and 1c. Although alternative mechanisms were considered,12
ultimately this process was shown to entail a simple lengthening
of the N1-C6 bond in the five-membered C-H activated ring.
For isomer 1c a transition state was located, TS1c-6c (E ) +24.4
kcal/mol), which shows significant elongation of both the N1-C6
distance (2.24 Å; cf. 1.49 Å in 1c) and the bond to the metalated
arm (Ru-C7 ) 2.44 Å; cf. 2.25 Å in 1c). The weaker Ru-C7
interaction results in a shortening of the Ru-H1 bond by 0.06 Å
to 1.60 Å. The initial C-N cleavage product is the propene
complex 6c (E ) +16.3 kcal/mol), and the shortening of the
C6-C7 bond throughout this process reflects an increase in double
bond character (1c, 1.55 Å; TS1c-6c, 1.40 Å; 6c, 1.38 Å). 6c also
features a C2-metalated imidazole ligand with asymmetric C2-N1
and C2-N3 bonds (1.35 and 1.41 Å, respectively).
From 6c the final observed C-N activation product, 2, would
be formed by protonation of N1 and displacement of propene by
chloride. The fact that no propene adducts have been observed
experimentally suggests that propene loss is facile, and this is also
consistent with the long computed distances to the propene ligand
in 6c (Ru-C6 ) 2.54 Å, Ru-C7 ) 2.47 Å). The C-N cleavage
step therefore controls the reactivity of the system, and for 1c this
occurs with an activation barrier of 20.1 kcal/mol.
C-N bond cleavage in 1a proceeds via transition state TS1a-6a
(E ) +29.9 kcal/mol) to give propene adduct 6a (E ) +17.7
kcal/mol). The geometric changes throughout this process are
similar to those described above for 1c, and the product complex
6a again features a weakly bound propene ligand (Ru-C6 )
2.48 Å, Ru-C7 ) 2.43 Å) and a shortened distance to the CO
originally trans to the metalated arm (Ru-C8 ) 1.88 Å (1a),
1.84 Å (6a)). The combination of experimental and computa-
tional results now allows us to propose the pathway shown in
Scheme 4 for the conversion of 1a to the C-N activated species
2. The lower energy of 1a means it will be the dominant species
Further experiments probed the reactivity of 1c. First, C-N
activation of 1c (determined by monitoring the formation of
1
propene by H NMR spectroscopy) occurred at the same rate
in the absence of a HCl source, although under these conditions
a mixture of unidentified Ru-containing products was observed.
This suggests that the role of HCl is simply to trap out the
products 2 and 3 after cleavage of the C-N bond.11 Second,
heating solutions of 1c and [ItBuH]Cl in the presence or absence
of ca. 30 equiv of PPh3 revealed no difference in the rate of
either C-N activation or isomerization. Moreover, an equivalent
reaction of 1c and [ItBuH]Cl in the presence of ca. 30 equiv of
P(p-tolyl)3 showed no incorporation of the tolylphosphine in
the products. Therefore, neither the C-N activation nor the prior
isomerization reactions involve phosphine dissociation.
Density Functional Theory (DFT) Calculations. We have
undertaken DFT calculations to model the C-N activation process
Scheme 3
9
18410 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

C-N Bond Activation in Ruthenium Carbene Complexes
A R T I C L E S
Figure 1. Computed reaction profiles (kcal/mol) for C-N activation in 1a and 1c. Selected distances are given in angstroms. Phenyl substituents on the
phosphine ligands have been removed for clarity.
Scheme 4
in solution; however, isomerization to 1c allows the system to
access the lower energy C-N cleavage pathway via TS1c-6c
metalated imidazole-propene complex 6c loses propene and
is trapped by HCl to generate 2 in the final step of the reaction.
.
C-N activation via this dual isomerization/cleavage process has
an overall barrier (relative to 1a) of 24.4 kcal/mol, significantly
lower than direct reaction from 1a (29.9 kcal/mol). The C2-
To understand why C-N bond cleavage is so much easier
in 1c compared to 1a, a range of model systems was considered
in which the substituents at N1 (R), at the NHC C4/C5 backbone
positions (R′), and in the phosphine (R′′) were varied (see Table
1). In all cases the same trend was seen, with ∆Eq(1c) being
8-10 kcal/mol lower than ∆Eq(1a). This is particularly impor-
tant with model 1 (R ) R′ ) R′′ ) H), where all significant
ligand steric effects have been removed. The more facile C-N
cleavage in isomer 1c must therefore be primarily electronic in
origin, and we suggest it arises from the unfavorable trans
disposition of the high-trans-influence hydride and alkyl ligands
present in 1c. This causes an electronic destabilization that is
relieved upon lengthening of the Ru-C7 bond in the C-N
(10) Ha¨ller, L. J. L.; Page, M. J.; Macgregor, S. A.; Mahon, M. F.;
Whittlesey, M. K. J. Am. Chem. Soc. 2009, 131, 4604.
(11) When [ItBuH]BF4 was used instead of [ItBuH]Cl to provide a source
of HBF4 rather than HCl, propene was generated but neither 2 nor 3
was formed.
(12) Alternative pathways based on initial formation of imidazolium via
C-H reductive elimination proved inaccessible or very high in energy.
In addition, introduction of an imidazolium cation as a proton source
did not affect the computed C-N cleavage activation barriers. This
latter observation is consistent with the fact that C-N activation of
1c occurs at the same rate in the absence of a HCl source.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18411

A R T I C L E S
Ha¨ller et al.
Table 1. Computed Energetics (kcal/mol) for C-N Activation with
Different Models of 1a and 1c
model
1a TS1a-6a ∆Eq(1a) 1c TS1c-6c ∆Eq(1c) ∆∆Eq
(1) R ) R′ ) R′′ ) H 0.0 27.7
27.7 8.7 28.2
29.9 7.5 28.9
19.5
21.4
8.2
8.5
(2) R ) Me;
R′ ) R′′ ) H
0.0 29.9
i
(3) R ) Pr;
0.0 30.8
30.8 7.5 29.3
31.2 8.4 30.1
29.9 4.3 24.4
21.9
21.8
20.1
8.9
9.4
9.8
R′ ) R′′ ) H
i
(4) R ) Pr; R′ ) Me; 0.0 31.2
R′′ ) H
i
(5) R ) Pr; R′ ) Me; 0.0 29.9
R′′ ) Ph
Figure 2. Molecular structure of 1adppe. Thermal ellipsoids are represented
at 30% probability. All hydrogen atoms (except those on the activated
carbene arm and Ru-H) are omitted for clarity. Selected bond lengths (Å)
and angles (deg): Ru(1)-C(1), 1.864(2); Ru(1)-C(2), 2.064(2); Ru(1)-C(6),
2.208(2); Ru(1)-P(1), 2.3429(6); Ru(1)-P(2), 2.2978(7); P(1)-Ru(1)-P(2),
85.21(2); C(1)-Ru(1)-C(2), 96.69(9); C(2)-Ru(1)-C(6), 76.92(9);
C(2)-Ru(1)-P(2), 169.02(6).
cleavage transition state. In 1a CO is trans to the metalated arm,
but the lower trans influence of this ligand means that the initial
electronic destabilization is less significant and so results in a
higher activation barrier.13
The results for model 1 show that, unlike the full
experimental system (model 5), the oVerall barrier to C-N
activation is lower for the direct reaction from 1a (27.7 kcal/
mol) rather than the isomerization/C-N cleavage route via
1c (28.2 kcal/mol). The major difference between these two
systems is in the relative energy of isomer 1c. With model
1 1c lies 8.7 kcal/mol above 1a, and this large difference
outweighs the lower barrier for the C-N bond cleavage step
found with 1c. In model 5, the bulky PPh3 ligands facilitate
the initial isomerization and 1c (with trans-PPh3 ligands) is
now only 4.3 kcal/mol above 1a (with cis-PPh3 ligands). As
a result, the barrier for the isomerization/C-N cleavage route
in the full system is 5.5 kcal/mol lower than that for the direct
reaction from 1a. The intermediate models 2-4 show that
changes in the N3 and C4/C5 substituents cause only subtle
changes in the relative energies of 1a and 1c and small
increases in the C-N cleavage barriers for both isomers. In
going from model 4 to model 5, the introduction of the full
PPh3 ligands stabilizes both transition states, suggesting that
relief of steric strain also plays some role in promoting C-N
activation in the full experimental systems.
In summary, two major factors come together to facilitate
the isomerization/C-N cleavage process in 1a: (i) the use of
bulky PPh3 ligands to facilitate isomerization to 1c and (ii) the
electronic activation of 1c toward C-N cleavage by placing a
high-trans-influence hydride ligand trans to the cyclometalated
alkyl group.
Ancillary Ligand and NHC Substituent Effects on C-N
Activation. The breadth of the C-N activation reaction was
probed by studying the reactivity of analogues of 1 in which
(i) the two PPh3 groups were replaced by a chelating dppe ligand
(1dppe), (ii) the NHC backbone substituents R′ were changed to
Ph (1iPr/Ph) and H (1iPr/H), and (iii) the N-iPr substituents were
changed to N-Et (1Et/Me) and N-tBu (1tBu/Me) groups. Scheme 5
shows the range of complexes that were investigated in the form
of their a isomers.4c,i
1. C-N Bond Activation in 1adppe. The dppe complex 1adppe
was isolated in 68% yield following the reaction of
Ru(dppe)(PPh3)(CO)HCl with 2 equiv of IiPr2Me2 at 70 °C. The
isomer a geometry was confirmed by the appearance of large
and small phosphorus couplings to the single hydride resonance
at δ -5.8 (JHP ) 108, 19 Hz) and the carbenic carbon signal at
δ 190 (JCP ) 85, 10 Hz), as well as by an X-ray crystal structure
(Figure 2). In comparison to the structure of 1a,4i the presence
of the dppe ligand results in a much more acute P-Ru-P angle
(1adppe, 85.21(2)°; 1a, 102.152(15)°) and shorter Ru-P bond
lengths (1adppe, 2.2978(7), 2.3429(6) Å; 1a, 2.3445(4), 2.4357(4)
Å).
In contrast to 1a, thermolysis of a benzene solution of 1adppe
and [ItBuH]Cl at 70 °C for several days gave no C-N activated
products. In this case, a reaction path involving a trans-
phosphine isomer equivalent to 1c is not accessible due to the
presence of the chelating dppe ligand. However, the necessary
trans-H-Ru-alkyl arrangement can be achieved in isomer
1ddppe, in which the H and CO have exchanged positions
compared to those in 1adppe. DFT calculations on C-N cleavage
in 1ddppe predict a barrier of 26.2 kcal/mol, somewhat higher
than that computed for 1a, but potentially still accessible at
higher temperatures. Indeed, heating 1adppe at 120 °C for 24 h
Scheme 5
9
18412 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

C-N Bond Activation in Ruthenium Carbene Complexes
A R T I C L E S
Scheme 6
in the presence of [ItBuH]Cl did lead to C-N activation with
formation of both propene and 1-isopropyl-4,5-dimethylimida-
zole, although no clearly identifiable ruthenium-containing
species were formed. The computed barrier for direct C-N
cleavage in 1adppe had a much higher computed barrier of 32.2
kcal/mol, and this re-emphasizes the role of the trans-hydride
ligand in promoting this process.
activation, benzene solutions of 1a and 1aPr/Ph/1bPr/Ph were
heated with 2 equiv of [ItBuH]Cl at 70 °C side by side over a
period of ca. 60 h. Depletion of the starting hydride signal and
the appearance of propene were measured intermittently by
proton NMR against an internal standard of ferrocene and found
i
i
i
to follow the order 1aPr/Ph > 1a.
i
When C-N activation in 1Pr/Ph was modeled with DFT
i
i
i
i
2. C-N Bond Activation in 1Pr/Ph and 1Pr/H. The attempted
calculations as for 1Pr/Me above, isomer 1aPr/Ph was found to
be the most stable form, in agreement with the experimental
observations. The lowest energy C-N activation transition state
is again accessed through isomer c (E ) +4.0 kcal/mol) and
proceeds with an overall barrier of 21.4 kcal/mol, lower than
i
synthesis of 1Pr/Ph via reaction of Ru(PPh3)3(CO)HCl with an
excess of the free carbene IiPr2Ph2 at 70 °C formed a mixture
of 1Pr/Ph, Ru(IiPr2Ph2)(PPh3)2(CO)H2 (7), and Ru(IiPr2Ph2)2-
i
(PPh3)(CO)H2 (8), as well as propene and 1-isopropyl-4,5-
diphenylimidazole. When the reaction was performed under 1
atm of H2, only 7 and 8 were formed.14 Isolation of 7 followed
by reaction with the hydrogen acceptor CH2dCHSiMe3 for
i
found for 1Pr/Me (+24.4 kcal/mol) and therefore consistent with
the more rapid reaction seen for the former.
i
Experimental studies on 1Pr/H, which has only hydrogen
atoms at the backbone C4 and C5 positions, revealed behavior
i
several days at room temperature afforded 1Pr/Ph in 63% isolated
yield (Scheme 6). 1H and 31P{1H} NMR spectroscopy revealed
quite different from that of 1Pr/Ph and 1Pr/Me. First, 1Pr/H has
been shown previously to exist in solution as the b isomer.10
When it was heated with [ItBuH]Cl at 70 °C for 48 h, the
formation of both propene and 1-isopropylimidazole was
observed, indicating that C-N activation was still occurring,
although only trace amounts of the corresponding C- and
i
i
i
i
that 1Pr/Ph was present in solution as a 3:2 mixture of the two
i
i
isomers 1aPr/Ph and 1bPr/Ph. Thus, the hydride region of the
proton NMR spectrum exhibited a doublet of doublets for
i
1aPr/Ph at δ -7.6 (JHP ) 103, 28 Hz) and a triplet for
1bPr/Ph at δ -7.3 (JHP ) 24 Hz). As expected, the 31P{1H} NMR
i
spectrum showed two doublet resonances at δ 56.7 and 36.5
N-bound ruthenium products 2Pr/H and 3Pr/H were formed.17
The predominant ruthenium-containing product was now
Ru(PPh3)3(CO)H2, implying that carbene loss becomes a major
pathway when R′ ) H. DFT calculations confirm that isomer
1b is more stable than 1a (by 0.8 kcal/mol); however, the
computed barrier via the standard isomerization/C-N cleavage
i
i
i
(JPP ) 16 Hz) arising from 1aPr/Ph and an AB pattern for
i
1bPr/Ph centered at δ 59.7.
The X-ray crystal structure of the major isomer 1aPr/Ph (Figure
i
3) revealed the expected distorted octahedral geometry. The
i
metrical parameters of 1aPr/Ph were very similar to those of 1a,
showing that backbone phenyl groups have only a minimal
impact on structure.
i
i
Thermolysis of the isomeric mixture of 1aPr/Ph/1bPr/Ph with
[I BuH]Cl in benzene at 70 °C for 6 h gave a mixture of 2Pr/Ph
t
i
and the N-bound tautomer 3Pr/Ph (Scheme 7).15 Some of the
i
starting material remained, although all of this disappeared over
a period of 24 h.16 In an effort to provide at least a qualitative
order to the influence of the backbone groups on C-N
(13) For the same reasons, C-N cleavage in 1b (where CO is also trans to
alkyl) involves a high barrier of +27.3 kcal/mol, while C-N activation
in a fourth isomer, 1d (E ) +7.2 kcal/mol), related to 1a via exchange
of H and CO ligands, leads to a lower barrier of only 22.8 kcal/mol.
(14) The X-ray structure of 7 is given in the Supporting Information.
i
(15) The C-N activated complex 2Pr/Ph was identified by comparison of
its NMR spectrum to that of 2, with the Ru-H (δ-14.0) and N-H
(δ 11.7) proton resonances appearing at chemical shifts very similar
to those of 2 (Ru-H, δ-14.0; N-H, δ 10.2). The identity of the
i
N-bound tautomer 3Pr/Ph was confirmed by an independent synthesis
involving the reaction of Ru(PPh3)3(CO)HCl with 1-isopropyl-4,5-
diphenylimidazole (see the Supporting Information).
i
i
(16) The formation of 2Pr/Ph and 3Pr/Ph was accompanied by two other
Ru-phosphine species, a hydride-containing complex, 9, and a non-
hydridic compound, 10. Prolonged heating (g72 h) afforded 10 as
the major reaction product. Spectroscopic and structural details for
10, which contains a C-H activated phenyl group on the backbone
of the NHC ligand, are provided in the Supporting Information.
(17) The reaction of Ru(PPh3)3(CO)HCl and N-isopropylimidazole was used
i
Figure 3. Molecular structure of 1aPr/Ph. Thermal ellipsoids are represented
at 30% probability. Solvent and all hydrogen atoms (except those on the
activated carbene arm and Ru-H) are omitted for clarity. Selected bond
lengths (Å) and angles (deg): Ru(1)-C(1), 1.8640(18); Ru(1)-C(2),
2.0668(16); Ru(1)-C(10), 2.2058(16); Ru(1)-P(1), 2.4124(4); Ru(1)-P(2),
2.3457(4); P(1)-Ru(1)-P(2), 102.520(15); C(1)-Ru(1)-C(2), 95.56(7);
C(2)-Ru(1)-C(10), 76.94(6); C(2)-Ru(1)-P(2), 161.80(5).
i
for the independent synthesis of 3Pr/H. See the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18413

A R T I C L E S
Ha¨ller et al.
Scheme 7
process is only +24.1 kcal/mol (relative to 1b). This would be
NHC and by more highly substituted alkyl substituents on the
N atoms of the imidazolylidene rings.
i
inconsistent with the slower reaction of 1Pr/H compared to
i
1
Pr/Me. This result, in combination with the different behavior
Our study shows that N-alkyl NHC ligands that undergo
intramolecular C-H activation may be particularly susceptible
to irreversible C-N activation. In addition to forming the
cyclometalated ring that is necessary for alkene formation, this
process also produces hydride-containing species, and this high-
trans-influence ligand can promote the C-N cleavage reaction.
Given the propensity of NHC ligands to intramolecular C-H
activation on a wide varietry of metal centers3,4 and the
continuing interest of many groups in using bulky carbenes to
stabilize low-coordinate complexes,19 our findings would suggest
that many more examples of C-N activation will be forthcoming.
i
of 1Pr/H seen experimentally, suggests that the detailed mech-
anism of C-N activation may well be different in this case.
t
3. C-N Bond Activation in 1Et/Me and 1 Bu/Me. The influence
of the N-substituent was established by turning to the N-ethyl
complex 1Et/Me, which also exists in solution as the b isomer.
When 1bEt/Me was heated in the presence of [ItBuH]Cl, <10%
conversion to the C- and N-bound products 2Et/Me and 3Et/Me
was found even after one week at 70 °C.18 DFT calculations
found isomers 1a and 1b to be almost equienergetic in this case,
while the computed overall barrier to C-N activation via TS1c-6c
was 25.6 kcal/mol. This higher barrier is consistent with the
i
lower reactivity seen for 1Et/Me compared to either 1Pr/Me or
Experimental Section
i
1Pr/Ph and reflects the nature of the alkene being formed;
General Comments. All manipulations were carried out using
standard Schlenk, high-vacuum, and glovebox techniques. Solvents
were purified using an MBraun SPS solvent system (toluene, Et2O,
hexane) or under a nitrogen atmosphere from sodium benzophenone
ketyl (benzene) or Mg/I2 (EtOH). Deuterated solvents (Fluorochem)
were vacuum transferred from potassium (C6D6) or CaH2 (CD2Cl2).
Ru(IiPr2Me2)′(PPh3)2(CO)H (1),1i,5 Ru(dppe)(PPh3)(CO)HCl,20
Ru(PPh3)3(CO)HCl,21 Ru(IEt2Me2)′(PPh3)2(CO)H (1Et/Me),4c
hyperconjugation effects should help stabilize the incipient
i
i
propene moiety in TS1c-6c derived from 1Pr/Me (or 1Pr/Ph), but
no such stabilization is available in the equivalent transition
state involving ethene formation from 1Et/Me
.
t
t
The N- Bu analogue, 1Bu/Me, would help substantiate this idea,
as the formation of 2-methylpropene should further favor C-N
activation. Although this species could not be accessed experi-
mentally, calculations were able to define a barrier of 21.3 kcal/
23
IiPr2Me2,22 and [IiPr2Ph2H]BF4 were prepared according to the
literature. NMR spectra were recorded on Bruker Avance 400 and
500 MHz NMR spectrometers at 298 K with 1H and 13C{1H} spectra
referenced as follows: C6D6 at δ 7.15 (1H), δ 128.0 (13C); CD2Cl2
at δ 5.32 (1H), δ 54.0 (13C). 31P{1H} NMR chemical shifts were
referenced externally to 85% H3PO4 (δ 0.0). IR spectra were
recorded on a Nicolet Nexus FTIR spectrometer. Mass spectrometry
was undertaken using a micrOTOF electrospray time-of-flight (ESI-
TOF) mass spectrometer (Bruker Daltonik GmbH). Elemental
analyses were performed by Elemental Microanalysis Ltd., Oke-
hampton, Devon, U.K.
t
mol relative to the most stable isomer 1a Bu/Me, significantly
i
lower than for either 1aPr/Me (24.4 kcal/mol) or 1aEt/Me (25.6
kcal/mol).
Conclusions
A combination of experimental studies and DFT calculations
has been used to show that C-N bond activation of the
N-heterocyclic carbene ligand in the C-H activated complex
Ru(IiPr2Me2)′(PPh3)2(CO)H (IiPr2Me2 ) 1,3-diisopropyl-4,5-
dimethylimidazol-2-ylidene) involves initial isomerization to a
trans-phosphine isomer, 1c, in which the hydride ligand is trans
to the metalated carbene arm. C-N cleavage then gives an
intermediate propene complex that features a C2-bound imida-
zole ligand. Under the reaction conditions this species reacts
with imidazolium chloride as a source of HCl to afford the
experimentally observed C-N activated product 2 (Scheme 1).
C-N activation in 1c is promoted by an electronic destabi-
lization that arises from the trans arrangement of high-trans-
influence hydride and alkyl ligands. This is relieved in the
transition state by C-N bond cleavage within the five-membered
C-H activated ring. The preliminary isomerization step is also
promoted by bulky PPh3 ligands that facilitate a trans-phosphine
geometry. We have also established that C-N cleavage is
promoted by Ph substituents at the backbone positions of the
Ru(IiPr2Me2)′(dppe)(CO)H (1adppe). A solution of Ru(dppe)-
(PPh3)(CO)HCl (588 mg, 0.711 mmol) and IiPr2Me2 (277 mg, 1.536
mmol) in toluene (10 mL) was heated at 70 °C overnight to yield
a white precipitate of the imidazolium salt [IiPr2Me2H]Cl and a
pale yellow solution. The solution was filtered, and hexane (50 mL)
was added to the eluent to precipitate a beige solid of the product,
which was filtered, washed with hexane (2 × 3 mL), and dried in
vacuo. Yield: 345 mg, 68%. 1H NMR (500 MHz, C6D6): δ 8.03 (t,
2H, J ) 8.5 Hz, C6H5), 7.99 (t, 2H, J ) 8.5 Hz, C6H5), 7.77 (t, 2H,
(19) See, for example: (a) Arduengo, A. J., III; Gamper, S. F.; Calabrese,
J. C.; Davidson, F. J. Am. Chem. Soc. 1994, 116, 4391. (b) Yamashita,
M.; Goto, K.; Kawashima, T. J. Am. Chem. Soc. 2005, 127, 7294. (c)
Laitar, D. S.; Mu¨ller, P.; Gray, T. G.; Sadighi, J. P. Organometallics
2005, 24, 4503. (d) van der Eide, E. F.; Romero, P. E.; Piers, W. E.
J. Am. Chem. Soc. 2008, 130, 4485. (e) Miyazaki, S.; Koga, Y.;
Matsumoto, T.; Matsubara, K. Chem. Commun. 2010, 46, 1932.
(20) Santos, A.; Lo´pez, J.; Montoya, J.; Noheda, P.; Romero, A.; Echa-
varren, A. M. Organometallics 1994, 13, 3605.
(18) Under more forcing conditions (60 h, 95 °C) formation of the N-bound
tautomer 3Et/Me was achieved. The identity of the product was
confirmed by an X-ray crystal structure following an independent
synthesis from Ru(PPh3)3(CO)HCl with 1-ethyl-4,5-dimethylimidazole.
See the Supporting Information.
(21) Ahmad, N.; Levison, J. J.; Robinson, S. D.; Uttley, M. F. Inorg. Synth.
1974, 15, 45.
(22) Ku¨hn, N.; Kratz, T. Synthesis 1993, 561.
(23) Chianese, A. R.; Kovacevic, A.; Zeglis, B. M.; Faller, J. W.; Crabtree,
R. H. Organometallics 2004, 23, 2461.
9
18414 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

C-N Bond Activation in Ruthenium Carbene Complexes
A R T I C L E S
Table 2. Crystal Data and Structure Refinement for Compounds
J ) 8.5 Hz, C6H5), 7.15 (m, 4H, C6H5), 7.10-6.90 (m, 10H, C6H5),
i
3
1adppe and 1aPr/Ph
5.63 (sept, 1H, JHH ) 7.0 Hz, CH(CH3)2), 3.92 (m, 1H,
CH(CH3)(CH2)), 2.36 (m, 2H, PCH2), 2.09 (m, 1H, PCH2), 1.95
(m, 1H, PCH2), 1.78 (s, 3H, NCCH3), 1.70 (s, 3H, NCCH3), 1.45
i
1a Pr/Ph
1adppe
empirical formula
fw
C38H44N2OP2Ru
707.76
C64H60N2OP2Ru
1036.15
3
3
(d, 3H, JHH ) 6.3 Hz, CH(CH3)), 1.38 (d, 3H, JHH ) 7.0 Hz,
3
CH(CH3)), 0.92 (d, 3H, JHH ) 7.0 Hz, CH(CH3)), 0.71 (m, 1H,
cryst syst
space group
a (Å)
b (Å)
c (Å)
triclinic
P1
triclinic
P1
2
CHH′), 0.30 (br t, 1H, J ) 10.2 Hz, CHH′), -5.78 (dd, 1H, JHP
j
j
) 107.9 Hz, JHP ) 19.2 Hz, RuH). 31P{1H} NMR (202 MHz,
2
11.2630(2)
12.6030(2)
12.7410(2)
99.643(1)
94.862(1)
104.420(1)
1711.56(5)
2
10.1400(1)
13.6920(2)
19.2510(3)
81.982(1)
86.649(1)
84.854(1)
2633.09(6)
2
C6D6): δ 76.8 (d, 2JPP ) 5.0 Hz), 58.5 (d, 2JPP ) 5.0 Hz). 13C{1H}
2
2
NMR (125 MHz, C6D6): δ 208.5 (dd, JCP ) 9 Hz, JCP ) 5 Hz,
CO), 189.8 (dd, 2JCP ) 85 Hz, 2JCP ) 10 Hz, NCN), 143.1 (d, JCP
R (deg)
ꢀ (deg)
γ (deg)
U (Å3)
Z
) 22 Hz, PC6H5), 140.8 (d, JCP ) 34 Hz, PC6H5), 138.3 (d, JCP
)
38 Hz, PC6H5), 136.2 (d, JCP ) 23 Hz, PC6H5), 133.3 (d, JCP ) 10
Hz, PC6H5), 132.9 (d, JCP ) 12 Hz, PC6H5), 132.6 (d, JCP ) 11
Hz, PC6H5), 129.2 (s, PC6H5), 128.8 (s, PC6H5), 128.7 (s, PC6H5),
124.0 (s, NCCH3), 122.5 (s, NCCH3), 58.8 (d, JCP ) 6 Hz,
CH(CH3)), 54.0 (s, CH(CH3)2), 30.4 (dd, JCP ) 22 Hz, JCP ) 18
Hz, PCH2), 28.1 (dd, JCP ) 24 Hz, JCP ) 21 Hz, PCH2), 22.8 (s,
Dc (g cm-3
)
1.373
0.584
736
1.307
0.403
1080
µ (mm-1
F(000)
)
cryst size (mm)
θ min, max for data
collection (deg)
index ranges
0.30 × 0.25 × 0.25
0.30 × 0.25 × 0.25
CH(CH3)), 21.9 (s, CH(CH3)), 21.8 (s, CH(CH3)), 20.5 (dd, 2JCP
)
3.88, 30.00
3.79, 27.53
9 Hz, 2JCP ) 7 Hz, RuCH2), 11.0 (s, NCCH3), 9.6 (s, NCCH3). IR
(C6H6, cm-1): 1885 (νCO). Anal. Found (Calcd) for C38H44N2OP2Ru:
-15 e h e +15,
-17 e k e +17,
-17 e l e +17
37121
-13 e h e +13,
-17 e k e +17,
-24 e l e +25
51926
C, 64.10 (64.48); H, 6.09 (6.27); N, 3.94 (3.96).
i
i
i
Ru(I Pr2Ph2)′(PPh3)2(CO)H (1aPr/Ph/1bPr/Ph). A benzene (10
24
no. of reflns collected
no. of independent
reflns, Rint
mL) solution of Ru(IiPr2Ph2)(PPh3)2(CO)H2 (7; 90 mg, 0.094
9889, 0.0861
11986, 0.0349
mmol) and H2CdCHSiMe3 (200 µL, 1.36 mmol) in benzene was
stirred at room temperature for 72 h. The volatiles were removed
under vacuum, and the residue was washed with hexane (2 × 3
no. of reflns obsd
(>2σ)
7010
10606
i
i
mL) to yield a white precipitate of 1aPr/Ph and 1bPr/Ph (57 mg,
data completeness
abs correction
max, min transmission
no. of data/restraints/
params
0.993
0.989
63%). Recrystallization of the precipitate from benzene/hexane
multiscan
0.89, 0.79
9889/4/416
multiscan
0.908, 0.852
11986/1/640
allowed 1aPr/Ph to be isolated. 1H NMR (500 MHz, C6D6): δ 7.72
i
(m, 6H, C6H5), 7.49 (t, 6H, J ) 8.1 Hz, C6H5), 7.17 (d, 2H, J )
8.1 Hz, C6H5), 7.10-6.93 (m, 22H, C6H5), 6.88 (m, 4H, C6H5),
3
goodness-of-fit on F2
final R1, wR2 [I >
2σ(I)]
0.984
0.0430, 0.0837
1.074
0.0291, 0.0713
5.59 (sept, 1H, JHH ) 7.0 Hz, CH(CH3)2), 4.88 (m, 1H,
CH(CH3)(CH2)), 2.19 (m, 1H, CHH′), 1.71 (d, 3H, 3JHH ) 7.0 Hz,
3
CH(CH3)), 1.15 (d, 3H, JHH ) 6.2 Hz, CH(CH3)), 0.60 (br t, 1H,
final R1, wR2 (all
data)
0.0797, 0.0939
0.0364, 0.0761
J ) 6.7 Hz, CHH′), 0.23 (d, 3H, 3JHH ) 7.0 Hz, CH(CH3)), -7.55
2
2
(dd, 1H, JHP ) 103.4 Hz, JHP ) 28.2 Hz, RuH). 31P{1H} NMR
(202 MHz, C6D6): δ 56.7 (d, 2JPP ) 16.1 Hz), 36.5 (d, 2JPP ) 16.1
Hz). 13C{1H} NMR (125 MHz, C6D6): δ 206.7 (dd, 2JCP ) 13 Hz,
2JCP ) 6 Hz, CO), 191.0 (dd, 2JCP ) 83 Hz, 2JCP ) 10 Hz, NCN),
139.8 (C6H5), 139.6 (d, JCP ) 9 Hz, PC6H5), 135.2 (d, JCP ) 11
Hz, PC6H5), 134.6 (s, C6H5), 134.5 (s, C6H5), 132.4 (s, NCC6H5),
131.7 (s, NCC6H5), 130.4 (s, C6H5), 128.9 (s, PC6H5), 128.3 (d,
JCP ) 8 Hz, PC6H5), 128.1 (s, C6H5), 128.0 (s, C6H5), 58.9 (s,
largest diff peak,
0.597, -0.919
0.484, -0.574
hole (e Å-3
)
Supplementary Publications CCDC 756039 and 756040. Copies
of the data can be obtained free of charge on application to CCDC,
12 Union Rd., Cambridge CB2 1EZ, U.K. [fax, (+44) 1223 336033;
e-mail, deposit@ccdc.cam.ac.uk].
CH(CH3)), 55.5 (s, CH(CH3)2), 25.5 (s, CH(CH3)), 23.8 (t, 2JCP
)
Computational Details. DFT calculations were run with Gauss-
ian 03 (revision D.01)26 using the BP86 functional. Ru and P centers
were described with the Stuttgart RECPs and associated basis sets,27
with added d-orbital polarization on P.28 6-31G** basis sets were
used for all other atoms.29 All stationary points were fully
characterized via analytical frequency calculations as either minima
(all positive eigenvalues) or transition states (one imaginary
eigenvalue), and IRC calculations and subsequent geometry opti-
mizations were used to confirm the minima linked by each transition
state. Reported energies include a correction for zero-point energies.
7 Hz, RuCH2), 22.0 (s, CH(CH3)), 21.9 (s, CH(CH3)). IR (C6H6,
cm-1): 1889 (νCO). Anal. Found (Calcd) for C58H55N2OP2Ru·C6H6:
C, 73.81 (74.11); H, 5.82 (5.93); N, 2.65 (2.70).
X-ray Crystallography. Single crystals of compounds 1adppe
i
and 1aPr/Ph were analyzed at 150 K using Mo KR radiation on a
Nonius Kappa CCD diffractometer. Details of the data collections,
solutions, and refinements are given in Table 2. The structures were
solved using SHELXS-9725 and refined using full-matrix least-
squares analysis in SHELXL-97.25 Refinements were generally
straightforward with the following exceptions and points of note.
H1, H5, H6a, and H6B in 1adppe were located and refined at
distances of 1.6 Å (for H1), and 0.9 Å (for the latter three
hydrogens) from the relevant parent atoms. The asymmetric unit
The following protocol was adopted to ensure that the lowest
energy conformations were located for each stationary point. For
each initial structure (derived from experiment or built from
chemical intuition) a molecular dynamics (MD) simulation was
i
in 1aPr/Ph contains one molecule of benzene in addition to one
molecule of the ruthenium complex. As for 1adppe, the hydride (H1)
was located and refined at 1.6 Å from Ru1. H10a and H10b were
ultimately included at calculated positions proximate to those
evident in the penultimate difference Fourier electron density map.
Crystallographic data for compounds 1adppe and 1aPr/Ph have been
deposited with the Cambridge Crystallographic Data Centre as
(26) Frisch, M. J.; et al. Gaussian 03, revision D.01; Gaussian, Inc.:
Wallingford, CT, 2004.
(27) Andrae, D.; Ha¨ussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123.
i
(28) Ho¨llwarth, A.; Bo¨hme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.;
Jonas, V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 237.
(24) See the Supporting Information.
(25) Sheldrick, G. M. Acta. Crystallogr. 1990, 467-473, A46. Sheldrick,
G. M. SHELXL-97, a computer program for crystal structure
refinement; University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(29) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18415

A R T I C L E S
Ha¨ller et al.
performed using the Tinker program30 and the MM3 force field.
These runs allowed movement of the hydrocarbyl substituents on
the phosphine and NHC ligands, with all other bonds being fixed.
MD simulations were run for 1 ns with a 1 fs time step in an NVT
ensemble, with coordinates being collected every picosecond to
generate 1000 structures. The trajectories were propagated using
the modified Beeman integration algorithm,31 and a Berendsen
thermostat32 was used to keep the temperature around 1000 K. We
found this relatively high temperature was needed to span the
conformational space in the most efficient way. The 1000 generated
structures were then optimized with the MM3 force field with the
same geometry constraints as above. Energetically unique structures
(typically between 15 and 60 conformations for each stationary
point) were then selected for optimization at the DFT level using
the “opt)loose” option in G03 and an SCF convergence of 10-6
au. The lowest energy structure and any others within approximately
1 kcal/mol (typically up to three structures) were then optimized
with regular convergence criteria and characterized via frequency
analyses. Sampling a larger range of structures from the MD
simulation (for example, 3000 or 10000 structures) did not result
in any new conformers. To test functional dependence, the lowest
energy conformers were reoptimized with the M06 and B97-D
approaches available with Gaussian 09 (revision A.02);33 these did
not reveal any significant changes to the trends obtained with the
BP86 results.
Acknowledgment. We thank the EPSRC for financial support
(L.J.L.H., M.J.P., S.E., M.A.N.) and Johnson Matthey plc for the
kind loan of hydrated ruthenium trichloride.
Supporting Information Available: CIF files giving X-ray
i
crystallographic data for 1adppe and 1aPr/Ph, experimental data
i
i
for 7, 3Pr/Ph, 10, 3Pr/H, and 3Et/Me, structural data for 7, 10, and
Et/Me, computed geometries and energies for all stationary
3
points, and full refs 26 and 33. This material is available free
of charge via the Internet at http://pubs.acs.org.
(30) (a) Ren, P.; Ponder, J. W. J. Phys. Chem. B 2003, 107, 5933. (b) Ren,
P.; Ponder, J. W. J. Comput. Chem. 2002, 23, 1497. (c) Pappu, R. V.;
Hart, R. K.; Ponder, J. W. J. Phys. Chem. B 1998, 102, 9725.
(31) Beeman, D. J. Comput. Phys. 1976, 20, 130.
(32) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola,
A.; Haak, J. R. J. Chem. Phys. 1984, 81, 3684.
JA109702H
(33) Frisch, M. J.; et al. Gaussian 09, revision A.02; Gaussian, Inc.,
Wallingford, CT, 2009.
9
18416 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010