Double Ci-H bond activation of C(sp3)H2 groups for the preparation of complexes with back-to-back bisimidazolinylidenes

Article abstract of DOI:10.1002/anie.201101806

Two-headed carbenes: N-heterocyclic carbene complexes of rhodium and iridium were obtained by the double Ci-H activation of CH₂ groups in N heterocycles (see structure: C gray, N blue, Cl green, Ir pink). The reaction constitutes a valuable approach to new carbene ligands with unprecedented architectures. Copyright

Full text of DOI:10.1002/anie.201101806

Communications
DOI: 10.1002/anie.201101806
N-Heterocyclic Carbenes
3
À
Double C H Bond Activation of C(sp )H2 Groups for the Preparation
of Complexes with Back-to-Back Bisimidazolinylidenes**
Amparo Prades, Macarena Poyatos, Josꢀ A. Mata, and Eduardo Peris*
It is now very well recognized that N-heterocyclic carbenes
(NHCs) have found an important place in the fields of
organometallic chemistry and homogeneous catalysis.
Despite their increasing popularity, the coordination of
NHCs to transition-metal complexes is limited to a small
number of procedures, most of which require the use of
azolium salts as carbene precursors.[1] Therefore, the design of
new NHCs with novel topologies and improved chemical
properties is basically restricted to the availability of their
related azolium precursors. In the last few years, several
attempts to obtain NHCs from nontraditional precursors have
been reported. For example, Hahn and co-workers reported
an elegant procedure in which the NHC ligand is generated
through a template-controlled cyclization of b-functionalized
isocyanides. This approach provides ready access to NH,NH-
substituted “protic” NHC ligands.[2]
Because we are interested obtaining NHC-based com-
plexes with new topologies and improved catalytic activities,
we are very interested in finding alternative routes to NHC
complexes to widen the library of ligand precursors used, so
that we can have access to NHC complexes that may not be
available from azolium salts. In our search for improved
homogeneous catalysts, we prepared a series of homo- and
heterodimetallic complexes based on the ligand 1,2,4-trime-
thyltriazolyldiylidene.[3] The use of heterodimetallic com-
plexes of Ir/Rh and Ir/Pd enabled us to study their activity in
catalytic tandem processes in which each metal mediated a
mechanistically distinct reaction.[4] Some related dimetallic
systems in which the linker between the two metal fragments
is a bis(benzimidazolylidene) ligand were also reported by
Bielawski and co-workers,[5] although in their case only
homometallic complexes were described. The bisylidene
ligands described by Bielawski and co-workers and by us
have in common their preparation from bisazolium salts that
were readily accessed by known literature procedures.[6]
Apart from these examples, despite a close review of the
literature dealing with N heterocycles, we have not found any
further simple bisazolium salts that are suitable precursors to
bisylidene linkers between two metal centers. In this regard,
we believe that the design of new discrete multitopic carbenes
that are poised to bind multiple transition metals may require
new strategies for NHC generation.
À
In certain cases, carbenes can be generated by double C
H bond activation of C(sp3)H2 groups.[7–9] Although most of
the known examples involve the transformation of cyclic
ethers into traditional Fischer carbenes,[8] there is one
example in which the double C(sp3) dehydrogenation of a
cyclic H2C(NRCH2)2 species afforded an Ru–NHC com-
plex.[9] To check the general applicability of this alternative
synthetic route to carbene complexes, we investigated the
possibility that the N heterocycle 1 (Scheme 1) might react
with iridium or rhodium complexes if a hydrogen acceptor
was added to the reaction medium to facilitate the formation
of the N-heterocyclic carbene.
Scheme 1. Reaction of 1 with rhodium–diolefin and iridium–diolefin
complexes in CH3CN at reflux to give carbene complexes 2 and 3 with
formation of the corresponding monoolefin. Bn=benzyl, NBE=nor-
bornene.
To our surprise, the reaction of 1 (R = Bn) with [{MCl-
(cod)}2] (M = Rh and Ir) in CH3CN at reflux afforded 2a and
3 in moderate yields (38% for 2a, 45% for 3) without the
need to add a hydrogen acceptor. GC analysis of the reaction
products revealed that cyclooctene (COE) was formed, and
that the amount of the monoolefin corresponded to the yields
observed for 2a and 3. This result clearly suggests that 1,5-
cyclooctadiene (COD) acted as an internal hydrogen acceptor
and was released from the metal as COE. Cyclooctane was
not detected among the reaction products, so COD is only
capable of accepting one molecule of H2. For stoichiometric
reasons, the reaction yield is restricted to a maximum of 50%,
unless an external hydrogen acceptor is added to the reaction
medium. We performed several experiments in which we
[*] A. Prades, Dr. M. Poyatos, Dr. J. A. Mata, Prof. E. Peris
Departamento de Quꢀmica Inorgꢁnica y Orgꢁnica
Universitat Jaume I
Avenida Sos Baynat, 12071 Castellꢂn (Spain)
Fax: (+34)96-472-8214
E-mail: eperis@qio.uji.es
[**] We gratefully acknowledge financial support from the MEC of Spain
(CTQ2008-04460/BQU) and Bancaixa (P1.1B2010-02, P1.1A2008-
02). We also thank the Ramꢂn y Cajal program (M.P.). A.P. thanks
the Ministerio de Ciencia e Innovaciꢂn for a fellowship. We are
grateful to the Serveis Centrals d’Instrumentaciꢂ Cientꢀfica (SCIC)
of the Universitat Jaume I for providing us with spectroscopic and
X-ray facilities.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201101806.
7666
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7666 –7669

added different classical hydrogen acceptors, such as tert-
butylethene, acetone, or cyclohexanone, but we did not
improve the yields of 2a and 3. We also did not observe the
formation of the corresponding hydrogenated partners of the
hydrogen acceptors used. However, when we added an extra
amount of COD (5 equiv), the reaction products were
obtained in yields higher than 50% (60% for 2a and 65%
for 3), which implies that the chelating nature of the hydrogen
acceptor is needed to facilitate the process. Again, the
formation of COE matched perfectly with the yields of the
complexes. To confirm that chelating olefins are needed to
promote the process, we carried out a reaction with [{RhCl-
(nbd)}2] as the metal source and also added 2,5-norborna-
diene (NBD; 5 equiv) to the reaction mixture. Under the
same reaction conditions as those described above, we
obtained the nbd-based complex 2b in high yield (67%). In
contrast, when the reaction was performed with [{MCl-
(coe)2}2] (M = Rh, Ir), in the presence of excess COE,
metalation of the diamino heterocycle did not take place.
Although we did not perform a detailed mechanistic
study, we believe that the formation of 2a and 3 from
Scheme 2. Expected steps involved in the formation of 2a and 3 from
1 and [{MCl(cod)}2].
toluene at reflux, the bimetallic compound 7 with a bis-NHC
bridging ligand was obtained. Compound 7 can also be
obtained from 6 (Scheme 3b).
the reaction of 1 and [{MCl(cod)}2] (M = Rh and Ir)
implies the steps depicted in Scheme 2. First, the
3
À
oxidative addition of a C(sp ) H bond to the metal
should afford an MIII hydride A, which may evolve to
À
intermediate B through olefin insertion into the M H
bond and a elimination of the N heterocycle. The
final step would involve the release of the monoolefin
by reductive elimination and the coordination of
COD to form the final product (2a or 3). Although we
do not have experimental proof of the formation of
the intermediates A and B, we believe that their
formation is strongly supported by previously
reported results. For example, Shaw and co-workers
reported the formation of a pincer-carbene (PCP)
À
iridium complex by C H oxidative addition and
subsequent a elimination with the release of dihydro-
gen.[10] In the same context, Crabtree and co-workers
performed a detailed mechanistic study in which they
showed that the formation of an iridium(III) carbene
from a 2-dimethylaminopyridine ligand proceeded
through a reversible a elimination, which in their case
proved to be extremely facile.[11] The oxidative
addition of neutral diamino heterocycles is also a
Scheme 3. Synthesis of the tetraazabicyclooctane 4 and its conversion into
mono- and biscarbene complexes.
known process, as recently described by Hahn and co-
All compounds 5–7 were characterized by means of NMR
spectroscopy and mass spectrometry (see the Supporting
Information for full details). The 1H NMR spectra of 5 and 6
revealed that the ligand is coordinated to only one metal
fragment, as shown by the two doublets due to the inequi-
valent NCH2N protons of the uncoordinated ring. The
13C NMR spectra confirmed that metalation had occurred,
with signals at d = 214 (doublet, 1JRh,C = 47 Hz; 5) and
workers.[12]
We believed that the unprecedented reaction shown in
Scheme 1 should have some important implications for the
formation of new NHC complexes that are not accessible
from the classical azolium carbene precursors. In this regard,
the tetraazabicyclooctane 4,[13] can be obtained in high yield
by the reaction shown in Scheme 3a. The reaction of 4 with
[{MCl(cod)}2] (M = Rh and Ir) in acetonitrile at reflux
afforded the corresponding mono-NHC species 5 and 6,
although in very low yields (< 30%). When the same reaction
was performed in the presence of COD, the products were
obtained in much higher yields (55% for 6 and 65% for 5)
with the concomitant production of the same amount of COE.
When the reaction with [{IrCl(cod)}2] was carried out in
À
207.7 ppm (6) assigned to the M Ccarbene carbon atom. The
1H NMR spectrum of 7 confirmed the twofold symmetry of
the compound. The signals due to the two NCH2N protons
disappeared, which indicates that double metalation had
occurred. In the 13C NMR spectrum, the signal due to the two
À
equivalent Ir Ccarbene carbon atoms appeared at d =
211.2 ppm.
Angew. Chem. Int. Ed. 2011, 50, 7666 –7669
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7667

Communications
Figure 2. Molecular structure of 7 (ellipsoids at 30% probability). All
hydrogen and solvent (hexane) atoms have been omitted for clarity.
Selected distances [ꢃ] and angles [8]: Ir1–C1 2.017(7), Ir1–Cl1
2.3509(19), Ir1–C46 2.107(7), Ir1–C47 2.114(8), Ir1–C42 2.202(7), Ir1–
C43 2.211(7), Ir2–C3 2.018(7), Ir2–Cl2 2.345(2), Ir2–C34 2.078(9), Ir2–
C33 2.087(9), Ir2–C38 2.202(7), Ir2–C37 2.204(7); C1-Ir1-Cl1 90.86(18),
C3-Ir2-Cl2 92.47(19).
average angle of 96.58 with respect to the metal coordination
planes.
Compounds 5, 6, and 7 were transformed into the
corresponding carbonylated complexes by treatment with
CO in dichloromethane (Scheme 3). Each carbonylation
yielded two rotamers that we could not separate and that
we attributed to the restricted rotation of the carbene about
Figure 1. Two perspectives of the molecular structure of 6 (ellipsoids
at 30% probability). All hydrogen atoms have been omitted for clarity.
Selected distances [ꢃ] and angles [8]: Ir1–C1 2.021(5), Ir1–Cl1
2.3541(17), Ir1–C40 2.093(6), Ir1–C33 2.096(6), Ir1–C36 2.192(6), Ir1–
C37 2.195(6); C1-Ir1-Cl1 90.31(16).
À
the M Ccarbene bond. The most stable situation arises from the
orientation of the two benzyl groups close to one of the
smaller CO ligands; two different rotamers are expected for
complexes 8 and 9 (Scheme 4a). Variable-temperature NMR
spectroscopic experiments (in [D8]toluene between 25 and
808C) together with spin-polarization-transfer (SPT) NMR
spectroscopic experiments enabled us to confirm that these
two rotamers do not interconvert. In the case of complex 10,
three different rotamers are expected (Scheme 4b). All three
The molecular structures of 6 and 7 were confirmed by
means of X-ray diffraction studies. The structure of com-
pound 6 (Figure 1)[14] confirms the coordination of the bicyclic
N-heterocyclic carbene to the IrI center. One chloride and one
cod ligand complete the coordination sphere about
the metal center. The Ir–Ccarbene distance is 2.021 ꢀ.
Because the bicyclic ligand is folded along the
À
connecting C(2) C(4) bond, the four N-benzyl
groups occupy the less sterically strained orienta-
tion, on the outside of the folded structure. The two
benzyl groups at the metalated ring are closer to the
chloride ligand and thus avoid the bulkier cod
ligand.
The molecular structure of 7 (Figure 2) confirms
the bimetallic nature of the compound. The mole-
cule has pseudo-C2v symmetry. The bicyclic bis-
NHC bridges two iridium fragments, which each
complete their coordination sphere with a chloride
and a cod ligand. The Ir–Ccarbene distances are
approximately 2.02 ꢀ. Owing to the angular
nature of the bicyclic ligand, the through-space
distance between the two metal centers is 7.166 ꢀ,
whereas the through-ligand distance is 8.5 ꢀ. As
observed for the monometallic complex 6, the N-
benzyl groups point toward the chloride ligands.
Scheme 4. Rotamers observed for a) the monocarbene complexes 8 (M=Rh) and
The heterocyclic biscarbene ligand displays an 9 (M=Ir), and b) the biscarbene complex 10.
7668 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7666 –7669

were detected in the corresponding 1H NMR spectrum.
Interestingly, this situation does not apply for the cod-
substituted complexes 5, 6, and 7, for which the complexes
with the benzyl groups close to the chloride ligand are
strongly preferred (see the molecular diagrams in Figures 1
and 2) over the related orientation of the benzyl groups close
to the bulkier cod ligand. This situation arises from the
Enders, H. Gielen, J. Organomet. Chem. 2001, 617, 70 – 80; T.
Weskamp, V. P. W. Bꢁhm, W. A. Herrmann, J. Organomet.
Chem. 2000, 600, 12 – 22; F. E. Hahn, M. C. Jahnke, Angew.
Chem. 2008, 120, 3166 – 3216; Angew. Chem. Int. Ed. 2008, 47,
3122 – 3172.
[2] A. Flores-Figueroa, O. Kaufhold, K. O. Feldmann, F. E. Hahn,
Dalton Trans. 2009, 9334 – 9342; A. Flores-Figueroa, T. Pape,
K. O. Feldmann, F. E. Hahn, Chem. Commun. 2010, 46, 324 –
326; A. Flores-Figueroa, T. Pape, J. J. Weigand, F. E. Hahn, Eur.
J. Inorg. Chem. 2010, 2907 – 2910; F. E. Hahn, C. G. Plumed, M.
Mꢂnder, T. Lꢂgger, Chem. Eur. J. 2004, 10, 6285 – 6293; O.
Kaufhold, A. Flores-Figueroa, T. Pape, F. E. Hahn, Organo-
metallics 2009, 28, 896 – 901; F. E. Hahn, M. Tamm, J. Organo-
met. Chem. 1993, 456, C11 – C14.
[3] E. Mas-Marzꢃ, J. A. Mata, E. Peris, Angew. Chem. 2007, 119,
3803 – 3805; Angew. Chem. Int. Ed. 2007, 46, 3729 – 3731.
[4] A. Zanardi, R. Corberan, J. A. Mata, E. Peris, Organometallics
2008, 27, 3570 – 3576; A. Zanardi, J. A. Mata, E. Peris, J. Am.
Chem. Soc. 2009, 131, 14531 – 14537; A. Zanardi, J. A. Mata, E.
Peris, Chem. Eur. J. 2010, 16, 13109 – 13115; A. Zanardi, J. A.
Mata, E. Peris, Chem. Eur. J. 2010, 16, 10502 – 10506.
[5] A. J. Boydston, C. W. Bielawski, Dalton Trans. 2006, 4073 – 4077;
A. J. Boydston, J. D. Rice, M. D. Sanderson, O. L. Dykhno, C. W.
Bielawski, Organometallics 2006, 25, 6087 – 6098; D. M. Khra-
mov, A. J. Boydston, C. W. Bielawski, Angew. Chem. 2006, 118,
6332 – 6335; Angew. Chem. Int. Ed. 2006, 45, 6186 – 6189; M. D.
Sanderson, J. W. Kamplain, C. W. Bielawski, J. Am. Chem. Soc.
2006, 128, 16514 – 16515.
À
restricted rotation of the benzyl groups about the N C bonds
owing to the special topology of the ligand. The four benzyl
groups are forced to point away from the inner cavity formed
by the two planes of the central bicyclic heterocycle. For the
monoazole-derived complexes 2 and 3, the situation is
completely different because the benzyl groups are free to
À
rotate about the N C bonds. Thus, the benzyl groups can
À
adopt the two possible orientations by free N C rotation,
À
whereas M C rotation is still restricted.
The existence of the different isomers does not change the
local symmetry about the metal centers nor the electron-
donating power of the ligands. For this reason, the IR spectra
of 8, 9, and 10 display only two CO stretching bands, as
expected for the cis orientation of the carbonyl ligands. For
the monometallic compounds, these bands appear at 2077 and
1995 cmÀ1 for 8 and 2063 and 1978 cmÀ1 for 9 and correlate
well with those of analogous previously described saturated
mono-NHC complexes of Rh[15] and Ir.[16] The CO stretching
bands of the bimetallic compound 10 appear at 2071 and
1986 cmÀ1 and thus indicate that the double coordination of
the biscarbene ligand slightly lowers the electron-donating
capacity of the carbene with respect to that of its mono-
coordinated form.
[6] T. J. Curphey, K. S. Prasad, J. Org. Chem. 1972, 37, 2259 – 2266;
O. Guerret, S. Sole, H. Gornitzka, M. Teichert, G. Trinquier, G.
Bertrand, J. Am. Chem. Soc. 1997, 119, 6668 – 6669.
[7] H. Werner, Angew. Chem. 2010, 122, 4822 – 4837; Angew. Chem.
Int. Ed. 2010, 49, 4714 – 4728; E. Clot, J. Y. Chen, D. H. Lee, S. Y.
Sung, L. N. Appelhans, J. W. Faller, R. H. Crabtree, O. Eisen-
stein, J. Am. Chem. Soc. 2004, 126, 8795 – 8804.
In conclusion, we have shown that saturated N-hetero-
cyclic carbenes of rhodium and iridium can be readily
prepared by the simple treatment of cyclic neutral N hetero-
cycles with widely used [{MCl(diolefin)}2] metal sources. The
reactions can be triggered by the addition of an external
amount of 1,5-cyclooctadiene or 2,5-norbornadiene, which act
as hydrogen acceptors, and offer several advantages over the
use of regular preligand-activating agents when azolium salts
are used. This method to generate NHC complexes enables
the ready preparation of new metal compounds with sophis-
ticated architectures, as demonstrated with complexes 5–10.
The possibility of preparing dimetallic complexes suggests
that we may soon find a way to prepare heterodimetallic
species, potentially useful for the design of new tandem
catalytic processes. Studies to extend the use of our new
biscarbene ligand to other metal species and explore the
catalytic properties of the resulting complexes are underway.
[8] J. N. Coalter III, G. Ferrando, K. G. Caulton, New J. Chem. 2000,
24, 835 – 836; E. Gutꢄerrez-Puebla, A. Monge, M. C. Nicasio, P. J.
Pꢅrez, M. L. Poveda, E. Carmona, Chem. Eur. J. 1998, 4, 2225 –
2236; H. F. Luecke, B. A. Arndtsen, P. Burger, R. G. Bergman, J.
Am. Chem. Soc. 1996, 118, 2517 – 2518; C. Slugovc, K. Mereiter,
S. Trofimenko, E. Carmona, Angew. Chem. 2000, 112, 2242 –
2244; Angew. Chem. Int. Ed. 2000, 39, 2158 – 2160.
[9] V. M. Ho, L. A. Watson, J. C. Huffman, K. G. Caulton, New J.
Chem. 2003, 27, 1446 – 1450.
[10] H. D. Empsall, E. M. Hyde, R. Markham, W. S. McDonald,
M. C. Norton, B. L. Shaw, B. Weeks, J. Chem. Soc. Chem.
Commun. 1977, 589 – 590.
[11] D. H. Lee, J. Y. Chen, J. W. Faller, R. H. Crabtree, Chem.
Commun. 2001, 213 – 214.
[12] T. Kꢁsterke, T. Pape, F. E. Hahn, J. Am. Chem. Soc. 2011, 133,
2112 – 2115.
[13] A. T. Nielsen, R. A. Nissan, A. P. Chafin, R. D. Gilardi, C. F.
George, J. Org. Chem. 1992, 57, 6756 – 6759; W. M. Koppes, M.
Chaykovsky, H. G. Adolph, R. Gilardi, C. George, J. Org. Chem.
1987, 52, 1113 – 1119.
[14] CCDC 816602 (6) and 816603 (7) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Received: March 14, 2011
Revised: May 11, 2011
Published online: June 29, 2011
[15] K. Denk, P. Sirsch, W. A. Herrmann, J. Organomet. Chem. 2002,
649, 219 – 224.
[16] Y. H. Chang, C. F. Fu, Y. H. Liu, S. M. Peng, J. T. Chen, S. T. Liu,
Dalton Trans. 2009, 861 – 867.
À
Keywords: C H activation · dimetallic compounds · iridium ·
N-heterocyclic carbenes · rhodium
.
[1] D. J. Cardin, B. Cetinkaya, M. F. Lappert, Chem. Rev. 1972, 72,
545 – 574; E. Peris, Top. Organomet. Chem. 2007, 21, 83 – 116; D.
Angew. Chem. Int. Ed. 2011, 50, 7666 –7669
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7669