Synthesis and reactivity of cationic triruthenium clusters derived from 2-methyl- and 4-methylpyrimidines: From conventional cyclometalated ligands to novel types of N-heterocyclic carbenes

Article abstract of DOI:10.1002/chem.201203815

The methylation of the uncoordinated nitrogen atom of the cyclometalated triruthenium cluster complexes [Ru₃(μ-H)(μ- κ²N¹,C⁶-2-Mepyr)(CO)₁₀] (1; 2-MepyrH=2-methylpyrimidine) and [Ru₃(μ-H)(μ- κ²N¹,C⁶-4-Mepyr)(CO)₁₀] (9; 4-MepyrH=4-methylpyrimidine) gives two similar cationic complexes, [Ru ₃(μ-H)(μ-κ²N¹,C⁶-2,3- Me₂pyr)(CO)₁₀]⁺(2⁺) and [Ru ₃(μ-H)(μ-κ²N¹,C⁶-3,4- Me₂pyr)(CO)₁₀]⁺ (9⁺), respectively, whose heterocyclic ligands belong to a novel type of N-heterocyclic carbenes (NHCs) that have the C_carbene atom in 6-position of a pyrimidine framework. The position of the C-methyl group in the ligands of complexes 2 ⁺ (on C²) and 9⁺ (on C⁴) is of key importance for the outcome of their reactions with K[N(SiMe₃) ₂], K-selectride, and cobaltocene. Although these reagents react with 2⁺ to give [Ru₃(μ-H)(μ-κ²N ¹,C⁶-2-CH₂-3-Mepyr)(CO)₁₀] (3; deprotonation of the C²-Me group), [Ru₃(μ-H) (μ₃-κ³N¹,C⁵,C ⁶-4-H-2,3-Me₂pyr)(CO)₉] (4; hydride addition at C⁴), and [Ru₆(μ-H)₂{μ₆- κ⁶N¹,N^1′,C⁵,C ^5′,C⁶,C^6′-4,4′-bis(2,3- Me₂pyr)}(CO)₁₈] (5; reductive dimerization at C ⁴), respectively, similar reactions with 9⁺ have only allowed the isolation of [Ru₃(μ-H)(μ₃- κ²N¹,C⁶-2-H-3,4-Me₂pyr)(CO) ₉] (11; hydride addition at C²). Compounds 3 and 11 also contain novel six-membered ring NHC ligands. Theoretical studies have established that the deprotonation of 2⁺ and 9⁺ (that have ligand-based LUMOs) are charge-controlled processes and that both the composition of the LUMOs of these cationic complexes and the steric protection of their ligand ring atoms govern the regioselectivity of their nucleophilic addition and reduction reactions. A methyl makes the difference: New N-heterocyclic carbenes (NHCs) having the C_carbene atom in 6-position of a pyrimidine framework have been identified as ligands of cationic triruthenium cluster complexes (see scheme). These ligands can be modified through processes that are governed by atomic charges (deprotonation) or by a combination of two factors: the composition of the LUMO of the parent complex and the steric protection of its ligand ring atoms (hydride addition and chemical reduction). Copyright

Full text of DOI:10.1002/chem.201203815

DOI: 10.1002/chem.201203815
Synthesis and Reactivity of Cationic Triruthenium Clusters Derived from
2-Methyl- and 4-Methylpyrimidines: From Conventional
Cyclometalated Ligands to Novel Types of N-Heterocyclic Carbenes
Javier A. Cabeza,*[a] Pablo Garcꢀa-ꢁlvarez,[a] Enrique Pꢂrez-CarreÇo,[b] and
Vanessa Pruneda[a]
Abstract: The methylation of the unco-
ordinated nitrogen atom of the cyclo-
metalated triruthenium cluster com-
plexes [Ru
CO)10] (1; 2-MepyrH=2-methylpyrimi-
(on C4) is of key importance for the
outcome of their reactions with K[N-
tions with 9+ have only allowed the
isolation of [Ru3A
(m-H)(m3-k2N1,C6-2-H-
CTHUNGTRENNUNG
A
3,4-Me2pyr)(CO)9] (11; hydride addi-
tion at C2). Compounds 3 and 11 also
contain novel six-membered ring NHC
ligands. Theoretical studies have estab-
lished that the deprotonation of 2+ and
9+ (that have ligand-based LUMOs)
are charge-controlled processes and
that both the composition of the
LUMOs of these cationic complexes
and the steric protection of their ligand
ring atoms govern the regioselectivity
of their nucleophilic addition and re-
duction reactions.
3ACHTUNGTRENNUNG
(m-H)(m-k2N1,C6-2-Mepyr)(-
CTHUNGTRENNUNG
dine) and [Ru
pyr)(CO)10]
3A
(9;
CTHUNGTRENNUNG
methylpyrimidine) gives two similar
cationic complexes, [Ru3A(m-H)(m-
k2N1,C6-2,3-Me2pyr)(CO)10]+(2+) and
[Ru3A
(m-H)(m-k2N1,C6-3,4-Me2pyr)-
E
(4; hydride addition at C4), and [Ru
CHTUNGTNER(NUNG m-
H)2{m6-k6N1,N1’,C5,C5’,C6,C6’-4,4’-bis(2,3-
Me2pyr)}(CO)18] (5; reductive dimeri-
zation at C4), respectively, similar reac-
CHTUNGTRENNUNG
(CO)10]+ (9+), respectively, whose het-
erocyclic ligands belong to a novel type
of N-heterocyclic carbenes (NHCs)
that have the Ccarbene atom in 6-position
of a pyrimidine framework. The posi-
tion of the C-methyl group in the li-
gands of complexes 2+ (on C2) and 9+
Keywords: cluster compounds · N-
heterocyclic carbenes · pyrimidyli-
denes
· reductive dimerization ·
ruthenium
Introduction
and pyrimidine[7] that contain two nitrogen atoms in a six-
membered ring.
In the last few years, N-heterocyclic carbenes (NHCs) de-
rived from aromatic heterocycles having one nitrogen atom
in a six-membered ring, such as pyridylidenes, quinolyli-
denes, etc., have become important ligands in coordination
chemistry and catalysis.[1–3] In this field, we have reported
that pyrid-2-ylidenes and quinol-2-ylidenes can be generated
in situ by deprotonation of N-methylpyridinium[4,5] and N-
methylquinolinium[5] cations and that they can be trapped
by [Ru3(CO)12]. However, this methodology (in situ depro-
tonation of “inium” cations) does not work for NHCs for-
mally derived from aromatic diazines, such as pyrazine[6]
Notwithstanding the great potential that metal complexes
containing aromatic diazine-derived NHC ligands may have
in coordination chemistry and catalysis and despite the first
synthesis of such complexes was published as early as 1985
by Crociani et al., who reported the N-protonation of 2-
pyrazyl- and 2-pyrimidylpalladium(II) complexes,[8] their
chemistry has not blossomed until very recently: Lassaletta
et al. reported in 2010 the synthesis of Rh and Ir complexes
containing NHCs derived from phthalazine by treatment of
N-alkylphthalazinium cations with [M2ACHTUNTGRENNUG(m-OtBu)2ACHTUNGTRENNUNG(COD)2]
(M=Rh, Ir);[9] Ganter et al. described in 2010 that the oxi-
À
dative addition of the C X (X=Cl, I) bond of N-alkyl-2-
halopyrimidinium cations to palladium(0) complexes gives
cationic palladium(II) derivatives that contain an N-alkyl-
pyrimid-2-ylidene ligand;[10] and our group has recently
shown that triruthenium complexes containing C,N-cyclome-
talated bridging ligands derived from pyrazine,[6] quinoxa-
line,[6] and pyrimidine,[7] can be methylated at their uncoor-
dinated nitrogen atom to give cationic products containing
novel NHC ligands that can be subsequently modified by re-
actions with nucleophiles, which attack a ring carbon atom,
and reducing reagents, which form short-living radical spe-
[a] Prof. J. A. Cabeza, Dr. P. Garcꢀa-ꢁlvarez, Dr. V. Pruneda
Departamento de Quꢀmica Orgꢂnica e Inorgꢂnica-IUQOEM
Universidad de Oviedo-CSIC, 33071 Oviedo (Spain)
Fax : (+34)985103446
E-mail: jac@uniovi.es
[b] Dr. E. Pꢃrez-CarreÇo
Departamento de Quꢀmica Fꢀsica y Analꢀtica
Universidad de Oviedo, 33071 Oviedo (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201203815.
3426
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436

FULL PAPER
cies that spontaneously undergo dimerization. Representa-
tive examples of cluster complexes of this family involving
NHC ligands derived from pyrimidine are shown in
Scheme 1.
those of their cationic precursors. A DFT theoretical study
has rationalized the experimental results.
Results and Discussion
Compounds derived from 2-methylpyrimidine: With the ob-
jective of preparing metal complexes having pyrimidylidene-
derived ligands in which their Ccarbene atom is not between
the N atoms in the pyrimidine skeleton (2-position) and
knowing that the pyrimid-2-ylidene-derived complexes
shown in Scheme 1 originally arise from a reaction of
[Ru3(CO)12] with pyrimidine (pyrH),[7] we first decided to
use [Ru3(CO)12] and 2-methylpyrimidine (2-MepyrH) as
precursors because the methyl group of the latter hampers
the metalation of its C2 carbon atom and, therefore, its reac-
tion with [Ru3(CO)12] could lead to an N1,C6-cyclometalated
product.
The N1,C6-cyclometalted triruthenium complex [Ru
3ACHTUNGTRENNUNG(m-
H)(m-k2N1,C6-2-Mepyr)(CO)10] (1) was satisfactorily pre-
pared by treating [Ru3(CO)12] with 2-methylpyrimidine
(THF, heating to reflux, 5 h). Its reaction with methyl tri-
flate (dichloromethane, heating to reflux, 4 h) allowed the
synthesis of the N-methyl derivative [Ru3ACHTUNGTRENNUNG
(m-H)(m-k2N1,C6-
OTf=triflate)
2,3-Me2pyr)(CO)10]CF3SO3
(Scheme 2).
(2·OTf;
Scheme 1. Syntheses of some tri- and hexaruthenium clusters containing
*
NHC ligands derived from pyrimidine (examples taken from ref. [7];
CO).
=
Concerning pyrimidylidenes, it should be noted that metal
complexes with NHCs formally arising from monodeproto-
nation at the C2 carbon atom of N1,N3-dialkylated pyrimi-
dines remain unknown, although a theoretical work has sug-
gested their feasibility,[11] and, apart from the palladium(II)
complexes mentioned in the above paragraph[8,10] and the
cluster complexes depicted in Scheme 1,[7] only a few other
unsaturated NHCs related to pyrimidylidenes have been re-
ported, namely, a diazadiborine NHC ligand described by
Roesler and coworkers,[12] and some NHCs based on a pyri-
midine-4,6-dione core reported by the Lavigne et al.[13]and
Bielawski et al.[14]
Scheme 2. Synthesis of compounds 1 and 2·OTf.
The composition of 1 and 2·OTf was established by their
microanalyses and FAB mass spectra. Their 1H and
13C{1H} NMR spectra confirmed the presence of one hy-
dride, a 1,6-cyclometalated 2-methylpyrimidine, and ten CO
ligands; however, as often observed for complexes with
pyrid-4-ylidene ligands,[2a,15] the 13C NMR shift of the metal-
bound heterocyclic C atoms of 1 and 2+ (d=187.1 and
188.0 ppm, respectively) was not a clear indicator of the car-
bene character of their ligands. The structures proposed for
these compounds in Scheme 3 are also supported by the nCO
region of their solution IR spectra, which are comparable to
With only one exception,[7] all hitherto known NHC li-
gands derived from pyrimidines have the Ccarbene atom locat-
ed in 2-position of the pyrimidine skeleton, that is, between
the two nitrogen atoms.
We now report a family of metal complexes containing
NHC ligands derived from pyrimidines whose Ccarbene atom
is not the C2 carbon atom of the pyrimidine skeleton. In par-
ticular, this paper describes the synthesis of cationic triru-
thenium clusters containing 2-methyl- or 4-methylpyrimid-6-
ylidene ligands and that the position of the C-methyl sub-
stituent on the N-heterocyclic ring of these complexes has a
decisive influence in the outcome of their reactions with
bases and reducing reagents, which lead to neutral products
with ligands (sometimes new NHCs) that are different from
those previously reported for the related complexes [Ru
(m-
R
H)(m-k2N1,C6-pyr)(CO)10] (pyrH=pyrimidine) and [Ru
H)(m-k2N1,C6-3-Mepyr)(CO)10]+,[7] which do not contain C-
methyl groups, the absorptions of the neutral complexes
being observed at lower wavenumbers than those of the cat-
ionic compounds.
Chem. Eur. J. 2013, 19, 3426 – 3436
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3427

J. A. Cabeza et al.
with imidazol-2-ylidene[16] or pyrid-2-ylidene[4] ligands (2.1–
2.2 ꢅ). Therefore, the heterocyclic ligand of 2+ is clearly an
NHC.
The reaction of 2·OTf with K[N
philic and strong base, resulted in deprotonation of its C2-
Me group, leading to the neutral complex [Ru3A(m-H)(m-
ACHTUNGTREN(NGNU SiMe3)2], a non-nucleo-
CHTUNGTRENNUNG
k2N1,C6-2-CH2-3-Mepyr)(CO)10] (3; Scheme 3), which was
isolated in 75% yield as a thermally unstable garnet solid (it
slowly decomposed at room temperature). A small amount
of compound 1 (ꢀ4%), arising from the abstraction of the
N-methyl group by the base, was also formed in this reac-
tion.
The limited thermal stability of compound 3 prevented
the preparation of X-ray-quality crystals, but it was ade-
quately characterized by other analytical and spectroscopic
techniques. In particular, the CH2 group is observed as two
doublets at d=3.69 and 3.44 ppm (J=3.2 Hz) in the
1H NMR spectrum and as a singlet at d=72.3 ppm in the
13C{1H} NMR spectrum. The 13C shift of the metalated C
atom is d=187.5 ppm. In order to know the ligand intera-
tomic distances, the molecular structure of this complex was
calculated by DFT methods (Figure 3). The shortest ligand
Scheme 3. Synthesis of compound 3. A small amount of 1 (4%) is also
formed in this reaction.
The four-electron donor ligand present in 2+ can be con-
sidered as a resonance hybrid of the three canonical forms
depicted in Figure 1, one being neutral (B) and two (A and
C) being zwitterionic. In order to establish the contribution
of these resonance forms to the ligand structure and as no
Figure 1. Resonance structures for the four-electron donor heterocyclic
ligand of 2+.
crystals of 2·OTf suitable for X-ray diffraction analysis were
available, the structure of the cation 2+ was calculated by
DFT methods (Figure 2). An inspection of the ligand intera-
tomic distances confirmed an important contribution of res-
onance structure B, because the shortest ligand edges of 2+
1
2
4
5
À
À
(N
C
and C C ) correspond to those having double
6
À
bonds in B. The C Ru distance, 2.067 ꢅ, is shorter than the
CNHC Ru distances found in triruthenium carbonyl clusters
À
Figure 3. DFT-optimized structure of 3 with selected interatomic distan-
ces in [ꢅ].
1
6
À
edge, N C , is that attached to the metal atoms and is
1
2
2
À
À
flanked by two long distances on one side, N C and C
N , and one on the other, C C . Curiously, the N C and
3
5
6
3
4
À
À
4
5
À
C
C distances are very similar and shorter than expected
6
À
for a single bond. The C Ru distance, 2.113 ꢅ, is longer
than that calculated for 2+, 2.067 ꢅ. Therefore, although the
three-electron donor heterocyclic ligand of compound 3 can
be considered as a carbene (represented by resonance struc-
tures D and E in Figure 4), the available structural data sug-
gest that the bonding distribution is best (but not only) rep-
resented by the non-carbenic structure F in Figure 4.
C-Alkyl deprotonation of metal-free C-alkyl-N-heterocy-
clic “inium” salts has been previously used to make useful
intermediates in heterocyclic syntheses,[17] but C-alkyl depro-
Figure 2. DFT-optimized structure of the cationic complex 2+ with select-
ed interatomic distances in [ꢅ].
3428
www.chemeurj.org
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436

Synthesis and Reactivity of Cationic Triruthenium Clusters
FULL PAPER
Figure 4. Carbenic (D, E) and non-carbenic (F) electron distributions for
the three-electron donor ligand of 3.
tonation of their transition-metal complexes has only been
reported for triruthenium clusters derived from C-methyl-
pyrazinium salts.[18]
When 2·OTf was treated with potassium tris(sec-butyl)hy-
dridoborate (K-selectride), a 1:3:15 mixture (1H NMR iden-
tification of the crude reaction solution) of compounds 1, 3,
and [Ru3ACHTUNGTRENNUNG
(m-H)(m3-k3N1,C5,C6-4-H-2,3-Me2pyr)(CO)9] (4) was
immediately formed, indicating that K-selectride, although
acting as demethylating (to give compound 1) and deproto-
nating agent (to give compound 3), prefers to act as a hy-
dride donor (to give compound 4). A chromatographic
workup allowed the isolation of 4 in 59% yield (Scheme 4).
Figure 5. X-ray diffraction structure of 4 (only one of the two analogous
but independent molecules present in the crystal is shown). Selected
À
À
À
bond lengths [ꢅ]: Ru1 Ru2 2.7342(7), Ru1 Ru3 2.8434(7), Ru2 Ru3
À
À
À
À
2.8734(7), Ru1 N1 2.148(5), Ru2 C1 2.069(7), Ru(3) C1 2.312(6), Ru3
À
À
À
À
C2 2.338(6), N1 C1 1.398(9), N1 C4 1.317(8), N2 C3 1.475(9), N2 C4
1.324(9), N2 C5 1.458(8), C1 C2 1.397(9), C2 C3 1.513(9), C4 C6
1.513(9).
À
À
À
À
Scheme 4. Synthesis of compound 4. Small amounts of compounds 1
(5%) and 3 (15%) are also formed in this reaction.
Figure 6. Resonance structures for the five-electron donor ligand of 4.
1H NMR spectra observed at high temperatures is proposed
in Scheme 5. Such a process interconverts the two asymmet-
ric enantiomers of compound 4 through a symmetric (CS) in-
termediate species, in which the ligand ring is perpendicular
to the Ru3 plane.
The solid-state molecular structure of compound 4 was
determined by X-ray diffraction (Figure 5). Clearly, the mol-
ecule results from a nucleophilic addition of a hydride to
the ligand C4 carbon atom of 2+. This should place a double
5
6
À
À
bond in the ligand C C edge (C2 C1 in Figure 5) that sub-
sequently displaces one CO from the Ru(CO)4 fragment to
give the final product after an easy rearrangement of the re-
sulting ligand over the Ru3 triangle.[6] The ligand ring is
roughly planar, forming a dihedral angle of 59.7(1)8 with the
2
1
À
Ru3 plane. The short and similar lengths of the C N and
2
3
À
À
À
C
N distances (C4 N1 and C4 N2 in Figure 5) and the
slightly pyramidal character of N1 indicate that this five-
electron donor ligand is a resonance hybrid of the structures
depicted in Figure 6.
Scheme 5. Dynamic process that explains the 1H NMR spectra of com-
pound 4.
In solution at room temperature, compound 4 is not rigid.
Most resonances of its 1H NMR spectrum at 293 K
(Figure 7) are broad, standing out a very broad peak at d=
3.9 ppm that includes both protons of the CH2 group of the
ligand. However, at 213 K, the spectrum shows inequivalent
CH2 protons, as one would expect for the rigid molecular
structure determined in solid state by X-ray diffraction. A
reasonable dynamic process that would account for the
It was expected that a one-electron reduction of 2+ would
C
generate the neutral odd-electron species 2 . However, this
species was found to be unstable with regard to spontaneous
dimerization because the chemical reduction of 2·OTf with
one equivalent of cobaltocene afforded the hexanuclear
cluster derivative [Ru6ACTHNUTRGNEUNG
(m-H)2{m6-k6N1,N1’,C5,C5’,C6,C6’-4,4’-
Chem. Eur. J. 2013, 19, 3426 – 3436
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3429

J. A. Cabeza et al.
Figure 9. X-ray diffraction structure of 5. Selected bond lengths [ꢅ]:
À
À
À
À
Ru1 Ru2 2.7298(5), Ru1 Ru3 2.8470(5), Ru2 Ru3 2.8713(5), Ru1 N1
À
À
À
À
2.146(4), Ru2 C1 2.073(4), Ru3 C1 2.312(4), Ru3 C2 2.307(4), N1 C1
À
À
À
À
1.408(5), N1 C4 1.316(6), N2 C3 1.482(5), N2 C4 1.345(6), N2 C5
1.462(6), C1 C2 1.397(6), C2 C3 1.508(6), C3 C9 1.565(9), C4 C6
1.489(6).
À
À
À
À
Figure 7. Variable-temperature 1H NMR spectra (CD2Cl2, 400 MHz) of
compound 4 (*=impurities).
prises two closed trimetallic Ru3(CO)9 moieties, each edge-
bridged by a hydride and face-capped by a ligand that arises
À
from the intermolecular formation of a C C bond between
bis(2,3-Me2pyr)}(CO)18] (5), which was isolated as an air-
stable yellow solid (Scheme 6). Accordingly, the cyclic vol-
tammograms of 2·OTf in dichloromethane at various scan
rates (Figure 8) show an irreversible reduction process, indi-
cating that the molecules of the reduced odd-electron spe-
the C4 atoms of the heterocyclic ligands of two molecules of
C
its radical precursor 2 . Therefore, the formation of complex
5 not only implies the reductive dimerization of its cationic
precursor 2+ but also the loss of a CO molecule per each
trinuclear subunit. Although the observed conformation of
the two trinuclear subunits with respect to the C
minimizes steric hindrance between them, the very long dis-
4
4’
À
C
cies 2 have a very short half-life in solution, not long
C bond
enough to allow them to be re-oxidized to the cationic spe-
cies 2+ even at fast scan rates.
tance of this bond (C3 C9 in Figure 9), 1.565(9) ꢅ, reflects
À
The molecular structure of compound 5 was determined
by X-ray diffraction methods (Figure 9). The molecule com-
a strong steric repulsion. The structure of each trinuclear
subunit of 5 is essentially analogous to that of complex 4,
but its sp3-hybridized carbon atom (C4), instead of being at-
tached to two hydrogen atoms (as in compound 4), is attach-
ed to one hydrogen atom and to the C4 atom of another tri-
nuclear subunit.
As expected from its structure, the nCO region of the solu-
tion IR spectrum of 5 is comparable to that of complex 4.
Its room temperature 1H NMR spectrum also contains
broad peaks that sharpen upon lowering the temperature
(as occurs with compound 4, see above).
Although the two faces of the heterocyclic ligand of the
Scheme 6. Synthesis of compound 5.
C
reduced odd-electron species 2 are not equivalent, the struc-
C
ture of compound 5 indicates that the dimerization of 2 se-
lectively takes place through the exo face of its N-heterocy-
clic ligand, probably because one of the axial CO ligands of
its Ru(CO)4 moiety protects the endo face from external at-
C
tacks. As 2 is asymmetric, it is in fact a mixture of two enan-
tiomers. The fact that 5 does not have a rac configuration
C
(C2 symmetry) implies that the dimerization of 2 takes
place between monomers with opposed chirality, that is, the
enantiomers. Dextro–dextro and levo–levo couplings would
lead to rac dimers (C2 symmetry).
In addition to the pyrimidine-derived cluster shown in
Figure 8. Cyclic voltammograms of 2·OTf acquired at 208C in dichloro-
methane solutions containing 0.1m [NEt4]
A
Scheme 1,[7] the reductive intermolecular formation of C C
À
3430
www.chemeurj.org
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436

Synthesis and Reactivity of Cationic Triruthenium Clusters
FULL PAPER
bonds between ligands of metal complexes has been previ-
ously observed only for pyrazine-, quinoxaline-, and 1,5-
naphthyridine-derived triruthenium clusters[19] and for a few
mononuclear complexes.[20] In general, these metal com-
plexes are cationic and have ligand-based LUMOs. By con-
trast, the reductive dimerization of complexes with metal-
À
based LUMOs leads to dimers with a new metal metal
bond.[21]
It has been previously shown that the reactions of the cat-
Scheme 8. Synthesis of compound 9·OTf.
ionic pyrimid-2-ylidene-derivedcluster [Ru3ACTHNUTRGNEUNG
(m-H)(m-k2N1,C2-
3-Mepyr)(CO)10]+ with K-selectride and cobaltocene also
result in the addition of a hydride to the heterocyclic ligand
Me2pyr)(CO)10]CF3SO3 (9·OTf) (Scheme 8), whose structur-
al (see Figure 10), analytical, and spectroscopic data are en-
tirely analogous, but not identical, to those of 2·OTf (they
differ in the position of the C-Me group) and, therefore,
these data do not merit any additional comment.
and
(Scheme 1).[7] However, the different coordination of the li-
3A
a
reductive dimerization product, respectively
CHTUNGTRENNUNG
gands in 2+ (k2N1,C6) and [Ru (m-H)(m-k2N1,C2-3-Mepyr)(-
CO)10]+ (k2N1,C2) determines that the ligands of their reac-
tion (hydride addition and reduction) products are very dif-
ferent.
Compounds derived from 4-methylpyrimidine: With the ob-
jective of determining whether or not the nature of the reac-
tion products is affected by the position of the C-alkyl group
in the pyrimidine-derived ligand, and considering that the
reactivity (nucleophilic addition and reductive dimerization)
of 2+ is centered on its ligand C4 carbon atom, we decided
to perform a comparative study of the reactivity of a cation-
ic cluster similar to 2+ but having a methyl group on the C4
carbon atom of the pyrimidine ring (2+ has a C-Me group
on the C2 carbon atom). Such a complex was prepared in
two steps form [Ru3(CO)12], 4-mehylpyrimidine, and methyl
triflate.
The N1,C6-cyclometalated triruthenium complex [Ru
3ACTHNUTRGENUG(N m-
H)(m-k2N1,C6-4-Mepyr)(CO)10] (6) was prepared by treating
[Ru3(CO)12] with 4-methylpyrimidine (THF, heating to
Figure 10. DFT-optimized structure of the cationic complex 9+ with se-
lected interatomic distances in [ꢅ].
reflux, 1.5 h). Small amounts of the isomeric clusters [Ru
3ACHTUNGERTN(NUNG m-
H)(m-k2N1,C2-4-Mepyr)(CO)10] (7) and [Ru (m-H)(m-k2N3,C2-
3A
4-Mepyr)(CO)10] (8), which were also formed in this reac-
tion, were separated from complex 6 by column chromatog-
raphy (Scheme 7).
The treatment of 9·OTf with K[NACHTNUTRGNEUNG(SiMe3)2], under similar
conditions to those previously used for 2·OTf, led to a mix-
ture of products (at least five small hydride resonances were
observed in the 1H NMR spectrum of an aliquot of the
crude reaction solution) that decomposed on chromato-
graphic supports and that could not be separated and char-
acterized.
The reaction of compound 6 with methyl triflate (di-
chloromethane, room temperature, 48 h) allowed the synthe-
sis of the N-methyl derivative [Ru3ACTHNUTRGNEUNG
(m-H)(m-k2N1,C6-3,4-
The reaction of 9·OTf with K-selectride immediately
formed the neutral decacarbonyl derivative [Ru3ACHTUNGTRENNUNG(m-H)(m-
k2N1,C6-2-H-3,4-Me2pyr)(CO)10] (10), but this compound
slowly decomposed in solution at room temperature to give
the nonacarbonyl product [Ru3ACTHNUTRGNEUNG
(m-H)(m3-k2N1,C6-2-H-3,4-
Me2pyr)(CO)9] (11) (Scheme 9).
Compound 10 was only characterized by spectroscopic
techniques due to limited thermal stability. However, its
structure was unambiguously inferred from its IR spectrum,
whose nCO region is nearly identical to that of complex 6 (in-
dicating a similar arrangement of the CO ligands), and from
1
its H NMR spectrum, which, besides the resonances of the
Scheme 7. Reaction of [Ru3(CO)12] with 4-methylpyrimidine.
N-heterocyclic ligand (that are related to those of compound
Chem. Eur. J. 2013, 19, 3426 – 3436
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3431

J. A. Cabeza et al.
tached to the C4 carbon atom. A hydride ligand spans the
À
À
same Ru Ru edge as the N1 atom. The short Ru3 C4 dis-
tance, 2.056(8) ꢅ, supports the carbenic character of the het-
erocyclic ligand. An analysis of the heterocyclic ligand in-
teratomic distances indicates that it can be described as a
resonance hybrid of the carbenic structures I and J, depicted
in Figure 12, with a greater contribution of the former.
Therefore, the heterocyclic ligand of compound 11 is the
first example of another unprecedented type of NHC.
Figure 12. Resonance structures for the five-electron donor ligand of 11.
Scheme 9. Synthesis of compound 11.
11, see below), shows a singlet at d=À14.72 ppm, confirm-
ing that the environment of the hydride ligand is compara-
ble to that of compound 6, whose hydride NMR shift ap-
pears at d=À14.57 ppm.
In solution, the CH2 group of the heterocyclic ligand of
compound 11 flips freely from one side to the other of the
ligand plane because it appears as a sharp singlet (at d=
1
4.01 ppm) in the room temperature H NMR spectrum. The
The structure of compound 11 was established by X-ray
diffraction analysis (Figure 11). The molecule results from a
nucleophilic addition of a hydride to the heterocyclic C2
carbon atom of 9+ and from the loss of a CO ligand. The
high 13C NMR shift of the metalated carbon atom, d=
215.2 ppm, also confirms the carbenic character of the
ligand.
As occurred with 2·OTf (Figure 8), the cyclic voltammo-
grams of 9·OTf also showed an irreversible reduction wave,
but in the case of 9·OTf the reduction occurs at a lower po-
tential (Ered =À0.95 V at 200 VsÀ1 for 2·OTf and À1.11 V at
200 VsÀ1 for 9·OTf). However, the reaction of 9·OTf with
cobaltocene led to a complicated product mixture that de-
composed on chromatographic supports, some compound 6
and a small amount of [Ru3(CO)12] being the only products
that could be isolated from a chromatographic workup on
silica gel of the reaction solution.
Therefore, the presence of the methyl group on the heter-
ocyclic C4 atom of compound 9·OTf seriously affects the re-
activity of this system, which is completely different from
that of 2·OTf under similar reaction conditions.
Atomic charge and molecular orbital calculations. Some the-
oretical calculations, aimed at shedding more light on the
origin of the regioselectivity of each reaction and, conse-
quently, on the question why a change in the position of the
ligand C-Me group results in a completely different reactivi-
ty of compounds 2·OTf and 9·OTf, were carried out.
Figure 11. X-ray diffraction structure of 11. Selected bond lengths [ꢅ]:
À
À
À
À
Ru1 Ru2 2.7733(9), Ru1 Ru3 2.7593(9), Ru2 Ru3 2.7556(9), Ru1 N1
À
À
À
À
2.171(7), Ru2 N1 2.119(8), Ru3 C4 2.056(8), N1 C1 1.46(1), N1 C4
Concerning the reactions with K[NACTHNUTRGNEU(NG SiMe3)2], which result-
ed in compound 3 (deprotonation of the C2-Me group) in
the case of 2+ but in an uncharacterized mixture of products
in the case of 9+, we reasoned that the higher the atomic
charge of a hydrogen atom the easier its deprotonation
would be. Consequently, we computed the natural bond or-
bital (NBO)[22] atomic charges of the heterocyclic ligand hy-
drogen atoms of the cations 2+ and 9+. The results are
shown in Figure 13 (DFT calculations at the B3LYP/
À
À
À
À
1.38(1), N2 C1 1.47(1), N2 C2 1.34(1), N2 C5 1.46(1), C2 C3 1.39(1),
À
À
C2 C6 1.48(1), C3 C4 1.42(1).
new ligand is not planar, because the N1-C1-N2 plane forms
a dihedral angle of 43.7(6)8 with the plane defined by the
N1, N2, C2, C3, and C4 atoms, and caps a face of the Ru3
triangle in such a way that an edge is bridged by the N1
atom of the ligand, whereas the remaining Ru atom is at-
3432
www.chemeurj.org
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436

Synthesis and Reactivity of Cationic Triruthenium Clusters
FULL PAPER
ophile to the C6 atom is sterically impeded by the nearby
carbonyl ligands, the C4 atom of 2+ can be easily accessed
by the nucleophile but the C4 atom of 9+ is protected by the
C4-Me group. Consequently, the nucleophile attacks the C4
atom of 2+ (to give compound 4) but not the C4 atom of 9+.
The fact that C2 carbon atom is the only atom of 9+ that,
contributing to the LUMO, is not sterically impeded ac-
counts for the regioselectivity of the reaction of 9·OTf with
K-selectride, in which the nucleophile attacks the C2 atom
(to give compound 11).
Figure 13. NBO atomic charges of the hydrogen atoms (left) and the
non-hydrogen atoms (right) of the heterocyclic ligands of the cationic
complexes 2+ and 9+ (for clarity, only the heterocyclic ligand and two
Ru atoms of each cluster are shown).
This orbital-steric analysis can also explain the results of
the reactions of 2·OTf and 9·OTf with cobaltocene. It
should be noted that, upon addition of one electron, the
LUMOs of the cationic complexes 2+ and 9+ are converted
LanL2DZ/6-311+G** level[23–26]). Although the differences
between the charges of the ligand hydrogen atoms are
small,[18] the greatest charges are located on the hydrogen
atoms of their corresponding C-Me group. Therefore, both
complexes should deprotonate at their C-Me group. We be-
lieve that the different results obtained from the reactions
C
C
into the SOMOs of the radicals 2 and 9 , respectively, and
that the reduction does not significantly alter the atomic
composition of these orbitals (i.e., the atomic composition
of the LUMO of each cationic complex is very similar to
that of the SOMO of the corresponding neutral radical de-
rivative).[19] Once 2+ has added one electron, the resulting
of 2+ and 9+ with K[N
ACTHNUGTRENNUG(SiMe3)2] should arise from a very
C
different stability of the intermediate species that result
from the deprotonation of their C-Me groups, because these
neutral species should have very different distributions of
single and double bonds within their heterocyclic rings.
With regard to the results of the nucleophilic addition re-
actions of a hydride to 2+ and 9+, which give 4 (addition at
C4) and 11 (addition at C2), respectively, the atomic charges
of the heterocyclic non-hydrogen atoms of 2+ and 9+
(Figure 13) might suggest that the nucleophile attacks the
unsubstituted ring atom having the greatest atomic charge
(C4 in 2+ and C2 in 9+). However, the fact that the most
positively charged atoms of these cationic complexes are the
carbonyl carbon atoms,[6,7] with charges greater than 0.56 in
all cases, clearly indicates that these reactions are not charge
controlled.
Investigating the orbital control of these nucleophilic ad-
dition reactions, we also carried out molecular orbital calcu-
lations. Figure 14 shows that the LUMOs of 2+ and 9+ are
nearly identical, being mostly a combination of p-type orbi-
tals of ring atoms (not all) of the bridging ligand. The major
contributors to these molecular orbitals are the C6 and C4
carbon atoms. Interestingly, although an attack of the nucle-
radical (2 ) undergoes dimerization through the unsubstitut-
ed C4 atom, that is the least sterically shielded ligand ring
atom of those that contribute to the SOMO of this radical.
However, the radical resulting from the reduction of 9+ (9 )
cannot dimerize through any of its ligand ring atoms as they
are all sterically protected, including the unsubstituted C2
atom (various CO groups and the nearby N-Me group ob-
struct the approach of the C atom of 9 to the C atom of
another molecule of 9 , that is much bulkier than a hydride).
Thus, once formed, the radical 9 undergoes decomposition
through as yet unknown reaction pathways but different
from dimerization.
Therefore, theoretical calculations have also helped us to
rationalize the regioselectivity of reactions of 2+ and 9+
(nucleophilic addition and reduction) and establish that it is
governed by a combination of two factors: the composition
of the LUMOs of the cationic complexes and the steric pro-
tection of their ligand ring atoms.
C
2
2
C
C
C
Conclusion
The N-methylation of two neutral triruthenium cluster com-
plexes derived from 1,6-cyclometalated 2-methyl- and 4-
methylpyrimidines (compounds 1 and 6) with methyl triflate
has allowed the synthesis of two similar cationic complexes
(compounds 2+ and 9+) that contain a six-membered ring
NHC ligand that has the Ccarbene atom in the 6-position of a
pyrimidine framework.
It has been determined that the position of the C-methyl
group in the ligands of complexes 2+ and 9+ is of key im-
portance in the results of their reactions with K[NACTHNUGTRNEUNG(SiMe3)2],
K-selectride, and cobaltocene. Although these reagents
react with 2+ to give compounds 3 (deprotonation of the C2-
Me group), 4 (hydride addition at C4), and 5 (reductive di-
merization at C4), respectively, similar reactions with 9+
have only allowed the isolation of compound 11 (hydride
Figure 14. LUMOs of the cationic complexes 2+ (left) and 9+ (right),
shown at an isosurface value of 0.05.
Chem. Eur. J. 2013, 19, 3426 – 3436
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3433

J. A. Cabeza et al.
addition at C2). Compounds 3 and 11 also contain novel six-
membered ring NHC ligands.
mixture was allowed to reach room temperature and was filtered. A
1H NMR analysis of an aliquot of this solution revealed the presence of
compounds 1 (ꢀ4%) and 3 (ꢀ96%). The filtrate was evaporated to dry-
ness and the residue was washed with hexane (5 mL) to give compound 3
The molecular and electronic structures of 2+ and 9+
(that have ligand-based LUMOs) have been probed by
using quantum mechanical calculations. These theoretical
studies have established that the deprotonation reactions of
2+ and 9+ are charge-controlled processes and that both the
composition of their LUMOs and the steric protection of
their ligand ring atoms govern the regioselectivity of their
nucleophilic addition and reduction reactions.
1
as a garnet solid (25 mg, 75%). H NMR (300 MHz, CD2Cl2, 293 K): d=
6.41 (d, J=6.5 Hz, 1H; CH), 5.68 (d, J=6.5 Hz, 1H; CH), 3.69 (d, J=
3.2 Hz, 1H; CH2), 3.44 (d, J=3.2 Hz, 1H; CH2), 3.00 (s, 3H; NCH3),
À14.66 ppm (s, 1H; m-H); 13C{1H} and DEPT NMR (100 MHz, CD2Cl2,
293 K): d=208.3, 205.8, 201.4, 201.3, 199.1, 196.4, 195.5, 192.2, 191.5,
190.4 (10ꢆCO), 187.5 (C), 149.7 (C), 139.7 (CH), 110.7 (CH), 72.3
(CH2), 41.7 ppm (NCH3); IR (THF): n˜CO =2097 (w), 2058 (s), 2046 (vs),
2012 (s, br), 1993 (m, sh), 1974 (w, sh), 1950 cmÀ1 (vw, sh); MS (+-FAB):
m/z: 693 [M]+; elemental analysis calcd for C16H8N2O10Ru3 (M=691.45):
C 27.29, H 1.17, N 4.05; found: C 27.84, H 1.19, N 4.01.
Reaction of 2·OTf with K-selectride
Synthesis of [Ru3ACTHNUTRGNEUNG
(m-H)(m3-k3N1,C5,C6-4-H-2,3-Me2pyr)(CO)9] (4): A solu-
Experimental Section
tion of K-selectride (90 mL, 1.0m in THF, 0.090 mmol) was added to a
cold (193 K) solution of 2·OTf (75 mg, 0.089 mmol) in dichloromethane
(20 mL). The color of the solution immediately changed from yellow to
orange. The mixture was allowed to reach room temperature and was fil-
tered. A 1H NMR analysis of an aliquot of this solution revealed the
presence of compounds 1 (ꢀ5%), 3 (ꢀ15%), and 4 (ꢀ75%). The fil-
trate was concentrated to a small volume (2 mL) and was transferred to
a silica gel column (2ꢆ10 cm) packed in hexane. Hexane eluted a small
amount of [Ru3(CO)12]. Hexane/dichloromethane (4:1) eluted a pale
pink band that contained a small amount of a mixture of compounds.
Hexane/dichloromethane (3:1) eluted a yellow band that contained com-
pound 4 (35 mg, 59%). Finally, neat dichloromethane eluted another
yellow band that contained some compound 1. Data for compound 4:
1H NMR (300 MHz, CD2Cl2, 213 K): d=4.10 (dd, J=14.4, 4.8 Hz, 1H;
CH2), 3.47 (d, J=14.4 Hz, 1H; CH2), 2.83 (s, 3H; NCH3), 2.58 (d, J=
4.8 Hz, 1H; CH), 1.96 (s, 3H; CH3), À16.48 ppm (s, 1H; m-H); 13C{1H}
and DEPT NMR (100 MHz, CD2Cl2, 293 K): d=203.5 (br, COs,), 196.9
(br, COs), 159.4 (C), 154.3. (br, C), 50.1 (CH2), 38.7 (NCH3), 29.7 (CH),
20.9 ppm (CH3); IR (CH2Cl2): n˜CO =2077 (w), 2045 (s), 2024 (vs), 2002
(m), 1990 (m, br), 1974 (w, sh), 1950 cmÀ1 (vw, sh); MS (+-FAB): m/z:
667 [M]+; elemental analysis calcd for C15H10N2O9Ru3 (M=665.46): C
27.07, H 1.51, N 4.21; found: C 27.10, H 1.54, N 4.16.
General: Solvents were dried over NaACHTNUGTRENUNG[Ph2CO] (THF, diethyl ether, hy-
drocarbons) or CaH2 (dichloromethane) and distilled under nitrogen
prior to use. The reactions were carried out under nitrogen, by using
Schlenk-vacuum line techniques, and were routinely monitored by solu-
tion IR spectroscopy and by spot TLC on silica gel. All reaction products
were dried in vacuum for several hours prior to being weighted and ana-
lyzed. IR: Perkin–Elmer FT Paragon 1000X. NMR: Bruker AV-400,
NAV-400, and DPX-300, room temperature, residual solvent as internal
standard. Microanalyses: Perkin–Elmer 2400. MS: VG Autospec double-
focusing mass spectrometer operating in the FAB+ mode; ions were pro-
duced with a standard Cs+ gun at about 30 kV; 3-nitrobenzyl alcohol was
used as matrix; data given correspond to the most abundant molecular
ion isotopomer.
Synthesis of [Ru3ACHTUNGTRENNUNG
(m-H)(m-k2N1,C6-2-Mepyr)(CO)10] (1): A solution of
[Ru3(CO)12] (450 mg, 0.704 mmol) and 2-methylpyrimidine (30 mL,
0.258m in THF, 0.774 mmol) in THF (40 mL) was stirred under reflux for
5 h. The color of the reaction mixture changed from orange to garnet.
The solvent was removed under reduced pressure, the residue was ex-
tracted into dichloromethane (2 mL) and the resulting solution was trans-
ferred to a silica gel column (2ꢆ15 cm) packed in hexane. Hexane eluted
a small amount of [Ru3(CO)12]. Dichloromethane eluted a yellow band
that contained compound 1 (250 mg, 52%). 1H NMR (300 MHz, CD2Cl2,
293 K): d=7.88 (d, J=4.9 Hz, 1H; CH), 7.24 (d, J=4.9 Hz, 1H; CH),
2.71 (s, 3H; CH3), À14.26 ppm (s, 1H; m-H); 13C{1H} and DEPT NMR
(100 MHz, CD2Cl2, 293 K): d=207.9, 206.1, 201.1, 201.0, 196.1, 195.7,
195.5, 192.3, 191.2, 190.4 (10 COs), 187.1 (C), 166.6 (C), 150.1 (CH),
133.4 (CH), 26.6 ppm (CH3); IR (CH2Cl2): n˜CO =2101 (w), 2063 (s), 2052
(vs), 2025 (m, sh), 2014 (m, br), 1999 cmÀ1 (w, sh); MS (+-FAB): m/z:
679 [M]+; elemental analysis calcd for C15H6N2O10Ru3 (M=677.43): C
26.59, H 0.89, N 4.14; found: C 26.63, H 0.94, N 4.08.
Reaction of 2·OTf with cobaltocene
(m-H)2{m6-k6N1,N1’,C5,C5’,C6,C6’-4,4’-bis(2,3-Me2pyr)}-
Synthesis of [Ru6ACTHNUTRGNEUNG
(CO)18] (5): Cobaltocene (40 mg, 0.214 mmol) was added to a cold
(193 K) solution of 2·OTf (180 mg, 0.214 mmol) in THF (15 mL). The
color of the solution immediately changed from yellow to brown. The
mixture was allowed to reach room temperature and was filtered After
stirring for 2 h; the solvent was removed, the residue was extracted into
dichloromethane (2 mL) and the resulting solution was transferred to a
silica gel column (2ꢆ10 cm) packed in hexane. Hexane eluted a small
amount of [Ru3(CO)12]. Hexane/dichloromethane (4:1), hexane/dichloro-
methane (3:1), and hexane/dichloromethane (1:1) eluted several pale
bands that were not collected. Hexane/dichloromethane (1:3) eluted a
mixture of two compounds (NMR identification that contained complex
5 as its major component. Further elution of the column with dichloro-
methane afforded some compound 1. The mixture that contained com-
pound 5 was separated by TLC on silica gel. Hexane/dichloromethane
(2:3) eluted two yellow bands. The first one could not be characterized.
Synthesis of [Ru3ACHTUNGTRENNUNG ACHTUNGTNER(NUGN 2·OTf):
(m-H)(m-k2N1,C6--2,3-Me2pyr)(CO)10]CF3SO3
Methyl triflate (208 mL, 1.845 mmol) was added to a solution of com-
pound 1 (250 mg, 0.369 mmol) in dichloromethane (30 mL). The mixture
was stirred under reflux for 4 h. The color of the solution changed from
yellow to orange and the solution contained some yellow solid. The sol-
vent was evaporated at reduced pressure and the residue was washed
with diethyl ether (2ꢆ20 mL) to give 2·OTf as a yellow solid (225 mg,
72%). 1H NMR (300 MHz, CD2Cl2, 293 K): d=8.65 (d, J=6.5 Hz, 1H;
CH), 8.41 (d, J=6.5 Hz, 1H; CH), 4.52 (s, 3H; NCH3), 3.46 (s, 3H;
CH3), À14.05 ppm (s, 1H; m-H); 13C{1H} and DEPT NMR (100 MHz,
[D6]acetone, 293 K): d=225.9, 209.7, 203.7, 203.3, 197.3, 193.9, 193.6,
191.8, 188.3 (9ꢆCO; the solvent CO resonance obscures an additional
CO peak), 188.0 (C), 164.9 (C), 144.9 (CH), 137.4 (CH), 49.3 (NCH3),
26.2 ppm (CH3); IR (CH2Cl2): n˜CO =2109 (w), 2075 (s), 2064 (vs), 2040
(w), 2014 (m, br), 1993 cmÀ1 (w, sh); MS (+-FAB): m/z: 694
[MÀCF3SO3]+; elemental analysis calcd for C17H9F3N2O13Ru3S (M=
841.53): C 24.26, H 1.08, N 3.33; found: C 24.33, H 1.14, N 3.29.
The second band contained compound
5
(40 mg, 28%). 1H NMR
(300 MHz, CD2Cl2, 233 K): d=3.58 (brs, 1H; CH), 2.89 (s, 3H; NCH3),
2.47 (brs, 1H; CH), 1.92 (s, 3H; CH3), À16.66 ppm (s, 1H; m-H);13C{1H}
and DEPT NMR (100 MHz, CD2Cl2, 293 K): d=156.7 (C), 64.9 (CHR),
38.7 (NCH3), 23.2 (CH3), 20.9 ppm (CH); IR (CH2Cl2): n˜CO =2078 (w),
2049 (vs), 2028 (s), 2006 (m), 1993 (m, br), 1976 (w, sh), 1953 cmÀ1 (vw,
sh); MS (+-FAB): m/z: 1329 [M]+; elemental analysis calcd for
C30H18N4O18Ru6 (M=1328.90): C 27.11, H 1.37, N 4.22; found: C 27.16,
H 1.40, N 4.19.
Deprotonation of 2·OTf with K[N
Synthesis of [Ru3A
(m-H)(m-k2N1,C6-2-CH2-3-Mepyr)(CO)10] (3): A solution
of K[N(SiMe3)2] (95 mL, 0.5m in toluene, 0.048 mmol) was added to a
cold (193 K) solution of 2·OTf (40 mg, 0.048 mmol) in THF (20 mL). The
G
Reaction of [Ru3(CO)12] with 4-methylpyrimidine
C
Synthesis of [Ru
k2N1,C2-4-Mepyr)(CO)10] (7), and
(8): A solution of [Ru3(CO)12] (300 mg, 0.459 mmol) and 4-methylpyrimi-
3A
3ACHTUNGTRENNUNG(m-H)(m-
G
A
CHTUNGTRENNUNG
3434
www.chemeurj.org
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436

Synthesis and Reactivity of Cationic Triruthenium Clusters
FULL PAPER
dine (47 mL, 0.516 mmol) in THF (30 mL) was stirred under reflux for
1.5 h. The color of the solution changed from orange to dark red. The
solvent was removed under reduced pressure, the residue was extracted
into dichloromethane (2 mL), and the resulting solution was transferred
to a silica gel column (2ꢆ15 cm) packed in hexane. Hexane eluted a
small amount of [Ru3(CO)12]. Hexane/dichloromethane (10:1) eluted
three bands that, upon solvent removal, led to compounds 6 (yellow-
orange; 100 mg, 32%), 7 (yellow; 30 mg, 10%), and 8 (yellow;10 mg,
3%), in order of elution.
Data for compound 6: 1H NMR (300 MHz, CD2Cl2, 293 K): d=8.39 (s,
1H; CH), 7.38 (s, 1H; CH), 2.36 (s, 3H; CH3), À14.57 ppm (s, 1H; m-H);
13C{1H} and DEPT NMR (100 MHz, CD2Cl2, 293 K): d=207.9, 206.3,
201.1, 200.9, 196.1, 195.1, 192.6, 192.4, 191.4, 190.0 (10ꢆCO), 186.0 (C),
Data for compound 11: 1H NMR (300 MHz, CD2Cl2, 213 K): d=5.85 (s,
1H; CH), 4.01 (s, 2H; CH2), 3.09 (s, 3H; NCH3), 2.02 (s, 3H; CH3),
À15.32 ppm (s, 1H; m-H); 13C{1H} and DEPT NMR (100 MHz, CD2Cl2,
293 K): d=215.2 (C), 202.6 (br, COs,), 194.4 (br, COs), 152.2 (C), 110.5
(CH), 79.1 (CH2), 36.7 (NCH3), 18.1 ppm (CH3); IR (CH2Cl2): n˜CO =2079
(w), 2050 (s), 2025 (vs), 1997 (m, br), 1950 cmÀ1 (vw, sh); MS (+-FAB):
m/z: 667 [M]+; elemental analysis calcd for C15H10N2O9Ru3 (M=665.46):
C 27.07, H 1.51, N 4.21; found: C 27.12, H 1.56, N 4.17.
Electrochemical studies: Cyclic voltammetric measurements were carried
out with an AUTOLAB-III potentiostat connected to a home-made
single-compartment micro cell containing a 1 mm platinum disk as work-
ing electrode, a platinum wire as auxiliary electrode, and a silver wire as
pseudo-reference electrode. All measurements were carried out in deoxy-
genated dichloromethane (freshly distilled from CaH2) solutions contain-
161.6 (C), 159.9 (CH), 135.9 (CH), 23.4 ppm (CH3); IR (CH2Cl2): n˜CO
=
ing 0.1m [NEt4]ACHTNUGTRENUNG[BF4]. The concentration of each cluster complex analyte
2101 (w), 2064 (s), 2052 (vs), 2023 (m, sh), 2014 (m, br), 1998 (w, sh),
1983 cmÀ1 (vw, sh); MS (+-FAB): m/z: 679 [M]+; elemental analysis
calcd for C15H6N2O10Ru3 (M=677.43): C 26.59, H 0.89, N 4.14; found: C
26.64, H 0.95, N 4.09.
was 0.5ꢆ10À3 m. To measure the potentials, ferrocene (FeCp2; Cp=cyclo-
pentadienyl) was used as internal standard. The [FeCp2]0/+ couple was
referenced to E1/2 =0.59 V versus a saturated calomel electrode (SCE).
Data for compound 7: 1H NMR (300 MHz, CD2Cl2, 293 K): d=7.86 (d,
J=5.8 Hz, 1H; CH), 6.74 (d, J=5.8 Hz, 1H; CH), 2.46 (s, 3H; CH3),
À14.49 ppm (s, 1H; m-H); 13C{1H} and DEPT NMR (100 MHz, CD2Cl2,
293 K): d=208.4, 206.3, 201.3, 201.2, 195.9, 195.0, 193.4, 192.9, 191.8,
190.2 (10ꢆCO), 186.4 (C), 166.8 (C), 156.9 (CH), 117.6 (CH), 24.7 ppm
(CH3); IR (CH2Cl2): n˜CO =2101 (w), 2064 (s), 2052 (vs), 2023 (m, sh),
2012 (m, br), 1998 (w, sh), 1982 cmÀ1 (vw, sh); MS (+-FAB): m/z: 679
[M]+; elemental analysis calcd for C15H6N2O10Ru3 (M=677.43): C 26.59,
H 0.89, N 4.14; found: C 26.62, H 0.93, N 4.10.
Data for compound 8: 1H NMR (300 MHz, CD2Cl2, 293 K): d=8.26 (d,
J=5.0 Hz, 1H; CH), 6.80 (d, J=5.0 Hz, 1H; CH), 2.48 (s, 3H; CH3),
À14.24 ppm (s, 1H; m-H); IR (CH2Cl2): n˜CO =2101 (w), 2064 (s), 2052
(vs), 2023 (m, sh), 2012 (m, br), 1999 (w, sh), 1981 cmÀ1 (vw, sh); MS
(+-FAB): m/z: 679 [M]+; elemental analysis calcd for C15H6N2O10Ru3
(M=677.43): C 26.59, H 0.89, N 4.14; found: C 26.61, H 0.90, N 4.09.
Theoretical calculations: Density functional theory (DFT) calculations
were carried out by using the three-parameter hybrid exchange-correla-
tion functional of Becke[23] and the B3LYP non-local gradient correc-
tion.[24] The LanL2DZ basis set, with relativistic effective core potentials,
was used for the Ru atoms.[25] The basis set used for the remaining atoms
was the 6-311+G**.[26] All optimized structures were confirmed as
energy minima by analytical calculation of the frequencies (all positive
eigenvalues). Atomic charges and molecular orbital data were obtained
from the natural bond orbital (NBO) analysis of the data.[22]All calcula-
tions were carried out with the Gaussian 03 package.[27] Cartesian coordi-
nates for the atoms of all optimized structures are given in the Support-
ing Information.
X-ray diffraction analyses: Crystals of compounds 4, 5, and 11 were ana-
lyzed by X-ray diffraction methods. Diffraction data were collected on an
Oxford Diffraction Xcalibur Onyx Nova single-crystal diffractometer, by
using CuKa radiation. Empirical absorption corrections for compounds 5
and 11 were applied by using the SCALE3 ABSPACK algorithm as im-
plemented in the program CrysAlisPro RED.[28] The XABS2[29] empirical
absorption correction was applied for compound 4. The structures were
solved by using the program SIR-97.[30] Isotropic and full-matrix aniso-
tropic least-square refinements were carried out by using SHELXL.[31]
All non-hydrogen atoms were refined anisotropically. The hydride li-
gands of compound 5 were located in a Fourier map and the positions of
Synthesis of [Ru3ACHTUNGTRENNUNG ACHTUGNTRNE[NUGN CF3SO3] (9·OTf):
(m-H)(m-k2N1,C6-3,4-Me2pyr)(CO)10]
Methyl triflate (84 mL, 0.740 mmol) was added to a solution of compound
1 (100 mg, 0.148 mmol) in dichloromethane (30 mL) and the mixture was
stirred for 48 h. A dark yellow solid precipitated. The solvent was evapo-
rated at reduced pressure and the residue was washed with diethyl ether
(2ꢆ20 mL) to give 9·OTf as an ocher-yellow solid (100 mg, 80%).
1H NMR (300 MHz, [D6]acetone, 293 K): d=9.80 (s, 1H; CH), 8.44 (s,
1H; CH), 4.31 (s, 3H; NCH3), 2.81 (s, 3H; CH3), À14.52 ppm (s, 1H; m-
H); 13C{1H} and DEPT NMR (100 MHz, [D6]acetone, 293 K): d=217.0,
206.8, 200.4, 200.3, 194.2, 192.5, 190.9, 190.8, 188.5 (9ꢆCOs; the solvent
CO resonance obscures an additional CO peak), 184.3 (C), 154.4 (CH),
the hydride atoms of compounds
4 and 11 were calculated with
XHYDEX.[32] The remaining hydrogen atoms were set in calculated posi-
tions and refined riding on their parent atoms. The molecular plots were
made with the PLATON program package.[33] The WINGX program
system[34] was used throughout the structure determinations. Selected
crystal, measurement, and refinement data as well as ORTEP plots of
the three structures are given in the Supporting Information. CCDC-
906830 (4), CCDC-906831 (5), and CCDC-906829 (11) contain the sup-
plementary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
152.6 (C), 136.8 (CH), 41.7 (NCH3), 18.1 ppm (CH3); IR (CH2Cl2): n˜CO
=
2110 (w), 2078 (s), 2063 (vs), 2038 (w), 2014 (m, br), 1994 (w, sh); MS
(+-FAB): m/z: 694 [MÀCF3SO3]+; elemental analysis calcd for
C17H9F3N2O13Ru3S (M=841.53): C 24.26, H 1.08, N 3.33; found: C 24.31,
H 1.13, N 3.28.
Reaction of 9·OTf with K-selectride
(m-H)(m3-k2N1,C6-2-H-3,4-Me2pyr)(CO)9] (11): A solu-
Synthesis of [Ru3ACHTUNGTRENNUNG
tion of K-selectride (60 mL, 1.0m in THF, 0.059 mmol) was added to a
cold (193 K) solution of 9·OTf (50 mg, 0.059 mmol) in dichloromethane
(20 mL). The color of the solution immediately changed from yellow to
orange. An IR spectrum of this solution indicated that it contained the
Acknowledgements
neutral
decacarbonyl
derivative
[Ru3ACTHGNUTERNNUG
(m-H)(m-k2N1,C6-2-H-3,4-
Me2pyr)(CO)10] (10).The mixture was allowed to reach room tempera-
ture and was filtered. A 1H NMR analysis of an aliquot of this solution
revealed the presence of compounds 10 (ꢀ60%)and 11 (ꢀ40%). After
4 h compound 11 was the only product in the solution (1H NMR analy-
sis). The solvent was removed under reduced pressure and the residue
was washed with hexane (10 mL) to give compound 11 as an orange solid
(30 mg, 76%).
Data for compound 10: 1H NMR (300 MHz, CD2Cl2, 213 K): d=5.44 (s,
1H; CH), 4.41 (d, J=9.7 Hz, 1H; CH), 4.28 (d, J=9.7 Hz, 1H; CH), 2.86
(s, 3H; NCH3), 1.91 (s, 3H; CH3), À14.72 ppm (s, 1H; m-H); IR
(CH2Cl2): n˜CO =2094 (w), 2055 (vs), 2044 (s), 2011 (s), 1988 cmÀ1 (m, sh).
This work has been supported by the MICINN-FEDER projects
CTQ2010-14933, MAT2010-15094, and DELACIERVA-09-05. We are
also grateful to the Government of Principado de Asturias for a Severo
Ochoa fellowship (to V.P.).
[1] a) For a review that cites all review articles and books published
since 2007 on the chemistry of cyclic carbenes and related species,
see: M. Melaimi, M. Soleihavoup, G. Bertrand, Angew. Chem. 2010,
122, 8992–9032; Angew. Chem. Int. Ed. 2010, 49, 8810–8849; b) for
a recent review focusing on classical NHCs, see: F. E. Hahn, M. C.
Chem. Eur. J. 2013, 19, 3426 – 3436
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3435

J. A. Cabeza et al.
Jahnke, Angew. Chem. 2008, 120, 3166–3216; Angew. Chem. Int. Ed.
2008, 47, 3122–3172.
[18] J. A. Cabeza, P. Garcꢀa-ꢁlvarez, E. Pꢃrez-CarreÇo, V. Pruneda,
Dalton Trans. 2012, 41, 4313–4315.
[2] For relevant reviews on complexes of pyridylidenes and related
NHC ligands, see: a) O. Schuster, L. Yang, H. G. Raubenheimer, M.
Albrecht, Chem. Rev. 2009, 109, 3445–3478; b) H. G. Raubenheim-
er, S. Cronje, Dalton Trans. 2008, 1265–1272.
[3] For a recent review on the transition-metal cluster chemistry of
NHCs, see: J. A. Cabeza, P. Garcꢀa-ꢁlvarez, Chem. Soc. Rev. 2011,
40, 5389–5405.
[4] J. A. Cabeza, I. del Rꢀo, E. Pꢃrez-CarreÇo, M. G. Sꢂnchez-Vega, D.
Vꢂzquez-Garcꢀa, Angew. Chem. 2009, 121, 563–566; Angew. Chem.
Int. Ed. 2009, 48, 555–558.
[5] J. A. Cabeza, I. del Rꢀo, E. Pꢃrez-CarreÇo, M. G. Sꢂnchez-Vega, D.
Vꢂzquez-Garcꢀa, Organometallics 2010, 29, 4464–4471.
[19] J. A. Cabeza, I. del Rꢀo, E. Pꢃrez-CarreÇo, V. Pruneda, Chem. Eur.
J. 2010, 16, 5425–5436.
[20] a) E. O. Fischer, D. Wittmann, D. Himmelreich, D. Neugebauer,
Angew. Chem. 1982, 94, 451–452; Angew. Chem. Int. Ed. Engl.
1982, 21, 444–445; b) S. Gambarotta, F. Urso, C. Floriani, A. Chiesi-
Villa, C. Guastini, Inorg. Chem. 1983, 22, 3966–3972; c) E. Gallo, E.
Solari, N. Re, C. Floriani, A. Chiesi-Villa, C. Rizzoli, J. Am. Chem.
Soc. 1997, 119, 5144–5154; d) F. Franceschi, E. Solari, C. Floriani,
M. Rosi, A. Chiesi-Villa, C. Rizzoli, Chem. Eur. J. 1999, 5, 708–721;
e) M. Rosi, A. Sgamellotti, F. Franceschi, C. Floriani, Chem. Eur. J.
1999, 5, 2914–2920; f) E. A. Duplessis, P. A. Jelliss, C. C. Kirkpa-
trick, S. D. Minteer, K. M. Wampler, J. Organomet. Chem. 2006, 691,
4660–4666; g) J. J. Curley, T. Murahashi, C. C. Cummings, Inorg.
Chem. 2009, 48, 7181–7193.
[6] J. A. Cabeza, I. del Rꢀo, M. C. Goite, E. Pꢃrez-CarreÇo, V. Pruneda,
Chem. Eur. J. 2009, 15, 7339–7349.
[7] J. A. Cabeza, I. del Rꢀo, E. Pꢃrez-CarreÇo, V. Pruneda, Organome-
tallics 2011, 30, 1148–1156.
[8] a) B. Crociani, F. Dibianca, A. Giovenco, A. Scrivanti, J. Organomet.
Chem. 1985, 291, 259–272; b) B. Crociani, F. Dibianca, R. Bertani,
C. B. Castellani, Inorg. Chim. Acta 1985, 101, 161–169.
[21] See, for example: a) J. A. Cabeza, I. del Rꢀo, V. Riera, F. Grepioni,
Organometallics 1995, 14, 3124–3126; b) T. Murahashi, Y. Hashimo-
to, K. Chiyoda, M. Fujimoto, T. Uemura, R. Inoue, S. Ogoshi, H.
Kurosawa, J. Am. Chem. Soc. 2008, 130, 8586–8587.
[22] a) A. E. Reed, R. B. Weinstock, F. Weinhold, J. Chem. Phys. 1985,
83, 735–746; b) A. E. Reed, L. A. Curtis, F. Weinhold, Chem. Rev.
1988, 88, 899–926.
[9] A. Magriz, S. Gꢇmez-Bujedo, E. ꢁlvarez, R. Fernꢂndez, J. M. Lassa-
letta, Organometallics 2010, 29, 5941–5945.
[10] B. Hildebrandt, G. Reiß, C. Ganter, J. Organomet. Chem. 2010, 695,
474–477.
[11] H. Wang, M.-D. Su, J. Phys. Chem. A 2008, 112, 7689–7698.
[12] T. D. Forster, K. E. Krahulic, H. M. Tuononen, R. McDonald, M.
Parvez, R. Roesler, Angew. Chem. 2006, 118, 6504–6507; Angew.
Chem. Int. Ed. 2006, 45, 6356–6359.
[13] a) V. Cꢃsar, N. Lugan, G. Lavigne, J. Am. Chem. Soc. 2008, 130,
11286–11287; b) V. Cꢃsar, N. Lugan, G. Lavigne, Eur. J. Inorg.
Chem. 2010, 361–365; c) V. Cꢃsar, N. Lugan, G. Lavigne, Chem.
Eur. J. 2010, 16, 11432–11442.
[14] a) T. W. Hudnall, C. W. Bielawski, J. Am. Chem. Soc. 2009, 131,
16039–16041; b) T. W. Hudnall, J. P. Moerdyk, C. W. Bielawski,
Chem. Commun. 2010, 46, 4288–4290; c) T. W. Hudnall, E. J. Moor-
head, D. G. Gusev, C. W. Bielawski, J. Org. Chem. 2010, 75, 2763–
2766; d) J. P. Moerdyk, C. W. Bielawski, Organometallics 2011, 30,
2278–2284.
[23] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[24] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[25] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[26] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213–222.
[27] Gaussian 03, Revision C2, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J.
Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K.
Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakr-
zewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz,
Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G.
Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challa-
combe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonza-
lez, J. A. Pople, Gaussian Inc., Wallingford, CT, 2004.
[15] See, for example: a) E. Stander-Grobler, O. Schuster, G. Heyden-
rych, S. Cronje, E. Tosh, M. Albrecht, G. Frenking, H. G. Rauben-
heimer, Organometallics 2010, 29, 5821–5833; b) E. Stander-Gro-
bler, C. E. Strasser, O. Schuster, G. S. Cronje, H. G. Raubenheimer,
Inorg. Chim. Acta 2011, 376, 87–94.
[16] For examples of ruthenium carbonyl clusters with imidazol-2-ylidene
ligands, see: a) C. E. Ellul, M. F. Mahon, O. Saker, M. K. Whittlesey,
Angew. Chem. 2007, 119, 6459–6461; Angew. Chem. Int. Ed. 2007,
46, 6343–6345; b) M. R. Crittall, C. E. Ellul, M. F. Mahon, O. Saker,
M. K. Whittlesey, Dalton Trans. 2008, 4209–4211; c) C. Zhang, B.
Li, H. Song, S. Xu, B. Wang, Organometallics 2011, 30, 3029–3036;
d) J. A. Cabeza, I. del Rꢀo, D. Miguel, E. Pꢃrez-CarreÇo, M. G.
Sꢂnchez-Vega, Organometallics 2008, 27, 211–217; e) J. A. Cabeza,
I. del Rꢀo, D. Miguel, E. Pꢃrez-CarreÇo, M. G. Sꢂnchez-Vega,
Dalton Trans. 2008, 1937–1942; f) J. A. Cabeza, I. del Rꢀo, D.
Miguel, M. G. Sꢂnchez-Vega, Angew. Chem. 2008, 120, 1946–1948;
Angew. Chem. Int. Ed. 2008, 47, 1920–1922; g) J. A. Cabeza, M.
Damonte, P. Garcꢀa-ꢁlvarez, Organometallics 2011, 30, 2371–2376;
h) J. A. Cabeza, M. Damonte, P. Garcꢀa-ꢁlvarez, A. R. Kennedy, E.
Pꢃrez-CarreÇo, Organometallics 2011, 30, 826–833.
[28] CrysAlisPro RED, version 1.171.34.36, Oxford Diffraction Ltd.,
Oxford, 2010.
[29] S. Parkin, B. Moezzi, H. Hope, J. Appl. Crystallogr. 1995, 28, 53–56.
[30] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115–119.
[31] SHELXL, version 2008: G. M. Sheldrick, Acta Crystallogr. Sect. A
2008, 64, 112–122.
[32] A. G. Orpen, J. Chem. Soc. Dalton Trans. 1980, 2509–2516.
[33] PLATON, a multipurpose crystallographic tool, version 1.15, A. L.
Spek, University of Utrecht, Utrecht, 2008.
[34] WinGX, version 1.80.05 (2009): L. J. Farrugia, J. Appl. Crystallogr.
1999, 32, 837–838.
[17] See, for example: D. McHattie, R. Buchan, M. Fraser, P. V. S. Kong-
Tho-Lin, Heterocycles 1992, 34, 1759–1771, and references therein.
Received: October 25, 2012
Published online: January 17, 2013
3436
www.chemeurj.org
ꢄ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 3426 – 3436