N -acyliminoimidazolium ylides as precursors to anionic N -heterocyclic carbene ligands: Control of topology and reactivity

Article abstract of DOI:10.1021/om400078f

A family of N-heterocyclic carbene ligands based on N-acyl iminoimidazolium ylides has been developed and investigated. The ylides are stable and are easily obtained in two steps. Following deprotonation, they form bidentate anionic NHC ligands. The synth

Full text of DOI:10.1021/om400078f

Article
pubs.acs.org/Organometallics
N‑Acyliminoimidazolium Ylides as Precursors to Anionic
N‑Heterocyclic Carbene Ligands: Control of Topology and Reactivity
Hannah Guernon and Claude Y. Legault*
́ ́
Department of Chemistry, University of Sherbrooke, 2500 boulevard de l’Universite, Sherbrooke, Quebec J1K 2R1, Canada
S
* Supporting Information
ABSTRACT: A family of N-heterocyclic carbene ligands based on N-acyl
iminoimidazolium ylides has been developed and investigated. The ylides are
stable and are easily obtained in two steps. Following deprotonation, they form
bidentate anionic NHC ligands. The synthesis of silver(I) complexes was
explored. Different complex structures have been obtained, depending on the
steric and electronic properties of the starting ylide proligands. Dimeric
structures were usually obtained, but trimeric structures were observed under
certain conditions. DFT calculations were used to explain these results.
The electronic properties of the complexes can be modified by the choice of
the electron-withdrawing group attached to the anionic tether. The silver
complexes are proficient ligand transfer agents. The first example of a proligand/ligand exchange was demonstrated. The catalytic
activity of the silver complexes was evaluated in a model (3 + 2) heterocycloaddition of an imino ester with an activated alkene.
electrophiles such as acyl chlorides will give access to a wide
range of ylides (Scheme 1). Therefore, the steric and electronic
INTRODUCTION
■
After two decades of intense research on N-heterocyclic
carbenes (NHCs), it is now well-known that they are proficient
ligands that have the feature of being strong σ-donors,1 making
them perfect substitutes for phosphine ligands.2 The resulting
NHC complexes have often been found to offer enhanced
stability to oxygen, light, and water.3 The classical structure of
NHCs is formed by a 1,3-disubstituted imidazolylidene core that
can present a broad range of reactivity and stability, depending
on the electronic and steric properties of its substituents.1c While
the classical formally neutral NHCs have been widely studied,
there is new interest in a variant of NHC ligands, possessing an
anionic tether.4 Bidentate anionic ligands have already
demonstrated their great potential in catalysis, some precisely
implicating an important role of the tether in the catalytic
reaction.5 The anionic tether is used to rigidify the structure,
leading to increased stability of the complex. As is shown in this
article, it can also act as a base in a catalytic cycle.
Numerous kinds of anionic NHC ligands have been reported
in the literature, depending on the nature of the anionic tether,
such as alkoxide,6 aryloxide,7 amino and arylamino,8 amido,9 and
aryl.10 While most of these ligands and their corresponding
complexes are quite stable, and of catalytic interest, they present
a nondelocalized anion initially formed by a strong base.
Modulation of the electronic properties of those hard anions is
not readily possible. Charette et al. previously reported the proof
of concept of a new type of anionic NHC ligand based on a
N-benzoyliminoimidazolium ylide and the formation of silver(I)
and copper(II) complexes.11 While this type of ligand presents
great potential, a more thorough exploration is necessary. In
principle, the ylide proligands can be made in a simple and
modular way. Following amination of the corresponding mono-
N-substituted imidazole, coupling of the N-amino salt with
Scheme 1. Retrosynthetic Analysis of Ylide Formation
properties could be easily modulated to afford ligands with a
wide range of properties.
The use of ylides as ligand precursors has another practical
advantage. They are formal neutral compounds and can be
easily purified using standard flash chromatography. Herein
we report the development of an optimized synthetic route to
access a family of N-acyl iminoimidazolium ylides, with varied
steric and electronic properties. We also present new pro-
cedures to obtain the corresponding silver(I) complexes from
these ylides, as well as structural and reactivity analyses of five
new silver complexes showing different structures depending
on the nature of the ligand.
RESULTS AND DISCUSSION
■
Synthetic Aspects of the Ylides. The synthetic sequence
is initiated by the N-amination of corresponding N-substituted
imidazoles. O-(2,4-Dinitrophenyl)hydroxylamine was selected
for this transformation, as it is an efficient and stable aminating
reagent for heterocycles.12 Under the conditions described
in eq 1, excellent yields were obtained for both N-aryl- and
N-alkylimidazoles.
Received: January 29, 2013
Published: March 11, 2013
© 2013 American Chemical Society
1988
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
Under the reaction conditions, the 2,4-dinitrophenolate
counterion of the N-amino salt reacts rapidly with the acyl
chloride reagent to form the corresponding ester. The latter
does not show sufficient electrophilicity to react with the
N-amino salt to yield the final ylide, thus requiring the extra
acyl chloride equivalent.
To investigate the modulation of the electronic properties
of the anionic tether, the ylides 2a−c were synthesized by
the addition of trifluoroacetic anhydride (TFAA) to the cor-
responding salts 1a−c (Scheme 2). Again, rapid treatment with
an aqueous saturated solution of NaHCO3 following the
addition of the electrophile led to the formation of the desired
ylides in very good to quantitative yields. In contrast to the acyl
chlorides, the trifluoroacetyl ester resulting from the reaction of
TFAA with the 2,4-dinitrophenolate counterion is electrophilic
enough to react with the N-amino salt. As a result, only 1.5 equiv
of TFAA was required to obtain optimal yields.
For the ylide formation step, we elected to develop condi-
tions with broader scope, as the previously reported synthesis
of ylide 2d was difficult to expand to a wide range of acyl
chlorides (i.e., benzoyl chloride as the solvent, 90 °C, 12 h).11
A protocol relying on a near-stoichiometric quantity of the acyl
chloride was developed, inspired by conditions optimized for
the synthesis of N-acyliminopyridinium ylides.12 The results are
summarized in Table 1.
The structures of ylides 2a,d were confirmed by X-ray
diffraction analysis (Figures 1 and 2). In all cases, the N3−N4
Table 1. Protocol Optimization for the Synthesis of
N-Acyliminoimidazolium Ylides
c
d
entry
R
base
time (min)
yield (%)
a
1
2
3
4
5
6
Ph
NaOH
NaOH
60
10
10
5
20
35
75
95
90
96
a
Ph
b
Ph
K2CO3
b
b
b
Figure 1. ORTEP plot of ylide 2a. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(3)−N(4) =
1.40, N(3)−C(2) = 1.34, C(2)−N(1) = 1.34, C(5)−N(4) = 1.32,
C(5)−O(6) = 1.24; N(1)−C(2)−N(3) 107.5.
Ph
NaHCO3
NaHCO3
NaHCO3
4-F-Ph
t-Bu
5
5
a
b
A 1 M aqueous solution was used. A saturated aqueous solution
c
was used. Time allowed for reaction with the aqueous base solution.
d
Isolated yield.
In contrast to the pyridinium-based ylides, the N-acylimino-
imidazolium ylides were found to be very sensitive to basic
conditions, leading to heavy decomposition in the presence of
sodium hydroxide (commonly used for the synthesis of the
former), regardless of reaction times (entries 1 and 2, Table 1).
It was not possible to fully characterize specific degradation
products, but this drastic difference in the stability of
imidazolium-based ylides in comparison to their pyridinium
analogues seems to originate from the more reactive 2-position
on the imidazolium ring. As it was found that a base is required
for the reaction to proceed, weaker bases were tested.
Gratifyingly, the use of either K2CO3 (entry 3, Table 1) or
NaHCO3 (entry 4, Table 1) did lead to a noticeable increase in
yield, the latter affording an excellent yield (95%) of ylide 2d.
The optimized reaction conditions were applied to other
acyl chlorides, to afford ylides 2e,f in similar yields. In all cases,
2 equiv of acyl chloride is required to obtain optimal yields.
Figure 2. ORTEP plot of ylide 2d. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(3)−N(4) =
1.41, N(3)−C(2) = 1.34, C(2)−N(1) = 1.34, C(5)−N(4) = 1.33,
C(5)−O(6) = 1.26; N(1)−C(2)−N(3) = 108.4.
bond lengths are indicative of a mostly single-bond character.
Although charge delocalization can either be endocyclic or
exocyclic, the structures illustrate that the negative charge of
the ylide is mostly located on the exocyclic moiety (Scheme 3).
The electronic effect of the acyl groups seems to be negligible
by comparing the structural features of ylides 2a and 2d.
Scheme 2. Synthesis of N-Trifluoroacetyliminoimidazolium Ylides
1989
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
acetic acid was removed by azeotropic distillation in refluxing
benzene through a solid Na2CO3 trap. Complexes 3c,f were
successfully obtained by this method in yields of 90% and 88%,
respectively (Scheme 5).
Scheme 3. Possible Resonance Forms of Anionic Charge
Delocalization
Scheme 5. Alternate Method for the Synthesis of Silver(I)
Complexes
DFT calculations13 were done on ylides 2a and 2d to determine
the extent of charge separation (ESP charges). On ylide 2d, 78%
of the negative charge is found on the exocyclic moiety. In the
case of ylide 2a, 88% of the negative charge is on the exocyclic
moiety. This small variation in charge separation is in accord
with the structural properties mesured by X-ray crystallography.
The NMR chemical displacements of the C2 proton on the
imidazole ring of ylides 2a,d are 9.89 and 9.68 ppm, res-
pectively. Similar displacements are observed for ylides 2b,c,e,f.
These values are very similar to displacements observed for
classical imidazolium salts. This is again indicative of the very
small contribution of endocyclic resonance to these ylides.
Synthetic Aspects of the Complexes. It is well recognized
that NHC−silver complexes are proficient ligand transfer
agents.1b This is even more relevant in the case of anionic
NHC−silver complexes, as the transfer reaction results in the
direct exchange of a halogen group, leading to facile and
irreversible silver halide salt precipitation (vide infra). For this
reason, we elected to develop preparative methods to access
silver complexes from this new family of ylide proligands.
Previously optimized conditions12 for the formation of silver
complex 3d from ylide 2d proved to be unsuccessful for most
of the new ylides synthesized, leading to low conversions to
the corresponding silver complexes. Other conditions were
attempted, and gratifyingly metalation involving reaction of the
ylide with silver(I) oxide14 at room temperature afforded the
corresponding silver complexes in quantitative yields in most
cases. The only drawback of the method is the long reaction
time required (48 h) to obtain complete conversion. Purification
is accomplished by simple filtration of the excess silver(I) oxide.
Complexes 3a,b,d,e were easily obtained by this method
(Scheme 4).
As previously reported, 3d is very stable to oxygen and light
and is obtained as a dimeric complex.11 In this dimeric structure,
the anionic moiety of one ligand cooperatively binds the second
silver atom through the nitrogen (N-bonded dimer). The
N-bonded and O-bonded dimers of 3d were optimized using
DFT calculations,15 and a large preference for the former is
observed (Scheme 6). In the O-bonded dimer, chelation of the
Scheme 6. Energetic Comparison of the N-Bonded and
a
O-Bonded Dimers of 3d
a
Energies were obtained at the B3LYP/6-31G(d)+LANL2DZ(Ag)
level. Thermodynamic corrections are unscaled.
anionic moiety oxygen is absent, with Ag−O(b) distances of
over 3.0 Å.
Analogously to 3d, most complexes were found to be solids
particularly stable to oxygen and light. By slow diffusion of
hexanes in concentrated dichlororomethane or benzene solutions
of the complexes, crystals suitable for X-ray diffraction analysis
were obtained for four of the new complexes. Structural analyses
show most complexes to be dimers similar to 3d: 3a (Figure 3),
Scheme 4. Synthesis of Silver(I) Complexes
Unfortunately, this mild metalation method resulted in
significant degradation for ylides 2c,f, both bearing a tert-butyl
group on either the imidazole or the acyl group. An alternative
procedure was thus developed, involving silver acetate.
Although the method solved the degradation problem, only
low conversions were obtained. We hypothesized that the acetic
acid formed in the reaction could result in protodemetalation,
preventing complete conversion. Addition of various bases in the
reaction mixture unfortunately resulted also in low conversions.
To drive the reaction forward without the use of an in situ base,
Figure 3. ORTEP plot of complex 3a. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(1)−C(2) =
1.36, C(2)−N(3) = 1.34, N(3)−N(4) = 1.42, N(4)−C(5) = 1.35,
C(5)−O(6) = 1.21, C(6)−Ag(8) = 2.07, N(4)−Ag(7) = 2.15, Ag(7)−
Ag(8) = 2.79; C(2)−Ag(8)−Ag(7) = 80.4, N(4)−Ag(7)−Ag(8) = 87.4,
Ag(7)−N(4)−C(2)−Ag(8) = 16.6.
1990
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
3b (Figure 4), and 3f (Figure 5). One complex, however, was
found to be a trimer: 3c (Figure 6). This is particularly interesting,
Structurally speaking, the dimers are symmetrical structures,
with the two ligands in identical chemical environments.
Although no formal Ag−Ag bond can exist in the dimers, they
all show significant interactions between the two silver atoms,
considering the sum of their van der Waals radii (3.4 Å). These
interactions should diminish the reactivity of the coordinatively
unsaturated (14e) silver centers and explain the surprisingly
high stability of these complexes. Another particular feature
is the almost negligible effect of the exocyclic moiety on the
resulting coordination bond lengths. A comparison of C2−Ag8
and N4−Ag7 bond lengths in complexes 3a,b,f shows almost no
differences, despite large variation in terms of the steric bulk
(3a vs 3b) and electronic properties of the exocyclic moiety
(3a vs 3f). These similarities could explain why all these
complexes present similar stabilities. One noticeable variation
observed by comparing 3a and 3b is the orientation of the
fluorine atoms on the trifluoroacetyl groups. This particular
orientation is observed in 3b to minimize steric interactions
with the isopropyl groups.
In contrast to the other complexes, a trimer was observed for
complex 3c (Figure 6), where the three ligands and three silver
atoms are disposed to form 12-membered metallacycles. The
crystal structure shows the three ligands in different chemical
environments, with specific bond angles and torsion angles,
while keeping a linear geometry about the silver atom. In
solution, a very simple 1H NMR spectrum is obtained, however,
indicative of rapid structural rearrangement. The trimeric
structures also have an important consequence for the Ag−Ag
interactions. For complex 3c, the Ag−Ag distances measured
are much longer than those observed for the dimers, with
values of 3.10, 3.21, and 3.94 Å. These weaker Ag−Ag inter-
actions could explain why the trimer is less stable and more
prone to degradation in comparison to the dimers 3a,b,f. The
other important differences between the dimeric and trimeric
complexes are the Ag7−N4-C2−Ag8 dihedral angles. While
these angles in the dimers are less than 17°, they vary between
45 and 67° in trimer 3c. In order to explain the preference of
ylide 2c to form the silver trimer 3c, we turned to DFT cal-
culations.15 The free energies of reaction for the interconver-
sion between the dimer and trimer forms of the silver
complexes studied were computed for all ylides. The results
are summarized in Table 2.
Figure 4. ORTEP plot of complex 3b. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(1)−C(2) =
1.39, C(2)−N(3) = 1.36, N(3)−N(4) = 1.43, N(4)−C(5) = 1.33,
C(5)−O(6) = 1.23, C(2)−Ag(8) = 2.05, N(4)−Ag(7) = 2.17, Ag(7)−
Ag(8) = 2.77; C(2)−Ag(8)−Ag(7) = 79.6, N(4)−Ag(7)−Ag(8) = 90.7,
Ag(7)−N(4)−C(2)−Ag(8) = 10.8.
Figure 5. ORTEP plot of complex 3f. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(1)−C(2) =
1.36, C(2)−N(3) = 1.35, N(3)−N(4) = 1.42, N(4)−C(5) = 1.35,
C(5)−O(6) = 1.23, C(2)−Ag(8) = 2.08, N(4)−Ag(7) = 2.14, Ag(7)−
Ag(8) = 2.77; N(1)−C(2)−N(3) = 103.9, C(2)−Ag(8)−Ag(7) = 80.3,
N(4)−Ag(7)−Ag(8) = 89.7, Ag(7)−N(4)−C(2)−Ag(8) = 12.3.
Table 2. Free Energies of Reaction of the Dimer−Trimer
Interconversion Processes for Different Ligands
a
Figure 6. ORTEP plot of complex 3c. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): C(2)−N(1) =
1.36, C(2)−N(3) = 1.37, N(3)−N(4) = 1.42, N(4)−C(5) = 1.33,
C(5)−O(6) = 1.25, C(2)−Ag(8) = 2.08, N(4)−Ag(7) = 2.13, Ag(7)−
Ag(8) = 3.94, Ag(8)−Ag(9) = 3.21, Ag(7)−Ag(9) = 3.10; Ag(7)−
N(4)−C(2)−Ag(8) = −67.1.
entry
R1
R2
ΔGdi-tri (kcal/mol)
1
2
3
4
5
Mes
Mes
Mes
t-Bu
Ph
+6.5
+9.7
+0.7
+2.9
-6.1
CF3
CF3
CF3
2,6-(i-Pr)2Ph
t-Bu
as reports of confirmed trimeric silver complexes are scarce.16 In
either the dimers or trimer, cooperative N-bonding of the anionic
moiety is observed.
a
Energies were calculated at the B3LYP/6-31+G(d,p)+LANL2DZ-
(Ag)//B3LYP/6-31G(d)+LANL2DZ(Ag) level. Thermodynamic cor-
rections are unscaled.
1991
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
With the exception of entry 5, the dimeric forms are favored
in all cases. It is worth noting, however, that for entries 3 (ylide
2a), and 4 (ylide 2b), the free energy values computed are fairly
small, and hence the trimeric form could be an accessible or
observable state at room temperature. In the case of entry 5,
related to ylide 2c, a definite preference for the trimeric form
is observed. The results presented in Table 2 thus reproduce
well the preference observed by X-ray structural analysis. The
preferred trimeric structure of complex 3c could be explained by
the steric properties of both the N-substituent of the imidazole
and the exocyclic anionic moiety. While a mesityl group is
considered to be sterically demanding, its steric hindrance is
mostly perpendicular to the imidazole ring. In contrast, the tert-
butyl group possesses a spherical steric hindrance (Figure 7).
of an anionic halogen ligand is possible through irreversible
precipitation of a silver halide salt. Complex 3d was previously
used for the synthesis of a Cu(II) complex.11 To evaluate the
potential of the new complexes bearing a trifluoroacetyl group,
transmetalation was performed with silver complex 3a to form
nickel complex 4a (eq 2). The transmetalation process is mild,
and only simple filtration of the insoluble silver halide salt is
required to obtain pure complex 4a in quantitative yield.
An X-ray structural analysis of complex 4a confirmed its
monomeric nature and the square-planar geometry at the
Ni(II) center (Figure 8). In this case, the anionic NHC ligands
Figure 7. Rationale for the destabilization of the dimeric silver
complex from ylide 2c.
As such, the combination of both the tert-butyl and the
trifluoroacetyl groups on ylide 2c seems to be sufficient to push
the equilibrium toward the trimer.
1
Analyses of silver complexes 3d,e by H NMR are indicative
of a very symmetric system, showing only one set of signals
for both ligands and displaying no restrained rotation of either
the mesityl group or the exocyclic benzoyl group. No crystal
structure was obtained for complex 3e, but the strong
preference for a dimeric structure can be assumed due to its
high similarity with the spectroscopic data of complex 3d.
Additionally, mass spectroscopy did show the exclusive
presence of the dimer. Interestingly, although only the dimers
were present in the X-ray crystal structures of complexes 3a,b,f,
two distinct sets of signals by 1H NMR in solution are observed.
To determine if these signals are the result of restrained rotation
or species interconversion, the effect of concentration was
investigated on complex 3a. A 0.015 M CDCl3 solution of
complex 3a presented two distinct sets of signals, while a 0.25 M
CDCl3 solution resulted in the observation of only one set of
signals, reminiscent of the highly symmetric system of 3d. The
effect of concentration is a clear indication of species inter-
conversion in solution. A rationale can be found from the free
energy of reaction calculated for the dimer−trimer interconver-
sion of complex 3a (Table 2, entry 3). A very small preference
(0.7 kcal/mol) for the dimer is predicted. The dimeric and
trimeric complexes would then be energetically accessible at
room temperature, and dynamic equilibrium could be possible.
The interconversion rate would be directly proportional to
concentration, explaining coalescence of the two sets of signals
at higher concentration. Considering that the free energy values
in Table 2 are all fairly small, it is reasonable to assume that both
multimeric forms could be accessible at room temperature. The
highest preference found is for complex 3d (entry 2, Table 2,
+9.7 kcal/mol), in accord with the exclusive presence of one set
Figure 8. ORTEP plot of complex 4a. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(1)−C(2) =
1.35, C(2)−N(3) = 1.35, N(3)−N(4) = 1.38, N(4)−C(5) = 1.28,
C(5)−O(6) = 1.29, O(6)−Ni(7) = 1.85, Ni(7)−C(2) = 1.91; C(2)−
Ni(7)−O(6) = 90.9.
chelate the nickel center through internal O-bonding. This
illustrates another interesting feature of this family of ligands;
the anionic moiety possesses multiple binding sites to
accommodate different geometries.17
The reactivity of the silver complexes toward an oxidizing
agent such as iodine was evaluated.18 Treatment of complexes
3a,c with 1 equiv of molecular iodine per silver atom led to
rapid iododemetalation at room temperature, affording the
iodo-substituted ylides 5a,c in excellent yields (Scheme 7).
Scheme 7. Iododemetalation of Complexes 3a,c using
Molecular Iodine
1
of signals in H NMR, regardless of concentration.
The structure of ylide 5c was confirmed by X-ray diffraction
analysis (Figure 9). These derivatives could potentially be
exploited for metal complex synthesis through oxidative insertion
into the C−I bond.
Reactivity of the Complexes. As stated above, silver−
NHC complexes are proficient ligand transfer agents. They are
especially useful for anionic NHC ligands, as formal exchange
1992
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
pharmaceutical compounds.22 The mechanism is described in
Scheme 8.
Scheme 8. Mechanism of the Silver-Mediated (3 + 2)
Heterocycloaddition of Azomethine Ylides
Figure 9. ORTEP plot of ylide 5c. Ellipsoids are drawn at the 30%
probability level. Selected distances (Å) and angles (deg): N(1)−C(2) =
1.36, C(2)−I(7) = 2.09, C(2)−N(3) = 1.34, N(3)−N(4) = 1.43, N(4)−
C(5) = 1.31, C(5)−O(6) = 1.25; C(2)−N(3)−N(4)−C(5) = 83.9.
Ligand Exchange of the Complexes. Ligand exchange
is an alternative strategy to synthesize new complexes. Various
ligands have been reported to displace NHC ligands. For
example, being strong σ-donors, phosphines19 and other
Numerous groups have developed enantioselective variants
of this reaction, traditionally using chelating ligands based on
amine and phosphine combined with commercially available
silver salts. Because of the instability of the formed complexes,
they are generated in situ prior to reaction. Additionally, an
external base is usually required to facilitate the azomethine
ylide formation (Scheme 8). To the best of our knowledge,
silver−NHC complexes have never been employed for this
reaction.
Silver complexes 3a−f were evaluated as catalysts for the
(3 + 2) cycloaddition of imino ester 6 with dimethyl maleate,
to produce pyrrolidine 7. Silver(I) oxide and silver acetate were
evaluated for the sake of catalytic reference. They are known
to be highly active, due to the absence of donating ligands
reducing their Lewis acidity. The results are summarized in
Table 3. Due to the anionic nature of the NHC ligands, no
NHCs19,20 can serve to displace NHC ligands. However, the
b
displacing ligand is fully formed (i.e., free NHC), in all reported
cases. This involves, in the case of NHCs, the use of a strong
base to deprotonate the imidazolium salt proligand. Due to the
apparent dynamic nature of the silver complexes obtained from
the ylides, and the anionic nature of the resulting NHC ligands,
we envisioned that ligand−proligand exchange could be
possible. This would result in a simplified ligand exchange
strategy. To test this hypothesis, complex 3a was mixed with
2 equiv of ylide 2d in CDCl3 at room temperature. Gratifyingly,
complete exchange on the silver centers was observed, leading
to complete conversion in 22 h to complex 3d and 2 equiv of
ylide 2a (eq 3). To the best of our knowledge, this is the first
Table 3. Evaluation of Catalytic Activity of the Silver
Complexes in the Azomethine Ylide Cycloaddition
a
b
entry
catalyst
loading (Ag mol %)
time (h)
yield (%)
1
2
3
4
5
6
7
8
Ag2O
AgOAc
3a
4
4
4
4
4
4
4
4
0.75
1
91
89
93
90
92
96
93
92
3
3b
3.5
0.5
12
9
example of a direct NHC ligand−proligand exchange, without
the need for formation of a formal carbene prior to exchange.
The conversion toward complex 3d is explained by the stronger
donating ability of the benzoyl-substituted exocyclic anionic
moiety, in comparison to the trifluoroacetyl-substituted group.
This effectively leads to a better stabilization of coordinatively
unsaturated silver(I) centers. It is another demonstration of the
easy modulation of the electronic properties of the ligands.
Complexes in Catalysis. While silver−NHC complexes are
commonly used as ligand transfer agents, their use in catalytic
applications is scarce.1b Nonetheless, homogeneous silver-
mediated catalysis is broad and offers great potential, owing to
the Lewis acidity and oxophilicity of silver.20 Due to the
particularly high stability and well-defined structures of the silver
complexes synthesized, we investigated their potential as catalysts
in a model reaction in order to compare their activity. Silver is an
excellent transition metal for the catalysis of certain cycloaddi-
tion reactions.21 For example, the (3 + 2) heterocycloaddition of
in situ generated azomethine ylides with electrophilic alkenes
generates polysubstituted pyrrolidines, known to be important
3c
3d
3e
3f
12
a
Reaction time required to observe >95% conversion by GC/MS.
Isolated yield.
b
additional base was required. In all cases, reactions are very
clean and only the endo adduct 7 is observed. Lewis acidity
has a drastic effect on activity. Catalysts with more donating
anionic moieties (3d−f) require long reaction times to achieve
complete conversions. For similar N-substituents on the
imidazole, replacement of the benzoyl group for a trifluor-
oacetyl group on the anionic tether has a noticeable effect on
activity, as can be seen by comparing catalysts 3a,d (entry 3 vs
entry 6, Table 3). Complex 3c, possessing weakly coordinating
anionic groups and no Ag−Ag interactions due to its trimeric
nature in the solid state, was found to be the most active
catalyst (entry 5, Table 3). The topology of the complexes
1993
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994

Organometallics
Article
(3) (a) Praetorius, J. M.; Crudden, C. M. RSC Catal. Ser. 2011, 6,
77−118. (b) Mata, J. A.; Poyatos, M.; Peris, E. Coord. Chem. Rev. 2007,
251, 841−859.
(4) (a) Liddle, S. T.; Edworthy, I. S.; Arnold, P. L. Chem. Soc. Rev.
2007, 36, 1732−1744. (b) Liddle, S. T.; Arnold, P. L. Chem. Commun.
2006, 3959−3971.
could thus also have an important effect on activity. With their
stability to light and moisture and well-defined structures,
these new silver complexes demonstrate the potential to access
well-defined, bench-stable, chiral catalysts for this type of
transformation.
(5) (a) Grubbs, R. H.; Endo, K. J. Am. Chem. Soc. 2011, 133, 8525−
8527. (b) Patel, D.; Liddle, S. T.; Mungur, S. A.; Rodden, M.; Blake, A.
J.; Arnold, P. L. Chem. Commun. 2006, 1124−1126. (c) Specht, Z. G.;
Cortes-Llamas, S. A.; Tran, H. N.; van Niekerk, C. J.; Rancudo, K. T.;
Golen, J. A.; Moore, C. E.; Rheingold, A. L.; Dwyer, T. J.; Grotjahn, D.
B. Chem. Eur. J. 2011, 17, 6606−6609.
(6) (a) Arnold, P. L.; Scarisbrick, A. C. Organometallics 2004, 23,
2519−2521. (b) Waymouth, R. M.; Ketz, B. E.; Cole, A. P.
Organometallics 2004, 23, 2835−2837. (c) Clavier, H.; Coutable, L.;
Toupet, L.; Guillemin, J.-C.; Mauduit, M. J. Organomet. Chem. 2005,
690, 5237−5254.
(7) (a) Shen, Q.; Yao, H.; Zhang, Y.; Sun, H. Eur. J. Inorg. Chem.
2009, 1920−1925. (b) Li, W.; Sun, H.; Chen, M.; Wang, Z.; Hu, D.;
Shen, Q.; Zhang, Y. Organometallics 2005, 24, 5925−5928. (c) Dyson,
G.; Frison, J.-C.; Simonovic, S.; Whitwood, A. C.; Douthwaite, R. E.
Organometallics 2008, 27, 281−288.
CONCLUSIONS
■
In summary, optimized synthetic methods were developed in
order to access a broader family of N-acyl iminoimidazolium
ylide proligands. Six new anionic NHC−Ag(I) complexes were
synthesized and characterized. These surprisingly stable and
well-defined silver complexes have led to a better understanding
of the structure−property relationship of this class of ligand. The
complexes can serve as ligand transfer agents, as exemplified by
the rapid synthesis of a nickel(II) complex, to study their
properties on a variety of transition metals. The unprecedented
ligand−proligand exchange opens the door to alternative metal
complex synthesis methods and interesting self-assembly
systems. These new directions are currently under investigation.
While stable and isolable, the complexes have demonstrated
good catalytic activity for the silver-mediated (3 + 2) hetero-
cycloaddition of azomethine ylides with electrophilic alkenes.
The development of chiral variants of these ylides, as well as
the investigation of possible catalytic applications, is currently
underway in our laboratory.
(8) (a) Arnold, P. L.; Edworthy, I. S.; Blake, A. J.; Wilson, C.
Organometallics 2007, 26, 3684−3689. (b) Arnold, P. L.; Liddle, S. T.;
McMaster, J.; Jones, C.; Mills, D. P. J. Am. Chem. Soc. 2007, 129,
5360−5361. (c) Douthwaite, R. E.; Houghton, J.; Kariuki, B. M. Chem.
Commun. 2004, 698−699. (d) Fryzuk, M. D.; Spencer, L. P.; Winston,
S. Organometallics 2004, 23, 3372−3374.
(9) Sakaguchi, S.; Kawakami, M.; O’Neill, J.; Yoo, K. S.; Jung, K. W. J.
Organomet. Chem. 2010, 695, 195−200.
ASSOCIATED CONTENT
■
(10) (a) Hamilton, P. J. C.; Danopoulos, A. A.; Downing, S. P.;
Pogorzelec, P. J. Eur. J. Inorg. Chem. 2009, 1816−1824. (b) Sun, H.
M.; Hu, D. M.; Wang, Y. S.; Shen, Q.; Zhang, Y. J. Organomet. Chem.
2007, 692, 903−907.
S
* Supporting Information
Text, tables, figures, and CIF files giving details of the syntheses
and characterization data and NMR spectra for all new
compounds, selected crystallographic data for compounds
2a,d, 3a−c,f, 4a, and 5c, the full Gaussian reference, Cartesian
coordinates, and electronic and zero-point vibrational energies.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(11) Charette, A. B.; Legault, C. Y.; Kendall, C. Chem. Commun.
2005, 3826−3828.
(12) Charette, A. B.; Legault, C. J. Org. Chem. 2003, 68, 7119−7122.
(13) Structures were optimized at the B3LYP level, using the 6-
31+G(d,p) basis set. See the Supporting Information for details.
(14) (a) Eastham, G.; Kleinhenz, S.; Tulloch, A. A. D.; Danopoulos,
A. A.; Winston, S. Dalton Trans. 2000, 4499−4506.
AUTHOR INFORMATION
(15) See the Supporting Information for details.
■
(16) (a) Flores, J. A.; Komine, N.; Pal, K.; Pinter, B.; Pink, M.; Chen,
C.-H.; Caulton, K. G.; Mindiola, D. J. ACS Catal. 2012, 2, 2066−2078.
(b) Dominique, F. J.-B.; Gormitzka, H.; Hemmert, C. J. Organomet.
Chem. 2008, 693, 579−583. (c) Slaughter, L. M.; Khan, M. A.;
Wanniarachchi, Y. A. Organometallics 2004, 23, 5881−5884.
(17) Samantaray, M. K.; Pang, K.; Shalkh, M. M.; Ghosh, P. Inorg.
Chem. 2008, 47, 4153−4165.
(18) Fu, C.-F.; Lee, C.-C.; Liu, Y.-H.; Peng, S.-M.; Warsink, S.;
Elsevier, C. J.; Chen, J.-T.; Liu, S.-T. Inorg. Chem. 2010, 49, 3011−
3018.
Corresponding Author
*E-mail for C.Y.L.: claude.legault@usherbrooke.ca.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Natural Sciences and
Engineering Research Council (NSERC) of Canada, the Fonds
(19) (a) Zeller, A.; Bielert, F.; Haerter, P.; Herrmann, W. A.;
Strassner, T. J. Organomet. Chem. 2005, 690, 3292−3299. (b) Titcomb,
L. R.; Caddick, S.; Cloke, F. G. N.; Wilson, D. J.; McKerrecher, D.
Chem. Commun. 2001, 1388−1389. (c) Vieille-Petit, L.; Clavier, H.;
Linden, A.; Blumentritt, S.; Nolan, S. P.; Dorta, R. Organometallics
2010, 29, 775−788.
́
Quebecois de la Recherche-Nature et Technologies (FQRNT),
the Canada Foundation for Innovation (CFI), the FQRNT
Centre in Green Chemistry and Catalysis (CGCC), and the
Universite
provided by the Res
́
de Sherbrooke. Computational resources were
́
eau quebecois de calcul de haute performance
́
́
(RQCHP). We thank Daniel Fortin for the X-ray analysis and
A. B. Charette and Michel Gravel for useful discusssions.
(20) Naodovic, M.; Yamamoto, H. Chem. Rev. 2008, 108, 3132−
3148.
(21) (a) Pale, P.; Blanc, A.; Weibel, J.-M. Chem. Rev. 2008, 108,
́
́
́
3149−3173. (b) Rodrıguez-Garcıa, I.; Munoz-Dorado, M.; Alvarez-
REFERENCES
■
Corral, M. Chem. Rev. 2008, 108, 3174−3198.
(1) (a) Mathew, J.; Cherumuttathu, H. S. Inorg. Chem. 2010, 49,
4665−4669. (b) Garrison, J. C.; Youngs, W. J. Chem. Rev. 2005, 105,
3978−4008. (c) Jacobsen, H.; Correa, A.; Poater, A.; Costabile, C.;
Cavallo, L. Coord. Chem. Rev. 2009, 253, 687−703.
(2) (a) Herrmann, W. A. Angew. Chem., Int. Ed. 2002, 41, 1290−
1309. (b) Hahn, F. E.; Jahnke, M. C. Angew. Chem., Int. Ed. 2008, 47,
3122−3172.
̀
(22) Sansano, J. M.; Najera, C. Heterocycl. Chem. 2008, 12, 117−145.
1994
dx.doi.org/10.1021/om400078f | Organometallics 2013, 32, 1988−1994