4,5-bis(dialkylamino)-substituted imidazolium systems: Facile access to N-heterocyclic carbenes with self-umpolung option

Article abstract of DOI:10.1002/chem.201101999

The first synthesis of 4,5-bis- (dimethylamino)-substituted imidazolium compounds was developed, which is based on the reaction of a 1,2-diamino-1,2-bis(phosphonio)ethene with lithiated formamidines. This represents the first application of this class of

Full text of DOI:10.1002/chem.201101999

DOI: 10.1002/chem.201101999
4,5-Bis(dialkylamino)-Substituted Imidazolium Systems: Facile Access to
N-Heterocyclic Carbenes with Self-Umpolung Option
Stefan M. Huber,*[a] Frank W. Heinemann,[b] Pierre Audebert,[c] and Robert Weiss*[a]
Dedicated to Professor Robert West
Abstract: The first synthesis of 4,5-bis-
(dimethylamino)-substituted imidazoli-
um compounds was developed, which
is based on the reaction of a 1,2-diami-
NHCs are capable of self-umpolung if
sufficiently strong acceptor substituents
are introduced at the carbene center.
To test the self-umpolung capabilities
of the NHCs, various substituents were
attached to the carbene center and the
obtained compounds were character-
ized by single-crystal X-ray analysis as
well as quantum chemical computa-
tions. Strong acceptor substituents are
required to induce self-umpolung, such
as in the phosphonio-substituted deriv-
ative, for which partial self-umpolung
was found. The N,N’-bis(4-dimethyl-
no-1,2-bis(phosphonio)ethene
with
ACHTUNGTRENNaUNG minophenyl)-substituted imidazolium
lithiated formamidines. This represents
the first application of this class of
ethene derivatives for the preparation
of heterocycles. These N-heterocyclic
carbene (NHC) precursors show a re-
markably reduced basicity and nucleo-
philicity of their NMe2 groups, which is
due to the strong anomeric interactions
of the latter with the imidazolium core.
According to DFT calculations, these
compound represents a special case, as
it incorporates as much as three two-
step redox systems within the NHC
framework. This will probably result in
a high electronic flexibility of the cor-
responding nucleophilic carbenes, espe-
cially when they serve as ligands in
transition metal complexes.
Keywords: carbenes · cyclization ·
heterocycles · ligand design · redox
chemistry
Introduction
In previous work[1–5] a central focus of our research has been
on the development of a novel type of nucleophilic car-
benes, which incorporate the option of “self-umpolung”.
These systems should be able to adapt chameleon-like to
the electronic demands of an organic or inorganic substitu-
ent attached to them.[3a,4] The formal conceptual basis for
this is a “short-circuit-like” connection through p conjuga-
tion between a singlet carbene center OX’ and a two-stage
redox module in its RED form[6] according to Figure 1.
Structure A is in resonance with the electronically comple-
mentary structure B, in which the “carbene” carbon center
Figure 1. Schematic depiction of the general structural motif of a “self-
umpolung” carbene (and its derivatives).
is formally reduced to a di-ide by the electron-rich p
module (Figure 1).[3a,4,5]
[a] Dr. S. M. Huber, Prof. Dr. R. Weiss
Department of Chemistry und Pharmacy
Organische Chemie, Henkestrasse 42
91054 Erlangen (Germany)
This “redox resonance” indicates an electronic flexibility
of this type of carbenes, which in principle allows stabilizing
p interactions with both electron-donating and electron-ac-
cepting partners, as shown by A’ and B’ (being the extreme
cases, with intermediate electronic demands forming a con-
tinuous spectrum in-between A’ and B’).
In the past we have conceived carbenes 1–3 (Scheme 1),
which exemplify this concept. At the same time, the dinitr-
oxide-carbene 4 and its radical anion 5 already represent a
conceptual extension, as in these cases the carbene center is
coupled with two two-stage redox systems through p conju-
gation (under avoidance of cyclic p conjugation).
E-mail: weiss@chemie.uni-erlangen.de
stefan.m.huber@tum.de
[b] Dr. F. W. Heinemann
Department of Chemistry und Pharmacy
Anorganische Chemie, Egerlandstrasse 1
91058 Erlangen (Germany)
[c] Prof. Dr. P. Audebert
PPSM (CNRS UMR 8531)
Ecole Normale Supꢀrieure de Cachan
61, avenue de Prꢀsident Wilson, 94235 Cachan Cedex (France)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201101999.
13078
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13078 – 13086

FULL PAPER
Scheme 1. Closed-shell and open-shell examples of nucleophilic carbenes
with „self-umpolung“ option.
Considering the importance and prevalence of imidazole-
based N-heterocyclic carbenes (NHCs) in various areas of
modern chemistry,[7] we subsequently set out to develop a
structurally variable and synthetically efficient synthesis of
imidazolin-2-ylidenes with built-in p-redox modulation.
Prima facie, the 4,5-bis(dialkylamino)imidazolinylidene 6
(Scheme 2) fulfills the criteria for “NHC-chameleons”, as
defined above. Electronic structure calculations confirm this
view (see below).
Scheme 3. Synthesis of 4,5-bis(dimethylamino)-substituted imidazolium
systems. i) PPh3 (2 equiv), trimethylsilyl trifluoromethanesulfonate
(TMSOT, 2 equiv); ii) RN(Li)C(H)NR.
as d0,d2-bisnucleophiles (Scheme 3). The dicationic salt 8
serves as a reagent for the ethene-bis(dimethyliminium) syn-
[14]
thon 10 and thus indirectly also for the elusive C2O2 (11).
This synthetic route constitutes a novel and direct access to
the imidazolium core, which does not involve a quaterniza-
tion of an imidazole intermediate.
Interestingly, although 9a was also detected by mass spec-
trometry when using the ethene derivative 7 instead of 8 as
bis-electrophile in the reaction with the respective formami-
dine, the product seemed to have been formed in low yield
only and could not be isolated. This might in part be due to
the very reactive nature of 7, which necessitates its prepara-
tion in situ and thus makes it difficult to add the formami-
dine anion in an exact 1:1 stoichiometry. In addition, the bi-
s(phosphonio)derivative 8 might also offer an alternative
mechanistic route to the product that is not feasible (or not
as favorable) with ethene 7. After addition of the formami-
dine and stirring for several hours at room temperature, in-
termediate 14 (Scheme 4) was detected by mass spectrome-
try, whereas substitution of the second phosphane ligand re-
quired thermal activation. Formation of intermediate 14 can
be accomplished either by an elimination–addition route via
the ketene iminium 13 or by an addition–elimination path-
way via the phosphonium ylide 12. Although we cannot rule
out one of these alternatives based on the current experi-
mental data, the ylidic stabilization of the negative charge in
12 renders this intermediate much more favorable than a
hypothetical dichloro analogue (starting from 7). The substi-
tution of the second phosphane ligand should proceed via
the ketene iminium intermediate 15, as an addition–elimina-
tion pathway would require the localization of negative
charge between two amine substituents at the ethene back-
bone.
Scheme 2. Structural characteristics of 4,5-bis(dialkylamino)imidazolin-2-
ylidene 6.
In the limiting case I, this type of molecule can be con-
ceived as a cyclic p-conjugated complex between a (strongly
reducing)[8] tetraaminoethene substructure and divalent
carbon. In the redox-complementary resonance structure II,
these compounds can be considered as special examples of
bisonio-diides, which have been the subject of intense dis-
cussion lately.[9] The latter structure (II) points to a further
interpretation of 6, according to which this nucleophilic car-
bene could be viewed as a chelate complex between an oxa-
1
lylic acid bisamidine and a D carbon atom.[10]
Although several 4,5-bis(dialkylamino)-substituted, N-
monosubstituted imidazoles are known,[11] selective alkyla-
tions at the N3 position would pose problems due to the
presence of the two amino groups.[12] Herein, we report on
the synthesis of the previously unknown 4,5-bis(dimethyl-
ACHTUNGTRENNUNGamino)imidazolium system 9, the deprotonation of which
allows a general access to the desired class of compound 6.
Results and Discussion
The decisive building block for the synthesis of 9 turned out
to be the bis(phosphonio)diaminoethene 8, which prepara-
tion we recently reported.[13] As an a1,a2-biselectrophile that
is easy to handle and dose, ethene 8 can be conveniently
converted to the desired imidazolium system 9 by a novel
[3+2] cyclodisubstitution reaction with formamidine anions
Investigations by cyclic voltammetry yielded a reversible
oxidation wave for 9a at 0.95 V (vs. a saturated calomel
electrode (SCE)) as well as an irreversible reduction wave
at ꢀ0.86 V (vs. SCE). Compound 9c includes two further, p-
Chem. Eur. J. 2011, 17, 13078 – 13086
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13079

S. M. Huber, R. Weiss et al.
Figure 2. Molecular structure of 9c.[36] The counteranion and hydrogen
atoms are omitted for clarity. Selected bond lengths [A] and angles [8]:
ꢀ
ꢀ
ꢀ
ꢀ
N1 C1 1.325, N1 C2 1.411, C2 C3 1.362, C2 N3 1.397, N1-C1-N2 109.6,
N3-C2-C3 128.2.
This interaction is also con-
firmed by DFT calculations[18]
on a model system (which bears
methyl substituents on the imi-
dazolium nitrogen atoms):
a
Scheme 5. Anomeric n!s* in-
teraction in the imidazolium
system 9a.
natural bond orbital (NBO)
analysis[19] in combination with
a
second-order perturbation
analysis reveals a stabilization
of the molecule by about 11 kcalmolꢀ1 due to anomeric in-
Scheme 4. Possible mechanistic routes for the preparation of 9a–c.
Anions are omitted for clarity; L=PPh3.
[15]
ꢀ
ꢀ
teraction of the NMe2 lone pairs with the C5 N1 and C4
ꢀ
N3 s* orbitals. A further anomeric donation into the C4 C5
s* orbital yields an additional stabilization by about 9 kcal
molꢀ1. Such an n!s* charge transfer within the imidazoli-
um s framework has so far not been described. This effect
is—at least partially—enforced by mutual steric hindrance
of the two NMe2 groups when co-planar with the ring
system. This latter arrangement could only be approximated
with the two donor groups as part of a rigid ring system.[20]
Because of the donation of the nitrogen lone pairs of the
amino groups into the s backbone of the imidazolium core,
less electron density is available for external electrophiles.
This has noticeable consequences on the reactivity of 9a
and 9b towards electrophiles: for instance, 9a can not be
measurably protonated by either CF3COOH or one equiva-
lent of HOTf. Only by addition of a massive excess of
HOTf could protonation be achieved, but this also leads to
partial protonation of the second NMe2 group. Addition of
one equivalent of MeOTf to 9a resulted in only about 20%
methylation, apparently because an equilibrium is in place.
The unusually low reactivity of both dimethylamino
groups at the C4 and C5 positions is also reflected in the
methylation of 9c with two equivalents of MeOTf
(Scheme 6), which selectively occurs at both phenylic NMe2
groups and thus yields the air-stable tricationic species 16 as
the exclusive product in quantitative yield (for single-crystal
X-ray analysis see the Supporting Information).
phenylenediamine-like, redox systems and is spontaneously
oxidized in solution by aerial oxygen (although being rea-
sonably stable as a solid). When treating a solution of 9b or
9c in CH2Cl2 with 1.5 equivalents of SbCl5 at 08C, an instan-
taneous color change towards purple was observed, which is
indicative for the formation of the respective dication radi-
cal species. However, rapid decoloration followed, and no
defined product could be isolated.
The structures of the three products 9a–c were confirmed
by spectroscopic data and single-crystal X-ray analyses.
The crystals of 9a and 9b, grown at room temperature,
undergo a phase transition at approximately ꢀ208C and
could thus only be measured at ꢀ108C. As a consequence,
the dimethylamino groups are disordered in both cases, but
nevertheless a near-perpendicular orientation of the NMe2
groups relative to the imidazolium ring is identifiable. In the
case of (non-disordered) 9c, both NMe2 groups at the imida-
zolium core are also orientated perpendicularly to the ring
system (with the methyl groups above and below the ring
plane) and are considerably pyramidalized (Figure 2).
This spatial arrangement is characteristic for an anomeric
ꢀ
donation of the NMe2 lone pairs into the adjacent C N s*
orbitals of the imidazolium core (Scheme 5). Accordingly,
[15]
ꢀ
ꢀ
the N3 C4 and N1 C5 bond lengths of the imidazolium
Having established the synthesis of the imidazolium com-
pounds 9a–c, our next goal was to introduce various sub-
stituents at the “carbene” position C2 to test the self-umpo-
compounds 9a–c are markedly elongated compared to suita-
ꢀ
ble reference systems. More precisely, the N1 C5 bond
lengths for 9a–c lie in a range of 1.394–1.416 ꢂ, whereas
lung capabilities of these bisACTHNGUTERNNU(G amino)-substituted imidazoli-
that of 1,3-bis
G
um systems. In advance, orientating DFT calculations[18] on
a model system (bearing methyl substituents on the imidazo-
and that of its 4,5-dichloro derivative is 1.385 ꢂ.[17]
13080
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13078 – 13086

Access to N-Heterocyclic Carbenes with Self-Umpolung Option
FULL PAPER
Scheme 6. Methylation of 9c.
lium nitrogen atoms) were conducted with the prototypical
Scheme 8. Generation and exemplary derivatization of the electron-rich
carbenes 6a–c (or their lithiated forms 19a–c). Counteranions are omit-
ted for clarity. E=electrophile; R=Ph (a), 4-Me-C6H4 (b), 4-NMe2-C6H4
(c) (see text).
ꢀ
+
ꢀ
ꢀ
donor CH2 and acceptor CH2 as substituents at the C2
position to investigate the structural effects of self-umpo-
lung. Scheme 7 shows some relevant resonance structures
for these derivatives 17 and 18.
the typical range for NHCs.[21] Lithiation of 9a–c with n-bu-
tyllithium at ꢀ788C in THF generates 19a–c in quantitative
1
yield without any side products, as shown by H NMR spec-
troscopy. Consequently, the lithiated species 19a–c were
chosen as starting materials for the subsequent derivatiza-
tion with electrophiles. An overview of those electrophilic
substitution reactions on 19a–c is presented in Scheme 9.
Scheme 7. Model systems 17 and 18 for computational self-umpolung in-
vestigations.
The most pronounced structural differences between
those extreme cases are in accord with the resonance struc-
tures 17b and 18b. Thus, in a comparison of the “normal”
ꢀ
electronic situation 17 to the self-umpoled case 18, the C
[15]
ꢀ
ꢀ
NMe2 and the C4 N3/C5 N1
bond lengths are shorter
ꢀ
(1.400 vs. 1.328 and 1.416 vs. 1.345 ꢂ), whereas the C4 C5
bond length increases markedly (1.362 vs. 1.505 ꢂ). In addi-
tion, the orientation and geometry of the dimethylamino
groups changes from pyramidalized NMe2 groups that are
perpendicular to the ring system (17) to planar NMe2
groups that are as parallel to the p-ring system as (sterically)
possible (18).
Scheme 9. Electrophilic substitution reactions on 19a–c. i) CF3COOD,
TMSOTf; ii) MeI; iii) MeO-C6H4-C(O)Cl; iv) C6F6; v) Br2; vi) I2;
vii) (PPh3)2ACTHNUGTRENNUG(OTf)2. Im=4,5-bis(dimethylamino)imidazolium; Tol=tolyl.
Derivatization with electrophiles: Derivatization of the imi-
dazolium systems 9a–c at the C2 position could be achieved
by generation of their lithiated forms 19a–c (serving as syn-
thetic equivalents for carbenes 6a–c) with nBuLi and in situ
treatment with electrophiles (Scheme 8).
Carbene 6b could be generated by deprotonation of 9b
with KH in degassed THF at ꢀ208C. Its carbene carbon
atom shows a 13C NMR signal at d=190 ppm, which is in
As first simple derivatizations of 19a, the deuterated (20)
and methylated (21) compounds were prepared in 49 and
87% yield, respectively. Deuteration was achieved by treat-
ment with CF3COOD and subsequent anion exchange with
TMSOTf. The apparent incomplete nature of the exchange
might explain the overall low yield of the product, which
Chem. Eur. J. 2011, 17, 13078 – 13086
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13081

S. M. Huber, R. Weiss et al.
was isolated as the pure triflate salt. Methylation of 19a was
performed by using methyl iodide, which resulted in precipi-
tation of the pure iodide salt 21 from the reaction mixture.
Compound 21 was also characterized by single-crystal X-ray
analysis (see the Supporting Information). As expected, the
structural features of the imidazolium part of 21 are very
similar to those of 9a. The iodide counterion resides atop
the p-ring system of the cation in the solid-phase structure.
An acyl substituent served as the first p acceptor that was
introduced at the carbene position of the imidazolium sys-
tems. Accordingly, the lithiated species 19a was treated with
p-methoxybenzoylchloride at low temperatures (ꢀ508C) in
THF. After subsequent anion exchange with TMSOTf, the
triflate salt 22 was isolated in 67% overall yield. Related
carbonyl-substituted imidazolium systems are known[22] and
have been used as acylation reagents.[23] The C=O stretch
frequency of salt 22 occurs at n˜ =1664 cmꢀ1 and thus denotes
the imidazolium part as an electron donor towards the car-
bonyl group (compare C=O stretch frequencies of n˜ =1692
and 1650 cmꢀ1 for the corresponding aldehyde and amide,
respectively).[24] A single-crystal X-ray analysis of 22 (see
the Supporting Information), however, shows a near-perpen-
dicular orientation of the acyl substituent relative to the p-
ring system (torsion angle 738) and furthermore “non-self-
umpoled” structural characteristics of the imidazolium core.
In addition, a bond length of 1.525 ꢂ between the “carbene”
C atom and the carbonyl C atom as well as a C=O bond
length of 1.241 ꢂ show no indication of p donation from the
imidazolium part and are very similar to those of a related
1-acyl-4,5-bismethylated imidazolium compound.[25] Thus,
the imidazolium part acts merely as a s donor towards the
acyl moiety. DFT calculations[18] show that the perpendicular
orientation of the acyl part in 22 is mainly due to steric in-
teractions and that the formyl group as a very small acyl
substituent would orient itself almost co-planar with the ring
system. Interestingly, however, even the formyl substituent
is predicted by DFT to induce only a weak self-umpolung
effect on the imidazolium backbone.
Accordingly, we turned our attention to stronger electron
acceptors at the C2 position. Introduction of a pentafluoro-
phenyl moiety, for instance, would generate a substrate for
the SASAPOS-protocol[26] and would thus allow the intro-
duction of a pentaoniobenzene substituent at the C2 posi-
tion. Reaction of 19b with one equivalent of hexafluoroben-
zene, though, led to formation of the bis(imidazolium)deri-
vative 23, which was isolated in 29% yield (relative to hexa-
fluorobenzene). No other products were detected by mass
spectrometry. To the best of our knowledge, bis(imidazoli-
um)-substituted derivatives of hexafluorobenzene were pre-
viously unknown. Although bis(imidazolio)benzenes have
been reported,[27] they have not been characterized structur-
ally. In contrast, salt 23 was studied by single-crystal X-ray
analysis (see Figure 3; the crystal cell contains four inde-
pendent dications of 23, one of which is shown here).
Figure 3. Cation of 23 (hydrogen atoms are omitted for clarity);[36] a=
1.388, b=1.364, c=1.409, and d=1.336 ꢂ, angle of torsion (N2-C1-C23-
C24)=70.28 (mean values over all four dications).
to induce “self-umpolung by polarization”. The anions of 23
are coordinated by weak hydrogen bonds at the periphery
of the imidazolium ligands.
Subsequently, compound 23 was exposed to SASAPOS
conditions[26] to substitute all fluorine substituents by 4-di-
methylaminopyridinium ligands. Although indications for
the formation of the hexacationic product were obtained
(after six hours of heating to reflux in acetonitrile) by mass
spectrometry and elemental analysis, the isolated red solid
proved to be extremely labile to hydrolysis. Consequently,
attempts to characterize the compound by NMR spectrosco-
py and single-crystal X-ray analysis failed.
Because phosphonio ligands constitute particularly strong-
ly electron-withdrawing substituents by virtue of ylidic
charge stabilization, we also prepared the phosphonio-sub-
stituted derivative of 9b by treating its lithiated form with
PPh3ACHTNUGRTNEUNG
(OTf)2 (prepared in situ from PPh3O and Tf2O).[28] To
the best of our knowledge, no C2-phosphonio-substituted
imidazolium compounds have been previously reported.[29]
The overall yield of 26 was 22%. Interestingly, the 13C NMR
peak of the C4/C5 carbon atom of the imidazolium ring
occurs at d=143 ppm, about 8 ppm downfield shifted com-
pared to the parent compound 9b. No such shift was ob-
served for all previous derivatives of 9b. This effect might
be explained by a higher weight of the self-umpoled struc-
ture of type II (see Scheme 2). Although no suitable single
crystals for X-ray analysis could be obtained, DFT calcula-
tions[18] on a model system (see Figure 4) point towards (at
least partial) self-umpolung.
Structural data of the diamino-imidazolium core show no
indications of self-umpolung. Apparently, the inductive elec-
tron withdrawal of the tetrafluorobenzene part is too weak
Hence, the amino groups are now almost planar (sum of
angles: 3548) and are orientated at a 458 angle relative to
ꢀ
the imidazolium ring. In addition, the elongated C4 C5
13082
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13078 – 13086

Access to N-Heterocyclic Carbenes with Self-Umpolung Option
FULL PAPER
Figure 4. Model system for phosphonio derivative 26, as obtained by
DFT calculations; a=1.367, b=1.406, c=1.391, and d=1.366 ꢂ.
ꢀ
bond length (1.406 ꢂ) and the shortened C NMe2 bond
Figure 5. Crystal structure of compound 25a (hydrogen atoms are omit-
length (1.367 ꢂ) relative to the parent compound (compare
Figure 2 and the Supporting Information) are visible proof
of a partial self-umpolung according to Figure 1 (B’). An
NBO analysis of the model system shown in Figure 4 further
confirms this by the finding that the single most important
interaction of the NMe2 lone pairs with the imidazolium
ring is found to be an n!p* donation. All these data indi-
cate that the electronic structure of 9b is at least partially of
the self-umpoled variety.
ted for clarity).[36]
structural data of the imidazolium core are similar to that of
the parent compound.
Further crystals also formed in the mother liquor from
which 25a was precipitated. Here, the counteranions consist
of a 1:1 mixture of iodide and tri-iodide (see Figure 6). The
latter was apparently formed with a slight excess of iodine
during the preparation of 25a. In the crystal structure, the
iodide is coordinated to two cations through nearly-orthogo-
In order to be able to introduce nucleophiles at the car-
bene center, halogenated derivatives of 9a–c were also pre-
pared. Thus, the lithiated species were treated with one
equivalent of elemental bromine or iodine to give com-
pounds 24 (R=Tol) and 25a–c (R=Ph, Tol, 4-NMe2-C6H4).
The imidazolium bromo-bromide[30] 24 was isolated in 76%
yield and characterized by single-crystal X-ray analysis (see
the Supporting Information). To the best of our knowledge,
this represents the first structural data for such a system (al-
though C2-bromo-substituted imidazolium derivatives with
other anions have already been characterized).[31] The bro-
mide counteranion forms a halogen bond with the cation
ꢀ
nal (988) halogen bonds (with I I distances of 3.303 and
3.306 ꢂ). The tri-iodide, in contrast, is hydrogen bonded to
a dimethylamino group. Thus, even when already coordinat-
ed to an imidazolium cation, the iodide is still a stronger
Lewis base than the tri-iodide anion. This, in turn, means
that molecular iodine is a stronger halogen bond donor (i.e.,
s*-Lewis acid) than the cationic imidazolium system. The
overall structural motive of 25a, especially the perpendicu-
ꢀ
ꢀ
lar orientation of the C2 I2 and I1 C1 vectors at I3 (see
ꢀ [34]
Figure 6), is reminiscent of pentaiodide I5 .
This spatial
ꢀ
(featuring a Br Br distance of 3.283 ꢂ) and also sits atop
arrangement of two iodocarbenium ions relative to a halide
center is unprecedented, to the best of our knowledge.
the imidazolium ring of a second cation (n!p* interac-
tion).[32] The dimers thus formed might explain the unusual
solvation behavior of compound 24, which is only well solu-
ble in CH2Cl2.
The iodo-iodides 25a–c were obtained in good yields
(25a: 52, 25b: 57, 25c: 59%). Although 25b is insoluble in
all tested solvents, 25a and 25c are reasonable soluble in
CH2Cl2 and CHCl3 (similar to 24). Compound 25a could
also be characterized by single-crystal X-ray crystallography
(see Figure 5).
Similar to compound 24, a strong halogen bond between
ꢀ
the anion and the cation is found, with an I I distance of
3.357 ꢂ. This is in accord with structural data of already re-
ported imidazolium iodo-iodides.[33] The fact that the halo-
gen bond length is distinctly longer than for a C4,C5-unsub-
stituted iodo-iodide[33a] (3.26 ꢂ) and more like that of a
C4,C5-methylated system[33b] (3.35 ꢂ) might be attributable
to the anomeric donation of the NMe2 lone pairs, which
Figure 6. Crystal structure of the cation of 25a with a 1:1 mixture of
iodide and tri-iodide counteranions (most hydrogen atoms are omitted
for clarity).[36]
ꢀ
would render the C I s*orbital of the cation less susceptible
to halogen bond formation with the anion. The further
Chem. Eur. J. 2011, 17, 13078 – 13086
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13083

S. M. Huber, R. Weiss et al.
Derivatization of compound 24 with nucleophiles: Bromo-
bromide 24 proved to be a suitable starting material for the
introduction of nucleophiles at the C2 position of the imida-
zolium backbone. In order to introduce prototypical p-
donor and p-acceptor substituents, an amine (pyrrolidine)
and cyanide were chosen as nucleophiles, respectively
(Scheme 10).
the mechanism of the ring cleavage, we assume that proto-
nation of the pyrrolidino nitrogen leads to a further intensi-
fication of the anomeric donation of the NMe2 lone pairs
(because the cationic charge on the pyrrolidino substituent
should lower the acceptor orbitals on the imidazolium ring).
ꢀ
ꢀ
This apparently leads to cleavage of the C4 N3 (or C5 N1)
bond of the ring system, through formation of a ketene imi-
nium intermediate. Thus, the first step of the hydrolysis cor-
responds to the reversal of the formation of the imidazolium
ring system (see Scheme 4) and further illustrates the impor-
tance of the anomeric interaction for the chemistry of these
systems.
Finally, we introduced a cyano substituent at the former
carbene position by addition of NEt4CN to a solution of 24
in CH3CN. The product 28 was isolated in 69% yield. Struc-
tural data obtained by a single-crystal X-ray analysis (see
the Supporting Information) shows no apparent signs of
self-umpolung of the imidazolium system and thus indicates
that the cyano substituent is not a very strong p acceptor, in
accord with literature reports.[35] To the best of our knowl-
edge, these data also represent the first structural characteri-
zation of a C2-cyano-substituted imidazolium system.
Scheme 10. Nucleophilic substitution reactions on bromo-bromide 24. i) 2
pyrrolidine; ii) NEt4CN.
Thus, treatment of 24 with two equivalents of pyrrolidine
(with one of them acting as a base) led to the isolation of
the trisACHTUNGTRENNUNG(amino)-substituted imidazolium compound 27 in
62% yield. Orientating PM3 calculations showed that the
newly-introduced amino group is orientated perpendicular
to the imidazolium ring for steric reasons. Cyclic voltamme-
try measurements resulted in a reversible oxidation peak at
0.64 V (vs. SCE), about 0.31 V lower than that for the
parent compound. This shift can easily be explained by the
electron donation of the amino group. When we tried to
grow crystals of 27 by diffusion of ether into a solution of 27
in CH2Cl2, crystal material was indeed obtained. The crys-
tals, however, referred to a guanidine derivative 27*, as
shown in Figure 7 (as every second molecule is protonated
by HBr, the overall crystal structure represents a 1:1 co-
crystal between protonated and unprotonated amidine).
Because compound 27 had been unambiguously charac-
terized before, apparently a ring-cleavage reaction had
taken place. We were able to reproduce this result by treat-
ing a solution of 27 in CH2Cl2 with concentrated HCl (and
therefore we assume that the charge of CH2Cl2 that was
used for crystallization was contaminated with acid). As for
Summary on self-umpolung investigations: For all (acceptor-
substituted) derivatives of 9a and 9b that could be charac-
terized by single-crystal X-ray analysis, no indications of
self-umpolung of the former carbene center could be found.
However, DFT calculations indicate that the C2-phospho-
nio-substituted derivative 26 might show at least partial self-
umpolung (see above). In order to determine, which kind of
acceptor substituents at the C2 position would be needed to
induce self-umpolung of the imidazolium ring, further DFT
investigations,[18] coupled with NBO analyses, were under-
taken. Table 1 shows the internal stabilization energies that
Table 1. Internal stabilization energies [kcalmolꢀ1] for both types of lone
pair donation (see Scheme 11) for various substituted systems.
ꢀ
+
+
+
ꢀ
18.2
3.6
ꢀ
21.0
5.2
ꢀ
20.4
7.7
ꢀ
18.9
15.8
ꢀ
6.1
ꢀ
R
CH2
NH2
H
PMe3
N2
CH2
[a]
n!s*
n!p*
–
–
[a]
84.7
[a] According to the NBO analysis, the NMe2 lone pairs form double
bonds with the imidazolium core.
arise from n!s* versus n!p* donation of the NMe2 lone
pairs into the respective acceptor orbitals of the imidazoli-
um ring (as computed by second-order perturbation theory
analysis within the NBO routine) for various substituents at
the C2 position. Again, a simplified model system was used
(see Scheme 11).
Figure 7. Crystal structure of 27* after apparent hydrolysis of 27 (hydro-
gen atoms and the bromide counteranion are omitted for clarity).
Scheme 11. n!s* versus n!p* donation of the NMe2 lone pairs (com-
pare Table 1).
13084
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13078 – 13086

Access to N-Heterocyclic Carbenes with Self-Umpolung Option
FULL PAPER
As can be seen from Table 1, strong (cationic) acceptor
substituents like the diazonium group are necessary to
induce (almost) full self-umpolung and “switch on” the
amino groups for p donation into the imidazolium ring. Un-
metal complexes. Thus, manipulation of the spin states and
electron richness of these systems by redox processes should
be feasible. These considerations, however, lead well beyond
the scope of our present investigations.
+
+
ꢀ
ꢀ
fortunately, all efforts to introduce either N2 or CH2
have not been successful. For other substituents, especially
donors, the NMe2 lone pairs are “locked” in the anomeric
interaction with the s* orbitals of the ring.
Experimental Section
Exemplary synthesis of 9a: A solution of N,N’-diphenylformamidine
(1.061 g, 5.40 mmol) in THF (25 mL) was cooled to ꢀ788C, treated with
a solution of n-butyllithium in hexane (3.4 mL, 1.6m) and slowly warmed
to room temperature. The resulting clear, yellowish solution was added
to a suspension of compound 7 (5.052 g, 5.40 mmol) in THF (50 mL) and
the mixture was stirred for 20 h at room temperature. After heating to
reflux for 5 h a clear red solution was obtained. The solvent was evapo-
rated and the residue was twice recrystallized from CH2Cl2 with an about
5-fold amount of Et2O, filtered off, washed with Et2O, and dried in high
vacuo (yield 59%).
It is also evident from the calculations that steric hin-
drance of the NMe2 groups among each other prevents a
completely planar orientation of the amino groups in the
case of the self-umpoled structures. This issue might be re-
solved by connecting two of the methyl groups of adjacent
amino functions and thus forming a second ring system.
Transition metals represent another class of potentially
strong p-acceptor substituents. As the synthesis of such com-
plexes did not represent an immediate goal of our investiga-
tions, we only undertook some preliminary test reactions.
Thus, although we were able to confirm the formation of a
transition metal complex in the reaction of lithiated 9b with
PdCl2 and two equivalents of PBu3 (and in the reaction of
16 with 0.5 equivalents of Ag2O) by means of mass spec-
trometry, no clean product could be isolated in both cases.
For spectroscopic data and further experimental procedures see the Sup-
porting Information.
Acknowledgements
S.M.H. thanks the Studienstiftung des Deutschen Volkes for a doctoral
scholarship.
Conclusion
[1] a) R. Weiss, C. Priesner, Angew. Chem. 1978, 90, 491–492; Angew.
Chem. Int. Ed. Engl. 1978, 17, 457–458; b) R. Weiss, C. Priesner, H.
Wolf, Angew. Chem. 1979, 91, 505–506; Angew. Chem. Int. Ed.
Engl. 1979, 18, 472–473; c) R. Weiss, C. Priesner, H. Wolf, Angew.
Chem. 1978, 90, 486–487; Angew. Chem. Int. Ed. Engl. 1978, 17,
446–447; d) R. Weiss, M. Hertel, H. Wolf, Angew. Chem. 1979, 91,
506–507; Angew. Chem. Int. Ed. Engl. 1979, 18, 473–474.
[2] a) R. H. Lowack, R. Weiss, J. Am. Chem. Soc. 1990, 112, 333–338;
b) R. Weiss, R. H. Lowack, Angew. Chem. 1991, 103, 1183–1184;
Angew. Chem. Int. Ed. Engl. 1991, 30, 1162–1163; c) R. Weiss, R. H.
Lowack, Bull. Soc. Chim. Belg. 1991, 100, 483.
[3] a) R. Weiss, S. Reichel, M. Handke, F. Hampel, Angew. Chem. 1998,
110, 352–354; Angew. Chem. Int. Ed. 1998, 37, 344–347; b) R.
Weiss, S. Reichel, Eur. J. Inorg. Chem. 2000, 1935–1939; c) M. Hol-
zinger, O. Vostrowski, A. Hirsch, F. Hennrich, M. Kappes, R. Weiss,
F. Jellen, Angew. Chem. 2001, 113, 4132–4136; Angew. Chem. Int.
Ed. 2001, 40, 4002–4005.
[4] R. Weiss, N. Kraut, Angew. Chem. 2002, 114, 327–329; Angew.
Chem. Int. Ed. 2002, 41, 311–314.
[5] R. Weiss, N. Kraut, F. Hampel, J. Organomet. Chem. 2001, 617–618,
473–482.
[6] K. Deuchert, S. Hꢃnig, Angew. Chem. 1978, 90, 927–938; Angew.
Chem. Int. Ed. Engl. 1978, 17, 875–886.
[7] For selected recent reviews, see a) F. E. Hahn, M. C. Jahnke, Angew.
Chem. 2008, 120, 3166–3216; Angew. Chem. Int. Ed. 2008, 47, 3122–
3172; b) K. Oefele, E. Tosh, C. Taubmann, W. A. Herrmann, Chem.
Rev. 2009, 109, 3408–3444; c) M. Albrecht, Science 2009, 326, 532–
533; d) L. Mercs, M. Albrecht, Chem. Soc. Rev. 2010, 39, 1903–
1912.
[8] a) H. E. Winberg, J. E. Carnahan, D. C. Coffman, M. Brown, J. Am.
Chem. Soc. 1965, 87, 2055–2056; b) N. Wiberg, Angew. Chem. 1968,
80, 809–822; Angew. Chem. Int. Ed. Engl. 1968, 7, 766–779.
[9] a) R. Tonner, F. ꢄxler, B. Neumꢃller, W. Petz, G. Frenking, Angew.
Chem. 2006, 118, 8206; Angew. Chem. Int. Ed. 2006, 45, 8038; b) O.
Kaufhold, F. E. Hahn, Angew. Chem. 2008, 120, 4122; Angew.
Chem. Int. Ed. 2008, 47, 4057.
The first synthesis of 4,5-bis(dimethylamino)-substituted
imidazolium compounds also constitutes the first application
of the dication 8 for the preparation of heterocycles. As a
potent a1,a2-biselectrophile, the ethene derivative 8 should
also be applicable for the synthesis of further electron-rich
precursors of heterocyclic carbenes. Beyond these applica-
tions, compounds 8 possess a fascinating potential as syn-
thetic equivalent of C2O2.
The novel NHC precursors presented in this paper show a
remarkably reduced basicity and nucleophilicity of their
NMe2 groups, which is due to the strong anomeric interac-
tions of the latter with the imidazolium core. On the one
hand, this electron donation creates a predetermined break-
ꢀ
ing point of the ring system in form of the acceptor C N
bonds, whereas on the other hand the ring strain of the imi-
dazolium core is somewhat released by elongation of the
said bonds.
According to DFT calculations, these NHCs should, in
principle, be capable of self-umpolung. However, our inves-
tigations have shown that typical acceptor substituents at
the C2 position are not capable of inducing this effect. For
this electronic re-organization, very strong (cationic) accept-
or substituents are necessary, and consequently partial self-
umpolung was found in the case of the phosphonio-substi-
tuted derivative.
Imidazolium compound 9c represents a special case, as it
incorporates as much as three two-step redox systems within
the NHC framework. This will probably result in a high
electronic flexibility of the corresponding nucleophilic car-
benes, especially when they serve as ligands in transition
[10] See the discussion in reference [3a].
Chem. Eur. J. 2011, 17, 13078 – 13086
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13085

S. M. Huber, R. Weiss et al.
[11] a) P. M. Kochergin, A. M. Tsyganova, V. S. Shilikhunova, M. A.
Klykov, Chem. Heterocyl. Compds. 1971, 7, 648; b) P. M. Kochergin,
A. M. Tsyganova, V. S. Shilikhunova, M. A. Klykov, Khim. Geterot-
sikl. Soedin. 1971, 5, 689; c) S. Kulkarni, M. R. Grimmett, L. R.
Hanton, J. Simpson, Aust. J. Chem. 1987, 40, 1399; d) R. Stradi, G.
Verga, Synthesis 1977, 688; e) L. Citerio, E. Rivera, M. L. Saccarello,
R. Stradi, B. Givia, J. Heterocycl. Chem. 1980, 17, 97.
Chataigner, A. Harrison-Marchand, J. Maddalumo, Sci. Synth. 2005,
26, 1123.
[28] J. B. Hendrickson, S. M. Schwartzmann, Tetrahedron Lett. 1975, 16,
277.
[29] The isolated product showed one set of NMR signals, but was
shown by elemental analysis to contain two equivalents of the side
product LiOTf, which could not be separated off despite various at-
tempts.
[12] Also,
a 4,5-bis(methylphenylamino)-substituted imidazolone (and
the corresponding imidazolthione) have been reported: a) M.
Wenzel, R. Beckert, W. Gꢃnther, H. Gçrls, Eur. J. Org. Chem. 1998,
183; b) M. Wenzel, D. Lindauer, R. Beckert, R. Boese, E. Anders,
Chem. Ber. 1996, 129, 39.
[30] C2-Bromo-substituted imidazolium derivatives are known: a) N.
Kuhn, A. Abu-Rayyan, A. Al-Sheikh, K. Eichele, C. Maichle-
Meossmer, S. M., K. Sweidan, Z. Naturforsch. B 2005, 60, 294; b) N.
Kuhn, A. Abu-Rayyan, K. Eichele, S. Schwarz, M. Steimann, Inorg.
Chim. Acta 2004, 357, 1799.
[31] a) N. Kuhn, A. Abu-Rayyan, K. Eichele, C. Piludu, M. Steimann, Z.
Anorg. Allg. Chem. 2004, 630, 495; b) N. Kuhn, A. Al-Sheikh, S.
Schwarz, M. Steimann, Z. Naturforsch. B 2004, 59, 129; c) N. Kuhn,
M. Richter, M. Steimann, M. Strobele, K. Sweidan, Z. Anorg. Allg.
Chem. 2004, 630, 2054.
[13] R. Weiss, S. M. Huber, F. W. Heinemann, P. Audebert, F. G. Pꢃhlhof-
er, Angew. Chem. 2006, 118, 8228–8231; Angew. Chem. Int. Ed.
2006, 45, 8059–8062.
[14] L. M. Nxumalo, E. K. Ngidi, T. A. Ford, J. Mol. Struct. 2006, 786,
168–174, and references therein.
[15] The numbering of atoms used in the text refers to the usual nomen-
clature (with the carbene center being the C2 position) and not to
the labels shown in Figure 2.
[32] See also: M. Rechinger, Dissertation, University of Erlangen–Nur-
emberg (Germany), 1995.
[16] A. J. Arduengo III, F. Davidson, H. V. R. Dias, J. R. Goerlich, D.
Khasnis, W. J. Marshall, T. K. Prakasha, J. Am. Chem. Soc. 1997,
119, 12742.
[17] A. J. Arduengo III, S. F. Gamper, M. Tamm, J. C. Calabrese, F. Da-
vidson, H. A. Craig, J. Am. Chem. Soc. 1995, 117, 572.
[18] Gaussian 03W (see Ref. [37]), B3LYP 6–31G*.
[33] A. J. Arduengo III, M. Tamm, J. C. Calabrese, J. Am. Chem. Soc.
1994, 116, 3625.
[34] See, for example: a) K. Neupert-Laves, M. Dobler, Helv. Chim. Acta
1975, 58, 432; b) C. L. Raston, A. H. White, D. Petridis, D. Taylor, J.
Chem. Soc. Dalton Trans. 1980, 1928; c) W. Zhang, S. R. Wilson,
D. N. Hendrickson, Inorg. Chem. 1989, 28, 4160.
[19] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899–
926.
[35] a) E. Riedel, Modern Inorganic Chemistry, Wiley, New York, 2005;
b) W. F. Reynolds, P. Dais, D. W. MacIntyre, R. D. Topsom, S. Mar-
riott, E. von Nagy-Felsobuki, R. W. Taft, J. Am. Chem. Soc. 1983,
105, 378–384; c) K. Takeuchi, Y. Senzaki, K. Okamoto, J. Chem.
Soc. Chem. Commun. 1984, 111–112.
[20] This structural variation—probably difficult to achieve according to
the present protocol—remains an intriguing challenge, as it would
possess much better requisites for self-umpolung than compounds 9.
[21] A. J. Arduengo III, H. V. Rasika Dias, R. L. Harlow, M. Kline, J.
Am. Chem. Soc. 1992, 114, 5530.
[36] Ortep-3 for Windows: L. J. Farrugia, J. Appl. Crystallogr. 1997, 30,
565.
[22] a) A. R. Chianese, B. M. Zeglis, R. H. Crabtree, Chem. Commun.
2004, 2176; b) N. Kuhn, M. Steimann, M. Walker, Z. Naturforsch. B
2001, 56, 129; c) D. S. McGuinness, K. J. Cavell, Organometallics
2000, 19, 4918.
[23] P. P. O. Santos, C. L. Trindade, A. M. Lobo, S. Prabhakar, Tetrahe-
dron Lett. 1993, 34, 3793.
[24] a) S. Gunasekaran, S. Seshadri, S. Muthu, S. Kumaresan, R. Arunba-
laji, Spectrochim. Acta Part A Mol. Spectrosc. 2008, 70 A, 550–556;
b) N. A. Owston, A. J. Parker, J. M. J. Williams, Org. Lett. 2007, 9,
73–75.
[25] N. Kuhn, M. Steimann, M. Walker, Z. Naturforsch. B 2001, 56, 129.
[26] a) R. Weiss, N. J. Salomon, G. E. Miess, Angew. Chem. 1986, 98, 925;
Angew. Chem. Int. Ed. 1986, 25, 917; b) R. Weiss, F. G. Pꢃhlhofer, N.
Jux, K. Merz, Angew. Chem. 2002, 114, 3969; Angew. Chem. Int. Ed.
2002, 41, 3815.
[37] Gaussian 03, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, 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, V. Bakken, 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. Zakrzewski, 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. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[27] a) N. Sawa, M. Yasuda, Jpn. Kokai Tokkyo Koko 1968, 11; b) P.
Schneiders, J. Heinze, H. Baumgaertel, Synthesis 1970, 18; c) P.
Schneiders, J. Heinze, H. Baumgaertel, Ber. Bunsen-Ges. 1972, 76,
556; d) H. Frischkorn, E. Schinzel, Liebigs Ann. Chem. 1984, 1129;
e) Z. Shi, H. Gu, L.-L. Xu, Synth. Commun. 1996, 26, 3175; f) I.
Received: June 28, 2011
Published online: October 5, 2011
13086
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13078 – 13086