From a stable dianion to a stable carbenoid

Article abstract of DOI:10.1002/anie.200701588

(Chemical Equation Presented) Robust, but reactive: The first example of a Li/Cl carbenoid that is stable at room temperature has been prepared by mild oxidation of a geminal dianion (see scheme). The electronic reasons behind this unusual thermal stabili

Full text of DOI:10.1002/anie.200701588

Angewandte
Chemie
DOI: 10.1002/anie.200701588
Carbenoid Compounds
From a Stable Dianion to a Stable Carbenoid**
Thibault Cantat, Xavier Jacques, Louis Ricard, Xavier FrØdØric Le Goff, Nicolas MØzailles, and
Pascal Le Floch*
Organometallic compounds have demonstrated major utility
(Scheme 1, route B)[9] and validate this hypothesis by isolat-
ing the first Li/Hal carbenoid that is stable at room temper-
ature.
The oxidation of the geminal dianion 1 was undertaken by
using hexachloroethane in diethyl ether (Scheme 2). The
ꢀ
ꢀ
in organic synthesis for the formation of C C and C
heteroelement bonds.[1] Among these, a-halogenated deriva-
tives (typically R2CLiX) have attracted much attention for
their unusual electrophilic character. These molecules have
been named “carbenoids” since they can formally release the
carbene fragment when trapped with an electron-rich center
(alkenes, carbanions, etc.).[2,3] In the absence of any trapping
reagent, Li/halogen (Li/Hal) carbenoids decompose at low
temperature by a-elimination of LiX.[2,4] For instance, the
carbenoid LiCHCl2 decomposes around ꢀ1008C.[5] Never-
theless, extensive work has been devoted to the synthesis and
characterization of carbenoid derivatives. The general syn-
thetic method relies on the a-metalation of halogenoalkanes
(Scheme 1, route A). Owing to their instability, these mole-
Scheme 2.
reaction was monitored by 31P NMRspectroscopy and
showed the complete formation of a sole compound (2),
characterized by a singlet at d = 45.5 ppm (vs. d = 20.6 ppm
for 1), within a few minutes.[10] Compound 2 was isolated in
quantitative yield after removal of LiCl by centrifugation and
Scheme 1.
1
fully characterized by NMRspectroscopy ( H and 13C). This
cules have mostly been studied by NMRspectroscopy and
theoretical calculations.[2] To date, only a few X-ray structures
have been recorded: Boche and co-workers succeeded in
determining the X-ray structure of the unstable LiCHCl2·
(pyridine)3 carbenoid,[6] and so far the most stable carbenoid
was characterized by Niecke and co-workers as an ylene-
(phosphoranylidene)carbenoid that decomposes at ꢀ108C.[7]
We have recently illustrated the use of thiophosphinoyl
groups in the stabilization of highly reduced carbon centers
and isolated stable geminal dianion 1.[8] We postulated that a
similar strategy could also be applied to stabilize a carbenoid.
Herein we report on a new synthetic path to Li/Hal
carbenoids by mild oxidation of geminal dianions
reaction validates the relevance of route B for the synthesis of
carbenoids. Surprisingly, 2 proved to be stable in the solid
state and in solution up to 608C. Above this temperature,
decomposition of 2 to a mixture of intractable products was
observed. The electronic reasons behind this exceptional
stability were sought after. Carbenoids are usually highly
unstable species, and their electronic structure has been
investigated mainly by low-temperature NMRspectroscopy,
which was corroborated by theoretical calculations.[2a,11] The
most striking point for these species concerns the 13C NMR
chemical shift of the Li/Hal carbenoid (R2CLiX) center,
which is deshielded compared to that of the hydrogenated
species (R2CHX). This deshielding has been attributed to a
weakening of the carbon–halogen bond upon metalation
(paramagnetic contribution of the s*Cꢀ orbital to the
X
[*] T. Cantat, X. Jacques, Dr. L. Ricard, Dr. X. F. Le Goff, Dr. N. MØzailles,
Prof. P. Le Floch
13C NMRchemical shift). For example, a deshielding of
50.0 ppm has been measured by Seebach et al. for LiCHCl2 in
THF.[12] Further insights into the electronic structure of 2 were
provided by its 13C NMRspectrum. In 2, the central carbon
atom resonates at d = 38.5 ppm as a triplet (1JCP = 80.0 Hz).
For the purpose of comparison, the hydrogenated species (3)
was synthesized by trapping 2 with water. Compound 3 was
fully characterized by 1H, 13C, and 31P NMRspectroscopy and
X-ray diffraction analysis.[13] A shielding of d = 15.2 ppm was
clearly measured for the central 13C atom on going from 3
(d = 53.7 ppm) to 2. This observation reveals that in carbenoid
Laboratoire “HØtØroØlØments et Coordination”
UMRCNRS 7653
Ecole Polytechnique
91128 Palaiseau CØdex (France)
Fax: (+33)169-333-990
E-mail: lefloch@poly.polytechnique.fr
[**] The authors thank the CNRS and the Ecole Polytechnique for the
financial support and IDRIS for the allowance of computer time
(project no. 71616).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 5947 –5950
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5947

Communications
ꢀ
2 the paramagnetic contribution is lowered and the C Cl
bond remains strong.
agreement with previous results obtained for dianion 1, which
showed that in this system a pure p lone pair is more
efficiently stabilized by negative hyperconjugation than a
sp2-hybridized lone pair. As a consequence, the carbon
This feature was confirmed by a study of the metric
parameters of 2 (Figure 1).[13] Indeed the C Cl bond length in
ꢀ
ꢀ
s character in the C Cl bond remains high, and the carbon
atom directs a hybrid orbital with 21.9% s character towards
the chloride atom (18.6% in 3). For comparison, in the
carbenoid LiCH2Cl·(OMe2)3 most of the 2s character is
ꢀ
concentrated in the C Li bond (34.3%) so that only 11.0%
ꢀ
is left for the C Cl bond (vs. 19.0% in CH3Cl), thus
ꢀ
accounting for the weakness of the C Cl bond. Accordingly,
the C Cl Wiberg bond index remains constant between 3
(0.88) and 2 (0.84), and it significantly increases for the C P
bonds (from 0.81 in 3 to 1.09 in 2). These findings are in
ꢀ
ꢀ
agreement with the experimental data (shielding of the
ꢀ
carbon atom and short C Cl bond). To conclude, the stability
of carbenoid 2 comes from 1) the coordination of the lithium
cation by the thiophosphinoyl arms and 2) the stabilization of
the carbon lone pair in a pure p orbital, which leads to a rather
Figure 1. ORTEP plot of 2 (ellipsoids set as 50% probability). Selected
bond lengths [Š] and angle [8]: P1-C1 1.722(2), P2-C1 1.730(2), C1-Cl1
1.781(2), P1-S1 1.987(1), P2-S2 1.995(1); P1-C1-P2 127.7(1).
ꢀ
strong C Cl bond.
To determine whether carbenoid 2 may be used as a
ꢀ
carbene precursor (i.e., whether the C Cl bond may be
2 (1.781(2) Š) is as short as in 3 (1.796(4) Š).[14] The carbon
atom is planar (ꢀangles = 359.98) and strongly stabilized by
cleaved), we investigated its reactivity towards several
transition-metal complexes. Results from our studies with
palladium complexes are presented here. Compound 2 reacts
slowly with [Pd0(PPh3)4] in THF at room temperature to form
complex 4, LiCl, and triphenylphosphine, thus showing that
indeed carbenoid 2 is a convenient source of a carbene ligand
(Scheme 2). Interestingly, complex 4 can also be prepared by
using dianion 1 as the carbene precursor and [PdCl2-
(PPh3)2].[16,17] To the best of our knowledge, this reactivity is
the first example of the use of a Li/Hal carbenoid as a
precursor for the synthesis of a carbene complex. The
mechanism of this unprecedented transformation was inves-
tigated by DFT calculations. Two mechanisms (A and B,
ꢀ
the two phosphorus atoms (P C 1.726 Š (av) in 2 vs. 1.860 Š
(av) in 3). Moreover, the X-ray structure reveals the
stabilization of the lithium cation by the two sulfur atoms
and two molecules of diethyl ether. This coordination most
likely prevents LiCl elimination and therefore accounts for
the increased thermal stability of 2. Such an effect is
consistent with the higher stability of carbenoids observed
in coordinating solvents.[2a]
Theoretical calculations have been performed to ration-
alize these findings in detail. The geometry optimized on the
complete system is in excellent agreement with the exper-
imental structure.[15] Analysis of the carbon hybridization
state indicates that the carbon atom possesses a lone pair in a
pure p orbital, stabilized by negative hyperconjugation into
the phosphorus antibonding orbitals. This observation is in
ꢀ
Figure 2) can be proposed, depending on the nature of the C
Cl bond cleavage. Indeed, the metal center may act as a
nucleophile through displacement of the chloride anion
(nucleophilic substitution, mechanism A) or insert into the
Figure 2.
5948 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5947 –5950

Angewandte
Chemie
ꢀ
analysis (%) calcd for C33H40ClLiO2P2S2: C 62.21, H 6.33; found:
C 62.24, H 6.41.
For syntheses of 3 and 4, see the Supporting Information.
C Cl bond (oxidative addition, mechanism B). The results of
calculations that were carried out on a model system are
presented in Figure 2.[15]
As the coordination of the lithium cation may play an
important role in the reactivity of the carbenoid,[18] both
mechanisms have been computed in the presence and in the
absence of a [Li(OMe2)2]+ counterion. We have found that the
presence of [Li(OMe2)2]+ does not have much influence on
both energetic profiles and, for the sake of simplicity, only the
mechanisms in the absence of the counterion are presented
(see the Supporting Information for the mechanism in the
presence of the cation).[19] Formation of complex III is a
Received: April 11, 2007
Published online: July 2, 2007
Keywords: carbenes · carbenoids ·
.
density functional calculations · lithium · phosphorus
[1] a) Lithium Chemistry: ATheoretical and Experimental Overview
(Eds.: A.-M. Sapse, P. von R. Schleyer), Wiley, Chichester, 1995;
b) The Chemistry of Organolithium Compounds (Eds.: Z.
Rappoport, I. Marek), Wiley, Chichester, 2004; c) B. J. Wake-
field, The Chemistry of Organolithium Compounds, Pergamon,
New York, 1974; d) R. A. Gossage, J. Jastrzebski, G. van Koten,
Angew. Chem. 2005, 117, 1472 – 1478; Angew. Chem. Int. Ed.
2005, 44, 1448 – 1454.
slightly exergonic process (DGPCM,I–III = ꢀ3.7 kcalmolꢀ1
;
PCM = polarizable continuum model). Pathway A proceeds
in two steps, namely coordination of the thiophosphinoyl
arms of I to the palladium(0) fragment yielding complex II
and a subsequent SN2 step yielding III (via TSII–III). This
pathway requires an overall energy of 24.9 kcalmolꢀ1. The
alternative mechanism B relies on an oxidative addition step:
[2] a) G. Boche, J. C. W. Lohrenz, Chem. Rev. 2001, 101, 697 – 756,
and references therein; b) G. Kobrich, Angew. Chem. 1967, 79,
15 – 27; Angew. Chem. Int. Ed. Engl. 1967, 6, 41 – 52.
The active [Pd0(PH3)2] species[20] inserts into the C Cl bond to
ꢀ
[3] a) G. L. Closs, R. A. Moss, J. Am. Chem. Soc. 1964, 86, 4042;
b) S. E. Denmark, J. P. Edwards, S. R. Wilson, J. Am. Chem. Soc.
1991, 113, 723 – 725; c) H. E. Simmons, R. D. Smith, J. Am.
Chem. Soc. 1958, 80, 5323 – 5324; d) J. F. Fournier, S. Mathieu,
A. B. Charette, J. Am. Chem. Soc. 2005, 127, 13140 – 13141.
[4] Replacing the lithium center by another metal center (Mg, Zn,
Al, etc.) increases the stability of the carbenoid; this method-
ology has been successfully developed, for example, in the
Simmons–Smith reaction (see references [3b–d]).
[5] D. S. Matteson, D. Majumdar, Organometallics 1983, 2, 1529 –
1535.
[6] A. Muller, M. Marsch, K. Harms, J. C. W. Lohrenz, G. Boche,
Angew. Chem. 1996, 108, 1639 – 1640; Angew. Chem. Int. Ed.
Engl. 1996, 35, 1518 – 1520.
[7] E. Niecke, P. Becker, M. Nieger, D. Stalke, W. W. Schoeller,
Angew. Chem. 1995, 107, 2012 – 2015; Angew. Chem. Int. Ed.
Engl. 1995, 34, 1849 – 1852; M. Yoshifuji, S. Ito, Top. Curr. Chem.
2003, 223, 67 – 89.
[8] T. Cantat, L. Ricard, P. Le Floch, N. MØzailles, Organometallics
2006, 25, 4965 – 4976.
[9] Some carbenoids have been obtained by oxidation of 1,1-dizinc
and 1,1-dimagnesium species; see: I. Marek, Chem. Rev. 2000,
100, 2887; F. Chemla, I. Marek, J. F. Normant, Synlett 1993, 665;
I. Creton, I. Marek, J. F. Normant, Synthesis 1996, 1499.
[10] The bis(phosphonate) equivalent of 2 has already been observed
in solution by 31P NMRspectroscopy: B. Iorga, P. Savignac, J.
Organomet. Chem. 2001, 624, 203 – 207.
form complex IV (via TSI–IV). Complex III is then formed
after elimination of the chloride ion, and PH3 is formed upon
coordination of the thiophosphinoyl arms (via V). The
oxidative addition is the rate-determining step of this pathway
and requires an activation energy of 34.8 kcalmolꢀ1. There-
fore, comparison of the two energetic profiles clearly shows
that formation of the carbene complex results from an SN2
attack of the palladium(0) center, which displaces the chloride
anion (LiCl in the presence of lithium ions). This mechanism
definitely demonstrates the electrophilic nature of 2, a
characteristic property of carbenoids. According to this
mechanism, further developments of this reactivity to form
carbene complexes should focus on electron-rich metal
centers to promote the SN2 step.
In conclusion, we have developed a new synthetic
approach towards carbenoids on the basis of the oxidation
of a geminal dianion. The use of electron-withdrawing
substituents proved to be efficient for the stabilization of
the carbenoid center and therefore allowed us to synthesize
the first Li/Hal carbenoid that is stable at room temperature.
Reactivity investigations showed that Li/Hal carbenoids can
be used to form carbene complexes with electron-rich metal
centers.
[11] a) M. Buehl, N. Hommes, P. von R. Schleyer, U. Fleischer, W.
Kutzelnigg, J. Am. Chem. Soc. 1991, 113, 2459 – 2465; b) V.
Schulze, R. Lowe, S. Fau, R. W. Hoffmann, J. Chem. Soc. Perkin
Trans. 2 1998, 463 – 465; c) T. Koizumi, O. Kikuchi, J. Mol. Struct.
(Theochem) 1995, 336, 39 – 46; d) D. Seebach, H. Siegel, J.
Gabriel, R. Hassig, Helv. Chim. Acta 1980, 63, 2046 – 2053.
[12] D. Seebach, R. Hassig, J. Gabriel, Helv. Chim. Acta 1983, 66,
308 – 337.
Experimental Section
All experiments were carried out in a dry argon or nitrogen
atmosphere using distilled and degassed solvents.
2: Hexachloroethane (94.8 mg, 0.40 mmol) was added to a
solution of 1 (0.40 mmol) in diethyl ether (3 mL) at ꢀ408C and
stirred for 5 minutes while being allowed to room temperature. LiCl
was removed by centrifugation, and the solvents were then evapo-
rated to afford 2 as a yellow solid in 100% yield (260 mg, 0.40 mmol).
1H NMR(300 MHz, [D 8]THF, 258C): d = 7.11–7.35 (m, 12H; Hmeta
and Hpara), 7.98 ppm (brs, 8H; Hortho); 31P{1H} NMR(121.5 MHz,
[D8]THF, 258C, 85 % H3PO4 as external standard): d = 45.5 ppm (s);
13C{1H} NMR(75.465 MHz, CDCl 3, 258C, [D8]THF (d = 68.6 ppm) as
internal reference): d = 38.5 (t, 1J(C,P) = 80.0 Hz; PCP), 127.7 (t,
3J(C,P) = 6.1 Hz; Cmeta), 130.1 (s; Cpara), 133.6 (t, 2J(C,P) = 5.1 Hz;
Cortho), 139.1 ppm (AXX’, ꢀJ(C,P) = 129.0 Hz; Cipso). Elemental
[13] X-ray structure data : Nonius KappaCCD diffractometer, f and
w scans, MoKa radiation (l = 0.71073 Š), graphite monochroma-
tor, T= 150 K, structure solution with SIR97,[21a] refinement
against F2 with SHELXL-97[21b] with anisotropic thermal param-
eters for all non-hydrogen atoms, calculated hydrogen positions
with riding isotropic thermal parameters. Data collection for 2:
C33H40ClLiO2P2S2, Mr = 637.10, crystal dimensions 0.20 ” 0.20 ”
¯
0.10 mm, triclinic, space group P1, a = 10.539(1), b = 10.672(1),
c = 16.889(1) Š, a = 82.135(1), b = 72.243(1), g = 68.526(1)8, V=
, , F-
1682.8(2) Š3, Z = 2, 1calcd = 1.257 gcmꢀ3 m = 0.361 cmꢀ1
Angew. Chem. Int. Ed. 2007, 46, 5947 –5950
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 5949

Communications
(000) = 672, qmax = 27.488, hkl ranges: ꢀ13 to 13, ꢀ13 to 13, ꢀ15
to 21, 10774 reflections collected, 7683 unique reflections (Rint
Le Floch, F. Nief, N. Mezailles, Organometallics 2006, 25, 1329 –
1332; T. Cantat, F. Jaroschik, F. Nief, L. Ricard, N. Mezailles, P.
Le Floch, Chem. Commun. 2005, 5178 – 5180.
=
0.0156), 6072 reflections with I > 2s(I), 401 parameters refined,
GOF(F2) = 1.042, final R indices (R1 = ꢀ j j Fo j ꢀ j Fc j j /ꢀ j Fo j ,
[18] a) Z. F. Ke, C. Y. Zhao, D. L. Phillips, J. Org. Chem. 2007, 72,
848 – 860; b) C. Y. Zhao, D. Q. Wang, D. L. Phillips, J. Am.
Chem. Soc. 2003, 125, 15200 – 15209.
wR2 = [ꢀw(F2oꢀF2c)2/ꢀw(Fo2)2]1/2
,
R1 = 0.0376, wR2 = 0.1030,
max/min
residual
electron
density
0.778(0.048)/
ꢀ0.340(0.048) eŠꢀ3. For the data collection for 3, see the
Supporting Information. CCDC-643274 and -643275 contain
the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[19] Decoordination of the lithium cation from the carbenoid VI (see
the Supporting Information) in the presence of two molecules of
ether (Me2O) to yield I and [Li(OMe2)4]+ is slightly endergonic
(DGPCM = 4.7 kcalmolꢀ1).
[20] It has been shown that, in solution, [Pd(PPh3)4] dissociates to
form [Pd(PPh3)3] and PPh3. The oxidative addition step proceeds
via the coordinatively unsaturated [Pd(PPh3)2] complex; see: C.
Amatore, A. Jutand, J. Organomet. Chem. 1999, 576, 254 – 278;
A. Jutand, A. Mosleh, Organometallics 1995, 14, 1810 – 1817; C.
Amatore, A. Jutand, F. Khalil, M. A. Mbarki, L. Mottier,
Organometallics 1993, 12, 3168 – 3178.
[21] a) A. Altomare, M. C. Burla, M. Camalli, G. Cascarano, C.
Giacovazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R.
Spagna, SIR97, an integrated package of computer programs for
the solution and refinement of crystal structures using single-
crystal data, 1999; b) M. Sheldrick, SHELXL-97; Universitꢀt
Gꢁttingen, Gꢁttingen, Germany, 1998.
ꢀ
[14] In the carbenoid LiCHCl2·(pyridine)3, the C Cl bond is elon-
gated to 1.84 Š.
[15] See the Supporting Information for further computational
details.
[16] T. Cantat, N. Mezailles, L. Ricard, Y. Jean, P. Le Floch, Angew.
Chem. 2004, 116, 6542 – 6545; Angew. Chem. Int. Ed. 2004, 43,
6382 – 6385.
[17] For other examples of the use of geminal dianion 1 as a carbene
precursor, see: T. Cantat, M. Demange, N. Mezailles, L. Ricard,
Y. Jean, P. Le Floch, Organometallics 2005, 24, 4838 – 4841; T.
Cantat, L. Ricard, N. Mezailles, P. Le Floch, Organometallics
2006, 25, 6030 – 6038; T. Cantat, F. Jaroschik, L. Ricard, P.
5950
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5947 –5950