Synergic synthesis of benzannulated zincabicyclic complexes, α-Zincated N ylides, through sodium-TM EDA-mediated zincation of a haloarene

Article abstract of DOI:10.1002/anie.200904506

Double role for sodium: Within a synergic sodium TM P-zincate reagent sodium mediates the direct ortho-zincation of chlorobenzene and delivers the TMEDA ligandto nucleophilically attack the benzene ring, thereby generating novel open and cyclic zinc zwitt

Full text of DOI:10.1002/anie.200904506

Angewandte
Chemie
DOI: 10.1002/anie.200904506
Zwitterionic Zincation
Synergic Synthesis of Benzannulated Zincabicyclic Complexes,
a-Zincated N Ylides, through Sodium-TMEDA-Mediated Zincation of
a Haloarene**
David R. Armstrong, Liam Balloch, William Clegg, Sophie H. Dale, Pablo Garcꢀa-ꢁlvarez,
Eva Hevia, Lorna M. Hogg, Alan R. Kennedy, Robert E. Mulvey,* and Charles T. OꢂHara
The zinc–hydrogen exchange reaction has undergone a
remarkable transformation in recent times from obscurity to
a novel alternative, and in many cases the preferred option, to
long-established lithiation methods for generating aromatic
organometallic intermediates suitable for subsequent func-
tionalization. Simple zinc reagents (alkyls, amides), as kineti-
cally sluggish bases, are generally useless for such applica-
tions. Installing high metalation power within a zinc reagent
usually requires a more complex composition, in which
reactivity is boosted through cooperative effects between its
different components. Modified from their magnesiating
“turbo-Grignard” reagents, Knochelꢀs three-component sys-
tems [(TMP)2Zn·2MgCl2·2LiCl][1] and [{iPr(tBu)N}2Zn·
2MgCl2·2LiCl][2] are excellent complex zincators for both
aromatic and heteroaromatic substrates (TMP is 2,2,6,6-
tetramethylpiperidide). Formally a two-component lithium
amide–zinc alkyl mixed complex, “[LiZn(TMP)(tBu)2]”
introduced by Kondo and Uchiyama, is also a potent
chemo- and regioselective zincator for similar substrates.[3]
Our own group has contributed the related sodium TMP-
zincate [(TMEDA)·Na(m-TMP)(m-tBu)Zn(tBu)][4] (1), which
depending on the organic substrate can execute regioselective
ortho-, meta-, or dizincation in reactions that have been
structurally defined.[5]
by a novel four-step ortho-zincation, zincate (or sodium
chloride) elimination, azazincation–addition, amine a-zinca-
tion sequence. An intermediate along this path formed prior
to the final step (amine a-zincation), [{1-Zn(tBu)2}À-{2-
N(Me)2CH2CH2NMe2}+-C6H4] (Int), has also been isolated
from the reaction and crystallographically characterized.
In the few previous studies of reactions of aryl halides and
TMP-zincates,[6] no experimental evidence has been collected
on the chemistry taking place between the limits of the
starting materials and the zinc-free quenched products.
Therefore, to shed light on the critical metal (and bimetal)
activity stage of these reactions, our primary objective was to
get inside these limits by isolating and characterizing repre-
sentative zinc-containing intermediates. This was realized
through the synthesis of 2, which were isolated as small
colorless block crystals in 20.7% yield from the equimolar
reaction of 1 and chlorobenzene in hexane solution.
Since it was reported earlier[6] that reaction of [LiZn-
(TMP)(tBu)2] with haloarenes followed by electrophilic
trapping produced polyfunctional haloarenes (i.e., with
retention of the original halogen substituent), we expected 2
to be an ortho-zincated chlorobenzene derivative. Surpris-
ingly, however, chloride is absent and while its aromatic ring
appears to have been ortho-zincated,[7] 2 has several other
unexpected new bonding features. Thus, a nitrogen from
TMEDA that links to the zinc atom through a CH2 bridge and
a CH2CH2N(Me)2 bridge, is now attached to the ring.
Terminal methyl and tert-butyl groups on the aryl-attached
nitrogen and zinc atoms, respectively, complete the formula-
tion, which can be interpreted as a mesoionic zwitterion[8]
with a +NR3-CH2-ZnÀ b-dipole. Alternatively, 2 can be
classified as an a-zincated N ylide.
Here we report the first investigation of the surprising
reactivity of 1 towards an aryl halide, namely chlorobenzene.
Fully characterized by X-ray crystallography and NMR
spectroscopy, the unexpected final product isolated from
this reaction is the remarkable zwitterionic benzannulated
bicyclic zinc complex [{1-Zn(tBu)}À-{2-N(Me)(CH2)CH2-
CH2NMe2}+-C6H4] (2). Formation of 2 can be rationalized
All the anticipated different types of hydrogen atom in 2
are accounted for and well resolved in its 1H NMR spectrum
recorded from C6D6 solution. Belonging to metalated
TMEDA, the Me2N and Me’N resonances appear at d =
1.96 and 2.49 ppm, respectively, while all six H atoms of the
three unique CH2 groups are inequivalent having separate
resonances at d = 0.96/1.94, 1.63/2.44, and 2.22/2.34 ppm,
consistent with the rigid, bicyclic conformation of the
structure. The pairing most upfield (0.96/1.94 ppm) can be
attributed to the deprotonated NCH2 unit of TMEDA. Four
aromatic resonances at d = 8.32, 7.26, 7.07, and 6.67 ppm
denote the H atoms at the 6 (ortho to Zn), 5, 4, and 3 (ortho to
N) positions, and the tBu resonance at d = 1.76 ppm com-
pletes the assignment.
[*] Dr. D. R. Armstrong, L. Balloch, Dr. P. Garcꢀa-ꢁlvarez, Dr. E. Hevia,
Dr. L. M. Hogg, Dr. A. R. Kennedy, Prof. R. E. Mulvey, Dr. C. T. O’Hara
WestCHEM, Department of Pure and Applied Chemistry
University of Strathclyde, Glasgow G1 1XL (UK)
E-mail: r.e.mulvey@strath.ac.uk
Prof. W. Clegg, Dr. S. H. Dale
School of Chemistry, Newcastle University
Newcastle upon Tyne NE1 7RU (UK)
[**] This research was supported by the UK Engineering and Physical
Science Research Council and by a Marie Curie Intra European
Fellowship within the 7th European Community Framework Pro-
gramme (for P.G.A.). We thank Prof. M. G. Davidson (University of
Bath) for helpful discussions. TMEDA=N,N,N’,N’-tetramethyl-
ethylenediamine.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200904506.
Angew. Chem. Int. Ed. 2009, 48, 8675 –8678
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8675

Communications
Similarly the 13C NMR spectrum of 2 is easily fully
assignable (see Supporting Information for details). The
molecular structure of 2 (Figure 1)[9] is mononuclear, with
Figure 1. Molecular structure of 2 with selected atom labels. Displace-
ment ellipsoids are drawn at the 50% probability level. Key bond
À
À
À
lengths [ꢂ]: Zn C2 2.0615(15), Zn C10 2.0611(16), Zn C13
À
À
2.0245(15), Zn N2 2.3313(13), N1 C10 1.5374(19).
Scheme 1.
both Zn and N(Me’) atoms chiral as each is attached to four
different substituents, though overall the crystal structure is
centrosymmetric and hence the material is racemic.[10] Three
anionic C atoms (two aliphatic sp3 atoms from the tBu and
deprotonated CH2N groups and one aryl sp2 atom) bind to
zinc [lengths 2.0245(15)/2.0611(16) and 2.0615(15) ꢁ, respec-
tively] with a dative interaction from the N(Me2) atom
[length, 2.3313(13) ꢁ] completing its distorted tetrahedral
coordination. Ring constraints severely distort this geometry
with bond angles ranging from 84.54(6)8 (C2-Zn1-C10) to
134.05(6)8 (C10-Zn1-C13) and a mean value of 107.398. The
atoms Zn and N1 lie slightly to one side of the benzene ring
plane (by 0.138 and 0.090 ꢁ, respectively).
Uchiyama paper.[6] Subsequent addition of TMEDA and
tBu2Zn across the benzyne triple bond leads to Int with
elimination of NaCl. The unexpected attachment of TMEDA
might be interpreted as a type of complex-induced proximity
effect (CIPE)[12] due to the Na(TMEDA) unit lying close to
the C atom originally carrying the Cl substituent. By
repeating the reaction with 4-chlorotoluene in place of
chlorobenzene to get a fix on the position of zincation, we
observed zinc adding to both the 3 and 4 positions, which
supports the existence, at least in part, of such a benzyne
mechanism (see Supporting Information for details).
Scheme 1 proposes a pathway for the formation of
zincacyclic 2. ortho-Zincation of chlorobenzene would be
the first step, almost certainly through heteroleptic 1 func-
tioning kinetically as a TMP base with concomitant release of
TMP(H).[11] This would generate a sodium monoaryl-bisal-
kylzincate intermediate for which there are close analogies in
lithium-TMP chemistry,[11] but the lack of TMP in the later
intermediate Int is perhaps the strongest evidence. Next, the
Zn and Cl s-attachments to the benzene ring could both
cleave to generate an o-benzyne intermediate and
“(tBu)2Zn·NaCl·TMEDA”. Militating against this mechanism
are Uchiyamaꢀs aforementioned studies of “LiZn(TMP)-
(tBu)2”,[6] which reveal no evidence for benzyne formation
towards haloarenes, but that reaction must follow a different
pathway to ours as the halogen substituent is retained on the
aryl ring, whereas it is cleaved in our case.
Intermediate Int provides definitive proof of this stage of
the reaction sequence. Made by reducing the experimental
conditions (no refluxing of the solution), Int has a molecular
structure (Figure 2)[9] akin to that of 2 but with two tBu
ligands attached to zinc and with an intact TMEDA molecule
occupying an ortho-position next to, but not attached to the
Zn(tBu)2 unit. In the final step, one tBu ligand deprotonates
one NMe2 arm of TMEDA to build an N-CH2-Zn bridge,
while concomitantly reduced steric constraints around the Zn
center allow the free Me2N terminal of the benzene-tethered
TMEDA to datively bind to Zn to close a seven-atom
(ZnCCNCCN) ring. Usually a-metalations of amines[13] have
high kinetic barriers and would conventionally be considered
impossible for only mildly electropositive zinc, but here the
barrier must be reduced substantially by the close proximity
of the tethered tBu2Zn and TMEDA units which sets up a
favorable intramolecular Zn–H exchange reaction with loss of
BuH.[14]
Note that at this stage in our reaction the benzyne
molecule may not be “naked” but could be involved in a
benzyne–zincate intermediate in which one of the metals
interacts with the p-system of the benzyne triple bond.
Evidence for this type of cation–p intermediate comes from
DFT calculations on model zincates in the aforementioned
As aforementioned, loss of the halogen substituent
distinguishes this new alkali metal mediated zincation
(AMMZn) reaction from those of “LiZn(TMP)(tBu)2”,
which, when intercepted by electrophiles, give polysubstitut-
8676 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8675 –8678

Angewandte
Chemie
philes, and electrophiles to generate (metal-free) benzannu-
lated cyclic products. Formally 2, 3, and Int could be
considered the products of three-component couplings of
PhCl, R2Zn (R = tBu or Me), and TMEDA (with concomitant
elimination of RCl and H2), but such reaction mixtures do not
lead to zincacyclic products (there was no trace of a zinc-aryl
product even when harsh reflux conditions were applied). In
practice, NaTMP or LiTMP must be present, so it is the
cooperativity between the different components within 1 and
4 that is behind the construction of 2 and 3, hence these
reactions can be interpreted as new examples of “synergic
synthesis”. The intriguing question now is: Can this reaction
be extended to a wide range of metals and neutral nucleo-
philes to provide access to new families of metalacyclic, ylidic,
and zwitterionic compounds?
Figure 2. Molecular structure of pre-bicyclic, pre-ylidic intermediate Int
with selected atom labels. Displacement ellipsoids are drawn at the
50% probability level. Key bond lengths and distances [ꢂ]: Zn1-C1
Experimental Section
All reactions were carried out under a protective argon atmosphere.
Full details can be found in the Supporting Information.
À
À
À
À
2.079(4), Zn1 C13 2.041(5), Zn1 C17 2.036(5), N1 C6 1.542(5), N1
Synthesis of 2: A hexane solution of 1 (2 mmol) was prepared as
described previously.[4] Chlorobenzene (2 mmol, 0.2 mL) was added,
and the mixture was heated to reflux for 1 h to produce a slightly
cloudy yellow solution. The Schlenk tube was surrounded with a
Dewar flask of hot water to allow the solution to cool slowly. This
yielded small, colorless, X-ray quality crystals (0.13 g, 21%). To
obtain a higher yield, a mixture of 1 (2 mmol) and chlorobenzene
(1 mmol) was stirred for 48 h, which produced a powder of 2 (0.16 g,
51% based on consumption of chlorobenzene).
Synthesis of 3: A solution of TMPH (2 mmol, 0.34 mL) in 10 mL
of dry hexane was treated with 1.25 mL of nBuLi (1.6m solution in
hexane, 2 mmol). The solution was stirred at room temperature for
30 min. Me2Zn (1.0m solution in heptane, 2 mmol, 2 mL) was then
added, before the addition of TMEDA (2 mmol, 0.3 mL). The
mixture was stirred for 1 h to give a pale yellow solution. Chlor-
obenzene (2 mmol, 0.2 mL) was added, and the solution was heated at
reflux for 1 h resulting in a slightly cloudy yellow solution. The
solution was filtered through Celite to remove a small amount of
precipitate in an effort to aid crystallization. The yellow solution was
concentrated with the removal of solvent in vacuum and was
transferred to a refrigerator operating at 08C. A small amount of
highly hydrocarbon-soluble colorless, cubic crystals were deposited
(0.07 g, 13%).
À
C7 1.494(5), N1 C8 1.508(5), C7···C13 4.007(7), C7···C17 4.220(6),
C8···C13 4.939(7), C8···C17 6.235(7).
ed haloarenes. Emphasizing the strong solvent dependency of
AMMZn reactions, significantly the former was performed in
hexane solution containing TMEDA, whereas the latter were
performed in THF solution with no TMEDA. Zincate
reactions are also sensitive to the alkyl group on Zn.
Interestingly, Uchiyama reports that “LiZn(TMP)(Me)2”
behaves differently from “LiZn(TMP)(tBu)2” towards halo-
arenes, generating benzyne intermediates which can be
trapped by dienes to produce polysubstituted halogen-free
aromatic compounds.[6] The reaction leading to 2 thus
represents a new, third variation on this theme.
Probing alkyl effects in our system, significantly we
isolated the methyl analog of 2, [{1-Zn(Me)}À-{2-N(Me)-
(CH2)CH2CH2NMe2}+-C6H4] (3), from the reaction of
chlorobenzene
and
the
lithium
methylzincate
“(TMEDA)·LiZn(TMP)(Me)2”[15] (4) (its crystalline yield
was only 13% reflecting its high solubility; see Supporting
Information for full details). Here the construction of the N-
CH2-Zn bridge must involve a methyl deprotonation (through
a “Me-Zn” unit) of an N-Me group, a reaction normally even
more difficult than a tert-butyl deprotonation (through a
“tBu-Zn” unit). Determined by X-ray crystallography, the
molecular structure of 3 is essentially identical to that of 2
with Me instead of tBu (see Supporting Information).
It is worth commenting that the primary role of TMEDA
in organometallic chemistry[16] is as a didentate N donor
ligand, but here in the formation of 2 and 3 it has an unusual
three-fold function. It acts first as a neutral N nucleophile
towards an electron-deficient aromatic C atom; second as an
anionic C nucleophile towards Zn (through an unprecedented
direct zincation of TMEDA); and third as a neutral N ligand
towards Zn. Tertiary amines are known to attack benzynes
nucleophilically.[17]
Synthesis of Int: A hexane solution of 1 (2 mmol) was prepared as
described previously.[4] Chlorobenzene (2 mmol, 0.2 mL) was added,
and the slightly cloudy yellow solution was transferred to a freezer
operating at À288C. Overnight, this yielded a mixture of small,
yellow, X-ray quality crystals of Int and colorless crystals of 2 (crude
yield, 0.15 g, which equates to 12% of each as determined from an
NMR analysis).
Received: August 12, 2009
Revised: September 15, 2009
Published online: October 14, 2009
Keywords: alkali metals · cooperative effects ·
.
metallacyclic compounds · metalation · zinc
[1] a) S. H. Wunderlich, P. Knochel, Angew. Chem. 2007, 119, 7829;
Angew. Chem. Int. Ed. 2007, 46, 7685; b) Z. Dong, G. C.
Clososki, S. H. Wunderlich, A. Unsinn, J. Li, P. Knochel,
Chem. Eur. J. 2009, 15, 457.
Yoshida et al. have recently demonstrated[18] elegant
three-component couplings of benzynes, unsaturated nucleo-
Angew. Chem. Int. Ed. 2009, 48, 8675 –8678
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8677

Communications
[2] C. J. Rohbogner, S. H. Wunderlich, G. C. Clososki, P. Knochel,
T= 123(2) K, Oxford Xcalibur S diffractometer) and 218 refined
Eur. J. Org. Chem. 2009, 1781.
parameters, difference map between + 0.720 and À0.411 eꢁÀ3
.
[3] a) Y. Kondo, M. Shilai, M. Uchiyama, T. Sakamoto, J. Am.
Chem. Soc. 1999, 121, 3539; b) M. Uchiyama, Y. Matsumoto, D.
Nobuto, T. Furuyama, K. Yamaguchi, K. Morokuma, J. Am.
Chem. Soc. 2006, 128, 8748.
[4] P. C. Andrikopoulos, D. R. Armstrong, H. R. L. Barley, W.
Clegg, S. H. Dale, E. Hevia, G. W. Honeyman, A. R. Kennedy,
R. E. Mulvey, J. Am. Chem. Soc. 2005, 127, 6184.
CCDC 741894 (2), 741895 (3) and 743469 (Int) contain the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[10] A search of the Cambridge Crystallographic Database revealed
no precedent for this structure. There is one example of an a-
zincated phosphorus ylide, see: M. Steiner, H. Grꢂtzmacher, H.
Pritzkow, L. Zsolnai, Chem. Commun. 1998, 285.
[5] R. E. Mulvey, Acc. Chem. Res. 2009, 42, 743.
[6] a) T. Harada, M. Chiba, A. Oku, J. Org. Chem. 1999, 64, 8210;
b) M. Uchiyama, Y. Kobayashi, T. Furuyama, S. Nakamura, Y.
Kajihara, T. Moyashi, T. Sakamoto, Y. Kondo, K. Morokuma, J.
Am. Chem. Soc. 2008, 130, 472.
[7] For reviews of directed ortho-metalation, see: a) H. W.
Gschwend, H. R. Rodriguez, Org. React. 1979, 26, 1; b) C. G.
Hartung, V. Snieckus in Modern Arene Chemistry (Ed.: D.
Astruc), Wiley-VCH, New York, 2002, pp. 330 – 367.
[8] For a review of zwitterionic organometalates, see: R. Chauvin,
Eur. J. Inorg. Chem. 2000, 577.
[9] Crystal data for 2: C16H28N2Zn, Mr = 313.8, monoclinic, space
group P21/c, a = 9.7870(11), b = 9.4959(10), c = 18.276(2) ꢁ, b =
97.588(2)8, V= 1683.6(3) ꢁ3, Z = 4; R(F) = 0.026 for 3466
reflections with F2 > 2s, Rw(F2) = 0.066 for all 3862 unique data
(MoKa radiation, l = 0.71073 ꢁ, T= 150 K, Nonius KappaCCD
diffractometer) and 178 refined parameters, difference map
[11] W. Clegg, B. Conway, E. Hevia, M. D. McCall, L. Russo, R. E.
Mulvey, J. Am. Chem. Soc. 2009, 131, 2375.
[12] M. C. Whisler, S. MacNeil, V. Snieckus, P. Beck, Angew. Chem.
2004, 116, 2256; Angew. Chem. Int. Ed. 2004, 43, 2206.
[13] For examples, see: a) D. Bojer, I. Kamps, X. Tian, A. Hepp, T.
Pape, R. Frꢃhlich, N. W. Mitzel, Angew. Chem. 2007, 119, 4254;
Angew. Chem. Int. Ed. 2007, 46, 4176; b) V. H. Gessner, C.
Strohmann, J. Am. Chem. Soc. 2008, 130, 14412.
[14] It is noted that the positive charge of the N atom is also a factor
in making the a-CH3 more acidic, akin to Lewis acid (e.g.
borane)-activated amines, see: S. V. Kessar, P. Singh, Chem. Rev.
1997, 97, 721.
[15] D. V. Graham, E. Hevia, A. R. Kennedy, R. E. Mulvey, Organo-
metallics 2006, 25, 3297.
[16] For an enlightening discussion on the role of TMEDA in group 1
chemistry see: D. B. Collum, Acc. Chem. Res. 1992, 25, 448.
[17] A. R. Lepley, R. H. Becker, A. G. Giumanini, J. Org. Chem.
1971, 36, 1222.
between + 0.40 and À0.27 eꢁÀ3
. Crystal data for Int:
C20H38N2Zn, Mr = 371.9, orthorhombic, space group Pbca, a =
9.6006(8), b = 19.4973(13), c = 22.8263(16) ꢁ, V= 4272.8(5) ꢁ3,
Z = 8; R(F) = 0.067 for 2255 reflections with F2 > 2s, Rw(F2) =
0.123 for all 4168 unique data (MoKa radiation, l = 0.71073 ꢁ,
[18] T. Morishita, H. Fukushima, H. Yoshida, J. Ohshita, A. Kunai, J.
Org. Chem. 2008, 73, 5452, and references therein.
8678 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8675 –8678