Models for the reactive states of homocuprates: Syntheses, structures and reactivities of [Cu2Li2Mes4] and [Cu 3LiMes4]

Article abstract of DOI:10.1039/b612554a

The synthesis and characterisation of two novel tetranuclear and thermally-stable lithium arylcuprates, [Cu₂Li₂Mes ₄] and [Cu₃LiMes₄], are reported and [Cu ₃LiMes₄] is shown to be a highly active promoter for the 1,4-addition of organolithiums to enones. The Royal Society of Chemistry.

Full text of DOI:10.1039/b612554a

View Article Online / Journal Homepage / Table of Contents for this issue
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Models for the reactive states of homocuprates: syntheses, structures
and reactivities of [Cu2Li2Mes4] and [Cu3LiMes4]{
Robert P. Davies,* Stefan Hornauer and Andrew J. P. White
Received (in Cambridge, UK) 1st September 2006, Accepted 16th October 2006
First published as an Advance Article on the web 30th October 2006
DOI: 10.1039/b612554a
The synthesis and characterisation of two novel tetranuclear
and thermally-stable lithium arylcuprates, [Cu2Li2Mes4] and
[Cu3LiMes4], are reported and [Cu3LiMes4] is shown to be a
highly active promoter for the 1,4-addition of organolithiums
to enones.
Lithium homocuprates, LiCuR2 (R = alkyl, aryl), are one of the
oldest and most commonly used reagents for the generation of
their separation and isolation as pure products in yields of 27 and
carbon–carbon bonds via conjugate addition, and are thus key to
33% respectively (see ESI for experimental procedure{).
many synthetic endeavours.1 Although the structures and reaction
mechanisms of these reagents has in the past been a subject of
much debate, it is now generally accepted that their resting state
(and probably the reactive species too) is the neutral dimeric form
[R2CuLi]2 (1).1 Fundamental to our current understanding of
lithium homocuprates are the solid-state structural characterisa-
tions of [Cu2Li2Ph4(OEt2)2] (2),2 [Cu2Li2Ph4(SMe2)3] (3),3
[Cu2Li2(C6H4CH2NMe2-2)4] (4),4 [Cu2Li2(CH2SiMe3)4(OEt2)3]5
Although moisture sensitive, compounds 5 and 6 are unusually
thermally stable for lithium cuprates (decomp. 189, 174 uC
respectively). This can, at least in part, be attributed to their lack of
b-hydrogens, thus avoiding decomposition by b-hydride elimina-
tion pathways. Both 5 and 6 were fully characterised in the solid-
state using single-crystal X-ray diffraction{ and in solution using
multinuclear NMR spectroscopy.{
The solid-state structure of 5 (Fig. 1) is comparable in
stoichiometry to other previously reported neutral homocuprates
(1). However, the coordination mode of the Li cation differs from
that observed in previously reported diarylcuprates (2–4), which all
contain 3c–2e Li–C–Cu bonds centred upon the aryl ipso-carbons.
In 5 each Li is effectively sandwiched between two co-planar
mesityl groups, being g6 coordinated to one (mean Li–C =
6
and [Cu2Li2(CH2SiMe3)4(SMe2)2]‘ which all contain dimeric
structures of type 1 with the alkyl or aryl R group asymmetrically
bridging the Cu and Li centres to form a 3 centre 2 electron (3c–2e)
bond. In all of these structures the lithium is additionally
coordinated by either a solvent donor molecule (Et2O or Me2S)
or in the case of 4 an amino side-arm donor on the R group. Other
solid-state structural studies of lithium homocuprates include a
monomeric R2CuLi homocuprate where R is the very bulky
C6H3Mes2-2,6 ligand,7 and five ion-separated compounds contain-
ing [R2Cu]2 cuprate anions and Li+ cations solvated by 12-crown-
4 or other strongly coordinating ligands.5,8,9 In addition several
lithium homocuprate clusters with differing Cu to Li ratios have
been reported: [Li3Cu2Ph5(SMe2)4],3,10 [Li5Cu4Ph9(SMe2)4]3 and
the anionic clusters [Li2Cu3Ph6]211 and [LiCu4Ph6]2.12 However,
in contrast to the dimeric form 1, at present there is no direct evi-
dence to suggest that any of these non-stoichiometric species exist
in any quantifiable amount in lithium diorganocuprate solutions.
We now report on the synthesis, characterisation and reactivity
of two novel lithium homocuprates, [Cu2Li2Mes4] (5) and
[Cu3LiMes4] (6), prepared from the 1 : 1 reaction of CuMes with
LiMes (Mes = C6H2Me3-2,4,6) in toluene (Scheme 1). The
differing solubilities of 5 and 6 in hexane and toluene facilitated
˚
˚
2.316(4) A; range 2.271(4) to 2.365(4) A) and sitting almost directly
1
above (g ) the ipso-carbon of the other (Li–C11 = 2.129(4) A). The
˚
Ar2Cu units in 5 are close to planar, C–Cu–C = 178.34(7)u,
although the mesityl rings are twisted by approximately 9.6u to one
another. This contrasts with the less obtuse C–Cu–C bond angles
in 2–4 (range 157.7(1) to 168.0(2)u),2–4 and is closer in magnitude
to the C–Cu–C bond angles in the ion-separate species [Mes2Cu]2
(180.0(7)u)13 and [Ph2Cu]2 (178.5(4)u8 and 174.8(4)u14). Hence, in 5
the ipso-carbon sp2 lone pairs are almost exclusively bonded to the
Cu atoms giving 2c–2e Cu–C bonds, in preference to the 3c–2e
bonds observed in 2–4. This also results in a shortening of the
˚
Cu–C bond lengths (1.925(2), 1.936(2) A in 5) compared to those
2–4
in 2–4 (range 1.936(3) to 1.948(3) A):
13
cf. [CuMes2]2
˚
(1.915(9) A); [CuPh2]2 (range 1.900(11) to 1.931(11) A).
8,14
˚
˚
One possible explanation for the differing Cu–C bonding modes in
Department of Chemistry, Imperial College London, South Kensington,
London, UK SW7 2AZ. E-mail: r.davies@imperial.ac.uk;
Fax: +44 870 1300438; Tel: +44 207 5945754
{ Electronic supplementary information (ESI) available: Full experimental
details and characterizations of 5 and 6, details of the experimental
protocol for measuring the addition reaction yields, additional information
on the theoretical calculations including pictures of the 3D optimised
structures, and additional crystallographic details. See DOI: 10.1039/
b612554a
Scheme 1
304 | Chem. Commun., 2007, 304–306
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Fig. 1 The molecular structure of the Ci-symmetric complex 5.
5 and 2–4 is that the additional Li-coordinating Lewis-basic donor
ligands in 2–4 prevent the parallel stacking of the aryl groups due
to steric considerations, hence forcing them to adopt a more
orthogonal placement and favouring the formation of asymmetric
3c–2e Cu–C–Li bonds. Note however that Au–C 2c–2e bonds
have been reported in the related gold complex
[Au2Li2(C6H4CH2NMe2-2)4],15 and similar aryl bridging modes
have also been observed in heterobimetallic Li/Na/K–Mg/Zn
systems.16
Fig. 3 The molecular structure of one of the two independent complexes
present in the crystals of 6.
postulated that Cu3LiR4 species are present in small quantities in
lithium diorganocuprate solutions,18 as far as we are aware this is
the first characterised example of such a species in either the solid
state or solution. The Li cation in 6 is sandwiched directly between
two aromatic rings with g6,g6 coordination (mean C–Li =
In order to further examine the nature of the bonding in 5, we
carried out density functional calculations (B3LYP) on the three
possible conformers of [Cu2Li2Mes4] (Fig. 2).{ In agreement with
the structure obtained from X-ray crystallography, these calcula-
˚
˚
2.393(8) A; range 2.304(7) to 2.531(8) A). The Mes2Cu3 ‘half’ of
6 is similar in structure to that of Cu4Mes419 with the Mes groups
arranged approximately orthogonal to the Cu–Cu axes and
coordinated to two Cu centres via their ipso-carbon to give 3c–2e
Cu–C–Cu bonds. The C–Cu–C angle at Cu(3) is 159.60(16)/
159.53(15)u whereas the mean C–Cu–C angle at Cu1/2 is
171.07(16)u. The average Cu–C bond distance in 6 for these
3c–2e bonds (Cu1–C1, Cu2–C21, Cu3–C1 and Cu3–C21) is
tions showed the optimised g6 g1 conformer I to be lower in
,
energy than either the optimised g6,g6 (II) or g1,g1 (III)
conformers. In addition, natural population analysis revealed a
high positive natural charge on the Li centre in I (+0.945), which
reflects the highly electrostatic nature of the Li-aryl interactions in
this conformer. This compares to Li charges of +0.885 in II and
+0.831 in III. The lower natural charge on Li in III can be
accounted for by the additional participation of this atom in
asymmetric 3c–2e Cu–C–Li bonds.
˚
2.007(4) A (range 1.999(4) to 2.016(4) A): this is longer than the
˚
Cu–C bonds in 5 and comparable with Cu–C distances in
19
Cu4Mes4 (mean 1.993(10) A) and Cu5Mes5 (mean 1.96(3) A).
20
˚
˚
The remaining Cu–C bonds in 6 (Cu1–C11, Cu2–C31) are shorter
˚
(mean 1.936(4) A), and are therefore similar in length to the Cu–C
NMR spectroscopic studies{ are consistent with retention of the
g6,g1 dimeric structure of 5 in benzene solution. In particular, only
one singlet peak is observed in the 7Li spectra at 29.99 ppm. The
high chemical shift of this peak can be attributed to ring current
phenomena, whereby the Li sits in the magnetically anisotropic
environment of the two aryl rings.17
bonds in 5 and can likewise be considered 2c–2e bonds.
Similar to 5, 6 is highly stable in benzene giving just one peak in
its 7Li NMR spectrum at 211.02 ppm. The chemical shift for the
Li centre in 6 is further upfield than that in 5 since the Li is now
directly sandwiched in the middle of two eclipsed aromatic rings
and therefore experiences the maximum effect of the two ring
currents. Further NMR spectroscopic studies on separate samples
of 5 and 6 in toluene showed no signs of any solution equilibria
occurring, with both complexes retaining their respective tetra-
nuclear structures even after heating at 80 uC for 48 h. It was
previously postulated by van Koten and Noltes21 that Cu3LiAr4
species (such as 6) are present in lithium diarylcuprate solutions
due to interaggregate exchange between Cu2Li2Ar4 and Cu4Ar4
species (possibly via octanuclear intermediates). In agreement with
this hypothesis, reaction of 5 with Cu4Mes4 (monitored in situ
using 7Li NMR spectroscopy) was observed to result in the
formation of the Cu3LiAr4 species 6. It is noteworthy that this
reaction sits, as far as is detectable, on the side of 6 with no
evidence of formation of 5 and Cu4Mes4 from the back reaction.
Finally it has previously been reported that CuMes can act as a
reagent for promoting the 1,4-addition reactions of organolithium
The X-ray crystal structure{ of 6 contains two very similar
[Cu3LiMes4] cuprate complexes (one of which is shown in Fig. 3,
and the other in Fig. S5 in the ESI{) and a molecule of hexane.
This tetranuclear cuprate complex can best be considered as either
a dimer similar to 1 with one of the Li cations replaced by a Cu(I)
centre, or alternatively as a [Cu4Mes4] tetramer with one of the
Cu(I) centres replaced by Li. Although it has previously been
Fig. 2 Schematic diagram of the optimised structural conformers of
[Cu2Li2Mes4] at the B3LYP/631AS level{ (relative energies, kcal mol21).
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 304–306 | 305

View Article Online
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
higher 1,4-selectivity in this solvent. Further studies on the
structures and solution equilibria of 5 and 6 and their heteroleptic
derivatives in different solvent systems are currently ongoing.
We thank the donors of the American Chemical Society
Petroleum Research Fund for support of this research.
Table 1 Addition reactions of lithium organocuprate reagents
Reagent
Solvent
Yield 7 (%)a
Yield 8 (%)a
Notes and references
CuMes + nBuLi
Toluene
THF
Et2O
Toluene
THF
Et2O
Toluene
THF
Et2O
61
66
81
29
12
52
82
96
95
3
8
4
53
66
7
0
1
0
{ Crystal data for 5: C36H44Cu2Li2, M = 617.67, P21/n (no. 14),
˚
a = 13.3359(7), b = 8.6156(5), c = 14.4588(7) A, b = 106.459(5)u, V =
3
1593.19(15) A , Z = 2 (Ci symmetry), Dc = 1.288 g cm23, m(Mo-Ka) =
˚
1.357 mm21, T = 173 K, colourless plates; 5507 independent measured
reflections, R1 = 0.044, wR2 = 0.106, 3405 independent observed absorption
corrected reflections [|Fo| . 4s(|Fo|), 2hmax = 65u], 187 parameters. 6:
5 + 2nBuLi
6 + 2nBuLi
a
C39H51Cu3Li2, M = 717.36, P21/c (no. 14), a = 17.3561(18), b = 28.142(2),
3
˚
˚
c = 15.7048(11) A, b = 111.095(8)u, V = 7156.6(10) A , Z = 8 (2
independent complexes and one n-hexane in the asymmetric unit), Dc =
1.332 g cm23, m(Cu-Ka) = 2.231 mm21, T = 173(2) K, colourless needles;
13781 independent measured reflections, R1 = 0.036, wR2 = 0.050, 4347
independent observed absorption corrected reflections [|Fo| . 4s(|Fo|),
2hmax = 143u], 800 parameters. CCDC 619872 and 691873. For crystal-
lographic data in CIF or other electronic format see DOI: 10.1039/
b612554a
Yields were determined by GC against an internal standard of
decane and are based on nBuLi.
reagents (RLi) to enones.22 The reactive organocuprate reagent
responsible for the conjugate addition of the R group is thought to
be a heteroleptic complex of general formula [(MesCuR)Li]n, in
1 See N. Krause, Modern Organocopper Chemistry, Wiley-VCH,
Weinheim, 2002 and references therein.
which the Mes group acts as a non-transferable holding group.22
A
2 N. P. Lorenzen and E. Weiss, Angew. Chem., Int. Ed. Engl., 1990, 29,
300.
3 M. M. Olmstead and P. P. Power, J. Am. Chem. Soc., 1990, 112, 8008.
4 G. van Koten, J. T. B. H. Jastrzebski, F. Muller and C. H. Stam, J. Am.
Chem. Soc., 1985, 107, 697.
5 M. John, C. Auel, C. Behrens, M. Marsch, K. Harms, F. Bosold,
R. M. Gschwind, P. R. Rajamohanan and G. Boche, Chem.–Eur. J.,
2000, 6, 3060.
6 M. M. Olmstead and P. P. Power, Organometallics, 1990, 9, 1720.
7 M. Niemeyer, Organometallics, 1998, 17, 4649.
8 H. Hope, M. M. Olmstead, P. P. Power, J. Sandell and X. Xu, J. Am.
Chem. Soc., 1985, 107, 4337.
9 C. Eaborn, P. B. Hitchcock, J. D. Smith and A. C. Sullivan,
J. Organomet. Chem., 1984, 263, C23.
10 M. M. Olmstead and P. P. Power, J. Am. Chem. Soc., 1989, 111, 4135.
11 H. Hope, D. Oram and P. P. Power, J. Am. Chem. Soc., 1984, 106,
1149.
12 S. I. Khan, P. G. Edwards, H. S. H. Yuan and R. Bau, J. Am. Chem.
Soc., 1985, 107, 1682.
comparison of the application of the new lithium organocuprates 5
and 6 with CuMes in the promotion of 1,4-addition is given in
Table 1 for the reaction of nBuLi with cyclohexen-2-one. Note that
no transfer of the Mes group to the cyclic enone was observed in
any of the reactions.
The reactivity results (Table 1) show 6 to be highly active and
regioselective in promoting the conjugate addition reaction of
nBuLi, giving yields in excess of those achieved for CuMes, with
virtually zero contamination from 1,2-addition products. Possible
explanations for the improved reactivity of 6 over CuMes include:
the equilibrium reaction of 6 with nBuLi lies further on the side of
the reactive heterocuprate [MesCu(nBu)Li] than the reaction of
CuMes with nBuLi does; or alternatively the higher Mes:nBu ratio
results in the formation of more reactive species with different
stoichiometries, for example [Cu2Li2Mes3(nBu)].
13 P. Leoni, M. Pasquali and C. A. Ghilardi, J. Chem. Soc., Chem.
Commun., 1983, 240.
14 R. P. Davies and S. Hornauer, Eur. J. Inorg. Chem., 2005, 51.
15 G. van Koten, J. T. B. H. Jastrzebski, C. H. Stam and N. C. Niemann,
J. Am. Chem. Soc., 1984, 106, 1880.
16 R. E. Mulvey, Organometallics, 2006, 25, 1060.
17 H. Gu¨nther, in Encyclopedia of NMR, ed. D. M. Grant and R. K.
Harris, Wiley, Chichester, 1996, vol. 5, p. 2807.
18 G. van Koten and J. G. Noltes, J. Organomet. Chem., 1979, 174, 367.
19 H. Eriksson and M. Hakansson, Organometallics, 1997, 16, 4243.
20 E. M. Meyer, S. Gambarotta, C. Floriani, A. Chiesivilla and
C. Guastini, Organometallics, 1989, 8, 1067.
Regioselectivities for the addition reactions using 5 are
significantly worse than for 6 or CuMes, with a high degree of
1,2-addition suggesting incomplete heterocuprate formation and
direct addition of unreacted nBuLi with the cyclohexenone. This is
perhaps not surprising given this is the only one of the three
reagents in which the Li : Cu ratio is greater than 1 : 1, therefore
increasing the likelihood of excess nBuLi being present in solution.
The regioselectivity of 5 is also extremely solvent dependent, giving
predominately 1,2-addition in toluene or THF and 1,4-addition in
Et2O. Recent studies have shown organocuprates to be far more
reactive in Et2O than THF due to the formation of contact ion-
pairs,5 and this increased reactivity in Et2O can perhaps explain the
21 G. van Koten and J. G. Noltes, J. Am. Chem. Soc., 1979, 101, 6593.
22 T. Tsuda, T. Yazawa, K. Watanabe, T. Fujii and T. Saegusa, J. Org.
Chem., 1981, 46, 192.
306 | Chem. Commun., 2007, 304–306
This journal is ß The Royal Society of Chemistry 2007