Structural studies on a lithium organo-amidocuprate in the solid state and in solution

Article abstract of DOI:10.1002/anie.200700822

(Figure Presented) Head-to-tail or head-to-head? A lithium amidocuprate complex has been characterized in the solid state and shown to adopt a head-to-tail conformation (see picture). 2D NMR spectroscopic studies show the presence of several structural isomers in solution, including a head-to-head isomer, resulting from Schlenk equilibrium. The presence of these isomers is postulated to have a significant influence on heterocuprate reactivity.

Full text of DOI:10.1002/anie.200700822

Angewandte
Chemie
DOI: 10.1002/anie.200700822
Organocopper Reagents
Structural Studies on a Lithium Organo-Amidocuprate in the Solid
State and in Solution**
Robert P. Davies,* Stefan Hornauer, and Peter B. Hitchcock
Lithium organo-amidocuprates are an important class of
organocopper reagents with applications in carbon–carbon
bond formation.[1] They differ from homocuprates, [R2CuLi],
by replacement of one of the reactive organic groups (R) with
a nontransferable amido group (L). Although heterocuprates,
including amidocuprates, were originally developed to cut
wastage of the reactive R group, they have also attracted
much attention for applications in enantioselective synthesis
when the nontransferable ligand L is chiral.[2,3] Crucial to the
optimization of these species for high yields and enantiomeric
excesses is a thorough understanding of their structures and
mechanisms of reaction. However, despite these implications
and in contrast to the large body of work on lithium
homocuprates,[1] there have been
ZnL,[10a] MgL,[10a] MnR,[10b] and AlR2,[10c] although as far as
we are aware there are currently no examples of these metals
forming dimeric aggregates similar to 1.
Reaction of CuMes (Mes = 2,4,6-trimethylphenyl) with
LiN(CH2Ph)2 in toluene gave the organo-amidocuprate of
stoichiometry [CuLiMes{N(CH2Ph)2}] (we have previously
demonstrated how the mesityl group can act as a model for
the reactive R group in organocuprates, wherein its lack of b-
hydrogen atoms results in increased thermal stability and ease
of handling[11]). Crystals of 3 were obtained from this reaction
solution and were characterized using X-ray diffraction
studies (Figure 1).[12] This study constitutes the first solid-
state structure of a hetero-amidocuprate.
very few studies on the coordination
chemistry of lithium amidocuprates.
Early work by Dieter and Rossiter
led to the assumption that these
species adopt a dimeric aggregate 1
(Scheme 1) in both their resting and
Scheme 1. Structural
models for lithium
reactive states.[4–6] More recently,
Davidsson and co-workers con-
firmed, using NMR spectroscopy,
the existence of such a dimer for
the heterocuprate with R = nBu and
L = (R)-N-methyl-1-phenyl-2-(1-pyr-
organo-amidocuprates
(R=organic group;
L=amido group).
rolidinyl)ethanamine.[7] However, monomeric aggregates
such as 2 (Scheme 1, M = Cu) have also been suggested to
play a role in the reaction of heterocuprates with enones,
particularly in polar solvents.[8] Calculations by Yamanaka
and Nakamura have shown the importance of the trans effect
Figure 1. The molecular structure of the Ci-symmetric complex 3;
thermal ellipsoids are set at 50%; hydrogen atoms are omitted for
clarity.
ꢀ
and strong Li L bonds for the nontransferability of the
heteroatom groups in these reagents.[9] In addition, studies on
related lithium heterobimetallic species have recently
revealed monomeric structures analogous to 2 for M =
The structure of 3 is consistent with previous predictions
for lithium hetero-amidocuprates (1, Scheme 1); it is dimeric
and contains two [RCuL]ꢀ cuprate units lying antiparallel, or
head-to-tail, to one another. The molecule is located on an
inversion center, with both cuprates adopting a nearly linear
[*] Dr. R. P. Davies, S. Hornauer
Department of Chemistry
ꢀ
conformation (C15-Cu-N = 175.3(2) Š). The Cu Cbond
Imperial College London, South Kensington
London, SW72AZ (UK)
Fax: (+44)870-1300438
lengths (1.913(5) Š) are shorter than in the related homo-
cuprate [Cu2Li2Mes4] (1.925(2) Š, 1.936(2) Š),[11] but compa-
rable to those reported for the [CuMes2]ꢀ anion
E-mail: r.davies@imperial.ac.uk
(1.915(9) Š).[13] The Cu N bond in 3 (1.921(4) Š) is signifi-
ꢀ
Dr. P. B. Hitchcock
School of Chemistry
University of Sussex
Brighton, BN19QJ (UK)
ꢀ
cantly longer than the Cu N bonds observed in the homo-
bis(amido)cuprates [Cu{N(SiMePh2)2}2]ꢀ (mean 1.88 Š)[14]
and [{Li(dme)Cu(NHMes)(NHPh)}2] (mean 1.855 Š; dme =
dimethoxyethane).[15] The Li+ ions in 3 are coordinated to
[**] This work was supported by the American Chemical Society
Petroleum Research Fund. P. R. Haycock is gratefully acknowledged
for his assistance with the 2D NMRspectroscopy.
ꢀ
both an amido nitrogen atom (Li N 1.969(10) Š) and a
ꢀ
mesityl aryl ring (Li Cin the range 2.334(10) to 2.445(10) Š;
mean 2.386 Š). Similar h6 coordination of Li+ ions has been
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
reported in the homocuprate [Cu2Li2Mes4].[11]
Angew. Chem. Int. Ed. 2007, 46, 5191 –5194
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5191

Communications
Although the solid-state structure of 3 provides some
The largest peak in the 7Li NMR spectrum, signal C( d =
ꢀ5.79 ppm), shows strong NOE cross-peaks with mesityl
CH3, dibenzylamido CH2, and aromatic ArH protons. Con-
sequently, LiC must lie in close proximity to both mesityl and
amido groups, and can therefore be assigned as the head-to-
tail dimer 3 (Scheme 2), which is analogous to the solid-state
valuable insights into the bonding within organo-amidocup-
rates, virtually all cuprate addition reactions are carried out in
solution, and hence elucidation of the solution structures is
crucial to understanding the reaction mechanisms and
7
selectivity of these reagents. Li NMR spectroscopic studies
on a [D8]toluene solution of 3 revealed no less than five main
lithium environments (corresponding to signals A–E in
Figure 2), indicative of at least a partial break up of the
Figure 2. 7Li NMRspectrum of 3 in [D8]toluene (298 K). See text for
peak assignments.
Scheme 2. Predicted Schlenk equilibrium for the lithium hetero-amido-
cuprate [Cu2Li2Mes2(NR2)2] in toluene (R=CH2Ph).
dimeric form 3 and the presence of Schlenk-type equilibrium.
Cyroscopic relative molecular mass (RMM) measurements
on 3 in benzene (0.19m solution) gave an association value of
n = 1.8 for [CuLiMes{N(CH2Ph)2}]n, suggesting a predom-
inance of dimeric aggregates.
structure. This assignment also fits with its high upfield shift of
d = ꢀ5.79 ppm, which is symptomatic of a lithium cation
sitting directly above one aromatic ring.[17,18]
Assignment of the solution species, as represented by the
different lithium environments in Figure 2, has been aided by
the use of modified pulse field gradient (PFG) inverse-
detected 1H–7Li HOESY NMR spectroscopy (Figure 3).[16]
Although no NOE cross-peaks are observable for peak D
owing to its weak signal, its chemical shift of d = ꢀ9.93 ppm is
almost identical to that previously reported for the h1,h6
bis(mesityl)-coordinated Li centers in [Cu2Li2Mes4] (d =
ꢀ9.99 ppm).[11] Hence, LiD can be assigned as arising from
this homocuprate (6, Scheme 2). Peak E is at an even higher
upfield shift (d = ꢀ11.88 ppm), suggestive of a Li center
sandwiched in an h6,h6 mode between two mesityl groups,
and hence being subjected to the full magnetic anisotropic
effect of both aromatic rings. In addition, peak E exhibits
NOE cross-peaks solely to mesityl CH3 and aromatic protons.
In our view, the most probable assignment for LiE is therefore
the head-to-head dimer 4. The other lithium center in 4 can be
accounted for by the peaks at signal A (d = + 1.11 ppm), as
these fall within the correct range expected for a bis(amido)-
coordinated Li center.[18] Signal A, in fact, consists of several
broad overlapping peaks and may also comprise the bis-
(amido)-coordinated
Li centers
in
the
homo-bis-
(amido)cuprate 7, which would be expected to form con-
currently with 6. The lack of any observable NOE cross-peaks
for the signals at A can be attributed to the Li+ ions being
located in between two quadrupolar N nuclei.
Peak B lies at d = ꢀ4.06 ppm and is therefore in the
correct region for a Li+ ion coordinated to the p system of one
aromatic ring. This signal possesses weak NOE cross-peaks
with amido methylene protons and ortho-CH3 mesityl pro-
tons. We tentatively assign this peak to the mixed homocup-
rate species 5, in which, because of geometric restraints, the
Li+ ions are h1-coordinated to the mesityl group and hence
more downfield than the h6-coordinated Li centers in 3.
Although an alternative assignment would perhaps be a
monomeric species similar to 2, this possibility can be
1
Figure 3. a) Modified PFG inverse-detected H–7Li HOESY NMRspec-
trum of [Cu2Li2Mes2{N(CH2Ph)2}2] in [D8]toluene. b)–d) Individual
slices through the 7Li NMRchemical shifts at d=ꢀ11.88 ppm (b),
d=ꢀ5.79 ppm (c), and d=ꢀ4.13 ppm (d).
5192
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5191 –5194

Angewandte
Chemie
discounted since the upfield shift of d = ꢀ4.06 ppm suggests
LiB is located at least partially above the aromatic ring, which
is not possible in the monomer owing to steric requirements.
Unlabelled minor peaks in the 7Li NMR spectrum (which can
be seen in the ranges d = + 0.5 to + 1.5 ppm and d = ꢀ3.0 to
ꢀ4.0 ppm) are speculatively assigned as [Cu2Li2{N-
(CH2Ph)2}3Mes], [Cu2Li2{N(CH2Ph)2}Mes3], and monomeric
species. It is worthwhile to note that no peaks were observed
for the lithium species [Li{N(CH2Ph)2}]n[19] in either the 1H or
7Li NMR spectra, signifying that 3 does not disassemble back
into its monometallic precursors.
lowest-energy optimized monomeric structure (Figure S8 in
the Supporting Information) was calculated to be consider-
ably higher in energy than I (+ 61.27 kcalmolꢀ1 for two
monomeric units), making the formation of monomeric
species seem unlikely. However, it is possible that any
[CuLiMes{N(CH2Ph)2}] monomeric complex would contain
additional stabilization from Li···HCinteractions, similar to
those previously reported for [Li{N(CH2Ph)2}]n,[19,20] thus
making its formation more favorable than that for the
calculated [CuLiMe(NH2)] model system.
In summary, the first solid-state structure of an organo-
amidocuprate has been presented and shown to form a head-
to-tail dimeric aggregate consistent with previous theories. In
addition, DFT calculations have been carried out, which
predict this head-to-tail dimer to be favored over other
possible structural isomers for a simple model system.
However, NMR studies on [Cu2Li2Mes2{N(CH2Ph)2}2] have
shown a mixture of structural isomers to be present in toluene
solution as a result of Schlenk equilibrium, and these
structural isomers have been identified with the aid of
1
Inspection of the H–7Li HOESY NMR spectrum (Fig-
ure 3c) allows the assignment of the dibenzylamido CH2
protons in 3 as two doublets at d = 3.67 and 3.88 ppm; their
inequivalence presumably is due to the geometric constraints
posed by the dimeric structure. Comparison of the integrals of
these signals with the other methylene proton signals in the
1H NMR spectrum reveals the head-to-tail dimer 3 to be the
dominant species in solution, accounting for 64% of the
amido groups (298 K, 0.22m in [D8]toluene). Variable-tem-
perature NMR studies show this ratio to remain approx-
imately constant in the temperature range 298 to 166 K.
Variable-concentration studies reveal the amount of 3 present
to fall with increasing dilution to 51% for a 0.02m [D8]toluene
solution. This decrease may be due to dissociation of the
dimer to give monomeric species as suggested by the
cryoscopic RMM studies (see above).
1
modified PFG inverse-detected H–7Li HOESY NMR spec-
troscopy. To our knowledge, this is the first time such
structural isomers have been identified in solution. The
presence of more than one isomer in solution is potentially
significant for the mechanism of hetero-amidocuprate addi-
tion reactions, and is of particular relevance to asymmetric
addition reactions involving chiral hetero-amidocuprate
reagents where high enantiomeric excesses are sought.
Density functional calculations (B3LYP) were performed
on the model system [Cu2Li2Me2(NH2)2] to further investigate
the relative thermodynamic stabilities of the different lithium
hetero-amidocuprate structural isomers. The lowest energy of Experimental Section
All reactions were carried out in a protective nitrogen atmosphere.
all the optimized structural isomers (Figure 4) is the head-to-
3: sBuLi in hexanes (1.3m, 1.53 mL, 2 mmol) was added to a
solution of dibenzylamine (340 mg, 2 mmol) in toluene (5 mL) at 08C.
The solution was allowed to warm to room temperature and stirred
for 1 h to give a light-red precipitate. A solution of CuMes (365 mg,
2 mmol) in toluene (5 mL) was added dropwise to give a clear red
solution, which was filtered through celite and concentrated under
vacuum to approximately 2 mL. Storage at 48Cfor two days yielded
colorless crystals suitable for X-ray analysis (431 mg, 56%); m.p.
898C(decomp).
Received: February 23, 2007
Published online: May 25, 2007
Keywords: copper · cuprates · heteroleptic complexes · lithium ·
.
NMRspectroscopy
Figure 4. Schematic diagram of the optimized structural isomers of
[Cu2Li2Me2(NH2)2] at the B3LYP/631AS level (energies relative to I in
kcalmolꢀ1; see the Supporting Information for full details).
[1] Modern Organocopper Chemistry (Ed.: N. Krause), Wiley-VCH,
Weinheim, 2002, and references therein.
[2] a) B. E. Rossiter, N. M. Swingle, Chem. Rev. 1992, 92, 771 – 806;
b) N. Krause, A. Gerold, Angew. Chem. 1997, 109, 194 – 213;
Angew. Chem. Int. Ed. Engl. 1997, 36, 187 – 204; c) R. K. Dieter
in Modern Organocopper Chemistry (Ed.: N. Krause), Wiley-
VCH, Weinheim, 2002, pp. 79 – 144.
[3] For a recent example, see: S. H. Bertz, C. A. Ogle, A. Rastogi, J.
Am. Chem. Soc. 2005, 127, 1372 – 1373.
[4] R. K. Dieter, M. Tokles, J. Am. Chem. Soc. 1987, 109, 2040 –
2046.
tail dimer I. This result is in agreement with our solid-state
and solution structural studies on [Cu2Li2Mes2{N(CH2Ph)2}2],
for which this isomer was shown to be the dominant species.
The head-to-head dimer II is + 0.744 kcalmolꢀ1 higher in
energy, and the mixed homodimer III is + 1.763 kcalmolꢀ1
higher in energy. The optimized homodimers (IV and V) and
mixed homo- and heterodimers (VI and VII) were also
confirmed to be higher in energy than I (Figure 4). The
[5] R. K. Dieter, T. W. Hanks, B. Lagu, Organometallics 1992, 11,
3549 – 3554.
Angew. Chem. Int. Ed. 2007, 46, 5191 –5194
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5193

Communications
[6] B. E. Rossiter, M. Eguchi, G. Miao, N. M. Swingle, A. E.
free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[13] P. Leoni, M. Pasquali, C. A. Ghilardi, J. Chem. Soc. Chem.
Commun. 1983, 240 – 241.
[14] P. P. Power, K. Ruhlandt-Senge, S. C. Shoner, Inorg. Chem. 1991,
30, 5013 – 5015.
[15] P. Reiss, D. Fenske, Z. Anorg. Allg. Chem. 2000, 626, 1317 – 1331.
[16] a) W. Bauer, Magn. Reson. Chem. 1996, 34, 532 – 537; b) T. M.
Alam, D. M. Pedrotty, T. J. Boyle, Magn. Reson. Chem. 2002, 40,
361 – 365.
[17] W. Bauer in Lithium Chemistry: a Theoretical and Experimental
Overview (Eds.: A.-M. Sapse, P. von R. Schleyer), Wiley, New
York, 1995, pp. 125 – 172.
[18] H. Gꢀnther in Encyclopedia of Nuclear Magnetic Resonance,
Vol. 5 (Eds.: D. M. Grant R. K. Harris), Wiley, New York, 1996,
pp. 2807 – 2828.
[19] a) D. Reed, D. Barr, R. E. Mulvey, R. Snaith, J. Chem. Soc.
Dalton Trans. 1986, 557 – 564; b) D. R. Armstrong, R. E. Mulvey,
G. T. Walker, D. Barr, R. Snaith, W. Clegg, D. Reed, J. Chem.
Soc. Dalton Trans. 1988, 617 – 628.
Hernandez, D. Vickers, E. Fluckiger, R. G. Patterson, K. V.
Reddy, Tetrahedron 1993, 49, 965 – 986.
[7] J. Eriksson, P. I. Arvidsson, O. Davidsson, J. Am. Chem. Soc.
2000, 122, 9310 – 9311.
[8] J. Eriksson, O. Davidsson, Organometallics 2001, 20, 4763 – 4765.
[9] M. Yamanaka, E. Nakamura, J. Am. Chem. Soc. 2005, 127, 4697 –
4706.
[10] a) R. E. Mulvey, Organometallics 2006, 25, 1060 – 1075, and
references therein; b) J. Garcia-lvarez, A. R. Kennedy, J. Klett,
R. Mulvey, Angew. Chem. 2007, 119, 1123 – 1126; Angew. Chem.
Int. Ed. 2007, 46, 1105 – 1108; c) H. Naka, M. Uchiyama, Y.
Matsumoto, A. E. H. Wheatley, M. McPartlin, J. V. Morey, Y.
Kondo, J. Am. Chem. Soc. 2007, 129, 1921 – 1930.
[11] R. P. Davies, S. Hornauer, Chem. Commun. 2007, 304 – 306.
[12] Crystal data for 3: C46H50Cu2Li2N2, Mr = 771.84, triclinic, space
¯
group P1 (no. 2), a = 9.0311(4), b = 10.8223(4), c = 11.4544(6) Š,
a = 77.951(3), b = 71.413(2), g = 70.608(3)8, V= 994.28(8) Š3,
Z = 1, 1calcd = 1.29 gcmꢀ3, m(MoKa) = 1.10 mmꢀ1, T= 173(2) K;
3885 independent measured reflections, R1 = 0.061, wR2 = 0.177,
3214 independent observed reflections [ j Fo j > 4s(jFoj), 2qmax
=
[20] D. Barr, W. Clegg, R. E. Mulvey, R. Snaith, J. Chem. Soc. Chem.
Commun. 1984, 287 – 289.
528], 237 parameters. CCDC-637881 contains the supplementary
crystallographic data for this paper. These data can be obtained
5194
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5191 –5194