Structural characterization of magnesium organocuprates derived from Grignard reagents: CuI-based inverse crown ethers

Article abstract of DOI:10.1002/anie.200800684

(Figure Presented) Shared interests: The structures of the title compounds (see example) were found to be directly comparable to those of lithium organocuprates, with the formal replacement of the lithium cation with a magnesium halide cation. Furthermore, both types of organocuprates form contact ion pairs in weakly coordinating solvents and solvent-separated ion pairs in strongly coordinating solvents.

Full text of DOI:10.1002/anie.200800684

Communications
DOI: 10.1002/anie.200800684
Organocopper Reagents
Structural Characterization of Magnesium Organocuprates Derived
from Grignard Reagents: CuI-Based Inverse Crown Ethers**
Roberta Bomparola, Robert P. Davies,* Stefan Hornauer, and Andrew J. P. White
Organocuprates are excellent reagents for the formation of
carbon–carbon bonds and have been used extensively in
synthetic methodology since the 1960s.[1] They are typically
formed in situ through the reaction of either an organolithium
or magnesium Grignard reagent with a copper(I) salt. The
resting state and also reactive form of lithium organocuprates
in nonpolar or weakly coordinating solvents, such as Et2O, is
now commonly accepted to be a dimeric contact ion pair
(CIP) species I. In more strongly coordinating solvents, such
[({Ph2Cu}{Mg(OEt2)I})2] (1) were obtained (Scheme 1). X-
ray crystallography[9] showed 1 to exist as a centrosymmetric
ꢀ
ꢀ
ꢀ
dimer consisting of two cuprate [Ph Cu Ph] anions con-
Scheme 1. Synthesis of 1.
nected by Et2O-solvated (MgI)+ monocations (Figure 1). The
Mg centers in 1 are approximately tetrahedral with the Mg-
coordinated iodide centers lying one above and one below the
as THF, a less reactive solvent-separated ion pair (SSIP) II is
predominant.[2,3]
However, in stark contrast to recent advances in the
understanding of lithium organocuprates, the structures of the
equally ubiquitous magnesium organocuprates derived from
Grignard reagents remain unexplored. To our knowledge,
[Cu4MgPh6(OEt2)], prepared by the treatment of CuBr with
MgPh2 in Et2O, is the only structurally characterized example
of a magnesium organocuprate,[4] although the aryl copper
magnesium arenethiolate [{Cu4Mes4}{m-SAr}2{MgSAr}2]
(Mes = 2,4,6-Me3C6H2; Ar= 2-Me2NCH(Me)C6H4—formed
from the self-assembly of (CuSAr)3 and MgMes2—has also
been proposed as a possible active species in Mg-catalyzed
cuprate addition reactions.[5,6] Herein, we present the first
structural determination of magnesium organocuprates
derived from Grignard reagents. CuIMes was employed as
the organocopper starting material in all reactions because of
its high thermal stability and previous use in the preparation
of stable lithium homo- and heteroamidocuprates.[7,8]
The treatment of CuMes with PhMgI in Et2O/toluene
gave a clear solution from which colorless crystals of
Figure 1. The molecular structure of the Ci-symmetric complex 1.
central eight-membered Cu2Mg2C4 ring (Mg–I 2.7046(10) Š).
It is salient that the structure is therefore isostructural to the
previously reported lithium diphenylcuprate [({Ph2Cu}{Li-
(OEt2)})2],[10] with the lithium cations formally replaced by
(MgI)+ monocations. Thus, we believe these results provide
the first experimental evidence that magnesium Grignard
diorganocuprates are able to form analogous CIP structures
to those (of type I) now well-established for lithium diorga-
nocuprates.
[*] R. Bomparola, Dr. R. P. Davies, S. Hornauer, Dr. A. J. P. White
Department of Chemistry
The phenyl groups in 1 bridge the Cu and Mg centers to
form asymmetric three-center, two-electron (3c–2e) Mg C
Imperial College London
South Kensington, London, SW7 5AZ (UK)
Fax: ( +44)870-130-0438
ꢀ ꢀ
Cu bonds. The Cu–Cipso distances in 1 (Cu–C(7) 1.939(3), Cu–
C(1) 1.945(3) Š) are comparable to those in [({Ph2Cu}{Li-
(OEt2)})2][10] and other lithium diaryl cuprates (1.906(5)–
E-mail: r.davies@imperial.ac.uk
[**] This research was supported by the EPSRC and the American
Chemical Society Petroleum Research Fund.
1.948(3) Š),[10–12] all of which contain analogous 3c–2e Cu C
ꢀ ꢀ
Li bonds. However, at 156.97(13)8, the C-Cu-C bond angle in
1 is less obtuse than the equivalent angles in these lithium
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5812
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5812 –5815

Angewandte
Chemie
diaryl cuprates (mean angle: 162.88).[10–12] Furthermore, the
Mg–C distances in (Mg–C(1) 2.261(3), Mg–C(7A)
2.290(3) Š) are significantly shorter than those in the only
other magnesium organocuprate studied, [Cu4MgPh6(OEt2)]
(mean Mg–C distance: 2.35(1) Š), in which the Mg center
bridges three phenyl groups.[4]
Although crystalline 1 was obtained in moderate yield
(31% with respect to PhMgI) from the reaction in Scheme 1,
the fate of the Mes ligand remained unknown, and it was
apparent that at least one other product must also be formed
in the reaction mixture. We envisaged initially that this
product could take the form of an analogous [{(Mes2Cu)-
(MgI)}2] dimer; however, NMR spectroscopic experiments to
investigate this possibility proved inconclusive. To probe the
identity of this additional product, the reaction was replicated
with MesMgBr as the Grignard reagent (Scheme 2).
The cation in 2 is unexceptional: It contains an octahedral
1
Mg center coordinated to five THF molecules and one
bromide atom (Mg–Br 2.586(3) Š). However, the dinuclear
[Mes3Cu2]ꢀ anion is unique, with two terminal Mes groups,
which can be considered to form short 2c–2e bonds with the
Cu centers (Cu(1)–C(11) 1.923(8), Cu(2)–C(21) 1.925(8) Š),
and one Mes group, which bridges both Cu centers through an
ꢀ ꢀ
approximately symmetric 3c–2e Cu C Cu bond to give
longer Cu–C distances (Cu(1)–C(1) 2.020(7), Cu(2)–C(1)
2.003(8) Š). Structures of the type [Cu2R3]ꢀ were not known
previously in organocopper chemistry, although the organo-
iodocuprate [2,6-Mes2H3C6Cu2I2]ꢀ does have
a similar
arrangement, with iodine atoms instead of organo groups in
the terminal positions.[14] The bond angles in 2 at Cu(1) and
Cu(2) are 168.0(3) and 171.0(3)8, respectively, and the Cu-C-
Cu angle at the bridging ipso C atom is 75.1(3)8. This
arrangement gives rise to an intramolecular Cu···Cu distance
of 2.4515(14) Š. Although this Cu···Cu contact is very short, it
does not imply the presence of any significant Cu–Cu bonding
interactions: Recent computational studies on the nature of
any possible d10–d10 interactions between CuI centers showed
that such interactions between CuI centers of similar charge
are likely to be weak or nonexistent.[15]
Repetition of the reaction between CuMes and MesMgBr
in the less strongly coordinating medium Et2O/toluene gave a
different product: [Cu4Mg2Mes6Br2] (3; Scheme 2). X-ray
crystallographic studies[16] revealed the crystals of 3 to contain
two structurally identical but independent Ci-symmetric
complexes 3a and 3b in the asymmetric unit (see the
Supporting Information), one of which is shown in Figure 3.
Complex 3 adopts a CIP arrangement, formally an aggrega-
Scheme 2. Synthesis of 2 and 3.
The treatment of CuMes with MesMgBr in THF/toluene
gave a white powder, which was recrystallized from hot
toluene to give colorless crystals of [Cu2Mes3]ꢀ[MgBr(thf)5]+
(2). These crystals were characterized by X-ray crystallo-
graphy (Figure 2).[13] Complex 2 can be considered most
simply as a SSIP consisting of a magnesium-based cation and
copper aryl anion. The structure of 2 is therefore directly
comparable to the structure in THF of lithium diorganocup-
rates, which also adopt a SSIP arrangement under such
conditions (see II), except with the surprising replacement of
the mononuclear [R2Cu]ꢀ cuprate anion with a dinuclear
[R3Cu2]ꢀ unit.
Figure 3. The molecular structure of 3a, one of the two crystallo-
graphically independent Ci-symmetric complexes present in the crystals
of 3.
Figure 2. The molecular structure of 2.
Angew. Chem. Int. Ed. 2008, 47, 5812 –5815
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5813

Communications
tion of two (MgBr)+ cations and two [Cu2Mes3]ꢀ anions—that
more, the metal centers within these complexes have been
observed to have a unique synergic relationship, which has
resulted in the application of these complexes in some
unparalleled regioselective and/or polydeprotonation reac-
tions.[17] In this context, 1 can be considered as an eight-
membered [Cu2Mg2C4]2+ inverse crown ether and 3 as a 12-
membered [Cu4Mg2C6]2+ inverse crown ether. They constitute
the first examples of inverse crown ethers in which the alkali
metal has been replaced formally by a transition metal (CuI).
In 1, the iodine atoms are too large to fit within the
macrocycle and therefore lie directly outside the Cu2Mg2C4
ring, whereas in 3, the bromine atoms are present as “guest”
anions to give a central Mg2Br2 ring similar to, for example,
the Mg2H2 ring in the hydride-guest-anion complex
[Na2Mg2(H)2{N(iPr)2}4][21] or the Mg2(OnBu)2 ring in the
alkoxo-guest-anion complex [Na2Mg2(OnBu)2{N(iPr)2}4].[22]
In summary, the first structural characterization of
magnesium organocuprates derived from Grignard reagents
shows close parallels between these complexes and lithium
organocuprates. In particular, both types of organocuprates
adopt dimeric CIP structures in Et2O and SSIP structures in
THF. Whereas complex 1 is directly isostructural with its
lithium analogue [({Ph2Cu}{Li(OEt2)})2], complex 3 forms a
unique larger ring system as a result of the increased steric
requirements of the Mes group. Complexes 1 and 3 can also
be considered as rare examples of dicationic transition-metal-
containing inverse crown ethers based upon 8- and 12-
membered macrocycles, respectively.
is, of the same structural units present in the SSIP complex 2.
The structures of 2 and 3 therefore hint at close parallels
between the behavior of lithium organocuprates and
Grignard reagent derived magnesium organocuprates,
whereby CIP structures (I and 3) are favored in weakly
coordinating solvents, such as Et2O, and SSIP structures (II
and 2) are favored in more strongly coordinating solvents,
such as THF.
The Mg centers in 3 are approximately tetrahedral and
form a central four-membered Mg2Br2 ring with the two
ꢀ
bromine atoms, with the Mg Br bond distances in the range
2.5683(9)–2.5735(9) Š (mean distance: 2.5707 Š). The Mg–C
distances are in the range 2.260(2)–2.274(2) Š (mean dis-
ꢀ
tance: 2.268 Š); thus, the Mg C bonds are comparable in
length to those in 1.
Two different types of mesityl groups are present in 3: The
first type bridges two CuI atoms symmetrically to give a 3c–2e
bond with Cu–C distances in the range 1.981(2)–2.002(2) Š
(mean distance: 1.994 Š). These distances are similar to the
Cu–C distances in the anion of 2. However, the Cu-Cipso-Cu
angles (77.56(9), 78.40(9)8) are less acute than the equivalent
angle in 2, thus resulting in a longer Cu···Cu distance
(2.5044(6), 2.5136(5) Š). The second type of mesityl group
bridges a Mg and a Cu center asymmetrically to form shorter
Cu–C contacts (1.955(2)–1.971(2) Š; mean distance:
1.964 Š). The bond angles at the Cu centers in 3 lie in the
range 154.23(9)–158.72(9)8 (mean angle: 156.638) and thus
are less obtuse than the C-Cu-C angles in 2, presumably as a
result of the steric requirements of the central Mg2Br2 ring.
Complex 3 can also be considered as a dimer similar to 1,
but with the [Ph2Cu]ꢀ units in 1 replaced by [Mes3Cu2]ꢀ units.
The difference in the composition of these cuprate anions in 1
and 3 can be explained by the increased steric requirements of
the Mes group over that of the Ph group. Thus, the formation
of a possible dimeric [{(Mes2Cu)(MgBr)}2] structure is
prevented as a result of the steric clash between ortho CH3
groups on the Mes ring and bromine atoms. Lithium
dimesitylcuprate, however, has been shown to adopt a dimeric
[{(Mes2Cu)Li}2] motif, as Li+ is considerably less sterically
demanding than (MgBr)+.[7]
Experimental Section
All reactions were carried out under a protective nitrogen atmo-
sphere. CuIMes was prepared according to a literature procedure.[23]
1: A solution of CuIMes (730 mg, 4 mmol) in toluene (10 mL) was
added to a solution of phenylmagnesium iodide in diethylether
(0.79m, 4.6 mL, 3.64 mmol). The mixture was stirred for 30 min at
room temperature and then filtered through celite. Storage at 48C for
14 days yielded 1 (0.246 g, 31% based on PhMgI) as colorless crystals
suitable for X-ray crystallographic analysis. M.p.: 188–1928C.
2: A solution of CuIMes (365 mg, 2 mmol) in toluene (5 mL) was
added to a solution of mesitylmagnesium bromide in THF (0.50m,
4.0 mL, 2 mmol) to give a white precipitate. The solvent was
evaporated under reduced pressure, and the precipitate was redis-
solved in hot toluene (15 mL). The solution was allowed to cool to
room temperature, whereupon complex 2 (0.352 g, 56%) crystallized
as colorless crystals suitable for X-ray crystallographic analysis. M.p.:
1988C (decomp.).
It seems likely that the formation of 3 is also a key first
step in the reaction in THF. Thus, we propose that an
aggregate similar or identical to 3 is formed initially and then
solvated by THF to give the SSIP 2. This pathway would
explain the observation of [Cu2Mes3]ꢀ anions in 2, rather than
the expected [CuMes2]ꢀ anions.
3: A solution of CuIMes (365 mg, 2 mmol) in toluene (5 mL) was
added to a solution of mesitylmagnesium bromide in diethyl ether
(0.41m, 4.9 mL, 2 mmol) to give a white precipitate. The solvent was
evaporated under reduced pressure, and the precipitate was redis-
solved in a mixture of toluene (15 mL) and ether (2 mL). The solution
was allowed to stand at room temperature for 7 days, during which
time 3 (0.340 g, 40%) crystallized as colorless crystals suitable for X-
ray crystallographic analysis. M.p.: 1908C (decomp.).
It is also important to consider the structures of 1 and 3 in
the light of recent studies on “inverse crown ether” com-
plexes: a special class of heterobimetallic compounds typi-
cally containing dicationic eight-membered ([MI2MII L4]2+) or
2
twelve-membered macrocycles ([MI4MII L6]2+), in which MI is
2
an alkali metal (Li+, Na+, K+), MII is a divalent metal (Mg2+,
Ca2+, Zn2+, Mn2+), and L is an amido, alkyl, or
enolato ligand.[17–20] These macrocycles (see,
for example, III) are named inverse crown
ethers because of their ability to bind organic
or inorganic anions and their topological
similarity to organic crown ethers. Further-
Received: February 11, 2008
Published online: June 23, 2008
Keywords: copper · Grignard reagents ·
.
inverse crown compounds · magnesium · organocuprates
5814
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5812 –5815

Angewandte
Chemie
15.2710(7) Š, b = 90.545(4)8, V= 4772.4(4) Š3, Z = 4, 1calcd
=
1.321 gcmꢀ3, m(CuKa) = 2.526 mmꢀ1, T= 173 K, colorless nee-
dles; Oxford Diffraction Xcalibur PX Ultra diffractometer; 7347
independent measured reflections, F2 refinement, R1 = 0.072,
wR2 = 0.215, 4236 independent observed reflections (j Fo j > 4s-
(jFoj), 2qmax = 1268), 565 parameters.
[1] N. Krause, Modern Organocopper Chemistry, Wiley-VCH,
Weinheim, 2002, and references therein.
[2] R. M. Gschwind, P. R. Rajamohanan, M. John, G. Boche,
Organometallics 2000, 19, 2868 – 2873.
[3] M. John, C. Auel, C. Behrens, M. Marsch, K. Harms, F. Bosold,
R. M. Gschwind, P. R. Rajamohanan, G. Boche, Chem. Eur. J.
2000, 6, 3060 – 3068.
[4] S. I. Khan, P. G. Edwards, H. S. H. Yuan, R. Bau, J. Am. Chem.
Soc. 1985, 107, 1682 – 1684.
[14] C. S. Hwang, P. P. Power, Organometallics 1999, 18, 697 – 700.
[15] M. A. Carvajal, S. Alvarez, J. J. Novoa, Chem. Eur. J. 2004, 10,
2117 – 2132.
[16] Crystal data for 3: C54H66Br2Cu4Mg2·C7H8, M = 1269.80, triclinic,
¯
[5] D. M. Knotter, W. J. J. Smeets, A. L. Spek, G. van Koten, J. Am.
Chem. Soc. 1990, 112, 5895 – 5896.
P1 (no. 2), a = 13.120(2), b = 14.8765(19), c = 15.9145(19) Š, a =
76.343(11), b = 73.593(12), g = 71.735(12)8, V= 2792.2(6) Š3,
Z = 2 (two Ci symmetric molecules), 1calcd = 1.510 gcmꢀ3, m-
(MoKa) = 2.994 mmꢀ1, T= 173 K, colorless blocks, Oxford Dif-
fraction Xcalibur 3 diffractometer; 17745 independent mea-
sured reflections, F2 refinement, R1 = 0.032, wR2 = 0.077, 11727
independent observed absorption-corrected reflections [ j Fo j >
4s(jFoj), 2qmax = 658], 631 parameters.
[6] D. M. Knotter, D. M. Grove, W. J. J. Smeets, A. L. Spek, G.
van Koten, J. Am. Chem. Soc. 1992, 114, 3400 – 3410.
[7] R. P. Davies, S. Hornauer, Chem. Commun. 2007, 304 – 306.
[8] R. P. Davies, S. Hornauer, P. B. Hitchcock, Angew. Chem. 2007,
119, 5283 – 5286; Angew. Chem. Int. Ed. 2007, 46, 5191 – 5194.
[9] Crystal data for 1: C32H40Cu2I2Mg2O2, M = 886.14, monoclinic,
P21/n (no. 14), a = 10.0546(2), b = 17.0931(3), c = 10.3538(2) Š,
[17] R. E. Mulvey, F. Mongin, M. Uchiyama, Y. Kondo, Angew.
Chem. 2007, 119, 3876 – 3899; Angew. Chem. Int. Ed. 2007, 46,
3802 – 3824.
[18] R. E. Mulvey, Organometallics 2006, 25, 1060 – 1075.
[19] L. M. Carrella, W. Clegg, D. V. Graham, L. M. Hogg, A. R.
Kennedy, J. Klett, R. E. Mulvey, E. Rentschler, L. Russo, Angew.
Chem. 2007, 119, 4746 – 4750; Angew. Chem. Int. Ed. 2007, 46,
4662 – 4666.
b = 101.016(2)8, V= 1746.66(6) Š3, Z = 2 (Ci symmetry), 1calcd
=
1.685 gcmꢀ3, m(MoKa) = 3.047 mmꢀ1, T= 173 K, colorless octa-
hedra; Oxford Diffraction Xcalibur 3 diffractometer; 5771
independent measured reflections, F2 refinement, R1 = 0.039,
wR2 = 0.094, 4544 independent observed absorption-corrected
reflections (j Fo j > 4s(jFoj), 2qmax = 658), 216 parameters.
CCDC 676090 (1), 676091 (2), and 676092 (3) 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] N. P. Lorenzen, E. Weiss, Angew. Chem. 1990, 102, 322 – 324;
Angew. Chem. Int. Ed. Engl. 1990, 29, 300 – 302.
[20] A. R. Kennedy, J. Klett, R. E. Mulvey, S. Newton, D. S. Wright,
Chem. Commun. 2008, 308 – 310.
[21] D. J. Gallagher, K. W. Henderson, A. R. Kennedy, C. T. OꢀHara,
R. E. Mulvey, R. B. Rowlings, Chem. Commun. 2002, 376 – 377.
[22] K. J. Drewette, K. W. Henderson, A. R. Kennedy, R. E. Mulvey,
C. T. OꢀHara, R. B. Rowlings, Chem. Commun. 2002, 1176 –
1177.
[11] M. M. Olmstead, P. P. Power, J. Am. Chem. Soc. 1990, 112, 8008 –
8014.
[12] G. van Koten, J. T. B. H. Jastrzebski, F. Muller, C. H. Stam, J.
Am. Chem. Soc. 1985, 107, 697 – 698.
[23] E. M. Meyer, S. Gambarotta, C. Floriani, A. Chiesi-Villa, C.
Guastini, Organometallics 1989, 8, 1067 – 1079.
[13] Crystal data for 2: [C27H33Cu2][C20H40O5MgBr], M = 949.35,
monoclinic, P21/n (no. 14), a = 15.1972(6), b = 20.5649(9), c =
Angew. Chem. Int. Ed. 2008, 47, 5812 –5815
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5815