Copper clusters built on bulky amidinate ligands: Spin delocalization via superexchange rather than direct metal-metal bonding

Article abstract of DOI:10.1039/b412152j

Entry into a new class of tetra- and dicopper clusters was assisted by a fine steric tuning of bulky amidinate ligands that provide spin-delocalizing superexchange pathways in class III mixed-valence dusters, the properties of which are best understood without invoking metal-metal bonding. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b412152j

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Copper clusters built on bulky amidinate ligands: spin delocalization via
superexchange rather than direct metal–metal bonding{
Xuan Jiang, John C. Bollinger, Mu-Hyun Baik* and Dongwhan Lee*
Received (in Cambridge, UK) 6th August 2004, Accepted 17th November 2004
First published as an Advance Article on the web 10th January 2005
DOI: 10.1039/b412152j
related Cu4 cores have previously been obtained using carboxylate
(RCO22, R 5 C6H5; CF3)6 or triazinate (NNR 2, R 5 CH3)7
Entry into a new class of tetra- and dicopper clusters was
assisted by a fine steric tuning of bulky amidinate ligands that
provide spin-delocalizing superexchange pathways in class III
mixed-valence clusters, the properties of which are best under-
stood without invoking metal–metal bonding.
2
ligands, which are isoelectronic with amidinate. Despite the
extensive use of amidinate ligands to support essentially every
metal ion across the periodic table,8 tetracopper clusters of this
robust bidentate ligand were previously unknown.
Valence-delocalized transition metal clusters often participate in
electron transfer (ET) processes in living systems.1 The dinuclear
CuA sites in cytochrome c oxidase and nitrous oxide reductase
represent one such example, in which a rapid switching between
Cu2(I,I) and Cu2(I,II) states accompanies electron shuttling.2
Synthetic analogues of these dicopper sites have previously been
achieved using multiaza chelating ligands,3 conformationally rigid
or sterically demanding carboxylate ligands,4 or a thioether-
In 1, ortho hydrogen atoms on the N-phenyl groups are pointing
toward the phenyl rings of the neighboring ligands. Installing alkyl
groups on the 2,6-positions of the phenyl ring was thus expected to
maximize interligand steric crowding and assist the assembly of
lower nuclearity copper(I) clusters. A bulkier amidinate ligand
N,N9-bis(2,6-dimethylphenyl)benzamidinate was thus deployed to
prepare a dicopper(I) complex [Cu2(m-NMe2Ph2CPh)2] (2) (Fig. 1).{
…
˚
The relatively short Cu Cu separation of 2.4571(2) A in 2 is
5
…
appended amido ligand. The relatively short Cu Cu distances
˚
˚
comparable to those (2.414(1)–2.497(2) A) obtained for other rare
examples of bis(amidinate)dicopper(I) complexes.9
determined for both biological (y 2.5 A) and synthetic (2.36–
˚
2.42 A) systems have often been invoked to rationalize spin
In CH2Cl2, 2 undergoes two consecutive redox processes as
determined by cyclic voltammetry and square wave voltammetry
(Fig. S1).{10 The two quasi-reversible redox waves centered
delocalization via direct Cu–Cu bonding.2 It remains to be
established whether such a direct metal–metal interaction is a
prerequisite for spin delocalization in copper dimers that lack
single-atom bridging ligands. In this communication, we disclose a
synthetic strategy that allowed access to discrete tetra- and
dicopper amidinate complexes. A copper(I) dimer having two-
coordinate metal centers was obtained, which undergoes two
consecutive one-electron redox processes in solution. One-electron
chemical oxidation of this compound afforded class III
…
dicopper(I,II) complexes. Despite a short Cu Cu distance, spin
delocalization in this dicopper(I,II) core is dominated by super-
exchange through m-1,3 bridging ligands.
Anion metathesis of a simple copper(I) salt with N,N-
disubstituted amidinates in MeCN cleanly afforded neutral
compounds that precipitated from the reaction mixture in isolated
yields exceeding 85% (Scheme 1). Highly simplified 1H- and
13C-NMR resonances displayed by these products indicated a
symmetric ligand environment. The solid-state structure of the
tetracopper(I) complex [Cu4(m-NPh2CPh)4] (1) was determined by
X-ray crystallography on yellow crystals obtained by vapor
diffusion of pentanes into a saturated CH2Cl2 solution of this
material.{ As shown in Fig. 1, the core structure 1 features a
Scheme 1 Synthesis of copper amidinate complexes.
…
pseudo-rhombic Cu4 center, in which Cu Cu distances of
˚
2.5674(6)–2.6419(7) A are spanned by m-1,3-bridging amidinates
disposed alternately above and below the Cu4 plane. Topologically
{ Electronic supplementary information (ESI) available: details of the
synthetic procedures, crystallographic data, and DFT calculations. See
http://www.rsc.org/suppdata/cc/b4/b412152j/
*mbaik@indiana.edu (Mu-Hyun Baik)
dongwhan@indiana.edu (Dongwhan Lee)
Fig. 1 ORTEP diagrams of 1 (left) and 2 (right) with thermal ellipsoids
at 50% probability.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 1043–1045 | 1043

View Article Online
at 0.31 and 0.79 V were assigned to Cu2(I,I)/Cu2(I,II) and Cu2(I,II)/
Cu2(II,II) redox couples, respectively. High-level DFT calculations
coupled to a continuum solvation model11 support this assign-
ment. Our calculations place the two oxidation potentials at 0.461
and 0.986 V, thus overestimating both metal-based redox
potentials by 151 and 196 mV, respectively.§ The computed
potential difference between the two redox events of 525 mV,
however, is in excellent agreement with the experimental value of
480 mV." A large comproportionation constant of Kcom 5 1.3 6
108 for the equilibrium between Cu(I)Cu(I) + Cu(II)Cu(II) and
Cu(I)Cu(II) species12 indicates
a significant thermodynamic
stability of the electrochemically generated Cu2(I,II) complex,
prompting us to explore synthetic routes to access them.
Fig. 3 X-band EPR spectra (solid lines) of a frozen CH2Cl2–toluene (1 : 1,
v/v) sample of 3a measured at 77 K. Simulations (dashed lines) are
calculated for gx 5 2.022, gy 5 2.132, gz 5 2.199, Ax 5 19.4 G, Ay 5 20.9 G,
Az 5 73.7 G.
Initial attempts to oxidize 2 using various Ag(I) reagents did not
result in any isolable products. A dark purple–blue species
generated in CH2Cl2 rapidly decayed to yellow–brown product(s),
and insoluble materials deposited over time. The stability of this
putative Cu2(I,II) species, however, could be enhanced by conduct-
ing the reaction in the presence of coordinating solvents such as
MeCN or THF at low temperatures (Scheme 1). Analytically pure,
purple–blue (lmax y 680 nm; e . 2000 M21 cm21 per dimer in
CH2Cl2) crystals of 3a (73% isolated yield) and 3b (49% isolated
yield) were obtained at 235 uC and characterized by X-ray
crystallography.{ As shown in Fig. 2, the 3a cation reveals two
metal centers related by crystallographic inversion symmetry,
implicating the valence-delocalized nature of the dicopper(I,II)
calculations are also consistent with the valence-delocalized nature
of 2+, the one-electron oxidation product of 2. Mulliken spin
populations of 0.23 and 0.22 at each of the copper centers in 2+
(Fig. 4) indicate that roughly half of the ‘‘extra’’ electron is
distributed equally among the two copper centers whereas the
ligand environment absorbs the other half of the spin.{
Interestingly, almost 30% of the unpaired spin is delocalized over
the p-space of the N-aryl groups, whereas the four nitrogen atoms
accommodate a total of approximately 20% of the unpaired
electron. Remarkably, the amidinate carbon atoms and the
appended phenyl groups play no role in the spin dissipation
process.
Electronically, the two Cu(I)-d10 centers in the neutral dimer give
rise to 10 metal-dominated frontier orbitals that are all filled, thus
disqualifying any M–M bonding interaction. The removal of
electrons from the high lying MO that is M–M antibonding allows
in principle for the formation of a single bond. However, a crucial
requirement for oxidation having such impact is that there is
…
˚
core. A short Cu Cu separation of 2.4547(13) A in 3a is spanned
by two m-1,3 amidinate ligands and the remaining coordination site
on each metal center is occupied by an MeCN ligand. An
…
analogous THF complex 3b has comparable Cu Cu distances of
˚
2.4423(12) and 2.3974(11) A, as determined for the two
independent molecules in the asymmetric unit (Fig. S2). Cyclic
voltammetry of 3a displays a quasi-reversible redox wave centered
at E1/2 (Cu2(I,II)/Cu2(I,I)) 5 0.29 V.
The valence-delocalized nature of the dicopper(I,II) center was
further confirmed by EPR spectroscopic studies. The seven-line
splitting patterns observed at 77 K for a frozen solution sample of
3a (Fig. 3) is characteristic of a spin interacting with two copper
ions (I 5 3/2). This rhombic signal retains its structure at 4.5 K,
at which point features arising from superhyperfine coupling to
the nitrogen nuclei become more discernible (Fig. S3). DFT
Fig. 2 ORTEP diagram of 3a cation with thermal ellipsoids at 50%
probability (left) and isosurface plot (0.05 au) of the redox-active MO (the
b-LUMO) of 2+ (right).
Fig. 4 Mulliken spin density distribution in 2+. Regions of the molecule
that participate notably in delocalizing the unpaired electron are
highlighted in bold.
1044 | Chem. Commun., 2005, 1043–1045
This journal is ß The Royal Society of Chemistry 2005

View Article Online
reflections, 8469 unique (Rint 5 0.0939), R1 5 0.0448 for I . 2s(I),
wR2 5 0.0954. 3b?CH2Cl2: C56H66Cl2Cu2F6N4O2Sb, M 5 1260.91,
sufficient overlap of metal d-orbitals along the M–M vector. Fig. 2
shows the redox-active MO, the b-LUMO of 2+. It is mainly
composed of the Cu dz2 orbital with some mixing of the dx2–y2
orbital, where the z-axis is aligned with the Cu–Cu vector. This
MO is best characterized as non-bonding or slightly anti-bonding
˚
triclinic, a 5 12.562(3), b 5 21.015(4), c 5 21.138(5) A, a 5 90.899(6),
3
˚
b 5 104.009(8), c 5 91.598(7)u, V 5 5411(2) A , T 5 123(2) K, space group
¯
˚
P1, Z 5 4, l(Mo-Ka) 5 0.71073 A, 78414 reflections, 33033 unique
(Rint 5 0.087), R1 5 0.0582 for I . 2s(I), wR2 5 0.1254. CCDC 247397–
247400. See http://www.rsc.org/suppdata/cc/b4/b412152j/ for crystallo-
graphic data in .cif or other electronic format.
…
with respect to both the Cu Cu centers and along the Cu–N
§ A difference of 196 mV corresponds to 4.5 kcal mol21 in total free energy.
Considering the usual experimental uncertainties and the approximations
used to compute free energies in solution, the agreement between theory
and experiment is remarkable.
bonds. As a consequence, the first oxidation leads to only an
˚
insignificant decrease of the Cu–Cu distance by 0.081 A and a
˚
similarly small average contraction of the Cu–N bonds by 0.031 A.
The non-bonding nature of the redox active MO was further
confirmed by the Mayer bond-order analysis.13 Upon removal of
one electron from 2, the Cu–Cu bond order increases by a
negligibly small amount of 0.05 from 0.22 to 0.27, whereas the Cu–
N bond orders of 0.50 each remain constant.14 The bond order
invariance is maintained for the removal of the second electron
giving a formal Cu–Cu bond order of 0.25 for 22+. We note that
the removal of the second electron leads to a formal decrease of the
Cu–Cu bond order by 0.02, highlighting that bond order changes
of this magnitude are physically meaningless and should be
interpreted as invariant. The negligible structural change upon
metal oxidation promises low inner-sphere energy barriers for ET,
an important requirement for future technical exploitations of this
redox system.
" Owing to systematic error cancellations, we expect the potential
differences (DE) to be more accurate than the actual potentials.
1 E. I. Solomon, D. W. Randall and T. Glaser, Coord. Chem. Rev., 2000,
200–202, 595–632.
2 Y. Lu, in Comprehensive Coordination Chemistry II, eds. L. Que, Jr. and
W. B. Tolman, Elsevier, Oxford, UK, 2003, vol. 8, pp. 91–122 and
references cited therein.
3 (a) C. Harding, V. McKee and J. Nelson, J. Am. Chem. Soc., 1991, 113,
9684–9685; (b) M. E. Barr, P. H. Smith, W. E. Antholine and
B. Spencer, J. Chem. Soc., Chem. Commun., 1993, 1649–1652; (c)
C. Harding, J. Nelson, M. C. R. Symons and J. Wyatt, J. Chem. Soc.,
Chem. Commun., 1994, 2499–2500; (d) R. P. Houser, V. G. Young, Jr.
and W. B. Tolman, J. Am. Chem. Soc., 1996, 118, 2101–2102; (e) C. He
and S. J. Lippard, Inorg. Chem., 2000, 39, 5225–5231; (f) R. Gupta,
Z. H. Zhang, D. Powell, M. P. Hendrich and A. S. Borovik, Inorg.
Chem., 2002, 41, 5100–5106.
4 (a) D. D. LeCloux, R. Davydov and S. J. Lippard, J. Am. Chem. Soc.,
1998, 120, 6810–6811; (b) J. R. Hagadorn, T. I. Zahn, L. Que, Jr. and
W. B. Tolman, Dalton Trans., 2003, 1790–1794.
5 S. B. Harkins and J. C. Peters, J. Am. Chem. Soc., 2004, 126, 2885–2893.
6 (a) M. G. B. Drew, D. A. Edwards and R. Richards, J. Chem. Soc.,
Dalton Trans., 1977, 299–303; (b) P. F. Rodesiler and E. L. Amma,
J. Chem. Soc., Chem. Commun., 1974, 599–600.
7 (a) F. E. Brinckman, H. S. Haiss and R. A. Robb, Inorg. Chem., 1965,
4, 936–942; (b) J. E. O’Connor, G. A. Janusonis and E. R. Corey, Chem.
Commun., 1968, 445–446.
8 (a) J. Barker and M. Kilner, Coord. Chem. Rev., 1994, 133, 219–300; (b)
F. T. Edelmann, D. M. M. Freckmann and H. Schumann, Chem. Rev.,
2002, 102, 1851–1896 and references cited therein.
In summary, through the steric modification of the modular
ArCNAr9 2 platform, we devised unique synthetic strategies to
2
access tetra- and dicopper clusters from a homologous ligand set.
The latter compound undergoes one-electron oxidation to afford
well-characterized class III dicopper(I,II) complexes, in which spin-
delocalization is mediated by superexchange rather than direct
metal–metal bonding. We thank the National Science Foundation
(0116050 to Indiana University) and the Indiana University for
funding.
Xuan Jiang, John C. Bollinger, Mu-Hyun Baik* and Dongwhan Lee*
Department of Chemistry and School of Informatics, Indiana University,
Bloomington, IN 47405, USA. E-mail: mbaik@indiana.edu;
9 (a) S. Maier, W. Hiller, J. Stra¨hle, C. Ergezinger and K. Dehnicke,
Z. Naturforsch., Teil B, 1988, 43, 1628–1632; (b) F. A. Cotton, X. Feng,
M. Matusz and R. Poli, J. Am. Chem. Soc., 1988, 110, 7077–7083; (c)
B. S. Lim, A. Rahtu, J.-S. Park and R. G. Gordon, Inorg. Chem., 2003,
42, 7951–7958.
dongwhan@indiana.edu; Fax: 812 855 8300; Tel: 812 855 9364
10 All potentials referenced to Fc/Fc+.
Notes and references
11 (a) M.-H. Baik and R. A. Friesner, J. Phys. Chem. A, 2002, 106,
7407–7412; (b) M.-H. Baik, C. K. Schauer and T. Ziegler, J. Am. Chem.
Soc., 2002, 124, 11167–11181; (c) M.-H. Baik, J. S. Silverman, I. V. Yang,
P. A. Ropp, V. A. Szalai, W. Yang and H. H. Thorp, J. Phys. Chem. B,
2001, 105, 6437–6444.
{ Crystal data. 1?0.5CH2Cl2?0.5C5H12: C79H71ClCu4N4, M 5 1366.11,
˚
triclinic, a 5 13.666(3), b 5 14.295(2), c 5 19.401(5) A, a 5 83.271(12),
3
˚
b 5 74.874(14), c 5 74.361(12)u, V 5 3518.8(14) A , T 5 112(2) K, space
¯
˚
group P1, Z 5 1, l(Mo-Ka) 5 0.71073 A, 97467 reflections, 20526 unique
12 R. R. Gagne´, C. A. Koval, T. J. Smith and M. C. Cimolino, J. Am.
Chem. Soc., 1979, 101, 4571–4580.
(Rint 5 0.0595), R1 5 0.0437 for I . 2s(I), wR2 5 0.1172. 2: C46H46Cu2N4,
˚
M 5 781.95, monoclinic, a 5 10.9223(8), b 5 14.8184(11), c 5 11.8480(9) A,
3
˚
13 I. Mayer, Int. J. Quantum Chem., 1986, 29, 477–483.
14 A computed M–M bond order that is lower than 0.3 indicates that there
is no bond present. For an example of a dimer with a typical M–M
bond that changes notably as a function of the oxidation state, see:
M.-H. Baik, R. A. Friesner and G. Parkin, Polyhedron, 2004, 23,
2879–2900.
b 5 91.315(2)u, V 5 1917.1(2) A , T 5 119(2) K, space group P21/n, Z 5 2,
˚
l(Mo-Ka) 5 0.71073 A, 75170 reflections, 11896 unique (Rint 5 0.0341),
R1
C52H56Cl2Cu2F6N6Sb, M 5 1198.81, monoclinic, a 5 22.385(4),
5 0.0269 for I . 2s(I), wR2 5 0.0737. 3a?C2H4Cl2:
3
˚
˚
b 5 10.4127(18), c 5 22.665(4) A, b 5 103.516(4)u, V 5 5136.6(16) A ,
˚
T 5 115(2) K, space group I2/a, Z 5 4, l(Mo-Ka) 5 0.71073 A, 29600
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 1043–1045 | 1045