Synthesis, structural features, absorption spectra, redox behaviour and luminescence Properties of ruthenium(II) rack-type dinuclear complexes with ditopic, hydrazone-based ligands

Article abstract of DOI:10.1002/chem.200401261

The isomeric bis(tridentate) hydrazone ligand strands 1a-c react with [Ru(terpy)Cl₃] (terpy = 2,2':6'.2"-terpyridine) to give dinuclear rack-type compounds 2a-c, which were characterised by several techniques, including X-ray crystallography and NMR methods. The absorption spectra, redox behaviour and luminescence properties (both in fluid solution at room temperature and in rigid matrix at 77 K) of the ligand strands 1a-c and of the metal complexes 2a-c have been studied. Compounds 1a-c exhibit absorption spectra dominated by intense π-π* bands, which, in the case of 1b and 1c, extend within the visible region, while the absorption spectra of the rack-type complexes 2a-c show intense bands both the in the UV region, due to spin-allowed ligand-centred (LC) transitions, and in the visible, due to spin-allowed metal-to-ligand charge-transfer (MLCT) transitions. The energy position of these bands strongly depends on the ligand strand: in the case of 2a. the lowest energy MLCT band is around 470 nm, while in 2b and 2c, it lies beyond 600 nm. Ligands 1a-c undergo oxidation processes that involve orbitals based mainly on the CH₃-N-N= fragments. The complexes 2a-c undergo reversible metal-centred oxidalion, while reductions involve the hydrazone-based ligands: in 2b and 2c, the bridging ligand is reduced twice and in 2a once before reduction of the peripheral terpy ligands takes place. Ligands 1a-c exhibit luminescence from the lowest-lying ¹π-π* level. Only for complex 2a does emission occur; this may be attributed to a ³MLCT state involving the bridging ligand. Taken together, the results clearly indicate that the structural variations introduced translate into interesting differences in the spectroscopic, luminescence and redox properties of the ligand strands as well as of the rack-type metal complexes.

Full text of DOI:10.1002/chem.200401261

FULL PAPER
Synthesis, Structural Features, Absorption Spectra, Redox Behaviour and
Luminescence Properties of Ruthenium(ii) Rack-Type Dinuclear Complexes
with Ditopic, Hydrazone-Based Ligands
Adrian-Mihail Stadler,[a] Fausto Puntoriero,[b] Sebastiano Campagna,*[b]
Nathalie Kyritsakas,[c] Richard Welter,[d] and Jean-Marie Lehn*[a]
Abstract: The isomeric bis(tridentate)
hydrazone ligand strands 1a–c react
with [Ru(terpy)Cl3] (terpy=2,2’:6’,2’’-
terpyridine) to give dinuclear rack-type
compounds 2a–c, which were charac-
terised by several techniques, including
X-ray crystallography and NMR meth-
ods. The absorption spectra, redox be-
haviour and luminescence properties
(both in fluid solution at room temper-
ature and in rigid matrix at 77 K) of
the ligand strands 1a–c and of the
metal complexes 2a–c have been stud-
ied. Compounds 1a–c exhibit absorp-
tion spectra dominated by intense p–
p* bands, which, in the case of 1b and
1c, extend within the visible region,
while the absorption spectra of the
rack-type complexes 2a–c show intense
bands both the in the UV region, due
to spin-allowed ligand-centred (LC)
transitions, and in the visible, due to
spin-allowed metal-to-ligand charge-
transfer (MLCT) transitions. The
energy position of these bands strongly
depends on the ligand strand: in the
case of 2a, the lowest energy MLCT
band is around 470 nm, while in 2b
and 2c, it lies beyond 600 nm. Ligands
1a–c undergo oxidation processes that
involve orbitals based mainly on the
oxidation, while reductions involve the
hydrazone-based ligands: in 2b and 2c,
the bridging ligand is reduced twice
and in 2a once before reduction of the
peripheral terpy ligands takes place. Li-
gands 1a–c exhibit luminescence from
the lowest-lying 1p–p* level. Only for
complex 2a does emission occur; this
3
may be attributed to a MLCT state in-
volving the bridging ligand. Taken to-
gether, the results clearly indicate that
the structural variations introduced
translate into interesting differences in
the spectroscopic, luminescence and
redox properties of the ligand strands
as well as of the rack-type metal com-
plexes.
ꢀ ꢀ
CH3 N N= fragments. The complexes
2a–c undergo reversible metal-centred
Keywords: luminescence
· N li-
gands · redox chemistry · rutheni-
um · supramolecular chemistry
[a] A.-M. Stadler, Prof. Dr. J.-M. Lehn
Laboratoire de Chimie Supramolꢀculaire, ISIS
Universitꢀ Louis Pasteur, 8 Allꢀe Gaspard Monge
BP 70028, 67083 Strasbourg Cedex (France)
Fax : (+33)390-245-140
Introduction
Coordinative polynuclear heteroleptic complexes of ruthe-
nium(ii) with 2,2’:6’,2’’-terpyridine (terpy) and polytopic het-
erocyclic ligand strands are the focus of interest for both
their photophysical and electrochemical properties,[1,2] and
as precursors for topological stereoisomers.[3] Such reasons
have indeed stimulated large efforts to synthesize appropri-
ate bis(tridentate) ligands[4] that are able to generate such
metallosupramolecular architectures.[5]
E-mail: lehn@isis.u-strasbg.fr
[b] Dr. F. Puntoriero, Prof. Dr. S. Campagna
Dipartimento di Chimica Inorganica
Chimica Analitica e Chimica Fisica, Universitꢁ di Messina
Villaggio S. Agata, 96166 Messina (Italy)
Fax : (+39)90-393756
E-mail: photochem@chem.unime.it
In our laboratories, we are interested in developing facile
synthetic pathways for analogues of helical ligands py1-(py2-
pym)n-py2-py1 (py1 =2-substituted pyridine, py2 =2,6-disub-
stituted pyridine, pym=4,6-disubstituted pyrimidine), by re-
placing the 2,6-disubstituted pyridine ring by its isomorphic
equivalent, a hydrazone (hyz) group.[6,7] This synthetic strat-
egy has also the advantage of increasing the molecular di-
[c] N. Kyritsakas
Service Commun de Rayons X, Institut Le Bel
Universitꢀ Louis Pasteur, 4 Rue Blaise Pascal
67000 Strasbourg (France)
[d] Prof. Dr. R. Welter
Laboratoire de Densitꢀ Electronique et Composꢀs Mꢀtalliques
Institut Le Bel, Universitꢀ Louis Pasteur
4 Rue Blaise Pascal, 67000 Strasbourg (France)
Chem. Eur. J. 2005, 11, 3997 – 4009
DOI: 10.1002/chem.200401261
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3997

versity. For a ligand strand with two tridentate terpy-like
sites (py1-py2-pym-py2-py1), replacing py2 by a hydrazone
group generates two types of two-site symmetric isomeric li-
gands, because the central unit may be derived from a bis-
(hydrazino)pyrimidine or from a pyrimidine-dicarboxalde-
hyde. This isomerism means the change of the position of
the C=N bond. Thus, one gains synthetic efficiency and sim-
plicity, but also access to diverse molecular species, that is,
to two possibilities for py1-hyz-pym-hyz-py1 (1a and 1b), in-
stead of one for py1-py2-pym-py2-py1 (Scheme 1). The same
synthetic strategy may be applied for other two-site ligands,
such as the py1-py2-pz-py2-py1 (pz=2,5-disubstituted pyra-
zine) sequence,[8] the hydrazone analogue of which (1c) we
report herein.
The ligands 1a–c react with [Ru(terpy)Cl3] in protic sol-
vents, by heating to reflux, and give the corresponding dinu-
clear ruthenium(ii) rack-type complexes, thus showing a re-
markable stability of the C=N bond from the hydrazone
groups. It could be expected that the spectroscopic, photo-
physical and electrochemical properties of the isomeric com-
plexes (derived from these isomeric ligands) be different,
due to changes of the C=N bond positions. Here we report
the synthesis of the three ligands (1a–c) of such a class of
molecule and of the three corresponding dinuclear rack-type
ruthenium(ii) complexes (2a–c) derived from these species.
All the compounds have been characterized by several tech-
niques, including various NMR methods and mass spectrom-
etry. The rack-type complexes have also been investigated
by X-ray analysis. Absorption spectra, redox behaviour and
luminescence properties of the new metal complexes, as
well as of the free ligands, have been studied. The results in-
dicate that the molecular diversity introduced by the isomer-
ism translates into interesting differences as far as the spec-
troscopic, luminescence and redox properties of the new
compounds are concerned.
Results and Discussion
Ligands synthesis: Ligands 1a–c were synthesized
(Scheme 2) by a double chain-extension method consisting
of the condensation of 4,6-bis(1-methylhydrazino)-pyrimi-
dine with 2-pyridinecarboxaldehyde (in case of ligand 1a[7])
or of a dialdehyde (4 or 5 with 2-(1-methylhydrazino)-pyri-
dine (in case of ligands 1c[7] and 1b, respectively). Strands
1a and 1b have a C2 symmetry axis, while strand 1c has an
inversion centre as a symmetry element.
Reaction of 2,5-dimethylpyrazine with benzaldehyde, in
presence of benzoic anhydride, gave 2,5-distyrylpyrazine[9]
(6), the ozonolysis (O3, MeOH, ꢀ788C) of which, followed
by reduction with aqueous solution of sodium metabisulfite
(Na2S2O5), produced 2,5-pyrazinedicarboxaldehyde[10] (4).
Condensation of dialdehyde 4 with 2-(1-methylhydrazino)-
pyridine[11] (3) in EtOH, at reflux, gave the ligand 1c.
In ligands 1a–c, the hydrazone group is the isomorphic
equivalent[6,7] of a 2,6-disubstituted pyridine ring, so these li-
gands may be considered as analogues of ditopic ligands
with two tridentate terpy-like sites. They have coordinative
capacity to form heteroleptic ruthenium(ii) rack-type com-
plexes, playing the same role as the two-site ligand com-
posed of py–pym[1a–d] or py–pz[8] sequences.
Abstract in French: La rꢀaction des ligands bidentates isomꢁ-
res de type bishydrazone 1a–c avec [Ru(terpy)Cl3], conduit
aux complexes dinuclꢀaires 2a–c. Les ꢀtudes radiocristallo-
graphiques par diffractionole rayons X montrent que les
complexes 2a,b contenant le noyau pyrimidine 4,6-disubstituꢀ
comme unitꢀ centrale prꢀsentent une courbure, alors que le
complexe 2c, qui a pour unitꢀ centrale la pyrazine 2,5-disubs-
tituꢀe, est ꢂ peu prꢁs linꢀaire. Quand lꢃunitꢀ centrale est dꢀri-
vꢀe dꢃun dialdꢀhyde, les complexes sont verts (2b,c; absorp-
tion ꢂ 600 nm), alors que lorsquꢃelle est dꢀrivꢀe dꢃune bishy-
drazine, le complexe correspondant 2a est brun-rouge (ab-
sorption ꢂ 470 nm). Ce dernier est le seul complexe ꢂ prꢀsen-
ter une ꢀmission, alors que les ligands 1a–c sont tous
luminescents. Dꢃun point de vue ꢀlectrochimique, les ligands
1a–c participent ꢂ des processus irrꢀversibles dꢃoxydation.
Les complexes 2a–c subissent des oxydations rꢀversibles ou
quasi-rꢀversibles et prꢀsentent plusieurs rꢀductions rꢀversibles
dans le domaine de potentiel de ꢀ2.00 ꢂ +2.00 V/ECS. Ces
rꢀsultats mettent en lumiꢁre les modifications des propriꢀtꢀs
structurales et physico-chimiques produites dans les comple-
xes mꢀtalliques par le remplacement dꢃun noyau pyrimidine
(2a,b) par un noyau pyrazine (2c) ou par le changement de
position de la fonction hydrazone (2a et 2b,c). Ils permettent
dꢃenvisager la synthꢁse de complexes de plus grande taille ꢂ
sites multiples, prꢀsentant une courbure/linꢀaritꢀ et des pro-
priꢀtꢀs contrꢄlables par la nature du ligand.
Complex synthesis: Rack-type complexes of ruthenium(ii)
may be formed by octahedral coordination of a RuII cation
with three nitrogen atoms of a terpy ligand and three of a
hydrazone site, so that a dihydrazone ligand could generate
a dinuclear complex.
The complexes were prepared by following a pathway
that has been described for the synthesis of racks containing
py–pym ligands[1a–d] (Scheme 2). [Ru(terpy)Cl3][12] was treat-
ed with ligands 1a–c, in molar ratio 2.2:1, at reflux, in a mix-
ture of protic solvents that have reducing properties (donors
of electrons for the process RuIII +eꢀ!RuII). The used mix-
ture was composed of 1:1 v/v water and an alcohol (ethanol
or isopropanol), in which the reactants were heated at reflux
during 18–21 h. The results showed that the hydrazone
bonds were stable under these conditions, as complexation
was achieved with yields comparable to those of the reac-
tions involving the py–pym based ligands.[1a,d]
The obtained complexes have Clꢀ as a counterion and are
water soluble. They were precipitated from water by addi-
ꢀ
tion of PF6 as aqueous solution of ammonium hexafluoro-
phosphate (NH4PF6). Further purification was performed by
recrystallization from water and/or reprecipitation from
CH3CN with Et2O or chloroform.
3998
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
Scheme 1. The replacement of a 2,6-disubstituted pyridine (py2) by a hydrazone (hyz) group provides efficient access to analogue isomeric strands with
similar coordinative properties.
Scheme 2. Synthesis of the ligands 1a–c and of the complexes 2a–c. The notations for the protons are also indicated.
The complexes are coloured, solid materials (2a is brown-
red, 2b and 2c are deep green), soluble in acetonitrile, ace-
tone and nitromethane, but insoluble in toluene, diethyl
ether, diisopropyl ether, benzene and chloroform, a fact that
was used in crystallization experiments to obtain single crys-
tals. On heating, they decompose at temperatures greater
than 3008C.
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3999

J.-M. Lehn, S. Campagna et al.
Crystallographic studies: Single crystals suitable for X-ray
diffraction were obtained by slow diffusion of Et2O into a
solution of complexes 2a or 2b in CH3CN/CH3OH (9:1)
and of iPr2O in a solution of complex 2c in CH3CN. Crystals
of 2a are triclinic, crystals of 2b are monoclinic and those of
2c are triclinic. Complex 2a crystallizes as a dimer with four
molecules of acetonitrile, one molecule of methanol and
two molecules of water (2C48H40N14Ru2P4F24·4CH3CN·
CH3OH·2H2O), the crystal cell containing two such units.
Complex 2b cocrystallizes with three molecules of acetoni-
trile (C48H40N14Ru2P4F24·3CH3CN) and the crystal cell con-
tains four such units. Complex 2c has an inversion centre
and cocrystallizes with six molecules of acetonitrile, the crys-
tallographic unit is (C48H40N14Ru2·4PF6·6CH3CN)/2 and the
crystal cell contains two such units.
ferent terpy moieties is 8.22 ꢃ (Figure 1a). The cavity gener-
ated by the two terpy ligands is empty in the case of 2b. In
the case of 2a, one unit of the dimer has a water molecule
between two of the terminal pyridine rings of this cavity,
whereas in the other unit the cavity is empty.
The analysis of the pseudo-octahedral coordinative envi-
ronment of the ruthenium(ii) ions from the X-ray structures
ꢀ
of complexes 2a–c provided the Ru N bond lengths com-
piled in Table 1 (for the notations see also Figure 1b; for
complex 2a crystallized as a dimer, the numbering of Ru
and N coordinating atoms is shown on a monomer; for both
complexes 2a and 2b, the numbering of Ru atoms is the
same as in the corresponding crystallographic information
ꢀ
files (CIF)). The average Ru N length in the complexes is
2.038 ꢃ. In the tridentate sites the distance between the Ru
atom and the middle N atom is shorter (average value
All the measured distances are internuclear (centroid-to-
ꢀ
ꢀ
centroid) distances. The Ru Ru distance is 6.14 ꢃ for 2a
1.980 ꢃ) than the two other Ru N distances (average value
(average distance), 6.13 ꢃ for 2b and 6.89 ꢃ for 2c. For the
pyrimidine-based ligands, the coordinated two-site ligand
strand is not perfectly linear but curved, and its length is
17.66 ꢃ for 2a (average values: calculated curve radius r=
12.61 ꢃ; curvature k=1/r=7.93·10ꢀ2 ꢃꢀ1; arc central angle:
898) and 17.64 ꢃ for 2b (calculated curve radius r=12.84 ꢃ;
curvature k=7.79·10ꢀ2 ꢃꢀ1; arc central angle: 86.58), which
shows that the hydrazone two-site ligands have almost iden-
tical dimensions in these two racks (Figure 1a). These two
ruthenium(ii) racks are shorter and more curved than the
corresponding lead(ii) racks (average length 18.20 ꢃ[13]).
This results from pinching at the coordinated centres due to
the shorter binding distances of RuII in comparison with
lead(ii). Such pinching, resulting in curved structures, has
been documented in related complexes.[1b] With the pyra-
zine-based ligand, the complex (2c) is globally linear (Fig-
ure 1a) due to the compensation of the coordinative con-
traction of the two py1-hyz-pz tridentate sites thanks to the
inversion centre; its length is 17.31 ꢃ (here also less than
the corresponding lead(ii) complex, 18.10 ꢃ[13]).
2.067 ꢃ). The average distance between the middle Nhyz
atom and Ru (1.974 ꢃ) is comparable with the correspond-
ing one from the terpy site (1.986 ꢃ), thus confirming the
analogous coordinative properties of the two N atoms.
The average value of the N-Ru-N chelating angles is
78.98, that is, 10.18 less than the ideal angle for an octahedral
coordination (908). The values corresponding to each chelat-
ing angle are listed in Table 2.
1
NMR spectroscopy: The H NMR spectra of 1a–c are given
in Figure 2. The pyridine protons (marked A–D) and the
methyl group (E) from the ligand 1c give rise to signals that
are similar to the corresponding protons of ligand 1b; this is
because the terminal pyridine units in both ligands 1b and
1c are identical. The two protons of the pyrazine ring are
equivalent, while the protons of pyrimidine ring are not. Al-
though the pyrazine proton (G) it is not located on a carbon
atom situated between two N atoms, as the C2 proton from
the pyrimidine ring (proton H) is, its chemical shift has a
value (9.16 ppm) close to that of the pyrimidine C2 proton
(9.08 ppm). This could be explained by the strong deshield-
ing effect due to the proximity to an N atom and to the two
In both 2a and 2b, the two terpy ligands are crystallo-
graphically unequivalent and the observation of identical
NMR signals are due to fast NMR timescale motions in so-
lution, which make these terpy units equivalent. In complex
2c, the two terpy moieties are equivalent. The length of co-
ordinated terpy is 11.15 ꢃ. In 2a and 2b, the planes of terpy
ligands are not parallel and make an angle of about 328
(31.48 for 2a, average value, and 32.98 for 2b) in the longi-
tudinal section that contains the plane of the two-site ligand,
neither are they perpendicular to the plane of the two-site
ligand. In contrast in complex 2c, the terpy planes are paral-
lel and almost perpendicular to the plane of pyrazine-based
ligand. For the pyrimidine derivatives, the shortest distance
between the terpy planes is about 3 ꢃ (3.08 ꢃ for 2a, aver-
age value, and 2.95 ꢃ for 2b), but the other distances
(mainly larger than 4 ꢃ), as well as the nonparallel orienta-
tion of the two molecular planes in 2a and 2b, suggest that
there is no p-stacking between the aromatic rings of these
two molecules. In 2c, the distance between terpy planes is
6.70 ꢃ and the distance between the K protons of two dif-
ꢀ
ꢀ
para-conjugated C=N groups located on the pyrazine
ring.
The 1H NMR spectra of complexes 2a–c are relatively
complicated (Figure 2), due to close signals in the aromatic
region, and two-dimensional experiments (COSY, NOESY,
ROESY) were useful for the peak assignment. Some partic-
ularities should be noted.
All the complexes, as well as the two-site ligands, are “lat-
eral” and “alternate” structural isomers. Complexes 2a,b
have C2 symmetry, and 14 groups of NMR-equivalent pro-
tons; the complex 2c has an inversion centre (from the
ligand 1c) and it displays only 13 groups of NMR-equivalent
protons.
The presence of the two terpy ligands (with the magnetic
anisotropy of the aromatic rings), as well as the metal-ion
coordination induce shielding (dcomplexꢀdfreeligand =Dd<0) or
deshielding (Dd>0) effects. Thus, for complexes 2a and 2b,
the zone between the two terpy moieties should be the most
4000
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
ꢀ
Figure 1. a) Representation of the X-ray molecular structures of complexes 2a, 2b and 2c (PF6 ions and solvent molecules are omitted for clarity); r is
the curve radius and k is the curvature (k=1/r). b) Notations for coordinating N atoms and Ru atoms in complexes 2a–c.
shielded from the molecule, a fact confirmed by the ex-
tremely low chemical shift of the C2 pyrimidine proton
(proton H; see Figure 2 and Table 3). On the other hand, in
the complex 2c, due to the inversion centre of the ligand 1c,
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4001

J.-M. Lehn, S. Campagna et al.
[a]
ꢀ
Table 1. Ru N bond lengths [internuclear distances, ꢃ] from the X-ray structures of complexes 2a–c
.
the two terpy units are on op-
posite sides of the rack and this
makes their shielding effect on
proton G weaker than in the
case of proton H from 2a and
2b (for the Dd value, see
Table 3).
The shielding of the C5
proton of the pyrimidine ring
(proton G) is also an effect of
the coordination process, al-
ready observed in the case of
complexes of 1a and 1b with
lead(ii).[13] The chemical shift
variation is important (for the
Dd values, see Table 3). This is
due to the presence of the terpy
units (as for the protons D), but
also to a conformational change
of the ligand by coordination.
In the case of 1a, the coordina-
tion causes rotation around the
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Ru1 N1
Ru1 N2
Ru1 N3
Ru1 N4
Ru1 N5
Ru1 N6
Ru2 N12
Ru3 N18
Ru4 N24
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Ru2 N7
Ru2 N8
Ru2 N11
Ru2 N10
Ru2 N9
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Ru3 N13
Ru3 N14
Ru3 N15
Ru3 N16
Ru3 N17
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Ru4 N19
Ru4 N20
Ru4 N23
Ru4 N22
Ru4 N21
2a
1.976(6)
1.975(6)
1.978(6)
1.976(7)
2.069(6)
2.057(7)
2.079(6)
2.069(7)
2.070(7)
2.062(7)
2.056(7)
2.082(7)
1.976(6)
1.974(7)
1.990(7)
1.986(7)
2.058(7)
2.057(7)
2.066(8)
2.062(7)
2.066(6)
2.064(6)
2.056(6)
2.072(7)
2b
2c
1.963(4)
1.965(4)
2.060(4)
2.070(4)
2.081(5)
2.067(4)
1.989(4)
1.981(4)
2.059(5)
2.074(4)
2.052(4)
2.049(4)
1.977(3)
1.977(3)
2.071(4)
2.071(4)
2.080(3)
2.084(3)
1.995(3)
1.995(3)
2.084(3)
2.080(3)
2.060(3)
2.060(3)
[a] Notations: hydrazone C=N nitrogen atoms (Nhyz): N1, N7, N13, N19; pyrimidine nitrogen atoms: N6, N12,
N18, N24; terpy nitrogen atoms: N3, N4, N5, N9, N10, N11, N15, N16, N17, N21, N22, N23; pyridine nitrogen
atoms from the two-site ligand: N2, N8, N14, N20. [b] Only for 2a.
Table 2. Values of N-Ru-N chelating angles [8] from the X-ray structures of complexes 2a–c[a]
.
N1-Ru1-N2
N1-Ru1-N6
N3-Ru1-N4
N4-Ru1-N5
N7-Ru2-N8
N7-Ru2-N12
N13-Ru3-N18[b]
N19-Ru4-N24[b]
N10-Ru2-N11
N15-Ru3-N16[b]
N22-Ru4-N23[b]
N9-Ru2-N10
N16-Ru3-N17[b]
N21-Ru4-N22[b]
N13-Ru3-N14[b]
N19-Ru4-N20[b]
ꢀ
C4pym N bond and the place of
2a
79.1(3)
79.3(3)
78.8(3)
79.3(3)
78.9(3)
78.4(3)
78.5(3)
78.7(3)
79.4(3)
79.5(3)
78.7(3)
78.4(3)
79.2(3)
79.1(3)
78.5(3)
79.4(3)
ꢀ
the N N bond is taken by the
ꢀ
N CH3 bond, so that the de-
shielding C=N bond and the N
lone pair close to proton G are
replaced by a methyl group. In
the case of 1b, the coordination
causes rotation around the
2b
2c
79.5(2)
78.7(2)
79.3(2)
79.1(2)
79.4(2)
79.1(2)
79.1(2)
78.9(2)
78.5(1)
78.5(1)
79.1(1)
79.1(1)
78.3(1)
78.5(1)
78.5(1)
78.3(1)
ꢀ
C4pym C= bond move and the
place of the deshielding C=N
N
[a] For the notations, see Figure 1b. [b] Only for 2a.
bond and of the N lone pair is
ꢀ
taken by the =C H bond.
The coordination of rutheni-
um(ii), as well as the localiza-
tion of hydrazone group pro-
tons (CH3 and HC N) in the de-
=
shielding zone of the two termi-
nal pyridine rings of the terpys,
make the chemical shift of the
methyl and imine protons
higher in the complexes than in
the free ligands (for the Dd
values, see Table 3). The oppo-
site orientation of tridentate
sites in 2c generates a proximity
between the hydrazone group
and the three rings of the other
terpy ligand; this should induce
a
stronger shielding of the
methyl group and of the proton
of the hydrazone bound in 2c,
relative to that found in 2a or
2b (Figure 2 and Table 3).
Figure 2. 1H NMR (400 MHz) spectra and peak assignment for ligands 1a–c (solvent CDCl3), complexes 2a-c
(solvent CD3CN) and terpy (solvent CDCl3). For the notations of the protons, see Scheme 2.
With respect to the free two-
site ligands, the resonances of
4002
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
Table 3. Chemical shifts of the protons in free (1a–c) and complexed (in 2a–c) two-site ligands.
transoid conformation into an
all-cisoid one. In the case of
terpy, it produces the (L,M)
NOE correlations observed for
all complexes 2a–c. The shape
of the 2-hydrazone-site coordi-
nated ligand was established by
NOE correlations that do not
exist in the free ligand, but
appear after coordination: (D,
E) and (F, G); simultaneously,
the (D, G) correlation charac-
teristic to the free ligands 1a–c
HA
HB
HC
HD
CH3
Hhydrazone
HG
HH
[a]
1a dligand
8.61
7.25
7.73
7.68
8.15
7.82
3.70
4.15
0.45
3.74
4.14
0.40
3.73
3.96
0.23
7.87
8.93
1.06
7.64
8.74
1.10
7.74
8.28
0.54
7.91
6.49
8.47
5.02
ꢀ3.45
9.08
5.26
ꢀ3.82
–
[a]
2a dcomplex
6.92[c]
6.92[c]
Dd[b]
ꢀ1.69
ꢀ0.33
ꢀ0.05
ꢀ0.33
ꢀ1.42
1b dligand
2b dcomplex
Dd
8.29
6.75
6.92
6.65
7.66
7.83
8.42
7.67
0.01
7.66
7.65
7.20[c]
7.73
ꢀ1.54
ꢀ0.27
ꢀ0.63
ꢀ0.69
1c dligand
2c dcomplex
Dd
8.26
6.87
7.60
9.16
6.78
ꢀ1.48
6.63[c]
7.14
ꢀ0.46
7.15
ꢀ2.01
–
–
ꢀ0.24
ꢀ0.01
[a] In ppm. [b] Dd=dcomplexꢀdligand. [c] Central position of the multiplet.
the protons of the terminal pyridines change in the complex.
In the case of protons A, B and D, a shielding effect is ob-
served. The most shielded pyridine proton is proton B. The
most dramatic shift change due to the shielding is observed
for proton A. The chemical shift of proton C is the less af-
fected by the complex formation (Table 3).
The integration of terpy units into the complex by coordi-
nation causes a shielding of protons I, J and L, but a de-
shielding of protons K, M and N, relative to the free terpy
moiety (see Table 4). The stronger shielding effect is ob-
disappeared (for the NOESY of complex 2c, see Figure 3).
Moreover, correlations corresponding to a NOE between
the proton E from the hydrazone strand and the proton I of
the terpy unit, but also between the protons F and I were
observed; this confirms the space proximity between the
two coordinated terpy units and the 1a–c coordinated
strands in the complex 2a–c.
Redox behaviour: Cyclic and differential pulse voltamme-
tries have been employed to investigate the redox behaviour
of the new compounds. For solubility reasons, the free li-
gands were studied in 1,2-dichloroethane and the metal
Table 4. Chemical shifts of terpy protons in the free terpy ligand and in
complexes 2a–c.
HI
HJ
HK
HL
HM
HN
terpy
d[a]
8.71
7.33
7.86
8.62
8.45
7.96
2a
d
Dd[b]
7.89
ꢀ0.82
7.20
ꢀ0.13
7.89
0.03
8.31
ꢀ0.31
8.52
0.07
8.35
0.39
2b
2c
d
7.87
ꢀ0.84
7.20[c]
ꢀ0.13
7.90
0.04
8.32
ꢀ0.30
8.56
0.11
8.43
0.47
Dd
d
7.91
ꢀ0.80
7.27
ꢀ0.06
7.98
0.12
8.47[c]
ꢀ0.15
8.72
0.27
8.49[c]
0.53
Dd
[a] In ppm. [b] Dd=dcomplexꢀdfreeterpy. [c] Central position of the multiplet.
served for the proton I, the average shift decrease being of
around 0.82 ppm. The stronger deshielding is noted for the
proton N, the average increase of the chemical shift being
0.46 ppm. Proton K is, similarly to proton C, less affected by
the coordination. The proximity of the two terpy groups
also generates in the case of 2a and 2b a reciprocal shield-
ing of their protons, whose chemical shifts are generally in-
ferior to the corresponding ones of 2c, while the shifts of
terpy protons have almost identical values in 2a and 2b
(Table 4).
The shape of the coordinated two-site strands as repre-
sented in the structural formula were confirmed by 2D
NMR experiments, essentially by correlations due to the ex-
istence of NOEs (nuclear Overhauser effects) between sev-
eral key-protons closely involved in the conformational
change of the ligand before and after binding of the metal
ion. The coordination induces the transformation of the all-
Figure 3. 1H-1H NOE correlations for the complex 2c (NOESY experi-
ment, 300 MHz, CD3CN).
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4003

J.-M. Lehn, S. Campagna et al.
complexes in acetonitrile. Data are reported in Table 5. De-
tailed discussion of the various processes is as follows.
azine, so the orbitals of the redox-active unit of 1c are less
stabilized than those of 1a and 1b by the central ring. As a
consequence, 1c is oxidized more easily. Such a line of rea-
soning could also explain the difference between the oxida-
tion potentials of 1a and 1b (Table 5), since the redox-
active orbital would be closer to the electron-withdrawing
pyrimidine ring (and therefore more affected electronically
and stabilized) in 1a. The second oxidation process exhibit-
ed by all the three ligand strands is easily interpreted by
Free ligands: Compounds 1a–c undergo two irreversible oxi-
dation processes at relatively mild potentials and no reduc-
tion process takes place within the potential window investi-
gated (ꢀ2.00/+2.00 V vs SCE). As far as the first oxidation
process of 1a–c is concerned, while for 1a and 1b it take
place at similar potentials, that of 1c occurs at a significantly
less negative potential (Table 5). This suggests that the cen-
tral pyrazine ring has an important role in determining the
oxidation potential of 1c. Such a role can be interpreted in
a simple way as follows: in all the cases the orbital involved
in the oxidation process (i.e., the HOMO of the molecule)
ꢀ ꢀ
ꢀ ꢀ
considering that each ligand has two CH3 N N= frame-
works, so first oxidation is mainly centred on one of these
sites and the second on the remaining one. This assignment
agrees with the potential separation between the two oxida-
tion processes in the ligand strands, which is related to the
electronic interactions between the hydrazone-based moiet-
ies and decreases in the series 1a, 1b, 1c (Table 5). The dis-
tance between the redox active sites (inversely proportional
to the electronic interactions) within each ligand indeed de-
creases with the same trend.
should receive large contributions from
a CH3 N N=
framework (probably with the largest contribution involving
the nitrogen atom carrying the methyl group), and the other
moieties can be seen as substituents of the redox-active unit.
Pyrimidine is a better electron-withdrawing group than pyr-
Rack-type metal complexes:
All the rack-type complexes
2a–c undergo two reversible or
quasi-reversible oxidation pro-
Table 5. Absorption, luminescence, and redox data.
Absorption[a]
298 K, lmax [nm]
(e [mcmꢀ1])
Luminescence[b]
Redox Data[c]
298 K,[a] lmax [nm]
(t [ns])
77 K,[d] lmax [nm]
(t [ns])
E1/2 (ox)
E1/2 (red)
cesses and several reduction
processes in the potential
window examined (see Table 5
and Figure 4). The oxidation
processes are easily assigned as
being metal-centred. Actually,
the oxidation potentials are
close to those of other ruthe-
nium(ii) rack-type complex-
es,[1a–d] for which metal-centred
oxidations have been reported.
It has also to be considered that
the ligand-based orbitals are ex-
pected to be significantly stabi-
lized by coordination of the
(positively charged) metal cen-
tres, so that their oxidations,
which in the free ligands are
quite close to the potentials of
the oxidation of the metal
racks, are here expected to
occur to more positive poten-
tials.[14]
[V vs SCE]
[V vs SCE]
1a
1b
284 (26800)
314 (43200)
340 (36500)
395 (16)
390 (630)
+1.30[e]
+1.50[e]
242 (27600)
335 (45150)
379 (32200)
432 (8)
469 (560)
+1.23[e]
+1.33[e]
1c
2a
247 (30200)
409 (67700)
468 (10)
758 (30)
449 (542)
741 (335)
+1.04[e]
+1.10[e]
226 (58900)
270 (47000)
308 (70300)
330 (60300)
362 (52200)
434 (22400)
470 (20700)
+1.28
+1.53
+1.78
ꢀ0.98
ꢀ1.22
ꢀ1.35[f]
2b
240 (42500)
270 (48000)
304 (61100)
330 (42500)
430 (45700)
506 (12600)
630 (27000)
+1.36
+1.66
ꢀ0.45
ꢀ1.05
ꢀ1.46
ꢀ1.65[f]
Splitting of the oxidation
processes in symmetric dinu-
clear metal complexes is due to
1) electronic interaction be-
tween the metal-based orbitals
mediated by superexchange
through the bridging ligand or-
bitals, which depends (beside
other factors) on the energy
separation between the full
metal-based dp and the empty
2c
272 (56700)
280 (53300)
304 (72300)
440 (47400)
614 (40600)
772 (2700)
+1.27
+1.50
ꢀ0.50
ꢀ1.05
ꢀ1.49
ꢀ1.57[f]
[a] For 1a–c, data are in argon-purged dichloromethane; for 2a–c in argon-purged acetonitrile. [b] Lumines-
cence quantum yields at room temperature are not reported, since in all the cases they were lower than 10ꢀ4
and accurate determination was difficult. [c] For 1a–c the measurements were performed in argon-purged 1,2-
dichloroethane, for 2a–c in argon-purged acetonitrile. [d] In MeOH/EtOH 4:1 (v/v) matrix. [e] Irreversible
process. [f] Other irreversible processes take place at more negative potentials, but they are not discussed
here.
,
4004
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
duction processes of 2b and 2c are assigned to the first re-
duction of the two terpy ligands.
As far as the reduction behaviour of 2a is concerned, the
first reduction is still attributed, as for the other racks, to
the first reduction of the bridging ligand; however, it occurs
at a potential that is far more negative than that for first re-
duction of 2b and 2c. This is due to the nature of the orbital
involved, which in 2a is essentially localized on the bis-coor-
dinated pyrimidine, with negligible contribution from the
hydrazone moieties and missing extensive conjugation with
the C=N double bonds. In contrast to the other racks, the
second reduction of 2a cannot be assigned to a second re-
duction of the same pyrimidine site: in fact, potential sepa-
ration between first and second processes in 2a is too small
(240 mV) to be attributed to electron pairing. Therefore,
both the second and third processes are assigned to the
terpy units. Interestingly, the potential separation between
the two terpy-centred processes is 130 mV (Table 5), smaller
with respect to that found in the isomeric species 2b, in
spite of the structural similarity. Whereas a different interac-
tion through the bridging ligand cannot be excluded, cou-
lombic reasons could also play a role: as shown by X-ray de-
termination (see above and also Figure 1a); the distance be-
tween coordinated terpy groups is smaller in 2b than in
2a.[17]
Figure 4. Differential pulse voltammogram of 2b in acetonitrile solution.
The peak at about 0.4 V is ferrocene, used as an internal standard. Work-
ing electrode: glassy carbon (8 mm2, Amel) electrode. Counter electrode:
a Pt wire. Reference electrode: SCE separated with a fine glass frit. The
concentration of the complex was about 5ꢄ10ꢀ4 m. Tetrabutylammonium
hexafluorophosphate (0.05m) was used as supporting electrolyte. Scan
rate: 20 mVsꢀ1
.
bridging ligand-based p* orbitals, and 2) electrostatic inter-
actions.[15] In spite of the first-sight similitude among 2a–c,
these species are different one another from an electronic
viewpoint: complexes 2b and 2c differ from each other by
the central ring of the bridge (pyrimidine vs. pyrazine); this
also introduces differences in the distances between the
For all the complexes, further reduction processes take
place at more negative potentials, but they are ill-defined
and will not be discussed here.
ꢀ
metal centres (although relatively slight; compare the Ru
Ru distances reported in the crystallographic section, 6.13
vs. 6.89 ꢃ for 2b and 2c, respectively). On the other hand,
complex 2a differs from the other two rack-type complexes
because of the energy level and the much more localized
nature of the bridging ligand-centred LUMO (see later the
discussion of the reduction processes). It is hard to discuss
in full detail the differences in the oxidation splitting among
the complex series (250, 300 and 230 mV for 2a, 2b and 2c,
respectively), because they arise from various factors whose
relative contributions are difficult to evaluate. As a matter
of fact, the larger oxidation splitting of 2b relative to 2c is
certainly due to the nature of the central ring of the bridge,
while when compared to 2a can be mainly attributed to the
energy difference between the orbitals involved in the super-
exchange mechanism.
For the sake of convenience, we start the discussion of the
reduction processes from 2b and 2c. Both these species
show four one-electron reduction processes at very similar
potentials. The first two processes take place at potentials
less negative than the first terpy-based reduction occurring
in [Ru(terpy)2]2+ (ꢀ1.25 V vs SCE),[14] so they are both at-
tributed to the bridging ligands. In particular, reduction can
be assigned to the central part of the bridging ligand strand,
involving a delocalized orbital that receives dominant con-
tributions from the pyrimidine/pyrazine ring and the two
conjugate C=N double bonds. The difference in potential
between the first two processes (600 and 550 mV for 2b and
2c, respectively) also agrees with the electron-pairing
energy expected for such orbitals.[16] The third and fourth re-
Absorption spectra: The absorption spectra of the free li-
gands (solvent: dichloromethane) show moderately intense
bands in the 260–460 nm spectral region (see Figure 5a and
Table 5). For such ligands, both spin-allowed p–p* and n–p*
transitions are expected,[18,19] with the latter having lower ex-
tinction coefficients. The absorption spectrum of terpy dis-
plays one band peaking at 280 nm, due to the spin-allowed
p–p* transition, and a weak band around 340 nm, due to
n–p* transitions.[14,19] The relatively high absorption coeffi-
cients of the bands which dominate the spectra of 1a–c sug-
gest that such bands are due to spin-allowed p–p* transi-
tions. Figure 5a shows a progressive increase of the lmax of
the lowest energy band in the following order: 1a (340 nm),
1b (379 nm) and 1c (409 nm). This bathocromic effect can
be explained by the nature of the diazine rings and by the
increase of conjugation between the heteroaromatic ring
and the C=N bonds. On increasing the conjugation, the en-
ergetic difference between the p and p* levels decreases, so
the wavelength of the corresponding p–p* transitions in-
creases. In 1a, the lowest energy bands receive contributions
from p–p* transitions involving orbitals localized on the
central pyrimidine ring or in the “peripheral” pyridine–hy-
drazone moieties. In 1b, there are two chromophoric C=N
bonds in conjugation with the same pyrimidine ring, in meta
orientation. The lowest energy band is therefore here as-
signed to a p–p* transition delocalized over the central
moiety of the ligand and is red-shifted relative to that of 1a.
When the orientation of the two C=N bonds is para, as in
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4005

J.-M. Lehn, S. Campagna et al.
MLCT transitions involving the terpy ligands. As far as the
intense bands at l>600 nm of 2b and 2c are concerned,
they can be attributed to spin-allowed MLCT transitions in-
ꢀ
ꢀ
ꢀ
ꢀ
volving the N=C diazine C=N central moiety of the
bridge as the acceptor orbital, while the corresponding Ru-
to-bridge CT transition in 2a, involving the much more diffi-
cult-to-reduce “isolated” pyrimidine fragment of the bridg-
ing ligand, is probably responsible for the broad absorption
between 500 and 600 nm of 2a.
Finally, the spectrum of 2c also shows a relatively weak
band at about 775 nm (Table 5, Figure 5b). We tentatively
attribute such a weak band to a spin-forbidden singlet-triplet
MLCT transition. Actually, 3MLCT absorption bands are
hardly discernable in ruthenium(ii) complexes. Although
never mentioned in mononuclear complexes, a few cases
have been reported in multinuclear complexes. In particular,
non-negligible spin-forbidden MLCT absorption has been
mentioned in a tetranuclear species containing 2,3-bis(2’-
pyridyl)pyrazine as a bridging ligand.[21] Interestingly, that
bridging ligand and the one in 2c share a pyrazine ring as
the central subunit of the bridge. This could justify why the
spin-forbidden transition is visible in 2c but not in 2a and
2b.
Luminescence properties: The luminescence properties of
the free ligands 1a–c and of the rack-type complexes 2a–c
have been investigated in dichloromethane (free ligands)
and acetonitrile (complexes) fluid solution at room tempera-
ture and in an MeOH/EtOH 4:1 (v/v) rigid matrix at 77 K
(all the compounds). The free ligands exhibit luminescence
under all the experimental conditions used, while among the
metal complexes only 2a is luminescent, both at room tem-
perature and at 77 K. Data are collected in Table 5, while
some luminescence spectra are shown in Figure 6.
Luminescence energies and lifetimes of 1a–c suggest that
the excited state from which emission takes place is the
lowest lying singlet p–p* level. Actually, emission spectra
look like the spectral image (in particular, at room tempera-
ture) of the lowest lying p–p* absorption band. The emis-
sion maxima gradually move to lower energy in the series
1a, 1b, 1c (Table 5, Figure 6a), in agreement with the attri-
bution. That the emissive state also has a CT character, as
mentioned in the absorption spectra discussion, is supported
by the solvent dependence of the emission band: for exam-
ple, the room temperature emission maximum of 1a moves
from 432 nm in dichloromethane to 442 nm in methanol, so
indicating that the excited state has a larger dipolar nature
than the ground state. This partial CT character, which sug-
gests mixing with close lying n–p* states, also accounts for
the quite low emission quantum yields (lower than 10ꢀ4, see
Table 5), which would be surprisingly low if a pure p–p*
fluorescence was considered. At 77 K the luminescence life-
times of the free ligands increase substantially (Table 5, Fig-
ure 6a); the spectra become structured and weakly blue-
shifted. Altogether, these changes would suggest a reduced
contribution at 77 K of the CT character in the emissive
levels.
Figure 5. a) Absorption spectra of 1a (solid line), 1b (dashed line) and
1c (dotted line) in dichloromethane. b) Absorption spectra of 2a (solid
line), 2b (dashed line) and 2c (dotted line) in acetonitrile.
1c, the conjugation is improved and the p–p* transition is at
even lower energies. In all the ligands, the lower intensity n–
p* bands are probably obscured, so their positions are not
experimentally defined. It should be considered anyway that
in hydrazone molecules containing electron-withdrawing
groups (in the present case, diazine rings), the p–p* transi-
tions can have some charge-transfer character,[20] also in-
duced by mixing between p–p* and n–p* states. This can
have important effects in particular as far as the deactiva-
tion processes are concerned (see luminescence section).
The absorption spectra of the rack-type metal complexes
(Figure 5b, Table 5; solvent: acetonitrile) exhibit intense
bands both in the UV and in the visible region. For all the
complexes, the bands in the UV region receive main contri-
butions from spin-allowed ligand-centred (LC) transitions
(namely, p–p*) involving the terpy ligands (in particular, the
band peaking at about 310 nm, common to all the species, is
attributed to terpy-centered transitions), with the contribu-
tion from LC transitions involving the bridging ligands; this
becomes especially relevant in the case of 2a. The bands
peaking at 430 and 440 nm in the spectra of 2b and 2c, re-
spectively, are attributed to superposition of the lowest lying
LC transitions involving the bridging ligands and the spin-al-
lowed metal-to-ligand charge-transfer (MLCT) transitions
involving terpy ligands. In 2a, the lowest lying LC band
occurs in the UV region, so that the absorption between 400
and 500 nm probably receives contributions only from
4006
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
mainly centred on the substituted pyrimidine ring. The
slight blue-shift of the emission spectrum and its longer life-
time on passing from room temperature fluid solution to
77 K rigid matrix are in line with the CT attribution.
The absence of emission within the wavelength region in-
vestigated (l<850 nm) for 2b and 2c is not surprising: in
fact, on considering the energy of the singlet MLCT state as
inferred by absorption spectroscopy (>600 nm in both com-
plexes), the corresponding triplet levels should lie at wave-
length longer than 800 nm and could be out of the sensitivity
range of our equipment (note that the band assigned in 2c
to the singlet–triplet absorption lies close to 800 nm, so the
corresponding emission spectrum would be further displaced
in the near-IR region). In these conditions, we cannot ex-
clude that 2b and 2c can exhibit luminescence in the near-
IR region, although it is expected that because of the
energy-gap law, the eventual luminescence quantum yields
would be extremely low.
Conclusion
New RuII bimetallic isomeric complexes (2a–c) with ditopic,
isomeric hydrazone-based ligands were synthesized and
characterized by several techniques, including NMR meth-
ods and crystallography. The absorption spectra, redox be-
haviour and luminescence properties of the new complexes
and of the free ligands have been studied. The results indi-
cate that the molecular diversity introduced by the isomer-
ism translates into interesting differences as far as the spec-
troscopic, luminescence and redox properties of the ligands
and the rack-type new metal complexes are concerned. For
example, spin-allowed MLCT bands (l>600 nm) are pres-
ent in 2b and 2c, while the lowest energy absorption MLCT
band of 2a lies at significantly higher energy. Moreover,
Figure 6. a) Luminescence spectra of 1a (solid line), 1b (dashed line) and
1c (dotted line) in dichloromethane at room temperature and (inset) lu-
minescence spectrum of 1b in MeOH/EtOH 4:1 (v/v) at 77 K. b) Lumi-
nescence spectra of 2a at room temperature in acetonitrile fluid solution
(solid line) and at 77 K in an MeOH/EtOH 4:1 (v/v) rigid matrix (dashed
line).
Ruthenium(ii) polypyridine complexes are usually lumi-
3
nescent from their lowest lying MLCT state.[14,22] However,
complexes containing tridentate ligands (e.g., terpy), are
scarcely luminescent at room temperature relative to species
based on bidentate ligands (e.g., 2,2’-bipyridine).[14,22b] Re-
cently, this drawback was overcome by several strategies, in-
cluding the use of ligands having reduction potentials less
negative.[2,23] The presence of such ligands moves the corre-
3
while 2a exhibits a typical MLCT emission both in fluid so-
lution at room temperature and in rigid matrix at 77 K, com-
plexes 2b and 2c do not show any emission for l<850 nm.
In general, it is shown that by taking advantage of relatively
slight differences within the ditopic, isomeric hydrazone-
based ligands strands, predetermined molecular architec-
tures exhibiting significantly different properties can be ob-
tained.
3
sponding MLCT state to low energies, so decreasing the ef-
ficiency of the activated process to the “poisoning” (with
3
regard to the luminescence properties) MC level. This ap-
proach anyway is sometimes a two-edged sword, since very
low MLCT states have diminished luminescence properties
Preparation of tri- and poly-, homo- or heterometallic
complexes of the same family, as well as their decoration
with suitable groups able to modify their electronic proper-
ties in predetermined ways, could constitute further devel-
opments in this area of multicentre rack-type complexes. In
particular, the combination of the specific redox properties
with the difference in shape between the “lateral” com-
plexes 2a and 2b and the “alternate” complex 2c suggests
the possibility to generate extended multicentre metallosu-
pramolecular architectures of curved or linear, respectively,
“metallomolecular wires”.
3
because the Franck–Condon factors for radiationless decay
of the (potentially) luminescent level increase on decreasing
the excited state energy (i.e., the basis of the so-called
energy-gap law[24]). To obtain room-temperature emission
from ruthenium(ii) complexes, the above-mentioned effects
have to be taken into account and suitably balanced. The
rack-type complex 2a indeed shows a relatively long-lived
emission both at room temperature and at 77 K (Table 5,
Figure 6b). The spectral shape, energy and lifetime, together
with comparison with literature data,[14,22] indicate that such
an emission can be straightforwardly attributed to the
3
lowest-lying MLCT state, in which the acceptor orbital is
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4007

J.-M. Lehn, S. Campagna et al.
dimensions: 0.20ꢄ0.04ꢄ0.02 mm; 1calcd =1.70 gcmꢀ3
; F(000): 922; m=
Experimental Section
0.627 mmꢀ1; hkl limits: ꢀ12,10/ꢀ18,18/ꢀ22,22. q range: 2.5ꢁqꢁ30.068;
number of data measured: 16315; number of data with I>3s(I): 5635;
number of variables: 496; R=0.048; Rw =0.062; goodness-of-fit: 1.059;
Materials and general methods: The following compounds were prepared
as previously described: 1a,b,[7] 3,[11] 4,[10] 5,[10] 6[9] and [Ru(terpy)Cl3].[12]
The following reagents were purchased from commercial sources: RuCl3
(Aldrich, Avocado), terpy (Aldrich, Avocado), 2,5-dimethylpyrazine (Al-
drich), benzaldehyde (Aldrich), benzoic anhydride (Aldrich). 400 MHz
1H and 100 MHz 13C NMR spectra were recorded on a Bruker Ultra-
shield Avance 400 spectrometer and 300 MHz 1H and 75 MHz 13C NMR
spectra were recorded on a Bruker 300 spectrometer. The solvent residu-
al signal[25] was used as an internal reference for both 1H and 13C NMR
largest peak in final difference: 1.286 eꢃꢀ3
.
CCDC-257649 (2a), CCDC-257650 (2b) and CCDC-257651 (2c) contain
the supplementary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Pyrazine-2,5-dicarboxaldehydebis[methyl(pyridin-2-yl)hydrazone] (1c):
A solution of hydrazine derivative 3 (100 mg, 0.813 mmol, 2.12 equiv)
and dialdehyde 4 (52 mg, 0.382 mmol, 1 equiv) in EtOH (5 mL) was
heated to reflux for 3 h. Then, the mixture was cooled and filtered. The
precipitate was washed with EtOH and dried for 10 h under high
vacuum: 1c (86 mg, 65%). Yellow solid; m.p. 3058C; 1H NMR
(400 MHz, CDCl3):d=9.16 (s, 2H; HG), 8.26 (ddd, J=1.0, 2.0, 5.0 Hz,
2H; HA), 7.76 (dt, J=1, 8.7 Hz, 2H; HD), 7.74 (s, 2H; HF), 7.66 (ddd, J=
2.0, 7.0, 8.7 Hz, 2H; HC), 6.87 (ddd, J=1.0, 5.0, 7.0 Hz, 2H; HB),
3.73 ppm (s, 6H; HE); 13C NMR (75 MHz, CDCl3): d=157.26, 148.63,
147.19, 140.89, 137.95, 132.29, 116.86, 110.41, 29.89 ppm; IR (KBr): n˜ =
3004 (w), 2902 (w), 1590 (s), 1560 (s), 1483 (s), 1435 (s), 1349 (s), 1213
(s), 1131 (s), 1029 (m), 921 (s), 868 (m), 770 (s), 731 (m), 613 cmꢀ1 (w);
UV/Vis (CHCl3): lmax =247, 409 nm; FAB-MS: m/z (%): 347.2 (100)
[M+H]+; HRMS (FAB-MS): calcd for [C18H18N8 +H]+: 347.1733; found:
347.1737.
1
spectra. The following notation is used for the H NMR spectral splitting
patterns: singlet (s), doublet (d), triplet (t), multiplet (m). The 2D-NMR
used experiments were: COSY (correlation spectroscopy), NOESY (Nu-
clear Overhauser enhancement spectroscopy or nuclear Overhauser and
exchange spectroscopy), ROESY (rotating-frame Overhauser enhance-
ment (effect) spectroscopy); they were done on 300 MHz or 500 MHz
Bruker spectrometers. FAB-MS, EI-MS and ES-MS measurements were
performed by the Service de Spectromꢀtrie de Masse, Universitꢀ Louis
Pasteur. Melting points were recorded on a Bꢅchi Melting Point B-540
apparatus and are uncorrected. UV/Vis spectra were recorded on a
Varian-Cary-3 spectrometer or on a Jasco V-560 spectrometer. IR spectra
were recorded on a Perkin–Elmer 1600-FTIR spectrometer; KBr pellets.
Luminescence spectra were recorded on a Jobin-Yvon Spex Fluoromax P
fluorimeter equipped with a Hamamatsu R3896 photomultiplier. Emis-
sion spectra have been corrected by the use of software purchased with
the fluorimeter. Luminescence lifetimes were determined by time-corre-
lated single-photon-counting (TCSPC) with an Edinburgh OB900 spec-
trometer (light pulse: Hamamatsu PL2 laser diode, pulse width 59 ps at
408 nm; or nitrogen discharge, pulse width: 2 ns). Electrochemical experi-
ments were performed as previously described,[1f] with acetonitrile (metal
complexes) or 1,2-dichloroethane (free ligands) as solvents. Experimental
uncertainties are as follows: absorption maxima, 2 nm; emission maxima,
4 nm; emission lifetimes, 10%; redox potentials, 10 mV.
[{Ru(terpy)}2(1a)](PF6)4 (2a): [Ru(terpy)Cl3] (19.6 mg, 0.044 mmol,
2.2 equiv) and free ligand 1a (7 mg, 0.020 mmol, 1equiv) were dissolved
in ethanol/water (5 mL, 1:1 v/v). The mixture was heated under reflux for
19 h, then cooled to room temperature and filtered. Excess aqueous
NH4PF6 was added to the solution and the precipitate was collected. The
solid was purified by recrystallization from acetonitrile/Et2O to afford 2a
(23 mg, 71%) as a brown solid. M.p. >3008C; 1H NMR (400 MHz,
CD3CN): d=8.93 (s, 2H; HE), 8.52 (d, J=8.2 Hz, 4H; HM), 8.37 (t, J=
8.2 Hz, 2H; HN), 8.31 (d, J=7.8 Hz, 4H; HL), 7.93–7.88 (m, 8H; HK, HI),
7.84 (d, J=7.8 Hz, 2H; HD), 7.72–7.66 (m, 2H; HC), 7.23–7.19 (m, 4H;
HJ), 6.94–6.91 (m, 4H; HA, HB), 6.49 (s, 1H; HG), 5.02 (s, 1H; HH),
4.15 ppm (s, 6H; HF); IR (KBr): n˜ =3110 (w), 1614 (s), 1566 (w), 1527
(m), 1460 (s), 1391 (m), 1268 (m), 1122 (s), 1006 (w), 842 (s), 558 cmꢀ1
(s); UV/Vis (CH3CN): lmax =226, 270, 308, 330, 362, 434, 470 nm; ES-
MS: m/z calcd for [Ru2(1a)(terpy)2(PF6)3]+ =[C48H40F18N14P3Ru2]+:
Crystal structure determinations: The crystals were obtained by diffu-
sion–recrystallization by using pure acetonitrile (for 2c) or a CH3CN/
CH3OH 9:1 mixture (for 2a and 2b) as a solvent and Et2O (for 2a and
2b) or iPr2O (for 2c) as a nonsolvent.
The crystals were placed in oil, and a single crystal was selected, mounted
on a copper wire and placed in a low-temperature N2 stream (T=173 K).
The X-ray diffraction data were collected on a Nonius-Kappa-CCD dif-
fractometer with graphite monochromatized MoKa radiation (l=
0.71073 ꢃ), f scans. The structures were solved by direct methods and re-
fined (based on F2 with all independent data) by full-matrix least-squares
methods (OpenMolen package (for 2a and 2c) or SHELXL-97 (for 2b)).
1451.1,
found:
1451.1;
calcd
for
[Ru2(1a)(terpy)2(PF6)2]2+
=
[C48H40F12N14P2Ru2]2+
:
653.1, found: 652.8; calcd for [Ru2(1a)-
(terpy)2PF6]3+ =[C48H40F6N14PRu2]3+: 387.0, found: 386.8; calcd for [Ru2-
(1a)(terpy)2]4+ =[C48H40N14Ru2]4+: 254.1, found: 253.8.
[{Ru(terpy)}2(1b)](PF6)4 (2b): [Ru(terpy)Cl3] (19.6 mg, 0.044 mmol,
2.2 equiv) and free ligand 1b (7 mg, 0.020 mmol, 1 equiv) were dissolved
in ethanol/water (5 mL, 1:1 v/v). The mixture was heated under reflux for
20 h, then cooled to room temperature and filtered. Excess aqueous
NH4PF6 was added to the solution and the precipitate was collected. The
solid was purified by recrystallization from water, then from acetonitrile/
Data for 2a: Formula:
C105H100F48N32O3P8Ru4 (2C48H40N14Ru2·
8PF6·4CH3CN·CH3OH·2H2O); Mr: 3422.18; crystal system: triclinic;
space group: P1; cell dimensions: a=14.5964(1), b=20.5711(2), c=
¯
25.0751(2) ꢃ, a=102.860(5)8, b=105.329(5)8, g=93.126(5)8, V=
7027.5(1) ꢃ3; Z=2; colour: brown; crystal dimensions: 0.16ꢄ0.13ꢄ
0.08 mm; 1calcd =1.62 gcmꢀ3; F(000): 3412; m=0.635 mmꢀ1; hkl limits:
ꢀ19,20/ꢀ28,28/ꢀ35,35; q range: 2.5ꢁqꢁ30.048; number of data mea-
sured: 61742; number of data with I>3s(I): 25682; number of variables:
1661; R=0.085; Rw =0.114; goodness-of-fit: 1.317; largest peak in final
Et2O to afford 2b (20 mg, 62%) as
a green solid. M.p. >3008C;
1H NMR (400 MHz, CD3CN): d=8.74 (s, 2H; HF), 8.56 (d, J=8.1 Hz,
4H; HM), 8.43 (t, J=8.1 Hz, 2H; HN), 8.32 (d, J=8.1 Hz, 4H; HL), 7.90
(td, J=1.6, 8.1 Hz, 4H; HK), 7.87 (d, J=6.3 Hz, 4H; HI), 7.73 (s, 1H;
HG), 7.67 (ddd, J=1.6, 7.4, 8.8 Hz, 2H; HC), 7.23–7.19 (m, 6H; HD, HJ),
6.75 (dd, J=1.0, 5.5 Hz, 2H; HA), 6.65 (ddd, J=1.0, 5.5, 7.4 Hz, 2H; HB),
5.26 (s, 1H; HH), 4.14 ppm (s, 6H; HF); IR (KBr): n˜ =3111 (w), 1603 (m),
1491 (m), 1436 (s), 1382 (w), 1321 (m), 1149 (m), 1109 (w), 838 (s), 767
(m), 558 cmꢀ1 (s); UV/Vis (CH3CN): lmax =240, 270, 304, 330, 430, 506,
difference: 1.017 eꢃꢀ3
.
Data for 2b: Formula: C54H49N17Ru2P4F24 (C48H40N14Ru2·4PF6·3CH3CN);
Mr: 1718.12; crystal system: monoclinic; space group: P21/n; cell dimen-
sions: a=14.882(5), b=13.131(5), c=34.498(5) ꢃ, a=90.00(5)8, b=
100.31(5)8, g=90.00 (5)8, V=6633(4) ꢃ3; Z=4; colour: green; crystal di-
mensions: 0.10ꢄ0.08ꢄ0.05 mm; 1calcd =1.721 gcmꢀ3; F(000): 3424; m=
630 nm;
ES-MS:
m/z
calcd
for
[Ru2(1b)(terpy)2(PF6)3]+
=
[C48H40F18N14P3Ru2]+: 1451.1, found: 1451.0; calcd for [Ru2(1b)(terpy)2-
(PF6)2]2+ =[C48H40F12N14P2Ru2]2+: 653.1, found: 652.8; calcd for [Ru2(1b)-
(terpy)2PF6]3+ =[C48H40F6N14PRu2]3+: 387.0, found: 386.8; calcd for [Ru2-
(1b)(terpy)2]4+ =[C48H40N14Ru2]4+: 254.1, found: 253.8.
0.672 mmꢀ1
;
hkl limits: ꢀ20,20/0,18/0,48;
q
range: 1.20ꢁqꢁ30.058;
number of data measured: 19356; number of data with I>2s(I): 10601;
number of variables: 895; R=0.0523; Rw =0.1682; goodness-of-fit: 1.060.
Data for 2c: Formula: C30H29F12N10P2Ru ((C48H40N14Ru2·4PF6·6CH3CN)/
2); Mr: 920.63; crystal system: triclinic; space group: P1; cell dimensions:
a=8.8208(2), b=13.0232(3), c=15.8342(4) ꢃ, a=97.160(5)8, b=
95.309(5)8, g=90.199(5)8, V=1796.82(7) ꢃ3; Z=2; colour: green; crystal
[{Ru(terpy)}2(1c)](PF6)4 (2c): Ru(terpy)Cl3 (19.6 mg, 0.044 mmol,
2.2 equiv) and free ligand 1c (7 mg, 0.020 mmol, 1 equiv) were dissolved
in ethanol/water (5 mL, 1:1 v/v). The mixture was heated under reflux for
20 h, then cooled to room temperature and filtered. Excess aqueous
¯
4008
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3997 – 4009

Supramolecular Chemistry
FULL PAPER
NH4PF6 was added to the solution and the precipitate was collected. The
solid was purified by recrystallization from acetonitrile/Et2O to afford 2c
(19 mg, 59%) as a green solid. M.p. >3008C; 1H NMR (400 MHz,
CD3CN): d=8.73 (d, J=7.8 Hz, 4H; HM), 8.52–8.46 (m, 6H; HN, HL),
8.28 (s, 2H; HF), 7.98 (td, J=1.5, 7.8 Hz, 4H; HK), 7.91 (d, J=5.9 Hz,
4H; HI), 7.66 (ddd, J=1.0, 7.3, 8.8 Hz, 2H; HC), 7.27 (ddd, J=1.0, 5.4,
6.9 Hz, 4H; HJ), 7.15 (d, J=8.8 Hz, 2H; HD), 7.14 (s, 2H; HG), 6.78 (dd,
J=1.0, 6.0 Hz, 2H; HA), 6.65–6.61 (m, 2H; HB), 3.98 ppm (s, 6H; HE);
IR (KBr): n˜ =3095 (w), 1605 (m), 1509 (s), 1468 (s), 1404 (s), 1317 (m),
1198 (m), 1036 (w), 837 (s), 766 (s), 557 cmꢀ1 (s); UV/Vis (CH3CN):
lmax =272, 280, 304, 440, 614, 772 nm; ES-MS: m/z calcd for [Ru2(1c)-
(terpy)2(PF6)3]+ =[C48H40F18N14P3Ru2]+: 1451.1, found: 1451.0.
Schofield, E. Figgemeier, D. Fenske, M. Neuburger, J. G. Vos, M.
Zehnder, Chem. Eur. J. 2002, 8, 137.
[3] M. Benaglia, F. Ponzini, C. R. Woods, J. S. Siegel, Org. Lett. 2001, 3,
967.
[4] D. Brown, S. Murajan, Y. Jang, R. Thummel, Org. Lett. 2002, 4,
1253.
[5] J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives,
VCH, Weinheim, 1995.
[6] K. M. Gardinier, R. G. Khoury, J.-M. Lehn, Chem. Eur. J. 2000, 6,
4124.
[7] J.-L. Schmitt, A.-M. Stadler, N. Kyritsakas and J.-M. Lehn, Helv.
Chim. Acta 2003, 86, 1598.
[8] a) T. Bark, M. Dꢅggeli, H. Stoeckli-Evans, A. von Zelewsky, Angew.
Chem. 2001, 113, 2924; Angew. Chem. Int. Ed. 2001, 40, 2848; b) T.
Bark, H. Stoeckli-Evans, A. von Zelewsky, J. Chem. Soc. Perkin
Trans. 1, 2002, 1881.
[9] R. Franke, Ber. Dtsch. Chem. Ges. 1905, 38, 3724.
[10] a) R. H. Wiley, U. S. Pat. 4, 260,757, 1981; b) R. H. Wiley, J. Macro-
mol. Sci. Chem. 1987, A24, 1183.
[11] M. A. Baldo, G. Chessa, G. Marangoni, B. Pitteri, Synthesis 1987,
720.
[12] B. P. Sullivan, J. M. Calvert, T. J. Meyer, Inorg. Chem. 1980, 19,
1404.
[13] A.-M. Stadler, J.-M. Lehn, unpublished results.
[14] A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, A. von
Zelewsky, Coord. Chem. Rev. 1988, 84, 85, and references therein.
[15] a) J. E. Sutton, H. Taube, Inorg. Chem. 1981, 20, 3125; b) C. Creutz,
Prog. Inorg. Chem. 1983, 30, 1; c) N. S. Hush, Coord. Chem. Rev.
1985, 64, 135; d) S. Ernst, V. Kasack, W. Kaim, Inorg. Chem. 1988,
27, 1146; e) G. Giuffrida, S. Campagna, Coord. Chem. Rev. 1994,
135, 517; f) M. D. Ward, Chem. Soc. Rev. 1995, 24, 121; g) B. D. Yeo-
mans, L. S. Kelso, P. A. Tregloan, F. R. Keene, Eur. J. Inorg. Chem.
2001, 239.
[16] a) A. A. Vlceck, Coord. Chem. Rev. 1982, 43, 39; b) M. Marcaccio,
F. Paolucci, C. Paradisi, S. Roffia, C. Fontanesi, L. J. Yellowlees, S.
Serroni, S. Campagna, G. Denti, V. Balzani, J. Am. Chem. Soc. 1999,
121, 10081, and references therein.
[17] It should be considered that redox data refers to solution, while X-
ray determination indicates “frozen” solid-state situations. However
it appears reasonable that in the average, distance differences can
also exist in solution.
[18] H. H. Jaffꢀ, M. Orchin, Theory and Applications of Ultraviolet Spec-
troscopy, Wiley, 1964, Chapter 14.
[19] M. Klessinger, J. Michl, Excited States and Photochemistry of Organ-
ic Molecules, VCH, New York, 1995.
[20] A. Giannetto, G. Guglielmo, V. Ricevuto, A. Giuffrida, S. Campa-
gna, J. Photochem. Photobiol. A 1990, 53, 23.
[21] J. Andersson, F. Puntoriero, S. Serroni, A. Yartsev, T. Pascher, T. Po-
livka, S. Campagna, V. Sundstrçm, Chem. Phys. Lett. 2004, 386, 336.
[22] a) G. R. Crosby, Acc. Chem. Res. 1975, 8, 231; b) T. J. Meyer, Pure
Appl. Chem. 1986, 58, 1193.
[23] M. I. J. Polson, E. A. Medlycott, G. S. Hanan, L. Mikelsons, N. J.
Taylor, M. Watanabe, Y. Tanaka, F. Loiseau, R. Passalacqua, S.
Campagna, Chem. Eur. J. 2004, 10, 3640.
[24] a) W. Siebrand, J. Chem. Phys. 1967, 46, 440; b) J. V. Caspar, T. J.
Meyer, J. Phys. Chem. 1983, 87, 952.
2,5-Pyrazinedicarboxaldehyde (4): Compound 4 was prepared by ozono-
lysis of 6, as previously described.[10] It was purified by recrystallization
(reprecipitation) from a CHCl3/iPr2O mixture. Yield 20–30%; yellowish
solid; m.p. 124–1258C; 1H NMR (400 MHz, CDCl3): d=10.23 (s, 2H),
9.30 ppm (s, 2H); 13C NMR (100 MHz, CDCl3): d=191.79, 148.77,
143.41 ppm; UV/Vis (CHCl3): lmax =280 nm; EI-MS: m/z (%): 136.1
(100) [M]+, 107.1 (47) [Mꢀ29]+, 80.0 (12) [Mꢀ56]+, 52.1 (34) [Mꢀ84]+,
29.1 (11) [Mꢀ107]+; HRMS (EI-MS): calcd for [C6H4O2N2]+:136.0273;
found: 136.0275.
2,5-(E,E)-Distyrylpyrazine (6): Compound 6 was prepared by condensa-
tion of benzaldehyde with 2,5-dimethylpyrazine, as previously de-
scribed.[9] Yellow solid; m. p. 2248C; 1H NMR (300 MHz, CDCl3): d=
8.60 (s, 2H), 7.74 (d, J=16.1 Hz, 2H), 7.61 (d, J=7.2 Hz, 4H), 7.44–7.30
(m, 6H), 7.18 ppm (d, J=16.1 Hz, 2H); 13C NMR (75 MHz, CDCl3): d=
149.24, 143.42, 136.42, 134.53, 128.99, 127.44, 124.25 ppm (two overlaping
peaks); UV/Vis (CHCl3): lmax =305, 384 nm.
Acknowledgements
We thank Dr. Annie Marquis and Raymond Hueber for mass spectrome-
try analyses. A.-M.S. thanks le Ministꢆre de la Recherche, de la Jeunesse
et de la Technologie for a predoctoral fellowship. S.C. also acknowledges
MIUR and FIRB for financial support.
[1] a) G. S. Hanan, C. Arana, J.-M. Lehn, D. Fenske, Angew. Chem.
1995, 107, 1191; Angew. Chem. Int. Ed. Engl. 1995, 34, 1122; b) G. S.
Hanan, C. Arana, J.-M. Lehn, G. Baum, D. Fenske, Chem. Eur. J.
1996, 2, 1292; c) B. Hasenknopf, J. Hall, J.-M. Lehn, V. Balzani, A.
Credi, S. Campagna, New J. Chem. 1996, 20, 725; d) P. Ceroni, A.
Credi, V. Balzani, S. Campagna, G. S. Hanan, C. Arana, J.-M. Lehn,
Eur. J. Inorg. Chem. 1999, 9, 1409; e) M. Ruben, E. Breuning, M.
Barboiu, J.-P. Gisselbrecht, J.-M. Lehn, Chem. Eur. J. 2003, 9, 291;
f) D. M. Bassani, J, -M. Lehn, S. Serroni, F. Puntoriero, S. Campa-
gna, Chem. Eur. J. 2003, 9, 5936.
[2] a) M. Maestri, N. Armaroli, V. Balzani, E. C. Constable, A. M. W.
Cargill Thompson, Inorg. Chem. 1995, 34, 2759; b) J.-P. Sauvage, J. P.
Collin, J. C. Chambron, S. Guillerez, C. Coudret, V. Balzani, F. Bari-
gelletti, L. De Cola, L. Flamigni, Chem. Rev. 1994, 94, 993; c) A. El-
ghayoury, A. Harriman, A. Khatyr, R. Ziessel, Angew. Chem. 2000,
112, 191; Angew. Chem. Int. Ed. 2000, 39, 185; d) A. El-ghayoury, A.
Harriman, A. Khatyr, R. Ziessel, J. Phys. Chem. A 2000, 104, 1512;
e) F. Barigelletti, L. Flamigni, Chem. Soc. Rev. 2000, 29, 1; f) F. Loi-
seau, R. Passalacqua, S. Campagna, M. I. J. Polson, Y.-Q. Fang, G. S.
Hanan, Photochem. Photobiol. Sci. 2002, 1, 982; g) S. Encinas, L.
Flamigni, F. Barigelletti, E. C. Constable, C. E. Housecroft, E. R.
[25] H. E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 1997, 62,
7512.
Received: December 8, 2004
Published online: April 21, 2005
Chem. Eur. J. 2005, 11, 3997 – 4009
www.chemeurj.org
ꢂ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4009