6,6″-azobis(2,2′-bipyridine) and its dinuclear ruthenium complex: A comparative study with positional isomers

Article abstract of DOI:10.1246/bcsj.80.902

The azobis(bipyridine) ligand 6,6″-azobis(2,2′-bipyridine) (6-azobpy) was prepared as a new member of the family of azopolypyridine derivatives. This compound undergoes reversible trans/cis photoisomerization. Reaction of 6-azobpy and [Ru(bpy)₂Cl₂] afforded a dinuclear complex [{Ru(bpy)₂}₂(μ-6-azobpy)] ⁴⁺. An analogue with deuterated bpy fragments was also prepared to help clarify the ¹HNMR signals in the part of the 6-azobpy ligand. The analysis of ¹H NMK spectra clearly identified the structure of the complex, in which the Ru²⁺ ions are chelated by the bpy parts of the bridging 6-azobpy ligand. Electronic absorption and electrochemical properties are characterized by the presence of a low-lying π level of the bridging ligand, manifested by a metal-to-ligand charge-transfer band extending down to 800 nm and by two consecutive one-electron reduction waves at relatively less negative potentials in cyclic voltammetry. These properties were compared with those of previously reported positional isomers with the help of density functional theory (DFT), time-dependent DFT, and Z1NDO calculations. The analysis of the redox couple of Ru^3+/2+ showed that the redox potentials differ slightly between the diastereomers and that there is an intermetallic electronic interaction.

Full text of DOI:10.1246/bcsj.80.902

902 Bull. Chem. Soc. Jpn. Vol. 80, No. 5, 902–909 (2007)
ꢀ 2007 The Chemical Society of Japan
6,600-Azobis(2,20-bipyridine) and Its Dinuclear Ruthenium Complex:
A Comparative Study with Positional Isomers
ꢀ
Joe Otsuki, Izumi Kurihara, Arata Imai, Yuki Hamada, and Nobuyuki Omokawa
College of Science and Technology, Nihon University, 1-8-14 Kanda Surugadai, Chiyoda-ku, Tokyo 101-8308
Received July 14, 2006; E-mail: otsuki@chem.cst.nihon-u.ac.jp
The azobis(bipyridine) ligand 6,600-azobis(2,20-bipyridine) (6-azobpy) was prepared as a new member of the family
of azopolypyridine derivatives. This compound undergoes reversible trans/cis photoisomerization. Reaction of 6-azobpy
and [Ru(bpy)2Cl2] afforded a dinuclear complex [{Ru(bpy)2}2(m-6-azobpy)]4þ. An analogue with deuterated bpy frag-
ments was also prepared to help clarify the 1H NMR signals in the part of the 6-azobpy ligand. The analysis of 1H NMR
spectra clearly identified the structure of the complex, in which the Ru2þ ions are chelated by the bpy parts of the bridg-
ing 6-azobpy ligand. Electronic absorption and electrochemical properties are characterized by the presence of a low-
ꢀ
lying ꢀ level of the bridging ligand, manifested by a metal-to-ligand charge-transfer band extending down to
800 nm and by two consecutive one-electron reduction waves at relatively less negative potentials in cyclic voltammetry.
These properties were compared with those of previously reported positional isomers with the help of density functional
theory (DFT), time-dependent DFT, and ZINDO calculations. The analysis of the redox couple of Ru3þ=2þ showed that
the redox potentials differ slightly between the diastereomers and that there is an intermetallic electronic interaction.
Ruthenium polypyridine complexes have attracted intense
interest because of their favorable optical and electronic
Ru(II)(bpy)2
N
N
properties for light emission and electron/energy-transfer
processes.1 Multicentered oligo/polymetal complexes have
thus been prepared to obtain insight into electronic communi-
cation among metal centers within the supramolecular frame-
work. Understanding these oligo/polymetal complexes may
lay the foundation for the materialization of molecular scale
electronic/optical devices based on the rich redox/optical
properties of metal complexes.
Ru(II)(bpy)2
N
N
N
N
N
N
N
N
N
N
Ru(II)(bpy)2
Ru(II)(bpy)2
Ru(4-azobpy)Ru
Ru(5-azobpy)Ru
In multimetal complexes, the bridging ligand connecting
metal complex units plays a pivotal role in determining the
electronic communication between the metal complex units.1
A large number of bridging ligands have been developed
and tested from various points of views. Especially, ꢀ-conju-
gated bridges have attracted special attention, since the delo-
calized electronic structures in these bridges are advantageous
for mediating electronic communication between the redox/
photoactive metal complex units located at both ends. Further-
more, the structural rigidity of ꢀ-conjugated bridges affords
well-defined geometry.
N
(bpy)2Ru(II)
N
N
N
N
N
N
6
Ru(II)(bpy)2
=
Ru(II)(bpy)2
N
Ru(II)
N
5
4
N
Ru(6-azobpy)Ru
3
Chart 1.
characteristics may lead to novel properties.5 Herein, we re-
port the preparation and properties of the ligand and its dinu-
clear Ru complex, which are compared with those of 4- and 5-
azobpy ligands and their Ru complexes.
We have prepared and investigated a series of Ru and Os
complexes containing 4,400-azobis(2,20-bipyridine) (4-azobpy),2
5,500-azobis(2,20-bipyridine) (5-azobpy),3 or 40,40000-azobis-
(2,20:60,200-terpyridine) (4-azotpy)4 (Chart 1). Some of these
complexes, that is, complexes containing 4-azobpy or 4-
azotpy, behave as molecular light switches responding to re-
dox input, and a complex that contains 5-azobpy shows com-
plex-sensitized photoisomerization. We prepared a new mem-
ber of the family, i.e., 6,600-azobis(2,20-bipyridine) (6-azobpy)
in the present study. The dinuclear complex Ru(6-azobpy)Ru
is unique in the family because the metal–metal distance is
Experimental
Instruments and Measurements. 1H NMR spectra were re-
corded on a JEOL GX-400 spectrometer. Electron impact high-
resolution mass spectroscopy (EI-HRMS) was performed by the
Chemical Analysis Center of the School of Medicine, Nihon Uni-
versity. Electron impact mass spectra (EI-MS) were recorded on a
Shimadzu QP5050A mass spectrometer. Electrospray ionization
mass spectra (ESI-MS) were obtained with a JEOL JMS-600H
mass spectrometer. Elemental Analyses were performed by the
Chemical Analysis Center of the College of Science and Technol-
ogy, Nihon University.
2þ
the shortest and the surroundings of the Ru(bpy)3 core are
more sterically congested. Thus, our expectation was that these
Published on the web May 14, 2007; doi:10.1246/bcsj.80.902

J. Otsuki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
903
Absorption spectra were acquired with a Shimadzu UV-
2400PC spectrometer in a 1 ꢁ 1 cm quartz optical cell maintained
at 25 ꢂC with a Peltier thermostat. In isomerization experiments,
the light source of a Shimadzu RF-5300PC fluorimeter was used
as an excitation light, with a slit width of 20 nm.
N
SnMe3
Br
N
NO2
Pd(Ph3P)4
Cyclic voltammetry measurements were carried out with a
Hokuto Denko HZ-3000 voltammetric analyzer. The electrochem-
ical cell, maintained under N2, was equipped with a Pt disk work-
ing electrode, a Pt wire counter electrode, and an Ag/Agþ (0.01 M
AgNO3 and 0.1 M Bu4NPF6 in CH3CN) reference electrode.
These electrodes were purchased from BAS. The supporting elec-
trolyte solution was 0.1 M Bu4NPF6 in CH3CN or CH2Cl2. The
scan rate was 100 mV sꢃ1 unless otherwise noted. Potentials are
reported with respect to the internal ferrocenium/ferrocene
(Fcþ/Fc) couple added after each measurement. Digital simula-
tions of cyclic voltammograms were conducted using DigiSim,
BAS.
N
N
NO2
Zn/NaOH
N
N
N
N
N
N
1. Ru(bpy)2Cl2
2. NH4PF6
Molecular Orbital Calculations. Density functional theory
(DFT) calculations were performed using the Gaussian 03w
program package on a personal computer.6 We used the B3LYP
functional7 for all the calculations. The molecular orbital calcula-
tions for the uncomplexed ligands were carried out using the
6-31+G(d) basis set on the structures optimized by B3LYP/
6-31G(d). Time-dependent DFT calculations were performed for
the free ligands also using B3LYP/6-31+G(d). Electrical transi-
tions expected to occur at >280 nm were calculated. For the com-
plexes, Ru(4-azobpy)Ru and Ru(5-azobpy)Ru, B3LYP/LanL2DZ
was employed for geometry optimization and molecular orbital
calculations. The optimization was performed with a Ci symmetry
constraint. For the structure optimization of Ru(6-azobpy)Ru (ꢀꢀ
and meso), the basis set LanL2DZ with an effective core approx-
imation was used only for the Ru atoms, and the 3-21G set, which
has a smaller number of primitive Gaussians, was used for the oth-
er atoms. This was to save the calculation time, since the symme-
try constraint had to be lifted to find stable geometries for these
complexes. The final molecular orbital calculations on the opti-
mized structures were carried out using LanL2DZ for all atoms
to obtain consistent results with other complexes. Electronic spec-
tra for the dinuclear complexes were simulated by ZINDO8 imple-
mented in a CAChe program. The calculations were carried out
with configuration interaction level 8 and INDO/1 parameters
on the complexes optimized by DFT as described above.
Materials. The ligand, 6-azobpy, and its dinuclear Ru com-
plex, Ru(6-azobpy)Ru, were prepared according to Scheme 1.
Toluene was distilled over Na. Other solvents were used as re-
ceived.
(bpy)2Ru
N
N
N
N
N
· (PF6)4
N
Ru(bpy)2
Ru(6-azobpy)Ru
+
mononuclear complex
Scheme 1.
solution of 6-nitro-2,20-bipyridine (300 mg, 1.49 mmol) in THF
(60 mL) (CAUTION!: Zn powder might catch fire). An aqueous
NaOH solution (1 g/2.5 mL) was then added dropwise, and the
mixture was stirred at room temperature for 6 h. The reaction mix-
ture was filtered, and the solid was extracted with CH2Cl2. The
combined solution was evaporated to give a purple solid. This
was passed through an Al2O3 column with CH2Cl2 to remove
the purple material to afford an orange solid, which was crystal-
lized from benzene to give orange needles (170 mg, 0.50 mmol,
67%). 1H NMR (CDCl3): ꢁ 7.37 (2H, ddd, J ¼ 7:6 Hz, 4.8 Hz,
1.2 Hz, py-50), 7.88 (2H, td, J ¼ 7:6 Hz, 1.8 Hz, py-40), 7.96 (2H,
dd, J ¼ 7:6 Hz, 0.8 Hz, py-5), 8.08 (2H, t, J ¼ 7:6 Hz, py-4), 8.64
(2H, dd, J ¼ 7:6 Hz, 0.8 Hz, py-3), 8.67 (2H, ddd, J ¼ 7:6 Hz,
1.2 Hz, 0.8 Hz, py-30), 8.73 (2H, ddd, J ¼ 4:8 Hz, 1.2 Hz, 0.8 Hz,
py-60). EI-MS: m=z ¼ 339 (Mþ). Anal. Calcd for C20H14N6: C,
70.99; H, 4.17; N, 24.84%. Found: C, 70.91; H, 4.41; N, 24.70%.
6-Nitro-2,20-bipyridine.
2-Bromo-6-nitropyridine9 (1.69 g,
Ru(6-azobpy)Ru.
A solution of 6-azobpy (110 mg, 1.28
8.32 mmol) and [Pd(Ph3P)4] (0.96 g, 0.83 mmol) were put in a
flask, the air was then replaced by Ar using a vacuum/Ar mani-
fold, and toluene (110 mL) was added. A solution of 2-trimethyl-
stannylpyridine10 (2.04 g, 8.40 mmol) in toluene purged with Ar
was added, and the mixture was refluxed for 16 h. After the reac-
tion mixture was filtered to remove insoluble materials, the solvent
was evaporated and the residue chromatographed (SiO2, CH2Cl2)
to afford a slightly yellowish solid (1.19 g, 5.91 mmol, 71%).
1H NMR (CDCl3): ꢁ 7.40 (1H, ddd, J ¼ 8 Hz, 4.8 Hz, 0.8 Hz, py-
50), 7.89 (1H, td, J ¼ 8 Hz, 1.2 Hz, py-40), 8.15 (1H, t, J ¼ 8 Hz,
py-4), 8.25 (1H, d, J ¼ 8 Hz, py-5), 8.56 (1H, dd, J ¼ 8 Hz, 0.8
Hz, py-30), 8.71 (1H, dd, J ¼ 4:8 Hz, 1.2 Hz, py-60), 8.83 (1H,
d, J ¼ 8 Hz, py-3). EI-HRMS: calcd for C10H7N3O2: 201.0538;
found: 201.0536.
mmol) and cis-[Ru(bpy)2Cl2] 2(H2O)11 (666 mg, 0.32 mmol) in
ꢄ
EtOH (150 mL) was refluxed for 34 h under Ar. The solvent was
evaporated, and H2O was added to the residue. After removing
the insoluble materials by filtration, NH4PF6 was added, and the
mixture was stirred well. The precipitated black solid was chroma-
tographed (SiO2, CH3CN/0.4 M aqueous KNO3 (3/1)). NH4PF6
was added to the isolated fraction, and the solvent concentrated
ꢃ
to remove CH3CN and to precipitate out the PF6 salt of the de-
sired complex as a dark green solid (156 mg, 0.09 mmol, 28%). A
mononuclear complex, [Ru(6-azobpy)(bpy)2], was also obtained,
which has not been fully characterized. 1H NMR (CDCl3): see
Results and Discussion section. ES-MS: m=z ¼ 1600 (½M ꢃ
ðPF6Þꢅþ), 728 (½M ꢃ 2ðPF6Þꢅ2þ). Anal. Calcd for C60H46F24N14-
P4Ru2: C, 41.30; H, 2.66; N, 11.24%. Found: C, 41.10; H, 2.91:
N, 11.15%.
6-Azobpy. Zn powder (3.0 g) was added portionwise to a

904 Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
Azobis(bipyridine) and Its Ruthenium Complex
Fig. 2. Absorption spectra for 6-azobpy in the trans config-
uration (solid line) and in the photostationary states under
340 nm (dashed line) and 450 nm (dotted line) irradiation.
The spectra were measured in DMF (0.1 mM).
the low-intensity, low-energy absorption bands were assigned
ꢀ
to the nꢀ band. Unfortunately, calculations on the idealized
ꢀ
transition and did not provide information as to why only 4-
azobpy does not exhibit this absorption band.
planar geometry gave a zero oscillator strength for the nꢀ
On the other hand, for intense absorption bands at the short-
er wavelength region (280–420 nm), the calculations predicted
the spectral shape quite correctly, although the transition ener-
gies were a little underestimated. The lowest energy transition
ꢀ
in this spectral range is the HOMO to LUMO ꢀꢀ transition,
which was predicted to be around 380–410 nm by the calcula-
Fig. 1. Absorption spectra for azobpy’s in DMF. Vertical
bars are relative oscillator strengths according to time-de-
pendent DFT calculations. Asterisks indicate the calculat-
tion and observed around 340–370 nm. The transitions of high-
er energies around 300 nm are other ꢀꢀ transitions. The cor-
ꢀ
ꢀ
ed positions of nꢀ transitions.
rectly reproduced results in the calculations are as follows. (i)
4-Azobpy exhibits a very weak HOMO to LUMO transition,
while higher energy transitions are strong. (ii) 5-Azobpy ex-
hibits a strong HOMO to LUMO transition, while higher ener-
gy transitions are weak. (iii) 6-Azobpy exhibits a spectrum
with two peaks, in which the lower and higher energy peaks
may correspond to the HOMO to LUMO transition and other
Results and Discussion
Spectroscopic and Electronic Properties of the Ligands.
The absorption spectra for 4-, 5-, and 6-azobpy are shown in
Fig. 1. The spectra are distinctly different depending on the
substitution positions of the azo group. For 4-azobpy, only
one major peak appears at a short wavelength of 284 nm. For
5-azobpy, an intense peak appears at 369 in addition to a weak
absorption around 450 nm. For 6-azobpy, there are two rela-
tively intense peaks at 280 and 341 nm in addition to a weak
absorption around 450 nm. Time-dependent DFT calculations
were performed to shed light on the origin of these absorption
bands. We used B3LYP/6-31+G(d) for the time-dependent
DFT calculations on the structures optimized by B3LYP/
6-31G(d). The obtained transition wavelengths are indicated
by vertical bars, of which the height is proportional to the cal-
culated oscillator strength, in Fig. 1. The numerical data for
the transitions (Table S1) and pictorial illustration of HOMO
and LUMO (Fig. S1) are given in Supporting Information.
Comparison of the three spectra in Fig. 1 reveals that the
low intensity absorption around 450 nm appears for 5- and
6-azobpy’s, but not for 4-azobpy. The calculations showed
ꢀ
ꢀꢀ transitions, respectively.
Inspection of the molecular orbitals gave a hint for a very
small oscillator strength for the HOMO to LUMO transition
in 4-azobpy. As shown in Fig. S1, the LUMO of azobpy is
concentrated on the azo and inner pyridine moieties, while
the HOMO is on the bipyridines, but not on the azo moiety.
Thus, these orbitals only overlap on the inner pyridines, which
may account for the small oscillator strength. For 5-azobpy,
both of the HOMO and LUMO are delocalized over the entire
molecule. For 6-azobpy, the HOMO and LUMO are overlap-
ped each other on the inner pyridine and the azo moieties.
ꢀ
Upon excitation into the ꢀꢀ absorption band of trans-6-
ꢀ
ꢀ
azobpy, the ꢀꢀ band decreased, and the nꢀ band increased,
with isosbestic points at 314 and 405 nm, until reaching a pho-
tostationary state, as shown in Fig. 2. This is a typical trans-to-
cis photoisomerization for azobenzene and azopyridine deriv-
ꢀ
ꢀ
that there are nꢀ excited states at energy around 2.5 eV
atives.12,13 The ꢀꢀ peak decreased to become a mere should-
(ꢆ500 nm), which are shown by asterisks on the graph. Thus,
er of a shorter absorption band, implying the cis content under

J. Otsuki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
905
N
N
N
N
340 nm
450 nm
N
N
N
N
N
N
N
N
Scheme 2.
these conditions is large. Upon subsequent irradiation of the
nꢀ band at 450 nm, the absorption spectrum nearly changed
ꢀ
back to the original spectrum, which corresponds to the cis-to-
trans conversion, as shown in Scheme 2. Thus, 6-azobpy as
well as 5-azobpy3 underwent reversible photoisomerization,
while 4-azobpy did not show any spectral change upon light
irradiation.
Cyclic voltammogram for 6-azobpy in CH2Cl2 had a reduc-
tion peak at a potential similar to those of 4-2 and 5-azobpy3
ligands. The wave at ꢃ1:38 V vs Fcþ/Fc corresponds to the
first one-electron reduction. The second one-electron reduc-
tion,14,15 expected around ꢃ1:8 V may be located slightly out
of the potential window (ꢇ ꢃ2:0 V). The first reduction was re-
versible in the sense that the ratio of cathodic to anodic current
was near unity. However, the electron-transfer reaction seemed
slower than the diffusion-limited rate, since the peak separa-
Fig. 3. 1H NMR spectra for Ru(6-azobpy)Ru. (a) Ru(6-
azobpy)Ru (top) and Ru(6-azobpy)Ru with deuterated bi-
pyridines (bottom) in CD3CN. (b) Comparison of chemi-
cal shifts of ligands in CDCl3 and complexes in CD3CN.
tion is large: 300–370 mV at scan rates of 50–200 mV sꢃ1
.
Structure of Ru(6-azobpy)Ru. The ligand 6-azobpy was
reacted with the [Ru(bpy)2Cl2] fragment according to a stan-
dard procedure. This reaction afforded mononuclear and dinu-
clear complexes, from which the latter was isolated with col-
umn chromatography. The composition of the dinuclear com-
plex [{Ru(bpy)2}2(m-6-azobpy)] was confirmed unambiguous-
H-5 undergo moderate downfield shifts, while that for H-6 ex-
hibits an upfield shift as large as 1.0 ppm (see Chart 1 for pro-
ton numbering). The latter upfield shift may be attributed to a
ring current effect by the pyridine ring in the nearest bpy li-
gand. As for the 6-azobpy ligand in the dinuclear complex,
chemical shift changes of protons on the B ring, which is the
outer pyridine ring, were as expected from the results just men-
tioned for the model complex. This indicates that the chemical
environment of the B ring is close to that of [Ru(bpy)3]2þ com-
plex, consistent with a tris(bpy) type coordination. Especially,
the large upfield shift of H-B6 by 1.1 ppm gives solid support
for this type of coordination. The chemical shift changes of
protons on the A ring were similar, except for H-A5. An up-
field shift as large as 1.6 ppm was observed for H-A5. The
DFT optimized molecular model of the Ru(6-azobpy)Ru indi-
cated that this proton comes just on top of one of the auxiliary
bpy ligand, which is expected to exert a large ring current ef-
fect, as shown in Fig. 4. In the figure, protons with a particu-
larly large upfield shift and pyridine rings that cause the shift
are related by arrows. It is apparent from the model that each
of H-B6 and H-A5 lies on top of a pyridine ring. In particular,
H-A5 protons are in a van der Waals contact with the face of
the pyridine ring. Thus, this structure is fully consistent with
the 1H NMR data. Finally, it is noted that 1H NMR resonances
for the 6-azobpy ligand in Ru(6-azobpy)Ru are duplicated,
which is clearly seen particularly for H-B6 and H-A5. This
can be attributed to the diastereomers consisting of meso
(ꢁꢀ) and racemic (ꢁꢁ and ꢀꢀ) compounds. The integration
1
ly by ESI-MS and elemental analysis. However, the H NMR
spectrum, as shown in the top of Fig. 3a, is quite complicated
even with the help of two-dimensional 1H NMR techniques. A
part of the difficulty comes from the fact that the pattern of
1H NMR spectrum for Ru(6-azobpy)Ru is quite different from
those for other dinuclear complexes, since the bridging ligand
and the auxiliary ligands are mutually closer due to geometri-
cal reasons. Furthermore, caution must be taken in the structur-
al characterization of a complex containing the 6-azobpy li-
gand, because it is, in principle, possible that a metal ion is
chelated by nitrogen atoms from the pyridine ring and from
the azo group.
Therefore, we prepared its analogue having deuterated bi-
pyridine ligands by reacting 6-azobpy and [Ru(bpy-d8)2Cl2]
1
1
to simplify the H NMR spectra. The H NMR spectrum thus
obtained is shown in the lower part of Fig. 3a. In this spec-
trum, only protons on the 6-azobpy ligand are observed. To
see whether these chemical shift values are reasonable for
the structure shown in Chart 1, in which the two Ru2þ ions
are chelated by the two bpy fragments, the chemical shifts of
the complex in CD3CN were compared to those of the ligand
in CDCl3, as shown in the lower part of Fig. 3b. Also shown in
the figure (the top of Fig. 3b) are the chemical shifts of bpy in
CDCl3 and those of [Ru(bpy)3]2þ in CD3CN. Upon the forma-
tion of the Ru complex, the chemical shifts for H-3, H-4, and

906 Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
Azobis(bipyridine) and Its Ruthenium Complex
values indicate that the as-prepared sample contained the iso-
mers in a ratio of 6:4.
ed spectra, the transitions above (lower in energy) and below
(higher in energy) 360 nm corresponded to MLCT: d(Ru) !
ꢀ
tense transitions for the lower energy MLCT are characteristic
ꢀ
Spectroscopic and Electronic Properties of the Ru Dinu-
clear Complexes. The electronic absorption spectra for the
series of Ru(azobpy)Ru are shown in Fig. 5a, and the numer-
ical data are summarized in Table 1. Bands around 450 nm
were common for all of the complexes. From the analogy to
the parent complex [Ru(bpy)3]2þ,1 they may be assigned as
the metal-to-ligand charge transfer (MLCT) bands, in which
ꢀ (azobpy) and MLCT: d(Ru) ! ꢀ (bpy), respectively. In-
ꢀ
charge transfer, d(Ru) ! ꢀ (bpy), occurs on excitation. The
most distinctive features in this class of complexes were the
additional MLCT bands located to the lower energy side,
which extend down to 800 nm. These may also be assigned
as MLCT transitions, which are represented as d(Ru) !
ꢀ
ꢀ (azobpy). These assignments are supported by theoretical
calculations as described below.
The structure optimization and molecular orbital calcula-
tions were carried out by DFT calculations. The pictorial rep-
resentations for the HOMO and LUMO orbitals obtained by
B3LYP/LanL2D2 are presented in Fig. S2. For all of these
complexes, the HOMO was comprised of Ru dꢀ orbitals,
while the LUMO lies on the azobpy bridging ligand. Absorp-
tion spectra were simulated by using the ZINDO method. We
chose ZINDO, because it has been previously shown that it
gives reliable results for Ru imine complexes.16,17 The results
of spectral simulation are shown in Fig. 5b. In the graph, an
arbitrarily chosen linewidth of 3200 cmꢃ1 was allotted to each
transition. As can be seen from the simulated spectra, the spec-
tral shape was reproduced quite well, although the transition
energies were overestimated by ca. 7000 cmꢃ1. In the simulat-
Fig. 5. Absorption spectra for dinuclear azobpy complexes.
(a) Experimental spectra measured in CH3CN. Dotted
line: Ru(4-azobpy)Ru; dashed line: Ru(5-azobpy)Ru; solid
line: Ru(6-azobpy)Ru. (b) Absorptivity as calculated by
ZINDO. The linewidth of transitions were arbitrarily set
at 3200 cmꢃ1. Dotted line: Ru(4-azobpy)Ru; dashed line:
Ru(5-azobpy)Ru; solid line: ꢀꢀ-Ru(6-azobpy)Ru; meso-
Ru(6-azobpy)Ru: long dashed line.
Fig. 4. DFT optimized structure of meso-Ru(6-azobpy)Ru.
The protons on 6-azobpy that undergo a significant upfield
shift and the pyridine rings that cause the shift by the ring
current effect are related by the arrows.
Table 1. Absorption and Redox Potentials of Dinuclear Complexes Bridged by Azobpy Ligandsa)
Eꢂ0/V vs Fcþ/Fc
ꢂ
max/nm (" ꢁ 10ꢃ3/cmꢃ1 Mꢃ1
)
MLCT
LCb)
MLCT
(bpy)
MLCT
(azo)
bpy0=ꢃ1
azoꢃ1=ꢃ2
azo0=ꢃ1
Ru3þ=2þ
(ꢁEp/mV)c)
Ru(4-azobpy)Rud)
Ru(5-azobpy)Rue)
245 (55)
244 (39)
286 (140)
286 (99)
371 (39)
286 (137)
365 (23)
286 (86)
439 (22)
440 (19)
558 (22)
520 (9)
ꢆ ꢃ1:8
ꢆ ꢃ1:8
ꢃ1:45
ꢃ1:15
ꢃ0:92
ꢃ0:77
0.95(82, 83)
0.96(90, 81)
Ru(6-azobpy)Ru
[Ru(bpy)3]2þ
242 (60)
244 (29)
449 (23)
450 (14)
ꢈ504 (5)
ꢆ ꢃ2:0
ꢃ1:73
ꢃ1:29
ꢃ0:68
1.06(127, 70)
1.01
a) In CH3CN. b) Ligand centered. c) The first and second figures in parentheses are peak separations between the anodic and
cathodic peaks for the Ru3þ=2þ couple and the internal Fcþ/Fc couple, respectively. d) Ref. 2. e) Ref. 3.

J. Otsuki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
907
wave contains more than one reversible couples overlapped
with slightly shifted potentials.
Let us summarize relevant reactions before discussing more
about the large peak separation. The electrode reactions are
represented as:
RuIIRuII ꢀ RuIIRuIII þ eꢃ;
RuIIRuIII ꢀ RuIIIRuIII þ eꢃ;
ð1Þ
ð2Þ
wherein a simplified notation is used, e.g., RuIIRuII stands for
RuII(6-azobpy)RuII. In addition, there is a conproportionation
equilibrium in solution:
Fig. 6. Cyclic voltammogram for Ru(6-azobpy)Ru with
internal ferrocene in CH3CN.
RuIIRuII þ RuIIIRuIII ꢀ 2RuIIRuIII;
ð3Þ
ð4Þ
2
½RuIIRuIIIꢅ
Kc ¼
to Ru(4-azobpy)Ru both in experiment and in theory. For
Ru(5-azobpy)Ru and Ru(6-azobpy)Ru (both ꢀꢀ and meso),
the oscillator strengths for the lower energy transition were
smaller, reproducing correctly the trend observed experimen-
tally. All complexes exhibited transitions for the higher energy
½RuIIRuIIꢅ½RuIIIRuIIIꢅ
where Kc is the conproportionation constant. The conpropor-
tionation constant is related to the redox potentials for the re-
actions 1 (E1) and 2 (E2) by:
ꢀ
ꢀ
ꢁ
MLCT, i.e., d(Ru) ! ꢀ (bpy), in similar manners in terms of
energy (observed at ca. 450 nm; calculated at ca. 320 nm) and
intensity.
F
Kc ¼ exp
ðE2 ꢃ E1Þ ;
ð5Þ
RT
These dinuclear complexes, including Ru(6-azobpy)Ru, are
hardly luminescent. Reduction of the 6-azobpy ligand did
not lead to enhanced luminescence. Previously reported
Ru(5-azobpy)Ru also does not show any redox responsive lu-
minescence behavior.3 The behavior is in contrast to that of
Ru(4-azobpy)Ru, which shows a large luminescence enhance-
ment upon reduction of the bridging ligand.2
Redox Properties of Ru(6-azobpy)Ru. Figure 6 shows
the cyclic voltammogram for Ru(6-azobpy)Ru. The complex
has redox centers at the metal ions, 6-azobpy, and bpy. On
the cathodic scan, two waves corresponding to consecutive
one-electron reductions of the bridging ligand appeared at
ꢃ0:68 and ꢃ1:29 V, giving Ru(6-azobpyꢉꢃ)Ru and Ru(6-
azobpy2ꢃ)Ru, respectively. These reduced 6,600-azobpy spe-
cies in the complex are more stable when compared to the iso-
lated ligand by 0.7 eV by the presence of positive charges of
the pair of Ru2þ ions. The waves not only appeared at less neg-
ative potentials, but the reversibility of the waves improved as
compared to the isolated ligand. Especially, the first reduction
wave was completely reversible as far as the peak separation
indicates under scan rates up to 200 mV sꢃ1. Waves for the re-
duction of the terminal bpy ligands were obscured by a spike
peak probably due to adsorption/desorption of multiply re-
duced species to the electrode.
The Ru2þ/Ru3þ redox couple for Ru(6-azobpy)Ru appeared
at 1.06 V. The peak separation of the anodic and cathodic
peaks (ꢁEp) in this couple was 130 mV at a scan rate of
100 mV sꢃ1, distinctly larger than those of other complexes
as shown in Table 1. This is also much larger than the value
of ꢆ60 mV expected for a reversible electrode reaction for
multiple noninteracting equivalent redox centers.18 This can
be, in principle, due to a slow electrode reaction and/or over-
lapping redox couples at slightly differing potentials. The peak
separation was invariant at 127 ꢊ 3 mV under scan rates in the
range of 10–100 mV sꢃ1. This result indicated that the redox
reaction is fast to the extent that the electrode reaction is dif-
fusion limited. It leaves only the possibility that the redox
where F, R, and T are the Faraday constant, the gas constant,
and the absolute temperature, respectively. The conproportio-
nation constant and the difference in the redox potentials
are measures of the electronic interaction between the redox
centers.
There are two possible reasons for the redox couple to split.
One is that the splitting is due to electronic interactions be-
tween the two metallic centers. In this case, the equilibrium
is shifted to the right-hand side in Eq. 3, giving rise to Kc larg-
er than the noninteracting statistical value of 4. It was found
by digital simulation that the difference between E1 and E2
was 77 mV, which corresponds to Kc ¼ 21, accounting for the
127 mV peak separation. The other possibility is that the redox
potentials between the meso and racemic diastereomers are
different. This situation corresponds to the presence of two sets
of equilibria represented by Eqs. 1–5, with E1meso and E2meso
values being different from E1racemic and E2racemic values, respec-
tively.
We found that the content of each isomer varies crop by
crop in recrystallization. Needlelike crystals were obtained
by diffusing CH2Cl2 vapor into a CH3CN solution of Ru(6-
azobpy)Ru, which contains both meso and racemic diaster-
eomers. In our two different runs, we obtained samples con-
sisting of isomers in ratios of 72:28 and 46:54, which are la-
beled A and B, respectively. These samples were subjected
to cyclic voltammetry to investigate the effect of the composi-
tion change on the redox waves. If the large peak separation is
exclusively due to an intermetallic interaction, the composition
change should have no effect on the redox waves. If, on the
other hand, the redox potentials are different between the dia-
stereomers, the composition change should affect the redox
waves. On going from sample A to sample B, the midpoint
of the anodic and cathodic peaks in the Ru2þ/Ru3þ redox
wave showed a small but measurable positive shift of 6 mV.
That the composition change leads to shifts in redox peaks in-
dicates that the redox potentials in the meso and racemic dia-
stereomers are indeed different.

908 Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
Azobis(bipyridine) and Its Ruthenium Complex
The problem now is whether the different redox potentials
can be the sole reason for the large peak separation. To address
this issue, digital simulations were carried out according to two
sets of Eqs. 1–5 to see how peak positions change depending
on the isomeric composition, assuming that there is no inter-
metallic electronic interaction (Kcmeso ¼ Kcracemic ¼ 4). In the
simulation, the difference in the redox potentials between the
We thank I. Yoshikawa and Dr. T. Akasaka of the Univer-
sity of Tokyo for the ESI-MS measurements and help in spec-
troelectrochemistry, respectively. This work was partially sup-
ported by the ‘‘High-Tech Research Center’’ Project for Private
Universities: matching fund subsidy from the Ministry of
Education, Culture, Sports, Science and Technology, Japan.
Supporting Information
diastereomers was set as E1meso ꢃ E1racemic ¼ E2meso ꢃ E2racemic
¼
Absorption data and pictorial representation of molecular orbi-
tals in PDF format. This material is available free of charge on the
Web at: http://www.csj.jp/journals/bcsj/.
ꢊ77 mV to best account for the peak separation. On going
from the isomeric ratio of 72:28 to 46:54, the shape of the re-
dox waves changed, and the midpoint of the anodic and catho-
dic peaks shifted by 22 mV. From the magnitude of the ob-
served shifts, which were less than a third of what would be
expected for the limiting case simulated above, it was conclud-
ed that the different redox potentials in the meso and racemic
diastereomers can only explain a minor portion of the large
peak separation. The remaining major part of the large peak
separation then must be ascribed to an intermetallic electronic
interaction. It is difficult to evaluate the conproportionation
constant exactly, but it can be assumed that the value is some-
what smaller than the upper limit value of 21.19
References
1
For reviews on Ru polypyridine complexes, see: a) A.
Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser,
A. von Zelewsky, Coord. Chem. Rev. 1988, 84, 85. b) K.
Kalyanasundaram, Photochemistry of Polypyridine and Porphyrin
Complexes, Academic, London, 1992. c) L. De Cola, P. Belser,
Coord. Chem. Rev. 1998, 177, 301. d) C. Kaes, A. Katz, M. W.
Hosseini, Chem. Rev. 2000, 100, 3553. e) F. Barigelletti, L.
Flamigni, Chem. Soc. Rev. 2000, 29, 1. f) H. Hofmeier, U. S.
Schubert, Chem. Soc. Rev. 2004, 33, 373.
The reason for the detection of an intermetallic interaction
only for Ru(6-azobpy)Ru in the series listed in Table 1 may
be that the metal–metal distance (rMM) in Ru(6-azobpy)Ru is
the shortest among them. The observed (X-ray) or estimated
2
a) J. Otsuki, D.-M. Li, K. Sato, A. Nakagome, T. Takido, I.
Yoshikawa, T. Akasaka, K. Araki, Bull. Chem. Soc. Jpn. 2003,
76, 1185. b) J. Otsuki, M. Tsujino, T. Iizaki, K. Araki, M. Seno,
K. Takatera, T. Watanabe, J. Am. Chem. Soc. 1997, 119, 7895.
c) J. Otsuki, K. Sato, M. Tsujino, N. Okuda, K. Araki, M. Seno,
Chem. Lett. 1996, 847.
˚
(DFT) rMM values are as follows: 13.4 A for Ru(4-azobpy)Ru
˚
(DFT), 10.0–11.8 A for Ru(5-azobpy)Ru, depending on the
3
˚
conformation (X-ray ), and 8.2 A for Ru(6-azobpy)Ru (DFT).
3
J. Otsuki, N. Omokawa, K. Yoshiba, I. Yoshikawa, T.
Relevant values for related metal complexes were searched
in the literature for comparison purposes. There are only a
few reports on the metal–metal interactions across bpy–
spacer–bpy type bridging ligands. Regarding the peak separa-
tion, ꢁEp ¼ 100, 90, and 90 mV have been reported for di-
nuclear Ru 2,20-bipyridine complexes connected through a
carbon–carbon triple bond at 4-, 5-, and 6-positions, respec-
tively.20 For another analogue, in which the spacer is an ethyl-
enic double bond, ꢁEp ¼ 90 mV and Kc ¼ 15 have been re-
ported.21
Akasaka, T. Suenobu, T. Takido, K. Araki, S. Fukuzumi, Inorg.
Chem. 2003, 42, 3057.
4
a) T. Akasaka, J. Otsuki, K. Araki, Chem. Eur. J. 2002, 8,
130. b) T. Akasaka, T. Mutai, J. Otsuki, K. Araki, Dalton Trans.
2003, 1537.
5
Another aspect of our recent interest concerning oligo/
polymetallic complexes is their ability to bind DNA double heli-
ces. We needed a set of dinuclear complexes systematically varied
in structure. Studies on interactions between the complexes in
Chart 1 and DNA double helices are in progress. See the following
references: a) C. Metcalfe, J. A. Thomas, Chem. Soc. Rev. 2003,
32, 215. b) J. A. Smith, J. G. Collins, B. T. Patterson, F. R. Keene,
Dalton Trans. 2004, 1277. c) C. Rajput, R. Rutkaite, L. Swanson,
I. Haq, J. A. Thomas, Chem. Eur. J. 2006, 12, 4611.
Conclusion
We have prepared a bis-chelating ligand, 6-azobpy, as a
new member of the azobpy family. In this family of 4-,
5-, and 6-azobpy’s, the latter two ligands were found to
be photochromic. The dinuclear Ru complex [Ru2(bpy)4(6-
6
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi,
V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P.
Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez,
J. A. Pople, Gaussian 03, Revision C.02, Gaussian, Inc., Walling-
1
azobpy)]4þ was prepared. The H NMR signals derived from
the 6-azobpy part were extracted by using deuterated bpy aux-
iliary ligands. The simplified spectrum proved the structure of
the complex as the tris(bipyridine) type dinuclear complex
bridged by the 6-azobpy ligand. Absorption and electrochem-
ical properties were compared with those of previously report-
ed positional isomers, with the help of theoretical calculations.
From the analysis of Ru2þ=3þ redox couple in cyclic voltam-
metry for Ru(6-azobpy)Ru, it was found that the meso and rac-
emic compounds exhibit slightly different redox potentials. In
addition, a metal–metal electronic interaction was detected
with the conproportionation constant estimated to be a little
less than 21. A shorter metal–metal separation in Ru(6-
azobpy)Ru than those in the other positional isomers may be
responsible for the intermetallic electronic interaction.

J. Otsuki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 5 (2007)
909
ford CT, 2004.
16 S. I. Gorelsky, E. S. Dodsworth, A. B. P. Lever, A. A.
Vlcek, Coord. Chem. Rev. 1998, 174, 469.
17 M. F. Ryan, R. A. Metcalfe, A. B. P. Lever, M. Haga,
J. Chem. Soc., Dalton Trans. 2000, 2357.
18 J. B. Flanagan, S. Margel, A. J. Bard, F. C. Anson, J. Am.
Chem. Soc. 1978, 100, 4248.
7
8
A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
M. C. Zerner, ZINDO Program, Quantum Theory Project,
University of Florida, Gainsesville, FL, U.S.A., 1998.
J. W. Streef, H. J. den Hertog, Recueil 1969, 88, 1391.
9
10 G. V. Long, S. E. Boyd, M. M. Harding, I. E. Buys, T. W.
Hambley, J. Chem. Soc., Dalton Trans. 1993, 3175.
11 B. P. Sullivan, D. J. Salmon, T. J. Meyer, Inorg. Chem.
1978, 17, 3334.
12 E. V. Brown, G. R. Granneman, J. Am. Chem. Soc. 1975,
97, 621.
13 H. Rau, Studies in Organic Chemistry 40: Photochromism,
19 An attempt to resolve the stepwise redox reactions with
differential pulse voltammetry was unsuccessful. Also, we were
unable to observe spectral changes upon partial oxidation of the
Ru centers that could be attributed to the intervalence charge-
transfer transition, possibly due to the overlapping absorption tail
ꢀ
of the low-energy d(Ru) ! ꢀ (6-azobpy) MLCT band.
Molecules and Systems, ed. by H. Durr, H. Bouas-Laurent,
Elsevier, Amsterdam, 1990, p. 165.
20 V. Grosshenny, A. Harriman, F. M. Romero, R. Ziessel,
J. Phys. Chem. 1996, 100, 17472.
¨
14 A. J. Bellamy, I. S. MacKirdy, C. E. Niven, J. Chem. Soc.,
Perkin Trans. 2 1983, 183.
15 J. L. Sadler, A. J. Bard, J. Am. Chem. Soc. 1968, 90, 1979.
21 G. F. Strouse, J. R. Schoonover, R. Duesing, S. Boyde,
W. E. Jones, Jr., T. J. Meyer, Inorg. Chem. 1995, 34, 473.