Role of ruthenium oxidation states in ligand-to-ligand charge transfer processes

Article abstract of DOI:10.1021/ic200966f

We describe in this paper the properties of [Ru^II/III(bpy) ₂ClL]^+1/+2 and [Ru^II/III(bpy)₂L ₂]^+2/+3. L = ditolyl-3-pyridylamine (dt3pya) is a redox active ligand related to triary

Full text of DOI:10.1021/ic200966f

Article
pubs.acs.org/IC
Role of Ruthenium Oxidation States in Ligand-to-Ligand Charge
Transfer Processes
Cristina L. Ramírez,† Cesar N. Pegoraro,† Oscar Filevich,‡ Andrea Bruttomeso,§ Roberto Etchenique,‡
́
and Alejandro R. Parise*,†
†Departamento de Química, FCEN−Universidad Nacional de Mar del Plata, Funes 3350, 7600 Mar del Plata, Argentina
§
‡INQUIMAE−Departamento de Química Inorgan
́
ica, Analítica y Fisicoquímica and Departamento de Química Organ
́
ica,
FCEN−Universidad de Buenos Aires, Ciudad Universitaria, 1428 Buenos Aires, Argentina
S
* Supporting Information
ABSTRACT: We describe in this paper the properties of
[RuII/III(bpy)2ClL]+1/+2 and [RuII/III(bpy)2L2]+2/+3. L = ditolyl-
3-pyridylamine (dt3pya) is a redox active ligand related to
triarylamines, which is very similar to 3-aminopyridine except
for the reversible redox behavior. The monosubstituted
complex shows a metal-to-ligand charge-transfer (MLCT) at
502 nm, and reversible waves in acetonitrile at E0(RuIII/II) =
1.07 V, E0(L+/0) = 1.46 V (NHE). The disubstituted complex
shows an MLCT at 461 nm, a photorelease of dt3pya with quantum yield of 0.11 at 473 nm, and two reversible one-electron
overlapped waves at 1.39 V associated with one of the ligands (1.37 V) and RuIII/II (1.41 V). Further oxidation of the second
ligand at 1.80 V forms a 2,2′-bipiridine derivative, in an irreversible reaction similar to dimerization of triphenylamine to yield
tetraphenylbenzidine. In the dioxidized state, the spectroelectrochemistry of the disubstituted complex shows a ligand-to-ligand
charge transfer at 1425 nm, with a transition moment of 1.25 Å and an effective two-state coupling of 1200 cm−1. No charge
transfer between ligands was observed when Ru was in a 2+ oxidation state. We propose that a superexchange process would be
involved in ligand−metal−ligand charge transfer, when ligands and metals are engaged in complementary π interactions, as in
metal−ligand−metal complexes. Best orbital matching occurs when metallic donor fragments are combined with acceptor ligands
and vice versa. In our case, RuIII bridge (an acceptor) and two dt3pya (donors, one of them being oxidized) made the complex a
Robin−Day Class II system, while the RuII bridge (a donor, reduced) was not able to couple two dt3pya (also donors, one
oxidized).
electron transfer literature.15 The set of compounds studied so
INTRODUCTION
■
far is small because most redox-active ligands are also involved
in coupled irreversible chemical reactions that hinder molecular
interrogation (usually formation of multiple N−C bonds
accompanied by proton transfer after the electrochemical
step16). The analysis of these systems focused on ligand charge
transfer should not differ essentially from the classical metal−
ligand−metal system like the Creutz-Taube cation, because the
underlying electron transfer theory is independent of the nature
of the fragments involved in redox reactions.
Mixed-valence compounds have played a central role in the
understanding of how electron transfer occurs in a variety of
systems whose properties (color, electrical conductivity,
magnetism) derive from this phenomenon. This topic has
been systematized in a series of reviews,1−10 and its influence
on the field of electron transfer has been so strong that we are
currently able to trace applications of the intervalence concept
in fields so apparently unconnected as magnetoresistance,11
energy conversion,12 and molecular electronics.13 The most
important strength of intervalence compounds, such as the
Creutz-Taube cation, is that they allow us to spectroscopically
interrogate the system that actually performs the electron
transfer event. This observation is not possible while studying
homogeneous redox reactions like solution self-exchange
It has been proposed15 that a pure ligand−ligand charge
transfer (a concept analogous to a metal−metal intervalence
charge transfer) must have no influence of the central metal. In
this extreme situation, the metal locates the ligands in such a
way that direct orbital overlapping results in electronic
coupling. In other words, the metal has a structural connectivity
role but not an electronic (orbital) role. On the basis of the
experience of the historically metal-focused mixed valence
chemistry, this extreme situation of negligible coupling between
adjacent metal−ligand pairs and non-negligible coupling
Fe(H2O)62+/Fe(H2O)6 (a classic example), because reacting
ions are nonassociated most of the time and spectroscopies are
usually not sensitive enough to allow studying the small
concentration of associated reactants.
Although ligands are usually organic molecules, and organic
mixed valence systems have made important contributions to
this field,14 ligand-focused charge transfer and ligand−metal−
ligand mixed valence compounds are notoriously scarce in
3+
Received: May 9, 2011
Published: January 6, 2012
© 2012 American Chemical Society
1261
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
between remote metals seems to be rather the exception than
the rule. Even in metal localized mixed-valence compounds like
[Ru(NH3)5-4,4′-bipyridine-Ru(NH3)5]5+ there is always some
kind of interaction with the bridge, reflected in the appearance
of a metal-to-ligand charge-transfer (MLCT) beside a metal-to-
metal charge transfer (MMCT). In these cases, the interactions
between adjacent sites are responsible for an electron transfer
process involving remote metals, without a direct metal−metal
−Abs
T
dn /dt = I
(1 − 10
)Abs /Abs φ
R T
p
beam
where np are the moles of released ligand, Ibeam is the intensity of the
incident light in Einsteins/s, AbsT is the total absorbance of the
solution at the beam’s wavelength, AbsR is the absorbance of Ru-
(dt3pya)2 complex (see below) at the same wavelength, and Φ is the
photolabilization quantum yield.
Factor analysis of the spectroelectrochemistry experiment was made
as reported26 using a two successive redox steps (C + 1 e− → B, B + 1
e− → A each one with a different value of E0, with the analytical
concentration of the initial sample) as fitting model.
coupling (a superexchange bridge-mediated mechanism).
Here, we will describe the mono and dicoordination of
ditolyl-3-pyridylamine, which is a redox-active arylamine
derivative, into the fragment [RuIII/II(bpy)2]3+/2+ (bpy is 2,2′-
bipyridine). The family of coordination compounds derived
from [RuII(bpy)3]2+ has been widely studied for their
photochemical properties, especially the charge transfer
capability from the excited electronic states.17 Because of this
characteristic, the chemistry of these compounds has shown to
be attractive because of its potential application in systems for
turning visible radiation into chemical energy,18,19 a topic that
has become a priority because of the new social values
regarding the use of natural resources. In this direction, the use
of the fragment [RuII(bpy)(SCN)2]0 with bidentate ligands
derived from triarylamines (redox-active) that allows a fast in
situ regeneration of the RuIII fragment formed after charge
injection into the supporting semiconductor has been recently
reported.20 Triarylamines are a family of compounds which has
received considerable attention in recent years as hole transport
materials, because of their chemical robustness and versatility in
derivatization for controlling their redox potential.21 Since their
E0 is close to that of the [RuIII/II(bpy)x]+3/+2 series, these
fragments are interesting to be incorporated in redox
optoelectronic devices for electrical conduction.
Ditolyl-3-pyridylamine (dt3pya). One gram (5 mmol) of
ditolylamine, 1.04 g (5 mmol) of 3-iodopyridine, 0.8 g (7 mmol) of
potassium tert-butoxide, and 40 mg of a mixture of ratio 1:1 of CuI and
1−10 phenanthroline were dissolved in 10 mL of dry toluene. The
mixture was heated over 12 h, refluxing at 110−115 °C, cooled, and
then filtered to eliminate insoluble particles. The filtrate was adsorbed
on silica Merck 60 and eluted using a solvent gradient (100%
dichloromethane, dichloromethane ethyl acetate 98%/2%) in a 3 × 10
cm column. Yield: 0.9 g (65%). NMR 1H (500 MHz, see Figure 1 for
Figure 1. Ligand (ditolyl-3-pyridylamine, dt3pya) used in this study.
1
Numeration relates to H NMR characterization (see Experimental
Section).
atom numbering): δ = 8.13 (H-2, bd, J = 3.8 Hz, 1H), 7.09 (H-4, m,
1H), 7.29 (H-5, ddd, J = 1.4, 2.9, 8.4 Hz, 1H), 8.32 (H-6, bs, 1H), 6.98
(H-8, d, J = 8.2 Hz, 4H), 7.08 (H-9, d, J = 8.2 Hz, 4H), 2.34 (H-11, s,
6H). NMR 13C: δ = 142.0, 144.7, 123.5, 128.3, 144.3, 124.7, 130.2,
133.5, 20.8.
EXPERIMENTAL SECTION
■
[Ru(bpy)2Cl-dt3pya] (PF6) (Ru-dt3pya). A 100 mg portion of
[Ru(bpy)2Cl2] (0.2 mmol) was suspended in 5 mL of 96% ethanol
and 2 mL of water under argon atmosphere. Then, 55 mg of dt3pya
(0.2 mmol) and 200 mg of LiCl were added. The mixture was
protected from light and heated to 90 °C until the spectrum remained
unchanged. The solution was evaporated in darkness, redissolved in
water and precipitated by adding drops of saturated NH4PF6 solution
in water. The solid was filtered, redissolved in 3 mL of LiCl saturated
acetone and refluxed over 1 h. This solution was dropped into water
(with continuous stirring), rendering a red precipitate which was
filtered and washed several times with cold water. This procedure for
separation and storage had to be carried out in darkness because of the
photolability of dt3pya. Yield: 120 mg (70%). 1H NMR (500 MHz,
acetone-d6): δ = 2.33 (s, 6H), 6.78 (d, J = 7.6 Hz, 4H), 7.00 (d, J = 7.7
Hz, 2H), 7.04 (d, J = 7.6 Hz, 4H), 7.05 (d, J = 8.4 Hz, 1H), 7.11 (d, J
= 7.8 Hz, 1H), 7.22 (d, J = 8.7 Hz, 1H), 7.27 (s, 1H), 7.46 (d, J = 5.6
Hz, 1H), 7.51 (d, J = 4.2 Hz, 1H), 7.58 (m, 2H), 7.63 (q, J = 8.4, 8.4,
8.4 Hz, 2H), 8.06 (t, J = 7.8, 7.8 Hz, 1H), 8.00 (t, J = 7.8, 7.8 Hz, 1H),
8.11 (d, J = 8.0 Hz, 1H), 8.16 (d, J = 8.1 Hz, 1H), 8.33 (dd, J = 7.9, 2.4
Hz, 2H), 8.59 (d, J = 5.4 Hz, 1H), 9.85 (d, J = 4.9 Hz, 1H). ESI-MS:
m/z calcd for C39H34ClN6Ru ([Ru(bpy)2ClL]1+): 723.1577; found:
723.1578.
Reagents and solvents were purchased from Sigma-Aldrich, Anedra,
Dorwill, and Cicarelli. Solvents were purified as described elsewhere.22
[Ru(bpy)2Cl2]0 and [Ru(bpy)2(3-aminopyridine)2](PF6)2 were syn-
thesized as previously described.23,24 Electrochemical experiments
were measured in an anaerobic three-electrode cell with a silver wire as
a pseudoreference, a 50 μ platinum disk as working electrode, and a
platinum wire as a counterelectrode. Redox potentials were reported
against NHE (ferrocene was added as internal standard after
measurements25). Spectroelectrochemistry experiments were per-
formed with a four-electrode quartz thin layer cell using a graphite
felt counterelectrode, platinum woven wire working electrode (70 μ
wires), a silver pseudoreference and an additional small platinum
electrode to measure cyclic voltammograms prior to the electrolysis.
Electronic spectra were measured in a Shimadzu spectrophotometer
UV-3101PC, electrochemistry experiments with a potentiostat TEQ-3,
NMR spectra in a Bruker AM-500, and high resolution mass spectra in
a Bruker micrOTOF-QII.
The setup to measure the photolysis quantum efficiencies consisted
of three collinearly aligned lasers at 532 nm, 473 nm, and 405 nm
respectively. An Ocean Optics PC2000 diode-array spectrophotometer
was mounted normally to the laser light path as in a fluorescence
setup. A quartz cuvette was at the center of the arrangement in an
aluminum holder. This was thermostatized with a peltier element, an
LM35 temperature sensor, and custom feedback electronics. A
magnetic stirrer provided solution homogeneity. Laser power was
measured by directly hitting a light power meter and found to be
constant throughout the experiments for all three lasers. Full spectra
were acquired every 2 s with OOIChem software. Data fitting were
carried out offline by numerical integration of the following
expression:
[Ru(bpy)2(dt3pya)2](PF6)2 (Ru-(dt3pya)2). A 100 mg portion of
[Ru(bpy)2Cl2] (0.2 mmol) was suspended in 6 mL of water and
heated to 90 °C under argon atmosphere. Labilization of Cl− to form
[RuII(bpy)2(H2O)2]2+ (λmax = 480 nm) was completed after 1 h. Then,
110 mg (0.4 mmol) of dt3pya dissolved in 1.5 mL of 96% ethanol was
added. The resulting suspension was heated with stirring at 90 °C for
12 h in darkness, and kept in an argon atmosphere until full
displacement of water by dt3pya. The orange solution was filtered to
1262
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
Table 1. MLCT Maxima, Redox Potential and Photolabilization Quantum Yield for cis-[RuII(bpy)2XY] Complexes
a
a
a
b
+/0
+/0
X−Y ligand
λ MLCT (nm)
RuIII/II (V)
L1 (V)
L2 (V)
Φ
c
c
Cl2
553, 380
0.53
d
d
d
Cl− L
L2
502
1.07
1.46
d
d
e
d
e
d
d
461
1.39 (1.37 )
1.39 (1.41 )
1.80
0.11
0.08
f
g
g
d
(3ampy)2
463
1.42
a
b
c
d
e
f
g
In acetonitrile. In acetone. Ref 26. This paper. From factor analysis. 3-ampy is 3 aminopyridine. Ref 24.
illuminated.27 Consistently with this behavior, the NMR
spectrum of irradiated solutions of Ru-(dt3pya)2 showed the
appearance of the methyl signals of the free ligand dt3pya at
2.31 ppm, as well as a decrease in the integration of the
coordinated ligand at 2.34 ppm (Supporting Information S1).
Changes in the UV−vis spectrum of the solution during
irradiation are shown in Figure 2, from which we fit a value of ϕ
= 0.11 (quantum yield) for dt3pya release in Ru-(dt3pya)2.
remove insoluble materials, and the filtrate was precipitated by adding
drops of saturated NH4PF6 solution. The suspension was filtered (or
centrifuged) and washed several times with cold water. The procedure
for separation and storage had to be carried out in the dark because of
1
the photolability of dt3pya. Yield: 200 mg (80%). H NMR (500
MHz, acetone-d6) δ = 2,34 (s, 12H), 6.94 (m, 8H), 7.21 (m, 10H),
7.31 (m, 2H), 7.38 (dd, J = 8.6, 2.5 Hz, 2H), 7.68 (d, J = 2.5 Hz, 2H),
7.78 (d, J = 5.0, 2H), 7.98 (d, J = 5.0, 2H), 7.98 (td, J = 6.8, 1.4 Hz,
2H), 8.05 (td, J = 8.0, 1.4 Hz, 2H), 8.34 (td, J = 8.0, 1.4 Hz, 2H), 8.58
(d, J = 8.0 Hz, 2H), 8.69 (d, J = 8.0 Hz, 2H), 8.89 (d, J = 5.0 Hz, 2H).
ESI-MS: m/z calcd for C58H52N8Ru ([Ru(bpy)2L2]2+): 481.1679;
found: 481.1666.
RESULTS AND DISCUSSION
■
Figure 1 shows the ligand dt3pya (ditolyl-3-pyridylamine) used
1
in this study. The H NMR spectrum of Ru-dt3pya showed a
complex set of signals (34 protons in 19 signals), reflecting the
asymmetry of substitution at the cis positions (Cl− and dt3pya).
In contrast, the spectrum of Ru-(dt3pya)2 showed 52 protons
in only 14 signals, corresponding to a more symmetrical
compound. This is consistent with the rotation of coordinated
dt3pya which, for NMR spectroscopy time scales, dynamically
averages the field seen by other protons and presents both cis
substitution sites as indistinguishable. A similar effect was
already observed in mono and disubstituted analogues with 4-
aminopyridine.27 A preliminary observation of possible
Figure 2. UV−vis spectra of Ru-(dt3pya)2 in acetone during
irradiation at 473 nm. Inset: moles of released ligand vs time.
28
coordinated dt3pya movements in Ru-(dt3pya)2 shows that
the rotation of the ligand is possible, but needs to be concerted
with the other ligand to generate enough space to allow full
movement (especially regarding the independent rotation of
the pyridine).29
The quantum efficiencies for labilization depend on the
stationary population of the eg dissociative state during the de-
excitation pathway following irradiation. This state increases its
participation in the decay mechanism as the π acceptor ability
of ligands moves the MLCT state (and the triplet emitter state)
above eg.31 In practice, this means that π acceptor ligands
(which blue-shift the energy of the MLCT RuII → bpy*)
propitiate photolabilization. By contrast, π donor ligands (CN−,
SCN−, for example) increase the eg state energy above the
triplet emitter state, which makes these complexes substitu-
tionally photoinert, suitable for practical uses in charge transfer
devices. An example of this is the reported electron injection
into TiO2 from the bpy centered excited state of the
[RuII(bpy)(SCN)2]0 fragment.20 In this system, the ruthenium
center is bound to an arylamine-like ligand that plays the role of
a temporary sacrificial electron-donating reagent. The techno-
logical uses of this kind of compounds involve achieving a
device with components in a spatial organization (usually layers
in a high surface electrode) so that the movement of electrons
generated from light absorption has a defined direction and
sense.
Table 1 shows a comparison of basic information between
[Ru(bpy)2Cl2], Ru-dt3pya, and Ru-(dt3pya)2 in terms of
MLCT (acetonitrile), photolabilization quantum yield (ace-
tone), and redox potential (acetonitrile), together with
[Ru(bpy)2(3-aminopyridine)2]2+. Electronic spectroscopy
showed that MLCT RuII → bpy* charge transfer shifted to
higher energies as Cl− was replaced by dt3pya (553, 502, and
461 nm for Cl2, Cl-dt3pya, and dt3pya2, respectively). This is a
common motif for acceptor ligands because of the stabilization
of RuII t2g orbitals after mixing with π* of the ligand,30 and it is
also reflected in a redox potential shift to oxidant values (0.53,
1.07, 1.41 V for the same complexes). Comparsion of Ru-
(dt3pya)2 with [Ru(bpy)2(3-aminopyridine)2]2+ reveals a very
similar MLCT (461 nm vs 463 nm), E0 (1.41 V vs 1.42 V, see
bellow), and photolabilization quantum yield (0.11 vs 0.08, see
bellow) that suggests conjugation with phenyls has little effect
on the π donor and π acceptor capabilities of dt3pya compared
with unsubstituted 3-aminopyridine.
The chemistry of [RuII(bpy)2L2]2+ (L = aliphatic or aromatic
amine) is characterized by the photolabilization of ligand L,31
which is responsible for the need to avoid light during the
synthesis of these compounds. This property has made them
promising as scheduled-delivery neurotransmitter agents in
living tissues, with verified physiological activity only when
The cyclic (CV) and square wave (SWV) voltammograms of
Ru-dt3pya in Figures 3a−b show two peaks associated to bpy
coligands reduction at −1.19 V and −1.42 V. At oxidizing
potentials Ru-dt3pya showed two independent one-electron
redox waves at 1.07 and 1.46 V (same current as reduction
waves). Since the spectroelectrochemistry of Ru-dt3pya at
1263
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
waves were found at oxidant potentials. One appears at 1.39 V,
and it shows in the CV a current peak twice the intensity than
each bpy monoelectronic reduction. Since the diffusion
coefficient for the initial specie in the voltammogram is the
same for both current waves (reduction and oxidation), we
infer that the wave at 1.39 V involves two electrons. The second
wave in Ru-(dt3pya)2 appears at 1.80 V and was not
reversible.35
Since we have three redox centers in this complex and the
wave at 1.39 V seems to be a bielectronic one, we need to
assign which of the redox centers were first oxidized to
understand the intervalence properties that we will see below.
Only two scenarios are expected for the dioxidized specie: in
scenario “a” one ligand and the metal are oxidized, in scenario
“b” both ligands are oxidized.
Supporting scenario “a”, from Table 1 we see that dt3pya
oxidation in Ru-dt3pya appears at 1.46 V and Ru oxidation in
[Ru(bpy)2(3ampy)2]2+ appears at 1.42 V, both values very close
to the two electron wave at 1.39 V. This comparison suggests
that one ligand and the ruthenium center would be the first
oxidized fragments.
On the other hand, it is established that the contributions of
the ligands to the redox potential of RuIII/II center can be
factored into σ and π interactions with the metal, and are
approximately additive.36 This is verified, for example, in
successive replacements of Cl− by pyridine31 in the following
series: [Ru(bpy)2Cl2] (0.53 V), [Ru (bpy)2ClPy]+ (1.02 V),
and [Ru(bpy)2(Py)2]2+ (1.53 V). Since the redox potential shift
between [Ru(bpy)2Cl2]0 and Ru-dt3pya was ≈0.5 V, we
estimate that the resulting shift when replacing both Cl− ligands
by two dt3pya should locate the redox couple RuIII/II at about
1.61 V, which falls halfway between the first oxidation wave at
1.39 V and the second one at 1.80 V. This estimation is
consistent with the simultaneous oxidation of both dt3pya
(scenario “b”) at 1.39 V and the Ru oxidation at a potential
higher than 1.61 V (1.80 V), because it would be coordinated
to two positive (then, more acceptor after oxidation) ligands.
Next, we look at the spectroscopic information.
Figure 3. Voltammograms of Ru-dt3pya in acetonitrile 0.1 M
TBAPF6: (a) CV (100 mV/s), (b) SWV (25 Hz, 2 mV/step).
Figure 4 showed that a decay of the MLCT RuII → bpy is
associated with the first wave at 1.07 V, we assigned this process
Figure 6a−b shows four spectra of the spectroelectrochem-
istry experiment of Ru-(dt3pya)2 inside the 1.39 V redox wave
and the absorbance changes at 712 y 1425 nm. It is clear that
the first spectral change in this wave (spectrum b) is the
appearance of a band at 712 nm without any absorption in the
NIR.37 This band is characteristic of triarylamine radicals,
similar to the one found at 720 nm in the second oxidation of
Ru-dt3pya, and supports that the first process at the
bielectronic wave is the oxidation of the coordinated ligand,
yielding RuII-dt3pya+-dt3pya. This assignment is consistent
with both scenarios depicted above.
Figure 4. Spectroelectrochemistry of Ru-dt3pya in acetonitrile, 0.1 M
TBAPF6, from 0.9 to 1.7 V.
Following the spectroelectrochemistry experiment at 1.39 V,
the absorbance profiles in Figure 6b show that the second
oxidation occurs a few millivolts after the first one, and that it is
clearly associated with the NIR band at 1425 nm. After full
development of this band, the spectrum starts changing in a
way that depends not only on potential but mainly of time. We
associate this process to the irreversibility of the wave at 1.80 V,
and it will be discussed below.
to a RuII oxidation. The later appearance of a band at 720 nm,
similar to that of radicals derived from tritolylamine,32 suggests
(consistently) that the second redox process at 1.46 V
corresponds to the oxidation of a redox-active ligand dt3pya.
The voltammograms of Ru-(dt3pya)2 in Figure 5a−b
showed two bpy reductions at −1.06 V and −1.25 V. The
higher reduction potential compared to Ru-dt3pya reflects the
effect that metal electron density is shared with a more
extended π system (two dt3pya for backbonding, instead of
one) and then bpy coligands are slightly less electron rich. A
similar shift has been observed for a series of reported
complexes,33 although with lower sensitivity.34 Second, two
Factor analysis with NIPALS algorithm26 was used to process
data from the spectroelectrochemistry experiment, fitting the
spectra and concentrations profiles obtained by the method to
the simplest model of two successive redox steps (see
Experimental Section). Values of 1.37 and 1.41 V for each
redox process were found, and the three spectra of pure
1264
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
Figure 5. Voltammograms of Ru-(dt3pya)2 in acetonitrile 0.1 M TBAPF6: (a) CV (500 mV/s), (b) SWV (25 Hz, 2 mV/step), (c) SWV (25 Hz, 2
mV/step) starting a 2 V downward, and (d) CV (20 mV/s).
calculated species are shown in Figure 7a−b, along with fitted
concentration profiles. The first specie is assigned to the Ru-
(dt3pya)2 initial complex, with absorption at 461 nm (MLCT
RuII → bpy*). The second specie shows absorption at 712 nm
and a band at 450 nm. The third specie shows a broad
absorption at 1425 nm and a small band at 620 nm. The
absorption at 450 nm is displaced to 430 nm and resembles the
shape of [RuIII(bpy)3]3+ system, although not conclusively.
Consistently with both scenarios regarding the assignment of
the dioxidized specie, it is clear that the first process in the two
overlapped electron waves at 1.39 V (1.37 V from the factor
analysis fitting) yields RuII-dt3pya+-dt3pya because we found a
radical absorption at 712 nm that precedes any other spectral
change. This system does not show a strong NIR intervalence
band, so we infer that the RuII center is not a good coupling
bridge to electronically communicate an intervalence system
formed by two coordinated redox-active donor ligands, one of
them being oxidized. It is remarkable that the innocuous
electronic role played by RuII in the dt3pya+-dt3pya
coordinated intervalence system contrasts with the redox
splitting effect observed with the metal in the same oxidation
state (2+) but now in the bpy−-bpy system, with a ΔE0 = 0.2 V
(see bpy reductions in Figures 3b and 5b).
Once the first process is clarified, it is in the second electron
process where scenarios “a” and “b” actually differ. One
possibility is that the second ligand gets oxidized while keeping
ruthenium at 2+ state (scenario “a”). In such a case, then,
Figure 6. Spectroelectrochemistry of Ru-(dt3pya)2, 0.1 M TBAPF6:
oxidation of both dt3pya would be separated by only 40 mV
because of the RuII effect. Arylamines mixed valence
(a) from 1.15 to 1.45 V, (b) Absorbance profiles at 712 nm (white
circles) and 1425 nm (black circles).
compounds have shown empirically smaller separations of
redox potential compared to classical metal mixed-valence
1265
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
electronic structure of this complex between the intermediate-
coupled benzidines and the full coupled ones (compounds 5a
and 4 in ref 40). For this reason, we assign the dioxidized specie
to RuIII-dt3pya+-dt3pya (scenario “b”) and, according to the
spectrum, we infer that the Ru3+ effectively couples both
arylamines. This system mediated by Ru3+ contrasts with the
isolated arylamine system bridged by Ru2+.41
The second oxidation wave in Ru-(dt3pya)2 (third electron
from the starting compound) appeared at 1.80 V in SWV and
showed to be irreversible in the CV experiment at low speed
(50 mV/s in the 50 μm electrode). It also made the first wave
at 1.39 V irreversible, which showed a reversible electro-
chemistry provided the potential did not exceed 1.80 V. This
irreversibility is clearly seen in the SWV reversing scan (from
2.00 V downward, Figure 5c) because, at decreased speeds, a
new chemical specie with its own redox activity appeared at
1.59 V. At faster scan rates (1 V/s), both the SWV and CV
showed quasi-reversible voltammograms, although not fully
reversible. At very slow sweeps (10 mV/s), as the potential in
the CV increased beyond 1.80 V, the overall current decreased
and the voltammogram returned to cathodic potentials with no
current (Figure 5d). The electrode surface could be regenerated
from this passivation by a fast cycling (5 cycles) between −2.00
and 2.00 V. This effect was related to the time spent at
oxidizing potentials rather than to the potential itself, as long as
it remained over 1.80 V.
The fact that the electrochemical irreversibility was
manifested only when the two ligands dt3pya were
simultaneously oxidized suggests that a coupled chemical
process, similar to the dimerization of triphenylamine (TPA),
occurs inside the coordination sphere.42 In the oxidized radical
TPA, a Lewis base (acetonitrile, for example) removes protons
in positions 4 to the arylamine nitrogen, letting the unpaired
electrons expose to the solution. A bimolecular process
involving two identical radicals results in the formation of a
carbon-to-carbon covalent bond to yield a tetraphenylbenzidine
derivative. In our system, this would give the ligand 5,5′-
bis(ditolylamino)-2,2′-bipyridine (Scheme 1).
Figure 7. Information from factor analysis using Nipals algoritm (see
text): (a) Spectra of RuII-(dt3pya)2 (green), RuII-dt3pya+-dt3pya
(red, absorbance reduced by a factor of 2) and RuIII-dt3pya+-dt3pya
(blue), (b) Relative concentration profiles vs potential for the same
species (same colors).
compounds38,39 but, even for small separations, a noticeable
absorption in the NIR is always observed for the single oxidized
mixed valence system. Moreover, no intervalence band would
be observed for the doubly oxidized system. Any of those
observations are verified in the spectra of Figure 6a, so we
conclude that scenario “a” is not possible.
Scheme 1. Oxidation Scheme of Coordinated Ligands
beyond 1.80 V
The other possibility (scenario “b”) is that ruthenium
oxidizes at 1.41 V to 3+ state, so the dioxidized specie would
be RuIII-dt3pya+-dt3pya. This would make the spectrum with
bands at 1425 and 620 nm the intervalence electronic
spectroscopy of a system of two aryalmines bridged by Ru3+.
The most closely related spectroscopic example to our proposal
of ligand-to-ligand charge transfer between arylamines is a set
of organic compounds consisting of substituted tetraphenyl-
benzidines (TPB, a triarylamine dimer) containing methyl
groups at the central benzidine rings.40 The steric hindrance of
the central methyl groups twists the molecule and leads to a
progressive loss of planarity in the conjugated π system. This
structural feature is associated with the coexistence of both the
intervalence transition at the NIR (at 1876 nm for the
intermediate-coupled benzidines and 1467 nm for the full
coupled system) and a displacement of the arylamine radical
absorption (from ca. 750 nm in free tritolylamine to 685 nm in
the intermediate-coupled benzidine, and 484 nm in the full
coupled one). The effect is rationalized by assuming a variable
π orbital overlap between arylamine fragments, depending on
the angle between phenyl planes. The spectrum of dioxidized
Ru-(dt3pya)2, with bands at 1435 and 620 nm, locates the
Using the MLCT RuII → bpy* in Ru-dt3pya as absorptivity
calibration (ε = 7750 M−1 cm−1), we estimated 6050 M−1 cm−1
for molar absorptivity of the intervalence band (both in
acetonitrile). This band gives an oscillator strength of f = 0.12
and a transition moment of 1.25 eÅ. Assuming an average
diabatic separation distance of 7.5 Å,43 we estimated a dipole
moment change upon absorption of 7.1 eÅ (95% of the diabatic
distance), a mixture of 3% between diabatic states, and an
effective ligand coupling of 1200 cm−1, all these figures being
under an effective two-state electronic model.5,44
Lambert and Launay have analyzed intervalence systems
containing arylamines based on a quantitative model of three
electronic interacting diabatic states,45,46 with individual
1266
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
coupled oscillators representing the influence of the reaction
coordinate on the energies of the associated electronic states
(see especially Figure 4b of ref 45). Applying a three-state
analysis to our system involves considering (a) the dioxidized
and diabatic states RuII-dt3pya+-dt3pya+, RuIII-dt3pya+-
dt3pya, and RuIII-dt3pya-dt3pya+; (b) an electronic coupling
only between chemical bond adjacent sites; and (c)
reorganization energies λm (for the fragment [Ru(bpy)2]) and
λl (for each of the two ligands dt3pya). While it is not possible
to obtain all necessary data to analyze our system quantitatively,
we can make several valuable observations at a qualitative level
using this approach.
The single-oxidized species Ru-dt3pya generated in the
spectroelectrochemistry experiment only shows absorbance
toward the UV (above 400 nm) where intraligand bpy π→ π*
bands appear. Since we were unable to locate the LMCT
dt3pya → RuIII in the visible-NIR region, we estimated this
metal−ligand coupling element from measurements on related
compounds. A large number of charge transfer parameters in
the series [RuII/III(NH3)5L]2+/3+ by electroabsorption (Stark)
spectroscopy have been reported.44 From the LMCT transition
in [RuIII(NH3)5-4-aminopyridine]3+ at 511 nm, it follows a
metal−ligand coupling of 3400 cm−1 (0.42 eV), similar to other
RuIII with pyridinic ligands. This coupling value is an excess
estimation for the chromophore RuIII-dt3pya because RuIII
orbitals are partially involved in π interactions with bpy
coligands and are less available to overlap with the π system of
dt3pya. As the ligand, the delocalization of dt3pya because of
phenyl groups compared to 4-aminopyridine also limits the π
orbital overlapping with the metal. For these two reasons, we
assume that the metal−ligand coupling (in Ru-dt3pya as in Ru-
(dt3pya)2) will be lower than in the [RuIII(NH3)5-4-amino-
pyridine]3+.
mV separation between the first and second dt3pya in Ru-
(dt3pya)2 (from 1.37 to 1.80 V) suggests an odd
communication effect of d5 RuIII that does not match the
information of electronic spectroscopy (a coupled although
localized system). Unlike linear (usually trans) intervalence
compounds, dt3pya in Ru-(dt3pya)2 are in adjacent positions,
with solvent in between. Analytical solvation models used in
electron transfer do not consider dielectric polarization inside
the intervalence cavity, where it is expected that the electric
field has a much stronger effect aligning the solvent dipoles
compared to the electric field outside the system. Since these
models assume linear species, we are not currently able to
estimate quantitatively the electrostatic and solvation influence
in ΔE to asses whether this splitting in redox potential is
reasonable or not.
CONCLUSIONS
■
We here found qualitative evidence of the different capabilities
of the bridges [RuII(bpy)2]2+ and [RuIII(bpy)2]3+ to electroni-
cally connect an intervalence system consisting of two
arylamine-like redox-active ligands. The mixed valence behavior
of this ligand−metal−ligand system seems to follow the same
rules as the classical metal−ligand−metal system, since the best
bridges for coupling sites are those with complementary π
orbital interactions, for example, the combination of either π*
empty acceptor bridges like pyrazine and donor d6 metals such
as [RuII(NH3)5]2+ or π donors bridges like dicyanobenzene
with the oxidized acceptor fragment [RuIII(NH3)5]3+.6 Applied
into our system Ru-(dt3pya)2, the π donor properties of
neutral arylamines (popular hole transport material for this
reason) are matched to the d5 acceptor ability of [RuIII(bpy)2]3+
to give a Class II compound dt3pya-dt3pya+ by the
superexchange mechanism. In contrast, the donor fragment
[RuII(bpy)2]2+ is not able to interact either with neutral or with
oxidized arylamine, giving a Class I intervalence system.
However, the same RuII fragment is able to interact with π*
empty acceptor orbitals in bpy and, when one of the ligands is
reduced, they become a Class II intervalence system where RuII
is an efficient communicating bridge.
It has been estimated that the internal reorganization energy
of tritolylamine (a reasonable approximation of dt3pya) at
DFT-B3LYP-6-31G* level is 0.14 eV.47 This value is about half
the internal reorganization energy in octahedral d6/5 low spin
coordination compounds of similar size, which is roughly
estimated at 0.24 eV.5 Considering that the external
reorganization adds about 0.7 eV for a charge-separated specie
like Ru-(dt3pya)2 in a polar solvent like acetonitrile,5 both λl
and λm result larger than the estimated coupling. This places the
Ru-(dt3pya)2 system at Class 2 (according to Robin and Day’s
classification) but coupled enough to stay in the adiabatic
regime.5 This is consistent with the properties of the
intervalence transition analyzed above (3% delocalization).
The fact that the spectrum of dioxidized Ru-(dt3pya)2
greatly resembles that of partially twisted benzidines suggests
that the unpaired electron circulates primarily by a minimum
energy at the potential surface associated with RuIII-dt3pya+-
dt3pya and RuIII-dt3pya-dt3pya+ states or, in Lambert’s
nomenclature,45 the triarylamine radical cation state centers.
Moreover, the minimum on that surface energy associated with
the remaining state RuII-dt3pya+-dt3pya+ (the bridge state in
terms of Lambert’s model) is either not accessible at the
experiment temperature or, at least, the concentration is not
spectroscopically detectable. This is only possible if the diabatic
state RuII-dt3pya+-dt3pya+ is at a higher energy than the other
two states involving RuIII. Therefore, the surface for Ru-
(dt3pya)2 is similar to the one depicted by Lambert in Figure 4
of ref 45.
Future efforts will focus on obtaining derivatives with CN−,
SCN−, and carboxypyridine to prevent photolabilization and
allow studying the photophysics of the ruthenium-arylamine
interaction in the context of solar cells. Ligand derivatization
will allow changing arylamines and ruthenium redox potentials
to continue exploring the role of the metal oxidation state in
the ligand−ligand mixed valence properties. We will also
explore the reactivity of coordinated diphenyl-3-aminopyridine
(the use of phenyl instead of tolyl would form polymers), and
diphenyl-4-aminopyridine (which should not participate in
irreversible reactions after oxidation). Theoretical studies will
also help in a deeper and quantitative understanding of the
electronic structure and intervalence properties of these
compounds.
ASSOCIATED CONTENT
■
S
* Supporting Information
A graph containing NMR signals at methyl resonance region
before and after irradiation a 473 nm, with integrations. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Whereas delocalized benzidine40 reaches about 300 mV in
ΔE and localized twisted benzidine about 150 mV, almost 400
1267
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268

Inorganic Chemistry
Article
(26) (a) Scarminio, I.; Kubista, M. Anal. Chem. 1993, 65, 409−416.
(b) Kubista, M.; Sjoeback, R.; Albinsson, B. Anal. Chem. 1993, 65,
994−998.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: aparise@mdp.edu.ar.
́
(27) Zayat, L.; Calero, C.; Albores, P.; Baraldo, L.; Etchenique, R. J.
Am. Chem. Soc. 2003, 125, 882−883.
(28) Ru-N(pyridine) bond rotation with fixed inter atomic distances
calculated by DFT-B3LYP-LANL2DZ.
ACKNOWLEDGMENTS
■
(29) All attempts to obtain single crystals to clarify this issue were so
far unsuccessful. However, crystal packaging could not represent the
dynamic behavior of the ligands in solution, and electrostatic
interactions could make some specific configuration preferable.
(30) Not to be confused with the π* of bpy coligands, which remains
at about the same energy in different complexes. Indeed, the bpy−/0
redox potential does not change significantly in a series of RuII
complexes. The reason why MLCT RuII → bpy* shifts is mainly the
stabilization of t2g orbitals.
(31) Pinnick, D. V.; Durham, B. Inorg. Chem. 1984, 23, 1440−1445.
(32) Sreenath, K.; Suneesh, C. V.; Gopidas, K. R.; Flowers, R. A. II J.
Phys. Chem. A 2009, 113, 6477−6483.
(33) Caspar, J.; Meyer, T. J. Inorg. Chem. 1983, 22, 2444−2453.
(34) For the same set of ligands, bpy+/0 redox potential spans along
0.2 V while RuIII/II spans 0.8V. See ref 33.
(35) Note that the voltammogram in Figure 5a does not reach the
wave at 1.80 V, while the pulsed experiment in Figure 5b does. This
wave makes the process at 1.39 V become irreversible. See Figures 5c−
d.
(36) (a) Fielder, S. S.; Osborne, M. C.; Lever, A. B. P.; Pietro, W. J. J.
Am. Chem. Soc. 1995, 117, 6990−6993. (b) Dodsworth, E. S.; Vlcek, A.
A.; Lever, A. B. P. Inorg. Chem. 1994, 33, 1045−1049.
(37) The impurity at 1.05 V in the SWV voltammogram (Figure 5b)
is not responsible for the first appearance of the 712 nm. We
equilibrated the spectrum at 1.20 V and only a small change at the 461
nm MLCT band was seen.
(38) (a) Kattnig, D. R.; Mladenova, B.; Grampp, G.; Kaiser, C.;
Heckmann, A.; Lambert, C. J. Phys. Chem. C 2009, 113, 2983−2995.
(39) Ramirez, C. L.; Pegoraro, C.; Trupp, L.; Bruttomesso, A.;
Amorebieta, V. T.; Vera, D. M. A.; Parise, A. R. Phys. Chem. Chem.
Phys., 2011, 132007620080.
(40) Low, P. J.; Paterson, M. A. J.; Goeta, A. E.; Yufit, D. S.; Howard,
J. A. K.; Cherryman, J. C.; Tackleyb, R.; Brown, B. J. Mater. Chem.
2004, 14, 2516−2523.
This research was supported by the National Agency for
Science and Technology Promotion (ANPCYT, PICT 213).
A.R.P. and R.E. are members of CONICET. C.L.R. and O.F.
thank CONICET for graduate fellowships. We thank Nestor
́
Katz for reading the manuscript and valuable discussions.
REFERENCES
■
(1) Robin, M. B.; Day, P. Adv. Inorg. Chem. Radiochem. 1967, 10,
247−422.
(2) Allen, G. C. Hush, N. S. Prog. Inorg. Chem. 1967, 8, 357−389.
(3) Hush, N. S. Prog. Inorg. Chem. 1967, 8, 391−444.
(4) Day, P. Int. Rev. Phys. Chem. 1981, 1, 149−193.
(5) Creutz, C. Prog. Inorg. Chem. 1983, 30, 1−73.
(6) Crutchley, R. J. Adv. Inorg. Chem. 1994, 41, 273−325.
(7) Brown, D. B., Ed.; Mixed-Valence Compounds: Theory and
Applications in Chemistry, Physics, Geology, and Biology; Reidel
Publishing Company: Dordrecht, The Netherlands, 1980; NATO
ASI Series D.
(8) Prassides, K., Ed.; Mixed Valence Systems: Applications in
Chemistry, Physics and Biology; Kluwer Academic Publishers:
Dordrecht, The Netherlands, 1991; NATO ASI Series C, Vol 343.
(9) (a) Demadis, K. D.; Hartshorn, C. M.; Meyer, T. J. Chem. Rev.
2001, 101, 2655−2686. (b) Brunschwig, B. S.; Creutz, C.; Sutin, N.
Chem. Soc. Rev. 2002, 31, 168−184.
(10) Day, P.; Hush, N. S.; Clark, R. J. H. Philos. Trans. R. Soc., A
2008, 366, 5−14 , and references in the same issue.
(11) Soncini, A.; Mallah, T.; Chibotaru, L. F. J. Am. Chem. Soc. 2010,
132, 8106−8114.
(12) Nocera, D. G. Inorg. Chem. 2009, 48, 10001−10017.
(13) (a) Lu, Y.; Quardokus, R.; Lent, C. S.; Justaud, F.; Lapinte, C.;
Kandel, S. A. J. Am. Chem. Soc. 2010, 132, 13519−13524. (b) Lent, C.
S.; Isaksen, B.; Lieberman, M. J. Am. Chem. Soc. 2003, 125, 1056−
1063.
(41) We could not follow the suggestion of Referee #1 regarding
measuring the EPR spectrum of the dioxidized specie because we did
not find a suitable chemical reagent for oxidation of Ru-(dt3pya)2 at
1.4 V.
(42) Nelson, R. F.; Fritsch, J. M.; Marcoux, L. S.; Donald, W.; Leedy,
D. W.; Adams, R. N. J. Am. Chem. Soc. 1966, 88, 3498−3503.
(43) This is a very rough estimation of the diabatic distance because
dt3pya is rotating, and redox centers are not at a fixed position. We
estimate that separation between the central nitrogens of dt3pya span
from 5 to 10 Å.
(14) Nelsen, S. F.; Ismagilov, R. F.; Trieber, D. A. Science 1997, 278,
846.
(15) Acosta, A.; Zink, L. I.; Cheon, J. Inorg. Chem. 2000, 39, 427−
432.
(16) Metcalfe, R. A.; Lever, A. B. P. Inorg. Chem. 1997, 36, 4762−
4771.
(17) (a) Balzani, V.; Bergamini, G.; Marchioni, F.; Ceroni, P. Coord.
Chem. Rev. 2006, 250, 1254−1266. (b) Kalyanasundaram, K. Coord.
Chem. Rev. 1982, 46, 159−244. (c) Juris, A.; Balzani, V.; Barigelletti,
F.; Campagna, S.; Belser, P.; von Zelewsky, A. Coord. Chem. Rev. 1988,
84, 85−277.
(44) Shin, Y. K.; Brunschwig, B. S.; Creutz, C.; Sutin, N. J. Phys.
Chem. 1996, 100, 8157−8169.
(18) Concepcion, J. J.; Jurss, J. W.; Brennaman, M. K.; Hoertz, P. G.;
Patrocinio, A. O. T.; Iha, N. I. M.; Templeton, J. L.; Meyer, T. J.. Acc.
Chem. Res. 2009, 42, 1954−1965.
(45) Lambert, C.; Noll, G.; Schelter, J. Nat. Mater. 2002, 1, 69−73.
̈
(46) Launay, J. P.; Coudret, C.; Hortholary, C. J. Phys. Chem. B 2007,
111, 6788−6797.
(19) Gratzel, M. Acc. Chem. Res. 2009, 42, 1788−1798.
̈
(47) Lin, B. C.; Cheng, C. P.; Lao, Z. P. M. J. Phys. Chem. A 2003,
107 (26), 5241−5251.
(20) Jin, Z.; Masuda, H.; Yamanaka, N.; Minami, M.; Nakamura, T.;
Nishikitani, Y. Phys. Chem. C 2009, 113, 2618−2623.
(21) Jiang, Z.; Ye, T.; Yang, C.; Yang, D.; Zhu, M.; Zhong, C.; Qin, J.;
Ma, D. Chem. Mater. 2011, 23, 771−777.
(22) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Pergamon Press: Oxford, U.K., 1988.
(23) (a) Sullivan, B. P.; Salmon, D. J.; Meyer, T. J. Inorg. Chem. 1978,
17, 3334. (b) Viala, C.; Coudret, C. Inorg. Chim. Acta 2006, 359, 984−
989.
(24) Ellis, C. D.; Margerum, L. D.; Murray, R. W.; Meyer, T. J. Inorg.
Chem. 1983, 22, 1283−1291.
(25) Noviandri, I.; Brown, K.; Leming, D. S. J. Phys. Chem. B 1999,
103, 6713−6722.
1268
dx.doi.org/10.1021/ic200966f | Inorg. Chem. 2012, 51, 1261−1268