Photoinduced electron transfer in linear triarylamine-photosensitizer- anthraquinone triads with ruthenium(II), osmium(II), and iridium(III)

Article abstract of DOI:10.1021/ic300558s

A rigid rod-like organic molecular ensemble comprised of a triarylamine electron donor, a 2,2′-bipyridine (bpy) ligand, and a 9,10-anthraquinone acceptor was synthesized and reacted with suitable metal precursors to yield triads with Ru(bpy)₃²⁺, Os(bpy)₃²⁺, and [Ir(2-(p-tolyl)pyridine)₂(bpy)]⁺ photosensitizers. Photoexcitation of these triads leads to long-lived charge-separated states (τ = 80-1300 ns) containing a triarylamine cation and an anthraquinone anion, as observed by transient absorption spectroscopy. From a combined electrochemical and optical spectroscopic study, the thermodynamics and kinetics for the individual photoinduced charge-separation and thermal charge-recombination events were determined; in some cases, measurements on suitable donor-sensitizer or sensitizer-acceptor dyads were necessary. In the case of the ruthenium and iridium triads, the fully charge-separated state is formed in nearly quantitative yield.

Full text of DOI:10.1021/ic300558s

Article
pubs.acs.org/IC
Photoinduced Electron Transfer in Linear Triarylamine−
Photosensitizer−Anthraquinone Triads with Ruthenium(II),
Osmium(II), and Iridium(III)
Jihane Hankache,† Marja Niemi,‡ Helge Lemmetyinen,*,‡ and Oliver S. Wenger*,†
†Georg-August-Universitat Gottingen, Institut fur Anorganische Chemie, Tammannstrasse 4, D-37077 Gottingen, Germany
̈
̈
̈
̈
‡Tampere University of Technology, Department of Chemistry and Bioengineering, P.O. Box 541, FIN-33101 Tampere, Finland
ABSTRACT: A rigid rod-like organic molecular ensemble comprised of a
triarylamine electron donor, a 2,2′-bipyridine (bpy) ligand, and a 9,10-
anthraquinone acceptor was synthesized and reacted with suitable metal
precursors to yield triads with Ru(bpy)32+, Os(bpy)32+, and [Ir(2-(p-
tolyl)pyridine)2(bpy)]+ photosensitizers. Photoexcitation of these triads
leads to long-lived charge-separated states (τ = 80−1300 ns) containing a
triarylamine cation and an anthraquinone anion, as observed by transient
absorption spectroscopy. From a combined electrochemical and optical
spectroscopic study, the thermodynamics and kinetics for the individual
photoinduced charge-separation and thermal charge-recombination events were determined; in some cases, measurements on
suitable donor−sensitizer or sensitizer−acceptor dyads were necessary. In the case of the ruthenium and iridium triads, the fully
charge-separated state is formed in nearly quantitative yield.
3
2,2′-bipyridine); particularly its short MLCT (metal-to-ligand
charge transfer) lifetime is less than optimal for photoinduced
electron transfer chemistry.11 Despite the need for shorter
INTRODUCTION
■
The construction of molecular triads with linear alignment of
an electron donor, photosensitizer, and an electron acceptor is
of long-standing interest.1 A key advantage of linear rigid rod-
like constructs is a maximum separation distance of the
electron−hole pair in the charge-separated state. The tradi-
tional approach to obtaining linear triads for vectorial electron
transfer with d6 metal photosensitizers involves the use of
2,2′;6′,2″-terpyridine (tpy) ligands, which are substituted at the
4′-position of the central pyridine ring with appropriate electron
donors or acceptors (Scheme 1a).2−10
3+
excitation wavelengths, isoelectronic Ir(tpy)2 complexes
represent an attractive alternative from a photophysical point
of view, but they are not at all easy to synthesize.8,12,13 An
interesting, newly discovered alternative option is the bis-
(diquinolinyl)pyridine ligand, which is structurally similar to
tpy (Scheme 1b) but ligates to ruthenium(II) with a
significantly larger bite angle, which in turn results in more
favorable photophysical properties of the complex.14−17
Numerous molecular electron transfer triads based on the
2+
2+
However, the Ru(tpy)2 photosensitizer has rather poor
Ru(bpy)3 photosensitizer have been explored. However,
photophysical properties when compared to Ru(bpy)32+ (bpy =
when substituting one bpy ligand with an electron donor
while equipping a second bpy ligand with an electron acceptor
(Scheme 1c), one is often confronted with the problem of
isomerism, and analysis of the electron transfer kinetics may
become tricky.18−28 Moreover, the resulting molecular
constructs are not linear. A viable solution to this problem is
to attach the donor and the acceptor at the 5- and 5′-positions
of a given bpy ligand (Scheme 1d). We recently communicated
preliminary results on what we believe to be the first rigid rod-
Scheme 1. Possible Constructs of Molecular Triads
Incorporating d6 Metal Complexes (Here, Ru(II)) as
Photosensitizers between Electron Donors (D) and Electron
Acceptors (A)
like (linear) triad based on the Ru(bpy)3 photosensitizer.29
2+
Here, we present a more detailed study including transient
absorption data at higher temporal resolution and an extension
2+
of the work to analogous triads with Os(bpy)3 and a
cyclometalated iridium(III) complex as photosensitizing units.
Specifically, we synthesized and investigated the three triads
from Scheme 2 along with a series of suitable molecular dyads.
Received: March 15, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Scheme 2. The Molecules Investigated in This Work
a
TAA = triarylamine; AQ = anthraquinone.
The triads are comprised of a triarylamine (TAA) electron
donor, a d6 metal diimine photosensitizer (RuII, OsII, IrIII), and
a 9,10-anthraquinone (AQ) unit, which is acting as a terminal
electron acceptor. The dyads contain either only the donor and
the sensitizer or the sensitizer and the acceptor.
Product characterization data are given in the Experimental
Section.
Optical Absorption and Luminescence Spectroscopy.
Figure 1 shows optical absorption spectra of the individual
compounds from Scheme 2 in acetonitrile solution at room-
temperature. The ruthenium complexes from panel (a) exhibit
metal-to-ligand charge transfer (MLCT) absorption bands
centered around 450 nm and a bpy-localized π−π* absorption
band at 290 nm. The AQ unit has relatively low-lying
absorptions, which account for some of the additional
extinction observable between 310 and 380 nm in the TAA−
RuII−AQ and RuII−AQ compounds.
However, although the p-xylene spacer leads to significantly
less π-conjugated systems than unsubstituted p-phenylene
bridges,33,34 some of the extinction in the 310−380 nm spectral
range is likely to be caused by the molecular bridge and/or an
increase of π-conjugation in the overall system. The TAA unit
absorbs predominantly at shorter wavelengths and contributes
substantially to the extinction below 300 nm.35,36
RESULTS AND DISCUSSION
■
Synthesis. The synthesis of the rigid rod-like triarylamine−
2,2′-bipyridine−anthraquinone unit was described in detail in
our prior communication.29 Briefly, the synthetic strategy is
based on 5,5′-dibromo-2,2′-bipyridine as a starting material to
which 4-(trimethylsilyl)phenylboronic acid was attached on
both sides in a Suzuki cross-coupling reaction. After
trimethylsilyl-halogen exchange, the resulting molecule can be
coupled to anthraquinone-2-boronic acid pinacol ester in a
Suzuki-type cross-coupling reaction. A subsequent palladi-
um(0)-catalyzed N−C coupling reaction with 4,4′-dimethoxy-
diphenylamine then yields the desired molecular rod in 33%
overall yield. The syntheses of the dyads departed from 5-
bromo-2,2′-bipyridine and relied on the same coupling strategy
involving 4-(trimethylsilyl)phenylboronic acid as a first
coupling partner, trimethylsilyl-halogen exchange, followed by
a reaction either with anthraquinone-2-boronic acid pinacol
ester or with 4,4′-dimethoxydiphenylamine. Detailed synthetic
protocols can be found in the Supporting Information of our
prior communication.29 Complexation of the dyad and triad
Expectedly, the osmium complexes from panel (b) exhibit
the same spectral features as the isoelectronic ruthenium
1
compounds, only that the MLCT bands are red-shifted and
3
that the MLCT absorptions between 520 and 700 nm now
become easily detectable as a consequence of the relaxation of
spin selection rule. As in the case of ruthenium, the AQ unit
causes the TAA−OsII−AQ and OsII−AQ molecules to absorb
more strongly between 310 and 380 nm than the TAA−OsII
and OsII molecules.
ligands to Ru(bpy)2Cl2, Os(bpy)2Cl2, and [Ir(2-(p-tolyl)-
30,31
pyridine)2Cl]2
occurred following standard protocols.32
B
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. Luminescence spectra of the 12 molecules from Scheme 2 in
aerated CH3CN at 25 °C. Excitation occurred at 450 nm in the case of
the ruthenium and osmium molecules (a, b) and at 380 nm in the case
of the iridium molecules (c). Relative emission intensities were
corrected for differences in sample absorbance at the excitation
wavelength, and the final intensities were normalized to a value of 1.0
for the reference complexes (RuII, OsII, IrIII).
Figure 1. Optical absorption spectra of the 12 molecules from Scheme
2 in CH3CN at 25 °C.
The cyclometalated iridium complexes in panel (c) exhibit
3
1MLCT and MLCT absorptions in the 400−500 nm spectral
range, and to the higher energy side these absorptions merge
directly into intraligand π−π* absorptions. Thus, it is difficult
to perform a clear distinction between absorptions that involve
the metal center and absorptions which do not. Be that as it
may, the global appearance of all absorption spectra in Figure 1
is that expected for d6 metal complexes of this type.11,37−42 The
dyad and triad spectra of the ruthenium and osmium triads
correspond more or less (but not precisely) to the sum of the
absorption spectra of the individual molecular components,
indicating that the overall systems are electronically weakly
coupled. At donor−photosensitizer and photosensitizer−
acceptor distances of roughly 4.3 Å (i. e., the length of one
p-xylene spacer), this is to be expected. However, for the TAA−
Ir dyad and the TAA−Ir−AQ triad, the absorption spectra in
Figure 1c reveal more pronounced interaction between the
individual molecular moieties: There is a broad absorption
band around 450 nm, in a spectral region where none of the
individual components absorbs. Similar observations have been
lying triplet excited states (∼2.7 eV for AQ, ∼3.2 eV for
TAA);45,46 hence excited-state deactivation by triplet−triplet
2+
energy transfer from the Ru(bpy)3 3MLCT excited state at
2.12 eV to either one of these two moieties is thermodynami-
cally unlikely.47−50 Subsequent sections will demonstrate that
electron transfer from the TAA unit to the ruthenium complex
is in fact the predominant excited-state deactivation pathway in
the TAA−RuII and TAA−RuII−AQ molecules.
Expectedly, the 3MLCT emissions of the osmium com-
pounds in Figure 2b (excited at 450 nm) are all significantly
red-shifted with respect to the ruthenium complexes in Figure
2a. The emission intensities of the OsII reference complex and
the TAA−OsII dyad are similar, while those of the OsII−AQ
dyad and the TAA−OsII−AQ triad are nearly a factor of 2 less
3
2+
intense. Since the emissive MLCT state of the Os(bpy)3
complex is at even lower energy (1.79 eV)51 than that of
Ru(bpy)32+, triplet−triplet energy transfer is even less probable
in this case. Indeed, the subsequent sections will demonstrate
that electron transfer from photoexcited osmium to AQ is an
important excited-state deactivation channel.
43
3+
made previously in a TAA−Ir(tpy)2 dyad.
Steady-State Luminescence Spectroscopy. All of the
compounds from Scheme 2 are emissive when irradiating
∼10−5 M (aerated) acetonitrile solutions of them with blue or
UV light, albeit with widely varying luminescence intensities.
Figure 2a shows the emission spectra of the four ruthenium
compounds as detected after excitation at 450 nm. The
luminescence intensity of the RuII reference complex has been
normalized artificially to a value of 1.0; all other luminescence
intensities are scaled relative to this reference point.44 The
RuII−AQ dyad exhibits an emission intensity practically on par
with that of the ruthenium reference complex, while the TAA−
RuII and TAA−RuII−AQ molecules emit an order of magnitude
weaker. Both the AQ and TAA units have energetically high
Figure 2c shows the luminescence spectra obtained from the
iridium complexes after excitation at 380 nm. For cyclo-
metalated iridium(III) complexes of this type, the emission is
40,42
3
commonly of mixed MLCT/intraligand π−π* character.
Given the comparatively high energy of the emissive triplet
states, excited-state deactivation by triplet−triplet energy
transfer is an energetically more viable option for the iridium
dyads and triads than for the ruthenium and osmium
compounds. Nevertheless, the subsequent paragraphs of this
paper will show that the strong emission quenching observed in
C
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 3. Cyclic voltammograms of the molecules from Scheme 2 in CH3CN in the presence of 0.1 M TBAPF6 as a supporting electrolyte. The
waves at 0.0 V vs Fc+/Fc (vertical dashed lines) are due to ferrocene, which was added in small quantities for internal voltage calibration.
the TAA−IrIII, IrIII−AQ, and TAA−IrIII−AQ molecules is
predominantly the consequence of efficient excited-state
deactivation by photoinduced electron transfer.
Table 1. Reduction Potentials for the Individual Redox-
Active Components of the Ruthenium Molecules from
a
Scheme 2
To summarize this paragraph on the steady-state lumines-
cence properties, we note that the ruthenium emission is
quenched significantly in presence of the TAA donor while AQ
has a weak influence, the osmium luminescence is quenched to a
noticeable extent in the presence of AQ while TAA has a weak
influence, and the iridium luminescence is strongly susceptible
to the presence of both TAA and AQ.
Electrochemical Investigations and Energy Level
Structure of the Triads. Figure 3 shows the cyclic
voltammograms of the compounds in Scheme 2 as measured
in acetonitrile solution in the presence of 0.1 M tetrabutyl-
ammonium hexafluorophosphate (TBAPF6) as a supporting
electrolyte. The reversible waves at 0.0 V vs Fc+/Fc (dashed
vertical lines) are due to ferrocene, which was added in small
quantities to the solutions for internal voltage calibration.
Vertical superposition of voltammograms from the reference
complex (red traces), TAA−metal dyad (oranges traces),
metal−AQ dyad (green traces), and TAA−metal−AQ triad
(blue traces) in each of the three panels (a, b, c) permits
unambiguous assignment of the individual redox waves. The
reduction potentials extracted from the ruthenium compounds
in Figure 3a are summarized in Table 1.
Oxidation of Ru(II) to Ru(III) occurs at a potential of 0.9 V
vs Fc+/Fc, in line with prior investigations.38,39,52 Three bpy-
localized reductions of the metal complex occur between −1.70
and −2.15 V vs Fc+/Fc, also in agreement with literature
values.53,54 Although these are clearly ligand-centered reduc-
tions, for convenience we will later designate the first of these
reduction processes as a reduction of the ruthenium(II)
complex to a ruthenium(I) species (RuII/RuI). Oxidation of
the TAA unit occurs at 0.30 V vs Fc+/Fc, and reduction of AQ
is at −1.27 V vs Fc+/Fc, both in agreement with previously
reported redox potentials for these moieties.19,35,36,55−57 The
fact that the redox potentials of all individual molecular
components are nearly identical to those reported for their
isolated counterparts is another indication that the molecular
dyads and triads from Scheme 2 are electronically weakly
coupled systems.
TAA−RuII
RuII−AQ
TAA−RuII−AQ
2+
Ru(bpy)3
0.89
Ru(III/II)
TAA+/0
AQ0/−
0.89
0.30
0.90
0.92
0.30
−1.28
−1.72
−1.90
−2.14
−1.27
−1.73
−1.86
−2.13
bpy0/−
−1.72
−1.91
−2.15
−1.70
−1.90
−2.13
bpy0/−
bpy0/−
a
All values were extracted from the data in Figure 3a and are reported
versus the ferrocenium/ferrocene (Fc+/Fc) couple in acetonitrile
solution. bpy- and AQ-localized one-electron reductions as well as
TAA- and metal-localized one-electron oxidations are found to exhibit
peak separations near the expected 59 mV, but the oxidation processes
have higher peak currents in the oxidative than in the reductive sweep
(ratios vary between 1:1 and ∼30:1).
The osmium data in Figure 3b lead us to similar conclusions.
Not surprisingly, the TAA, AQ, and bpy redox potentials are
hardly affected by the change in metal (Table 2); only the metal
oxidation process is susceptible to the replacement of Ru(II) by
Os(II). The conversion of Os(II) to Os(III) occurs at ∼0.5 V
Table 2. Reduction Potentials for the Individual Redox-
Active Components of the Osmium Molecules from Scheme
a
2
2+
Os(bpy)3
0.45
TAA−OsII
OsII−AQ
TAA−OsII−AQ
Os(III/II)
TAA+/0
AQ0/−
0.47
0.30
0.46
0.48
0.30
−1.29
−1.65
−1.85
−2.17
−1.28
−1.62
−1.83
−2.14
bpy0/−
−1.67
−1.86
−2.16
−1.63
−1.84
−2.13
bpy0/−
bpy0/−
a
All values were extracted from the data in Figure 3b and are reported
versus the ferrocenium/ferrocene (Fc+/Fc) couple in acetonitrile
solution. All one-electron redox waves exhibit peak separations near
the expected 59 mV. All redox waves exhibit current peak ratios near
1:1 when comparing the current peaks of oxidative and reductive
sweeps.
D
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
vs Fc+/Fc, again in line with previous investigations.51 As noted
above for ruthenium, reduction of the Os(bpy)3 complex
the organic redox-active moieties (TAA and AQ) are essentially
the same in all of the dyads and triads (Tables 1−3).
On the basis of the electrochemical data from Figure 3 and
Tables 1−3, one can establish the energy level diagram for the
triads shown in Scheme 3. In doing so, we have neglected any
effects arising from distance-dependent donor−acceptor
interactions and have simply calculated energies for the
individual states from differences in reduction potentials.58
This procedure obviously yields crude estimates at best, and
even though we indicate energies to two digits of electron volts,
we note that the error bars associated with these values are on
the order of 0.1 eV. The energies of the emissive excited states
of the individual metal complexes (2.12 eV, 1.79 eV, 2.37 eV)
represent the commonly used literature values.11,31,38,39,51 In
Scheme 3, these excited states are designated as *RuII, *OsII,
and *IrIII.
2+
occurs predominantly at the bpy ligands, but for convenience
we will later designate the one-electron reduced osmium
complex as OsI.
The cyclic voltammograms for the iridium complexes in
Figure 3c are less rich on the reductive side than the ruthenium
and osmium data because there is only one (instead of three)
ligand-based reduction process in the potential window
considered here. The wave at −1.8 V vs Fc+/Fc is assigned
to a bpy-localized reduction process (Table 3), whereas the 2-
Table 3. Reduction Potentials for the Individual Redox-
Active Components of the Iridium Molecules from Scheme
a
2
IrIII
TAA−IrIII
IrIII−AQ
TAA−IrIII−AQ
Common to all three triads is the presence of a charge-
separated state near 1.6 eV containing oxidized triarylamine
(TAA+), the metal complex in its initial state (RuII, OsII, or
IrIII), and reduced anthraquinone (AQ−). The energy of this
final charge-separated state is obviously independent of the
metal. Importantly, this state is energetically below all of the
initially excited metal-localized emissive states; consequently,
the final charge-separated state is energetically accessible
irrespective of whether the ruthenium, osmium, or iridium
triad is considered.
Ir(IV/III)
TAA+/0
AQ0/−
0.92
0.84
0.30
0.84
0.84
0.30
−1.28
−1.80
−1.26
−1.75
bpy0/−
−1.79
−1.76
a
All values were extracted from the data in Figure 3c and are reported
versus the ferrocenium/ferrocene (Fc+/Fc) couple in acetonitrile
solution. The bpy- and AQ-localized reductions exhibit good
reversibility. TAA- and iridium-localized one-electron oxidation
waves show peak separations close to the expected 59 mW, but the
ratio between the current peaks in oxidative and reductive sweeps
ranges from ∼1:1 to ∼40:1.
In all three triads, there are two possibilities for the formation
of the final charge-separated state: (i) reductive quenching of
the initially excited *RuII, *OsII, and *IrIII species by TAA,
followed by electron transfer from the now reduced RuI, OsI, or
IrII complexes to AQ, or (ii) oxidative quenching of the initially
excited *RuII, *OsII, and *IrIII species by AQ, followed by
electron transfer from TAA to the now oxidized RuIII, OsIII, or
IrIV complexes. As seen from Scheme 3a, in the ruthenium
system, possibility i is likely to dominate because the TAA+−
RuI−AQ state at 2.03 eV can be formed in an exergonic step
from the initially excited TAA−*RuII−AQ level (at 2.12 eV),
while formation of the TAA−RuIII−AQ− state (at 2.19 eV) is
thermodynamically uphill. These thermodynamic considera-
tions are consistent with the observation of a very weak
emission quenching in the RuII−AQ dyad with respect to the
RuII reference complex (Figure 2a), and a significant
luminescence quenching in the TAA−RuII dyad and TAA−
RuII−AQ triad relative to RuII.
(p-tolyl)pyridine ligands are apparently reduced only at
significantly more negative potentials. By analogy to what was
noted above for ruthenium and osmium, below we will
designate the reduced iridium complex as IrII. Oxidation of
the metal complex occurs at ∼0.84 V vs Fc+/Fc, producing a
species that will be designated as IrIV. This potential is 80 mV
lower for the iridium triad and dyads compared to that of the
free IrIII complex (second row of Table 3). For the ruthenium
and osmium compounds, the difference between the metal
oxidation potentials of the free complexes and those of the
dyads and triads ranges from 0 to 30 mV (Tables 1 and 2).
Thus, its seems that interaction of the iridium photosensitizing
unit with the TAA and AQ moieties is somewhat stronger than
in the case of the ruthenium and osmium dyads and triads. This
finding is in line with those from optical absorption
spectroscopy (see above). The electrochemical potentials of
Scheme 3. Energy Level Scheme Showing the Relevant Photoexcited and Charge-Separated States Which Can Be Formed in the
Three Triads from Scheme 2
E
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
In the case of the osmium triad (Scheme 3b), the
thermodynamic situation is different, and the oxidative
quenching of the initially excited *OsII state is more probable:
The TAA+−OsI−AQ state is at 1.92 eV, which is roughly 0.13
3
eV above the initially excited MLCT state. By contrast, the
TAA−OsIII−AQ− level is at 1.76 eV, energetically close to the
3
2+
lowest MLCT state of the Os(bpy)3 unit (at 1.79 eV);
hence, the oxidative quenching appears thermodynamically
more viable than the reductive quenching. Again, the
luminescence data from Figure 2 is consistent with our
energetic considerations: The emission intensity of the
molecules containing AQ (OsII−AQ dyad and TAA−OsII−
AQ triad) are weaker than those of the molecules that do not
contain this oxidant (OsII, TAA−OsII dyad), supporting the
hypothesis that excited-state deactivation by electron transfer to
anthraquinone is a more efficient process than reductive
excited-state quenching by triarylamine in this case.
For the iridium triad, both the oxidative and reductive
excited-state quenching processes are thermodynamically
downhill from the initial TAA−*IrIII−AQ state at 2.37 eV:
The TAA−IrIV−AQ− level is estimated to be at 2.10 eV; the
TAA+−IrII−AQ state is calculated to lie at 2.05 eV. The
observation of strong emission quenchings in both iridium
dyads (including the triad) relative to the IrIII reference
complex is consistent with this energy level structure:
irrespective of whether TAA or AQ is attached to the metal
complex, nonradiative excited-state deactivation becomes
efficient.
Nanosecond Transient Absorption. Figure 4 provides
direct experimental evidence for the formation of the final
charge-separated states containing oxidized TAA and reduced
AQ. The series of transient absorption spectra shown in Figure
4a−c was measured using ∼10−5 M solutions of the TAA−
RuII−AQ (a), TAA−OsII−AQ (b), and TAA−IrIII−AQ (c)
triads in deoxygenated acetonitrile.
Figure 4. Panels a−c show transient absorption spectra measured on
CH3CN solutions of TAA−RuII−AQ (a), TAA−OsII−AQ (b), and
TAA−IrIII−AQ (c) in a 200-ns time window starting immediately after
excitation with ∼10-ns laser pulses at 532 nm (a, b) or 355 nm (c).
Panel d shows a series of absorption spectra from a CH2Cl2 solution of
a triarylamine reference molecule (structure shown in the inset)
measured after increasing time intervals following the application of an
electrochemical potential more positive than 0.5 V vs Fc+/Fc (in
presence of 0.1 M TBAPF6). Panel e shows a series of absorption
spectra obtained from a CH2Cl2 solution of 9,10-anthraquinone
obtained in an analogous spectro-electrochemical experiment using
potentials more negative than −1.2 V vs Fc+/Fc.
In the case of the ruthenium and osmium systems, excitation
occurred at 532 nm, while the iridium compound was excited at
355 nm. In all cases, the laser pulses had a width of ∼10 ns.
Detection took place in a time window of 200 ns starting
immediately after the laser pulses. Under these experimental
conditions, one obtains similar transient absorption spectra for
all three triads. In each of the three spectra (Figure 4a−c), there
are three bands with maxima near 380 nm, 565 nm, and 770
nm. On the basis of the spectro-electrochemical data in Figure
4d and e, the three bands can be readily assigned. Figure 4d
shows a series of absorption spectra which were obtained while
applying an electrochemical potential more positive than 0.5 V
vs Fc+/Fc to a CH2Cl2 solution of a triarylamine reference
molecule (chemical structure shown in the inset). From this
series of spectra (obtained after different time intervals after
initiating the oxidation process; using 0.1 M TBAPF6 as an
electrolyte), we learn that the transient absorption band located
around 770 nm is due to the oxidized amine.35,36,55,59,60 The
absorption spectra in Figure 4e were measured while applying
an electrochemical potential more negative than −1.2 V vs Fc+/
Fc to a CH2Cl2 solution of 9,10-anthraquinone in the presence
of 0.1 M TBAPF6. From this series of spectra, we learn that the
transient absorption bands located at 380 and 565 nm are due
to reduced anthraquinone.19,56,61,62 Thus, the observation of a
fully charge-separated state in all three triads is beyond
question, and this finding is in line with the energy level
diagram from Scheme 3, in which we have come to the
conclusion that such a final charge-separated state is
thermodynamically accessible from the initially excited metal-
localized state in all three triads.
Figure 5 shows the decays of the transient absorption
intensities at 380 nm (black traces), 565 nm (blue traces), and
770 nm (green traces) in deoxygenated acetonitrile solution. In
all three triads, we observe decays which are single exponential
over at least 1 order of magnitude, and in all cases the decays at
the three above-mentioned wavelengths yield nearly identical
lifetimes. This is consistent with the notion that the TAA+ and
AQ− species disappear jointly in a thermal charge-recombina-
tion event. The average lifetimes of the fully charge-separated
states extracted from fits to the experimental decay data in
Figure 5 are 1.3 μs in the case of the TAA+−RuII−AQ− state,29
80 ns for the TAA+−OsII−AQ− state, and 890 ns in the case of
the TAA+−IrIII−AQ− state (all in deoxygenated CH3CN at 25
°C).63
Given the fact that the final charge-separated state involves
electron−hole separation formally over a 22-Å distance, a
lifetime in the 100-ns-to-μs regime is not particularly
surprising,1 even if the effective electron transfer distance may
be somewhat shorter as a consequence of partial hole or
electron delocalization onto the p-xylene bridging ele-
ments.64−66 What is surprising, however, is the observation of
a markedly shorter lifetime for the osmium triad compared to
the ruthenium and iridium systems. We can only speculate what
F
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 6. (a) Temporal evolution of the transient absorption signal at
770 nm (TAA+ formation) after excitation of the TAA−RuII−AQ triad
at 400 nm with laser pulses of 150 fs width. (b) Time profile for the
transient absorption signal at 550 nm (AQ− formation) in the same
experiment. (c) Blue trace: Decay of the TAA+ absorption at 770 nm
in the TAA−RuII dyad after excitation at 450 nm with ∼10-ns laser
pulses. Red trace: Instrument response curve. (d) Black trace: Decay
of the luminescence emitted by RuII at 610 nm in deoxygenated
CH3CN after excitation at 450 nm. Green trace: decay of the same
luminescence in RuII−AQ in deoxygenated CH3CN.
Figure 5. Decays of the transient absorption intensities from Figure 4
at three different wavelengths in the TAA−RuII−AQ (a), TAA−OsII−
AQ (b), and TAA−IrIII−AQ (c) triads in deoxygenated CH3CN
solution at 25 °C. Excitation occurred at 532 nm for the ruthenium
and osmium triads and at 355 nm for the iridium system.
the origin of this effect might be, but it seems possible that the
energetic proximity of the TAA−OsIII−AQ− state, only 0.18 eV
above the final charge-separated state, may play a role. In the
ruthenium and iridium triads, the final charge-separated states
are energetically well below all of the other states (>0.45 eV);
hence, thermal one-step back-electron transfers might be less
likely to occur in these systems than in the osmium triad. Be
that as it may, we conclude from this section that the final
charge-separated states containing TAA+ cations and AQ−
anions are formed in all three triads from Scheme 2. In the
following, we focus on the kinetics and quantum yields of their
formation.
Kinetics and Quantum Yields for Formation of
Charge-Separated States in the Ruthenium Triad. Figure
6a shows the build-up of the transient absorption intensity at
770 nm after photoexcitation of the TAA−RuII−AQ triad in
acetonitrile at 400 nm. The pulse width in this case was 150 fs;
hence, we are able to monitor the formation of TAA+ at high
temporal resolution in this experiment. The rise time for this
transient absorption is 9 ps, and we may conclude that the
TAA+−RuI−AQ state at 2.03 eV is formed with a time constant
of 1.1 × 1011 s−1 from the initially excited TAA−*RuII−AQ
state at 2.12 eV. In Scheme 3 and Table 4, this particular
electron transfer event, corresponding to a reductive Ru-
(bpy)32+ 3MLCT excited-state quenching by TAA, is denoted
as process “1”.
Figure 6b shows the temporal evolution of the transient
absorption intensity for the same sample as in Figure 6a and in
the same experiment, but detected at 550 nm. The rise time in
this case is 50 ps, and we conclude that AQ− is formed with a
time constant of 2 × 1010 s−1. Given the rapid kinetics for the
formation of the TAA+−RuI−AQ state at 2.03 eV, a time
constant of 2 × 1010 s−1 is attributed to process “3” in Scheme
3, i.e., the electron transfer from RuI to AQ while maintaining
the hole at the TAA+ site. Already after ∼200 ps the transient
absorption intensities at 770 and 550 nm have both reached
their maxima, indicating that the final charge-separated state
Table 4. Rate Constants for the Individual Intramolecular
Processes Shown in Schemes 3 and 4 As Extracted from the
Data in Figures 5−8 (Deoxygenated CH3CN Solution,
25°C)
a
reaction step no.
TAA−RuII−AQ
TAA−OsII−AQ
TAA−IrIII−AQ
1
2
3
4
5
6
1.1 × 1011 s−1
2.1 × 106 s−1
2.0 × 1010 s−1
6.7 × 107 s−1
7.7 × 105 s−1
1.2 × 106 s−1
<5.3 × 106 s−1
3.5 × 1012 s−1
3.3 × 1011 s−1
2.5 × 1010 s−1
1.5 × 107 s−1
1.1 × 106 s−1
4.4 × 106 s−1
∼108 s−1
∼108 s−1
3.8 × 107 s−1
1.3 × 107 s−1
5.3 × 107 s−1
a
Refers to the reaction steps marked by the numbered arrows in
Schemes 3 and 4.
(TAA+−RuII−AQ−) at 1.57 eV is completely formed at this
point. As discussed in the prior section, this state has a lifetime
of 1.3 μs in deoxygenated CH3CN, corresponding to a rate
constant of 7.7 × 105 s−1 for process “5” in Scheme 3a (Table
4).
In the TAA−RuII dyad, the TAA+ radical cation absorption at
770 nm decays with a lifetime of 15 ns (blue trace in Figure 6c).
We infer from this observation that, in the triad, process “4”
(Scheme 3a) proceeds with a rate constant of 6.7 × 107 s−1.
Thus, once the TAA+−RuI−AQ state at 2.03 eV is formed, the
system is much more likely to undergo ruthenium-to-
anthraquinone electron transfer (k3 = 2 × 1010 s−1) than
ruthenium-to-triarylamine back-electron transfer (k4 = 6.7 ×
107 s−1).
In order to estimate the quantum yield for the formation of
the TAA+−RuII−AQ− state at 1.57 eV out of the initially
excited TAA−*RuII−AQ state, two pieces of information are
yet missing: (i) the rate constant for the reductive excited-state
G
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
quenching to form the TAA−RuIII−AQ− state at 2.19 eV
(process “2” in Scheme 3a) and (ii) the decay rate constant of
the Ru(bpy)32+ 3MLCT excited state (process “6” in Scheme
3a). The rate constant for process “6” may simply be estimated
2+
from the lifetime of the isolated Ru(bpy)3 complex in
deoxygenated acetonitrile (1.2 × 106 s−1). The rate constant for
process “2” is more difficult to obtain. We have found that the
only viable possibility in this case involves the use of time-
resolved luminescence spectroscopy: Figure 6d compares the
luminescence decays of the RuII reference complex (black
trace) and the RuII−AQ dyad (green trace), detected at 610 nm
after the excitation at 532 nm with laser pulses of ∼10 ns width.
The luminescence decays with a lifetime of 830 ns in the case of
the reference complex, and with a lifetime of 300 ns in the case
of the RuII−AQ dyad. The difference between the two
luminescence decay rate constants (2.1 × 106 s−1) is taken as
the rate constant for process “2” in Scheme 3. Unfortunately,
the AQ− anion cannot be detected for the RuII−AQ dyad,57
presumably because of rapid thermal back-electron transfer in
the opposite sense; this appears to be not an uncommon
problem for ruthenium−quinone dyads.25,57,74 In the triad,
AQ− is formed after TAA+ (50 ps vs 9 ps, see above); hence in
the picosecond transient absorption data (Figure 6a, b), one
observes the formation of the final charge-separated state at
1.57 eV rather than the TAA−RuIII−AQ− state at 2.19 eV.
With numerical estimates for the rate constants of the
processes “1” through “6” from Scheme 3a at hand (second
column of Table 4), we estimate a quantum yield of 99.7% for
the formation of the final charge-separated state. Essentially
every photon put into the Ru(bpy)32+ 3MLCT state at 2.12 eV
thus leads to the formation of a molecule in the TAA+−RuII−
AQ− state at 1.57 eV.
Figure 7. (a) Rise of the transient absorption signals at 550 nm (green
trace) and 770 nm (red trace) after excitation of the TAA−OsII−AQ
triad at 532 nm with laser pulses of ∼10 ns width (CH3CN solution).
(b) Decay of the transient absorption signal of the OsII−AQ dyad at
550 nm (AQ− disappearance) after excitation at 532 nm with laser
pulses of ∼10 ns width (deoxygenated CH3CN solution).
constant of 20 ns (Figure 7a, red trace). Assuming process “2”
has k2 ≈ 108 s−1 and further assuming that the TAA−OsIII−
AQ− state at 1.76 eV must be formed before the fully charge-
separated state is accessible, we arrive at the conclusion that the
rate constant for process “3” is ∼108 s−1. The TAA+−OsII−
AQ− state at 1.58 eV then decays with a lifetime of 80 ns (see
prior section), corresponding to a rate constant of 1.3 × 107 s−1
for process “5” in Scheme 3b. The complete set of rate
constants for the TAA−OsII−AQ triad in the third column of
Table 4 leads us to the conclusion that the fully charge-
separated state is formed with a quantum yield of ∼46% out of
the initially excited Os(bpy)32+ 3MLCT state.
Kinetics and Quantum Yields for Formation of
Charge-Separated States in the Iridium Triad. For the
iridium triad, the situation is fundamentally different from that
for the ruthenium and osmium triads: we were unable to
selectively excite the metal complex in TAA−IrIII−AQ. Even at
the comparatively long wavelength of 420 nm there is an
absorption from the organic moieties; this is particularly
evident from a comparison of the (ground-state) absorption
spectrum of the IrIII reference complex with those of the TAA−
IrIII dyad and the TAA−IrIII−AQ triad (Figure 1c). Instead of
Scheme 3c, we therefore use an energy level diagram for the
TAA−IrIII−AQ triad which has been adapted to reflect this
additional complication (Scheme 4). In this more complex
scheme, we introduce an additional state named *(TAA−IrIII)−
AQ, which is supposed to reflect the possibility that initial
excitation may involve the entire triarylamine-iridium(III)
fragment and not just solely the IrIII complex. In addition to
the evidence from absorption spectra, evidence for electronic
interaction between the iridium complex and TAA was
obtained from the electrochemical measurements, where a
clear shift in the iridium oxidation potential was observed for
the dyads and the triad (see above). Furthermore, in Scheme 4
we omit the TAA−IrIV−AQ− state at 2.10 eV because there is
no experimental evidence for its formation in the triad, see
below.
Kinetics and Quantum Yields for Formation of
Charge-Separated States in the Osmium Triad. In the
case of the osmium triad, a completely different situation is
encountered. The reductive excited-state quenching by TAA is
endergonic in this sample, and we have been unable to find any
evidence for the formation of the TAA+−OsI−AQ state at 1.92
eV, from investigations of both the TAA−OsII dyad and the
TAA−OsII−AQ triad, in the pico- to millisecond time regimes.
We conclude from this observation that the rate constant for
process “1” in Scheme 3b amounts to less than 10% of the
inherent Os(bpy)32+ 3MLCT decay rate constant, which is 5.3
× 107 s−1 under these experimental conditions (deoxygenated
CH3CN at room temperature; process “6” in Scheme 3b).51
In transient absorption experiments performed with the
OsII−AQ dyad and the TAA−OsII−AQ triad, the radical anion
of AQ (monitored at 550 nm) is formed with a time constant of
∼10 ns (Figure 7a, green trace). There is a technical problem
associated with this finding: on the one hand, 10 ns is too long
to be measured accurately with our femtosecond equipment,
and on the other hand, 10 ns is too short to be detected on our
nanosecond setup with reliable accuracy. In this awkward
situation we tentatively attribute a rate constant of ∼108 s−1 to
process “2” from Scheme 3b but note that error bars are rather
large in this particular case.
The red trace in Figure 8a is the transient absorption
spectrum detected with a delay of 1 ps after the excitation of
TAA−IrIII−AQ in CH3CN at 420 nm. This spectrum shows an
absorption band at 770 nm, which we have identified above as
due to TAA+. There is an additional absorption between 500
and 650 nm, which cannot be accounted for by the oxidized
triarylamine unit (compare to Figure 4d). It appears plausible
to attribute this additional absorption to the reduced metal
complex, i.e., the IrII species. In the TAA−IrIII dyad, a similar
In the OsII−AQ dyad, the AQ− signal at 550 nm decays with
a time constant of 26 ns (Figure 7b), and consequently we
estimate a rate constant of 3.8 × 107 s−1 for process “4” in
Scheme 3b, which corresponds to back-electron transfer from
reduced anthraquinone to OsIII.
For the TAA−OsII−AQ triad, the transient absorption
intensity at 770 nm, due to TAA+, builds up with a time
H
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
3 ps (green trace in Figure 8a) must be chemically very similar
to each other. In Scheme 4, we designate the state formed after
0.4 ps as *(TAA+−IrII)−AQ, while the state formed after 3 ps is
denoted as TAA+−IrII−AQ. In other words, after 0.4 ps a
charge-separated state is already formed, but this state
undergoes subsequent electronic relaxation with a time
constant of 3 ps. In our opinion, this explanation makes
sense in view of the fact that photoexcitation of the TAA−IrIII−
AQ triad (contrary to the ruthenium and osmium systems)
cannot occur selectively at the metal center but seems to
involve the entire TAA−IrIII fragment, see above. Thus, in
Scheme 4, we attribute a rate constant of 3.5 × 1012 s−1 to
process “1” and a rate constant of 3.3 × 1011 s−1 to process “2”
(last column of Table 4).
Scheme 4. Energy Level Scheme Showing the Relevant
Photoexcited and Charge-Separated States Which Can Be
Formed in the Iridium Triad
Once formed, the relaxed TAA+−IrII−AQ state at 2.05 eV
can either undergo charge-recombination to the ground state
(process “4” in Scheme 4) or it can proceed to the final charge-
separated state at 1.56 eV (process “3”). Experiments on the
TAA−IrIII dyad indicate that the back-electron transfer between
the oxidized TAA and reduced iridium takes place with a time
constant of 67 ns; the respective transient absorption decay
data are shown in Figure 8c. We infer that in the triad, process
“4” occurs with a rate constant of 1.5 × 107 s−1 (last column of
Table 4).
Kinetic information regarding the formation of the fully
charge-separated state at 1.56 eV can be extracted from the
purple trace in Figure 8d, which shows the time profile of the
optical density at 550 nm, i.e., at one of the absorption band
maxima of the AQ− species. The respective time profile shows
an initial rapid rise and a decay due to the formation of the
*(TAA+−IrII)−AQ and TAA+−IrII−AQ states, which also
absorb at this wavelength (Figure 8a). Subsequently, there is
a slower rise with a time constant of 40 ps, which is attributed
to the build-up of the TAA+−IrIII−AQ− population. At the
same time, the optical density at 770 nm stays essentially
constant (blue trace in Figure 8d), consistent with the
formation of the fully charge-separated state. A rate constant
of 2.5 × 1010 s−1 is therefore attributed to process “3” in
Scheme 4.
The rate constant for process “5”, i.e., thermal charge-
recombination from the TAA+−IrIII−AQ− state at 1.56 eV, is
1.1 × 106 s−1 (lifetime of 890 ns, see prior section). The rate
constant for process “6”, i.e., relaxation of the photoexcited
iridium complex to the electronic ground state, is estimated
from the luminescence lifetime of the IrIII reference complex
(230 ns in oxygen-free acetonitrile; k6 = 4.4 × 106 s−1).31
On the basis of the rate constants for the individual
photophysical and photochemical processes in Scheme 4 (last
column of Table 4), we arrive at the conclusion that the fully
charge-separated state is formed in essentially quantitative yield
from the initially photoexcited state.
Figure 8. (a) Red trace: Transient absorption spectrum detected with
a delay of 1 ps after excitation of an acetonitrile solution of TAA−IrIII−
AQ at 420 nm. Green trace: transient absorption spectrum from the
same sample detected with a delay of 3 ps. (b) Time profile of the
transient absorption at 770 nm from the same sample after excitation
at 420 nm with laser pulses of 150 fs width. (c) Decay of the transient
absorption at 770 nm after excitation of the TAA−IrIII dyad (in
deoxygenated CH3CN) at 355 nm with ∼10-ns laser pulses. (d) Time
profiles of the transient absorption at 550 nm (purple trace) and 770
nm (blue trace) after 420-nm excitation of the TAA−IrIII−AQ triad in
CH3CN (laser pulse width: 150 fs).
SUMMARY AND CONCLUSIONS
transient absorption spectrum can be detected in a 200-ns time
window starting immediately after a 10-ns laser pulse (data not
shown).
■
Final charge-separated states containing an oxidized triaryl-
amine fragment and a reduced anthraquinone moiety are
formed in all three triads from Scheme 2, albeit with different
quantum yields and via differing reaction mechanisms involving
different kinetics. The thermodynamics of the photoinduced
charge-separation steps are such that reductive quenching of
the initially excited ruthenium state is clearly favored kinetically,
while oxidative quenching is predominant in the case of the
osmium system. In the iridium triad both the reductive and
oxidative excited-state quenching steps are thermodynamically
The time profile of the optical density at 770 nm after
excitation of the iridium triad at 420 nm with femtosecond laser
pulses is shown in Figure 8b. From the initial rise, we extract a
time constant of 0.4 ps. Subsequently, there is a decrease in the
ΔOD at this detection wavelength, occurring with a time
constant of 3 ps. The spectral changes that occur in this time
regime are minor (green trace in Figure 8a); hence, the
electronic states formed after 0.4 ps (red trace in Figure 8a) and
I
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
TAA−OsII. Obtained in 68% yield (53 mg) from 30 mg of free
ligand29 and 35 mg of Os(bpy)2Cl2. 1H NMR (300 MHz, CD2Cl2, 25
°C): δ [ppm] 1.79 (s, 3 H, CH3), 1.90 (s, 3 H, CH3), 3.73 (s, 6 H,
OCH3), 6.76 (m, 9 H), 6.95 (s, 1 H, xy), 7.36 (m, 5 H), 7.61 (m, 6 H),
7.86 (m, 6 H), 8.41 (m, 6 H). ES-MS: m/z 495.66 (calculated 495.66
for C52H45N7O2Os2+). Anal. Calcd. for C52H45N7O2OsP2F12: C, 48.79;
H, 3.54; N, 7.66. Found: C, 48.44; H, 3.47; N, 7.50. (The abbreviation
“xy” in the NMR data stands for aromatic protons of the p-xylene
units).
possible, but the reductive pathway dominates kinetically. In
the ruthenium and iridium systems, the driving forces
associated with the formation of initial charge-separated states
are sufficiently large to make photoinduced electron transfer
the dominant excited-state deactivation pathway, particularly in
view of the comparatively long 3MLCT lifetimes of the
2+
Ru(bpy)3 and [Ir(2-(p-tolyl)pyridine)2(bpy)]+ photosensi-
tizers. From the initial charge-separated states, the formation of
the final charge-separated state is kinetically favored versus
thermal recombination in all three cases, which may be a
manifestation of an inverted driving-force effect.47,67 These
favorable circumstances lead to the formation of the final
charge-separated state with quantum yields near unity, at least
in the case of the ruthenium and iridium systems. The osmium
OsII−AQ. Obtained in 49% yield (64 mg) from 50 mg of free
ligand29 and 61 mg of Os(bpy)2Cl2. 1H NMR (300 MHz, CD3CN, 25
°C): δ [ppm] 2.02 (s, 3 H, CH3), 2.25 (s, 3 H, CH3), 7.12 (s, 1 H, xy),
7.22 (s, 1 H, xy), 7.32 (m, 5 H), 7.56 (d, J = 1.5 Hz, 1 H), 7.67 (m, 4
H), 7.83 (m, 4 H), 7.90 (m, 6 H), 8.12 (d, J = 1.6 Hz, 1 H), 8.29 (m, 3
H), 8.50 (m, 6 H). ES-MS: m/z 485.133 (calculated 485.135 for
C52H38N6O2Os2+). Anal. Calcd. for C52H38N6O2OsP2F12·1.5 H2O: C,
48.56; H, 3.21; N, 6.53. Found C, 48.66; H, 3.13; N, 6.49.
TAA−OsII−AQ. Obtained in 72% yield from 30 mg of free ligand29
3
triad, by contrast, suffers from a much shorter MLCT lifetime
2+
of the Os(bpy)3 sensitizer and a significantly lower driving
1
force for formation of the initial charge-separated state, leading
to a quantum yield around 0.46 for formation of the fully
charge-separated state.
and 21 mg of Os(bpy)2Cl2. H NMR (300 MHz, CD2Cl2, 25 °C): δ
[ppm] 1.81 (s, 3 H, CH3), 1.91 (s, 3 H, CH3), 1.97 (s, 3 H, CH3), 2.34
(s, 3 H, CH3), 3.74 (s, 6 H, OCH3), 6.77 (m, 8 H, amine), 6.99 (s, 1
H), 7.14 (s, 1 H), 7.16 (s, 1 H), 7.35 (m, 2 H), 7.46 (m, 3 H), 7.57
(m, 2 H), 7.67 (m, 2 H), 7.73 (m, 1 H), 7.82 (m, 6 H), 7.92 (m, 4H),
8.20 (m, 1 H), 8.30 (m, 3 H), 8.45 (m, 4 H), 8.59 (m, 2 H). ES-MS:
m/z 650.71 (calculated 650.71 for C74H59N7O4Os2+). Anal. Calcd. for
C74H59N7O4OsP2F12·2H2O: C, 54.64; H, 3.90; N, 6.03. Found: C,
54.41; H, 3.79; N, 6.04.
The lifetimes of the fully charge-separated states are in the
microsecond regime in the ruthenium and iridium triads. Three
factors may be responsible for these slow recombination
kinetics: (i) an inverted driving-force effect,47,67 (ii) a long
electron−hole separation distance (∼22 Å),68,69 and (iii) spin
selection rule.70 In the case of the osmium triad, the lifetime of
the fully charge-separated state is more than an order of
magnitude shorter than in the ruthenium and iridium systems,
possibly because of a relatively small energy gap between the
TAA−OsIII−AQ− state and the TAA+−OsII−AQ− state. It thus
appears that a large energy gap (here, > 0.45 eV) to the
energetically next higher lying electronic state is another
important ingredient for obtaining a long-lived final charge-
separated state. We think this is an important new finding; one
would have expected much more similar lifetimes for the
charge-separated states of the three triads. The fact that we
were able to compare a nearly isostructural series of linear
donor−sensitizer−acceptor compounds is a significant advant-
age in this context.
TAA−IrIII. Obtained in 73% yield (28 mg) from refluxing 28 mg of
organic ligand29 with 30 mg of [Ir(2-(p-tolyl)pyridine)2Cl]2 in a
30
mixture of ethanol (10 mL) and chloroform (3 mL). After cooling to
room temperature and the addition of a saturated aqueous solution of
KPF6, a yellow-orange solid formed. This solid was filtered, washed
with water and diethylether, and dried under a vacuum. 1H NMR (300
MHz, CD2Cl2, 25 °C): δ [ppm] 1.80 (s, 3 H, CH3), 1.92 (s, 3 H,
CH3), 2.11 (m, 6 H, CH3), 3.73 (s, 6 H, OCH3), 6.08 (s, 1 H), 6.15 (s,
1 H), 6.78 (m, 9 H), 6.93 (m, 5 H), 7.43 (m, 1 H), 7.56 (m, 4 H), 7.74
(m, 2 H), 7.88 (m, 2 H), 8.05 (m, 4 H), 8.51 (m, 2 H). ES-MS: m/z
1016.35 (calculated 1016.35 for C56H49N5O2Ir+). Anal. Calcd. for
C56H49N5O2IrPF6·H2O: C, 57.04; H, 4.36; N, 5.94. Found: C, 57.10;
H, 4.22; N, 5.87.
IrIII−AQ. Obtained in 80% yield (24 mg) from 27 mg of organic
30
ligand29 and 30 mg of [Ir(2-(p-tolyl)pyridine)2Cl]2 following the
procedure described above for TAA−IrIII. 1H NMR (300 MHz,
CD3CN, 25 °C): δ [ppm] 1.99 (s, 3 H, CH3), 2.07 (s, 3 H, CH3), 2.11
(s, 3 H, CH3), 2.27 (s, 3 H, CH3), 6.12 (s, 1 H), 6.17 (s, 1 H), 6.88
(m, 2 H), 7.01 (m, 2 H), 7.16 (s, 1 H), 7.23 (s, 1 H), 7.52 (m, 1 H),
7.63 (m, 1 H), 7.69 (m, 3 H), 7.83 (m, 3 H), 7.90 (m, 2 H), 8.00 (m, 4
H), 8.17 (m, 3 H), 8.29 (m, 3 H), 8.57 (m, 2 H). ES-MS: m/z 995.29
(calculated 995.29 for C56H42N4O2Ir+). Anal. Calcd. for
C56H42N4O2IrPF6·0.3CHCl3: C, 57.50; H, 3.63; N, 4.76. Found: C,
57.74; H, 3.41; N, 4.73.
Future work on these systems will focus on the role of
coupling of intramolecular photoinduced electron transfer to
bimolecular proton transfer with reduced anthraquinone as a
proton-accepting site. Preliminary results from this work have
been communicated recently.71
EXPERIMENTAL SECTION
■
The syntheses of the organic moieties of the rigid rod-like molecular
triads and dyads (triarylamine-2,2′-bipyridine-anthraquinone unit for
the triad; triarylamine-2,2′-bipyridine and 2,2′-bipyridine-anthraqui-
none units for the dyads) were described in detail in the Supporting
Information to one of our previous publications.29 Reaction of the
individual functionalized bpy ligands with Ru(bpy)2Cl2, Os(bpy)2Cl2,
and [Ir(2-(p-tolyl)pyridine)2Cl]2 precursors occurred following stand-
ard protocols.30−32 Briefly, a mixture of the starting materials in
ethylene glycol was refluxed overnight under N2. After cooling to room
temperature, water was added, and the aqueous phase was extracted
with CH2Cl2. The organic phase was dried over MgSO4, and the
solvent was removed under reduced pressure. Product purification
occurred by column chromatography on silica gel using a mixture of
acetone/water/aqueous saturated KNO3 solution (90/9/1) as the
eluent. The desired product was precipitated from the aqueous
solution (after acetone removal) by the addition of saturated aqueous
KPF6 solution.
TAA−IrIII−AQ. Obtained in 87% yield (34 mg) from 46 mg of
organic ligand29 and 30 mg of [Ir(2-(p-tolyl)pyridine)2Cl]230 following
1
the procedure described above for TAA−IrIII. H NMR (300 MHz,
CD2Cl2, 25 °C): δ [ppm] 1.81 (s, 3 H, CH3), 1.93 (s, 3 H, CH3), 1.97
(s, 3 H, CH3), 2.10 (m, 6 H, CH3), 2.28 (s, 3 H, CH3), 3.75 (s, 6 H,
OCH3), 6.15 (m, 2 H), 6.78 (m, 8 H, C6H4), 6.88 (m, 2 H), 6.99 (m,
3 H), 7.17 (m, 2 H), 7.62 (m, 4 H), 7.77 (m, 3 H), 7.83 (m, 2 H), 7.89
(m, 2 H), 8.10 (m, 2 H), 8.16 (m, 2 H), 8.30 (m, 5 H), 8.58 (m, 2 H).
ES-MS: m/z 1326.45 (calculated 1326.45 for C78H63N5O4Ir+). Anal.
Calcd. for C78H63N5O4IrPF6: C, 63.66; H, 4.32; N, 4.76. Found: C,
64.00; H, 4.45; N, 4.78.
1H NMR spectroscopy was performed using Bruker Avance DRX
300 and Bruker B-ACS-120 spectrometers. A Finnigan MAT8200
instrument was employed for mass spectrometry, and elemental
analysis was performed on a Vario EL III CHNS analyzer from
Elementar. Cyclic voltammograms were obtained using a Versastat3−
200 potentiostat from Princeton Applied Research. A glassy carbon
disk was used as a working electrode. A silver wire served as a quasi-
reference electrode, and a second silver wire was used as a
Product characterization data for TAA−RuII−AQ, TAA−RuII, and
RuII−AQ (including ligands for the dyads) have been reported
previously.29,57 For all other (new) molecules, they are as follows:
J
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
counterelectrode. Voltage sweeps occurred at rates of 100 mV/s;
solutions were deoxygenated by bubbling N2 gas prior to measure-
ments. Optical absorption spectra were recorded on a Cary 300
spectrometer from Varian. Spectro-electrochemical experiments were
performed using the Cary 300 spectrometer, the potentiostat
mentioned above, and an optically transparent thin-layer (OTTLE)
cell from Specac.72 Steady-state luminescence spectra were measured
on a Fluorolog-3 instrument (FL322) from Horiba Jobin-Yvon,
equipped with a TBC-07C detector from Hamamatsu. Transient
absorption and time-resolved luminescence in the nanosecond time
domain was measured using an LP920-KS instrument from Edinburgh
Instruments. The detection system of the LP920-KS spectrometer
consisted of an R928 photomultiplier and an iCCD camera from
Andor. The excitation source was a Quantel Brilliant b laser
(frequency-doubled or -tripled). Prior to nanosecond time-resolved
measurements, samples were thoroughly deoxygenated by bubbling N2
gas through the solutions, or by using home-built quartz cuvettes and a
freeze−pump−thaw technique for oxygen removal. The sample
absorbance at the excitation wavelength was typically between 0.1
and 0.3. A pump−probe method for time-resolved absorption was
used to detect fast processes with a time resolution of 150 fs. The
femtosecond pulse generator (TISSA50, Avesta/CDP) was pumped
with a continuous wave Nd:YAG second harmonic laser (Verdi-V6,
Coherent). The femtosecond pulses were amplified with a Ti-Sapphire
amplifier (Avesta/CDP) pumped by a Nd:YAG laser (LF114, Solar
TII). After the amplifier, the beam was split in two separate beams.
The first part was passed through a second harmonic generator to
obtain excitation (pump) pulses at 400 or 420 nm, and the second part
was passed through a cuvette with water to generate a white light
continuum as the monitoring (probe) pulse. The excitation beam was
directed to a delay line (Avesta/CDP), enabling measurements of the
transient absorption spectra up to 1 ns after excitation. A
monochromator (Andor 0032) and a CCD camera (Newton
DU920N-BR-DD, Andor) were used to record the spectra. The
sample was placed in a rotating cuvette to prevent any degradation due
to the laser excitation. The obtained time-resolved absorption decay
curves were globally fitted to a sum of exponentials. The instrumental
setup and the data analysis procedure are described in more detail
elsewhere.73
(7) Dixon, I. M.; Collin, J.-P.; Sauvage, J.-P.; Barigelletti, F.; Flamigni,
L. Angew. Chem., Int. Ed. 2000, 39, 1292−1295.
(8) Dixon, I. M.; Collin, J.-P.; Sauvage, J.-P.; Flamigni, L.; Encinas, S.;
Barigelletti, F. Chem. Soc. Rev. 2000, 29, 385−391.
(9) Chakraborty, S.; Wadas, T. J.; Hester, H.; Schmehl, R.; Eisenberg,
R. Inorg. Chem. 2005, 44, 6865−6878.
(10) Goransson, E.; Boixel, J.; Monnereau, C.; Blart, E.; Pellegrin, Y.;
Becker, H. C.; Hammarstrom, L.; Odobel, F. Inorg. Chem. 2010, 49,
̈
9823−9832.
(11) Roundhill, D. M. Photochemistry and Photophysics of Metal
Complexes; Plenum Press: New York, 1994.
(12) Baranoff, E.; Collin, J.-P.; Flamigni, L.; Sauvage, J.-P. Chem. Soc.
Rev. 2004, 33, 147−155.
(13) Flamigni, L.; Collin, J.-P.; Sauvage, J.-P. Acc. Chem. Res. 2008,
41, 857−871.
(14) Abrahamsson, M.; Jager, M.; Osterman, T.; Eriksson, L.;
Persson, P.; Becker, H. C.; Johansson, O.; Hammarstrom, L. J. Am.
̈
Chem. Soc. 2006, 128, 12616−12617.
(15) Abrahamsson, M.; Jager, M.; Kumar, R. J.; Osterman, T.;
Persson, P.; Becker, H. C.; Johansson, O.; Hammarstrom, L. J. Am.
̈
Chem. Soc. 2008, 130, 15533−15542.
(16) Hammarstrom, L.; Johansson, O. Coord. Chem. Rev. 2010, 254,
̈
2546−2559.
(17) Kumar, R. J.; Karlsson, S.; Streich, D.; Jensen, A. R.; Jager, M.;
Becker, H. C.; Bergquist, J.; Johansson, O.; Hammarstrom, L. Chem.
̈
Eur. J. 2010, 16, 2830−2842.
(18) Cooley, L. F.; Larson, S. L.; Elliott, C. M.; Kelley, D. F. J. Phys.
Chem. 1991, 95, 10694−10700.
(19) Opperman, K. A.; Mecklenburg, S. L.; Meyer, T. J. Inorg. Chem.
1994, 33, 5295−5301.
(20) Larson, S. L.; Elliott, C. M.; Kelley, D. F. J. Phys. Chem. 1995,
99, 6530−6539.
(21) Treadway, J. A.; Chen, P. Y.; Rutherford, T. J.; Keene, F. R.;
Meyer, T. J. J. Phys. Chem. A 1997, 101, 6824−6826.
(22) Rutherford, T. J.; Keene, F. R. Inorg. Chem. 1997, 36, 2872−
2878.
(23) Klumpp, T.; Linsenmann, M.; Larson, S. L.; Limoges, B. R.;
Burssner, D.; Krissinel, E. B.; Elliott, C. M.; Steiner, U. E. J. Am. Chem.
̈
Soc. 1999, 121, 1076−1087.
AUTHOR INFORMATION
■
(24) Maxwell, K. A.; Sykora, M.; DeSimone, J. M.; Meyer, T. J. Inorg.
Chem. 2000, 39, 71−75.
Corresponding Author
*E-mail: helge.lemmetyinen@tut.fi (H.L.), oliver.wenger@
chemie.uni-goettingen.de (O.S.W.).
(25) Borgstrom, M.; Johansson, O.; Lomoth, R.; Baudin, H. B.;
̈
Wallin, S.; Sun, L. C.; Åkermark, B.; Hammarstrom, L. Inorg. Chem.
̈
2003, 42, 5173−5184.
(26) Falkenstrom, M.; Johansson, O.; Hammarstrom, L. Inorg. Chim.
Acta 2007, 360, 741−750.
Notes
̈
̈
The authors declare no competing financial interest.
(27) Wenger, O. S. Coord. Chem. Rev. 2009, 253, 1439−1457.
(28) Dupont, N.; Ran, Y. F.; Jia, H. P.; Grilj, J.; Ding, J.; Liu, S. X.;
Decurtins, S.; Hauser, A. Inorg. Chem. 2011, 50, 3295−3303.
(29) Hankache, J.; Wenger, O. S. Chem. Commun. 2011, 47, 10145−
10147.
ACKNOWLEDGMENTS
■
This work was supported by the Swiss National Science
Foundation (SNSF) through Grant 200021-117578, by the
Deutsche Forschungsgemeinschaft (DFG) through Grant
INST186/872-1, and by the Academy of Finland.
(30) Freys, J. C.; Bernardinelli, G.; Wenger, O. S. Chem. Commun.
2008, 4267−4269.
(31) Hanss, D.; Freys, J. C.; Bernardinelli, G.; Wenger, O. S. Eur. J.
Inorg. Chem. 2009, 4850−4859.
REFERENCES
■
(1) Balzani, V. Electron Transfer in Chemistry; VCH Wiley: Weinheim,
Germany, 2001; Vol. 3.
(32) Sullivan, B. P.; Salmon, D. J.; Meyer, T. J. Inorg. Chem. 1978, 17,
3334−3341.
(33) Hanss, D.; Wenger, O. S. Inorg. Chem. 2008, 47, 9081−9084.
(34) Hanss, D.; Wenger, O. S. Eur. J. Inorg. Chem. 2009, 3778−3790.
(35) Geiss, B.; Lambert, C. Chem. Commun. 2009, 1670−1672.
(36) Sreenath, K.; Suneesh, C. V.; Gopidas, K. R.; Flowers, R. A. J.
Phys. Chem. A 2009, 113, 6477−6483.
(37) Meyer, T. J. Pure Appl. Chem. 1986, 58, 1193−1206.
(38) Balzani, V.; Juris, A.; Venturi, M.; Campagna, S.; Serroni, S.
Chem. Rev. 1996, 96, 759−833.
(39) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.;
Von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85−277.
(2) Collin, J.-P.; Guillerez, S.; Sauvage, J.-P.; Barigelletti, F.; De Cola,
L.; Flamigni, L.; Balzani, V. Inorg. Chem. 1991, 30, 4230−4238.
(3) Collin, J.-P.; Guillerez, S.; Sauvage, J.-P.; Barigelletti, F.; Flamigni,
L.; De Cola, L.; Balzani, V. Coord. Chem. Rev. 1991, 111, 291−296.
(4) Collin, J.-P.; Guillerez, S.; Sauvage, J.-P.; Barigelletti, F.; De Cola,
L.; Flamigni, L.; Balzani, V. Inorg. Chem. 1992, 31, 4112−4117.
(5) Sauvage, J.-P.; Collin, J.-P.; Chambron, J.-C.; Guillerez, S.;
Coudret, C.; Balzani, V.; Barigelletti, F.; De Cola, L.; Flamigni, L.
Chem. Rev. 1994, 94, 993−1019.
(6) Collin, J.-P.; Dixon, I. M.; Sauvage, J.-P.; Williams, J. A. G.;
Barigelletti, F.; Flamigni, L. J. Am. Chem. Soc. 1999, 121, 5009−5016.
K
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(40) Lamansky, S.; Djurovich, P.; Murphy, D.; Abdel-Razzaq, F.;
Kwong, R.; Tsyba, I.; Bortz, M.; Mui, B.; Bau, R.; Thompson, M. E.
Inorg. Chem. 2001, 40, 1704−1711.
(72) Krejci
179−187.
̌
k, M.; Danek
̌
, M.; Hartl, F. J. Electroanal. Chem. 1991, 317,
(73) Tkachenko, N. V.; Rantala, L.; Tauber, A. Y.; Helaja, J.;
Hynninen, P. H.; Lemmetyinen, H. J. Am. Chem. Soc. 1999, 121,
9378−9387.
(41) McCusker, J. K. Acc. Chem. Res. 2003, 36, 876−887.
(42) Lowry, M. S.; Bernhard, S. Chem.Eur. J. 2006, 12, 7970−
7977.
(74) Hankache, J.; Hanss, D.; Wenger, O. S. J. Phys. Chem. A 2012,
116, 3347−3358.
(43) Baranoff, E.; Dixon, I. M.; Collin, J.-P.; Sauvage, J.-P.; Ventura,
B.; Flamigni, L. Inorg. Chem. 2004, 43, 3057−3066.
(44) The absorbance at the excitation wavelength was typically in the
range between 0.1 and 0.3. The individual luminescence intensities
were corrected for differences in absorbance at the excitation
wavelength.
(45) Turro, N. J. Molecular Photochemistry; W.A. Benjamin: New
York, 1967.
(46) Marsal, P.; Avilov, I.; da Silva, D. A.; Bredas, J. L.; Beljonne, D.
Chem. Phys. Lett. 2004, 392, 521−528.
(47) Schanze, K. S.; MacQueen, D. B.; Perkins, T. A.; Cabana, L. A.
Coord. Chem. Rev. 1993, 122, 63−89.
(48) Schoonover, J. R.; Dattelbaum, D. M.; Malko, A.; Klimov, V. I.;
Meyer, T. J.; Styers-Barnett, D. J.; Gannon, E. Z.; Granger, J. C.;
Aldridge, W. S.; Papanikolas, J. M. J. Phys. Chem. A 2005, 109, 2472−
2475.
(49) Simon, J. A.; Curry, S. L.; Schmehl, R. H.; Schatz, T. R.;
Piotrowiak, P.; Jin, X. Q.; Thummel, R. P. J. Am. Chem. Soc. 1997, 119,
11012−11022.
(50) Freys, J. C.; Wenger, O. S. Eur. J. Inorg. Chem. 2010, 5509−
5516.
(51) Creutz, C.; Chou, M.; Netzel, T. L.; Okumura, M.; Sutin, N. J.
Am. Chem. Soc. 1980, 102, 1309−1319.
(52) Walther, M. E.; Wenger, O. S. Inorg. Chem. 2011, 50, 10901−
10907.
(53) Furue, M.; Maruyama, K.; Oguni, T.; Naiki, M.; Kamachi, M.
Inorg. Chem. 1992, 31, 3792−3795.
(54) Anderson, P. A.; Keene, F. R.; Meyer, T. J.; Moss, J. A.; Strouse,
G. F.; Treadway, J. A. J. Chem. Soc., Dalton Trans. 2002, 3820−3831.
(55) Sreenath, K.; Thomas, T. G.; Gopidas, K. R. Org. Lett. 2011, 13,
1134−1137.
(56) Babaei, A.; Connor, P. A.; McQuillan, A. J.; Umapathy, S. J.
Chem. Educ. 1997, 74, 1200−1204.
(57) Hankache, J.; Wenger, O. S. Phys. Chem. Chem. Phys. 2012, 14,
2685−2692.
(58) Weller, A. Z. Phys. Chem. 1982, 133, 93−98.
(59) Lambert, C.; Noll, G. J. Am. Chem. Soc. 1999, 121, 8434−8442.
̈
(60) Hankache, J.; Wenger, O. S. Chem. Rev. 2011, 111, 5138−5178.
(61) Lewis, F. D.; Thazhathveetil, A. K.; Zeidan, T. A.; Vura-Weis, J.;
Wasielewski, M. R. J. Am. Chem. Soc. 2010, 132, 444−445.
(62) Mecklenburg, S. L.; McCafferty, D. G.; Schoonover, J. R.; Peek,
B. M.; Erickson, B. W.; Meyer, T. J. Inorg. Chem. 1994, 33, 2974−
2983.
(63) The shorter lifetime of the fully charge-separated state in the
osmium triad explains the lower signal-to-noise ratio in the transient
absorption spectrum of Figure 4b relative to those of Figure 4a/c.
(64) Lancaster, K.; Odom, S. A.; Jones, S. C.; Thayumanavan, S.;
́
Marder, S. R.; Bredas, J. L.; Coropceanu, V.; Barlow, S. J. Am. Chem.
Soc. 2009, 131, 1717−1723.
(65) Nelsen, S. F.; Konradsson, A. E.; Weaver, M. N.; Telo, J. P. J.
Am. Chem. Soc. 2003, 125, 12493−12501.
(66) Nelsen, S. F.; Weaver, M. N.; Zink, J. I.; Telo, J. P. J. Am. Chem.
Soc. 2005, 127, 10611−10622.
(67) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265−
322.
(68) Gray, H. B.; Winkler, J. R. Proc. Natl. Acad. Sci. U. S. A. 2005,
102, 3534−3539.
(69) Wenger, O. S. Acc. Chem. Res. 2011, 44, 25−35.
(70) Weiss, E. A.; Ahrens, M. J.; Sinks, L. E.; Ratner, M. A.;
Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 9510−9511.
(71) Hankache, J.; Wenger, O. S. Chem.Eur. J. 2012,
DOI: 10.1002/chem.201200199.
L
dx.doi.org/10.1021/ic300558s | Inorg. Chem. XXXX, XXX, XXX−XXX