Using intramolecular energy transfer to transform non-photoactive, visible-light-absorbing chromophores into sensitizers for photoredox reactions

Article abstract of DOI:10.1021/ja909785b

This work discusses the synthesis, photophysical behavior, and photoinduced electron-transfer reactivity of multichromophoric molecules having a visible-light-absorbing MLCT component coupled to a ligand with a localized excited state of the same spin multiplicity that serves to lengthen the excited-state lifetime of the complex significantly. The appropriate ligands were prepared by Wittig coupling of a bipyridine derivative with pyrenecarboxaldehyde. The modified ligand, a pyrene-vinyl-bipyridyl ensemble (pyrv-bpy), was then reacted with RuCl₃ to yield [(pyrv-bpy) ₂RuCl₂]. The complex has MLCT absorption out to 800 nm, and excitation results in the formation of a ligand-localized excited state with a lifetime long enough to undergo bimolecular electron-transfer reactions. The pyrenylvinyl "localized" excited state of the complex reacts via photoinduced electron transfer with a variety of viologen and diquat electron acceptors. The remarkable aspect of the electron-transfer process is that whereas the excited state can be considered to be ligand-localized the photoredox reaction almost certainly involves the direct formation of the one-electron-oxidized metal center.

Full text of DOI:10.1021/ja909785b

Published on Web 05/11/2010
Using Intramolecular Energy Transfer to Transform
non-Photoactive, Visible-Light-Absorbing Chromophores into
Sensitizers for Photoredox Reactions
Jing Gu, Jin Chen, and Russell H. Schmehl*
Department of Chemistry, Tulane UniVersity, New Orleans, Louisiana 70118
Received November 18, 2009; E-mail: russ@tulane.edu
Abstract: This work discusses the synthesis, photophysical behavior, and photoinduced electron-transfer
reactivity of multichromophoric molecules having a visible-light-absorbing MLCT component coupled to a
ligand with a localized excited state of the same spin multiplicity that serves to lengthen the excited-state
lifetime of the complex significantly. The appropriate ligands were prepared by Wittig coupling of a bipyridine
derivative with pyrenecarboxaldehyde. The modified ligand, a pyrene-vinyl-bipyridyl ensemble (pyrv-bpy),
was then reacted with RuCl3 to yield [(pyrv-bpy)2RuCl2]. The complex has MLCT absorption out to 800 nm,
and excitation results in the formation of a ligand-localized excited state with a lifetime long enough to
undergo bimolecular electron-transfer reactions. The pyrenylvinyl “localized” excited state of the complex
reacts via photoinduced electron transfer with a variety of viologen and diquat electron acceptors. The
remarkable aspect of the electron-transfer process is that whereas the excited state can be considered to
be ligand-localized the photoredox reaction almost certainly involves the direct formation of the one-electron-
oxidized metal center.
Introduction
energy triplet excited states (in the red or near-infrared) with
lifetimes of microseconds or longer in solution, transition-metal
Transition-metal complex chromophores have been widely
used as sensitizers for photoredox reactions and have become
significant as sensitizers for charge injection into semiconductors
in dye-sensitized solar cells.1-6 A desired characteristic for
chromophores employed in solar energy conversion schemes
is that they absorb light throughout the entire visible spectrum
and have excited-state energies high enough for useful light-
induced redox chemistry.7-10 Unfortunately, many complexes
that absorb well in the red via metal-to-ligand charge-transfer
transitions (MLCT) also have very short excited-state lifetimes
that serve to diminish their utility as sensitizers for light-induced
electron-transfer reactions. Reasons for the rapid nonradiative
relaxation in transition-metal complex chromophores are well
established and involve either relaxation through a manifold of
ligand-field (LF or MC) excited states or enhanced coupling of
excited-state and ground-state potential surfaces with decreasing
energy gaps (energy gap law limitations).11-14 Although there
are numerous purely organic chromophores with relatively low
complex chromophores generally have much shorter excited-
state lifetimes regardless of the differences in spin between the
ground and excited states because of much higher spin-orbit
coupling interactions and also because of the coupling of charge
transfer and ligand-field or metal-centered (LF or MC) excited
states (i.e., very high density of states).11-16 Nonetheless, the
development of transition-metal complex chromophores is
attractive because many complexes have fully reversible one-
electron oxidation and reduction potentials, making them very
useful as photoredox sensitizers, and many complexes exhibit
visible absorption that can be tuned by systematic ligand
variation.3,7,9,12,15
One approach to extending the excited-state lifetime of metal
complex chromophores is to couple excited states localized on
the metal complex with states of triplet spin multiplicity on a
covalently linked organic (often aromatic hydrocarbon) com-
ponent. The approach was demonstrated in the 1970s by
Wrighton17-20 and has been exploited in a variety of ways since
(1) Williams, J. A. G. Top. Curr. Chem. 2007, 281, 205–268.
(2) Kumaresan, D.; Shankar, K.; Vaidya, S.; Schmehl, R. H. Top. Curr.
Chem. 2007, 281, 101–142.
(12) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von
Zelewsky, A. Coord. Chem. ReV. 1988, 84, 85–277.
(13) Chen, Y. J.; Xie, P.; Heeg, M. J.; Endicott, J. F. Inorg. Chem. 2006,
45, 6282–6297.
(3) Alstrum-Acevedo, J. H.; Brennaman, M. K.; Meyer, T. J. Inorg. Chem.
2005, 44, 6802–6827.
(4) Nazeeruddin, M. K.; Gratzel, M. Struct. Bonding (Berlin) 2006, 123,
113–175.
(14) Chen, P.; Meyer, T. J. Chem. ReV. 1998, 98, 1439–1477.
(15) Vlcek, A. Coord. Chem. ReV. 2000, 200-202, 933–978.
(16) Caspar, J. V.; Meyer, T. J. Inorg. Chem. 1983, 22, 2444–2453.
(17) Wrighton, M. S.; Morse, D. L.; Pdungsap, L. J. Am. Chem. Soc. 1975,
97, 2073–2079.
(5) Li, B.; Wang, L.; Kang, B.; Wang, P.; Qiu, Y. Sol. Energy Mater.
Sol. Cells 2006, 90, 549–573.
(6) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2001, 34, 40–
48.
(18) Pdungsap, L.; Wrighton, M. S. J. Organomet. Chem. 1977, 127, 337–
347.
(7) Kudo, A.; Miseki, Y. Chem. Soc. ReV. 2009, 38, 253–278.
(8) Barber, J. Chem. Soc. ReV. 2009, 38, 185–196.
(9) Ardo, S.; Meyer, G. J. Chem. Soc. ReV. 2009, 38, 115–164.
(10) Hofmeier, H.; Schubert, U. S. Chem. Soc. ReV. 2004, 33, 373–399.
(11) Meyer, T. J. Pure Appl. Chem. 1986, 58, 1193–206.
(19) Giordano, P. J.; Fredericks, S. M.; Wrighton, M. S.; Morse, D. L.
J. Am. Chem. Soc. 1978, 100, 2257–2259.
(20) Fredericks, S. M.; Luong, J. C.; Wrighton, M. S. J. Am. Chem. Soc.
1979, 101, 7415–7417.
9
7338 J. AM. CHEM. SOC. 2010, 132, 7338–7346
10.1021/ja909785b  2010 American Chemical Society

Sensitizers for Photoredox Reactions
A R T I C L E S
Scheme 1. The Ru(II) cis-chloro complexes were prepared using
a variation of literature methods.35 RuCl3 ·nH2O was refluxed
in 2-methoxyethanol with 2 equiv of bipyridine under an Ar
atmosphere for 8 h. The complex was purified by reprecipitation
and characterized by UV-vis, cyclic voltammetry, and ESI mass
spectrometry (HRMS). The ethylenediamine complexes were
prepared from the cis-chloro derivative via reaction with excess
ethylene diamine (en) in refluxing ethanol, also under Ar. The
complexes were purified by reprecipitation and characterized
by the above techniques.
Figure 1. Energy-level diagram for complexes having nearly isoenergetic
MLCT and IL triplet states.
then to yield complexes of Ru(II),21-28 Os(II),29,30 and Pt(II)31,32
with very long excited-state lifetimes. Depending on the energy
gap between the states involved, the system either can be in
equilibrium or can be effectively localized on the pendant ligand
triplet (Figure 1).33,34 In this article, we report the exploitation
of intramolecular energy transfer from MLCT excited states to
the ligand-localized triplet state to achieve long-wavelength
visible-light-absorbing chromophores with significant excited-
state lifetime enhancements relative to that of the complex
having only the MLCT state. In addition, the work illustrates
that the complexes are active sensitizers for light-induced
electron-transfer reactions.
The chromophores are derivatives of Ru(II) bipyridine
complexes in which the bipyridine ligand has been substituted
with a vinyl pyrenyl moiety (pyrv-bpy, Figure 2, Scheme 1)
that provides a low-energy triplet excited state to serve as an
excitation energy reservoir in the complexes. The chromophores
reported here are [(bpy)2RuCl2] and [(bpy)2Ru(en)](PF6)2 (en
) ethylenediamine) and the corresponding pyrv-bpy complexes.
Spectrophotometry and Electrochemistry. Absorption spectra
of the cis-chloro and en complexes in acetonitrile are shown in
Figures 3 and 4, along with the spectrum of [(pyrv-bpy)3Zn]2+
.
The Zn complex is included because it illustrates the absorption
of the pyrv-bpy ligand when coordinated to a metal lacking
MLCT or ligand-field transitions. Absorption maxima and
absorbtivities are given in Table 1. Ligand pyr-bpy and complex
[(pyr-bpy)2RuCl2] are also included to illustrate the importance
of the ethylene substituent in the molecule.
The transition observed near 300 nm for both ligands and all
of the complexes can be readily assigned as a bipyridine-
localized π-π* transition, analogous to similar transitions
observed for nearly all bipyridine-containing M(II) complexes.12
For ligand pyr-bpy, the absorption spectrum includes a pyrene-
localized π-π* transition with a maximum at 342 nm, and the
pyrv-bpy ligand has a maximum much farther to the red, at
390 nm, reflecting electronic interaction between pyrene and
bipyridine. The long-wavelength maximum of the Zn complex
at 424 nm is also a ligand-localized π-π* transition; the lower
energy of the transition relative to pyrv-bpy may be related to
some degree of charge-transfer character between the pyrene
and the coordinated bipyridine, similar to that clearly observed
and characterized in terpyridyl phenylene vinylene complexes
of Zn(II).36
Results and Discussion
Ligand and Complex Syntheses. The ligand pyrv-bpy was
prepared from 4,4′-dimethyl-2,2′-bipyridine by NBS bromona-
tion of the methyl, the formation of diethyl phosphonate from
bromide, and the reaction of phosphonate with pyrene-1-
carboxaldehyde under Knovenagle conditions as shown in
The spectra of the Ru(II) complexes have two visible
absorption maxima; the longest-wavelength absorption maxi-
Figure 2. Representation of one stereoisomer of the pyrv-bpy and pyr-bpy ligands and the complexes.
J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010 7339
9

A R T I C L E S
Gu et al.
Scheme 1. Ligand Synthesis
mum (λ > 500 nm) can be assigned as an MLCT transition on
the basis of a literature precedent37 and computational analysis
(density functional calculations section). The absorption between
350 and 450 nm is very likely of MLCT origin for [(bpy)2-
RuCl2]38 but represents the overlap of the localized pyrv-bpy
ligand and MLCT transitions for the pyrv-bpy complex. The
absorptivity of both visible transitions is >104 M-1cm-1 for the
pyrv-bpy complexes at their maxima, indicating that the
transitions are strongly allowed.
The cis-chloro complexes exhibit slight solvatochromism
associated with the long-wavelength MLCT absorption transi-
tion; Table 3 presents the absorption maxima for both visible
transitions in a variety of solvents. The magnitude of the
solvatochromism is not large, ranging from 580 nm in aceto-
nitrile to 606 nm in benzonitrile, and the observed spectral
behavior does not seem to correlate with any measure of solvent
polarity such as Gutmann solvent donor or acceptor numbers
or the ET(30) scale.39
mmetry and differential pulse voltammetry; potentials are
included in Table 1. Ligands pyr-bpy and pyrv-bpy both exhibit
irreversible oxidation (as does pyrene); the oxidation is pyrene-
localized because the parent ligand, 2,2′-bipyridine (bpy), and
a model for pyrv-bpy, 4-(styryl)-2,2′-bipyridine, have no oxida-
tive waves at potentials less than +1.8 V versus SCE.12
The [(bpy)2RuCl2], [(pyr-bpy)2RuCl2], and [(pyrv-bpy)2RuCl2]
cis-chloro complexes all exhibit reversible one-electron oxida-
tion at <0.4 V versus SCE in acetonitrile (Table 1). On the basis
of earlier work by others, the oxidation is assigned to the
oxidation of Ru(II) to Ru(III).16 The pyrene-containing com-
plexes also have an irreversible oxidation wave in cyclic
voltammograms at potentials corresponding to pyrene oxidation
(1.4 to 1.6 V vs SCE).24 All of the complexes had several
overlapping quasi-reversible reductive waves between -1.4 and
-2.0 V. By analogy to complexes reported in the literature,
the waves correspond to the reduction of the bipyridine ligands.
Complex [(pyrv-bpy)2RuCl2] was examined by spectroelectro-
chemistry in order to determine the spectrum of and evaluate
the stability of the Ru(III) form of the complex. Figure 5 shows
visible absorption spectra obtained before and after oxidation
at +0.5 V (SCE). The two charge-transfer bands of the Ru(II)
complex disappear, and a new transition is observed with a
maximum of 435 nm. A precipitate formed during the oxidation
The redox behavior of the ligands and complexes was
examined in DMF solution by a combination of cyclic volta-
(21) Ford, W. E.; Rodgers, M. A. J. J. Phys. Chem. 1992, 96, 2917–2920.
(22) Tyson, D. S.; Castellano, F. N. J. Phys. Chem. A 1999, 103, 10955–
10960.
(23) Tyson, D. S.; Castellano, F. N. Abstracts of Papers, 220th ACS
National Meeting, Washington, DC, Aug 20-24, 2000.
(24) Simon, J. A.; Curry, S. L.; Schmehl, R. H.; Schatz, T. R.; Piotrowiak,
P.; Jin, X.; Thummel, R. P. J. Am. Chem. Soc. 1997, 119, 11012–
11022.
(25) Harriman, A.; Hissler, M.; Khatyr, A.; Ziessel, R. Chem. Commun.
1999, 735–736.
(26) Hissler, M.; Harriman, A.; Khatyr, A.; Ziessel, R. Chem.sEur. J. 1999,
5, 3366–3381.
(27) Morales, A. F.; Accorsi, G.; Armaroli, N.; Barigelletti, F.; Pope,
S. J. A.; Ward, M. D. Inorg. Chem. 2002, 41, 6711–6719.
(28) Benniston, A. C.; Harriman, A.; Lawrie, D. J.; Mayeux, A. Phys. Chem.
Chem. Phys. 2004, 6, 51–57.
(29) Maubert, B.; McClenaghan, N. D.; Indelli, M. T.; Campagna, S. J.
Phys. Chem. A 2003, 107, 447–455.
(30) Balazs, G. C.; del Guerzo, A.; Schmehl, R. H. Photochem. Photobiol.
Sci. 2005, 4, 89–94.
(31) Michalec, J. F.; Bejune, S. A.; McMillin, D. R. Inorg. Chem. 2000,
39, 2708–2709.
(32) Pomestchenko, I. E.; Luman, C. R.; Hissler, M.; Ziessel, R.; Castellano,
F. N. Inorg. Chem. 2003, 42, 1394–1396.
(33) Wang, X. Y.; Del Guerzo, A.; Schmehl, R. H. J. Photochem.
Photobiol., C 2004, 5, 55–77.
(34) Schmehl, R. Spectrum 2000, 13, 17–21.
(35) Leidner, C. R.; Murray, R. W.; Meyer, T. J. Inorg. Chem. 1987, 26,
882–891.
(36) Wang, X.-Y.; Del Guerzo, A.; Schmehl, R. H. Chem. Commun. 2002,
2344–2345.
(37) Klassen, D. M.; Crosby, G. A. J. Chem. Phys. 1968, 48, 1853–1858.
(38) Odongo, O. S.; Heeg, M. J.; Chen, Y. J.; Xie, P.; Endicott, J. F. Inorg.
Chem. 2008, 47, 7493–7511.
Figure 3. UV-vis absorption spectra in CH3CN of (A) [Ru(bpy)2Cl2] and
(B) [Zn(pyrv-bpy)3]2+ (yellow ---, ε/5) and [Ru(pyrv-bpy)2Cl2] (green -).
(39) Gutmann, V. The Donor Acceptor Approach to Molecular Interactions,
Springer, New York, 1978.
9
7340 J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010

Sensitizers for Photoredox Reactions
A R T I C L E S
Figure 5. Acetonitrile UV-vis spectra of [(pyrv-bpy)2Ru(II)Cl2] (---) and
[(pyrv-bpy)2Ru(III)Cl2]+ (-) obtained by oxidation at +0.5 V vs SCE.
Figure 4. UV-vis absorption spectra in CH3CN of (A) [Ru(pyrv-bpy)2Cl2]
and (B) [Zn(pyrv-bpy)3]2+ (yellow ---, ε/5) and [Ru(pyrv-bpy)2en]2+ (purple
-).
process, preventing the determination of the molar absorptivity
of the Ru(III) species, but reduction clearly led to the appearance
of the original absorption spectrum (diminished somewhat
because of precipitate formation).
Table 1. Electronic Spectral Data and Potentials for One-Electron
Oxidation of the Ligands and Complexes in Room-Temperature
Acetonitrile
Voltammetric analysis of complex [(pyrv-bpy)2Ru(en)]2+
between 0 and +1.8 V versus SCE revealed two irreversible
oxidative waves. In contrast to the cis-chloro complexes, no
reversible ruthenium-based oxidation was observed. The more
positive oxidation wave corresponded closely to the pyrene-
localized oxidation observed in [(pyrv-bpy)2RuCl2]. The lower
potential oxidation, with a peak potential of 0.97 V, is attributed
to the oxidation of the coordinated ethylenediamine (en) ligand.
This assignment is based upon the report of Meyer and co-
workers indicating that [(bpy)2Ru(en)]2+ is quasi-reversibly
oxidized in acetonitrile. Exhaustive electrolysis indicates that
each coordinated amine is oxidized to an imine as shown in
Scheme 2.40 Although we have not isolated the resulting imine
complex resulting from the multielectron oxidation of [(pyrv-
bpy)2Ru(en)]2+, our thinking at this time is that the irreversibility
of the oxidation results in the same chemistry as that of the
unsubstituted bipyridine complex reported by Meyer and co-
workers.
compound
Abs λmax, nm (log ε) E0(ox) V vs SCE trans. Abs τ, ns
pyrv-bpy
290 (4.50)
390 (4.65)
280 (4.65)
342 (4.41)
288 (4.66)
376 (4.20)
550 (4.19)
300 (4.59)
390 (4.41)
582 (4.10)
300 (4.61)
412 (4.55)
520 (4.35)
312
1.4a
pyr-bpy
1.6a
[(bpy)2RuCl2]
0.31b
<10
620
750
[(pyrv-bpy)2RuCl2]
0.21
1.4a
[(pyrv-bpy)2Ru(en)]2+
0.97a
1.4a
[(pyrv-bpy)3Zn]2+
[(pyr-bpy)2RuCl2]
424
344 (4.56)
564 (4.11)
0.3
<10
1.6a
a Irreversible; anodic peak potential reported; from ref 24. b Tanaka,
M.; Nagai, T.; Miki, E. Inorg. Chem. 1989, 28, 1704.
Excited-State Behavior. A detailed account of the lumines-
cence behavior of [(bpy)2RuCl2] and [(bpy)2Ru(en)]2+ in frozen
Table 2. Calculated and Experimental Values for Hydrocarbon, Ligand, and Complex Singlet-Triplet Energy Gaps and Comparison with
Experimental Valuesa
ligand/complex (solvent)
E singlet (hartrees)
-615.7731
E triplet (hartrees)
-615.6958
∆E (S-T) (cm-1
16 863b
)
∆E (nm)
% error
pyrene
593
593 (exp)
572
581(exp)
682
0
16 863 (exp)
stilbene
-540.7099
-924.2164
-540.6315
-924.1497
17 480b
1.1
2.3
17 200 (exp)
14 660c
pyrv-phenyl
(1-styryl-pyrene)
Pyrv-bpy
14 300 (exp)
14 440
10 580
699 (exp)
690
945
-1226.6896
-1500.3447
-1114.6422
-1226.6236
-1500.2995
-1114.5825
[Zn(pyvmb)(NH3)4]2+
[(bpy)2RuCl2] (EtOH)
13 100
763
8.4
4.0
14 283 (exp)
15 725
15 110 (exp)
700 (exp)
636
662 (exp)
[(bpy)2Ru(en)]2+ (CH3CN)
-1274.91049
-1274.07165
a Values are based upon the energies of geometry-optimized structures for the ground-state singlet and triplet states. See the text. b From Murov, V.
Handbook of Photochemistry; Dekker: New York, 1973; p 27. c From Arai, T.; Tokumaru, K. Chem. ReV. 1993, 93, 23.
9
J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010 7341

A R T I C L E S
Gu et al.
Scheme 2. Irreversible en Oxidation in [(bpy)2Ru(en)]2+
Table 3. Spectral Characteristics of [(pyrv-bpy)2RuCl2] in Various
Solvents at Room Temperature
IL maximum nm
(ε, M-1cm-1
MLCT maximum nm
(ε, M-1cm-1
τ
(TA), ns
solvent
)
)
benzonitrile
394
606
322
620
535
7.6
10
(2.53 × 104)
390
(1.51 × 104)
580
acetonitrile
n-butyronitrile
acetone
(2.56 × 104)
388
(1.72 × 104)
582
(2.58 × 104)
390
(1.718 × 104)
588
(2.56 × 104)
394
(1.700 × 104)
586
Figure 6. Transient difference spectra of [(bpy)2RuCl2] in CH3CN solution
at room temperature at 18 (-9-), 20 (-b-), and 22 ns (-2-) following
excitation at 550 nm. The lower trace shows the UV-vis spectrum of the
complex in acetonitrile.
DMF
(2.53 × 104)
(1.706 × 104)
matrices was reported recently by Endicott and co-workers13
and was also described by Crosby and Klassen in the late
1960s.41 The reported emission maxima of [(bpy)2RuCl2] and
[(bpy)2Ru(en)]2+ were 14 280 cm-1 (700 nm) and 15 110 cm-1
(662 nm), respectively; these values are close to those reported
earlier by Crosby and Klassen.41 Excited-state lifetimes were
not reported. We examined the transient absorption spectra in
solution at room temperature for [(bpy)2RuCl2] using nanosec-
ond laser flash photolysis. Transient spectra at different times
after excitation are shown in Figure 6. The spectrum exhibits
the absorption of the excited state between 350 and 400 nm
and bleaching of the ground state absorption in the visible. The
ground-state bleaching is clearly observed by comparison of
the transient spectrum with the UV-vis spectrum of the
complex, also shown in Figure 6. The excited-state absorption
is similar to that reported for [Ru(bpy)3]2+, which has been
assigned as absorption of the bipyridine anion radical associated
42,43
3
with the MLCT state of the complex.
The similarity of
the spectrum of [(bpy)2RuCl2] to that of the tris-bipyridine
complex suggests that the state observed in the transient
3
difference spectrum is the MLCT state. The lifetime of this
state is <10 ns and is not clearly resolvable with our equipment.
The transient difference spectrum of [(pyrv-bpy)2RuCl2]
sharply contrasts with that of [(bpy)2RuCl2]. As can be seen in
Figure 7, the spectrum exhibits excited-state absorption in the
UV and throughout the visible; bleaching of the ground-state
absorbance is observed only at wavelengths near the maximum
of the UV-vis spectrum. In addition, the excited-state lifetime
in N2-purged acetonitrile is 620 ns, which is at least 60 times
longer than the lifetime of [(bpy)2RuCl2] under similar condi-
tions. The transient spectrum resembles that of numerous other
Ru(II) diimine complexes having aromatic hydrocarbon sub-
stituents. The lowest-energy excited state of these complexes
Figure 7. Transient difference spectra of [(pyrv-bpy)2RuCl2] in CH3CN
solution at room temperature at 144 (-9-), 507 (-b-), and 640 ns (-2-)
following excitation at 550 nm. The lower trace shows the UV-vis spectrum
of the complex in acetonitrile.
are often based on the relationship of the reported diimine-linked
3
aromatic hydrocarbon π-π* state energy (from phosphores-
3
cence) to the energy of the MLCT state of closely related
“parent” chromophores that are luminescent. In this system, the
3
energy of the MLCT state is approximated from Endicott’s
recent publication of the emission of [(bpy)2RuCl2].13 However,
we have not been able to obtain either room-temperature or 77
K phosphorescence spectra of [(pyrv-bpy)2RuCl2], the pyrv-
bpy ligand, or the Zn(II) complex of the pyrv-bpy ligand. As a
result, we lack a definitive measure of the energy for the ligand-
localized triplet state. To address this problem, the energies of
the ligand-localized triplet states were estimated via density
functional methods described below.
3
is proposed to be either a ligand-localized π-π* state or a
triplet intraligand charge-transfer (3ILCT) state. The assignments
(40) Brown, G. A.; Weaver, T. R.; Keene, R. F.; Meyer, T. J.; Keenan,
W. R. J. Inorg. Chem. 1976, 15, 190–196.
(41) Klassen, D. M.; Crosby, G. A. J. Chem. Phys. 1968, 48, 1853–1855.
(42) Creutz, C.; Chou, M.; Netzel, T. L.; Okumura, M.; Sutin, N. J. Am.
Chem. Soc. 1980, 102, 1309–1319.
The ethylenediamine complex [(pyrv-bpy)2Ru(en)]2+ also
exhibits transient behavior indicating that the lowest-energy
excited state is ligand-localized. Figure 8 illustrates the transient
(43) Bensasson, R.; Salet, C.; Balzani, V. J. Am. Chem. Soc. 1976, 98,
3722–3724.
9
7342 J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010

Sensitizers for Photoredox Reactions
A R T I C L E S
geometries of the two states. Although this method is very
simple, it provides reliable values of the triplet-state energies
for the structurally related aromatic hydrocarbon references used.
Table 2 lists calculated and experimental values for three
aromatic hydrocarbons as well as calculated values for pyrv-
bpy and a Zn(II) complex containing pyrv-bpy, [(NH3)4Zn(pyrv-
bpy)]2+. The triplet energy for the Zn(II) complex was calculated
because the lowest-energy state is ligand-localized but reflects
an increased degree of pyrene-to-bipyridine charge-transfer
interaction following the coordination of the bipyridine ligand
to Zn(II). The result is that the triplet energy of the pyrv-bpy
coordinated to Zn(II) is significantly lower than the triplet energy
of the free ligand.
For the Ru(II) complexes, two approaches were employed
to estimate excited-state energies. A method similar to that for
the ligands was used, except that the ground state and lowest-
energy triplet were calculated for the complex in the presence
of solvent (using a polarized continuum model of Gaussian 03).
The hybrid B3LYP functional was used, and the basis set was
LANL2DZ for all of the atoms. We were unable to use a higher-
level basis set for the lighter elements because the computation
times became prohibitively long. By calculating the energies
of the solvated singlet and lowest-energy triplet, a measure of
the zero-zero energy is obtained. Our experience using this
method with various tris-diimine chromophores of known
emission energy is that the energy calculated in this way is
higher than the experimental luminescence maximum energy
by 0.22 ( 0.04 eV (1870 cm-1). In the case of the cis-chloro
complex, the computationally determined energy is lower than
the experimental energy reported by Endicott (Table 2).
An alternate approach to estimating the energy of the lowest-
energy triplet was via time-dependent density functional calcula-
tions on the geometry-optimized singlet ground states of the
chromophores. In this case also, the method used was B3LYP
and the basis set for the light elements was 6-31G(d,p) whereas
that for Ru and Cl was LANL2DZ. Time-dependent DFT
calculations for the ground-state optimized geometry included
acetonitrile solvent using the polarized continuum model
included with Gaussian 03. The result provides a measure of
the vertical transition energies for the complexes (ground-state
singlet to lowest-energy triplet vertical transitions). For
[(bpy)2RuCl2], the lowest-energy transition is at 690 nm (14 500
cm-1), slightly higher than the experimental value (Table 2).
The orbitals of the transition (from the CI coefficient expansions)
clearly indicate the lowest singlet-triplet transition to be Ru
dπ to bpy π* MLCT. The ground-state to lowest-energy triplet
transition for the en complex, [(bpy)2Ru(en)]2+, is also higher
in energy than the experimental values at 602 nm (16 600 cm-1),
and as with the cis-chloro complex, the transition is very clearly
MLCT. Time-dependent analysis of the [(bpy)(pyrv-bpy)RuCl2]
complex indicates that the transition to the lowest-energy triplet
lies at 750 nm (13 300 cm-1), which is in the middle of the
triplet energy range derived from triplet quenching experiments
(vide supra). For this molecule, the CI coefficient expansion
indicates that the transition has contributions from MLLCT, Ru
dπ to Pyrv-bpy π*, and a rather diffuse dπ, Pyrv-bpy π-to-
Pyrv-bpy π* transition. The influence of the styryl pyrene
substituent is readily apparent in lowering the energy of the
transition and in increasing the degree of ligand localization in
the transition.
Figure 8. Transient difference spectra of [(pyrv-bpy)2Ru(en)]2+ in CH3CN
solution at room temperature at 130 (black -0-), 450 (-b-), 840 (blue -0-),
and 1300 ns (green -0-) following excitation at 550 nm. The lower trace
shows the UV-vis spectrum of the complex in acetonitrile.
difference and ground-state spectra for this complex in room-
temperature CH3CN. This complex also has bleaching at
wavelengths where ground-state absorbance is strongest and
excited-state absorbance at wavelengths longer than 550 nm.
Triplet-State Energy Approximation via Quenching Stu-
dies and Density Functional Methods. A key difficulty in
characterizing the lowest-energy long-lived excited state of these
complexes is the fact that we have not directly observed
phosphorescence for the pyrv-bpy ligand or either of the
complexes. In addition, attempts to obtain the ligand triplet state
energy using time-resolved photoacoustic calorimetry (via
collaboration) yielded no definitive energy. As a result, we have
employed triplet quenching kinetic measurements and hybrid
density functional calculations to bridge the gap between
structurally related systems with known triplet excited-state
energies and the pyrv-bpy complexes reported here. Triplet
quenching studies were carried out on the coordinated ligand
in the complex [(pyrv-bpy)2RuCl2] because no triplet formation
was observed upon direct photolysis of the ligand; thus, the
assumption is made that the triplet energy of this complex is
the same as the triplet energy of the ligand. Rate constants for
quenching of the transient absorption of the complex with
tetracene (ET ) 10 250 cm-1), perylene (ET ) 12 400 cm-1),
and 9,10-dibromoanthracene (ET ) 14 060 cm-1) were found
to be 3 × 109, 2 × 109, and 2 × 107 M-1 s-1, respectively. No
quenching was observed with anthracene (ET ) 14 900 cm-1).
The implication is that the triplet-state energy is between 12 500
and 14 000 cm-1. On the basis of the luminescence results of
Endicott, the excited state of the complex is almost certainly
3
lower than that of the MLCT state of the parent complex
[(bpy)2RuCl2] (>14 000 cm-1). With the limited number of data
points, we decided not to fit the data to a kinetic free-energy
dependence in an attempt to determine the triplet energy.
The computational approach employed was to estimate the
ligand ground state-lowest-energy triplet state zero-zero
energies using the hybrid B3LYP functional with an intermediate
basis set of 6-31G(d,p).44 The energy difference was taken as
the difference in energy of the gas-phase or solution-optimized
The calculations help support the concept that the lowest-
energy transition on the pyrv-bpy containing complexes is, to
a significant extent, localized on the pyrv-bpy ligand. In addition,
(44) Frisch, M. J., et al. Gaussian 03, ReVision D.02; Gaussian, Inc.:
Wallingford, CT, 2004.
9
J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010 7343

A R T I C L E S
Gu et al.
Figure 9. Transient difference spectra of [(pyrv-bpy)2RuCl2] in (A) acetone between 10 and 30 ns and (B) benzonitrile between 20 and 500 ns.
the energy gap between this lowest-energy triplet state and the
higher-energy MLCT state is certainly greater for the en
complex than for the cis-chloro complex.
functional results for the Zn(II) complex of pyrv-bpy (Table 2)
to be 1.3 eV (10 500 cm-1). If the excited state of the complex
is ligand-localized and the excited state is to be used in the
reduction of an electron acceptor, then a question that arises as
to whether the oxidation of the chromophore will occur on the
ligand or at the more easily oxidized metal center. There is a
very large difference in the potentials for oxidation of the pyrv-
bpy ligand (∼1.4 V) and the Ru(II) center (0.2 V). If excited-
state electron transfer initially results in the oxidation of the
pyrv-bpy ligand, then the excited state will be a very weak
reducing agent. If we use 1.3 eV as the energy of the pyrv-bpy
ligand-localized state and 1.4 V versus SCE as the E0 for the
one-electron reduction of the coordinated pyrv-bpy ligand cation
radical, then the potential for oxidation of the excited state (eq
1) is negative and the ligand-localized excited state should not
be able to reduce methyl viologen.
Solvatochromism in Transient Spectra of [(pyrv-bpy)2-
RuCl2]. The electronic spectra of [(bpy)2RuCl2] exhibit solva-
tochromism in the Ru-to-bpy MLCT transition. A question that
arises relates to the relative energy of MLCT and ligand-
localized excited states as a function of solvent. Figure 9 shows
spectra of the complex in benzonitrile and acetone; the ben-
zonitrile spectrum clearly shows the transient spectrum indica-
tive of the ligand-localized excited state. The difference
spectrum in acetone has a very short excited-state lifetime and
bleaching of the MLCT absorption between 550 and 600 nm
and a unique transient absorbance at 470 nm. In all of the nitrile
solvents examined, the transient spectrum reflects a ligand-
localized excited state (long lifetime and TA with strong
absorbance in the red) whereas in acetone and DMF the
photoactive excited state was MLCT (short lifetime with 470
nm absorbance and bleaching in the green). The spectrum in
benzonitrile (Figure 9B) shows that, within a few nanoseconds
of laser excitation, a short-lived absorbance feature at 470 nm
is observed that yields to the longer-lived ligand-localized triplet
transient spectrum. The complex is soluble in only a limited
number of solvents, and it was not possible to make a
meaningful correlation of the relative-state energy with the
solvent donor number or other solvent parameters. A final note
relating to this is that the photolysis of the complex in CH2Cl2
resulted in a color change from green to yellow; the final
spectrum is similar to that of the one-electron oxidized complex
(Figure 5). The implication of this observation is that the
photoexcited pyrv-bpy complex is capable of reducing the
solvent, although we have not examined this in detail.
[(pyrv+•-bpy)2Ru(II)Cl2] + e-
h
[(pyrv-bpy)(*pyrv-bpy)Ru(II)Cl2] E0 ≈ 0.1 V (1)
If, however, the excited-state electron transfer involves the direct
oxidation of Ru(II), then the excited-state potential (eq 2), as
expressed below, may be as positive as +1.1 V versus SCE.
[(pyrv-bpy)2Ru(III)Cl2] + e-
h
(pyrv-bpy)(*pyrv-bpy)Ru(II)Cl2] E0 ≈ -1.1 V (2)
To address this question, we examined the quenching of the
transient absorbance of [(pyrv-bpy)2RuCl2] with a variety of
electron-accepting quaternary ammonium salts in acetonitrile.
-1
Rate constants, given in Table 4, are all greater than 108 M
s-1 except for that for quencher 5 (E0 ) -1.1 V vs SCE), which
was somewhat slower. Thus, quenching is still facile for electron
acceptors with a potential of -0.65 V (quencher 4), clearly
Photoinduced Electron Transfer of [(pyrv-bpy)2RuCl2] to
Electron Acceptors. The energy of the ligand-localized excited
state of [(pyrv-bpy)2RuCl2] can be approximated from density
9
7344 J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010

Sensitizers for Photoredox Reactions
A R T I C L E S
pletely offset by the bleaching of the oxidized Ru complex. The
lower segment of Figure 10 shows the experimental spectrum
of the radical ion products and the spectrum calculated by first
generating the spectrum of the known components (the reduced
viologen minus the ground-state Ru(II) complex), assuming the
transient spectrum of the radical ion products at 600 nm consists
of only reduced viologen and bleached Ru(II), using the
absorptivity to calculate the concentration of radical ions, and
finally using this concentration to generate the transient differ-
ence spectrum in terms of molar absorptivity. The result is that
the transient absorbance data provides direct evidence that
quenching of the reactive excited state of [(pyrv-bpy)2RuCl2]
by electron acceptors results in the formation of the one-electron-
oxidized Ru complex because the transient species absorbing
at 440 nm mirrors that of the one-electron-oxidized Ru complex
observed spectroelectrochemically. It should be noted that,
although complexes [(bpy)2RuCl2] and [(bpy)2Ru(en)] do not
produce radical ions with low concentrations of acceptors such
as viologens (5 mM or less), it is possible to observe photoredox
products with these chromophores using a much higher quencher
concentration (∼0.1 M).
Figure 10. Transient spectrum of the cis chloro complex in the presence
of 5 mM methyl viologen in nitrogen-purged acetonitrile at different times
after excitation between 50 ns and 500 ns. The final spectrum represents
the spectrum of the charge-separated species. The lower frame represents
a comparison of the experimental (triangles) and calculated transient
spectrum (solid line) of the radical ion products generated using the method
described in the text.
Summary
This work demonstrates that multichromophoric molecules
having a visible-light-absorbing MLCT component coupled to
a UV-absorbing ligand-localized excited state of the same spin
multiplicity can result in sensitization of the ligand-localized
state. The result is an excited state with a lifetime long enough
to undergo bimolecular electron-transfer reactions, as opposed
to the MLCT state of the unmodified complex. With complex
[(pyrv-bpy)2RuCl2], the pyrv “localized” excited state formed
reacts via photoinduced electron transfer with a variety of
electron acceptors. The remarkable aspect of the electron-transfer
process is that, whereas the excited state can be considered to
be ligand-localized, the photoredox reaction almost certainly
involves the direct formation of the one-electron-oxidized metal
center.
Table 4. Viologen and Pyridine Compounds Used as Quenchers
Table 5. Rate Constants for Quenching [(pyrv-bpy)2RuCl2] with
Quaternary Ammonium Electron Acceptors in Acetonitrile at Room
Temperature
Experimental Section
Materials. All starting materials were purchased either from
Aldrich or Strem Chemicals and were used without further
purification. All solvents, including NMR solvents, were used as
obtained. Acetonitrile for electrochemistry was dried by refluxing
over CaH2 under a N2 atmosphere and distilled immediately before
use. Dimethylformamide (DMF) for electrochemistry was dried over
P2O5 overnight and distilled immediately prior to use. Silica columns
were run in the open air using 32-63 µm silica.
quencher
kq × 10-8, M-1 s-1
E0 (Q2+/ Q+), V vs SCE
3
2
1
4
6
5
6.7
4.9
6.81
19.4
1.1
-0.15
-0.24
-0.45
-0.66
-0.84
-1.1
0.74
Spectroscopy. NMR spectra were recorded on a Varian 400 MHz
NMR. Chemical shifts are referenced internally to the solvent signal.
All UV-vis absorption spectra were obtained on a Hewlett-Packard
8452A diode array spectrophotometer. ESI mass spectra were
obtained using a Bruker microTOF spectrometer.
indicating that the reactive excited state of the complex reacts
directly to form the Ru(III) species (and not the pyrv-bpy cation
radical). Further support for this is given in Figure 10, which
shows the transient absorption spectrum of [(pyrv-bpy)2RuCl2]
in acetonitrile containing 5 mM methyl viologen. The spectra
obtained at different times after excitation show the disappear-
ance of the excited state and the formation of a long-lived
transient that exhibits a strong absorption transient signal at 440
nm and small signals for absorption between 350 and 400 nm,
bleaching between 500 and 550 nm, and weak absorbance
further to the red. In general, photoinduced electron transfer
involving methyl viologen results in a strong absorbance
transient at 390 nm and broad absorption between 570 and 700
nm. In this case, the ground-state absorption of the Ru complex
chromophore is very similar to that of reduced methyl viologen
and thus the absorbance of reduced viologen is nearly com-
Electrochemistry. Cyclic voltammetry and differential pulse
voltammetry were obtained using a CH Instruments 630 electro-
chemical workstation. The complexes were dissolved in freshly
distilled, nitrogen-purged solvent with 0.1 M tetrabuylammonium
hexafluorophosphate (TBAH) as the supporting electrolyte. The
working electrode was a freshly polished glassy carbon disk (0.2
cm2). Potentials were measured versus a nonaqueous Ag/Ag+
reference electrode (CH Instruments) but also included ferrocene
is as an internal reference. Spectroelectrochemical analysis was
carried out using the CH Instruments workstation for potential
control and an Ocean Optics USB 2000 spectrophotometer for
spectrophotometric analysis. The cell for spectroelectrochemical
measurements was home-built, featuring a 0.2 cm path length and
9
J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010 7345

A R T I C L E S
Gu et al.
a Pt grid working electrode (50 lines/in) at a Pt coil counter electrode
and a Ag wire pseudoreference.
mg(0.189 mmol) of LiCl in a round-bottomed flask, and 10 mL of
CH3OCH2CH2OH was added. After heating to 125 °C with stirring
for 8 h under an atmosphere of argon, the reaction mixture was
cooled to room temperature and placed in an ice bath overnight.
The dark-brown mixture was filtered, affording a black precipitate
that was washed with water (3 × 5 mL) and diethyl ether (3 × 5
mL). The crude product was redissolved in a minimal volume of
DMSO and precipitated by addition to 20 mL of diethyl ether. The
black-powdered product (25 mg, 52%) was collected by suction
filtration. ESI mass: m/e calcd for C58H40Cl2N4Ru, 964.1678; found,
964.1739. (See Supporting Information for the isotope distribution.)
[(pyrv-bpy)2Ru(en)][PF6]2.38 A 20 mg (0.02 mmol) sample of
[Ru(pyrv-bpy)2Cl2] was suspended in 10 mL of 1:1 (v/v)
water-methanol followed by the addition of 80 mg (1.33 mmol)
of ethylene diamine (en). The resulting mixed solution was refluxed
at 100 °C for 2 h under an atmosphere of argon. After cooling to
room temperature, 5 mL of a saturated aqueous solution of NH4PF6
was added to the reaction mixture. The red precipitate that formed
was collected by suction filtration and washed with distilled water
(3 × 5 mL), followed by diethyl ether (3 × 5 mL). The black-
powdered product (11 mg, 44%) was obtained by filtration. ESI
Mass: m/e calcd for C60H48N6Ru, 477.1494; found, 477.1548.
Time-Resolved Transient Absorption Spectra. Transient ab-
sorption measurements were performed using a Brilliant B Q-
switched Nd:YAG laser-pumped OPO (Opotek) as the pump source
at a right angle to the analyzing light source. Excitation pulses were
generally <5 ns between 410-600 nm. An Applied Photophysics
LKS.60 laser flash photolysis spectrometer was used for detection;
the instrument is equipped with a 150 W pulsed Xe lamp as the
analyzing light source, a single grating monochromator after the
sample, and PMT detection. The output was recorded on an Agilent
Infiniium transient digitizer and analyzed with Applied Photophysics
LKS.60 software.
Syntheses. The preparation of [(bpy)2RuCl2] and [Ru(bpy)2-
(en)](PF6)2 followed modifications of literature procedures and is
included in the Supporting Information.38,40
4-(Pyrenylvinyl)-4′-methyl-2, 2′-bipyridine (pyrv-bpy).45 Po-
tassium tert-butoxide (0.23 g, 2 mmol) was added to a THF (20
mL) solution of 1-pyrenecarboxaldehyde (0.23 g, 0.99 mmol) and
2 (0.30 g, 0.937 mmol). The resulting mixture was allowed to stir
for 24 h at ambient temperature. The solvent was then removed
under reduced pressure. The crude residue was dissolved in a
minimal volume of dichloromethane and dispersed in 30 mL of
hexane. The precipitate was collected by filtration and washed with
distilled water (3 × 5 mL), followed by ether (3 × 5 mL). This
Acknowledgment. We thank the U.S. Department of Energy,
Office of Chemical Sciences (grant DE-FG-02-96ER14617) for
supporting this research. We also thank the National Science
Foundation (grant CHE0619770) for funding for the ESI mass
spectrometer. J.G. thanks the IBM Corporation for a fellowship in
Computational Science administered through the Tulane Center for
Computational Science.
1
gave 0.18 g of light-yellow product 3 (yield 48.5%). H NMR: δ
8.70 (d, JH-H ) 4.5 Hz, 1H), 8.67 (s, 1H), 8.60 (d, JH-H ) 4.5 Hz,
1H), 8.53 (dd, Jvinyl-H-H ) 16 Hz, JH-H ) 6.8 Hz, 2H), 8.33 (d,
JH-H ) 8 Hz, 1H), 8.29 (s, 1H), 8.21-8.19 (m, 4H), 8.08 (d, JH-H
) 2.4 Hz, 2H), 8.015 (t, JH-H ) 8 Hz, 1H), 8.54 (d, JH-H ) 4.5
Hz, 1H), 7.36 (d, 3Jvinyl-H-H ) 16 Hz,1H), 7.18 (d, JH-H ) 4.5 Hz,
1H), 2.47 (s, 3H).
1
Supporting Information Available: Detailed syntheses, H
[(pyrv-bpy)2RuCl2].35 A 12 mg (0.05 mmol) sample of
RuCl3 ·3H2O was mixed with 40 mg (0.1 mmol) of 3 and 8
NMR and mass spectral data, sample Stern-Volmer quenching
data, and extended references. This material is available free
of charge via the Internet at http://pubs.acs.org.
(45) Viau, L.; Maury, O.; Le Bozec, H. Tetrahedron Lett. 2004, 45, 125–
130.
JA909785B
9
7346 J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010