Ruthenium complexes of 2-(2′-pyridyl)benzimidazole as photosensitizers for dye-sensitized solar cells

Article abstract of DOI:10.1039/b208289f

N-Alkylated carboxylic acid derivatives of 2-(2′-pyridyl)benzimidazole (pbimH) with different chain-lengths (pbim(CH)_nCO₂H where n = 1-3) and their ruthenium complexes [Ru(bpy)₂(pbim)](PF₆)₂ have been synthesized and characterized. 2D COSY and NOESY NMR spectroscopy were used to aid the assignment of the pbim NMR spectrum. The effect of chain-length on the cyclic voltammetry (CV) was studied and the voltammetry of the parent pbimH complex was re-investigated. The ability of the carboxylic acid groups to bind to TiO₂ coated electrodes was confirmed by the observation of a symmetrical, surface-confined Ru^III/II wave, while the specular reflectance IR revealed a band at 1620 cm^-1 due to the bound carboxylate (COO ... Ti) group. The efficiencies of solar cells using these sensitizers were rather low, due to the distance between the sensitizer and the surface and the inefficient coupling of the charge-separated excited state to the surface. A fall in the cell open-circuit voltage with chain length reflected this distance effect. Time-resolved luminescence spectroscopy indicated that rapid electron injection into the TiO₂ conduction band was occurring (<30 ns), but this is not fast enough to compete effectively with alternative excited state processes.

Full text of DOI:10.1039/b208289f

View Article Online / Journal Homepage / Table of Contents for this issue
Ruthenium complexes of 2-(2Ј-pyridyl)benzimidazole as
photosensitizers for dye-sensitized solar cells†
Hunan Yi, Joe A. Crayston* and John T. S. Irvine
School of Chemistry, University of St. Andrews, St. Andrews, Fife, UK KY16 9ST
Received 27th August 2002, Accepted 26th November 2002
First published as an Advance Article on the web 13th January 2003
N-Alkylated carboxylic acid derivatives of 2-(2Ј-pyridyl)benzimidazole (pbimH) with different chain-lengths
(pbim(CH)nCO2H where n = 1–3) and their ruthenium complexes [Ru(bpy)2(pbim)](PF6)2 have been synthesized and
characterized. 2D COSY and NOESY NMR spectroscopy were used to aid the assignment of the pbim NMR
spectrum. The effect of chain-length on the cyclic voltammetry (CV) was studied and the voltammetry of the parent
pbimH complex was re-investigated. The ability of the carboxylic acid groups to bind to TiO2 coated electrodes was
confirmed by the observation of a symmetrical, surface-confined RuIII/II wave, while the specular reflectance IR
revealed a band at 1620 cmϪ1 due to the bound carboxylate (COO ؒ ؒ ؒ Ti) group. The efficiencies of solar cells using
these sensitizers were rather low, due to the distance between the sensitizer and the surface and the inefficient
coupling of the charge-separated excited state to the surface. A fall in the cell open-circuit voltage with chain length
reflected this distance effect. Time-resolved luminescence spectroscopy indicated that rapid electron injection into the
TiO2 conduction band was occurring (<30 ns), but this is not fast enough to compete effectively with alternative
excited state processes.
field and flows through an external cell to perform useful work.
The back-reaction (rate contant, ker1) is unproductive and
Introduction
In the last decade, an order of magnitude increase in solar
should be minimised. Most cells have high efficiency as ker1
Ӷ
energy conversion efficiencies at dye-sensitized photoelectro-
chemical cells has been realized by attaching ruthenium poly-
pyridyl complexes to high surface area TiO2 electrodes.1 This
discovery has provoked many research studies, which have been
reviewed extensively.2–11 The currently accepted mode of work-
ing of the cell is shown in Fig. 1. After light absorption by the
kinj. The oxidized dye is reduced by an electron donor (IϪ) pres-
Ϫ
ent in the electrolyte (k2). Reduction of the oxidized donor (I3
)
occurs at the counter electrode and the solar cell is therefore
regenerative. A variety of sensitizers has been reported to date,
yet still the most efficient and stable ones are based on Ru()
polypyridyl coordination compounds.1,9,12,13 In particular, the
benchmark complex RuN3, 1, has attracted the most attention.1
This compound possesses metal-to-ligand charge transfer
(MLCT) absorption bands that harvest a large fraction of vis-
ible light. It was the most efficient sensitizer until [Ru(tcterpy)-
(SCN)2]SCN (tcterpy = 4,4Ј,4Љ-tricarboxy-2,2Ј:6Љ,2Љ-terpyrid-
ine) was reported.14 The dye should have an anchoring group
such as silanyl, carboxyl and phosphonato which reacts spon-
taneously with surface hydroxyl groups of oxide surfaces to
form ester linkages that exhibit good stability. The aim of our
work is to improve the efficiency of dye-sensitized solar cells by
designing new sensitizer molecules and by improving the TiO2–
sensitizer interaction. There are many requirements for the dye,
some conflicting.2 In addition to low cost and high stability in
the oxidized, ground, and excited states there are the following
requirements: (1) an excited state π* level that lies at higher
energy (i.e. more negative potential) than the semiconductor
Fig. 1 Schematic representation of electron transitions in a dye-
sensitized solar cell. k0 = rate constant for decay of excited state (S*
S); kinj = rate constant for electron injection into TiO2 (S*
eϪTiO );
2
ker1 = rate constant for electron recombination with oxidized dye
(eϪ ϩ Sϩ); ker2 = rate constant for electron recombination with redox
mediator (eϪ ϩ I3).
dye the excited dye can decay to the ground state radiatively or
non-radiatively (with rate constant k0) or inject an electron into
the semiconductor conduction and become oxidized:
S* ϩ SC
Sϩ ϩ ecbϪ(TiO2) (charge injection) (kinj) (1)
conduction band edge; (2) a positive ground state Ru3ϩ/2ϩ
-
Charge injection can happen in less than 1 ps. The electron is
then swept to the semiconductor bulk by the surface electric
oxidation potential to ensure rapid oxidation by the I3Ϫ donor;
and (3) intense absorption in the solar region, accompanied by
formation of the excited state upon light absorption, regardless
of the excitation wavelength. One method of moving λmax to
longer wavelength is to make E(Ru3ϩ/2ϩ) less positive (use good π
donorligandstoraisethet2g level), butthiswilltendtoslowdown
the reaction with donor (point (2) above). The ligand π* level can
† Electronic supplementary information (ESI) available: details of
the luminescence measurements and solar cell construction and
testing; synthesis details; 1H–1H COSY and NOESY 2D spectra
for [(bpy)2Ru(pbimH)](PF6)2. See http://www.rsc.org/suppdata/dt/b2/
b208289f/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
685

View Article Online
also be lowered by adding substituents. However, too low a level
would inhibit electron injection into the conduction band (point
(1)). Finally, to make the absorption bands more intense,
extended ligands with phenyl group substituents may be used in
order to increase the transition dipole moment.
acetonitrile solution containing ruthenium complexes (5 × 10Ϫ4
M) and TBAP (0.1 M) as a supporting electrolyte. Measure-
ments were carried out using a saturated calomel electrode
(SCE) as the reference with a Luggin capillary and a Pt wire as
the counter electrode. The reference electrode, separated from
the voltammetric cell by a salt bridge, was Ag/AgNO3 (0.01 M
in CH3CN). The electronic spectrum was recorded on a Perkin-
Elmer spectrometer Lambda 14P spectrophotometer. Details
of the luminescence measurements and solar cell construction
and testing may be found in the ESI.†
Another factor that may be important is the distance between
the anchoring group and the semiconductor. A particular
advantage of interfacial charge-separated states at semi-
conductor materials is that the injected electron can be collected
as an electrical response. This forms the basis for new appli-
cations that exploit both the electronic and optical properties
of the sensitized materials, such as charge storage, displays,
chemical sensing and optical switching. To achieve a charge-
separated state to improve the efficiency of energy transfer,
some work has been done on slowing down the charge
recombination process. One strategy is to increase charge-
separation lifetimes by exploring the performance of a more
complex molecular sensitizer Ru(dcpy)2(bpy-PTZ), with a
covalently bound electron donor phenothiazine (PTZ) to
quench the Ru3ϩ hole formed after charge injection.15 This
transfers the hole farther away from the TiO2 surface in order to
reduce the likelihood of back electron transfer. Another strat-
egy is that the distance between the semiconductor electrode
and the dye-sensitizer can be optimized by introducing a bridg-
ing group between them, i.e. –CH2, or –C6H4. In such a way, a
charge separated state can also be achieved after an electron is
injected from the donor to the acceptor. Thus, we are develop-
ing a series of ruthenium polypyridyl complexes with anchoring
groups of different chain-length (2a–d).
Ligand preparations
A representative procedure is given here, see the ESI for a full
listing.
Ethyl(2-pyridin-2-yl-benzimidazol-1-yl)acetate L2Ј. Method
A. To a suspension of 2-(2Ј-pyridyl)benzimidazole (2 g,
10.2 mmol) in 10 ml DMF was added Cs2CO3 (0.33 g, 1 mmol)
and K2CO3 (1.41 g, 10.2 mmol). After stirring for 5 min, ethyl
bromoacetate (1.30 ml, 11.5 mmol) was added dropwise. The
mixture was stirred for 24 h and then evaporated to dryness.
The residue was extracted with ethyl acetate (100 ml) and then
washed with water (3 × 30 ml) and brine (3 × 30 ml), dried in
anhydrous Na2SO4 overnight and the solvent was evaporated.
The product was recrystallized from ethyl acetate–hexane to
afford the product (1.5 g, 52%).
Method B. 2-(2Ј-Pyridyl)benzimidazole (2 g, 10.2 mmol) was
dissolved in 20 ml anhydrous THF under N2. The solution
was cooled to Ϫ78 ЊC and butyllithium (4.5 ml × 2.5 mol LϪ1
,
11.0 mmol) was added with stirring. After addition, the mixture
was kept at Ϫ78 ЊC for 3 h and then ethyl bromoacetate
(1.30 ml, 11.5 mmol) was added dropwise. It was allowed to
warm to room temperature and then refluxed for 3 h under
anhydrous conditions. The mixture was poured into 50 ml water
and extracted with ethyl acetate (3 × 50 ml). The organic phase
was washed with water (3 × 30 ml) and brine (3 × 30 ml), dried
in anhydrous Na2SO4 overnight and the solvent was evapor-
ated. The product was recrystallized from ethyl acetate–hexane
1
(2.0 g, 70%). H NMR (CDCl3) δ 9.34 (d, 1 H), 8.74 (d, 1 H),
8.31 (t, 1 H), 8.12 (t, 1 H), 7.6 (m, 4 H), 5.78 (s, 2 H), 4.22
(q, 2 H), 1.21 (t, 3 H). Anal. calc. for C14H11N3O2: C, 68.33; H,
5.34; N, 14.94. Found: C, 68.34; H, 5.57; N, 14.92%.
In this way, we can find a reasonable chain-length connecting
the Ru complex to the TiO2 nanocrystalline electrode surface to
see if there is an optimum distance which minimizes back elec-
tron transfer without sacrificing a high rate of charge injection.
The pbim ligand possesses an NH group which can be deproton-
ated and derivatised.16,17 The alkylated pbim ligand features in
a number of mixed-ligand Ru sensitizer complexes.18–20 Alkyl-
ation of pbim with pendant carboxylic acid groups of different
spacer lengths is our proposed way to adjust the distance
between the semiconductor and the dye-sensitizer to find the
optimum distance.
Synthesis of complexes
More details are available as ESI.
[Ru(bpy)2(L1)](PF6)2 (2a) (L1 ؍ pbimH). Ru(bpy)2Cl2ؒH2O
(1.0 g, 1.9 mmol) and 2-(2Ј-pyridyl)benzimidazole (0.37 g,
1.9 mmol) were dissolved in 25 ml ethanol and refluxed for 4 h
under argon. After reaction, the solution was cooled to room
temperature and excess KPF6 aqueous solution was added. The
precipitate was collected by filtration and the crude product was
recrystallized from methanol–acetone (1.64 g, 95%).5 UV-Vis
(λmax) 460 nm (ε = 7500 MϪ1 cmϪ1). 1H NMR (DMSO-d6) δ 8.95
(t, 3 H), 8.87 (d, 1 H), 8.72 (d, 1 H), 8.39 (t, 2 H), 8.28 (p, 2 H),
8.22 (t, 1 H), 8.10 (d, 1 H), 7.98 (d, 1 H), 7.95 (d, 1 H), 7.92
(d, 2 H), 7.88 (d, 1 H), 7.73 (t, 1 H), 7.67 (br, 3 H), 7.62 (t, 1 H),
7.52 (t, 1 H), 7.19 (t, 1 H), 5.82 (d, 1 H). Anal. calc. for
C32H25F12N7P2: C, 42.77; H, 2.78; N, 10.91. Found: C, 42.79; H,
2.25; N, 10.73%.
Experimental
Chemicals and materials
RuCl3ؒ3H2O (Aldrich), 2-(2Ј-pyridyl)benzimidazole (Aldrich),
2,2Ј-bipyridine (Aldrich), ethyl bromoacetate (Aldrich), ethyl
acrylate (Fluka), ethyl bromobutylate, butyllithium (Aldrich),
tetra-n-butylammonium hexafluorophosphate (Fluka, electro-
chemical grade) and DMF (Aldrich) were used as received.
THF was refluxed from potassium.
[Ru(bpy)2(L2)](PF6)2 (2b) (L2 ؍ pbimCH2CO2H). This was
prepared by following the same procedure as that for the prep-
aration of [(bpy)2Ru(pbimH)](PF6)2 except that pbimCH2-
1
CO2H was used as the starting material, 93% yield. H NMR
Measurements
(DMSO-d6) δ 8.85 (t, 3 H), 8.76 (d, 1 H), 8.46 (d, 1 H), 8.22
(t, 1 H), 8.10 (br, 4 H), 7.97 (d, 1 H), 7.91 (d, 1 H), 7.82 (d, 1 H),
7.72 (br, 3 H), 7.48 (br, 6 H), 7.17 (t, 1 H), 5.68 (t, 1 H), 5.56
(s, 2 H). Anal. calc. for RuC34H27F12N7O2P2: C, 42.69; H, 2.84;
N, 10.25. Found: C, 42.78; H, 2.60; N, 10.66%.
Cyclic voltammograms was recorded using an EG and G PARC
273 A potentiostat/galvanostat controlled by version 4.11 of
the Electrochemistry Research Software running on a PC. All
the experiments were carried out in degassed HPLC grade
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
686

View Article Online
Table 1 1H chemical shift assignment (δ) of [(bpy)2Ru(pbimH)](PF6)2 in DMSO-d6
H3 H4 H6
H5
Benzene
8.85 (HB3, HA3, HD3), 8.25 (HE4, HB4), 8.14 7.96 (d, HB6), 7.84 7.60 (t, HB5), 7.53 7.40 (t, Hc), 7.08 (t, Hb),
8.75 (d, HC3), 8.59 (HA4, HD4), 8.09 (t, (d, HC6), 7.82 (d, (HE5, HA5, HC5), 7.49 5.68 (d, Ha)
(d, HE3
)
HC4
)
HD6), 7.78 (HA6, Hd), (t, HD5),
7.73 (d, HE6
)
Isolation of the mono-substituted products was achieved by
addition of an aqueous solution of KPF6 to the reaction mix-
ture. Purification of the crude products was accomplished in
most cases by recrystallization from dry methanol–acetone. The
final products tend to be insoluble in most solvents including
water, but dissolve well in acetone.
Results
The synthesis of the carboxylic acid derivatives of pyridylbenz-
imidazole was hampered by the unreactivity of the NH group
(pKa ca. 14 for parent benzimidazole)21 and the possibility of
pyridine quaternization. After deprotonation double alkylation
of the benzimidazole ring is prevented as otherwise the arom-
aticity of the five-membered ring would be lost.22 There seemed
to be no obvious and established route to the missing member
of the series, pbimCO2H, save for reaction with ethyl chloro-
formate.23 However, this failed because when hydrolysed from
its ester form, decarboxylation occurred. Reactions to produce
the other members of the series were carried out in DMF
solution with potassium carbonate as the base, in which the
bromo-substituted ester starting material was at risk of being
hydrolysed. Nevertheless, the reaction was successful in all cases
except for the synthesis of pbim(CH2)2CO2Et (L3Ј). A possible
explanation is that the desired SN2 reaction is competitive with
the elimination reaction to give ethyl acrylate under strongly
basic conditions. It might be easier to carry out the deproton-
ation after complexation, since the pKa falls to ca. 6.824 but
this was not attempted. Instead, this ligand was prepared by
reacting 2-(2Ј-pyridyl)benzimidazole with ethyl acrylate in
a Michael addition reaction with 1,3,4,6,7,8-hexahydro-2H-
pyrimido[1,2-a]pyrimidine as the catalyst (Scheme 1).
NMR Spectroscopy of the pbim complex
The 1H NMR spectrum of [(bpy)2Ru(pbimH)](PF6)2 in DMSO-
d6 is shown in Fig. 2. The chemical shift assignment of the
Fig. 2 1H NMR of the L1 complex [(bpy)2Ru(pbimH)](PF6)2 in
DMSO-d6 (below) and with expanded regions of the pyridyl protons
(above).
protons of the L1 complex (bpy)2Ru(pbimH)(PF6)2 is listed in
Table 1. The assignment of the chemical shift of the protons
was determined by 1H–1H COSY and NOESY 2D spectroscopy
(see ESI). The correlation of each proton was observed in the
1H–1H COSY spectrum, which distinguishes the protons in dif-
ferent positions (H3, H4, H5, H6) on different pyridine rings
(A, B, C, D, E) as defined in the structure in Table 1.
The chemical shifts of the protons on the pbim benzene
rings, however, separate at higher field, especially Ha at 5.68
ppm because of the ring current effect with the benzene ring
facing pyridine A in the complex. First of all, we can determine
the chemical shift of the protons on the benzene ring of pbimH
since they are far apart from other protons from the COSY
spectrum. Ha is upshifted at 5.68 ppm because of the ring cur-
rent we mentioned earlier which arises when pbimH is coordin-
ated to the ruthenium center. By reading across to the left
(or down) from the Ha (5.68 ppm) peak in the COSY spectrum,
we can immediately assign Hb, Hc and Hd peaks. The chemical
shifts of the bipyridine protons separate in groups as H3, H4,
H6 and H5, and each group contains the protons from the
bipyridine rings and the pyridine ring from pbimH. Bipyridine
H3 and H3Јcoupling was observed in the NOESY spectrum,
which holds a key to the assignments of the rest of the protons
by walking around the pyridine rings from the COSY spectrum.
The only H3 signal (δ 8.70, measured in MeCN) without a
cross peak must be HE3, so from the COSY spectrum we can
Scheme 1
It was decided to use the acid form, rather than the ester, to
anchor the complex to the TiO2 surface. Although the TiO2
surface is capable of catalyzing the hydrolysis of esters, IR evi-
dence indicates that only partial hydrolysis occurs,25 and some
loss of solar cell efficiency has been noticed using the ester
rather than the acid.26 Attempts to make the homoleptic
[Ru(L)3]2ϩcomplexes were unsuccessful starting from RuCl3,
using the method for pbimH, presumably due to the potential
for the ligands to bind via the oxygen of the carboxylate group.
Instead, the ligands were reacted with cis-[Ru(bpy)2Cl2]2ϩ and
purified using published methods:27
(2)
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
687

View Article Online
Table 2 Data from UV absorption and emission spectra of the [(bpy)2Ru(pbimH)](PF6)2 complex and its derivatives measured in CH3CN
Emission/nm
(free solution)
Emission/nm
(on TiO2)
Compound
λmax/nm (ε/MϪ1 cmϪ1
)
[(bpy)2Ru(pbimH)](PF6)2
[(bpy)2Ru(pbimC2)](PF6)2
[(bpy)2Ru(pbimC3)](PF6)2
[(bpy)2Ru(pbimC4)](PF6)2
242 (22400), 285 (42800), 315 (17200), 431 (sh), 455 (8400)
243 (21600), 287 (45200), 315 (18400), 432 (sh), 458 (9600)
243 (22500), 283 (46500), 314 (18800), 431 (sh), 457 (15400)
240 (28000), 280 (48800), 315 (19200), 431 (sh), 457 (10600)
653
662
661
653
—
643
648
647
Note: pbimH = L1, pbimC2 = L2, pbimC3 = L3, pbimC4 = L4.
Table 3 Voltammetric data obtained for the oxidation of ruthenium
2-(2Ј-pyridyl)benzimidazole complexes (1 mM) in acetonitrile at a
platinum electrode. The scan rate was at 100 mV sϪ1. Potential
referenced to added ferrocene (∆Ep = 78 mV)
determine all the E ring proton assignments. The unique signal
at δ 8.87 must correspond to the unique pyridine ring C trans to
the benzimidazole group. Then we can use the COSY to assign
all the C and D protons. A similar process yields A and B
assignments.
Compound
Epox/mV
Epred/mV
E1/2/mV
Electronic absorption spectra
[(bpy)2Ru(pbimH)](PF6)2
[(bpy)2Ru(pbimC2)](PF6)2
[(bpy)2Ru(pbimC3)](PF6)2
[(bpy)2Ru(pbimC4)](PF6)2
825
848
820
809
747
768
746
733
786
808
783
771
Data from the absorption spectra for the complexes obtained in
acetonitrile in the visible region are summarized in Table 2. The
low energy absorption band at about λmax = 455 nm could be
assigned to the π* (bpy)
dπ (Ru) metal-to-ligand charge
transfer transition. The higher energy bands, 240 and 285, 315
nm, are the pyridylbenzimidazole and bipyridyl intraligand
π–π* transitions respectively. The visible absorption band of
the pbim complex is blue-shifted relative to that of [Ru(dcpy)2-
(SCN)2] (λmax = 537 nm). The absorption spectrum of the
pbimH complex [(bpy)2Ru(L1)](PF6)2 is shown in Fig. 3.
Fig. 4 Emission spectra of [(bpy)2Ru(pbimC2)](PF6)2 (5 × 10Ϫ5 M) in
MeCN at T = 298 K when excited at 400 nm, (1) free solution and (2)
absorbed on TiO2.
Fig. 3 Absorption spectra of [(bpy)2Ru(L1)](PF6)2 in ethanol–H2O
(20 : 1) in the absence (1) and in the presence (2) of NaOH (mole
ratio to the complex = 2 : 1) at 25 ЊC.
When 100 µl of 0.01 M NaOH was introduced into the solu-
tion, as shown below, the shoulder at 315 nm became a full peak
at 325 nm (Fig. 3), indicating that the concentration of the
deprotonated form of pbimH increased. The peak position at
455 nm in the visible light region decreased while a new peak at
440 nm started to take shape, however, leading to lower energy
shifts of the absorption maxima on deprotonation of the
coordinated pbimH ligand. Similar phenomena were observed
by Haga24 when pimH (2-(2Ј-pyridyl)imidazole) and biimH
(2,2Ј-biimdazole) were used as the coordinating ligands upon
deprotonation. Data from the luminescence spectra of [(bpy)2-
Ru(pbimH)](PF6)2 and its alkylated derivatives are shown in
Table 2. The emission maxima are almost identical, near
660 nm in free solution and about 645 nm adsorbed on a TiO2
film, when excited at 400 nm at room temperature (Fig. 4)
Fig. 5 Voltammetric oxidation of [(bpy)2Ru(pbimC2)](PF6)2 (1 mM)
complex in acetonitrile at a platinum electrode. Ferrocene standard
added. Scan rate 100 mV sϪ1
.
trode is [(bpy)2Ru(pbimC4)](PF6)2 > [(bpy)2Ru(pbimC3)](PF6)2
(≈ [(bpy)2Ru(pbimH)](PF6)2) > [(bpy)2Ru(pbimC2)](PF6)2. This
is because of the electron withdrawing nature of the carboxylic
acid group –CH2CO2H. The electron density of the metal
center is lower than that of [(bpy)2Ru(pbimH)](PF6)2 and thus
it is more difficult to oxidize. For complex [(bpy)2Ru(pbimC4)]-
(PF6)2, however, because the carboxylic group is further away,
–CH2CH2CH2CO2H becomes electron donating rather than
withdrawing, which increases the electron density of the metal
center. A similar observation was reported for the effect of
alkylation of the related tridentate bis(benzimidazole) pyridine
(dbip) ligand.28 The redox potentials are not so different from
that of the pbimH complex. In 2b the CO2H groups is closest to
the pbim ring and this has the effect of raising the potential.
For [Ru(bpy–(CH2)3CO2H)(dmbpy)2]3ϩ the Ru(/) couple was
shifted 60 mV more cathodic than [Ru(bpy–CO2H)(dmbpy)2].27
As the chain-length increases the potential falls, as expected for
Electrochemistry
The data for the oxidation of the pbim complexes are summar-
ized in Table 3. The cyclic voltammogram of the oxidation of
[(bpy)2Ru(pbimC2)](PF6)2 in MeCN is shown in Fig. 5, with
ferrocene added as the potential calibrator. The effect of
N-alkylation of pbim complexes on the Ru3ϩ/2ϩ shows that the
tendency of the pbim complexes to be oxidized at the elec-
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
688

View Article Online
Table 4 Voltammetric data obtained for the reduction of the ruthenium 2-(2Ј-pyridyl)benzimidazole complexes (1 mM) in acetonitrile at a platinum
electrode. The scan rate was 100 mV sϪ1
Redox potential/V (vs. Fc/Fcϩ)
ox
ox
ox
Compound
Epred (Ep
)
E1/2
Epred (Ep
)
E1/2
Epred (Ep
)
E1/2
[(bpy)2Ru(pbimH)](PF6)2
[(bpy)2Ru(pbimC2)](PF6)2
[(bpy)2Ru(pbimC3)](PF6)2
[(bpy)2Ru(pbimC4)](PF6)2
Ϫ1.617 (—)
Ϫ1.901 (Ϫ1.815)
Ϫ2.030 (Ϫ1.950)
Ϫ2.006 (Ϫ1.926)
Ϫ2.003 (Ϫ1.933)
Ϫ1.853
Ϫ1.990
Ϫ1.898
Ϫ1.968
Ϫ2.133 (Ϫ2.061)
Ϫ2.292 (Ϫ2.222)
Ϫ2.254 (Ϫ2.180)
Ϫ2.259 (Ϫ2.179)
Ϫ2.097
Ϫ2.257
Ϫ2.217
Ϫ2.219
Ϫ1.820 (Ϫ1.744)
Ϫ1.808 (Ϫ1.730)
Ϫ1.805 (Ϫ1.731)
Ϫ1.782
Ϫ1.701
Ϫ1.768
the electron-donating effect of an alkyl group. For the related
tridentate bis(benzimidazole) pyridine (dbip) ligand the shift
was 60 mV cathodic.28
[(bpy)2Ru(pbimC2)]2ϩ
[(bpy)2Ru(pbimC2)]3ϩ ϩ eϪ (3)
The alkylated pbim complexes display three one-electron
reduction waves (Table 4; CV shown for the pbimC2complex in
Fig. 6). These are assigned to the reduction of each of the two
bpy ligands, followed by the reduction of the pbim ligands:
[(bpy)2Ru(pbimC2)]2ϩ ϩ eϪ
[(bpy Ϫ)(bpy)Ru(pbimC2)]ϩ (4)
ؒ
[(bpy Ϫ)(bpy)Ru(pbimC2)]ϩϩ eϪ
ؒ
[(bpy Ϫ)2Ru(pbimC2)]0 (5)
ؒ
Fig.
Ru(pbimC2)](PF6)2 (trace b) (1 mM) in acetonitrile at a platinum
electrode. Scan rate 100 mV sϪ1
6
Reduction of [(bpy)2Ru(pbimH)](PF6)2 (trace a), [(bpy)2-
Ϫ
0
Ϫ
Ϫ
[(bpy )2Ru(pbimC2)] ϩ e [(bpy )2Ru(pbimC2 Ϫ)]Ϫ (6)
ؒ
ؒ
ؒ
.
The potential for the reduction of the pbim ligand accords
with the expected trends in the ligand π* energies. In agreement
with this assignment, the third reduction process shows the
most variation between the complexes; that of [Ru(bpy)2-
(pbimC2)](PF6)2 is slightly more negative because of the
electron withdrawing effect from the carboxylic acid group.
In comparison, the complex [RuL2]2ϩ, L = 2,6-bis(benz-
imidazole)pyridine, does not have such straightforward electro-
chemistry, showing one irreversible wave at Ϫ1.70 V vs.
Ag/ Agϩ.28
Somewhat different behaviour is observed for the pbimH
complex (Fig. 6). Only two reversible waves are observed, and
these are preceded by an irreversible reduction. None of these
waves were reported in the original work by Haga, although in a
footnote it was claimed that additional waves at in this region
disappeared when dry alumina was added to the cell.29 We pro-
pose that the irreversible reduction is due to the reduction of
released protons from the pbim ligand (overall reaction shown
in eqn. (7)):
Fig. 7 Reduction of [(bpy)2Ru(pbimH)](PF6)2 (1 mM) in acetonitrile
at carbon (trace a) and Au (trace b) electrodes at 100 mV sϪ1
.
the electrode. The analogous sensitizers that do not possess
such groups display surface coverages that are at least an order
of magnitude lower.25 All of the studied complexes adsorbed
strongly onto oxide surfaces with the exception of the pbimH
parent complex.
[(bpy)2Ru(pbimH)]2ϩ ϩ H2O ϩ 2eϪ
(1) IR and UV-Vis spectroscopy. The infrared spectra of the
TiO2 semiconductor and [(bpy)2Ru(pbimC2)](PF6)2 absorbed
onto the TiO2 semiconductor showed that the asymmetric
stretching ν(CO2Ϫ) at 1605 cmϪ1 indicated the formation of the
ester bond between the carboxylic and the TiO2 hydroxyl
groups. Goodenough et al suggested that a dehydrative coup-
ling reaction between the sensitizer, 4,4Ј-dicarboxyl-2,2Ј-bi-
pyridine, for example, and the surface hydroxyl groups on
rutile TiO2 would yield a ester linkage on the surface with
enhanced electronic coupling between the π* orbital of the bi-
pyridine ring and the Ti 3d orbital manifold of the semi-
conductor.31 The UV-Vis spectrum of the adsorbed complexes
were virtually identical to the solution spectrum. Despite this it
has been shown in other work that the photocurrent action of
sensitizers may reveal a significant red shift upon surface
attachment, consistent with surface stabilization of the MLCT
excited state.1,32
[(bpy)2Ru(pbim)]ϩ ϩ H2(g) ϩ OHϪ (7)
This explanation is related to that proposed for similar cyclic
voltammetry behaviour noted for RuN3 (1) by Wolfbauer et
al.30 They showed that the proton reduction reaction was highly
dependent on the nature of the electrode. If gold or carbon is
used the proton reduction is suppressed to some extent as
shown in Fig. 7. Deprotonation of the pbimH ligand cathodi-
cally shifts the redox potentials of the bpy reduction by about
100 mV. The pbim ligand reduction is shifted by about 350 mV
compared to the alkylated ligand complexes to Ϫ2.6 V. This
region is close to the solvent reduction and the wave appears as
a double wave possibly overlapping with a metal-based reduc-
tion process.
Surface attachment chemistry
In solar cells, the carboxylic acid or ester functional groups are
necessary for achieving a high surface coverage, apart from the
porosity of the nanocrystalline semiconductor that was used for
(2) Cyclic voltammetry. Cyclic voltammograms of the
oxidation of [(bpy)2Ru(pbim)](PF6)2 on a TiO2 semiconductor
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
689

View Article Online
Table 5 Lifetimes (ns) of the dye-sensitizers obtained from both free solution and onto TiO2 films (Γv ≈ 6.0 × 10Ϫ5 mol cmϪ2). Errors 15%, or
greater at the shorter times
MeCN solution
ZrO2 films
TiO2 films
Compound
τ1
τ2
τav
τ1
τ2
τav
τ1
τ2
τav
[(bpy)2Ru(pbimC2)](PF6)2
[(bpy)2Ru(pbimC3)](PF6)2
[(bpy)2Ru(pbimC4)](PF6)2
224
206
297
27.6
18.2
180
148
188
237
367
388
386
81
102
147
299
308
284
88.6
138
109
11.5
16.3
11.1
45.3
66.1
46.1
Fig. 9 Normalized time-resolved photoluminescence spectra recorded
at 650 nm after pulsed at 400 nm excitation of [(bpy)2Ru(pbimC2)]-
(PF6)2 on (1) ZrO2, (2) in free solution and (3) on TiO2 films in
acetonitrile solution.
Fig. 8 Cyclic voltammograms of [(bpy)2Ru(pbimC2)](PF6)2 absorbed
onto a TiO2 semiconductor electrode in 0.1 M tBu4NPF6 CH2Cl2
solution. The data were recorded at scan rates of 20, 40, 80 and 100 mV
sϪ1
.
Much reduced lifetimes were observed on the TiO2 semi-
conductor electrode due to electron injection. The same phen-
omena were observed for pbimC3 and pbimC4 complexes. Our
results are comparable with those of [Ru(bpy)2(4,4Ј-(PO3H2)2-
bpy)](PF6)2 tested on TiO2 film (27 ns) and on ZrO2 film
(960 ns).33 The shorter lifetimes detected on TiO2 are consistent
with efficient charge injection, and it is possible to obtain a
charge injection rate from the difference in the reciprocal life-
electrode are shown in Fig. 8. The complex [(bpy)2Ru(pbimC2)]-
(PF6)2 displays a reversible Ru(/) redox process by cyclic
voltammetry both in fluid solution and when anchored to the
nanocrystalline TiO2 film. The peak-to-peak separation for
the surface-bound sensitizer is almost twice that in the free
solution, for example 186 mV for anchored dye and 80 mV in
solution at 100 mV sϪ1. The redox process is therefore quasi-
reversible. The redox potential, however, decreased from 808 to
708 mV. This means that, once anchored to the semiconductor,
the dye-sensitizer became more easily oxidized. The area under
the peak gives the surface coverage of electroactive complex
(Γe, eqn. (8)).
36
times on TiO2 and ZrO2 but the errors were rather large at
these short timescales close to the limit of the instrument and
no clear trends were discerned from these data. In addition, the
traces may be composites of several kinetic processes such as:
(1) luminescence decay from complexes adsorbed at multiple
sites; (2) luminescence from impurities present in the system
(e.g. released ligands); and (3) charge injection from molecules
not directly attached to the surface. Despite these caveats, the
data suggests that charge injection is possible using our com-
plexes, and the behaviour as solar cell sensitizers should
resemble that of other pendant chain complexes such as
[(bmp)2RuL]2ϩ (bmp = 4,4Ј-dimethyl-2,2Ј-bipyridyl, L = 4-(3-
carboxypropyl)-2,2Ј-bipyridyl).27 These complexes have similar
redox potentials and so the driving force for the back reaction
and the reaction with iodide should be comparable.
(8)
(9)
where Q is the charge, n is the number of electrons transferred,
F is the Faraday constant, S is the area of the electrode; A is the
absorbance at the peak maximum and ε is the extinction co-
efficient. Compared with that calculated from the UV-Vis
absorbance (Γv, eqn. (9)), more than 90% of ruthenium centres
are redox-inactive. As suggested previously33 this inactivity
probably rules out lateral redox hopping between complexes on
the surface of TiO2. However, a recent paper has found much
higher electroactivity in amine dye coatings, indicating lateral
charge transfer.34
Solar energy conversion using the complexes
Disappointingly, the IPCE values that we observed for Grätzel-
type solar cells using the pbim complexes as sensitizers are very
small, <1%. The sensitivity of our equipment at this level was
not sufficient to make a detailed comparison of the complexes.
The open-circuit voltage (Voc) and short-circuit current (Isc)
tested under direct sunlight (AM 1.5) of the solar cells based on
the dye-sensitizer pbim derivative complexes are shown in Table
6. The open-circuit voltage is known to be directly related to the
rate of charge injection relative to the rates of removal of the
injected electron (e.g. by the back reaction).15 The VOC data
clearly fall as: [(bpy)2Ru(pbimC2)](PF6)2 > [(bpy)2Ru(pbimC3)]-
(PF6)2 > [(bpy)2Ru(pbimC4)](PF6)2, the expected trend as the
chain-length increases. It appears that the apparently efficient
charge injection rates suggested by the photoluminescence data
are in fact slower than standard dyes and do not compete well
with intramolecular deactivation processes of the excited state.
Time-resolved photoluminescence spectroscopy
Normalized photoluminescence quenching decays of the dye-
sensitizer [(bpy)2Ru(pbimC2)](PF6)2 both in free solution and
on ZrO2 and TiO2 films are shown in Fig. 9. Data from experi-
ments on the other dye-sensitizer are collected in Table 5. The
lifetimes were obtained from the second-order exponential
decay fit of the spectra. On ZrO2 films, no electron transfer
occurs because the energy level of the conduction band is
higher than that of the excited state of the dye-sensitizer.35
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
690

View Article Online
Table 6 Data from the solar cell of the pbim derivatives of ruthenium
complexes tested under direct sunlight (Γv ≈ 6.0 × 10Ϫ5 mol cmϪ2
St. Andrews for funding Miss Gemma Holliday who helped
with the electrochemical analysis.
)
Compound
Voc/mV
Isc/µA
References
[(bpy)2Ru(pbimC2)](PF6)2
[(bpy)2Ru(pbimC3)](PF6)2
[(bpy)2Ru(pbimC4)](PF6)2
273
177
135
96
60
47
1 M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker,
E. Muller, P. Liska, N. Vlachopoulos and M. Grätzel, J. Am. Chem.
Soc., 1993, 115, 6382–6390.
2 K. Kalyanasundaram and M. Grätzel, Coord. Chem. Rev., 1998,
177, 347–414.
3 J. J. Kelly and D. Vanmaekelbergh, Electrochim. Acta, 1998, 43,
2773–2780.
4 J. E. Moser, P. Bonnote and M. Grätzel, Coord. Chem. Rev., 1998,
171, 245–250.
5 L. M. Peter, E. A. Ponomarev, G. Franco and N. J. Shaw,
Electrochim. Acta, 1999, 45, 549–560.
6 C. A. Bignozzi, R. Argazzi and C. J. Kleverlaan, Chem. Soc. Rev.,
2000, 29, 87–96.
7 C. G. Garcia, J. F. de Lima and N. Y. M. Iha, Coord. Chem. Rev.,
2000, 196, 219–247.
8 C. A. Kelly and G. J. Meyer, Coord. Chem. Rev., 2001, 211, 295–315.
9 A. Hagfeldt and M. Grätzel, Chem. Rev., 1995, 95, 49–68.
10 A. Hagfeldt and M. Grätzel, Acc. Chem. Res., 2000, 33, 269–277.
11 J. M. Stipkala, F. N. Castellano, T. A. Heimer, C. A. Kelly, K. J. T.
Livi and G. J. Meyer, Chem. Mater., 1997, 9, 2341–2353.
12 B. O’Regan and M. Grätzel, Nature, 1991, 353, 737–740.
13 A. Zaban, S. Ferrere and B. A. Gregg, J. Phys. Chem. B, 1998, 102,
452–460.
Further experiments such as time-resolved transient absorption
measurements on the ps timescale will be necessary to confirm
this point.
Low IPCE values have also been observed for sensitizers with
pendant linkage groups; for example, modification of RuN3 by
inserting a phenyl group between the bpy and the carboxylic
acid group was found to decrease the IPCE from over 80% to
only 8%,37 while a propyl group led to a ca. 50% decrease.27
Lowering of the IPCE to 2% or less was observed using proline-
type anchors of length 14–16 Å instead of the usual carboxyl-
ated bpy ligands.33 In contrast, for a series of merocyanine
dye-sensitizers the efficiency increased as the chain length
increased to 18–20 carbon atoms. It was proposed that the back
electron transfer reaction was dominant in this case.38 An inter-
esting study was carried out on [ReCl(CO)3(bpy–(CH2)2CO2H)]
sensitizers.39 It was shown that inserting the first CH2 group
led to a 200-fold drop—much greater than predicted from
Marcus theory of electron-transfer distance dependence.
Thereafter, however, increasing the chain length decreased the
rate in line with theory. A similar trend was observed for Fe
sensitizers.40
14 M. K. Nazeeruddin, P. Pechy and M. Grätzel, Chem. Commun.,
1997, 1705–1706.
15 R. Argazzi, C. A. Bignozzi, T. A. Heimer, F. N. Castellano and
G. J. Meyer, J. Am. Chem. Soc., 1995, 117, 11815–11816.
16 A. Yoshimura, K. Nozaki, N. Ikeda and T. Ohno, Bull. Chem. Soc.
Jpn., 1996, 69, 2791–2799.
A further, perhaps more serious problem with the pbim com-
plexes is that the π* energy level of the pbim ligand is much
higher than that of the other bpy ligands. Evidence for this lies
in the assignment of the first two reduction potentials as being
bpy based. This means that in the excited state the electron
resides on the bpy ligand, rather than the pbim. As a con-
sequence the electron is very far from the anchoring point to the
surface of the TiO2. It has been shown that this is deleterious
to the IPCE; for example, in terpy/biquinoline mixed ligand
sensitizers the IPCE falls from 75.6% to 1.74%.37 The use
of alternative ligands to bipyridine or the preparation of
homoleptic pbim complexes may solve this problem.
17 S. A. Thomson, J. A. Josey, R. Cadilla, M. D. Gaul, C. F. Hassman,
M. J. Luzzio, A. J. Pipe, K. L. Reed, D. J. Ricca, R. W. Wiethe and
S. A. Noble, Tetrahedron, 1995, 51, 6179–6194.
18 M. K. Nazeeruddin, E. Muller, R. Humphry-Baker, N. Vlacho-
poulos and M. Grätzel, J. Chem. Soc., Dalton Trans., 1997, 4571–
4578.
19 O. Kohle, S. Ruile and M. Grätzel, Inorg. Chem., 1996, 35, 4779–
4787.
20 S. Ruile, O. Kohle, P. Pechy and M. Grätzel, Inorg. Chim. Acta, 1997,
261, 129–140.
21 R. K. Boggess and R. B. Martin, Inorg. Chem., 1974, 13, 1525.
22 M. R. Grimmett, in Comprehensive Heterocyclic Chemistry,
A. R. Katritzky and C. W. Rees, eds., 1979, pp. 382–390.
23 K. H. Mayer, Synthesis, 1975, 673.
24 M. A. Haga, Inorg. Chim. Acta, 1983, 75, 29.
25 P. Qu and G. J. Meyer, Langmuir, 2001, 17, 6720–6728.
26 Y. J. Hou, P. H. Xie, B. W. Zhang, Y. Cao, X. R. Xiao and
W. B. Wang, Inorg. Chem., 1999, 38, 6320.
27 T. A. Heimer, S. T. Darcangelis, F. Farzad, J. M. Stipkala and
G. J. Meyer, Inorg. Chem., 1996, 35, 5319–5324.
28 X. M. Xiao, M. A. Haga, T. Matsumurainoue, Y. Ru, A. W. Addison
and K. Kano, J. Chem. Soc., Dalton Trans., 1993, 2477–2484.
29 M. A. Haga and A. Tsunemitsu, Inorg. Chim. Acta, 1989, 164, 137.
30 G. Wolfbauer, A. M. Bond, G. B. Deacon, D. MacFarlane and
L. Spiccia, J. Am. Chem. Soc., 2000, 122, 130.
31 S. Anderson, E. C. Constable, M. P. Dare-Edwards,
J. B. Goodenough, A. Hamnett, K. R. Seddon and R. D. Wright,
Nature, 1979, 280, 571.
32 D. A. Gulino and H. G. Drickamer, J. Phys. Chem., 1984, 88, 1173.
33 S. A. Trammell, J. A. Moss, J. C. Yang, B. M. Nakhle, C. A. Slate,
F. Odobel, M. Sykora, B. W. Erickson and T. J. Meyer, Inorg. Chem.,
1999, 38, 3665.
34 P. Bonhote, E. Gogniat, S. Tingry, C. Barbe, N. Vlachopoulos,
F. Lenzmann, P. Comte and M. Grätzel, J. Phys. Chem. B, 1998, 102,
1498.
35 Y. Tachibana, J. E. Moser, M. Grätzel, D. R. Klug and J. R.
Durrant, J. Phys. Chem., 1996, 100, 20056.
36 R. Argazzi, C. A. Bignozzi, T. A. Heimer, F. N. Castellano and
G. J. Meyer, Inorg. Chem., 1994, 33, 5741.
Conclusions
The new carboxylated pbim complexes display well-resolved
NMR spectra which can be fully assigned by 2D NMR spectro-
scopy. The pbimH complex electrochemistry has been reinvesti-
gated and evidence was presented for reduction of the deproton-
ated ligand. The carboxylated pbim complexes show good
electrochemical reversibility, with slightly more positive redox
potentials than the pbimH parent complex. In general, the
spectroscopic (light-harvesting) and electrochemical properties
of the carboxylated complexes (relating to electron collection
efficiency) were very similar. Hence, any difference in solar-cell
sensitization performance must be due to changes in the length
of the spacer. Unfortunately, the higher π* energy level of the
pbim ligand means that the electron in the MLCT excited state
resides on bpy, well away from the anchoring point. Con-
sequently, the IPCE values are too low to differentiate between
the complexes, but a fall in the open-circuit potential as the
chain length increases was observed.
37 B. W. Jing, H. Zhang, M. H. Zhang, Z. H. Lu and T. Shen, J. Mater.
Chem., 1998, 8, 2055.
Acknowledgements
38 K. Sayama, S. Tsukagoshi, K. Hara, Y. Ohga, A. Shinpou, Y. Abe,
S. Suga and H. Arakawa, J. Phys. Chem. B, 2002, 106, 1363.
39 J. B. Asbury, E. C. Hao, Y. Q. Wang and T. Q. Lian, J. Phys. Chem.
B, 2000, 104, 11957.
We thank Dr D. T. Richens (School of Chemistry), Dr Bruce
Sinclair (School of Physics) and Dr Ebinazar Namdas (Organic
Semiconductor Centre) and the CVCP for an ORS scholarship
to H. Y. We thank the Carnegie Fund and the University
40 S. Ferrere, Chem. Mater., 2000, 12, 1083.
D a l t o n T r a n s . , 2 0 0 3 , 6 8 5 – 6 9 1
691