Minireviews N.

Robertson

DOI: 10.1002/anie.200503083
Dye-Sensitized Solar Cells

Optimizing Dyes for Dye-Sensitized Solar Cells

Neil Robertson*
Keywords:
dyes/pigments · energy conversion · redox chemistry ·
semiconductors · sensitizers

Dye-sensitized solar cells (DSSCs) have emerged as an important

cheap photovoltaic technology. Charge separation is initiated at the
dye, bound at the interface of an inorganic semiconductor and a hole-
transport material. Careful design of the dye can minimize loss
mechanisms and improve light harvesting. Mass application of DSSCs
is currently limited by manufacturing complexity and long-term
stability associated with the liquid redox electrolyte used in the most-
efficient cells. In this Minireview, dye design is discussed in the context
of novel alternatives to the standard liquid electrolyte. Rapid progress adsorb, which is crucial for efficient
is being made in improving the efficiencies of such solid and quasi- light harvesting. The porous TiO2 layer
is interpenetrated by a hole-transport
solid DSSCs which promises cheap, efficient, and robust photovoltaic
material (HTM), which may be a
systems. redox electrolyte in solution or a
solid-state or quasi-solid-state (gel)
material. Excitation of the dye leads
1. Introduction to the injection of electrons from the excited dye to the
conduction band of the TiO2. The ground state of the dye is
1.1. Design of Dye-Sensitized Solar Cells regenerated through reduction by the HTM to give the

Increasing energy demands and concerns over global

warming have led to a greater focus on renewable energy
sources in recent years. The conversion of solar energy is
likely to play a key role as one of the technologies that can
replace fossil fuels in the generation of mass energy. However,
the current high cost of solar panels made from traditional
inorganic semiconductors[1] imposes a restriction on their
mass usage. Alternative cheaper solar energy technologies are
therefore under intensive study, and in this context dye-
sensitized solar cells (DSSCs) have emerged as an important
class of photovoltaic device. DSSCs are currently undergoing
rapid development in an effort to obtain robust, efficient, and
cheap devices that are suitable for practical use.[2–4] An outline
of the operation of a DSSC is illustrated in Figure 1. The
system comprises a dye that is bound to the surface of an
inorganic semiconductor. Typically nanocrystalline TiO2 is
used as it provides a large surface area to which the dye can

Figure 1. Outline of the operation of a DSSC (D = dye; V [V] vs

[*] Dr. N. Robertson standard calomel electrode (SCE)). The green arrows represent pro-
School of Chemistry cesses required for photovoltaic function: k1 = charge injection,
University of Edinburgh k2 = dye regeneration, k3 = charge collection at the conducting glass
King’s Buildings electrode, and k4 = charge collection at the Pt electrode. The red arrows
West Mains Road represent loss mechanisms: k5 = charge recombination with the hole-
Edinburgh EH9 3JJ (UK) transport material (HTM; dark current), k6 = charge recombination
Fax: (+ 44) 131-650-4743 with the oxidized dye (D+), and k7 = decay of the excited state of the
E-mail: neil.robertson@ed.ac.uk dye (D*).

2338  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345
 Angewandte
Dye-Sensitized Solar Cells Chemie

required charge separation. Charges migrate and are collect- overview of dye characteristics and illustrate the ways in
ed at a transparent conducting electrode (electrons) and Pt which dye design has been used to enhance the efficiency of
electrode (holes). A number of parameters are used to cells. The key challenge at present is to obtain efficiencies that
characterize the detailed performance of a photovoltaic are comparable to that for the optimized I /I3 /volatile
cell;[2] however, in this Minireview discussion is limited solvent cell by using an HTM that is more suitable for mass
mainly to the overall efficiency of conversion of solar-to- production and long-term stability. The design of dyes will
electrical energy of the cell (h) and also to the incident photon therefore be presented in this context in the quest for
to current efficiency (IPCE), which gives a measure of the photovoltaic (PV) cells that are efficient, robust, and cheap
efficiency as a function of wavelength of the incident light. to manufacture.
Following the initial development of this type of cell by
O3Regan and Gr6tzel,[5] there has been further extensive
study and optimization of the design, including modifications 1.2. General Design of Dyes
to the nanocrystalline semiconductor, the redox electrolyte,
and the dye. Currently, the most efficient DSSCs show Progress in the optimization of the dye component of the
efficiencies of over 10 %, which is sufficiently high to be of cell has been made through systematic variation of the
practical utility. The cells that display the highest efficiency, ligands, metal, and other substituent groups in candidate
however, use an HTM comprising an I /I3 redox electrolyte transition-metal complexes.[2, 15, 16, 17] This systematic study has
solution that gives rise to poor long-term stability and resulted in the development of mononuclear[2] and polynuc-
manufacturing complexity.[6] The involvement of volatile I2 lear[18] dyes based on metals such as RuII,[19–21] OsII,[4, 22, 23]
and volatile solvent requires the cells to be sealed, and PtII,[24, 25] ReI,[26] CuI,[27] and FeII.[28] Besides transition-metal
additionally, the I /I3 redox electrolyte can be corrosive complexes, a range of organic molecules have been explored,
towards the Pt electrode. Although some commercialization with recent examples including coumarin,[29] squaraine,[30]
of DSSCs has begun, new aspects of cell design are being indoline,[31] hemicyanine,[32] and other conjugated donor–
intensively explored to continue to address these limitations acceptor organic dyes,[33–36] and the best efficiency reported
and open up DSSC technology to much wider exploitation. was 8 % (see Figure 7 a in Section 2.5).[31] Porphyrin dyes[37, 38]
Thus, a number of alternative redox mediators and electrolyte and phthalocyanine dyes[39] have also been explored.
systems have been explored, including I /I3 in either solid The dyes used in DSSC technology must conform to a
polymer,[7, 8] gel,[6] ionic liquid,[9] or plastic crystal[10] systems; number of essential design requirements in order to function.
solid inorganic materials;[11] CoII/CoIII[12] and SeCN / They must bind strongly to TiO2 by means of an anchoring
(SeCN)3 redox couples;[9] and hole-conducting organic group, typically carboxylic or phosphonic acid groups, to
polymers[13] and small organic molecules.[14] However, a ensure efficient electron injection into the TiO2 conducting
reduced efficiency has so far been achieved for such cells; band and to prevent gradual leaching by the electrolyte. The
for example, the maximum cell efficiencies observed for the LUMO of the dye must be sufficiently high in energy for
gel-electrolyte systems, organic HTM systems, and ionic efficient charge injection into the TiO2, and the HOMO must
liquid systems are around 6, 4, and 8 %, respectively. In each be sufficiently low in energy for efficient regeneration of the
of these cases, the efficiencies are less than that for the oxidized dye by the HTM. The dye must absorb solar
optimized I /I3 /volatile solvent cell as a result of factors such radiation strongly with absorption bands in the visible or
as reduced hole mobility, poorer electron-transfer kinetics, near-IR region, preferably covering a broad range of wave-
and poorer contact at the dye–HTM interface. Much is now lengths. Electron transfer from the dye to the TiO2 must also
understood concerning loss mechanisms that arise within the be rapid in comparison with decay to the ground state of the
system, and key processes are indicated in Figure 1. dye. Dyes that show emission in the solution state at room
Some recent overviews that discuss general aspects of temperature have typically been used, although this is not
DSSCs[1, 2, 4] have appeared along with a series of articles that essential.[40]
give a detailed account of several specific topics within the The family of complexes [{(4,4’-CO2H)2(bipy)}2RuX2]
field.[3] The aim of this Minireview is to provide a general (bipy = 2,2’-bipyridyl; X = Cl, Br, I, CN, NCS) perform
well;[19] for example, the dye [{(4,4’-CO2H)2bipy}2Ru(NCS)2]
Neil Robertson studied chemistry at the (N3) and the doubly deprotonated analogue (N719, Figure 2)
University of Edinburgh and obtained his give a solar-to-electrical energy conversion efficiency of over
PhD there in 1992. He then carried out 10 %. Use of a terpyridyl ligand led to the so-called “black
postdoctoral work at the Freie Universit$t dye” (Figure 2), which gives a very high IPCE across the
Berlin and University of Wales, Bangor,
wavelength range 400–700 nm and a cell efficiency of over
where he first developed an interest in the
electronic and magnetic properties of molec- 10 %.[21, 41] Dyes are often referred to by codes (as indicated in
ular materials. After a Royal Society of the figures), which will be used in the remainder of the article.
Edinburgh/BP Research Fellowship at Edin-
burgh (1996–1999) and a lectureship at
Imperial College London, he took up his 2. Current Developments in Dye Design
current position (2001) as a senior lecturer
in chemistry at the University of Edinburgh.
As previously mentioned, the key limitations of DSSCs
arise from the I /I3 liquid electrolyte and this has neces-

Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 2339
Angewandte
Chemie N. Robertson

sitated the exploration of other cell designs that use an Recent work has illustrated that this approach can be
alternative HTM. Initial research into these modified cells extended further and that dyes with attached electron-donor
with novel designs typically used dyes that were previously groups can locate the cationic charge at a greater distant from
optimized for a cell with an I /I3 solution electrolyte as the the TiO2 surface.[43, 44] For instance, the substituted porphyrin
HTM (e.g. N3 or N719). As studies develop, however, it is shown in Figure 3 a attaches itself to the TiO2 surface through
apparent that innovation in transport materials should be
carried out in conjunction with the development of new dyes
to maximize the improvements in efficiency and stability.
Thus, the exploration of new dye designs specifically for solid-
state and quasi-solid-state cells is currently an area of high
priority that has started to receive increased attention. Key
approaches to this area of research are described in Sec-
tion 2.1–2.5.

2.1. Enhanced Charge Separation in the Dye: Minimizing Charge

Recombination

The complex N3 (Figure 2) is a good illustration of the

role that charge separation in the dye plays in controlling the
kinetics of electron transfer. Absorption of the dye in the low-
energy visible region involves an MLCT (metal-to-ligand
charge transfer) transition that places the excited electron on
the diimine, which is directly attached to the TiO2. The result
is ultrafast charge injection (see Section 2.2), however, the
positive charge density that remains on the dye is distributed
over the metal and also to some extent over the NCS ligands.
The resulting spatial separation of the positive charge density Figure 3. Examples of dyes with attached triarylamine electron-donor
on the dye and the injected electrons has the crucial effect of groups: a) a porphyrin dye[44] and b) Ru dye N845.[45]
retarding the rate of charge recombination between the
injected electrons and the dye cation, which is a key loss
mechanism (Figure 1, k6). Indeed, it has been shown that its carboxylic acid group, and p–p* photoexcitation leads to
charge-recombination dynamics are closely dependent on this charge injection.[44] It was found that recombination of the
separation and, in contrast, show very little dependence on injected electron with the dye was an order of magnitude
the thermodynamic driving force for the recombination, as slower than for a comparable dye that lacked the electron-
determined by the reduction potential of the dye cation.[42] donor triphenylamine groups. This difference was attributed
to the location of the cationic charge largely on the triphenyl-
amine moieties of the dye and the consequently larger
physical separation of the cationic charge from the TiO2
surface. This approach was extended to the ruthenium dye
N845 (Figure 3 b), which also contains an appended triaryl-
amine moiety.[45] In this case, a 1000-fold retardation of the
recombination dynamics was attributed to the 4 L increase in
distance between the cationic center of charge and the TiO2
surface (as estimated from semi-empirical calculations) in
comparison with N719 (Figure 2).

2.2. Methods to Attach the Dye to TiO2

The majority of dyes are linked to the TiO2 semiconductor

through acidic groups—mostly carboxylic acid or, less com-
monly, phosphonic acid linkers[46]—although a variety of
other moieties also have been used.[47, 48] Carboxylic acid
groups can form ester linkages with the surface of the metal
oxide to provide a strongly bound dye and good electronic
Figure 2. Examples of some Ru–polypyridyl dyes used in DSSCs that communication between the two parts. However, the link can
give cell efficiencies of over 10 %. TBA = tetra-n-butylammonium. be hydrolyzed through the presence of water, an important
factor in terms of the stability of the cell (see Section 2.3). The

2340 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345
 Angewandte
Dye-Sensitized Solar Cells Chemie

dyes that have shown the highest cell efficiencies have used minimized, with charge injection only just fast enough to
carboxylic acid linkers, although a recent study into phos- compete with excited-state decay such that loss through
phonic acid linkers reported the highest efficiency ( 8 %) for charge recombination is also minimized. Variation of the
any dye with a non-carboxylic acid linker (Z955, Figure 4). In nature of the linker group between the dye and the TiO2 can
this case, modification of the linker from carboxylic to play a crucial role in this optimization and hence on the cell
phosphonic acid groups resulted in efficiencies achieved.
an interrelated series of changes in In an attempt to control these interfacial electron-transfer
the characteristics of the device; for processes, Haque et al. studied an azobenzene dye encapsu-
example, a blue shift of the absorp- lated within a cyclodextrin molecule attached to TiO2.[53] The
tion maxima, good stability of the cyclodextrin comprises a hydrophilic outer layer, which is
device, and slower charge-recombi- suitable for adsorption onto the TiO2 surface, and a hydro-
nation kinetics for Z955 relative to phobic inner surface. The spatial separation of the dye from
the analogous dye with carboxylic the TiO2 increased, however, the charge injection yield was
acid linkers (Z907, see Figure 5, Sec- comparable with a non-encapsulated dye analogue, in keep-
tion 2.3). The increased number of ing with the kinetic redundancy argument. Importantly, the
protons on the phosphonic acid increased separation also led to a significantly slower rate for
Figure 4. Ru–polypyridyl groups compared with carboxylic k6, offering potential gains in cell efficiency. As well as these
dye Z955 with phosphon- acid groups has been suggested as a effects, the encapsulation approach has the possibility to
ic acid linker groups.[46]
factor in modifying the performance enhance photochemical stability, redox reversibility, and
of the dye, as it has been observed electroluminescent efficiency.
that the efficiency of the solar cell These recent examples illustrate that much is now under-
can be influenced by changing the protonation of the acid stood concerning the optimization of kinetic parameters for
groups. This is attributed to the effect of the bound dye on the interfacial electron transfer. These studies, however, have
energy of the TiO2 conducting band, such that N719, which is typically involved use of the I /I3 solution redox electrolyte,
deprotonated, gives a higher cell efficiency than the proton- and an important next step will involve the exploitation of
ated analogue N3 (Figure 2).[41] these ideas in improving the lower efficiencies currently
Besides the protonation state of the carboxylic acid group, achieved in cells with alternative solid and quasi-solid hole-
the position of these linker groups on the bipyridyl moiety has transport materials.
been explored.[49, 25] The large majority of studies involving
bipyridyl groups employ 4,4’-substituted derivatives. The
study of a Ru complex with carboxylic acid groups in the 2.3. Hydrophobic Dyes
3,3’-positions of the bipyridyl revealed a decreased efficiency
of the cell,[49] whereas a study of Pt dyes showed the opposite Dyes attached to TiO2 through carboxylic acid groups are
effect with a slightly improved efficiency for the complex with susceptible to desorption from the surface under the action of
3,3’-substituted bipyridyl.[25] As substitution at the 3,3’-posi- trace quantities of water and has serious consequences on the
tions necessitates some twisting of the ligand, the consequen- long-term stability of the resultant solar cells. It has been
ces of the electronic alteration to the dye when bound to TiO2 reported that dyes with attached
may be difficult to predict, however, it seems that further hydrophobic chains (e.g. Z907, Fig-
study of complexes with 3,3’-substituted bipyridyls is merited. ure 5) can display an enhanced sta-
Charge injection from Ru–polypyridyl dyes linked bility towards desorption from TiO2
through carboxylic acid groups is extremely rapid, and for induced by water in the liquid or gel
dye-sensitized TiO2 covered with an inert solvent it occurs on electrolyte.[54, 6] A cell based on Z907
a femto- to picosecond timescale from the excited singlet and using a polymer gel I /I3 electrolyte
triplet states of the dye, respectively.[50] In a complete was shown to combine a cell effi-
DSSC,[51, 52] charge injection occurs on the picosecond time- ciency of 6.1 % with excellent stabil-
scale, with the composition of the redox electrolyte playing an ity to both prolonged thermal stress
important role in modifying the energetics of the TiO2 and light soaking that matched the
conduction band and hence the charge-injection rate. This criteria required for the outdoor use Figure 5. The amphiphil-
extremely rapid process leads to injection yields that ap- of solar cells.[6] Besides the role ic dye Z907.
proach 100 % for many dyes. However, as the excited-state played by the gel electrolyte, the
lifetime of dyes such as N3 can be as long as 50 ns, high stability was attributed to the hydro-
injection yields may still be achieved with an injection rate phobic properties of the dye which enhanced stability towards
that is several orders of magnitude slower than those water-induced desorption. This effect was illustrated with an
observed. The unnecessarily rapid charge injection is referred analogous cell, which used N719 in place of Z907 as the dye,
to as “kinetic redundancy” and is important because the rates that displayed much poorer stability to thermal stress.
of charge injection (k1) and the charge-recombination loss A significant loss mechanism involves recombination of
process (k6) are correlated, as both are influenced by the electrons from the TiO2 conduction band directly with the
strength of electronic communication of the dye with the HTM, known as the dark current (k5). In addition to the
TiO2.[52] For an optimum device kinetic redundancy should be above considerations, it has also been suggested that dyes

Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 2341
Angewandte
Chemie N. Robertson

with attached hydrophobic groups can inhibit k5 by forming a

hydrophobic network that impedes the interaction between
I3 and the TiO2 surface and minimizes this loss. Evidence
indicates that N719 itself acts to suppress the dark current by
forming a blocking layer on the TiO2 surface,[55, 25] and it may
be that further addition of steric bulk to the dye might
enhance this role.
Another benefit of amphiphilic dyes arises in cells where
an organic HTM has been employed to replace the liquid Figure 6. a) Ru complex with trans NCS ligands.[57] b) Example of an
electrolyte. In such cells, there is generally a poor interaction Os complex dye with a good incident photon to current efficiency at
between the dye and the organic HTM[56] that leads to weaker long wavelengths.[23]
electronic communication and slower regeneration kinetics
(k2) of the dye. Dyes with attached hydrophobic chains can
enhance the wettability of the TiO2 by the hole-conducting singlet character through the larger spin–orbit coupling in
polymer enhancing interaction with the dye. Interfacial heavier elements (heavy-atom effect).[23] For example, for the
contact between the dye and the HTM is also important in osmium complex shown in Figure 6 b this has led to better
the context of solid polymer electrolytes,[7] particularly for IPCE values at longer wavelengths than for a comparable Ru
high-molecular-weight polymers that do not penetrate well complex dye, although over the whole spectral region the Ru
into the nanopores of the TiO2 semiconductor. It is also complex performed better.
interesting to note in this context that the typically slower k2 Another approach to long-wavelength sensitization in-
displayed with an organic HTM may lead to a requirement for volves the use of phthalocyanine dyes, which are known to
dyes that have better long-term stability in their oxidized display an intense absorption in the Q band at low energy as
form.[23] well as a higher energy Soret band. However, studies on
phthalocyanines are hampered by poor solubility and also by
their tendency to aggregate on the TiO2 surface which leads to
2.4. Extending the Spectral Coverage of Dyes deactivation of the excited state of the dye. A recent example
employed a titanium phthalocyanine dye with axial ligation to
Currently, sensitizers such as N3 and N719 (Figure 2) enable binding to TiO2. Bulky terminal tert-butyl groups were
show comparatively low IPCEs in the red and near-infrared also included to prevent aggregation of the dye and to
(NIR) region of the electromagnetic spectrum. Control of the improve solubility.[58] As expected for this class of molecule,
HOMO and LUMO levels of a dye is required to develop an extremely intense Q band absorption (e =
better red-absorbing dyes, as illustrated by the development 135 000 cm 1m 1) at lmax = 702 nm was observed. Although
of the “black dye” shown in Figure 2. Manipulation of the in practice it was found that this particular dye was not able to
absorption spectrum of the dye also allows the possibility to efficiently inject electrons following excitation in the Q band,
develop solar cells that absorb in the NIR and are transparent a key observation was the lack of aggregation of the dye when
to visible light, thus allowing their use as photovoltaic adsorbed onto TiO2 which opens up the possibility to design
windows on buildings. To function as a DSSC, however, the other phthalocyanine sensitizers that exploit the intense
LUMO must remain sufficiently higher than the edge of the absorption at low energy.
conduction band of TiO2 for efficient charge injection while Similar considerations arise in the case of porphyrin dyes,
the HOMO must remain sufficiently below the redox level of which also display an intense, low-energy Q band absorption
the HTM for efficient regeneration of the dye. The lower and show kinetics for charge-injection and charge-recombi-
energy of longer-wavelength photons makes the development nation processes that are comparable to those of the best Ru–
of appropriate red-absorbing dyes that adhere to these polypyridyl dyes. Again a serious limitation arises from the
requirements a challenge, as the HOMO–LUMO gap is tendency of the dye to aggregate, although poor electronic
narrower. One manner in which this has been approached communication between the dye core and the carboxylate
involves the study of Ru–polypyridyl dyes related to N3 that linker may also play a role. Recent reports of porphyrin dyes
are constrained to show trans geometry of the NCS ligands with a conjugated carboxylate linker revealed negligible
(for example, as shown in Figure 6 a).[57] Such trans-Ru– evidence of aggregation and, thus, efficiencies of up to 5.6 %,
polypyridyl complexes typically show lower-energy absorp- the highest known value for any porphyrin dye.[38] Signifi-
tion in comparison with the cis analogues. The use of geo- cantly, these dyes show comparatively high molar extinction
metrically restrained ligands is required to prevent photo- coefficients, for example, e = 18 500 cm 1m 1 at lmax = 622 nm,
induced isomerization of the trans form to the cis isomer. This suggesting that further optimization of this family may be a
approach has led to dyes that show absorption bands across fruitful source of novel long-wavelength-absorbing dyes.
the entire visible and NIR regions of the spectrum, and initial The use of several dyes as cosensitizers has been explored
studies have suggested that these dyes may perform better to extend the spectral region of the sensitizing layer. In the
than N3 once their performance has been optimized. case of organic dyes, this approach may also overcome the
It has been noted that in osmium complexes, spin- typically narrow absorption bands observed for these dyes. A
forbidden singlet–triplet MLCT excitation can show signifi- difficulty of this approach involves the typical decrease in the
cant intensity at low energy as a result of the mixing of some sensitizing efficiency of the individual dyes upon mixing with

2342 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345
 Angewandte
Dye-Sensitized Solar Cells Chemie

another cosensitizer, as might be expected from simple

considerations of the quantity of dye on the TiO2 surface. A
recent report[59] demonstrated that cosensitization using three
organic dyes with complementary absorption spectra can lead
to an overall efficiency of 6.5 %. In this case, all three dyes in
the mixture actually displayed higher IPCE values in their
own spectral region than they did when used alone. Although
this appears counterintuitive, it was attributed to two possible
effects: 1) more complete packing of the three dyes on the
semiconductor surface which blocks the dark current (k5) and
2) an isolating effect that causes a decrease in the aggregation
of each dye, thus minimizing loss through the associated
deactivation of the excited state. This recent example thus
demonstrates that a cosensitization approach can lead to
enhanced spectral coverage and better enable the use of
organic dyes, which typically display high extinction coef-
ficients (see Section 2.5).
In an alternative approach to co-sensitization, both a Ru–
polypyridyl dye and a Ru–phthalocyanine dye were used
together, however, in this case a secondary layer of metal
oxide was deposited onto the Ru–polypyridyl nanoparticles
before attachment of the Ru–phthalocyanine layer.[60] This Figure 7. a, b) Organic dyes with high extinction coefficient that are
approach allowed near-monolayer coverage of both dyes, used to a) give a high-efficiency DSSC[31] and b) make thin DSSCs with
rather than a competition between the two for adsorption organic HTMs.[61] c) Ru dye K19, which was designed to give an
sites, and resulted in an electron-transfer cascade whereby the increased extinction coefficient for the MLCT band.[62]
charge center of the dye cation was moved away from the
TiO2 surface by transfer between the two dye layers. The
observation of efficient current generation through excitation stress and light soaking displayed by Z907 that was attributed
of both dyes indicates that this is an important new approach to the hydrophobic spectator ligand. In addition, however,
in the development of panchromatic systems. K19 shows a higher extinction coefficient for the low-energy
band at l = 543 nm of e = 18 200 cm 1m 1 compared with e =
12 200 and 14 000 cm 1m 1 for Z907 and N719 (Figure 2),
2.5. Enhancing Molar Extinction Coefficients respectively. A comparison of these three dyes demonstrates
that under the same conditions K19 reveals the highest
The use of dyes with a higher molar extinction coefficient efficiency of 7.0 %, and interestingly this was achieved using a
clearly allows increased light harvesting of a given film low-vapor-pressure electrolyte. A related approach also
thickness or, alternatively, thinner dye-sensitized films to be involved the use of 4,4’-bis(carboxyvinyl)-2,2’-bipyridyl li-
used which results in better efficiencies from decreased losses gands as extended delocalized units that increase the ex-
during charge transport through the nanocrystalline TiO2. tinction coefficient of the dye. In this case, the extended bipy
High extinction coefficients have been achieved by using ligand was used to link to TiO2 rather than as the spectator
organic dyes[34, 29] rather than transition-metal complexes, ligand.[63, 64] As well as an increase in extinction coefficient, a
however, the former typically suffer other disadvantages such red shift of the maxima of the MLCT band was observed and
as narrow absorption bands that limit the light-harvesting relates to extending the spectral coverage of the dye, as
ability (see Section 2.4).[59] Enhancement of the extinction discussed in Section 2.4.
coefficient is particularly important in the context of cells that
use organic materials as the HTM. For these devices the
thickness of the film is crucial, as the limited charge-carrier 3. Summary and Outlook
mobility in the organic HTM leads to significant charge-
carrier recombination (k5) and a much lower efficiency. The Recently, much effort has been directed towards the
use of a highly absorbing organic dye (Figure 7 b) allowed a optimization of all aspects of DSSCs, including the inorganic
reduction in these losses through fabrication of a much nanocrystalline semiconductor and the hole-transport mate-
thinner device, leading to the greatest efficiency recorded for rial. Often, initial studies in this area have involved using the
an organic HTM cell of over 4 %.[61] well-established dyes that worked best for previous DSSC
Another recent approach involved the use of extended designs. The well-known Ru complex dyes N3, N719, and the
delocalized ligands for Ru complexes, for example, with the “black dye” (Figure 2) exhibit high efficiencies in the
dye K19 (Figure 7 c).[62] This dye is related to the analogous standard dye-sensitized solar cell with an I /I3 solution
amphiphilic dye Z907 (Figure 5) but comprises additional electrolyte, and it is questionable whether in these devices
stilbene units conjugated onto the hydrophobic ligand. their performance will be significantly bettered. Dye-sensi-
Consequently, this dye retains the high stability to thermal tized solar cells formed with alternative hole-transport

Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 2343
Angewandte
Chemie N. Robertson

materials, however, may offer different energy levels, differ- [20] A. Islam, H. Sugihara, K. Hara, L. P. Singh, R. Katoh, M.
ent hydrophobicity/hydrophilicity properties, and different Yanagida, Y. Takahashi, S. Murata, H. Arakawa, J. Photochem.
electron-transfer kinetics, and show a decrease in hole Photobiol. A 2001, 145, 135.
[21] M. K. Nazeeruddin, P. Pechy, M. Gr6tzel, Chem. Commun. 1997,
transport. A need has emerged to optimize the dye in
1705.
conjunction with other design factors to best exploit and be
[22] D. Kuciauskas, M. S. Freund, H. B. Gray, J. R. Winkler, N. S.
fully compatible with other cell modifications that have taken Lewis, J. Phys. Chem. B 2001, 105, 392.
place. For example: higher molar extinction coefficients may [23] R. Argazzi, G. Larramona, C. Contado, C. A. Bignozzi, J.
be crucial to allow thinner cells and the dominance of Ru– Photochem. Photobiol. A 2004, 164, 15.
polypyridyl sensitizers may be challenged by organic, phtha- [24] A. Islam, H. Sugihara, K. Hara, L. P. Singh, R. Katoh, M.
locyanine, or porphyrin dyes in this type of device; the use of Yanagida, Y. Takahashi, S. Murata, H. Arakawa, Inorg. Chem.
hydrophobic groups to enhance the stability of the device has 2001, 40, 5371.
become well established; near-IR dyes would allow photo- [25] E. A. M. Geary, L. J. Yellowlees, L. A. Jack, I. D. H. Oswald, S.
Parsons, N. Hirata, J. R. Durrant, N. Robertson, Inorg. Chem.
voltaic windows to be developed; and cheaper dyes that do
2005, 44, 242.
not contain expensive transition metals may also become [26] G. M. Hasselmann, G. J. Meyer, Z. Phys. Chem. 1999, 212, 39.
more important in the context of commercialization. Dye- [27] N. Alonso-Vante, J.-F. Nierengarten, J.-P. Sauvage, J. Chem. Soc.
sensitized solar cells are beginning to be exploited as a Dalton Trans. 1994, 1649.
commercial technology, and further developments in dye [28] P. M. Jayaweera, S. S. Palayangoda, K. Tennakone, J. Photo-
design will play a crucial part in the ongoing optimization of chem. Photobiol. A 2001, 140, 173.
these devices. [29] K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara,
Y. Dan-oh, C. Kasada, A. Shinpo, S. Suga, J. Phys. Chem. B 2005,
109, 15 476.
The University of Edinburgh and the EPSRC (Supergen
[30] S. Alex, U. Santhosh, S. Das, J. Photochem. Photobiol. A 2005,
Project) are acknowledged for their recent support in this area. 172, 63.
[31] T. Horiuchi, H. Miura, K. Sumioka, S. Uchida, J. Am. Chem. Soc.
Received: August 30, 2005 2004, 126, 12 218.
Published online: March 9, 2006 [32] Y.-S. Chen, C. Li, Z.-H. Zeng, W.-B. Wang, X.-S. Wang, B.-W.
Zhang, J. Mater. Chem. 2005, 15, 1654.
[33] K. Hara, T. Sato, R. Katoh, A. Furube, T. Yoshihara, M. Murai,
[1] M. Gr6tzel, Nature 2001, 414, 338. M. Kurashige, S. Ito, A. Shinpo, S. Suga, H. Arakawa, Adv.
[2] M. K. Nazeeruddin, M. Gr6tzel in Comprehensive Coordination Funct. Mater. 2005, 15, 246.
Chemistry II, Vol. 9 (Eds.: J. A. McCleverty, T. J. Meyer), [34] K. R. J. Thomas, J. T. Lin, Y.-C. Hsu, K.-C. Ho, Chem. Commun.
Elsevier, Dordrecht, 2004, chap. 16. 2005, 4098.
[3] Special issues on DSSC: Coord. Chem. Rev. 2004, 248, 1161 – [35] T. Kitamura, M. Ikeda, K. Shigaki, T. Inoue, N. A. Anderson, X.
1530. Ai, T. Lian, S. Yanagida, Chem. Mater. 2004, 16, 1806.
[4] M. Gr6tzel, J. Photochem. Photobiol. C 2003, 4, 145. [36] K. Hara, M. Kurashige, S. Ito, A. Shinpo, S. Suga, K. Sayama, H.
[5] B. O3Regan, M. Gr6tzel, Nature 1991, 353, 737. Arakawa, Chem. Commun. 2003, 252.
[6] P. Wang, S. M. Zakeeruddin, J. E. Moser, M. K. Nazeeruddin, T. [37] W. M. Campbell, A. K. Burrell, D. L. Officer, K. W. Jolley,
Sekiguchi, M. Gr6tzel, Nat. Mater. 2003, 2, 402. Coord. Chem. Rev. 2004, 248, 1363.
[7] M.-S. Kang, J. H. Kim, Y. J. Kim, J. Won, N.-G. Park, Y. S. Kang, [38] Q. Wang, W. M. Campbell, E. E. Bonfantani, K. W. Jolley, D. L.
Chem. Commun. 2005, 889. Officer, P. J. Walsh, K. Gordon, R. Humphrey-Baker, M. K.
[8] M.-S. Kang, Y. J. Kim, J. Won, Y. S. Kang, Chem. Commun. 2005, Nazeeruddin, M. Gr6tzel, J. Phys. Chem. B 2005, 109, 15 397.
2686.
[39] T. Komori, Y. Amao, J. Porphyrins Phthalocyanines 2003, 7, 131.
[9] P. Wang, S. M. Zakeeruddin, J. E. Moser, R. Humphrey-Baker,
[40] S. A. Haque, E. Palomares, B. M. Cho, A. N. M. Green, N.
M. Gr6tzel, J. Am. Chem. Soc. 2004, 126, 7164.
Hirata, D. R. Klug, J. R. Durrant, J. Am. Chem. Soc. 2005, 127,
[10] P. Wang, Q. Dai, S. M. Zakeeruddin, M. Forsyth, D. R. MacFar-
3456.
lane, M. Gr6tzel, J. Am. Chem. Soc. 2004, 126, 13 590.
[41] M. K. Nazeeruddin, S. M. Zakeeruddin, R. Humphrey-Baker,
[11] Q.-B. Meng, K. Takahashi, X.-T. Zhang, I. Sutanto, T. N. Rao, O.
M. Jirousek, P. Liska, N. Vlachopoulos, V. Shklover, C.-H.
Sato, A. Fujishima, Langmuir 2003, 19, 3572.
Fischer, M. Gr6tzel, Inorg. Chem. 1999, 38, 6298.
[12] H. Nusbaumer, S. M. Zakeeruddin, J.-E. Moser, M. Gr6tzel,
[42] J. N. Clifford, E. Palomares, M. K. Nazeeruddin, M. Gr6tzel, J.
Chem. Eur. J. 2003, 9, 3756.
[13] Y. Saito, T. Azechi, T. Kitamura, Y. Hasegawa, Y. Wada, S. Nelson, X. Li, N. J. Long, J. R. Durrant, J. Am. Chem. Soc. 2004,
Yanagida, Coord. Chem. Rev. 2004, 248, 1469. 126, 5225.
[14] J. KrTger, R. Plass, L. Cevey, M. Piccirelli, M. Gr6tzel, U. Bach, [43] R. Argazzi, C. A. Bignozzi, J. Am. Chem. Soc. 1995, 117, 11 815.
Appl. Phys. Lett. 2001, 79, 2085. [44] J. N. Clifford, G. Yahioglu, L. R. Milgrom, J. R. Durrant, Chem.
[15] L. Spiccia, G. B. Deacon, C. M. Kepert, Coord. Chem. Rev. 2004, Commun. 2002, 1260.
248, 1329. [45] N. Hirata, J.-J. Lagref, E. J. Palomares, J. R. Durrant, M. K.
[16] R. Argazzi, N. Y. M. Iha, H. Zabri, F. Odobel, C. A. Bignozzi, Nazeeruddin, M. Gr6tzel, D. Di Censo, Chem. Eur. J. 2004, 10,
Coord. Chem. Rev. 2004, 248, 1299. 595.
[17] A. S. Polo, M. K. Itokazu, N. Y. M. Iha, Coord. Chem. Rev. 2004, [46] P. Wang, C. Klein, J. E. Moser, R. Humphrey-Baker, N.-L.
248, 1343. Cevey-Ha, R. Charvet, P. Comte, S. M. Zakeeruddin, M.
[18] C. A. Bignozzi, R. Argazzi, C. J. Kleverlaan, Chem. Soc. Rev. Gr6tzel, J. Phys. Chem. B 2004, 108, 17 553.
2000, 29, 87. [47] E. Galoppini, Coord. Chem. Rev. 2004, 248, 1283.
[19] M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. [48] S. Altobello, C. A. Bignozzi, S. Caramori, G. Larramona, S.
MTller, P. Liska, N. Vlachopolous, M. Gr6tzel, J. Am. Chem. Soc. Quici, G. Marzanni, L. R. Lakhmiri, J. Photochem. Photobiol. A
1993, 115, 6382. 2004, 166, 91.

2344 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345
 Angewandte
Dye-Sensitized Solar Cells Chemie

[49] P.-H. Xie, Y.-J. Hou, T.-X. Wei, B.-W. Zhang, Y. Cao, C.-H. [57] T. Renouard, R.-A. Fallahpour, M. K. Nazeeruddin, R. Hum-
Huang, Inorg. Chim. Acta 2000, 308, 73. phrey-Baker, S. I. Gorelsky, A. B. P. Lever, M. Gr6tzel, Inorg.
[50] G. Benko, J. Kallioinen, J. E. I. Korppi-Tommola, A. P. Yartsev, Chem. 2002, 41, 367.
V. SundstrUm, J. Am. Chem. Soc. 2002, 124, 489. [58] E. Palomares, M. V. Marinez-Diaz, S. A. Haque, T. Torres, J. R.
[51] Y. Tachibana, J. E. Moser, M. Gr6tzel, D. R. Klug, J. R. Durrant, Durrant, Chem. Commun. 2004, 2112.
J. Phys. Chem. 1996, 100, 20 056. [59] Y. Chen, Z. Zeng, C. Li, W. Wang, X. Wang, B. Zhang, New J.
[52] S. A. Haque, E. Palomares, B. M. Cho, A. N. M. Green, N. Chem. 2005, 29, 773.
Hirata, D. R. Klug, J. R. Durrant, J. Am. Chem. Soc. 2005, 127, [60] J. N. Clifford, E. Palomares, M. K. Nazeeruddin, R. Thampi, M.
3456. Gr6tzel, J. R. Durrant, J. Am. Chem. Soc. 2004, 126, 5670.
[53] S. A. Haque, J. S. Park, M. Srinivasarao, J. R. Durrant, Adv. [61] L. Schmidt-Mende, U. Bach, R. Humphrey-Baker, T. Horiuchi,
Mater. 2004, 16, 1177. H. Miura, S. Ito, S. Uchida, M. Gr6tzel, Adv. Mater. 2005, 17, 813.
[54] S. M. Zakeeruddin, M. K. Nazeeruddin, R. Humphry-Baker, P. [62] P. Wang, C. Klein, R. Humphrey-Baker, S. M. Zakeeruddin, M.
Pechy, P. Quagliotto, C. Barolo, G. Viscardi, M. Gr6tzel, Gr6tzel, J. Am. Chem. Soc. 2005, 127, 808.
Langmuir 2002, 18, 952. [63] M. K. Nazeeruddin, C. Klein, P. Liska, M. Gr6tzel, Coord. Chem.
[55] S. Ito, P. Liska, P. Comte, R. Charvet, P. Pechy, U. Bach, L. Rev. 2005, 249, 1460.
Schmidt-Mende, S. M. Zakeeruddin, A. Kay, M. K. Nazeerud-
[64] C. Klein, M. K. Nazeeruddin, P. Liska, D. Di Censo, N. Hirata, E.
din, M. Gr6tzel, Chem. Commun. 2005, 4351.
Palomares, J. R. Durrant, M. Gr6tzel, Inorg. Chem. 2005, 44, 178.
[56] Y. Saito, N. Fukuri, R. Senadeera, T. Kitamura, Y. Wada, S.
Yanagida, Electrochem. Commun. 2004, 6, 71.

Angew. Chem. Int. Ed. 2006, 45, 2338 – 2345  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 2345