Modification of nonlinear optical dyes for dye sensitized solar cells: A new use for a familiar acceptor

Article abstract of DOI:10.1039/c0jm03774e

The chemistry of electron deficient π-acceptors offers unique challenges in the rational design and synthesis of organic dyes for use in solar cells. We have synthesized 2-cyano-2-(3-cyano-4-((E)-4-(dibutylamino)styryl)-5,5- dimethylfuran-2(5H)-ylidene)et

Full text of DOI:10.1039/c0jm03774e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 4242
www.rsc.org/materials
PAPER
Modification of nonlinear optical dyes for dye sensitized solar cells: a new use
for a familiar acceptor†
a
a
b
Dinesh G. Patel, Nick M. Bastianon, Paul Tongwa, Janelle M. Leger, Tatiana V. Timofeeva
a
b
and Glenn P. Bartholomew‡*a
Received 3rd November 2010, Accepted 21st December 2010
DOI: 10.1039/c0jm03774e
The chemistry of electron deficient p-acceptors offers unique challenges in the rational design and synthesis
of organic dyes for use in solar cells. We have synthesized 2-cyano-2-(3-cyano-4-((E)-4-
(
dibutylamino)styryl)-5,5-dimethylfuran-2(5H)-ylidene)ethanoic acid (2), a dye containing an carboxylated
version of the strongly electron withdrawing TCF group by way of stepwise synthesis and isolation of
,5-dihydro-2-imino-4,5,5-trimethylfuran-3-carbonitrile (5), previously only available as an intermediate
2
available via microwave chemistry. We use this dye to sensitize nanocrystalline TiO in order to examine its
2
efficacy indye-sensitizedsolar cells. Crystal structures of anintermediateand thefinal dye arepresented and
the isomerization about the terminal alkene is discussed in addition to preliminary solar cell data.
Introduction
Strong p-electron acceptor groups play an important role in
organic photonic and electronic materials. Such groups modulate
the extent of intramolecular charge transfer (ICT) from a donor,
which can give rise to significant nonlinear optical polarizabil-
ities. Dipolar chromophores optimized for quadratic nonlinear
optical (NLO) response, as a result of substantial non-symmetric
ICT, exhibit red-shifted absorption maxima and substantially
larger molar extinction coefficients relative to similar chromo-
A continuing challenge in the development of highly efficient
1
,2
(
>10%) TiO -based dye sensitized solar cells is improving the
2
overlap between the sensitizing dye absorption profile and the
3,4
solar spectrum. Additionally, the intermediate molar extinction
4
4
ꢁ1
ꢁ1 5,6
coefficients (1.4 ꢀ 10 to 8.5 ꢀ 10 M cm ) of (bipyridyl)Ru(II)
derivatives typically utilized with TiO nanocrystalline systems
2
7
necessitate additional engineering in the form of scattering layers
22
phores with negligible ICT. While this intense red absorption
and additional complexation of dye at the surface to increase the
8
amount of light absorbed. While spectral overlap and extinction
5,9,10
coefficients of Ru-based dyes have improved,
can limit the performance of NLO chromophores in terms of
23
transparency in the visible region, it is intriguing for solar cell
alternatives to
applications. Other characteristics of successful NLO chromo-
phores are thermal and chemical stability. These materials are
designed for use under high photon flux, high electric fields, and
consequently elevated temperatures. Finally, if a design strategy
can be found to adapt this class of materials as surface sensitizers,
there is a large body of work to mine for new potential structures.
Critical to the success of these sensitizers is appending a func-
transition metal dyes are sought for further improvement of effi-
ciency. Recently, there have been a number of reports describing
11–15
all-organic dyes ranging from coumarin derivatives
16–20
moderately dipolar donor–acceptor systems.
to
The majority of
these dyes are promising, with an indoline based dye having an
efficiency of just over 9% when interfaced with an optimized
21
2
electrolyte and TiO film thickness.
2
tional group capable of anchoring to TiO forming a metal ester
24
via an appended carboxylic acid. Many of the all-organic dyes
mentioned above utilize a cyanoacrylic acid terminus that acts
1
1,19,25,26
both as an acceptor and an anchor.
introducing a stronger acceptor that is also able to anchor to
TiO and would introduce a red-shifted absorption band relative
to the best performing dyes, have an increased molar absorption
We were interested in
a
Department of Chemistry, University of Washington, Seattle, WA, 98195,
USA. E-mail: glenn.bartholomew@sirigen.com
b
Department of Chemistry, New Mexico Highlands University, Las Vegas,
NM, 87701, USA
2
5
ꢁ1
ꢁ1
†
Electronic supplementary information (ESI) available: Synthetic
coefficient (>1 ꢀ 10 M cm ), and provide an advantageous
procedures for tert-butyl-2-cyanoacetate and compound 9, NMR
spectra for compounds 2, 5, 8–11, and tert-butyl-2-cyanoacetate, cyclic
voltammetry data for 2 and 11, crystallography data for 10 and 11,
and J–V curves for solar cells discussed in the text. Crystal structure
data for compounds 10 and 11 has been deposited with the Cambridge
Crystallographic Data Center (numbers 789672 and 789673). See DOI:
arrangement of the molecular dipole in relation to the TiO
3
2
surface to enhance charge transfer. Specifically, our efforts are
focused on the synthesis of compounds incorporating 2-cyano-2-
(
3-cyano-4,5,5-trimethylfuran-2(5H)-ylidene)ethanoic acid (1).
This compound is the mono-hydrolyzed analog of 2-dicyano-
1
‡
0.1039/c0jm03774e
Present address: Sirigen Inc, 7330 Carroll Road, Suite 150, San Diego,
methylen-3-cyano-4,5,5-trimethyl-2,5-dihydrofuran
a strong acceptor used in NLO chromophore design
(TCF),
2
7,28
CA 92121 USA.
and
4
242 | J. Mater. Chem., 2011, 21, 4242–4250
This journal is ª The Royal Society of Chemistry 2011

View Article Online
2
9
31
32
more recently for long wavelength single molecule emitters.
preparation of the TCF acceptor and the EZ-FTC dye, our
initial strategy was to hydrolyze one of the terminal nitrile groups
on EZ-FTC to the corresponding carboxylic acid. However,
attempts at hydrolysis using either acid or base catalysis led to
significant decomposition and/or was nonselective. Another
entry to our target was to replace one of the cyano groups at the
terminus of the acceptor with a carboxyl group earlier in the
synthesis. TCF is typically synthesized through the reaction of
3-hydroxy-3-methyl-2-butanone (3) with two equivalents of
malononitrile under basic conditions (Scheme 1). The resonance
stabilized deprotonated malononitrile reacts with the carbonyl in
3 to give 2-(3-hydroxy-3-methylbutan-2-ylidene)propanedinitrile
(4) after a net dehydration. Cyclization under these conditions is
spontaneous and affords 2-imino-3-cyano-4,5,5-trimethyl-2,5-
dihydrofuran (5). Condensation of 5 with a second equivalent of
malononitrile yields TCF. We aimed to intercept 5 to allow
condensation with cyanoacetic acid to give 1 (Scheme 1).
However, upon formation of 5 a second condensation with
unreacted malononitrile occurs rapidly.
The molecular structure of TCF is shown in Chart 1 along with
the structure of EZ-FTC, a prototypical NLO dye containing the
30
TCF acceptor. If the acceptor group can be effectively modified
to allow for anchoring to TiO while maintaining most of its ICT
2
character, dyes of this type can be rationally designed and
synthesized that optimize absorption, band gap, and electron
affinities/ionization potentials.
In this contribution we detail a well-defined synthesis of a dye
incorporating 1. Of particular interest are some of the non-intuitive
conditions for several of the steps required to obtain the carboxylic
acidgroupappendedtosuchastrong electronacceptorunit. Weuse
this modified acceptor in a donor–acceptor molecule, 2-cyano-2-(3-
cyano-4-((E)-4-(dibutylamino)styryl)-5,5-dimethylfuran-2(5H)-yli-
dene)ethanoic acid (2), to examine the efficacy of this approach. We
also examine the performance of 2 relative to cis-di(thiocyanato)-
0
bis(2,2-bipyridyl-4,4 -dicarboxylate)-ruthenium(II) (N3, Chart 1)
one of the first efficient Ru-based dyes for sensitizing nanocrystal-
2
line TiO in photoelectrochemical cells. We show that the relatively
simple modified NLO chromophore performs comparably to N3
under similar conditions. Theseresults areencouraging andprovide
a synthetic entry to an entire class of dyes previously examined for
nonlinear optical applications.
A microwave procedure that allows for the formation and
30,31
isolation of 5 has been reported.
In our lab, this microwave
reaction was not reproducible with respect to yield and purity. In
addition, a non-reactive lactone (6) was also obtained which
made the reaction mixture difficult to purify, resulting in even
lower yields (Scheme 1).
Results and discussion
As an alternative to the microwave synthetic route, we pursued
a step-wise condensation utilizing a protecting group for the
hydroxy group on 3, thus providing control over cyclization and
preventing a second condensation with malononitrile. To this
end, we protected 3 with tert-butyldimethylsilyl chloride to
Synthetic approach
The primary synthetic challenge was to find an efficient route to
1, which would then allow coupling to a range of aldehyde-
containing donor groups. Given the high yield and simple
33,34
give 3-(tert-butyldimethylsilyloxy)-3-methylbutan-2-one (7).
During the workup unreacted tert-butyldimethylsilyl chloride is
converted to the silyl alcohol and does not interfere in the next
condensation step, so the crude mixture containing 7 is used
without purification and is condensed with a single equivalent
34
of malononitrile to give 2-(3-(tert-butyldimethylsilyloxy)-3-
Chart 1 Molecular structures of TCF, EZ-FTC, modified acceptor 1,
dye 2, and N3.
Scheme 1 Microwave synthesis of TCF and 5, and formation of lactone
6; condensate 4 is not isolable.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4242–4250 | 4243

View Article Online
methylbutan-2-ylidene)propanedinitrile (8). Upon deprotection
reactivity of a t-butyl protected ester. The t-butyl group is more
electron donating than the ethyl group and through inductive
effects makes the ester carbonyl more reactive toward the tri-
methylsilyl iodide. Furthermore, stability of the t-butyl carbo-
cation might impart an added driving force for carbon–oxygen
bond fission. In fact, we did observe deprotection when the
t-butyl group was employed. We hypothesize that the acid
catalyzed deprotection did not work for the ethyl ester because
35
with aqueous HF, cyclization occurs to give 5 (Scheme 2).
Attempts at using typical tetrabutylammonium fluoride
(
TBAF) deprotection conditions (stirring 8 in 1 M TBAF in
36
THF) gave poor yields (ꢂ30%) that ultimately could not be
isolated from significant amounts of polymeric material formed
in the reaction. However, treatment with aqueous HF and
sonication gives high yields (96%) with no byproducts. This
result seems counter-intuitive considering 5 is known to readily
40
the acid we chose, HBr, was not acidic enough to protonate the
ester in this intensely electron poor substrate. While it might
make sense to try a stronger acid such as H SO , work by
31
condense with water under acidic conditions to give lactone
30,31
6.
We speculate that the aqueous medium limits the basicity
2
4
40
of the fluoride anion thus preventing unwanted base-catalyzed
side reactions. Additionally, hydrogen bonding between water
and HF significantly reduces the nucleophilicity of water inhib-
iting the formation of the unwanted lactone byproduct. Under
more standard conditions utilizing TBAF and THF, TBAF is
likely better solubilized and the lack of hydrogen bonding leads
to a much more reactive naked fluoride anion that catalyzes
intermolecular condensations, resulting in the formation of
polymeric material. Interestingly, without sonication the reac-
tion of 8 in aqueous HF proceeds sluggishly (52% yield after
stirring for 48 hours) with a significant amount of lactone 6 also
produced.
Landini and Rolla indicates that the bisulfate anion is not
nucleophilic enough to lead to ester cleavage. It is unclear as to
why base catalyzed hydrolysis was ineffective although given the
crystal structure data for 10 and 11, it is possible that hydroxide
anion was unable to approach the carbonyl carbon for the
reaction to take place.
To obtain tert-butyl 2-cyano-2-(3-cyano-4,5,5-trimethylfuran-
2(5H)-ylidene)ethanoate (10), which is the t-butyl protected
analogue of 1, 5 is condensed with neat tert-butyl cyanoacetate.
Running the reaction using tert-butyl cyanoacetate as the solvent
affords 10 in 66% yield. A Knoevenagel condensation of 10 with
9 gives tert-butyl 2-cyano-2-(3-cyano-4-(4-(dibutylamino)styryl)-
5,5-dimethylfuran-2(5H)-ylidene)ethanoate (11) in good yield
after purification on silica. Deprotection of 11 is then easily
accomplished using trimethylsilyl bromide and sodium iodide in
85% yield (Scheme 3).
With a reliable means to prepare pure, gram quantities of 5, we
sought a means to incorporate the cyanoacetic acid moiety. The
most direct method involves condensation of cyanoacetic acid
with 5 using acetic acid as the solvent. Despite extensive prece-
dence for the analogous condensation of cyanoacetic acid with
19,20,37,38
carbonyl groups,
we were never able to isolate 1 in
NMR analysis
significant yield, again, likely due to the unusually electron-
deficient substrate. Alternatively, Jen and co-workers demon-
In TCF and TCF containing molecules replacement of a single
cyano substituent with a carboxyl group on the exocyclic
carbon–carbon double bond introduces the possibility of
regioisomers (shown for 10 and 11 in Scheme 4). These isomers
are observable by NMR spectroscopy and by TLC (silica, 1 : 1
hexanes/ethyl acetate) for 10, 11, and 2 with the E- and Z-isomers
30
strated condensation of 5 with ethyl cyanoacetate in 40% yield,
which would afford the ethyl-protected analogue of 1. We
obtained similar yields using oil bath heating. Knoevenagel
condensation of this ethyl ester analogue with 4-dibutylamino-
benzaldehyde (9) gives a dye that has a protected carboxylic acid.
Subsequent attempts at hydrolysis of dyes containing the ethyl
ester acceptor, both acid- and base-catalyzed, gave either no
reaction or an intractable mixture of products. A powerful and
general method for the deprotection of esters using trimethylsilyl
having distinct chemical shifts and R values. In all the previous
f
39
chloride and sodium iodide introduced by Olah et al. was also
attempted. However, these conditions gave no reaction and the
starting ethyl protected acid was recovered. Given that nucleo-
philic attack by the ester carbonyl carbon on the trimethylsilyl
iodide generated in situ is a key first step, we chose to examine the
Scheme 2 Stepwise synthesis of compound 5.
Scheme 3 Synthesis of ester acceptor 10 and dye 2.
4
244 | J. Mater. Chem., 2011, 21, 4242–4250
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Scheme 4 E/Z isomerization in acceptor 10 and dye 11.
schemes, the Z-isomer is shown, but it must be assumed that both
isomers are reaction products and present in solutions used for
solar cell fabrication and spectroscopic measurements. From
isomeric mixtures, a single isomer can be isolated by the growth
of single crystals and enriched samples can be obtained by
column chromatography, but room temperature interconversion
1
occurs in solution as determined by H NMR spectroscopy.
30
Using the work of Liu and co-wokers as a guide for spec-
troscopic analysis we had expected only to obtain the Z-isomer
and consequently only three singlet peaks should appear in the
1
H NMR spectrum for acceptor 10. To our surprise, we observed
41
two sets of peaks (a total of six peaks) from samples that started
as chromatography fractions containing a single spot by TLC.
This observation suggests that the barrier to rotation about the
exocyclic double bond is on the order of kT and that 10 isom-
erizes during typical handling and storage conditions. In fact,
chromatography samples that are initially pure will show two
spots by TLC if the sample is allowed to remain at room
temperature for several hours.
We were able to grow single crystals of 10 that were used for
both X-ray diffraction experiments and NMR analysis. The
NMR proton resonances, when correlated with the X-ray data,
show that the Z-isomer constitutes ꢂ78% of the sample at
thermal equilibrium. A spectroscopic comparison of the isomeric
mixture with that of pure E-isomer 10 is shown in Fig. 1. It is
surprising that even though the Z-isomer is the major constit-
uent, the E-isomer crystallizes preferentially which suggests that
this form is less soluble.
1
Fig. 1 The H NMR spectrum of 10 at thermal equilibrium is shown (a)
along with the spectrum of the E-isomer of 10 obtained from single
crystals (b); the peak at 2.10 ppm is identified as residual acetic acid.
a bidentate adsorption mode, as has been shown for cyanoacrylic
43
acid containing dyes, the Z-isomer maximizes the titania–
donor distance as it adopts an orthogonal orientation as shown
in Fig. 2. The same bidentate adsorption mode brings the donor
in closer proximity to the metal oxide surface in the E-isomer,
which would be more favorable for charge recombination.
Protected dye 11 and deprotected dye 2 are also present as E
and Z-isomers observable by NMR spectroscopy. For 2 the ratio
is close to 1 : 1 at thermal equilibrium. As we were not able to
successfully grow single crystals of 2 we cannot definitively
correlate the peaks to a specific isomer. For 11, however, single
crystals were grown and found to be of the Z-isomer, which is
43
Recent work by Liang and co-workers for a series of cyanoa-
crylic acid and rhodanine-3-acetic acid dyes has shown that
orientation of the dye at the TiO
oriented dyes such as the E-isomer of 2 allow the redox electro-
lyte to approach the TiO surface thereby increasing charge
2
interface is critical and that side
1
also found to be the major isomer by H NMR at ꢂ64% of the
isomeric mixture at thermal equilibrium.
2
To our knowledge, this isomerization has not been reported in
asymmetrically substituted TCF containing molecules or in
cyanoacrylic acid terminated dyes used for dye-sensitized solar
cells, where E- and Z-isomers are also possible. It is expected that
the different isomers of 2 would both adsorb onto titania through
the carboxylic acid groups; however, only the Z-isomer is
expected to afford the desired maximum distance between the
donor group and allow for maximal packing of dye molecules on
the titania surface.
recombination and decreasing cell efficiency.
X-Ray crystallography
Single crystals of 10 and 11 were both grown by slow evaporation
of a solution of isomeric (E/Z) mixtures in hexanes/methylene
chloride and examined by X-ray diffraction techniques. Crystals
for 11 were of the Z-isomer while 10 crystallized as the E-isomer.
1
These same single crystals were subjected to H NMR analysis
The isomerization observed by NMR has implications for
device performance. It has been shown that with NLO dyes such
allowing the correlation of proton signals to the corresponding
isomer. ORTEP diagrams for 10 and 11 are shown in Fig. 3. A
complete table of bond lengths and angles and crystallographic
parameters is available in the ESI section (Tables S1 and S2 in
ESI†) though bond lengths and angles of interest are
discussed below.
42
as 2, the HOMO is largely localized on the donor; to prevent
unwanted charge recombination between the oxidized dye and
the injected electron, it is thus advantageous to have the donor
unit at a maximum distance from the TiO
2
surface. Assuming
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4242–4250 | 4245

View Article Online
Dye molecule 2 is structurally similar to a TCF containing
46
analogue that differed only in the substitution about the amine
nitrogen; in both cases, the diaklylaminobenzene donor is
slightly out of plane with the rest of the molecule. The angle
between the planes defined the furan ring and the benzene ring
ꢃ
ꢃ
for 11 is 9.13 , slightly less than the 13.97 reported for the TCF
46
analog. This modest angle between planes coupled with the
ꢃ
dihedral angle defined by the alkene linker (176.07 for C14–
ꢃ
C15–C16–C17 which is close to 180 that would be fully planar)
suggests there exists effective conjugation between the acceptor
and donor portions of the molecule. The average C–N bond
ꢀ
lengths in the nitrile groups are longer in 11 (average of 1.148 A)
ꢀ
when compared to 2 (average of 1.146 A), indicating effective
Fig. 2 Representation showing how the isomerization affects the
orientation of 2 adsorbed on TiO ; the Z-isomer is expected to have
a near-orthogonal orientation thus maximizing the distance between the
TiO surface while the E-isomer is expected to be nearly parallel to the
donation of electron density from the dialkylamino group to the
2
acceptor portion of the molecule in 11. However, the N1–C9
ꢀ
bond length in 11 at 1.368 A is longer than the analogous bond
2
surface bringing the donor unit in close proximity to the titania.
ꢀ
length in the TCF containing molecule at 1.359 A, which further
supports the idea that 10 is a good acceptor, but not quite as
strong as TCF.
Photophysical characterization
The UV-Vis absorption profiles of 11 and 2 were examined in
acetonitrile and found to have strong absorption in the visible
with lmax values of 555 nm and 562 nm, respectively. For 2, this is
a 28 nm redshift relative to N3 (534 nm in ethanol solution) and
an almost fourfold increase in extinction coefficient with 55 000
ꢁ1
ꢁ1
ꢁ1
ꢁ1
M
cm for 2 in acetonitrile versus 14 200 M cm for N3 in
ethanol. On loss of the t-butyl group, there is little modification
of the absorption band. The spectra of 2 and 11 are available in
the ESI section (Fig. S1†).
While we were hoping to achieve a better overlap between the
dye absorption profile and the solar emission spectrum, the ICT
interaction between the dialkylaminobenzene donor and the
modified acceptor in 2 and 11 is only moderate. As shown for
NLO dyes, the inclusion of a donor bridge such as thiophene
would allow for enhanced ICT and shift the absorption band
4
7
Fig. 3 Molecular structures of 10 (top) and 11 (bottom) with thermal
ellipsoids shown at 50% probability; hydrogen atoms have been omitted
for clarity.
maximum to longer wavelengths. Indeed, following these
preliminary studies we are moving in that direction by synthe-
sizing an analog of EZ-FTC containing 1 as well as dyes utilizing
1
directly bound through a conjugated bridge to strong electron
Acceptor molecule 10 is highly planar indicating good
p-orbital overlap throughout the molecule. The carbonyl group,
however, is twisted slightly out of the plane defined by the furan
donors.
Electrochemistry
ꢃ
ring with a dihedral angle of 15.58 for O3–C5–C6–C8. This is in
contrast to the TCF molecule, which has a mirror plane of
We used cyclic voltammetry (CV) to assign the position of the
HOMO and LUMO levels for dye 2 (cyclic voltammograms are
available in the ESI†). To better approximate the energy levels of
the dye, we chose to perform our experiments by adsorbing 2
symmetry containing the furan ring and three acceptor cyano-
44
groups. Of interest is the endocyclic C12–C14 carbon–carbon
ꢀ
double bond at a length of 1.340 A; this length is slightly longer
ꢀ
45
than typical carbon–carbon double bond at 1.32 A and is
expected given the conjugation between this double bond and the
cyano and carbonyl electron acceptor groups. In the TCF
2
onto a TiO coated ITO electrode, which was then immersed in
electrolyte solution. Under these conditions, we assign the
HOMO of 2 at approximately 5.5 eV below vacuum, when using
ꢀ
44
+
Fc/Fc as an internal standard and taking the Fc/Fc redox
+
acceptor the analogous bond is similar at 1.343 A. The presence
of the carboxyl group leads to a slight lengthening of the C–N
48
couple as 5.12 eV below vacuum.
ꢀ
triple bonds with an average distance of 1.146(2) A in 10 relative
We were not able to observe a clear onset of reduction for 2
ꢀ
44
to 1.136(4) A in TCF. This perturbation relative to TCF could
be due to the carboxyl group being less electron withdrawing
than the cyano group thus leading to an overall lengthening of
the C–N bonds.
under our experimental conditions (either in solution or adsor-
bed on TiO
(ꢂ670 nm) to calculate the position of the LUMO. We obtain
a value of 3.6 eV for the LUMO of 2 when using E
¼ 1.85 eV.
2
) and thus have used the onset of absorption
g
4
246 | J. Mater. Chem., 2011, 21, 4242–4250
This journal is ª The Royal Society of Chemistry 2011

View Article Online
a
Table 2 Efficiency of 2 and N3 without TBP
This lies above the conduction band edge for TiO , typically
2
49,50
taken as 4.0–4.2 eV
be as high as 3.74 eV below vacuum. As observed in other cases
where the HOMO level used with E derived by UV-Vis spec-
below vacuum but has been reported to
ꢁ
2
51
Dye
J
sc/mA cm
VOC/V
FF
h (%)
g
2
N3
ꢁ11.70
ꢁ12.26
0.36
0.38
0.60
0.49
2.55
2.33
troscopy, the calculated position of the LUMO may be higher
than estimated. Given the position of the LUMO, however,
a
The electrolyte contained 0.9 M LiI and 0.03 M I
-t-butylpyridine. TiO film thickness was 3.9 mm.
2
in acetonitrile. TBP is
2
electron transfer into the TiO conduction band should be facile.
4
2
As a comparison, the electrochemistry of 11 was also exam-
ined. In solution we assign a HOMO level of 5.5 eV below
vacuum, exactly as was obtained for 2 adsorbed on a TiO
2
product of dye aggregation or recombination of the oxidized dye
or the electrolyte with injected electrons as has been reported for
other systems. When considering the fact that the E isomer
brings the amine moiety closer to the metal oxide surface, we
believe that our cells likely suffer from unwanted charge
recombination.
electrode. We were also able to obtain reproducible reduction
scans, although the reduction process is not reversible. Using the
onset of reduction, compound 11 has a LUMO level of 4.0 eV
below vacuum, slightly lower than the value calculated for 2 but
close to the value calculated for N3 by ultraviolet photoelectron
51
spectroscopy (UPS) and the energy of absorption at lmax
. The
To test our hypothesis, we decided to examine the efficiency of
cells employing an electrolyte with a higher lithium ion concen-
tration while also moving to thinner films to take advantage of
the high molar extinction coefficient of 2. The lithium ion is
known to effectively adsorb onto titanium dioxide thereby
leading to a decrease in the conduction band level. However, this
adsorption also leads to an effective screening of injected elec-
trons, which may aid in decreasing losses resulting from charge
recombination. The use of t-butylpyridine (TBP) as an electro-
lyte additive is known to increase the energy at the surface of
loss of the t-butyl group is not expected to lead to significant
energy level perturbation; therefore, we believe that the energy
levels derived from 11 are an excellent approximation for the
2
position of the frontier orbitals of 2 adsorbed onto TiO .
Solar cell performance
2
The effectiveness of 2 as a sensitizer of TiO in a dye-sensitized
solar cell was examined. Nanocrystalline titanium dioxide elec-
trodes were fabricated in house and solar cells assembled as
detailed in the Experimental section. Initial experiments were
performed using a thicker (7.2 mm thick) titanium dioxide film
and the widely reported redox electrolyte optimized for Ru-based
TiO conduction band, which leads to improvements in the open
2
53,54
circuit voltage (VOC) for Ru-based dyes,
but we chose to
exclude TBP in our studies involving higher lithium concentra-
tions. Indeed, with TBP omitted and with the lithium iodide
increase from 0.3 M to 0.9 M we did see a drop in VOC of 110 mV
for cells employing 2, but also observed a nearly four-fold
increase in JSC affording cells with an efficiency of 2.55%. This
performance was similar to N3 under identical conditions at
2.33%. Given that increased lithium ion concentration gave an
increase in the value of JSC, we are confident that charge
recombination is the major hinderance to realizing the full
potential of dye 2.
dyes (0.3 M LiI /0.03 M I
to optimize the cell parameters for 2 we also examined cells with
thinner (3.9 mm thickness) TiO films and an electrolyte with
2
/0.5 M t-butylpyridine). As we desired
2
a higher lithium concentration and no t-butylpyridine. Thinner
films were expected to be better suited to use with a dye having
52
a high extinction coefficient such as 2.
The data obtained using the prototypical electrolyte system
and thicker titanium dioxide films are shown in Table 1. We
obtain an efficiency of 1.02% for 2 compared to 5.55% for N3
under identical conditions; this low performance was unex-
pected. While the difference in VOC values between the two
sensitizers is only 180 mV, there is a four-fold difference in Jsc
between 2 and N3. We attribute the lower VOC value to
a downward shift in the titanium dioxide conduction band due to
the presence of E and Z-isomers of 2. The adsorption of the
Conclusions
In conclusion, we have developed an alternative, well-defined
synthesis for modification of the strong acceptor group TCF to
E-isomer would have an orientation relative to the TiO
2
surface
allow for surface binding to TiO . Synthesis of a dye, 2, with
2
43
(
Fig. 2) similar to that of previously examined rhodanine dyes;
strong ICT character that employs the modified acceptor 1 is
detailed along with some unusual chemistry related to the highly
electron-deficient nature of the substrates involved. Unlike other
studies involving cyanoacrylate acceptor groups, we report E/Z
isomerization about the cyanoacrylate double bond with these
isomers being observable by NMR spectroscopy. Our prelimi-
nary results with dye-sensitized photoelectrochemical cells in
direct comparison with N3 demonstrate that dye designs typi-
cally thought of as optimized NLO chromophores may prove to
be significant materials in dye sensitization of metal oxides for
solar energy applications. However, the ability of the molecule to
thermally isomerize may give rise to unwanted charge recombi-
nation and a downward shift of the titanium dioxide conduction
band leading to decreases in cell efficiency.
it is this slanted orientation and the associated molecular dipole
having a small vertical component responsible for the lower VOC
in the rhodanine dyes. The low short circuit current may be the
a
Table 1 Efficiency of 2 and N3 employed in nanocrystalline TiO2 dye-
sensitized solar cells
ꢁ
2
Dye
Jsc/mA cm
V
OC/V
FF
h (%)
2
N3
ꢁ3.17
0.47
0.65
0.67
0.67
1.02
5.55
ꢁ12.89
a
The electrolyte employed consisted of 0.3 M LiI/0.03 M I
film thickness was 7.2 mm.
2
/0.5 M
t-butylpyridine; TiO
2
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4242–4250 | 4247

View Article Online
radiation using a 150 W Oriel Xe lamp with a KG5 filter and
calibrated with a standard silicon photovoltaic cell traceable to
a primary NREL standard. The global efficiency was determined
according to eqn (1) and (2).
Experimental section
Materials and methods
All reagents were purchased from Acros or Aldrich and used as
received unless stated otherwise. N3 was obtained for Solaronix and
used without further purification. DMF and THF were dried by
passing argon degassed solvent through activated alumina columns.
Chromatography was carried out on a CombiFlash Companion
purification system using pre-packed silica columns. The major band
was collected and all fractions containing a single spot by TLC were
combined and solvent removed. Further purification of impure
fractions was carried out if necessary to significantly improve yields.
References are provided for compounds that have been previously
synthesized and for analogous compounds. NMR spectra were run
JSCVOC
hglobal
¼
FF
(1)
P
S
P
Max
FF ¼
(2)
JSCVOC
1
on a Brucker AV300 operating at 300 MHz for H experiments and
Synthetic details
2-(3-(tert-Butyldimethylsilyloxy)-3-methylbutan-2-ylidene)-
1
3
7
5 MHz for C experiments. All spectra are referenced to the
55
deuterated solvent used as an internal standard. Compounds 2, 10,
and 11 show E/Z-isomerism about the exocyclic double bond. Single
crystals of isomerically pure 10 and 11 were obtained and used to
3
3,34
propanedinitrile (8)
. In a 100 mL flask was combined N,N-
dimethylaminopyridine (0.290 g, 0.38 mmol), imidazole
(4.858 g, 71.31 mmol), tert-butyldimethylsilyl chloride (8.598
g, 57.04 mmol), and DMF (25 mL). After stirring at room
temperature for 10 min, 3-hydroxy-3-methyl-2-butanone
1
13
correlate H and C NMR spectral peaks with isomerization state.
For 2 and 11 it is assumed that the dibutylamino peaks for the two
1
3
isomers in the C NMR spectrum overlap, consistent with the
number of signals observed. High-resolution mass spectra were
obtained on either a Bruker APEX III 47e or a Sciex Qstar XL mass
spectrometer. Cyclic voltammetry experiments were performed on
an EG&G Princeton Applied Research model 362 scanning poten-
tiostat with two 2010 Keithley multimeters in acetonitrile using tet-
rabutylammonium hexafluorophosphate (0.1 M concentration) as
the supporting electrolyte. The analyte concentration was typically
ꢂ3 mM. A Pt button was used as the working electrode, a Pt flag as
(
92%, 5.0 mL, 48 mmol) was added slowly in one portion and
ꢃ
the reaction heated to 100 C overnight (ꢂ18 hours). After
cooling to room temperature and diluting with ether (400
mL), the product was isolated by washing with water (3 ꢀ 50
mL) and brine (50 mL). The ether phase was dried over
4
MgSO , filtered, and solvent removed under reduced pres-
sure. The resulting light brown liquid (crude compound 7)
was added to benzene (25 mL) containing malononitrile
(
3.458 g, 52.35 mmol), 3-aminophenol (0.52 g, 4.76 mmol),
the counter electrode, and an Ag/AgNO
3
electrode was purchased
and HOAc (1 mL). Using a Dean–Stark setup to remove
ꢃ
from BAS used as the reference electrode. Scans are referenced to the
+
48
water, the reaction was heated to 105 C overnight, then
Fc/Fc internal redox couple (5.12 eV relative to vacuum). Single
crystal X-ray diffraction experiments were performed on a Bruker
SMART APEX2 CCD diffractometer and corrected for absorption
cooled to room temperature and diluted with methylene
chloride (200 mL). After washing with 1 M aqueous HCl (2 ꢀ
56
50 mL), saturated aqueous NaHCO (2 ꢀ 50 mL), and brine
3
using the SUDABS program. All calculations were carried out
5
7
(50 mL), the organic phase was dried over MgSO and solvent
4
using the SHELXTL program.
Fabrication and characterization of solar cells
TiO nanoparticles were synthesized according to literature
removed under reduced pressure to obtain a brown oil that
was chromatographed on silica eluting with a hexanes/
methylene chloride gradient. A white, crystalline solid was
1
obtained upon removal of solvent (10.535 g, 84%). H NMR
2
58
reports then mixed with alpha-terpineol and ethyl cellulose.
(300 MHz, CDCl ) d 2.32 (s, 3H), 1.60 (s, 6H), 0.91 (s, 9H),
3
1
0.19 (s, 6H); C NMR (75 MHz, CDCl ) d 188.2, 113.0, 112.4,
3
Films were deposited on FTO (with a compact layer of TiO ) via
2
3
59
spincoating to give the desired thickness. The resulting films
ꢃ
83.8, 77.1, 28.7, 25.9, 21.6, 18.3, ꢁ2.1; HRMS (ESI) m/z calcd.
+
for C H N OSi 265.1736, found 265.1735.
were sintered at 450 C for 45 minutes then impregnated with
ꢃ
1
4
25
2
0
.05 M TiCl
4
at 70 C for 30 minutes followed by another sin-
ꢃ
tering at 450 C for 30 minutes. After sintering from the TiCl
ꢃ
4
2-Imino-4,5,5-trimethyl-2,5-dihydrofuran-3-carbonitrile (5). In
a polyethylene bottle was placed 8 (4.000 g, 15.13 mmol),
acetonitrile (7 mL), and HF (20 mL, 48% aqueous). The biphasic
reaction was sonicated until TLC indicated that no starting
material remained (4 hours), then poured into aqueous KOH
(150 mL, 5 M) and extracted with methylene chloride (10 ꢀ
50 mL). The combined organic phases were washed with brine
step, the substrates were cooled to 80 C and immersed in
ꢁ4
a sealed container of 3 ꢀ 10 M dye solution in ethanol. After
soaking overnight, the stained photoelectrodes were removed
and rinsed with ethanol and dried under a stream of nitrogen. A
spacer (SX-1170, Solaronix) was cut into strips and used between
the photoelectrode and platinized counter electrode. Platinized
counter electrodes were made by spin coating
a
hexa-
(100 mL), dried over MgSO , filtered, and solvent removed under
4
ꢃ
chloroplatinic acid on FTO followed by calcinations at 370 C
for 30 minutes. Film thickness was measured on a Tencor Alpha-
Step 500 profilometer. The assembled cells were stored in
a desiccator prior to testing. Current–voltage characteristics were
reduced pressure to give a white solid whose spectra matched
30
that previously reported in the literature (2.147 g, 95%).
1
Additional purification was unnecessary. H NMR (300 MHz,
1
3
CDCl
3
) d 6.99 (s, 1H), 2.14 (s, 3H), 1.41 (s, 6H); C NMR
) d 176.4, 164.1, 111.1, 105.3, 90.2, 24.5, 13.0.
ꢁ2
recorded at 100 mW cm under simulated AM 1.5 solar
(75 MHz, CDCl
3
4
248 | J. Mater. Chem., 2011, 21, 4242–4250
This journal is ª The Royal Society of Chemistry 2011

View Article Online
tert-Butyl 2-cyano-2-(3-cyano-4,5,5-trimethylfuran-2(5H)-yli-
dene)ethanoate (10). In a round bottom flask with air condenser
was added 5 (1.000 g, 6.65 mmol) and tert-butyl-2-cyanoacetate
with methylene chloride (3 ꢀ 10 mL). Using a stream of
nitrogen the volume of the combined extracts was reduced
(approximately 5–10 mL) and the product precipitated by slow
addition of hexane. A dark powder was isolated by suction
(
5 mL). The faint yellow solution was placed under vacuum with
stirring for an hour to remove residual water, then backfilled
ꢃ
filtration (0.075 g, 85%) that was pure by spectroscopic anal-
1
ysis. E- and Z-isomers H NMR (300 MHz, CDCl
with argon and stirred at 100 C for 48 hours after which time the
3
) d 7.80 (d,
reaction was cooled to room temperature. Residual tert-butyl-2-
cyanoacetate was removed under reduced pressure (5 mTorr)
with mild heating. The remaining brown solid was chromato-
graphed on silica eluting with a hexanes/ethyl acetate gradient.
Upon removal of solvent, a light brownish-white solid was
J ¼ 16.2 Hz, 1H), 7.78 (d, J ¼ 16.2 Hz, 1H), 7.69 (d, J ¼ 9.0 Hz,
2H), 7.68 (d, J ¼ 9.0 Hz, 2H), 6.88 (d, J ¼ 16.2 Hz, 1H), 6.84 (d,
J ¼ 16.2 Hz, 1H), 6.78 (d, J ¼ 9.0 Hz, 2H), 6.77 (d, J ¼ 9.0 Hz,
1
3
3
2H); E- and Z-isomers C NMR (75 MHz, CDCl ) d 176.1,
174.4, 174.3, 172.5, 164.0, 163.6, 151.4, 151.3, 147.6, 147.1,
132.1, 121.6, 121.4, 116.2, 115.7, 112.7, 112.5, 111.9, 111.9,
108.7, 108.5, 96.4, 95.7, 94.8, 93.7, 75.1, 75.0, 50.0, 29.0, 26.1,
obtained (1.214 g, 66%). Two isomers were present by TLC
1
silica, 1 : 1 hexanes/ethyl acetate) and H NMR, but the
(
+
E-isomer could be preferentially crystallized from hexane/
26.0, 19.5, 13.8. HRMS (ESI) m/z calcd. for C H N NaO
3
2
6
31
3
1
methylene chloride. The H and C NMR spectra for this isomer
13
456.2263, found 456.2258.
1
are reported. E-isomer H NMR (300 MHz, CDCl
3
) d 2.29
s, 2H), 1.51 (s, 6H), 1.47 (s, 9H); E-isomer C NMR (75 MHz,
CDCl ) d 182.5, 172.5, 161.0, 114.5, 110.5, 105.9, 95.2, 82.9, 80.6,
7.8, 24.2, 13.9; Z-isomer H NMR (300 MHz, CDCl
s, 3H), 1.60 (s, 6H), 1.52 (s, 9H). HRMS (ESI) m/z calcd. for
13
(
Acknowledgements
3
1
We wish to thank the University of Washington and the NSF
CHE 0610193 and STC-MDITR No. DMR 0120967 and No.
2
3
) d 2.32
(
(
+
DMR 0934212) for funding. We also wish to thank Dr Martin
Sadilek for help with obtaining mass spectra, Dr Philip Sullivan
for helpful discussions regarding acceptor chemistry, and Dr
Deanna Rodovsky for helpful discussions regarding device
testing.
15 19 2 3
C H N O 275.1396, found 275.1400.
tert-Butyl-2-cyano-2-(3-cyano-4-(4-(dibutylamino)styryl)-5,5-
dimethylfuran-2(5H)-ylidene)ethanoate (11). flask with
Dean–Stark setup was charged with 9 (0.300 g, 1.29 mmol), 3-
A
aminophenol (0.056 g, 0.51 mmol), HOAc (0.15 mL), and
ꢃ
benzene (10 mL). The mixture was heated to 105 C and after 5
References
mL of benzene/water collected in the Dean–Stark trap 10 (0.352
g, 1.28 mmol) dissolved in benzene (5 mL) was added drop-wise
over 30 min. The benzene/water azeotrope was periodically
drained during the addition. After stirring for 2 h the reaction
was cooled, solvent removed under reduced pressure, and the
dark residue chromatographed on silica eluting with a gradient
of hexanes/ethyl acetate. Upon removal of solvent a dark red
1
M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker,
E. M u€ ller, P. Liska, N. Vlachopoulos and M. Gr €a tzel, J. Am.
Chem. Soc., 1993, 115, 6382.
B. O’Regan and M. Gr €a tzel, Nature, 1991, 353, 737.
J.-H. Yum, P. Walter, S. Huber, D. Rentsch, T. Geiger, F. N u€ esch,
F. De Angelis, M. Gr €a tzel and M. K. Nazeeruddin, J. Am. Chem.
Soc., 2007, 129, 10320.
2
3
4 A. Kukrek, D. Wang, Y. Hou, R. Zong and R. P. Thummel, Inorg.
Chem., 2006, 45, 10131.
1
solid was obtained (0.486 g, 77%); Z-isomer H NMR (300
5
C. S. Karthikeyan, H. Wietasch and M. Thelakkat, Adv. Mater., 2007,
9, 1091.
6 P. Wang, C. Klein, R. Baker-Humphry, S. M. Zakeeruddin and
M. Gr
MHz, CDCl ) d 7.49 (d, 2H, J ¼ 9.0 Hz), 7.43 (d, 1H, J ¼ 18.0
3
1
Hz), 6.83 (d, 1H, J ¼ 15.0 Hz), 6.64 (d, 2H, J ¼ 9.0 Hz), 3.37 (t,
€
4
H, J ¼ 7.5 Hz), 1.76 (s, 6H), 1.61 (m, 4H), 1.53 (s, 9H), 1.38
13
(
m, 4H), 0.98 (t, 6H, J ¼ 7.5 Hz); Z-isomer C NMR (75 MHz,
CDCl ) d 173.0, 172.0, 161.6, 151.4, 146.1, 131.6, 121.3, 115.2,
12.2, 111.7, 109.0, 96.9, 95.7, 81.5, 77.9, 50.8, 29.3, 28.1, 26.8,
2
3
8
J. Kr u€ ger, R. Plass, M. Gr €a tzel and H.-J. Matthieu, Appl. Phys. Lett.,
2002, 81, 367.
1
2
1
9 C.-Y. Chen, S.-J. Wu, C.-G. Wu, J.-G. Chen and K.-C. Ho, Angew.
Chem., Int. Ed., 2006, 45, 5822.
3
0.1, 13.8; E-isomer H NMR (300 MHz, CDCl ) d 7.51 (d, 2H,
J ¼ 9.0 Hz), 7.38 (d, 1H, J ¼ 15.0 Hz), 6.96 (d, 1H, J ¼ 15.0 Hz),
1
0 M. K. Nazeeruddin, P. P ꢁe chy and M. Gr €a tzel, Chem. Commun., 1997,
6
1
.64 (d, 2H, J ¼ 9.0 Hz), 3.37 (t, 4H, J ¼ 7.5 Hz), 1.73 (s, 6H),
1
11 K. Hara, M. Kurashige, Y. Dan-oh, C. Kasada, A. Shinpo, S. Suga,
705.
.61 (m, 4H), 1.56 (s, 9H), 1.38 (m, 4H), 0.98 (t, 6H, J ¼ 7.5 Hz).
+
K. Sayama and H. Arakawa, New J. Chem., 2003, 27, 783.
2 K. Hara, T. Sato, R. Katoh, A. Furube, Y. Ohga, A. Shinpo, S. Suga,
HRMS (ESI) m/z calcd. for C H N O 490.3070, found
3
3
0
40
3
1
4
90.3064.
K. Sayama, H. Sugihara and H. Arakawa, J. Phys. Chem. B, 2003,
1
07, 597.
3 K. Hara, K. Sayama, Y. Ohga, A. Shinpo, S. Suga and H. Arakawa,
Chem. Commun., 2001, 569.
4 K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara,
Y. Dan-oh, C. Kasada, A. Shinpo and S. Suga, J. Phys. Chem. B,
2005, 109, 15476.
1
1
2
-Cyano-2-(3-cyano-4-((E)-4-(dibutylamino)styryl)-5,5-dime-
thylfuran-2(5H)-ylidene)ethanoic acid (2). Following a proce-
3
9
dure similar to that by Olah and co-workers, a flask
containing 11 (0.100 g, 0.20 mmol) and NaI (0.122 g, 0.81
mmol) was placed under vacuum overnight, then backfilled
with argon. Acetonitrile (2 mL) was added and the solution
1
5 Z.-S. Wang, Y. Cui, K. Hara, Y. Dan-oh, C. Kasada and A. Shinpo,
Adv. Mater., 2007, 19, 1138.
1
6 D. P. Hagberg, T. Edvinsson, T. Marinado, G. Boschloo, A. Hagfeldt
and L. Sun, Chem. Commun., 2006, 2245.
ꢃ
cooled to 0 C. Trimethylsilyl bromide (0.11 mL, 0.82 mmol)
was then added dropwise over ꢂ1 minute. Upon complete
17 T. Horiuchi, H. Miura, K. Sumioka and S. Uchida, J. Am. Chem.
Soc., 2004, 126, 12218.
addition, the reaction was allowed to stir and slowly warm to
ꢃ
1
8 S. Kim, J. K. Lee, S. O. Kang, J. Ko, J.-H. Yum, I. Fantacci,
F. D. Angelis, D. D. Censo, M. K. Nazeeruddin and M. Gr €a tzel,
J. Am. Chem. Soc., 2006, 128, 16701.
room temperature overnight. After re-cooling to 0 C, cold
water (5 mL) was added slowly. The product was extracted
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4242–4250 | 4249

View Article Online
1
2
2
9 K. R. J. Thomas, J. T. Lin, Y.-C. Hsu and K.-C. Ho, Chem. Commun.,
41 When we synthesized the ethyl protected version of compound 10 as
was done in reference 30, we also observe two sets of peaks
corresponding to E and Z-isomer formation. Consequently, we do
not believe that isomerization is due to the use of the t-butyl
protecting group.
42 T. Kinninbrugh, S. Bhattacharjee, P. Sullivan, C. Isborn,
B. H. Robinson and B. E. Eichinger, J. Phys. Chem. B, 2006, 110,
13512.
43 Y. Liang, B. Peng and J. Chen, J. Phys. Chem. C, 2010, 114, 10992.
44 S. Y. Li, Y. Y. Song, Z. L. You, Y. W. Wen and J. G. Qin, Acta
Crystallogr., Sect. E: Struct. Rep. Online, 2005, 61, O2093–O2095.
45 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen
and R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1987, S1–S19.
46 S. Li, M. Li, J. G. Qin, M. L. Tong, X. M. Chen, T. Liu, Y. Fu,
S. X. Wu and Z. M. Su, CrystEngComm, 2009, 11, 589.
47 D. M. Casmier, P. A. Sullivan, O. Clot, K. A. Firestone,
B. E. Eichinger, S. Lee, S. Heller, A. Brumbaugh, B. Millard and
L. R. Dalton, Proc. SPIE, 2004, 5351, 243.
2005, 4098.
0 M. Velusamy, K. R. J. Thomas, J. T. Lin, Y.-C. Hsu and K.-C. Ho,
Org. Lett., 2005, 7, 1899.
1 S. Ito, S. M. Zakeeruddin, R. Humphry-Baker, P. Liska, R. Charvet,
P. Comte, M. K. Nazeeruddin, P. P ꢁe chy, M. Takata, H. Miura,
S. Uchida and M. Gr €a tzel, Adv. Mater., 2006, 18, 1202.
2 F. Meyers, S. R. Marder and J. W. Perry, in Chemistry of Advanced
Materials: An Overview, ed. L. V. Interrante and M. J. Mampden-
Smith, Wiley-VCH, Inc., New York, 1998, pp. 207–269.
3 G. Mignani, M. Barzoukas, J. Zyss, G. Soula, F. Balegroune,
D. Grandjean and D. Josse, Organometallics, 1991, 10, 3660.
4 E. Galoppini, Coord. Chem. Rev., 2004, 248, 1238.
5 A. Furube, R. Katoh, K. Hara, T. Sato, S. Murata, H. Arakawa and
M. Tachiya, J. Phys. Chem. B, 2005, 109, 16406.
6 K. Hara, T. Sato, R. Katoh, A. Furube, T. Yoshihara, M. Marai,
M. Kurashige, S. Ito, A. Shinpo, S. Suga and H. Arakawa, Adv.
Funct. Mater., 2005, 15, 246.
2
2
2
2
2
27 L. R. Dalton, P. A. Sullivan and D. H. Bale, Chem. Rev., 2010, 110, 25.
28 P. A. Sullivan and L. R. Dalton, Acc. Chem. Res., 2010, 43, 10.
29 Z. K. Lu, N. Liu, S. J. Lord, S. D. Bunge, W. E. Moerner and
R. J. Twieg, Chem. Mater., 2009, 21, 797.
48 B. Thompson, PhD Dissertation, University of Florida, Gainesville,
FL, 2005.
49 K. Kalyanasundaram and M. Gr €a tzel, Coord. Chem. Rev., 1998, 77,
347.
30 S. Liu, M. A. Haller, H. Ma, L. R. Dalton, S.-H. Jang and A. K.-
Y. Jen, Adv. Mater., 2003, 15, 603.
50 V. S. Saji, K. Zong and M. Pyo, J. Photochem. Photobiol., A, 2010,
212, 81.
3
3
1 D. Villemin and L. Liao, Synth. Commun., 2001, 31, 1771.
2 F. Wang, A. S. Ren, M. He, M. S. Lee, A. W. Harper, L. R. Dalton,
H. Zhang, S. M. Garner, A. Chen and W. H. Steier, Polym. Prepr.
51 J. H. Snook, L. A. Samuelson, J. Kumar, Y.-G. Kim and J. E. Witten,
Org. Electron., 2005, 6, 55.
52 L. Schmidt-Mende and M. Gratzel, Thin Solid Films, 2006, 500, 296.
53 G. Boschloo, L. Haggman and A. Hagfeldt, J. Phys. Chem. B, 2006,
110, 13144.
54 G. Schlichth o€ rl, S. Y. Huang, J. Sprague and A. J. Frank, J. Phys.
Chem. B, 1997, 101, 8141.
55 H. E. Gottlieb, V. Kotlyar and A. Nudelman, J. Org. Chem., 1997, 62,
7512.
(Am. Chem. Soc., Div. Polym. Chem.), 1999, 39, 1065.
3
3
3
3
3 D. G. McCarthy, C. C. Collins, J. P. O’Driscoll and S. E. Lawrence, J.
Chem. Soc., Perkin Trans. 1, 1999, 3667.
4 F. S. Prout, A. A. Abdel-latif and M. R. Kamal, J. Chem. Eng. Data,
1
963, 8, 597.
5 E. W. Collington, H. Finch and I. J. Smith, Tetrahedron Lett., 1985,
6, 681.
2
56 G. M. Sheldrick, Universit €a t G o€ ttingen, Germany, 2004.
57 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112.
6 T. W. Greene and P. G. M. Wuts, Protective Groups in Organic
Synthesis, John Wiley and Sons, 1999.
3
3
7 H. Maehr and J. M. Smallheer, J. Org. Chem., 1981, 46, 1752.
8 D. Prajapati and J. S. Sandhu, J. Chem. Soc., Perkin Trans. 1, 1993,
58 C. J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann,
V. Shklover and M. Gratzel, J. Am. Ceram. Soc., 1997, 80,
3157.
739.
3
9 G. A. Olah, S. C. Narang, B. G. B. Gupta and R. Malhotra, J. Org.
Chem., 1979, 44, 1247.
0 D. Landini and F. Rolla, J. Org. Chem., 1982, 47, 154.
59 A. C. Arango, L. R. Johnson, V. N. Bliznyuk, Z. Schlesinger,
S. A. Carter and H.-H. H o€ rnhold, Adv. Mater., 2000, 12,
1689.
4
4
250 | J. Mater. Chem., 2011, 21, 4242–4250
This journal is ª The Royal Society of Chemistry 2011