Modified phthalocyanines for efficient near-IR sensitization of nanostructured TiO2 electrode

Article abstract of DOI:10.1021/ja0178012

A zinc phthalocyanine with tyrosine substituents (ZnPcTyr), modified for efficient far-red/near-IR performance in dye-sensitized nanostructured TiO₂ solar cells, and its reference, glycine-substituted zinc phthalocyanine (ZnPcGly), were synthes

Full text of DOI:10.1021/ja0178012

Published on Web 04/06/2002
Modified Phthalocyanines for Efficient Near-IR Sensitization of
Nanostructured TiO2 Electrode
Jianjun He,† Ga´bor Benko¨,† Ferenc Korodi,‡ Toma´s Pol´ıvka,† Reiner Lomoth,§
Bjo¨rn Åkermark,‡ Licheng Sun,*,‡ Anders Hagfeldt,§ and Villy Sundstro¨m*,†
Contribution from the Department of Chemical Physics, Lund UniVersity, S-221 00 Lund,
Sweden, Department of Organic Chemistry, Stockholm UniVersity, S-106 91 Stockholm, Sweden,
and Department of Physical Chemistry, Uppsala UniVersity, S-751 21 Uppsala, Sweden
Received December 18, 2001
Abstract: A zinc phthalocyanine with tyrosine substituents (ZnPcTyr), modified for efficient far-red/near-
IR performance in dye-sensitized nanostructured TiO2 solar cells, and its reference, glycine-substituted
zinc phthalocyanine (ZnPcGly), were synthesized and characterized. The compounds were studied
spectroscopically, electrochemically, and photoelectrochemically. Incorporating tyrosine groups into
phthalocyanine makes the dye ethanol-soluble and reduces surface aggregation as a result of steric effects.
The performance of a solar cell based on ZnPcTyr is much better than that based on ZnPcGly. Addition of
3R,7R-dihydroxy-5â-cholic acid (cheno) and 4-tert-butylpyridine (TBP) to the dye solution when preparing
a dye-sensitized TiO2 electrode diminishes significantly the surface aggregation and, therefore, improves
the performance of solar cells based on these phthalocyanines. The highest monochromatic incident photo-
to-current conversion efficiency (IPCE) of ∼24% at 690 nm and an overall conversion efficiency (η) of
0.54% were achieved for a cell based on a ZnPcTyr-sensitized TiO2 electrode. Addition of TBP in the
electrolyte decreases the IPCE and η considerably, although it increases the open-circuit photovoltage.
Time-resolved transient absorption measurements of interfacial electron-transfer kinetics in a ZnPcTyr-
sensitized nanostructured TiO2 thin film show that electron injection from the excited state of the dye into
the conduction band of TiO2 is completed in ∼500 fs and that more than half of the injected electrons
recombines with the oxidized dye molecules in ∼300 ps. In addition to surface aggregation, the very fast
electron recombination is most likely responsible for the low performance of the solar cell based on ZnPcTyr.
Introduction
The DSSC is a sophisticated system, which is not yet
understood in detail. To increase the conversion efficiency as
well as the stability, almost all the elements of the solar cell
device need further improvement. As for the sensitizing dye,
there are several theoretical requirements that it has to comply
with.5 First, the lowest unoccupied molecular orbital (LUMO)
of the dye should have a higher energy than the conduction
band edge of the semiconductor and have a good orbital overlap
to facilitate electron injection. Ideally, the dye should have
intensive absorption in the whole solar spectrum. Second, the
dye should be attached strongly to the semiconductor surface
and inject electrons into the conduction band with a quantum
yield of unity, while the charge recombination between the
injected electron and the oxidized dye should be slow enough
for the electron transport to the external circuit. Third, the redox
The dye-sensitized photoelectrochemical solar cell (DSSC)
is one of the recent possibilities of harvesting solar energy by
converting it into electricity. The heart of this cell is a
photoanode, which is based on a nanoporous nanocrystalline,
commonly called nanostructured, TiO2 film covered by a
monolayer of a sensitizing dye.1-3 Upon light excitation, the
adsorbed dye molecules inject electrons from their excited states
into the conduction band of the semiconductor. The electrons
are brought back to the oxidized dye through an external circuit,
a platinum counter electrode, and a redox system (typically
I-/I3-). The use of a nanostructured TiO2 film together with
the Ru(dcbpy)2(NCS)2 [dcbpy ) 4,4′-dicarboxy-2,2′-bipyridine]
(RuN3 for short) dye introduced by Gra¨tzel and co-workers1,2
was a breakthrough in terms of a commercial application. An
overall conversion efficiency of ∼10% has been achieved. Since
then, extensive attention has been paid to this research field.4
(4) For example: (a) Hagfeldt, A.; Gra¨tzel, M. Acc. Chem. Res. 2000, 33,
269. (b) Kelly, C. A.; Meyer, G. J. Coord. Chem. ReV. 2001, 211, 295. (c)
Bonhoˆte, P.; Moser, J.-E.; Humphry-Baker, R.; Vlachopoulos, N.; Zakeer-
uddin, S. M.; Walder, L.; Gra¨tzel, M. J. Am. Chem. Soc. 1999, 121, 1324.
(d) Kalyanasundaram, K.; Gra¨tzel, M. Coord. Chem. ReV. 1998, 77, 347.
(e) Huang, S. Y.; Schlichtho¨rl, G., Nozik, A. J.; Gra¨tzel, M.; Frank, A. J.
J. Phys. Chem. B 1997, 101, 2576. (f) Hara, K.; Sayama, K.; Ohga, Y.;
Shinpo, A.; Suga, S.; Arakawa, H. Chem. Commun. 2001, 569. (g)
Schwarzburg, K.; Willig, F. J. Phys. Chem. B 1999, 103, 5743.
(5) Nazeeruddin, M. K.; Pe´chy, P.; Renouard, T.; Zakeeruddin, S. M.;
Humphry-Baker, R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover,
V.; Spiccia, L.; Deacon, G. B.; Bignozzi, C. A.; Gra¨tzel, M. J. Am. Chem.
Soc. 2001, 123, 1613.
* Corresponding author. E-mail: Villy.Sundstrom@chemphys.lu.se.
† Lund University.
‡ Stockholm University.
§ Uppsala University.
(1) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737.
(2) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mu¨ller, E.;
Liska, P.; Vlachopoulos, N.; Gra¨tzel, M. J. Am. Chem. Soc. 1993, 115,
6382.
(3) Hagfeldt, A.; Gra¨tzel, M. Chem. ReV. 1995, 95, 49.
9
4922
J. AM. CHEM. SOC. 2002, 124, 4922-4932
10.1021/ja0178012 CCC: $22.00 © 2002 American Chemical Society

Modified Phthalocyanines for NIR Sensitization
A R T I C L E S
Chart 1
potential of the dye should be sufficiently more positive than
that of the redox couple in the electrolyte so that the dye can
be regenerated rapidly by electron transfer from the reduced
species in the electrolyte. Finally, the dye should be chemically
stable for long time exposure to natural sunlight. Another,
mainly practical, requirement is that the dye should be soluble
in solvents that are compatible with the semiconductor and
favorable for adsorption of a nonaggregated monolayer of the
dye on the nanoparticle surface. This latter requirement is
important because the dye molecules that are connected directly
to the semiconductor are more effective in injecting electrons
into the semiconductor as compared to those in outer layers.
For an aggregated sensitizer, the excess of dye molecules
absorbs light and acts as a filter. Since it is very difficult to
calculate all the necessary parameters and then synthesize a dye
fulfilling all of them, the method of “trials and failures” is still
an important approach to finding appropriate sensitizing dyes.
To date, ruthenium(II) polypyridyl complexes have proved
to be the best sensitizers for dye-sensitized nanostructured TiO2
solar cells.1,2 However, the main drawback of those sensitizers
is the lack of absorption in the far-red/near-IR region of the
solar spectrum.5 To further improve the performance of these
devices, it is imperative to enhance their response in the above-
mentioned wavelength region. Phthalocyanines possess intensive
absorption in the far-red/near-IR region; are known for their
excellent chemical, light, and thermal stability; and have the
appropriate redox properties for sensitization of large band-gap
semiconductors, e.g., TiO2,6 making them attractive for DSSC.
However, there are some problems to be solved for phthalo-
cyanines to be used in solar cell applications. The typical
phthalocyanines explored for sensitization of large band-gap
semiconductors are free-base or metallic ones, substituted by
carboxylic or sulfonic acid groups for attachment to the
semiconductor surface.7 They are poorly soluble in organic
solvents, for example, ethanol and chloroform, which makes it
difficult to synthesize, separate, and purify these kinds of
phthalocyanines. Another major problem with phthalocyanines
is their strong tendency to aggregate on the semiconductor
surface, which to some extent results in very low IPCE (typically
<4%) of solar cells.7 Gra¨tzel and co-workers8 reported a
strikingly high IPCE of 45% in the near-infrared region for a
sandwich solar cell based on a zinc tetracarboxyl phthalocyanine
(ZnTcPc)-sensitized nanostructured TiO2 electrode when surface
aggregation of the sensitizer was avoided. In contrast to the
typical phthalocyanines with carboxyl groups, phthalocyanines
with ester groups9 are readily synthesized and purified in good
yields, as they are reasonably soluble in moderately polar
solvents, for example, chloroform. However, they could not be
attached to nanostructured TiO2 film by means of ordinary
methods.2 Moreover, they were surprisingly resistant toward
hydrolysis by, for example, NaOH. Recently, we9 reported a
novel method to attach phthalocyanines substituted by ester
groups to nanostructured TiO2 film by pretreatment of the TiO2
film with a Lewis base ((CH3)3COLi). The highest IPCE for a
TiO2 electrode coated by a zinc phthalocyanine substituted by
four ester groups (ZnPcBu) was 4.3% at 690 nm.9 In our present
study, a zinc phthalocyanine substituted with tyrosine groups
(ZnPcTyr; see molecular structure in Chart 1) was modified
for efficient far-red/near-IR sensitization of a nanostructured
TiO2 solar cell electrode. This compound and a reference, zinc
phthalocyanine with glycine substituents (ZnPcGly; see molec-
ular structure in Chart 1), were synthesized and characterized.
The substituted phthalocyanines are more easily synthesized,
separated, and purified in good yields than typical phthalocya-
nines,7,8 since they can be obtained readily by hydrolysis from
their corresponding esters, which are separated to high purity
by means of column chromatography. Because of the incorpora-
tion of tyrosine groups into the macrocycle of zinc phthalocya-
nine, ZnPcTyr has a good solubility in ethanol, which is a typical
solvent for sensitizers when preparing dye-sensitized nanostruc-
tured TiO2 electrodes. Substituting tyrosine groups into phtha-
locyanine reduces considerably surface aggregation as a result
of steric effects. It has been found recently10 that tyrosine plays
an important role in the photosynthetic oxygen-evolving pho-
tosystem II (PSII) reaction center; i.e., after a multistep downhill
photoinduced electron transfer, the oxidized chlorophyll special
pair (P680+) recaptures an electron by oxidizing a nearby tyrosine
to a tyrosyl radical, which then abstracts a hydrogen atom from
a water molecule coordinated to a manganese cluster. Therefore,
another potential function of incorporating redox-active tyrosine
(6) Phthalocyanines: Properties and applications; Leznoff, C. C., Lever, A.
B. P., Eds.; VCH: New York, 1993 (Vol. 1), 1993 (Vol. 2), 1993 (Vol. 3),
1996 (Vol. 4).
(7) (a) Fang, J.; Su, L.; Wu, J.; Shen, Y.; Lu, Z. New J. Chem. 1997, 21,
1303. (b) Shen, Y.; Wang, L.; Lu, Z.; Wei, Y.; Zhou, Q.; Mao, H.; Xu, H.
Thin Solid Films 1995, 257, 144. (c) Fang, J.; Wu, J.; Lu, X.; Shen, Y.;
Lu, Z. Chem. Phys. Lett. 1997, 270, 145. (d) Yanagi, H.; Chen, S.; Lee, P.
A.; Nebesny, K. W.; Armstrong, N. R.; Fujishima, A. J. Phys. Chem. 1996,
100, 5447. (e) Deng, H.; Lu, Z.; Mao, H.; Xu, H. Chem. Phys. 1997, 221,
323.
(8) Nazeeruddin, M. K.; Humphry-Baker, R.; Gra¨tzel, M.; Wo¨hrle, D.;
Schnurpfeil, G.; Schneider, G.; Hirth, A.; Trombach, N. J. Porphyrins
Phthalocyanines 1999, 3, 230.
(9) He, J.; Hagfeldt, A.; Lindquist, S.-E.; Grennberg, H.; Korodi, F.; Sun, L.;
A° kermark, B. Langmuir 2001, 17, 2743.
(10) (a) Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer,
K.; Klein, M. P. Science 1993, 260, 675. (b) Tommos, C.; Tang, X.;
Warncke, K.; Hoganson, C. W.; Styring, S.; McCracken, J.; Diner, B. A.;
Babcock, G. T. J. Am. Chem. Soc. 1995, 117, 10325.
9
J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002 4923

A R T I C L E S
He et al.
4 × CH2), 5.00 (br, 4H, 4 × CH), 6.87 (m, 8H, Ar(Tyr3,5)-H), 7.38
(m, 8H, Ar(Tyr2,6)-H), 8.60 (br, 4H, Ar(Pc)-H), 9.20 (br, 4H, Ar(Pc)-
H), 9.32 (d, 4H, J ) 10.8 Hz, NH), 9.62 (m, 4H, Ar(Pc)-H). MS:
1651.55 ([M + Na]+), 1628.86 ([M]+, calcd for C88H84N12O12Zn:
1628.54).
groups into phthalocyanine is to transfer an electron from
tyrosine to the oxidized ZnPc, thereby preventing back-electron
transfer. The performance of a sandwich solar cell based on
ZnPcTyr-coated nanostructured TiO2 electrodes prepared by
different methods was compared with those based on ZnPcGly-
TiO2 electrodes. The relationship between solar cell performance
and the extent of surface aggregation was investigated. The
effects on solar cell performance of TBP in the electrolyte and
of some additives in the dye solution for preparing dye-sensitized
TiO2 electrode were demonstrated. Furthermore, to explain the
lower efficiency of a solar cell based on phthalocyanine com-
pared to that based on the RuN3 dye, electron injection and
recombination kinetics of ZnPcTyr-sensitized TiO2 thin film
were investigated by femtosecond pump-probe spectroscopy.
The results were compared to the results obtained with RuN3-
sensitized TiO2 thin film.
2,9,16,23-Tetra[1-carbonyl-2-(4-hydroxyphenyl)ethylaminocar-
boxyl]phthalocyaninato Zinc(II) (ZnPcTyr). The mixture of 2a (0.163
g, 0.1 mmol) and 5 M aqueous sodium hydroxide solution (0.2 mL, 1
mmol) was refluxed in 10 mL of acetone for 1 h. After cooling to
room temperature, the precipitate was collected and washed with
acetone. This sodium salt was dissolved in water (10 mL) and then 1
mL of 1 M aqueous hydrochloric acid solution was added to precipitate
the product, which was collected and washed with water and diethyl
ether to get 82 mg (59%) of the title compound. 1H NMR (DMSO-d6)
δ (ppm): 3.39 (br, 8H, 4 × CH2), 4.90 (br, 4H, 4 × CH), 6.82 (m, 8H,
Ar (Tyr3, 5)-H), 7.37 (m, 8H, Ar(Tyr2,6)-H), 8.71 (br, 4H, Ar(Pc)-H),
9.26 (br, 4H, Ar(Pc)-H), 9.50 (br, 4H, NH), 9.94 (br, 4H, Ar(Pc)-H).
MS: 1405.39 ([M]+, calcd for C72H53N12O12Zn: 1405.29).
4-Carbonyl(N-glycine ethyl ester)phthalonitrile (1b). Crude acid
chloride (ca. 10 mmol) was dissolved into acetonitrile (10 mL) and
added dropwise to the mixture of glycine ethyl ester hydrochloride
(1.4 g, 10 mmol), triethylamine (4.04 g, 40 mmol), and 10 mL of
acetonitrile at 0-5 °C while stirring and cooling with ice-water bath
during 30 min. The reaction mixture was stirred at the same temper-
ature for an additional 10 min and then 20 mL of water was added
and the solution extracted with chloroform (20 mL × 3). The or-
ganic phases were dried over sodium carbonate and then the solution
was evaporated. The residue was purified by column chromatography
eluting with a chloroform/methanol (95/5, v/v) solution to give 2.0 g
Experimental Section
1. Synthesis. Materials and Structural Assignments. Starting
materials, except for 3,4-dicyanobenzoyl chloride, were purchased from
Lancaster. Crude acid chloride (3,4-dicyanobenzoyl chloride) was
synthesized by refluxing 3,4-dicyanobenzoic acid11 in thionyl chloride
and chloroform, followed by evaporation and used immediately without
further purification.
Melting points were measured on a Gallenkamp melting point
apparatus and are uncorrected. Thin-layer chromatography (TLC) and
column chromatography were performed by using silica gel adsorbent
(Kieselgel F254 and Kieselgel 60, 0.040-0.063 mm, respectively).
Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker
AM 400 or Varian AS 400 instrument. The splitting of the signal is
described as singlets (s), doublets (d), triplets (t), quartets (q), multiplets
(m), or broad (br). MALDI-TOF MS investigations were carried out
using a BRUKER BIFLEX III instrument in reflector mode, and 3,5-
dimethoxy-4-hydroxy-trans-cinnamic acid (sinapinic acid) was used as
a matrix in all cases.
1
(78%) of the title compound. Mp: 124-126 °C. H NMR (CDCl3) δ
(ppm): 1.31 (t, 3H, J ) 6.8 Hz, CH3), 4.21-4.30 (m, 4H, CH2 +
CH2), 7.04 (br, 1H, NH), 7.92(d, 1H, J ) 8.1 Hz, Ar(6)-H), 8.16 (dd,
1H, J1)8.1 Hz, J2 )1.5 Hz, Ar(5)-H), 8.27 (d, 1H, J ) 1.5 Hz,
Ar(3)-H).
2,9,16,23-Tetra(n-butoxycarbonylmethylaminocarbonyl)-
phthalocyaninato Zinc(II) (2b). The mixture of 1b (0.77 g, 3 mmol),
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, 0.684 g, 4.5 mmol), zinc
acetate dihydrate (0.165 g, 0.75 mmol), and butanol (15 mL) was
refluxed for 1 h. After cooling to room temperature, hexane (45 mL)
was added dropwise, and the reaction mixture was stirred for 2 h. The
precipitate was collected, washed with diethyl ether (3×), and then
dried. The crude product was subjected to column chromatography and
eluted with a chloroform/ethanol (9/1, v/v) solution. The major fraction
was collected, dried, and then washed with ether to afford pure 2b at
a yield of 304 mg (34%). 1H NMR (DMSO-d6) δ (ppm): 1.06 (t, 12H,
J ) 6.6 Hz, 4 × CH3), 1.57 (m, 8H, 4 × CH2), 1.80 (m, 8H, 4 ×
CH2), 4.34 (t, 8H, J ) 6.0 Hz, 4 × CH2), 4.46 (s, 8H, 4 × CH2), 8.66
(m, 4H, Ar-H), 8.98 (m, 4H, Ar-H). 9.48 (m, 4H, NH). 9.64 (m, 4H,
Ar-H). 13C NMR (DMSO-d6) δ (ppm): 14.66, 19.70, 31.34, 42.86,
65.22, 122.26, 122.75, 129.10, 135.14, 137.67, 139.76, 151.89, 152.32,
167.87, 170.99. MS: 1205.45 ([M+H]+, calcd for C60H61N12O12Zn:
1205.38).
4-Carboxy(N-tyrosine ethyl ester)phthalonitrile (1a). Crude acid
chloride (ca. 10 mmol) was dissolved in acetonitrile (10 mL) and added
dropwise to the mixture of tyrosine ethyl ester (2.45 g, 10 mmol),
triethylamine (3.03 g, 40 mmol), and 10 mL of acetonitrile at 0-5 °C
while stirring and cooling with ice-water bath during 20 min. The
reaction mixture was stirred at the same temperature for an additional
10 min, and then 20 mL of water was added and extracted with
dichloromethane (20 mL × 3). The organic phases were dried over
sodium carbonate and then evaporated. The residue was purified by
column chromatography using chloroform/methanol (95/5, v/v) as an
1
eluent. Yield: 1.9 g (52%). Mp: 122-124 °C. H NMR (CDCl3) δ
(ppm): 1.34 (t, 3H, J ) 7.1 Hz, CH3), 3.20 (m, 2H, CH2), 4.27 (q, 2H,
J ) 7.1 Hz, CH2), 5.01 (m, 1H, CH), 6.72 (d, 2H, J ) 8.4 Hz, Ar
(Tyr-3, 5)-H), 6.95 (d, 2H, J ) 8.4 Hz, Ar(Tyr-2,6)-H), 7.88 (d, 1H, J
) 8.1 Hz, Ar(nitrile-6)-H), 8.05 (dd, 1H, J1)8.1 Hz, J2)1.5 Hz, Ar-
(nitrile-5)-H), 8.16 (d, 1H, J ) 1.5 Hz, Ar(nitrile-3)-H).
2,9,16,23-Tetra[1-(n-butoxycarbonyl)-2-(4-hydroxyphenyl)eth-
ylaminocarbonyl]phthalocyaninato Zinc(II) (2a). The mixture of 1a
(0.726 g, 2 mmol), DBU (0.456 g, 3 mmol), zinc(II) acetate dihydrate
(0.11 g, 0.5 mmol), and butanol (10 mL) was refluxed for 1 h. After
cooling to room temperature, 10 mL of diethyl ether was added and
the mixture stirred for an additional 2 h. The precipitate was collected
and washed by diethyl ether (3×). The crude product was purified by
column chromatography eluted with a chloroform/ethanol (9/1, v/v)
solution to get 292 mg (36%) of the title compound. 1H NMR (DMSO-
d6) δ (ppm): 1.00 (t, 12H, J ) 6.7 Hz, 4 × CH3), 1.45 (m, 8H, 4 ×
CH2), 1.70 (m, 8H, 4 × CH2), 3.36 (br, 8H, 4 × CH2), 4.25 (m, 8H,
2,9,16,23-Tetra(carbonylmethylaminocarboxyl)phthalo-
cyaninato zinc(II) (ZnPcGly). The mixture of 2b (120 mg, 0.1 mmol)
and a 5 M aqueous sodium hydroxide solution (0.2 mL, 1 mmol) was
refluxed in 10 mL of acetone for 1 h. After cooling to room tempera-
ture the precipitate was collected and washed with acetone. This
sodium salt was dissolved in water (10 mL) and then 1 mL of 1 M
aqueous hydrochloric acid solution was added to precipitate the
product. The precipitate was collected and washed with water and
1
diethyl ether. Yield: 87 mg (89%). H NMR (DMSO-d6) δ (ppm):
4.29 (s, 8H, 4 × CH2), 8.74 (br, 4H, Ar-H), 9.38 (br, 4H, Ar-H),
9.64 (br, 4H, 4 × NH), 9.87 (br, 4H, Ar-H). 13C NMR (DMSO-d6) δ
(ppm): 42.69, 122.23, 123.00, 129.35, 135.53, 138.56, 140.63, 153.31,
154.24, 167.69, 172.28. MS: 981.21 ([M + H]+, calcd for C44H29N12O12-
Zn: 981.13).
(11) George, R.; Snow, A. W. J. Heterocycl. Chem. 1995, 32, 495.
9
4924 J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002

Modified Phthalocyanines for NIR Sensitization
A R T I C L E S
2. Electrochemical Measurements. The electrochemical experi-
ments were performed at room temperature. A conventional three-
electrode cell with separate compartments for counter and reference
electrodes was employed in these experiments. A 3-mm glassy carbon
disk acted as the working electrode, a nonaqueous Ag/Ag+ electrode
(10 mM AgNO3 in acetonitrile) as the reference at a potential of 0.30
V vs SCE,12 and a platinum wire as the counter electrode. The glassy
carbon electrode was polished with Al2O3 paste and cleaned thoroughly
in an ultrasonic water bath before each measurement. The electrolyte
was 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6) in dry
acetonitrile/dimethyl sulfoxide (DMSO) (7/1, v/v). The concentration
of the dyes in the electrolyte was ∼0.5 mM. The potentiostat was an
Autolab µII interfaced to a personal computer for data acquisition and
instrument control.
sandwich-type cells. The working electrode of the dye-sensitized TiO2
film on conducting glass was squeezed together with a platinized
conducting glass using a spring and illuminated from the substrate side.
The electrolyte, typically 0.5 M LiI/0.05 M I2 in propylene carbonate,
was attracted into the cavities of the dye-sensitized TiO2 electrode by
capillary forces. A 450 W xenon lamp with a monochromator was used
as light source for the photocurrent action spectra measurements. The
cell was operated in the short-circuit mode. The IPCE values were
determined at 10 nm intervals between 400 and 800 nm. The IPCE
was then calculated according to the following equation:
1240iph[µA]
IPCE )
(1)
P[µW]λ[nm]
where iph and P are the photocurrent and power of the incident radiation
per unit area and λ is the wavelength of the monochromatic light. No
corrections were made for absorption and reflection in the substrate.
A 1000 W xenon lamp sun-simulator with a 10 cm water filter was
used as a light source for the photocurrent-photovoltage characteristics.
The light intensity was measured by a pyranometer (Kipp & Zonen
CM 11). The active electrode area was typically 0.25 cm2. The fill
factor is defined by
3. Film Preparations and Dye-Coating. Colloidal TiO2 paste
(Ti-Nanoxide T) was purchased from Solaronix SA. According to the
supplier’s specification, the particle size of TiO2 is ca. 13 nm and the
paste is well-suited for preparation of transparent photoelectrodes for
dye-sensitized solar cells. For obtaining a porous film of uniform
thickness the following procedure was used. The colloidal TiO2
suspension was spread onto transparent conducting glass sheets (Libbey
Owens Ford, fluorine-doped SnO2 glass, sheet resistance 8 Ω/0) using
Scotch tape as a spacer. A thin film was obtained by raking off the
excess of suspension with a glass rod. After removing the tape and
air-drying, the sample was sintered in air at 450 °C for 30 min to form
a transparent TiO2 film electrode. This procedure was repeated two
more times to obtain a thicker film. The thickness of the film, recorded
by Dektak3 Surface Profile Measuring System, was about 7.6 µm. For
femtosecond pump-probe measurements a ∼2 µm thick nanostructured
TiO2 film was prepared on a ∼60-80 µm microscope cover slip,
according to the above-mentioned process.
J
Ph(max)UPh(max)
FF )
(2)
JSCUOC
where JPh(max) and UPh(max) are the photocurrent and photovoltage for
maximum power output and JSC and UOC are the short-circuit photo-
current and open-circuit photovoltage. The overall conversion efficiency
(η) is defined by the following equation:
J
SCUOCFF
η )
(3)
Pin
Dye-coating of the TiO2 film was carried out by soaking the film in
the dye solution immediately after the high-temperature annealing and
while it was still warm (∼80 °C). ZnPcGly- and ZnPcTyr-sensitized
TiO2 film electrodes (denoted ZnPcGly-TiO2 and ZnPcTyr-TiO2)
were prepared by dipping the TiO2 films into a 5.0 × 10-5 M ZnPcGly-
or ZnPcTyr-ethanol solution (containing 3% (v/v) DMSO for ZnPcGly)
for 10 h, followed by rinsing with ethanol several times. To decrease
dye aggregation on the semiconductor surface, 10 mmol cheno and
7% (v/v) TBP were added to the above dye solutions and the dye-
sensitized TiO2 film electrodes were obtained by dipping the TiO2 films
into the dye solutions with additives for 15 h and rinsing with ethanol
immediately after withdrawal from the dye solutions. Sensitized
electrodes prepared from dye solutions with additives are denoted
(ZnPcGly-TiO2)add and (ZnPcTyr-TiO2)add. Dye-coating of the TiO2
film for pump-probe measurements was performed by immersing the
film into the above ZnPcTyr-ethanol solution with additives for 10 h.
The film was rinsed with ethanol, dried at room temperature, covered
by acetonitrile and another microscope slip, and finally sealed. Steady-
state absorption spectra of the dye-sensitized TiO2 films in acetonitrile
were recorded before and after laser experiments to ensure that no
degradation or modification of the samples had occurred in sample
preparation or during measurements. The optical density of the sample
was ∼0.65 at the excitation wavelength of 700 nm. All the dye-
sensitized electrodes were stored in ethanol and kept in the dark before
all measurements.
Here Pin is the power of incident white light.
5. Femtosecond Spectrometer. The femtosecond spectrometer used
here was based on a regeneratively amplified mode-locked Ti:sapphire
laser system working at 5 kHz repetition rate in combination with an
optical parametic amplifier (OPA).13 A detailed description of this setup
has been presented previously.14 Excitation pulses with a typical energy
of ∼0.2 µJ at 700 nm was focused to a spot of ∼400 µm diameter at
the sample position. The excitation pulse intensity of ∼1014 photon/
cm2 used was far from the saturation limit considering that the optical
cross section of the dye is ∼1.5 × 10-16 cm-2. The probe and reference
pulses were taken from a white-light continuum generated by a part of
the amplified fundamental 800 nm laser beam, which was sent into a
sapphire plate. In the region from 1200 to 2000 nm, another OPA was
used to generate the probe light. The polarization of pump and probe
beams was kept at the magic angle configuration in all experiments.
For data acquisition, a monochromator and three photodiodes in a single
shot detection system were used. Transient absorption kinetics at a
definite wavelength was recorded by keeping the detection wavelength
fixed and scanning the pump-probe delay time. The instrument
response function of ∼140-150 fs (fwhm), depending slightly on the
probe wavelength, was determined by sum frequency cross correlation
in a BBO crystal. The accurate zero-time delay at all measured
wavelengths was independently determined by sum frequency cross-
correlation measurements and the nonresonant “spike signal” in a 1
mm thick glass plate. Measured kinetics were analyzed with the
deconvolution software Spectra Solve 2.01, LASTEK Pty. Ltd. 1997.
All experiments were conducted at room temperature.
4. Optical and Photoelectric Measurements. Absorption spectra
of the dye-sensitized TiO2 electrodes were recorded on a JASCO V530
UV/vis spectrophotometer with a bare TiO2 film as a reference. These
measurements were carried out for dry film electrodes, and no
corrections were made for optical effects due to the presence of the
electrolyte. Photocurrent action spectra and photocurrent-photovoltage
characteristics of the dye-sensitized TiO2 electrodes were measured with
Results and Discussion
1. Synthesis. The desired phthalocyanine dyes, ZnPcTyr and
ZnPcGly, were synthesized as outlined in Scheme 1, by using
(13) Chachisvilis, M.; Ku¨hn, O.; Pullerits, T.; Sundstro¨m, V. J. Phys. Chem. B
1997, 101, 7275.
(14) Benko¨, G.; Hilgendorff, M.; Yartsev, A. P.; Sundstro¨m, V. J. Phys. Chem.
B 2001, 105, 967.
(12) Handbook of Analytical Chemistry; Meites L., Ed.; Mcgrow Hill, New York,
1963.
9
J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002 4925

A R T I C L E S
He et al.
Scheme 1
DBU as catalyst followed by saponification. The purpose of
introducing amino acids as substituents is to provide carboxylic
acids as attaching groups to the TiO2 surface. Both ZnPcTyr
and ZnPcGly were obtained in quite good yield. During the
formation of the macrocycles, butoxy groups replaced ethoxy
groups by transesterification with the solvent. The butyl esters
(2a and 2b) could be purified by means of column chromatog-
raphy and eluted with a chloroform/ethanol solution. Further-
more, they were hydrolyzed successfully in both cases.
obtained by DPV, owing to smaller charging effects.16 Relative
voltammetric peak heights and DPV peak areas indicate that
oxidation III of ZnPcTyr (Figure 1B) is a one-electron process,
while the third oxidation of ZnPcGly (Figure 1A) seems to
involve more than one electron. The reversible peaks I and II
of ZnPcGly (Figure 1A) can be assigned to oxidations generating
the π-cation radical and the dication.17 The oxidation potentials
are 0.17 V vs Ag/AgNO3 (0.47 V vs SCE) and 0.36 V vs Ag/
AgNO3 (0.66 V vs SCE), respectively. The additional peaks
(III and IV) observed at 0.72 and 0.96 V vs Ag/AgNO3 or 1.02
and 1.26 V vs SCE, respectively, have not been reported before
for zinc phthalocyanines and are not observed for ZnPcTyr
(Figure 1B). From our present data, no assignment can be made
for these peaks. In the case of ZnPcTyr (Figure 1B), three
oxidations at 0.25, 0.40, and 0.55 V vs Ag/AgNO3 or 0.55, 0.70,
and 0.85 V vs SCE are observed. By comparison with the
oxidation potentials of ZnPcGly, the first and second oxidation
peak can be assigned to the oxidations of the marocycle of
ZnPcTyr. The oxidation potentials of ZnPcTyr (0.55 and 0.70
V vs SCE) are more positive than those of ZnPcGly (0.47 and
0.66 V vs SCE), owing to the electron-withdrawing effect of
the tyrosine group.18 The third oxidation peak of ZnPcTyr is
assigned to the oxidation of one of the tyrosines on the basis of
a reported oxidation potential of 0.84 V vs SCE19 for tyrosine
in acetonitrile, which is in good agreement with the value of
0.85 V vs SCE observed with ZnPcTyr. It seems unlikely that
the third peak of ZnPcTyr is due to the same oxidation process
Phthalocyanines were characterized by NMR and mass
spectrometry. MALDI-mass showed satisfactory results in all
1
cases. H NMR also agreed with the desired structure in all
cases. Almost all the aromatic peaks became multiplets and/or
broad after the formation of phthalocyanine rings, due to the
existence of statistical isomers and possible aggregation. These
observations supported the achievement of the designed struc-
tures.
2. Electrochemical Studies. Both ZnPcTyr and ZnPcGly
were found to be soluble in an acetonitrile/DMSO (7/1, v/v)
mixed solvent to a maximum concentration of ∼0.5-1 mM.
The oxidation potentials were determined from half-wave
potentials (E1/2 ) (Eoxid + Ered)/2) by cyclic voltammetry (CV)
or peak potentials (Ep) by differential pulse voltammetry (DPV).
Voltammograms were recorded at a glassy carbon electrode in
the acetonitrile/DMSO solvent containing 0.1 M tetrabutylam-
monium hexafluorophosphate (TBAPF6) that is redox passive
in our scanning region (dashed lines in Figure 1). The cyclic
voltammograms (CVs, dotted lines) and differential pulse
voltammograms (DPVs, solid lines) of ZnPcGly and ZnPcTyr
are shown in parts A and B of Figure 1, respectively. The cyclic
voltammograms are not well-resolved and are partly distorted,
presumably due to aggregation of the poorly soluble phthalo-
cyanines. The separation between anodic and cathodic CV peaks
is difficult to determine accurately, but the oxidations I and II
appear to be reversible one-electron processes for both ZnPcGly
and ZnPcTyr.15 Reasonably well resolved voltammograms were
(15) Kissinger, P. T.; Heineman, W. R. J. Chem. Educ. 1983, 60, 702.
(16) Bard, A. L.; Faulkner, L. R. Electrochemical Methods: Fundamentals and
Applications; Wiley: New York, 1980.
(17) (a) Nyokong, T.; Gasyna, Z.; Stillman, M. J. Inorg. Chem. 1987, 26, 548.
(b) Manivannan, V.; Nevin, W. A.; Leznoff, C. C.; Lever, A. B. P. J. Coord.
Chem. 1988, 19, 139.
(18) Zhou, Y.; Wang, Z.; Wang, D.; Sugiura, K.; Sakata, Y. J. Porphyrins
Phthalocyanines 2001, 5, 523.
(19) Sun, L.; Burkitt, M.; Tamm, M.; Ramond, M. K.; Abrahamsson, M.;
LeGourrie´rec, D.; Frapart, Y.; Magnuson, A.; Kene´z, P. H.; Brandt, P.;
Tran, A.; Hammarstro¨m, L.; Styring S.; A° kermark, B. J. Am. Chem. Soc.
1999, 121, 6834.
9
4926 J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002

Modified Phthalocyanines for NIR Sensitization
A R T I C L E S
Figure 2. Absorption spectra of ZnPcTyr in ethanol at different concentra-
tions. (1) 1.0 × 10-6 M; (2) 1.0 × 10-5 M, and (3) 2.0 × 10-5 M.
As mentioned in the Introduction, incorporating the redox-
active tyrosine groups into phthalocyanine has the potential
function of blocking back-electron transfer from the conduction
band of TiO2 to the oxidized sensitizer by electron transfer from
tyrosine to the oxidized ZnPc. However, since the oxidation
potential of tyrosine is more positive than that of the zinc
phthalocyanine macrocycle, it is evident that this process is
thermodynamically impossible.
3. Steady-State Spectroscopies. ZnPcGly is insoluble in pure
ethanol. However, it is soluble in ethanol with the aid of DMSO
(3%, v/v). In contrast, ZnPcTyr has an appreciable solubility
in ethanol, suggesting that the tyrosine groups incorporated into
the zinc phthalocyanine macrocycle are crucial for obtaining
good solubility in ethanol. Figure 2 shows the absorption spectra
of ZnPcTyr in ethanol at different concentrations. In dilute
solution (curve 1) ZnPcTyr is present mainly as monomers,
characterized by the sharp absorption bands in the Soret (350
nm) and in the Q-band region (680 nm).20,21 On increasing the
dye concentration (curve 2), a new absorption band at around
635 nm appears and becomes stronger compared to the 680-
nm band (curve 3) with further increase of the dye concentration.
This blue-shifted band with a maximum at 635 nm can be
assigned as the characteristic Q-band of the ZnPcTyr face-to-
Figure 1. Differential pulse voltammograms (solid lines) and cyclic
voltammograms (dotted lines) of ∼0.5 mM ZnPcGly (A) and ZnPcTyr (B).
Dry acetonitrile/DMSO (7/1, v/v) containing 0.1 M TBAPF6 was used as
supporting electrolyte. Dashed lines are the DPVs of 0.1M TBAPF6 in dry
acetonitrile/DMSO (7/1, v/v). I, II, III and IV represent the first, second,
third, and fourth oxidations, respectively. Scans are initiated at -0.05 or 0
V versus Ag/AgNO3 in positive direction at the rates of 50 mV/s. In the
case of DPV, the modulation time is 50 ms, interval time 100 ms, step
potential 5 mV, and modulation amplitude 25 mV. The solutions were
purged with argon and stirred for over 15 min prior to the measurements.
The solutions were kept under argon during the measurements.
that gives rise to the third peak of ZnPcGly, since the latter
occurs at substantially higher potential (1.02 V vs SCE) and
involves more than one electron. That the further oxidations of
ZnPcGly (Figure 1A, peaks III and IV, 1.02 and 1.26 V vs SCE)
are not observed with ZnPcTyr is probably due to the reactivity
of the tyrosine radical formed from ZnPcTyr at 0.85 V vs SCE
(Figure 1B, peak III).
With respect to dye-sensitization of large band-gap semicon-
ductors, e.g. TiO2, the first oxidation process is of interest. From
the first oxidation potentials of ZnPcGly and ZnPcTyr and their
0-0 transition energy (E0-0) of 1.82 eV assessed from the
respective absorption and emission spectral data (not shown),
the energy levels of the singlet excited states (LUMO) of
ZnPcGly and ZnPcTyr were determined to be -1.35 and -1.27
V vs SCE, respectively, whereas the energy level of the
conduction band edge of TiO2 is ca. -0.74 V vs SCE.3 This
makes electron injection from the excited state of ZnPcGly or
ZnPcTyr into the conduction band of TiO2 thermody-
namically possible. Furthermore, the HOMO levels of the dyes
face dimer or higher order aggregate (H-aggregate).21,22
A
similar behavior was also found for ZnPcGly in ethanol
containing 3% (v/v) DMSO (not shown).
Absorption spectra of the dye-sensitized TiO2 electrodes and
a bare TiO2 film are shown in Figure 3. Curves 1 and 2 are the
absorption spectra of ZnPcTyr-TiO2 and ZnPcGly-TiO2
electrodes, respectively, prepared from their corresponding
dye-ethanol solutions under identical conditions except that
3% (v/v) DMSO was added to the solution of ZnPcGly. Curves
3 and 4 are the absorption spectra of (ZnPcTyr-TiO2)add and
(ZnPcGly-TiO2)add electrodes prepared under identical condi-
tions (except for 3% (v/v) DMSO was added to in the ZnPcGly
solution) from their corresponding dye-ethanol solutions with
additives (10 mmol cheno and 7% (v/v) TBP). Curve 5 is the
absorption spectrum of a bare TiO2 film. It is interesting to see
that adsorption of dye onto the TiO2 film makes the film more
(20) Phthalocyanine Materials, Mckeown, N. B., Ed.; Cambridge University
Press: Cambridge, 1998.
+
+
(EZnPcGly/ZnPcGly and EZnPcTyr/ZnPcTyr ) are lower than the energy
level of the redox couple I-/I3 (0.2 V vs SCE3) in the
-
(21) Schlettwein, D.; Oekermann, T.; Yoshida, T.; Tochimoto, M.; Minoura,
H. J. Electroanal. Chem. 2000, 481, 42.
electrolyte, enabling the dye regeneration by electron transfer
from I-.
(22) (a) Nevin, W. A.; Liu, W.; Lever, A. B. P. Can. J. Chem. 1987, 65, 855.
(b) Snow, A. W.; Jarvis, N. L. J. Am. Chem. Soc. 1984, 106, 4706.
9
J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002 4927

A R T I C L E S
He et al.
Figure 3. Absorption spectra of the dye-sensitized nanostructured TiO2
film electrodes (1-4) referenced by a bare TiO2 film and of a TiO2 film
(5) with conducting glass as a reference. (1) and (3) are the spectra of
ZnPcTyr-TiO2 and (ZnPcTyr-TiO2)add electrodes prepared from ZnPcTyr
solutions without and with additives. (2) and (4) are the spectra of ZnPcGly-
TiO2 and (ZnPcGly-TiO2)add electrodes prepared from ZnPcGly solutions
without and with additives.
Table 1. Optical Data of the Dye-Sensitized TiO2 Electrodes and
Photovoltaic Data of the Sandwich Solar Cells Based on These
Electrodes
IPCE/%
A /Am (690 nm) (690 nm)
φ/%
JSC/mA
cm
2
electrodes
UOC/V FF/% η/%
a
ZnPcTyr-TiO2
ZnPcGly-TiO2
0.80
1.01
6.9
4.0
24.2
5.5
6.9
4.1
24.3
5.7
1.27 0.36 64 0.29
0.36 0.31 65 0.07
2.25 0.36 67 0.54
0.61 0.32 68 0.13
(ZnPcTyr-TiO2)add 0.59
(ZnPcGly-TiO2)add 0.83
Figure 4. Photocurrent action spectra of sandwich solar cells based on the
same ZnPcTyr-TiO2 (1) and ZnPcGly-TiO2 (2) electrodes (A), and
(ZnPcTyr-TiO2)add (1) and (ZnPcGly-TiO2)add (2) electrodes (B) as those
in Figure 3.
transparent in the 400-500 nm region, indicating that the dye-
sensitized TiO2 electrodes have less light scattering than the
bare electrodes in this spectral region. Each spectrum of the
dye-sensitized TiO2 electrode shows two Q-bands, one near the
position of the monomer Q-band and the other near the position
of the Q-band of the aggregate of the dye in solution (Figure
2). Each system is therefore characterized by adsorption to the
TiO2 surface of a mixture between monomers and aggregates.
The two Q-bands of ZnPcTyr and the Q-band of the ZnPcGly
monomer red-shift by ∼5 nm compared to the dyes in ethanol
solution (Figure 2), indicating a strong interaction between the
dyes and the semiconductor surface.23 The Q-band of the
ZnPcGly aggregate blue-shifts by ∼2 nm, which may be induced
by a heavier surface aggregation of ZnPcGly compared to that
of ZnPcTyr. To estimate the extent of surface aggregation of
the above four dye-sensitized TiO2 electrodes, the ratio of the
absorbance at the aggregate band maximum (633 nm for
ZnPcGly and 640 nm for ZnPcTyr) (Aa) and that of monomer
(685 nm) (Am) of each system was calculated (see Table 1).
We can see that the Aa/Am of the ZnPcTyr-TiO2 electrode (0.80)
is lower than that of the ZnPcGly-TiO2 electrode (1.01),
implying that the extent of surface aggregation for ZnPcTyr is
less than that for ZnPcGly. It is conceivable that the presence
of tyrosine groups in ZnPcTyr reduces the formation of dye
H-aggregates on the TiO2 surface because of steric effects. It
was found8 that addition of some chemicals, such as cheno and
TBP, to the zinc phthalocyanine solution from which the dye-
sensitized TiO2 electrode was prepared significantly diminishes
the surface aggregation of the sensitizer. The cheno molecules
are expected to coadsorb with the dye molecules to avoid to
some extent the dye aggregation on the semiconductor surface,
while the TBP will coordinate to the axial position of the zinc
phthalocyanine, preventing stacking of the dye molecules. It is
seen from Figure 3 and Table 1 that the electrodes (ZnPcTyr-
TiO2)add and (ZnPcGly-TiO2)add prepared from their corre-
sponding dye solutions with these additives have less extent of
surface aggregation than ZnPcTyr-TiO2 and ZnPcGly-TiO2,
respectively. By addition of these additives, the Aa/Am of the
ZnPcTyr-TiO2 electrode is decreased by 26% (from 0.80 to
0.59), while the Aa/Am of ZnPcGly-TiO2 electrode is reduced
by 18% (from 1.01 to 0.83). The larger Aa/Am decrease for the
ZnPcTyr-TiO2 electrode as compared to the ZnPcGly-TiO2
electrode can also be explained by steric effects of tyrosine
groups in ZnPcTyr.
4. Photovoltaic Studies. The four dye-sensitized TiO2
electrodes were employed as working electrodes in sandwich
solar cells, and the performance of the cells was characterized
by spectral photocurrent responses and photocurrent-photo-
voltage characteristics. Figure 4a shows the photocurrent action
spectra of solar cells with ZnPcTyr-TiO2 and ZnPcGly-TiO2
as working electrodes, while the photocurrent action spectra of
the cells based on (ZnPcTyr-TiO2)add and (ZnPcGly-TiO2)add
working electrodes are shown in Figure 4b. The photocurrent
action spectra resemble the absorption spectra (Figure 3) except
(23) Kamat, P. V. Chem. ReV. 1993, 93, 267.
9
4928 J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002

Modified Phthalocyanines for NIR Sensitization
A R T I C L E S
sensitized TiO2 electrodes prepared from their corresponding
dye solutions with additives, it is seen that the IPCE, φ, and η
values for the (ZnPcTyr-TiO2)add electrode are also much higher
than those of the (ZnPcGly-TiO2)add electrode. Furthermore,
the (ZnPcTyr-TiO2)add and (ZnPcGly-TiO2)add electrodes have
higher IPCE, φ, and η values as compared to the corresponding
electrodes prepared without additives to the dye solutions. By
comparison to the absorption spectra of the dye-sensitized TiO2
electrodes (Figure 3), these results indicate that surface ag-
gregation of the phthalocyanines plays an important role in the
performance of the solar cells. Reducing surface aggregation
leads to a considerable increase in solar cell performance, which
can be explained as suppression of quenching processes due to
energy transfer25 or charge-transfer reactions26 between ag-
gregated molecules and/or between molecules in the aggregates
and monomers. Due to the steric effects of the tyrosine groups
in ZnPcTyr, this sensitizer has a lower degree of aggregation
on the surface compared to ZnPcGly (Figure 3 and Table 1),
which leads to a much better performance of solar cells based
on ZnPcTyr-sensitized TiO2 electrodes. Addition of cheno and
TBP to the dye solutions decreases the surface aggregation and
enhances significantly the solar cell performance, especially for
ZnPcTyr. The (ZnPcTyr-TiO2)add electrode has the highest
IPCE of 24.2% at 690 nm, corresponding to a φ of 24.3%, JSC
of 2.25 mA/cm2, UOC of 360 mV, and FF of 67%, giving a η
of 0.54%. This is one of the highest efficiencies reported for a
photovoltaic device based on phthalocyanine.
5. Effect of TBP in the Electrolyte. It has been demonstrated
previously that pretreatment of dye-sensitized nanostructured
TiO2 electrodes with TBP2 or adding TBP in the electrolyte27
enhances dramatically the open-circuit photovoltage without
apparent loss in short-circuit photocurrent. The general explana-
tion is that the adsorption of TBP on the free area of the TiO2
electrode exposed to redox electrolyte suppresses the dark
current, arising from the triiodide reduction by conduction band
electrons at the semiconductor electrolyte junction (eq 5).
Figure 5. Photocurrent-photovoltage characteristics of sandwich solar cells
based on the same ZnPcTyr-TiO2 (1) and ZnPcGly-TiO2 (2) electrodes
(A), and (ZnPcTyr-TiO2)add (1) and (ZnPcGly-TiO2)add (2) electrodes (B)
as those in Figure 3. The light intensity was 100 mW/cm2.
for a slight red-shift by ca. 5 nm. The highest IPCE values at
690 nm for each system are summarized in Table 1. Taking
into consideration the different light-harvesting efficiencies of
the various systems, the quantum efficiency (φ), defined as
electrons measured in the external circuit per absorbed photon
at a definite wavelength as in the following equation,24 is
determined to provide a appropriate comparison.
-
ket[I3
2ecb-(TiO2) + I3
]8 3I-
(5)
-
For regenerative photoelectrochemical systems, eq 6 holds:2
IPCE
1 - R - T
Iinj
kT
φ )
(4)
UOC
)
ln
(6)
-
( ) (
)
e
ncbket[I3 ]
In eq 4, R and T are the reflectance and transmission of the
electrodes, and R can be considered to be 0 because the dye-
sensitized TiO2 electrodes scatter little light (see Figure 3 and
the text in section 3). The quantum efficiencies obtained at 690
nm are listed in Table 1. It should be noted that the quantum
efficiency is very close to the IPCE value, because the
absorptance (1 - R - T) is close to 1 for all electrodes. Figure
5 shows the photocurrent-photovoltage characteristics of
sandwich solar cells based on the four dye-sensitized TiO2
electrodes illuminated by a sun simulator. The short-circuit
photocurrent (JSC), open-circuit photovoltage (UOC), fill factor
(FF), and overall conversion efficiency (η) for each system are
also summarized in Table 1. We can see that all the IPCE, φ,
and η for the ZnPcTyr-TiO2 electrode are higher than those
of the ZnPcGly-TiO2 electrode. When comparing the dye-
where k and T are the Boltzmann constant and absolute
temperature, Iinj is the flux of charge resulting from electron
injection from the sensitizing dye, ncb is the concentration of
electrons at the TiO2 surface, and ket is the rate constant for
triiodide reduction by conduction band electrons. A decrease
of the rate constant for triiodide reduction should result in an
increase in the open-circuit photovoltage. More recently, Hag-
(25) (a) Schneider, G.; Wo¨hrle, D.; Spiller, W.; Stark, J.; Schulz-Ekloff, G.
Photochem. Photobiol. 1994, 60, 333. (b) Spiller, W.; Kliesch, H.; Wo¨hrle,
D.; Hackbarth, S.; Roeder, B. J. Porphyrins. Phthalocyanines 1998, 2, 145.
(26) (a) Klessinger, M.; Michl, J. Excited States and Photochemistry of Organic
Molecules; VCH: New York, 1995; (b) Rodgers, M. A. J. Photosensiti-
zation; Moreno, G., Pottier, R. H., Truscott, T. G., Eds.; Springer: Berlin,
1988.
(27) (a) Go´mez, M. M.; Lu, J.; Solis, J. L.; Olsson, E.; Hagfeldt, A.; Granqvist,
C. G. J. Phys. Chem. B 2000, 104, 8712. (b) Park, N.-G.; van de Lagemaat,
J.; Frank, A. J. J. Phys. Chem. B 2000, 104, 8989. (c) Sayama, K.; Hara,
K.; Mori, N.; Satsuki, M.; Suga, S.; Tsukagoshi, S.; Abe, Y.; Sugihara,
H.; Arakawa, H. Chem. Commun. 2000, 1173.
(24) O’Regan, B.; Schwartz, D. T. Chem. Mater. 1995, 7, 1349.
9
J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002 4929

A R T I C L E S
He et al.
and IPCE with addition of TBP to the electrolyte is not known
at present.
6. Ultrafast Spectroscopic Measurements. As demonstrated
above, the best solar cell presented here is based on a
(ZnPcTyr-TiO2)add electrode. It has a quantum efficiency of
∼24% at 690 nm and an overall conversion efficiency of 0.54%.
Aiming for a good performance, one can compare the (ZnPc-
Tyr-TiO2)add and RuN3-based cells, since the solar cell based
on RuN3-sensitized TiO2 shows a quantum efficiency of ∼100%
and an overall conversion efficiency of ∼10%.4a,d It is well-
known that besides other important properties, the several orders
of magnitude difference between the time constants of electron
injection (ranging from ∼50 fs to tens of picoseconds29) and
recombination (on the nanosecond to millisecond time scales30),
make the RuN3-TiO2 system one of the most efficient light-
to-energy converters available for DSSC. The apparent differ-
ence between the two DSSCs naturally raises the question
whether the relatively low performance of (ZnPcTyr-TiO2)add
DSSC may be caused by either slow electron injection or/and
fast charge recombination. It is important to point out that in
DSSC the overall conversion yield is the result of a kinetic
competition between reactions of the oxidized dye molecules
with conduction band electrons and redox species. It is crucial
that the recombination with conduction band electrons should
be slower than interception of the dye cation by a redox
mediator. In RuN3-based solar cells, this later process was
shown to take place within about 10 ns,31 although there are
systems where part of the oxidized dye reduction occurs already
on the tens of picoseconds time scale.14
Figure 6. Photocurrent action spectra of sandwich solar cells based on a
(ZnPcTyr-TiO2)add film electrode using 0.5 M LiI/0.05 M I2 in propylene
carbonate (1) and 0.5 M LiI/0.05 M I2/0.2 M TBP in propylene carbonate
(2) as electrolyte.
For clarifying the issue of whether electron injection or
recombination controls the overall conversion yield in our solar
cell, the kinetics of a ZnPcTyr-sensitized TiO2 thin film was
studied by femtosecond pump-probe spectroscopy, and the
results were compared with those obtained for a RuN3-coated
TiO2 thin film. Here, only a limited number of kinetics
monitoring the interfacial electron-transfer reactions in our
sample is presented. The detailed femtosecond transient absorp-
tion study focusing on TiO2 film sensitized by ZnPcTyr and
containing acetonitrile as solvent will be published elsewhere.32
In transient absorption spectroscopy, the electron injection
process can be monitored by the formation of the electron-
transfer products, the injected electrons in the conduction band
Figure 7. Photocurrent-photovoltage characteristics of sandwich solar cells
based on the same (ZnPcTyr-TiO2)add film electrode as that in Figure 6
using 0.5 M LiI/0.05 M I2 in propylene carbonate (1) and 0.5 M LiI/0.05
M I2/0.2 M TBP in propylene carbonate (2) as electrolyte.
feldt and co-workers28 demonstrated another function of TBP
added in the electrolyte in a RuN3-sensitized nanostructured
TiO2 solar cell: suppression of the loss of thiocyanato (SCN-)
groups from the dye so as to improve the stability of the cell.
To further improve the performance of the solar cells based on
of the semiconductor (e-TiO ), or the oxidized dye molecule
2
(ZnPcTyr+). The decay of the electron-transfer products and
the recovery of the ground-state bleach (GSR) are indicative of
the charge recombination reaction. For the ZnPcTyr-sensitized
TiO2 thin film, we have chosen an excitation wavelength of
700 nm to ensure a selective excitation of the monomers (see
curve 3 in Figure 3). Figure 8 shows the transient absorption
kinetics recorded at 1800 nm. On the basis of the extinction
coefficient of conduction band electrons in the near-IR,33 the
rise of the weak induced absorption is assigned to the arrival
-
our phthalocyanine dyes, TBP was added to the I-/I3 redox
electrolyte. Figure 6 shows the photocurrent action spectra of
the sandwich solar cells based on a ZnPcTyr-TiO2 electrode
using 0.5 M LiI/0.05 M I2 (curve 1) or 0.5 M LiI/0.05 M I2/0.2
M TBP (curve 2) in propylene carbonate as the electrolyte. The
photocurrent-photovoltage characteristics of the same cells are
shown in Figure 7. We can see that the open-circuit photovoltage
increases by ∼20%, but the short-circuit photocurrent decreases
by ∼50% when adding TBP to the electrolyte; overall, this leads
to a decrease in the conversion efficiency from 0.54% to 0.31%.
The IPCE values also decrease, for example by ∼35% at 690
nm. Similar behavior was also found for the cells based on the
other three phthalocyanine-sensitized TiO2 electrodes (not
shown). The origin of the decreased short-circuit photocurrent
(29) Benko¨, G.; Kallioinen, J.; Korppi-Tommola, J. E.; Yartsev, A. P.;
Sundstro¨m, V. J. Am. Chem. Soc. 2002, 124, 489. and references therein.
(30) Haque, S. A.; Tachibana, Y.; Willis, R. L.; Moser, J. E.; Gra¨tzel, M.; Klug,
D. R.; Durrant, J. R. J. Phys. Chem. B 2000, 104, 538.
(31) Gra¨tzel, M.; Moser, J.-E. Solar Energy Conversion. In Electron Transfer
in Chemistry; Balzani, V., Gould, I., Eds.; Wiley-VCH: Weinheim, 2001;
Vol. V, p 589, and references therein.
(32) Benko¨, G.; He, J.; Yartsev, A. P.; Pol´ıvka, T.; Sundstro¨m, V. To be
published.
(33) (a) van’t Spijker, H.; O’Regan, B.; Goossens, A. J. Phys. Chem. B 2001,
105, 7220. (b) Benko¨, G.; Yartsev, A. P.; Sundstro¨m, V. Unpublished
results.
(28) Greijer, H.; Lindgren J.; Hagfeldt, A. J. Phys. Chem. B 2001, 105, 6314.
9
4930 J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002

Modified Phthalocyanines for NIR Sensitization
A R T I C L E S
Figure 9. Transient absorption kinetics of the ZnPcTyr-sensitized TiO2
thin film at 780 nm (oxidized dye) and 650 nm (GSR). For comparison,
kinetics of the oxidized dye in the RuN3-sensitized TiO2 film in acetonitrile
is shown. The kinetics for RuN3 was taken from our previous report.36
Symbols are experimental data points, while the solid curves are fits. See
the text for details.
Figure 8. Rise kinetics of conduction band electrons in a ZnPcTyr-TiO2
thin film at 1800 nm induced by a ∼100 fs pulse at 700 nm. Solid circles
are experimental data points and the dashed line represents the instru-
ment response function, while the solid line is a kinetic fit. See the text for
details.
of electrons in the conduction band of the TiO2 and shows that
the overall injection is over in ∼500 fs. The best fit to the
experimental data yields the following time constants and
amplitudes (in parentheses)srise within the laser pulse (25%),
delayed rise 160 fs (75%)sand decays0.6 ps (40%), 50 ps
(40%), >1 ns (20%). The pulse-limited rise may be due to either
a very fast electron injection, which we are not capable of
resolving with our spectrometer, and/or excited state absorption
of the phthalocyanine present at this wavelength. The main
component (160 fs) we assign solely to electron injection.
Kinetics recorded at around 780 nm (data not shown) show the
formation of the oxidized dye34,35 with very similar rise time
as for the conduction band electrons.
than half of the injected electrons recombine with the oxidized
dye in 300 ps (Figure 9). For RuN3-TiO2, there is no recom-
bination in 300 ps, as recombination occurs at much longer
times.30 The comparison of the two systems allows us to con-
clude that the most important condition necessary for efficient
light-to-current conversion (very fast injection combined with
slow recombination) is not established in the ZnPcTyr-TiO2
electrode. The next important condition (ZnPcTyr+ reaction with
I-) was not studied in the present work. Since recombination
is very fast (a major part of it occurs on the picosecond time
scale), the original state of the dye will be restored mainly by
conduction band electrons, before the reaction between oxidized
dye and iodide could take place. As a result, the very fast
electron recombination probably leads to the low quantum
efficiency and low overall conversion efficiency of the ZnPcTyr-
based DSSC.
As pointed out above, the decay of e-TiO , ZnPcTyr+ signals,
2
and GSR measured at 1800, 780, and 650 nm, respectively, can
be used to characterize the recombination process in the
ZnPcTyr-sensitized TiO2 film. The normalized kinetics of the
two later signals is shown in Figure 9. The decays can be
fitted with the identical time constants of 50 ps (12%), 250 ps
(65%), and >1 ns (23%), which suggests a widely distributed
recombination rate ranging from a few tens of picoseconds up
to longer times. The fit to the decay of the electron absorption
signal (see Figure 8) also contains a subpicosecond time con-
stant, which is attributed to electron thermalization and trapping
dynamics within the TiO2, a process that takes place after
injection.31
Comparing electron injection and recombination kinetics in
ZnPcTyr and RuN3-sensitized TiO2 films (comparison shown
in Figure 9), one can see that electron injection for the former
system proceeds in ∼500 fs (see Figure 8), while for the latter
system it is completed only after ∼100 ps.29 Therefore, ZnPcTyr
is slightly better than RuN3 from the electron injection point
of view, and the charge injection does not control the con-
version yield. However, the recombination in ZnPCTyr-TiO2
is considerably faster than it is in RuN3-TiO2. In fact, more
Conclusions and Perspectives
The synthesis of a modified zinc phthalocyanine with tyrosine
substituents (ZnPcTyr) and its reference, glycine-substituted zinc
phthalocyanine (ZnPcGly), is described. Incorporation of ty-
rosine groups makes the dye ethanol-soluble and decreases
considerably surface aggregation of the sensitizer due to steric
effects and as a result improves the solar cell performance.
Addition of cheno and TBP to the sensitizing dye solutions
diminishes surface aggregation of the sensitizers, especially for
ZnPcTyr, and therefore enhances significantly the solar cell
performance. Since the oxidation potential of tyrosine is more
positive than that of the zinc phthalocyanine macrocycle,
electron transfer from tyrosine to the oxidized phthalocyanine
is thermodynamically impossible and therefore not a possible
route to block back-electron transfer from the TiO2 conduction
band to the oxidized sensitizer. An IPCE of ∼24% at 690 nm
and an overall conversion efficiency of 0.54% were achieved
for a cell based on a ZnPcTyr-sensitized TiO2 electrode, which
is one of the best results reported for a photovoltaic device based
on phthalocyanine. Adding TBP to the electrolyte decreases the
IPCE, JSC, and η considerably, although it increases somewhat
the UOC. In comparison with the data for a RuN3-sensitized
TiO2 thin film, electron injection is somewhat faster (completes
(34) It was found that the ZnPcTyr+ has absorption in the near-IR spectral region,
similarly to Ru phthalocyanines.35
(35) Charlesworth, P.; Truscott, T. G.; Brooks, R. C.; Wilson, B. C. J.
Photochem. Photobiol. B: Biol. 1994, 26, 277.
(36) Kallioinen, J.; Benko¨, G.; Sundstro¨m, V.; Korppi-Tommola, J. E.; Yartsev,
A. P. Submitted for publication.
9
J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002 4931

A R T I C L E S
He et al.
in ∼500 fs), charge recombination is much faster, and more
than half of the injected electrons recombine with the oxidized
dye in ∼300 ps. Surface aggregation and fast charge recombina-
tion are responsible for the low IPCE, φ, and η of the solar cell
based on ZnPcTyr.
To further improve the performance of phthalocyanine-based
TiO2 solar cells, some other redox-active substituents than
tyrosine should be incorporated into the phthalocyanine mac-
rocycle. Besides making the dye soluble in organic solvents and
avoiding surface aggregation, the substituent should have an
oxidation potential more negative than that of phthalocyanine
to suppress the back-electron transfer from the conduction band
of the semiconductor to the oxidized dye.
Acknowledgment. Financial support was provided by the
Swedish National Energy Administration, the Knut and Alice
Wallenberg Foundation, and the Swedish Research Council. We
thank Dr. Arkady P. Yartsev at the Department of Chemical
Physics, Lund University, for helpful discussions. We also thank
Masaru Yanagisawa at the Department of Organic Chemistry,
Stockholm University, for the help with the manuscript.
JA0178012
9
4932 J. AM. CHEM. SOC. VOL. 124, NO. 17, 2002