Regenerative dye-sensitized photoelectrochemical cells have been under intense investigation for the last decade. The low production costs and good efficiency for energy conversion, reaching 10 % in some cases, make such devices a promising alternative for the development of a new generation of solar cells¹.
In our Laboratory, we have assembled a solid-state version of the regenerative dye-TiO₂ solar cell using a polymer electrolyte based on the elastomer poly(epichlorohydrin-co-ethylene oxide), poly (EO-EPI), from Daiso Co. Ltd., Osaka, complexed with NaI and I₂^2-4. This polymer electrolyte makes the assembly of the cell much easier and eliminates several problems related to the use of liquid electrolytes. However, it reduces the efficiency of the cell in comparison to those with liquid electrolytes. The best efficiency for energy conversion (h) that we have obtained for a solid-state TiO₂ / dye cell (active area = 1 cm^-2) was h = 2.6 % under 10 mW cm^-2 (h = 1.6 % under 100 mW cm^-2). Thus, such devices have a potential application in conditions of low illumination.²
Presently, we are searching for a lower cost and broader applicability for TiO₂ / dye photocells, using the same polymer electrolyte and PET-ITO flexible electrodes, i.e., a film of poly(ethylene terephthalate) coated with tin-doped indium oxide. Since PET-ITO degrades above 150 ^oC, alternative methodologies were tested for the preparation flexible photoelectrodes.
A small aliquot of TiO₂ suspension (Ti nanoxide-T, Solaronix) was spread on the conductivesurface of the PET-ITO and, for comparison, also on glass-ITO electrodes. They were exposed to UV radiation (high pressure Hg lamp) during different periods of time and heated at 130-140 ^oC during 2h in a dry box. After sensitization with a solution of the dye cis-bis (isothiocyanato)bis(2,2'-bipyridyl-4,4'-dicarboxylato)-ruthenium(II) (Solaronix), a film of the polymer electrolyte was deposited onto the photoelectrode by casting, using a solution of poly (EO-EPI) with NaI, LiI and I₂ in acetone. The assembly of the cell (1 cm²) was finalized by pressing the counter electrode, a Pt film deposited on PET-ITO by sputtering. The cells were investigated by current-potential and Electrochemical Impedance Spectroscopy (EIS) measurements under different light intensities.
The best results were obtained using an electrode exposed during 15 min to UV radiation, followed by heating before sensitization with the dye. FTIR measurements revealed that this treatment degraded, in some extension, the surfactants used to prepare the TiO₂ suspension. Such films were mechanically stable and presented an intense adsorption of the dye sensitizer. On the other hand, longer UV exposition reduced the adherence of the TiO₂ film.
The performance of this cell, and of a similar cell assembled with glass electrodes using the same procedure is compared in Table. Results previously obtained with a cell prepared without using the UV radiation, i.e., TiO₂ (2h 140 ^oC)³, as well as for a cell prepared with a different TiO₂ suspension TiO₂’(4h 140 ^oC) ⁴ are also shown. Open circuit voltage V_OC, short-circuit current I_SC, fill factor ff and efficiency h were estimated from I-V curves obtained after irradiating the cells for 4 days.

Photoelectrode	Intensity mW cm-2	VOC V	ISC mA cm-2	ff	h %
PET-ITO|TiO2	100	0.80	0.54	0.54	0.23
(15min UV+2h 130 oC)	10	0.72	0.060	0.71	0.32
glass-ITO|TiO2	100	0.76	0.46	0.71	0.25
(15min UV+2h 130 oC)	10	0.71	0.11	0.71	0.53
PET-ITO|TiO2	100	0.72	0.40	0.42	0.12
(2h 130 oC) 3	10	0.65	0.056	0.61	0.22
PET-ITO|TiO2 ‘	100	0.83	0.19	0.56	0.09
(4h 140 oC)	10	0.76	0.034	0.44	0.11

The stability of the cells assembled with TiO₂ films (15 min UV + 2 h 140 ^oC) deposited on PET and glass electrodes was investigated during 50 days by I-V and EIS measurements. The performance presented in the Table was maintained from the 1^st to the 4th day after assembling the cells and then decayed. From the 5^th until the 14^th day, the V_OC values were almost constant but photocurrent and efficiency decreased considerably. The efficiency remained unchanged at 0.17 % from the 14^th until the 45^th day and decayed to 0.10 % in the 50^th day (10 mW cm^-2).
EIS measurements revealed that the series resistance in such solar cells increased with time, affecting the stability and lowering the cell efficiency. This effect was not so evident for cells assembled by the same procedure but using glass electrodes. Therefore, the flexible electrode not only limits the preparation of the porous TiO₂ photoelectrode, as well as brings a large series resistance to the device. However, these results are very promising for developing solid-state and flexible solar cells with a lower cost and broader applicability.

Acknowledgments
Authors acknowledge Daiso Co. Ltd. Osaka, and financial support from PRONEX/CNPq and FAPESP (grants 96/09983-0, 00/03086-3, 01/02454-1 and 98/10561-6).

References
1. Hagfeldt, A.; Grätzel, M. Chem. Rev. 95 (1995) 49.
2. Nogueira, A.F.; Durrant, J.R., De Paoli, M.-A. Adv. Mater. 13 (2001) 826.
3. Longo, Claudia; Cachet, H.; Nogueira, A.F.; De Paoli, M.-A. J. Phys. Chem, submitted.
4. De Paoli, M.-A.; Machado, D.A.; Nogueira, A.F.; Longo, Claudia, Electrochim. Acta. 46 (2001) 4243.