J. Phys. Chem. B 2001, 105, 12809-12815
12809
Quasi-Solid-State Dye-Sensitized TiO2 Solar Cells: Effective Charge Transport in Mesoporous Space Filled with Gel Electrolytes Containing Iodide and Iodine Wataru Kubo, Kei Murakoshi,† Takayuki Kitamura, Shigeo Yoshida,‡ Mitsuru Haruki, Kenji Hanabusa,§ Hirofusa Shirai,§ Yuji Wada, and Shozo Yanagida* Material and Life Science and Intense 14MeV Neutron Source Facility, Department of Electronic, Information and Energy Engineering, Graduate School of Engineering, Osaka UniVersity, Suita 565-0871 Japan, and Department of Functional Polymer Science, Faculty of Textile Science & Technology, Shinshu UniVersity, Ueda 386-8567, Japan ReceiVed: May 29, 2001; In Final Form: August 30, 2001
Quasi-solid-state dye-sensitized solar cells were fabricated using low-molecular-weight gelators. They showed comparable photoenergy conversion efficiencies to the liquid cell at high illumination intensity up to AM 1.5 (1 sun). Conductivity measurements of the electrolyte phases revealed that the gelation does not affect the conductivity of the electrolyte and that the conductivity increased with an increase of iodine in both gel electrolytes and liquid electrolyte. The formation of polyiodide ions, such as I3- and I5-, caused by addition of iodine was confirmed by Raman spectroscopic measurement. The self-diffusion of iodide species in the gel electrolyte was found about a quarter of that of I- in acetonitrile. The formation of less-mobile polyiodide ions in electrolyte increased the conductivity in the mesoporous phase, which should be rationalized as due to the Grotthuss-type electron exchange mechanism caused by rather packed polyiodide species in the electrolytes. The optimized quasi-solid-state cell showed the values of 0.67 V for open-circuit voltage, 12.8 mA cm-2 for short-circuit photocurrent density, and 5.91% for photoenergy conversion efficiency under AM 1.5 irradiation with higher durability.
Introduction Dye-sensitized solar cells have attractive features in high photon-to-current conversion efficiency and low production cost.1,2 Porous nanocrystalline semiconductor electrodes are currently of great interest in solar energy conversion systems.3 A characteristic feature of the porous nanocrystalline TiO2 electrodes used in that system is the high surface area that allows higher cross section for the photoexcitation of the monolayer dye adsorbed on the electrode while maintaining contact with the electrolyte. The redox electrolyte required in these solar cells is liquid organic solution consisting of iodide (I-) and triiodide (I3-) derived from I- and I2. The liquid electrolyte in cell presents several problems such as solvent evaporation and penetration of air and water caused by difficulty in long-term sealing especially in temperature cyclic tests. Several attempts were made to quasi-solidify the liquid electrolyte by using polymer matrixes4-7 or to replace the electrolytes with p-type semiconductors8-11 or conductive organic materials12-15 as a hole-transport layer of the cells. In those cases, imperfect filling of the nanosized pore of the electrodes cased the insufficient conversion efficiencies of the solid cells compare to those using liquid electrolyte. Very recently, low-molecular-weight compounds used to hearden organic solvents, which are called “gelators”, have * To whom correspondence should be sent. E-mail: yanagida@chem. eng.osaka-u.ac.jp. WWW: http://www.mls.eng.osaka-u.ac.jp/∼mol_pro, Fax: +81-6-6879-7875. † Present address: PRESTO, Japan Science and Technology Corporation; Department of Chemistry, Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531, Japan. ‡ Intense 14MeV Neutron Source Facility. § Department of Functional Polymer Science.
received considerable attention16 since they are potentially useful for several applications: environmentally, as hardeners of spilled liquids and cooking oils; industrially, as thickeners for paint; and medically as drug-delivery materials. Gelators of lowmolecular-weight compounds have unique characteristics of both good solubility upon heating and inducement of smooth gelation of organic liquids at low concentration. The gel electrolyte shows high conductivity comparable with that of liquid one.17 We reported that the dye-sensitized solar cell using the gel electrolytes was demonstrated comparable characteristics as the cell with liquid electrolyte at high light intensity such as AM 1.5 (one sun).18,19 The comparable characteristics of the quasisolid-state solar cells were ascribed to good contact of the electrolyte with the surface dye molecules. Because the gel electrolytes gain liquidlike viscosity at temperatures >TSG (solution-to-gel transition temperature), high-temperature fabrication made them possible for the electrolyte to penetrate well into mesoscopic space. However, it is interesting that the quasisolid-state electrolyte at working conditions, i.e., at room temperature, hardly interferes with conductivity of the cells. This fact implies the specific charge transport of the electrolytes that are composed by the I-/I3- redox couple. In this paper, we report on further studies on the quasi-solidstate dye-sensitized solar cells and the charge transport property of the gel-state I-/I3- redox electrolyte in comparison with the corresponding liquid-state electrolyte. Experimental Section Chemicals. Porous TiO2 films (thickness ) 12 µm, roughness factor ) 1200) were fabricated using anatase TiO2 nanocrystallites with a diameter of ca. 23 nm according to the previously
10.1021/jp012026y CCC: $20.00 © 2001 American Chemical Society Published on Web 11/30/2001
12810 J. Phys. Chem. B, Vol. 105, No. 51, 2001
Kubo et al.
CHART 1: Chemical Structure of Gelators
reported procedure.20 The dye, cis-dithiocyanate-N,N′-bis(4carboxylate-4-tetrabutylammoniumcarboxylate-2,2′-bipyridine)ruthenium(II) (N719), was synthesized from cis-dithiocyanateN,N′-bis(4,4′-dicarboxylate-2,2′-bipyridine)ruthenium(II) (N3). A detailed description for the preparation of the dyes was given elsewhere.21 All solvents, 2-methyl-2-propanol (extra pure grade, Wako Pure Chemical), 3-methoxypropionitrile (MPN: 99%, Tokyo Chemical Industry), and 4-tert-butylpyridine (BP: 98%, Aldrich), were purified by distillation before use. Acetonitrile (infinity pure grade, Wako Pure Chemical) and other spectroscopy grade solvents were used as delivered. Gelators and Their Gelation Ability. Gelators (Chart 1) were synthesized as reported previously.22-24 These gelators form thermoreversible physical gels from a variety of organic liquids at very low concentration. Characteristics of the gelators are originated from their superior ability in the formation of intermolecular hydrogen bond between the oxygen atom in the urethane group and the hydrogen atom in the amide group. Long alkyl chains in the molecules also contribute to form stiff selfaggregates via van der Waals force.25 The gel electrolytes for the cell were prepared as follows; typically, ca. 0.04 g (40 g L-1) of gelator was mixed with 1 mL liquid electrolyte containing iodine in a tube with a screw cap, and then the mixture was heated until the solid was dissolved. The liquid electrolyte containing 0.6 M of 1,2-dimethyl-3-propylimidazolium iodide (DMPImI: Shikoku Corp.), 0.1 M of iodine (I2: Wako Pure Chemical), 0.1 M of lithium iodide (LiI: Wako Pure Chemical), and 1 M of BP dissolved in MPN. The solutionto-gel transition temperatures (TSG) of the electrolyte were 6163 °C for gelator 1, 85-87 °C for gelator 2, 47-49 °C for gelator 3, and 58-60 °C for gelator 4, under the abovementioned concentration conditions. Although some of the gel electrolytes have slight turbidness, most gels looked transparent with ruby color due to iodine. Fabrication of the Cells. Porous TiO2 film electrode was immersed in a mixture of acetonitrile and 2-methyl-2-propanol (50:50 vol/vol) solution containing 5 × 10-4 M of ruthenium dye (N719) for overnight to adsorb dye molecules. The resulting electrode was rinsed with acetonitrile containing 10 mM of BP. After drying of the electrode, the porous surface was covered by platinized conducting glass (sheet resistance 10 Ω square-1,
F doped SnO2: Nippon Sheet Glass) as a counter electrode. Effective area of the cell electrode was 0.25 cm2. Melted gel electrolytes were injected from one of two holes made on the counter electrode with keeping temperature at ca. 100 °C. The viscosity of the gel electrolytes at ca. 100 °C was low enough to inject into the gap between the electrodes. For a long-term durability test, the gap between the porous dyed-TiO2 electrode and the platinized counter electrode was adjusted to be a few tens micrometers and sealed using thermal adhesive films (HIMILAN 1652: Mitsui-Dupont Polychemical) and the electrodes were sealed by heat. The injection holes were sealed by cover glass (Iwaki Glass), HIMILAN film, and epoxy resin (Torr seal: Varian Vacuum Products). All the steps of the cell fabrication were done in a glovebox (Miwa: type DBO 1.5) filled with dry nitrogen. Measurements. Photoelectrochemical measurements were performed according to the reported procedures.2 Photoenergy conversion efficiency was evaluated using a solar simulator (Yamashita Denso: YSS-80) as AM 1.5 light source and a computer-controlled digital multimeter (Keithley Instruments: model 166. Hewlett-Packard: model HP3478A.). We adapted Japanese Industrial Standard for amorphous solar cells in the cell efficiency determination.26 Photocurrent action spectrum measurements were carried out using a computer-controlled monochromator (Instruments S. A.: Triax 180) and a potentiostat (Hokuto Denko: HA-151). The light source was a 300-W halogen lump (Atago Bussan: XC-300). The Raman spectra of the electrolytes were measured by a Raman spectrometer (Perkin-Elmer: Spectrum GX2000R) equipped with a Nd:YAG laser (10-1600 mW). Impedance measurements were carried out using twoelectrode thin-layer cells. The cell consisted of two platinum sheets separated by a 0.48-mm-thick Teflon spacer. Effective areas of the electrodes were 0.27 cm2. The frequency dependence of the cell impedance was measured using a frequency response analyzer equipped with a potentiostat (Radiometer: PGZ301) with a modulation amplitude of 5 mV. Temperature of the electrolytes was controlled at 25 °C. Conductivity of the gel electrolytes was determined by the value of the real part of the cell impedance at the frequency, which gave minimum value of imaginary.27 The cell fabrication and the impedance measure-
Dye-Sensitized TiO2 Solar Cells
J. Phys. Chem. B, Vol. 105, No. 51, 2001 12811 TABLE 1: Performance of the Quasi-Solid-State Solar Cells electrolyte
VOC /V
JSC/mA cm-2
FF
η/%
liquid gel 1 gel 2 gel 3 gel 4
0.622 0.625 0.632 0.640 0.623
11.1 10.9 11.1 11.1 11.2
0.674 0.658 0.658 0.634 0.664
4.66 4.46 4.62 4.49 4.67
Figure 1. Photocurrent-voltage curves of the cells with electrolytes containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN solvent (liquid electrolyte: bold curve), with 40 g L-1 of gelator 1 (solid curve), gelator 2 (dotted curve), gelator 3 (short dash curve), and gelator 4 (long dash curve) under AM 1.5 irradiation.
ments were carried out in the glovebox. Details in the measurements were described previously.17 The self-diffusion constant of iodide in gel electrolyte was determined by open-ended capillary method28 with radioactive iodide isotope. An aqueous alkaline Na125I (37 MBq, 10 mL) solution purchased from Amersham Pharmaticia Biotech was diluted by 490 mL of MPN, then 10.4 mL of the diluted solution (0.77 MBq) was dissolved with 3.0 mL of MPN solution of 0.5 M NaI and 0.05 M I2 with 0.12 g (40 g L-1) gelator 2. The mixture was heated at 120 °C for 5 min to obtain a homogeneous solution. Glass capillaries (inside diameter d, 0.5 mm; length l, 2.0 cm; volume, 3.9 µL) were immersed with the radioactive solution electrolyte, and then cooled to give capillaries filled with radioactive gel electrolyte. Then one end of the capillaries was capped by rubber stopper and they were placed in a large excess amount of the nonradioactive and nongel electrolyte solution bath (50 mL) in the open end up. After a given time t, the five capillaries were picked out from the bath, and the outside of them was wiped out before radioactivity measurement. The radioactivity of the capillary filled by gel electrolyte was determined by an auto well gamma system (Aroka: ARC380). The temperature was controlled at 25 ( 0.5 °C. Natural decay of the 125I was also monitored in the separated gel electrolyte. Self-diffusion coefficient D was calculated from the following equation as reported in open-ended capillary method28
At/A0 ) 8/π2 exp(-π2Dt/4l2) where A0 and At are the total radioactivities of the gel electrolyte contained in capillary at time equal to 0 and t, respectively, and l denotes the length of capillary. In our conditions, the experimentally obtained D is the average of the D values of I-, I3-, and higher polyiodide. Results and Discussion Performance of Quasi-Solid-State Dye-Sensitized Solar Cells Fabricated Using Low-Molecular-Weight Gelators. The gelators shown in Chart 1 caused physical gelation of the electrolyte consisting of 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP dissolved in MPN, giving optically transparent gel-electrolytes with good strength. Figure 1 shows photocurrent-voltage curves for cells with a liquid electrolyte
Figure 2. Incident photon-to-current conversion efficiency of the cells with electrolytes containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN solvent (liquid electrolyte: bold curve), with 40 g L-1 of gelator 1 (solid curve), gelator 2 (dotted curve), gelator 3 (short dash curve), and gelator 4 (long dash curve).
and the gel electrolytes under AM 1.5 irradiation. Open circuit voltage (VOC), short circuit photocurrent density (JSC), fill factor (FF), and photoenergy conversion efficiency (η) of the cells are listed in Table 1. It is quite interesting that the performance of the cells using gel electrolytes were comparable to that using the liquid electrolyte. When the value of JSC is same in the cell using the gel electrolytes, FF decreases slightly, because the gel electrolytes introduced into the cell with the form of liquid solution show excellent penetration in the porous film. It is said that the presence of lithium ion in the electrolyte decreases the conduction band potential of TiO2, resulting in the decrease of VOC as observed for protonation.29 The stronger interaction of gelators with lithium ions48 may suppress such interaction, leading to a slight increase in the value of VOC (3-18 mV). As a consequence, the photoenergy conversion efficiencies of the quasisolid-state solar cells showed comparable values as the photoenergy conversion efficiency of the liquid cell. The quasi-solidstate solar cells also showed no decrease of the conversion efficiency in the light intensity ranging from 10 to 100 mW cm-2. The wavelength dependence of photon-to-current conversion efficiency (IPCE) of the quasi-solid-state cells was also quite similar to that of the liquid cell (Figure 2). The IPCE spectra of the cells had a peak at 530 nm due to the absorption of the Ru-sensitizer.21 The result implies that inhibitory contribution due to light absorption of the gelators is negligible in the present conditions. As expected, the quasi-solid-state cells showed excellent durability compared with the liquid solar cells. Figure 3 shows time-dependent change in the photoconversion efficiency of the liquid and the quasi-solid-state cells fabricated using four types of low-molecular-weight gelators. In this experiment, all cells were sealed using thermal adhesive films (HIMILAN 1652). The photoconversion efficiency was measured once per a few days after being stored under dark at room temperature. Efficiencies were normalized to the values on the fifth day,
12812 J. Phys. Chem. B, Vol. 105, No. 51, 2001
Figure 3. Time-course change of the normalized efficiency of the cells: with electrolytes containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN solvent (liquid electrolyte: bold curve), with 40 g L-1 of gelator 1 (solid curve), gelator 2 (dotted curve), gelator 3 (short dash curve), and gelator 4 (long dash curve).
Figure 4. Arrhenius plots for the conductivity of electrolytes containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN solvent (liquid electrolyte: closed circle), with 40 g L-1 of gelator 1 (open circle), gelator 2 (closed triangle), gelator 3 (open triangle), and gelator 4 (closed square).
because it took 5 days to get the averaged value of the cell performance. Although the conversion efficiencies of quasisolid-state cells were kept constant (gelators 1, 2, and 4) or increased by about 10% (gelator 3), the conversion efficiency of the liquid cell decreased by about 50% after 180 days. Thus, these results proved that the quasi-solidification using the lowmolecular-weight gelators should suppress leak or evaporation of low-boiling materials in the electrolyte,30 providing the useful fabrication technique of the dye-sensitized solar cells without loosing their intrinsic performances. Characteristics of Gel Electrolytes. For further optimization of the quasi-solid-state dye-sensitized solar cell using the lowmolecular-weight gelators, we characterized the electrochemical properties of the gel electrolytes. The employed electrolyte consisted of DMPImI, LiI, and I2, and BP as an additive and MPN as solvent, which was reported as a suitable electrolyte in the dye-sensitized solar cell.31 Figure 4 shows the plots of conductivity against 1/T for the gel electrolytes using gelators 1-4. The plot of conductivity for the liquid electrolyte without any gelators is also shown. The temperature dependence shows a classical Arrhenius behavior obeying the equation
σ ) σ0 exp(-EA/kT) where σ is the conductivity, EA is the activation energy, and k is Boltzmann’s constant. It is shown that while slight decrease in the conductivity was seen in gelator 1, the gelation using gelators 2, 3, and 4 does not decrease the conductivity significantly when compared to that of the liquid electrolyte in
Kubo et al.
Figure 5. Effect of the concentration of gelators on conductivity for gel electrolytes containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN solvent (electrolyte solution: bold curve), with 40 g L-1 of gelator 1 (open circle), gelator 2 (closed triangle), gelator 3 (open triangle), and gelator 4 (closed square).
the measured region of the temperature. Interestingly, all the gelator gave the almost equal activation energy (11.5-12.3 kJ mol-1) even when compared to that of the liquid electrolyte (11.6 kJ mol-1). It should be noted that the values of activation energy of gel electrolytes using gelators 1, 2, and 4 were not changed at the temperature where the systems showed solutionto-gel transitions (TSG were 61-63 °C for gelator 1, 85-87 °C for gelator 2, 47-49 °C for gelator 3, and 58-60 °C for gelator 4, as described in the Experimental Section). An apparent but slight change in the activation energy was observed only in the case of gelator 3 at around TSG. These results suggest that the channel for the charge transport in gel electrolyte is almost identical with that of liquid electrolyte and could not be altered in the intertwining self-assembled molecular network constructed by the gelators. Figure 5 shows the effect of the concentration of gelators on the conductivity. The conductivity was decreased slightly with increasing the concentration of gelator. Moderate intermolecular interaction of gelators keeps the charge-transfer channel of the electrolytes even at their high concentration. As shown in Figures 4 and 5, the gelators have superior abilities in the formation of the quasi-solid-state electrolyte without largely affecting the electrolyte conductivity. Accordingly, no remarkable difference was observed in the cell performance, giving the almost identical performance of the quasi-solid-state solar cells with the liquid cell. These facts suggest that not only ion migration process but also other charge transport mechanisms should contribute to the hole transport in the dye-sensitized solar cell. Effect of I2 on Conductivity of the Electrolyte. The effect of gelator 2 concentration on conductivity for the electrolytes consisted of I2, DMPImI or LiI, and gelator 2, are shown in Figure 6. As the gelator concentration increased, the conductivity of the electrolyte decreases. The decrease was much more apparent in the system of LiI than that of DMPImI. The difference between the cations reflects the mobility of the ions in the molecular network of the gelator.48 When I2 is introduced into the electrolyte, the conductivity and the concentration dependence changed remarkably. As shown in Figure 6, the addition of I2 into the electrolyte resulted in the improvement of the conductivity. At high concentration region around 200 g L-1 of gelator 2, improvement of the conductivity by the addition of I2 was more than one order. The results suggest that existence of I2 causes drastic improvements in the charge transfer, which are able to work in the microscopic molecular networks of the gelators.
Dye-Sensitized TiO2 Solar Cells
Figure 6. Effect of the concentration of gelator 2 on conductivity for electrolytes containing 0.5 M of iodide salts (DMPImI: open circle. LiI: open triangle.) and electrolytes containing 0.5 M of iodide salts with 0.1 M of I2 (DMPImI: closed circle. LiI: closed triangle.).
Figure 7. Effect of concentration of iodine on conductivity for electrolytes: 0.5 M DMPImI, 0-0.5 M I2 in MPN solvent (liquid electrolyte: closed circle), with 40 g L-1 gelator 2 (gel electrolyte: open circle).
The same effect of I2 is also observed for the liquid electrolyte. Figure 7 shows the effect of the concentration of I2 on conductivity for liquid electrolyte and gel electrolyte. The improvements of the conductivity are clearly proved as due to the addition of I2 in both systems. The increments in conductivity were not linear and increase of the I2 concentration resulted in the drastic increments of the conductivity. Formation of Polyiodide Species at High Concentration of I2. To explain the increased conductivity at high concentration of I2, we examined formation of polyiodides (I2n+1-) in the electrolyte, because polyiodides would be formed at the high concentration of I2 in the presence of iodide ions. Polyiodide was considered to contribute to the charge transport process. Figure 8 shows Raman spectra of the electrolytes. In Figure 8a, the spectra of MPN solvent, a MPN solution containing 0.5 M of DMPImBr, and a MPN solution containing 40 g L-1 of gelator 2 are shown as controls. The control spectra show no apparent band in the Raman shift region of 50-250 cm-1 except for the broad band around 110 cm-1, which is attributable to the presence of MPN. In the spectra of liquid and gel electrolyte containing 0.5 M of DMPImI and I2 (Figure 8b,c), the apparent band was observed around 110 cm-1, which was assigned to symmetric stretch of I3-.32,33 The band intensity increased as of concentration of I2, while it was not observed in the absence of I2. At higher concentration of I2, another band around 145 cm-1 was observed as a shoulder of the main I3- band. This new band was assigned to the vibration mode of higher polyiodides, such as I5-, which have been reported observed at
J. Phys. Chem. B, Vol. 105, No. 51, 2001 12813 around 150 cm-1.34 The other band was too small to be attributed to vibrations of molecular iodine (I2), which are usually observed around 180-210 cm-1.35 The Raman spectral analysis of the liquid and the gel electrolytes proved that I3and higher polyiodide species are formed exclusively by the addition of an equal amount of I2 to the iodide in MPN. The difference of between solution and gel is negligible. Contribution of Electronic Conduction Process Caused by Polyiodide Species. The conductivity of electrolyte solution was generally expressed by diffusional ion migration, but it is known that the redox couple in electrolyte often causes electronic conduction process, and that electron exchange between redox couples plays a roll in electronic conduction.36-42 In dye-sensitized solar cells, the electron exchange should occur between I-, I3-, and polyiodides through chemical bond exchange.43 The following fact and findings suggest that the Grotthuss-type charge-transfer mechanism, where electron hopping and polyiodides bond exchange were coupled, should contribute to the effective conductance of the gel electrolyte, rationalizing the high conductivity of the electrolyte in the dyesensitized solar cells. First, in the electrolyte, the added I2 was consumed for forming polyiodides, and no molecular iodine was observed by Raman spectral analysis. The ionic strength that contributes to charge transport was unchanged. Further, the ionic radius of iodide species increased with the addition of I2 due to the formation of polyiodide. The viscosity of the electrolyte increases with the I2 concentration. The ionic component in charge transport process should be unchanged or decreased with increase of I2 in the electrolyte. Second, polyiodides have relatively low limiting molar conductivities (∆0) than the limiting molar conductivity of monoiodide due to their large ionic radius.44 Assuming that the conductivity be only determined by diffusion or migration of ionic species in the electrolytes, the formation of the large polyiodide ions especially in the gel electrolyte should have decreased the conductivity of the electrolyte. As shown in Figure 7, however, the conductivity increased drastically as the concentration of polyiodides. Third, in the concentrated iodide solution about 0.5 M, average separation between ion species is estimated about 8 Å.45 Taking into account the ionic radius of iodide, 2.2 Å, and the covalent radius of iodine, 1.3 Å,46 polyiodides and iodide species should be located in proximity each other in the electrolyte. This proximity situation suggests the possibility of the Grotthuss-type charge-transfer mechanism.43 (polyiodide bond exchange mechanism) The conductivity of the tetramethylammonium polyiodide (Me4N+ I2n+1-) at high concentration (>2.5 × 10-3 M) in acetonitrile was reported to be higher than that of tetramethylammonium iodide (Me4N+ I- at the concentration (> 2.5 × 10-3 M)).44 This fact will be explained similarly. Fourth, the diffusion of I3- in the mesoporous film was recently determined 1 order of magnitude lower than that in the free electrolyte.47 We now determined the self-diffusion coefficient of iodide species in the gel electrolyte by measuring a time-course decay of the radioactivity using the open-ended capillary method as shown in Figure 9. The self-diffusion coefficient D of iodide species (I-, I3-, and higher polyiodide) was 6.8 ( 0.5 × 10-6 cm2 s-1, which is about a quarter of D of I- in acetonitrile (2.72 × 10-5 cm2 s-1). This result suggests that the gel electrolyte provides a good channel for the selfdiffusion of iodide species.
12814 J. Phys. Chem. B, Vol. 105, No. 51, 2001
Kubo et al.
Figure 8. Raman spectra of the electrolytes. (a) Control solutions: MPN (bold curve), 0.5 M DMPImBr in MPN (solid curve), and 40 g L-1 gelator 2 solved in MPN (dotted curve). (b) Liquid electrolytes: MPN (control: bold curve), 0.5 M DMPImI with 0 M of I2 (solid curve), 0.01 M of I2 (long dash curve), 0.05 M of I2 (short dash curve), 0.1 M of I2 (broken curve), and 0.5 M I2 in MPN solvent (dotted curve). (c) Gel electrolytes: 40 g L-1 gelator 2 in MPN (control: bold curve), 0.5 M DMPImI with 0 M of I2 (solid curve), 0.01 M of I2 (long dash curve), 0.05 M of I2 (short dash curve), 0.1 M of I2 (broken curve), and 0.5 M of I2 (dotted curve).
Figure 9. Time-course decay of the normalized radioactive iodide in the gel electrolyte contains 0.5 M of NaI, 0.05 M I2, 40 g L-1 of gelator 2 in MPN solvent.
Optimization for the Dye-Sensitized Solar Cells. Gelator 1 was the most promising materials for the quasi-solid-state solar cell in our present experiment conditions. Moderate interaction between the gelator molecules gave the most durable cell that was composed of high concentration of gelators without loosing their performance. Figure 10 shows an I-V curve of the quasisolid-state solar cell using gelator 1, in which electrolyte consists of 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN with 10 g L-1 of gelator 1. The cell shows the best value of 0.67 V for VOC, 12.8 mA cm-2 for JSC, 0.67 for FF, and 5.91% for η. At this moment, this is the best performance of our quasi-solid-state solar cell using low-molecular-weight gelators. Conclusion Quasi-solid-sate solar cells were successfully fabricated with the low-molecular-weight gelators. The cells showed comparable photoenergy conversion efficiencies to the liquid cell. Successful application of the gelators to the quasi-solid-state solar cell should reflect perfect penetration of the gel electrolyte into the
Figure 10. Photocurrent-voltage curves of the cell with gel electrolyte containing 0.6 M of DMPImI, 0.1 M of I2, 0.1 M of LiI, and 1 M of BP in MPN with 10 g L-1 of gelator 1 under AM 1.5 irradiation.
porous space of the TiO2 electrode. The molecular aggregate of the gelator also supports the electrolyte without loosing their charge transport properties in the space of porous nanocrystalline electrode. Conductivity measurements showed that the gelation does not influence significantly the charge transport process in the electrolyte. Remarkable improvements in the conductivity at high concentration of iodine were observed in both liquid and gel electrolytes. The conversion of iodine to polyiodides, I2n+1-, was confirmed by Raman spectroscopic measurements of the electrolyte containing excess iodine. The effective charge transport in the gel electrolyte containing polyiodide species was rationalized by the Grotthuss-type electron exchange mechanism (polyiodide bond exchange mechanism). The gelation led to higher durability of the cell without loosing their intrinsic performance of the dye-sensitized solar cell. Present findings alleviate the concerned problem of the dye-sensitized solar cell for the durability due to the evaporation of the electrolyte solvent. High conductive solid electrolytes containing the iodide/iodine redox couple and the gelators can be applied not only to the solar cell, but also to wide variety of electrical devices.
Dye-Sensitized TiO2 Solar Cells Acknowledgment. We are grateful to Dr. S. Kuwabata (Department of Material Chemistry, Graduate School of Engineering, Osaka University) for valuable suggestions about the experiments and the interpretations of impedance measurements, as well as to Dr. Greg P. Smestad (Associate and Letters Editor, Solar Energy Materials and Solar Cells) for teaching us the fabrication of the cells, and to Prof. T. Yamamoto and Dr. Y. Yamaguchi (Radioisotope Research Center, Osaka University) for the conducting of isotope experiments. This work was partially supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture, Japan (Grant 11358006), and Grant-in-Aid for the Development of Innovative Technology (No. 12310) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. References and Notes (1) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737. (2) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mueller, 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. (4) Cao, F.; Oskam, G.; Searson, P. C. J. Phys. Chem. 1995, 99, 17071. (5) Matsumoto, M.; Miyazaki, H.; Matsuhiro, K.; Kumashiro, Y.; Takaoka, Y. Solid State Ionics 1996, 89, 263. (6) Matsumoto, M.; Wada, Y.; Kitamura, T.; Shigaki, K.; Inoue, T.; Ikeda, M.; Yanagida, S. Bull. Chem. Soc. Jpn. 2001, 74, 387. (7) Tennakone, K.; Sendeera, G. K. R.; Perera, V. P. S.; Kottegoda, I. R. M.; De Silva, L. A. A. Chem. Mater. 1999, 11, 2474. (8) O’Regan, B.; Schwartz, D. T. Chem. Mater. 1995, 7, 1349. (9) O’Regan, B.; Schwartz, D. T. J. Appl. Phys. 1996, 80, 4749. (10) Tennakone, K.; Kumara, G. R. R. A.; Kumarsinghe, A. R.; Wijayantha, K. G. U.; Sirimanne, P. M. Semicond. Sci. Technol. 1995, 10, 1689. (11) Tennakone, K.; Kumara, G. R. R. A.; Kumarashinghe, A. R.; Kottegoda, I. R. M.; Wijayantha, K. G. U.; Perera, V. P. S. J. Phys. D: Appl. Phys. 1998, 31, 1492. (12) Murakoshi, K.; Kogure, R.; Wada, Y.; Yanagida, S. Chem. Lett. 1997, 471. (13) Murakoshi, K.; Kogure, R.; Wada, Y.; Yanagida, S. Sol. Energy Mater. Sol. Cells 1998, 55, 113. (14) Hagen, J.; Schaffrath, W.; Otschik, P.; Fink, R.; Bacher, A.; Schmidt, H.; Haarer, D. Synth. Met. 1997, 89, 215. (15) Bach, U.; Lupo, D.; Comte, P.; Moser, J. E.; Weisso¨rtel, F.; Salbeck, J.; Spreitzer, H.; Gra¨tzel, M. Nature 1998, 395, 583. (16) Terech, P.; Weiss, R. G. Chem. ReV. 1997, 97, 3133. (17) Hanabusa, K.; Hiratsuka, K.; Kimura, M.; Shirai, H. Chem. Mater. 1999, 11, 649. (18) Kubo, W.; Murakoshi, K.; Kitamura, T.; Wada, Y.; Yanagida, S. Chem. Lett. 1998, 1241. (19) Yanagida, S.; Kambe, S.; Kubo, W.; Murakoshi, K.; Wada, Y.; Kitamura, T. Z. Phys. Chem. 1999, 212, 31.
J. Phys. Chem. B, Vol. 105, No. 51, 2001 12815 (20) Murakoshi, K.; Kano, G.; Wada, Y.; Yanagida, S.; Miyazaki, H.; Matsumoto, M.; Murasawa, S. J. Electroanal. Chem. 1995, 396, 27. (21) Nazeeruddin, M. K.; Zakeeruddin, S. M.; Humphry-Baker, R.; Jirousek, M.; Liska, P.; Vlachopoulos, N.; Shklover, V.; Fischer, C.; Gra¨tzel, M. Inorg. Chem. 1999, 38, 6298. (22) Hanabusa, K.; Tange, J.; Taguchi, Y.; Koyama, T.; Shirai, H. J. Chem. Soc., Chem. Commun. 1993, 390. (23) Hanabusa, K.; Tanaka, R.; Suzuki, M.; Kimura, M.; Shirai, H. AdV. Mater. 1997, 9, 1095. (24) Hanabusa, K.; Nakamura, H.; Kimura, M.; Shirai, H. Chem. Lett. 2000, 1070. (25) Hanabusa, K.; Shirai, H. Kobunshi Ronbunshu 1995, 52, 773. (26) Japanese Industrial Standard Committee Japanese Industrial Standard Solar Simulators for Amorphous Solar Cells and Modules; Japanese Standards Association: Tokyo, Japan, 1995; Vol. JIS C 8933. (27) Hagenmuller, P.; Gool, W. V. Solid Electrolytes; Academic Press: New York, 1978. (28) Anderson, J. S.; Saddington, K. J. Chem. Soc. 1949, S 381. (29) Zaban, A.; Ferrere, S.; Gregg, B. A. J. Phys. Chem. B 1998, 102, 452. (30) Kay, A.; Gra¨tzel, M. Sol. Energy Mater. Sol. Cells 1996, 44, 99. (31) Zakeeruddin, S. M.; Nazeeruddin, M. K.; Humphry-Baker, R.; Gra¨tzel, M.; Shklover, V. Inorg. Chem. 1998, 37, 5251. (32) Tadayyoni, M. A.; Gao, P.; Weaver, M. J. J. Electroanal. Chem. 1986, 198, 125. (33) Andrews, L.; Prochaska, E. S.; Loewenschuss, A. Inorg. Chem. 1980, 19, 463. (34) Loos, K. R.; Jones, A. C. J. Phys. Chem. 1974, 78, 2306. (35) Klaboe, P. J. Am. Chem. Soc. 1967, 89, 3667. (36) Dahms, H. J. Phys. Chem. 1968, 72, 362. (37) Ruff, I.; Friedrich, V. J. J. Phys. Chem. 1971, 75, 3297. (38) Ruff, I.; Friedrich, V. J.; Demeter, K.; Csaillag, K. J. Phys. Chem. 1971, 75, 3303. (39) Ruff, I.; Korosi-odor, I. Inorg. Chem. 1970, 9, 186. (40) Ruff, I. Electrochim. Acta 1970, 15, 1059. (41) Botar, L.; Ruff, I. Chem. Phys. Lett. 1986, 126, 348. (42) Botar, L.; Ruff, I. J. Chem. Phys. 1985, 83, 1292. (43) Papageorgiou, N.; Athanassov, Y.; Armand, M.; Bonhoˆte, P.; Pettersson, H.; Azam, A.; Gra¨tzel, M. J. Electrochem. Soc. 1996, 143, 3099. (44) Popov, A. I.; Rygg, R. H.; Skelly, N. E. J. Am. Chem. Soc. 1956, 78, 5740. (45) Turro, N. J. Modern Molecular Photochemistry; The Benjamin/ Cummings Publishing Compan, Inc.: Menlo Park, CA, 1978; p 319. (46) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 75th ed.; CRC Press Inc.: Boca Raton, 1994; p 9-19, 12-8. (47) Kebede, Z.; Lindquist, S. Sol. Energy Mater. Sol. Cells 1998, 51, 291. (48) Lithium ion should be immobilized by the interaction with the hydrogen bond site of gelator. The interaction between lithium ion and the site affects the ability of the gelation. The minimum concentration of gelator necessary for gelation significantly depends on the content of lithium ion in the electrolyte. For example, 10 g L-1 for 0.5 M of DMPImI, 5 g L-1 for 0.5 M of DMPImI with 0.1M of I2, 110 g L-1 for 0.5 M of LiI, and 60 g L-1 for 0.5 M of LiI with 0.1 M of I2 at 25 °C in MPN.