3276
Langmuir 2009, 25, 3276-3281
Effect of Cation Size on Solid Polymer Electrolyte Based Dye-Sensitized Solar Cells Bhaskar Bhattacharya,*,†,‡ Jun Young Lee,† Jianxin Geng,† Hee-Tae Jung,† and Jung-Ki Park*,† Department of Chemical and Biomolecular Engineering (BK 21 Graduate Program), Korea AdVanced Institute of Science and Technology (KAIST), 373-1 Guseong-dong, Yuseong-gu, Daejeon, 305-701, Republic of Korea, and Department of Physics, Hindustan College of Science and Technology, Farah, Mathura-281 122, India ReceiVed September 5, 2008. ReVised Manuscript ReceiVed January 3, 2009 We report the preparation of a series of solid polymer electrolytes with different cations but with the same anion. We have chosen a PEO:PEG blend (at 40:60 by w/w) and complexed it with LiI, NaI, KI, NH4I, 1-ethyl-3methylimidazolium iodide (EMII), and 1-hexyl-3-methylimidazolium iodide (HMII) with the same ion to monomer ratio. The role of different cations in terms of their size has been investigated on the conductivity of the polymer electrolyte films. The electrolyte properties have been explained in terms of the changes in the number of free charge carriers and the variation in crystallinity of the matrix. The cell performance of the dye-sensitized solar cells (DSSC) fabricated with these systems also showed strong dependence on the cation radii. The concept of ion intercalation and surface adsorption used for the liquid electrolyte systems has been extended to the polymer electrolytes and discussed in this paper.
1. Introduction Dye-sensitized solar cells (DSSC) have become a topic of international interest and challenge since the pioneering work of Gra¨tzel.1 A conversion efficiency of more than 10% has brought this class of devices near to the achievable goal for global energy solution. However, most of the works giving high efficiency utilize liquids as electrolytes that limit their outdoor application and practical utility. The use of solid polymer electrolyte in place of the liquid electrolyte has shown promise for DSSC applications.2,3 However, in the use of polymers, their low conductivity and partial crystalline nature come as the main hurdles. Several efforts have been made so far to reduce the crystallinity and at the same time to increase the conductivity. The use of different plasticizers4,5 and inorganic fillers like titania, silica, or alumina6,7 has already been reported to enhance the electrolyte property. In the last four decades, after the work by Wright et al. on the doping of alkali halides in poly(ethylene oxide) (PEO),8 various salts have been doped so far and their effects on the electrolyte properties have been explored.9,10 However, no comparative analysis on the use of different alkali and other monovalent salts in the polymer electrolytes, in view of their ionic radii, has been * To whom correspondence should be addressed. Tel: 82-42-350-3925. Fax: 82-42-350-3910. E-mail:
[email protected] (B.B), jungpark@ kaist.ac.kr (J.-K.P.). † Korea Advanced Institute of Science and Technology (KAIST). ‡ Hindustan College of Science and Technology.
(1) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737. (2) (a) Nogueira, A. F.; De Paoli, M.-A.; Montanari, I.; Monkhouse, R.; Nelson, J.; Durrant, J. R. J. Phys. Chem. B 2001, 105, 7517. (b) Nogueira, A. F.; Durrant, J. R.; De Paoli, M.-A. AdV. Mater. 2001, 13, 826. (3) Gunes, S.; Sariciftci, N. S. Inorg. Chim. Acta 2008, 361, 581. (4) Cameron, G. G.; Ingram, M. D.; Sarmouk, K. Eur. Polym. J. 1990, 26, 197. (5) MacCallum, J. R.; Vincent, C. A. Polymer Electrolyte ReViews 1; Elsevier: New York, 1989. (6) Wieczorek, W.; Such, K. H.; Cislike, W. Y.; Polcharski, J. Solid State Ionics 1989, 36, 255. (7) Croce, F.; Appetecchi, G. B.; Persi, L.; Scrosati, B. Nature 1998, 394, 456. (8) Fenton, D. E.; Parker, J. M.; Wright, P. V. Polymer. 1973, 14, 589. (9) Sorensen, P. R.; Jacobson, T. Electrochim. Acta 1982, 27, 1675. (10) Reddy, M. J.; Chu, P. P. Electrochim. Acta 2002, 47, 1189.
seen so far. In terms of the electrochemical device application of the solid polymer electrolytes, e.g., in the DSSC, use of different alkali halides in doping of the polymer electrolyte and providing the redox couple have been reported so far. The use of LiI, NaI, and KI has been reported in PEO, out of which KI is the most common in the literature.11-13 In this paper, we report a comparative study on the electrolyte and solar cell properties of the PEO-based polymer electrolyte with different cationiodide salts (MI), the cations (M+) being Li+, Na+, K+, NH4+, EMI+, and HMI+ with wide variation in their ionic radii. We have chosen the Li+ ion as the smallest ion and the largest (almost 5 times) as HMI+. It has been shown that the number of charge carriers, in this case, has almost no role in controlling the conductivity. The change in the electrolyte properties have been attributed to the change in the mobility of the ions due to the change in the crystallinity of the polymer electrolyte. The change in the solar cell performance has been explained in terms of the surface adsorption theory, which has been employed in the case of liquid electrolytes. The change in the flatband condition and hence the potential has been discussed in detail.
2. Experimental Section 2.1. Preparation of Solid Polymer Electrolyte. Poly(ethylene oxide) (PEO, Mw ∼ 5 × 106, Aldrich) and poly(ethylene glycol) (PEG, Mw ∼ 200, Aldrich) and all the iodide salts, viz., lithium iodide (LiI, Aldrich), sodium iodide (NaI, Aldrich), potassium iodide (KI, Aldrich), ammonium iodide (NH4I, Aldrich), 1-ethyl3-methylimidazolium iodide (EMII, Solaronix), and 1-hexyl-3methylimidazolium iodide (HMII, Solaronix), were used as procured. All chemicals were used without further purification. The desired amount of the polymers and the salt were weighed in a glovebox and dissolved in acetonitrile (ACN, Merck) stored over molecular sieves. The PEO:PEG ratio was fixed at 40:60 (by w/w) for all the (11) Upadhyaya, H. M.; Hirata, N.; Haque, S. A.; De Paoli, M.-A.; Durrant, J. R. Chem. Commun. 2006, 877. (12) Stergiopoulos, T.; Arabbtzis, I. M.; Katsaros, G.; Falaras, P. Nano Lett. 2002, 2, 1259. (13) Kalaignan, G. P.; Kang, M. S.; Kang, Y. S. Solid State Ionics 2006, 177, 1091.
10.1021/la8029177 CCC: $40.75 2009 American Chemical Society Published on Web 02/05/2009
Solid Polymer Electrolyte Based DSSC samples. Though it was possible to have good and stable films up to 20:80 ratio of the two polymers with LiI, for KI and imidazolium based salts, the films were very sticky and could not be taken off the Petri dish. Therefore, to maintain uniformity, we have chosen this ratio for which all the films were stable and could be dried up. For all the samples the [EO]:[M+] was fixed as 20:1. The mixtures were stirred for more than 10 h to ensure the complete complexation of the polymer with the salt. The viscous solutions were then cast on glass Petri dishes and dried in N2 atmosphere. For complete elimination of the solvent, the films were further dried in a vacuum oven for 2 days at room temperature. 2.2. Characterization of Polymer Electrolyte. The solid polymer electrolyte films were subjected to structural, thermal, and electrical characterization for their possible use in DSSC. Differential scanning calorimeter (DuPont TA 2000 DSC) was used to understand the thermal behavior of the complexes. Each sample was scanned at a heating rate of 10 °C/min within an appropriate temperature range under nitrogen atmosphere. Impedance spectroscopic techniques were used to evaluate the ionic conductivities of the polymer films. The polymer films were sandwiched between two polished stainless steel (SS) electrodes and were vacuum-packed in an aluminum plastic pouch to avoid any contamination. Ionic conductivities of polymer films were calculated from bulk resistance estimated from the by ac complex impedance plots obtained using a Solartron 1455 frequency response analyzer (FRA) over a frequency range of 100 Hz-1 MHz. The conductivity (σ) was calculated using the relation σ ) d/RbA, where d and A are the thickness of the prepared polymer film and area of the electrodes respectively, and Rb is bulk resistance obtained from the intercept on the real Z-axis of the impedance data in the complex plane. The dielectric constant of the samples has also been calculated from the capacitance value obtained from the impedance data at 1 MHz frequency. The structural features of the polymer electrolyte film surfaces were studied with an optical microscope (OM, Leica DM LB) under cross polarizer. The X-ray diffraction (XRD, Rigaku D/MAX-RC 12kW) pattern of the samples was obtained with 2θ values ranging from 10° to 80°. The results have been given as Supporting Information with this paper. 2.3. Fabrication of Cell Assembly. DSSCs were prepared by the method published earlier.14 A two-step casting method was employed to cast the polymer electrolyte film on the dye-soaked TiO2 film.15 No additional sealing was used. The drying of the films was ensured before subjecting the cells to any measurement. All the measurements were carried out the same day the specific cell was prepared to avoid the possibilities of any degradation and other reactions. Before measurements, the cells were stored under dark conditions at room temperature. At least three cells for each electrolyte were prepared and the stable result has been considered for comparisons/discussion. Photocurrent density-photovoltage (J-V) characteristics of the DSSCs were recorded with a computer-controlled digital source measure meter (Keithley 236) by applying external potential bias to the cell and measuring the photocurrent generated under a solar simulator (Oriel 91192 model) with an AM 1.5 filter and was calibrated prior to use. All cell voltages reported in the text below refer to the potential difference between the counter electrode and the photoanode and may be written as V ) ECE - EPA, where CE and PA represent counter electrode and the photoanode respectively. The cell performances in dark were recorded by placing the cell inside a black box and switching off the solar simulator.
3. Results and Discussion 3.1. Role of Cations in the Electrolyte. Figure 1 shows the conductivity of the PEO:PEG blend electrolytes complexed with (14) (a) Bhattacharya, B.; Tomar, S. K.; Park, J. K. Nanotechnology 2007, 18, 485711. (b) Lee, J. Y.; Bhattacharya, B.; Kim, D. W.; Park, J. K. J. Phys. Chem. C 2008, 112, 12576. (15) Kang, M. S.; Kim, J. H.; Kim, Y. J.; Won, J.; Park, N. G.; Kang, Y. S. Chem. Commun. 2005, 889.
Langmuir, Vol. 25, No. 5, 2009 3277
Figure 1. Conductivity isotherm of the polymer electrolyte films containing different MI salts at room temperature (M+ ) Li+, Na+, K+, NH4+, EMI+, and HMI+) with the ascending cation size.
salts containing different MI salts, where M ) Li+, Na+, K+, NH4+, EMI+, and HMI+. The conductivity is found to increase with increasing size of the cation. The sizes of the ions have been taken from Lange’s Handbook of Chemistry16 for the alkali metal ions and from literarure for the ammonium ion.17 For imidazolium salts, the radii have been calculated from the ion volumes reported by Krossing et al.,18 assuming the ions to be spherical. It is generally believed that the conductivity of the electrolyte with smaller cation should be higher, possibly due to the higher mobility of smaller cations. However, our results contradict this belief: we find that the conductivity of the films with smaller cation is less compared to that with the larger cation. In order to justify this behavior and to understand the reason for the increase in conductivity with larger cation size, we have tried to explore all the possibilities causing changes in the ionic conductivity. The total conductivity is given by the relation
σ ) nqµ where n is the number of dissociated charge carriers in the matrix, q is the charge carried by them, and µ is the mobility of the carriers. Therefore, any change in the number of charge carriers and/or in the mobility will result in changing the total conductivity value. We have calculated the relative number of dissociated (free) charge carriers available for conduction to verify the contribution of n or µ in the increase in the conductivity with the size of the cations. We have assumed that the salt gets completely dissociated when dissolved in the polymer matrix. According to the electrolyte dissociation model,19 the number of dissociated charge carriers (n) is given as
{ }
n ) n0 exp
-U 2εkBT
where, U is the dissociation energy of the salt and ε is the dielectric constant of the matrix at temperature T. The dielectric constants of the samples were calculated at 1 MHz frequency from the impedance data and are listed in Table 1. We have taken the dissociation energy values for the alkali salts from their (16) Dean, J. A. Lange’s Handbook of Cheistry, 14th ed.; McGraw-Hill: New York, 1992. (17) Gray, F. M. Solid Polymer Electrolytes: Fundamentals and Technological Applications: Wiley-VCH Publishers Inc.: New York, 1991. (18) Krossing, I.; Slattery, J. M.; Daguenet, C.; Dyson, P. J.; Oleinikova, A.; Weingartener, H. J. Am. Chem. Soc. 2006, 128, 13427. (19) Barker, R. E., Jr.; Thomas, C. R. J. Appl. Phys. 1964, 35, 3203.
3278 Langmuir, Vol. 25, No. 5, 2009
Bhattacharya et al.
Table 1. Dielectric Constant of the Polymer Electrolytes Containing Iodide Salts with Different Cationsa type of cation
cation radius (in Å)
dielectric constant
no salt Li+ Na+ K+ NH4+ EMI+ HMI+
0.76 1.02 1.38 1.5 3.33 3.85
76.56 722.95 674.52 596.34 574.16 121.41 525.20
a The size of the cations (as mentioned in text) has been adopted from refs 16, 17, and 18.
spectroscopic data,20 for NH4I the value comes from the literature,21 and for ionic liquids, we have used the interaction energy (of cations with the I- anion) values as calculated by Turner et al.22 Figure 2 shows the relative number of the carriers (n/n0) for the samples with different cation radii. We have fixed the cation to monomer ratio (M+/EO) for all the salts, and in polymer electrolytes complexed alkali halides the maximum contribution to the conductivity is known to be by the anions (I- in this case).23 Therefore, the number of charge carriers remains almost the same for Li+, Na+, or K+. For NH4+, as reported in literature,21 the cation (H+) transference number is greater than 0.7 with a mobility of the order of ∼10-6, but in our case, no significant change in the number of charge carriers have been observed. However, in case of the larger cations EMI+ and HMI+, we believe that the cation is very large and can rarely migrate through, and therefore, there is a decrease in the relative number of charge carriers. As stated by Watanabe,24 ionic liquids basically consists of charged cationic and anionic species which can diffuse and contribute to the conduction. However, at the same time they are much prone to the Coulombic interaction between the charged species due to their many molecular parameters. Therefore, the possibility of the formation of ion pairs and neutral ion aggregates is more when compared with the alkali halides. In that case, the cationic transference number, in any of the studied polymer salt complexes, will not exceed the 0.5 value and hence the system will remain predominantly anion conductor, which is desired for the DSSC applications. The formation of ion aggregates25 in these systems, viz., EMI+ and HMI+, is also evidenced by the decrease in the dielectric constant value of the matrix for these systems (see Table 1). The maximum value is for the Li salt, for which the number of free carriers is also high. As we have mentioned in eq 1, the contribution to the total conductivity may be due to the mobility as well. As we have shown in the above section, the role of the number of charge carriers is not significant, so the role of mobility should be responsible for the conductivity enhancement. The direct measurement of the mobility is almost impossible for the polymeric systems; also, it is an established fact that in polymer electrolytes, the ions can move only through the amorphous regions.26 Therefore, to get an idea about the contribution of the mobility, we have measured the change in enthalpy and calculated (20) Herzberg, G. Molecular Spectra and Molecular Structure I. Spectra of Diatomic Molecules; Krieger Publishing Co., Malabar, FL, 1989. (21) Maurya, K. K.; Srivastava, N.; Hashmi, S. A.; Chandra, S. J. Mater. Sci. 1992, 27, 6357. (22) Turner, E. A.; Pye, C. C.; Singer, R. D. J. Phys. Chem. A 2003, 107, 2277. (23) Cheradame, H.; LeNest, J. F.; Gandini, A.; Leveque, M. J. Power Sources 1985, 14, 27. (24) Noda, A.; Hayamizu, K.; Watanabe, M. J. Phys. Chem. B 2001, 105, 4603. (25) Singh, B.; Hundal, M. S.; Park, G. G.; Park, J. S.; Lee, W. Y.; Kim, C. S.; Yamada, K.; Sekhon, S. S. Solid State Ionics 2007, 178, 1404. (26) (a) Sato, H. Soild State Ionics 1998, 28/30, 333. (b) Halder, B.; Singru, R. M.; Maurya, K. K.; Chandra, S. Phys. ReV. B 1996, 54, 7143.
the relative change in the crystallinity of the films due to the iodide salts with different cations. The change in enthalpy or the area under the peak due to the melting of the crystalline parts is directly proportional to the amorphicity/crystallinity of the material. We have calculated the crystallinity of the films from the change in enthalpy values due to the melting of the crystalline part of the polymer and the polymer-salt complex as well. The crystallinity values of the samples are plotted against the cation size of the salts in Figure 3. As expected, the crystallinity of the PEO:PEG blend drops to less than its 77% due to the addition of Li salt. It can be seen that, as the size of the salt cation increases, the crystallinity decrease almost linearly from Li+ > Na+ > K+. However, this trend is not followed by the NH4+ salt, for which the crystallinity value is almost the same as for the Li+ salt. This may be due to the role of H+ ion, which has a smaller radius than the former.21 In case of larger cations, the crystallinity drops further down, as can be seen in Figure 3. A comparison of this figure with the conductivity values (Figure 1) shows almost direct correlation. The conductivity is higher for the systems with lower crystallinity and vice-versa. Therefore, in the case of our systems, it is the crystallinity of the medium which in turn relates with the mobility of the ions and is the main controlling factor for the conductivity. The role of the free charge carriers is only an addition to this and does not play the dominant role. 3.2. Role of Cations in the DSSC. The DSSC were fabricated using all the different polymer electrolytes. As discussed in section 2, extra iodine (10% w/w of the iodide salt) was also added for the formation of the redox couple. The J-V curve of the solar cells for all the samples showed the typical S-type nature in the dark, indicating the formation of the junction and corresponding barriers at the electron collecting photoanode. In the dark, the absence of appreciable current until the contacts start injecting heavily due to the forward biasing larger than the open circuit voltage indicates the Schottky-type behavior of the junction. Figure 4 shows the solar cell region of the characteristics for the DSSC fabricated with different polymer electrolytes under AM 1.5. Under illumination, the contribution of the photogenerated carriers in the total current is clearly seen. At the flat band condition, the photogenerated current is balanced to zero, whereas at the short circuit condition, the maximum generated photocurrent could be seen. The maximum power point is identified between these two points, and the corresponding efficiency is calculated. It is observed that the efficiency of the cells is much higher than those reported using only PEO as the host polymer. These results are comparable to that reported by Upadhyaya et al.11 with NaI (+I2) and plasticized PEO and much better than those reported using only PEO as the polymer electrolyte, where the efficiency is ∼0.01% only.13 As the charge transfer reaction, in all cases, is mediated by the I-/I3-, the total reaction at the Pt/electrolyte reaction for all the cells can be described as
3I- T I3- + 2eThe efficiency and short circuit current (Jsc) of the cells obtained from Figure 5 are plotted against the cation radii in Figure 5. The efficiency of the cells is found to vary drastically with the size of cation. The figure indicates an almost linear decrease in the photoconversion efficiency with the increasing cation radius. Similarly, the short circuit photocurrent density also follows the linear decreasing trend from Li+ to HMI+. It should be recalled here that for all of our samples the M+/EO (and hence I-/EO) ratio was fixed and the charge transfer mediator was the I-/I3couple. Therefore, all the variations in efficiency and/or Jsc are
Solid Polymer Electrolyte Based DSSC
Langmuir, Vol. 25, No. 5, 2009 3279
Figure 2. The relative number of charge carriers as calculated using eq 2 for different MI salts (M+ ) Li+, Na+, K+, NH4+, EMI+, and HMI+) complexed with polymer films.
Figure 3. Change in the crystallinity of the polymer electrolyte films with increasing size of the cation as calculated from the DSC data (for details, see the text).
Figure 5. The change in the DSSC parameters for cations with different ionic radii showing the cell efficiency ([) and the short circuit current (9) decreasing from Li+ to HMI+. The dotted line is only for guidance of the eye.
Figure 4. The solar cell characteristics under AM 1.5 illumination. The inset indicates the salts used in the solid polymer electrolyte as the electrolyte for the DSSC.
polymer electrolyte, as we have already shown (Figure 1) that with the increase in the size of the cation, the conductivity of the polymer electrolyte increases. Therefore, the decrease in the cell parameter must be related with the modification only in the cell properties due to the cation size. Similar observations showing a decrease in the photocurrent have been reported by Wolfbauer et al.27 in the CH3CN:H2O mixed liquid electrolyte. Different justifications have been given so far. But none of the reports, to our knowledge, has considered the effects of cations in solid polymer electrolytes. The cations may either intercalate, as very well established in the case of Li+ ions, or even get adsorbed at the surface of the polycrystalline TiO2 and thus can result in the possible changes in the energy band structure of the interface and hence the cell properties. In the following paragraphs, we have discussed both the possibilities that have been applied to the liquid electrolytes so far and proposed a model for the observed changes in the cell parameters with different cation size.
predominantly due to the cations. The decrease in the Jsc and efficiency value cannot be related with the conductivity of the
(27) Wolfbauer, G.; Bond, A. M.; Eklund, J. C.; MacFarlane, D. R. Sol. Energy Mater. Sol. Cells 2001, 70, 85.
3280 Langmuir, Vol. 25, No. 5, 2009
Bhattacharya et al.
Figure 6. The change in the flat band potential (Vfb′) due to the adsorbed/intercalated layer of the cations at the photoanode-polymer (redox) electrolyte interface.
3.2.1. Intercalation of Cations into the Dye Layer. For Li+, intercalation is a well-established phenomenon and would be expected to lead to a positive shift in the potential of the conduction and valance band edges at the electrode-electrolyte interface.28 Small cations, which can practically intercalate in the pores of the TiO2 electrode, will form charge traps and hence will modify the diffusion as well as the recombination rates. If the intercalating cation had a tendency to penetrate the crystal structure to a significant extent, it would be expected that electrostatic attraction between the counterion and the intercalating cation would lead to a smaller shift in the flat band potential of the TiO2. Thus, the flat band potential of the TiO2 electrode will increase with increasing size of the cation. In the case of bulky cations, which cannot penetrate into the dye molecular layer, or intercalate, the cations cannot practically screen the negatively charged TiO2 particles within the size of the dye molecule, and thus, the polyiodide anion (I3-) as an electron acceptor will experience a repulsive force to penetrate the dye layer.29 This causes a drop in the back electron (recombination) transfer and hence increases the photocurrent. However, the imidazolium salt, with flat morphology, can penetrate into the dye adsorbed layer and adsorb at the TiO2 surface, thereby reducing the electron lifetime and hence Jsc.30 3.2.2. Surface Adsorption on the TiO2. In the liquid electrolyte DSSC systems, the mechanism of the charge transport is known to occur in two different ways, electron hopping in the TiO2 film and by the Grotthuss mechanism in the case of polyiodides,31 which can be extended in the polymer electrolytes as well. Therefore, the diffusion of iodides (or polyiodides) will decide the Jsc. As calculated by Son et al.,32 as the size of the cation will be larger, it will have a larger charge density including its solvated cloud. Thus, the movement of the I- will become slower under such large charge density, causing smaller diffusion and hence a decrease in Jsc. In contrast, for smaller cations, the corresponding solvated ion will be smaller, allowing faster movement of iodides, and hence, the current density increases. These cations, on the other hand, may get adsorbed on the negatively charged TiO2 particles and form a neutral cluster. Thus, the band at the TiO2-cation interface will bend toward positive energy values, and hence, the flatband potential (Vfb) (28) Wang, H.; Bell, J.; Desilvestro, J.; Bertoz, M.; Evans, G. J. Phys. Chem. C 2007, 111, 15125. (29) Nakade, S.; Kanazaki, T.; Kambe, S.; Wada, Y.; Yanagida, S. Langmuir 1995, 21, 11414. (30) Dupont, J.; Suarez, P. A. Z. Phys. Chem. Chem. Phys. 2006, 8, 2441. (31) Kopidakis, N.; Schiff, E. A.; Park, N. G.; van de Lagemaat, J.; Frank, A. J. J. Phys. Chem. B 2000, 104, 3930. (32) Son, K. M.; Kang, M. G.; Vittal, R.; Lee, J.; Kim, K. J. J. Appl. Electrochem. 2008, 38, 1647.
will shift toward positive. A similar negative shift in the Vfb has also been reported due to the adsorption of fluoride ions at the TiO2 surface and penetration into the dye layer.33 Park et al.34 have shown with rutile TiO2 electrodes that, the smaller the cation, the smaller the solvated ion cloud, the faster the diffusion, and hence the larger the corresponding DSSC parameters. Redmond and Fitzmaurice35 have also been able to distinguish the cation size effect in a series of nonaqueous protic and aprotic solvents. According to their investigations, the flatband potential, in either of the cases, depends on the electrolyte cation and thereby affects the cell performance. Though, the formation of a space charge region, resulting in the band bending, necessitates a larger region of semiconductor (i.e., bulk). However, for nanosized semiconductor particles having surface adsorbed cations, this bending will be localized and more sensitive to the electrolyte environment than the semiconductor.27 Thus, the thickness and charge distribution of the Helmholtz layer will vary and hence will result in the cation-dependent potential gradient at the semiconductor surface. As suggested by Murray et al.,36 the adsorption dominates at lower concentrations and intercalation at higher concentrations of the cations. Therefore, the possibility of adsorption of the cations is more in our systems, where the concentration of the cations is less than 10-1M (assuming complete dissolution of the salt in polymer). Also, the free carriers due to these adsorbed ions maintain the electrical neutrality at the electrode-electrolyte interface. In order to maintain the charge neutrality, both the ions must codiffuse into the TiO2. The hindrance of diffusion of cations into the TiO2 layer due to their physical and electrostatic diameter will cause a decrease in the photocurrent. This effect will be pronounced with the ionic diameters of the adsorbent. The adsorption, on the other hand, may also increase the acceptor (I3-) concentration at the interfaces through electrostatic forces. The adsorption-induced acceptor level shifting has been considered as the responsible factor in the LiI/I2 acetonitrile system by Kelly et al.37 for the electron injection at the interface that depends upon the concentration of the cation. This will, in turn, increase the probability of the back electron transfer as a recombination process and hence lower the current. (33) (a) Cheng, X. F.; Leng, W. H.; Liu, D. P.; Xu, Y. M.; Zhang, J. Q.; Cao, C. N. J. Phys. Chem. C 2008, 112, 8725. (b) Cooper, G.; Turner, J. A.; Nozik, A. J. J. Electrochem. Soc. 1982, 129, 1973. (34) Park, N. G.; Chang, S. H.; van de Lagemaat, J.; Kim, K. J.; Frank, A. J. Bull. Korean Chem. Soc. 2000, 21, 985. (35) Redmond, G.; Fitzmaurice, D. J. Phys. Chem. 1993, 97, 1426. (36) Murray, D. J. ; Healy, T. W.; Fuerstenau, D. W. AdVances in Chemistry; Gould, R. F., Ed.; American Chemical Society: Washington, DC, 1968; Vol. 97, Chapter 7. (37) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Stipkala, J. M.; Meyer, G. J. Langmuir 1999, 15, 7047.
Solid Polymer Electrolyte Based DSSC
Figure 6 schematically shows the positive shifting of the bands under any of the above conditions and hence modification in the flatband potential Vfb to its new value Vfb′. The conduction band edge of the TiO2 will shift if it is affected (by either the adsorption or by intercalation of the cations) to its new position with respect to its earlier position (without the influence of any cation). Thus, the bending at the FTO/TiO2 or FTO/TiO2 (+ cation) will get modified, and accordingly, the Voc should change. However, in our case, the number of cations interacting (in any of the ways discussed above) is almost the same (Section 3.1 and Figure 2); therefore, the change in Voc is negligible. If the concentration of the cations is varied, the effects will be more pronounced and a clear shift in Voc could be observed. Change in Voc with increasing cation concentration has been reported by Watson and Meyer.38 The movement of the semiconductor flatband potential changes the degree of overlap between the sensitizer excited-state distribution function and the density of the semiconductor acceptor states. Thereby, the rate of electron injection gets modified and hence so do the cell parameters, viz., the photocurrent and efficiency. This explanation fits well with our measured values for the alkali and ammonium salts. For the ionic liquid containing systems, the behavior is different, which can be attributed to the nonfeasibility of the intercalation by the EMI+ or HMI+ due to their larger size. (38) Watson, D. F.; Meyer, G. J. Coor. Chem. ReV. 2004, 248, 1391.
Langmuir, Vol. 25, No. 5, 2009 3281
4. Conclusions The conductivity of the PEO:PEG blend solid polymer electrolyte complexed with various iodide salt is shown to depend on the size of the cation of the salt. The change in conductivity is found to be not due to the number of dissociated charge carriers but due to the change in crystallinity vis-a`-vis the mobility. The dye-sensitized solar cells fabricated using these solid polymer electrolytes showed better efficiency than those with only PEO. The solar cell parameters, mainly, the efficiency and short circuit current (Jsc), have been found to depend on the size of the cation. A model has been proposed to explain this cation dependency in terms of the intercalation and the adsorption of the cations at the polycrystalline TiO2 surface. Acknowledgment. This work was supported by the Brain Korea 21 (BK 21) project under the Ministry of Education, Science and Technology, Republic of Korea. Supporting Information Available: The DSC thermograms, polarized optical micrographs (POM), X-ray diffraction (XRD) of polymer electrolytes, the plots showing the variations in the Voc, and FF of the DSSC with the cation size. This material is available free of charge via the Internet at http://pubs.acs.org. LA8029177
