Portal de Periódicos da CAPES

Two Redox Couples are Better Than One: Improved Current and Fill Factor from Cobalt‐Based Electrolytes in Dye‐Sensitized Solar Cells

2014; Wiley; Volume: 4; Issue: 8 Linguagem: Inglês

10.1002/aenm.201301273

1614-6840

Jiayan Cong, Yan Hao, Licheng Sun, Lars Kloo,

Copper-based nanomaterials and applications

A tandem redox strategy is used in cobalt-based electrolytes. Co(bpy)32+/Co(bpy)33+ offers a high photovoltage at the photoanode, whereas the I−/I3− or Fc/Fc+ redox couples facilitate charge transfer at the counter electrode. Electron exchange in the electrolyte offers beneficial concentration gradients. The overall conversion efficiency is improved from 6.5% to 7.5%. Since Grätzel and O'Regan significantly improved the efficiency of dye-sensitized solar cells (DSCs) in 1991,1 large research efforts have been put towards realization of this promising technology. Although the recent improvement of the efficiency of DSCs, for a long time, the most efficient redox couple for DSCs has been the I−/I3− system.2-6 In 2010, the Hagfeldt and Sun groups combined an electrolyte based on metal-organic cobalt complexes with bulky organic dyes.7 The new strategy greatly improved the efficiency of DSCs and has opened the possibilities to reach conversion efficiencies well above 15%. The Grätzel group reported a record cell with an efficiency of 12.3% using a cobalt-based electrolyte in 2011.8 Since then, cobalt-complex redox systems have become the most promising type of redox system to replace the traditional I−/I3− system. However, there are still several problems that limit the efficiency of DSCs involving cobalt-based electrolytes. One of the most important challenges is mass-transport limitation. Because of the larger effective size of the cobalt complexes as compared to polyiodide species, the ion mobility of the cobalt complexes are considerably lower; in particular inside the mesoporous TiO2 film.9, 10 Recently, TiO2 films with a higher degree of porosity and larger pore size was proven to reduce the mass-transport problem.11, 12 However, for non-volatile solvent electrolytes, implicitly also of higher viscosity used in DSCs, mass transport is still a serious problem. Another significant challenge for cobalt-complex redox systems is the inefficient charge transfer at the counter electrode.10 Most recent studies have focused on finding new materials to replace the traditional catalytic Pt on the counter electrode. Graphene,13, 14 PEDOT15 and other materials, such as TiC,16 etc., have proven better performance than platinized fluorine-doped tin oxide (FTO) as counter electrode materials. Here, a tandem electrolyte system is used in an attempt to resolve some of the intrinsic problems arising from the use of cobalt-based electrolyte systems. Tandem redox systems have previously been used in DSCs,17, 18 and some of them have involved cobalt-complex redox systems in the electrolytes.19, 20 However, the presence of multiple redox species in the electrolyte makes it a complicated system to study and intertwined results to resolve. In 2006, Bignozzi et al. studied the tandem Co(DTB)32+/Fc (ferrocene) system (DTB = 4,4′-dimethyl-2,2′-bipyridine). They claimed that the photo-oxidized dye would be reduced/regenerated by the co-mediator Fc. The oxidized form (Fc+) could then be rapidly intercepted by the Co(II) in the Co(DTB)32+/Fc system.19 In 2010, Caramori et al. demonstrated that a combination of an [Fe(DMB)3]2+/[Fe(DMB)3]3+ co-mediator and a [Co(DTB)3]2+/[Co(DTB)3]3+ mediator could be used to improve the electron-collection efficiency in a DSC sensitized by a (thienylterpyridine)ruthenium complex (DMB = 4,4′-di-methyl-2,2′-bipyridine). The Fe-based redox system was assumed to mainly work on the TiO2 working-electrode side, whereas the cobalt-based redox system to work at the counter-electrode side.20 Here, a tandem-redox system has been carefully designed. By optimizing the redox species in the electrolyte, a cobalt-based redox couple, analogously to the Bignozzi systems, is expected to mainly operate as the dye-regenerating agent, and the other (tandem) redox system to mainly operate at the counter electrode. The major difference from previous tandem systems is building on the recent advances combining a relatively non-bulky cobalt complex redox system with a bulky, organic sensitizing dye.7, 8 If the expectations are realized, the new tandem redox electrolyte can maintain the high and desirable open-circuit voltage associated with cobalt-based redox systems and at the same time decrease the problematic charge-transfer resistance at the counter electrode without the need for a change of material. The new tandem redox system combined with an organic dye greatly improved the light-to-electricity conversion efficiency of the DSC devices from Bignozzi's system showing efficiencies of 0.7% to 7.5% in our system. The tandem redox system also so outperformed the DSCs containing the pure cobalt-based redox system. Moreover, the results obtained suggest that redox couples that have suitable potentials, but react inefficiently at the counter electrode, can be used in DSCs using a tandem strategy. The commonly used cobalt-based redox couple Co(bpy)32+/Co(bpy)33+ and I2 were added to generate a tandem redox electrolyte. Because I− can (and will) react with Co3+ in the electrolyte, initially I− was not added to the tandem redox system (Raman spectra were recorded to verify that this reaction takes place; more information can be found at the end of this paper). Four electrolytes CI0-CI3 were prepared. CI0 contains the cobalt-based redox system alone, and the composition used was 0.22 M Co(bpy)3(PF6)2, 0.05 M Co(bpy)3(PF6)3, 0.10 M LiClO4 and 0.20 M TBP in acetonitrile. The other three electrolytes CI1-CI3 also contained I2 of concentrations 0.001 M, 0.005 M and 0.010 M together with the components of CI0, respectively, forming three tandem redox electrolytes. The organic dye LEG4 (Figure 1) was used as sensitizer for all DSC devices. The photovoltaic properties of the DSCs assembled with the different electrolytes are shown in Table 1. Current density–voltage (J–V) curves are shown in Figure 2. The details of fabrication can be found in the Supporting Information. Under full Sun AM 1.5G illumination, all DSC devices containing the tandem redox electrolytes showed better short-circuit current density (Jsc), fill factor (ff), and efficiency (η) than those of the pure cobalt-based electrolyte CI0. Among the three tandem electrolytes, the DSCs containing the electrolyte CI1, initially containing 0.001 M I2, gave the highest conversion efficiency of 7.5%. However, a decrease in open-circuit voltage (Voc) was observed, when the concentration of I2 was increased. This decrease may be ascribed to the higher concentration of I− formed in the DSCs under working conditions. At a higher concentrations of I−, the probability/risk of iodide ions being involved in the regeneration/reduction of the oxidized dye at the working electrode increases. Considering the concentration dependence of the Nernstian equation it is also clear that the change in I− concentration will cause a more negative redox potential. These two aspects are likely to cause the observed decrease in Voc. The large increase in the Jsc and the ff was mainly caused by the facilitated charge-transfer reaction at the counter electrode effectuated by the I−/I3− co-redox couple (a detailed model for the mechanism of the tandem redox electrolyte is shown at the end of this communication). However, by controlling the concentration of I2 in the tandem redox electrolyte, the high DSC Voc can be maintained and, at the same time, the other photovoltaic properties can be optimized to offer a better over-all solar cell performance. The incident photon-to-electron conversion efficiencies (IPCEs) of the solar cells based on the different electrolytes are shown in Figure 3. No decrease in the IPCE was found in the range below 400 nm, which shows that the low concentration of I2 in the electrolyte did not shadow the dye in the DSC devices. From 300 nm to 700 nm, the IPCE of the DSCs containing the tandem redox electrolytes are higher than for the cobalt-only DSCs. This shows that the tandem redox electrolytes promote a high photon-to-electron injection into the TiO2 substrate. The IPCE spectra were recorded under lower light intensity than 1 Sun. The spectra show the IPCE values of individual devices, and consequently there may be some deviations from the photovoltaic characteristics shown in Table 1. In non-volatile solvents, such as 3-methoxypropionitrile (MPN), the advantages of tandem redox electrolytes will become more obvious. The photovoltaic properties of DSCs assembled with MPN-based electrolytes are shown in Table 2. The composition of CI0-MPN is 0.22 M Co(bpy)3(PF6)2, 0.05 M Co(bpy)3(PF6)3, 0.10 M LiClO4 and 0.20 M TBP in MPN, and CI3-MPN is analogous CI0-MPN albeit containing additional 0.010 M I2. The short-circuit current density, fill factor and conversion efficiency of DSCs containing the tandem redox electrolytes are all higher than for DSCs containing the pure cobalt-based electrolyte, especially the fill factor. The open-circuit voltage decrease of DSCs invoking the tandem redox electrolytes is less obvious than when using acetonitrile as electrolyte solvent. A possible reason is that the diffusion rate of I− in MPN is lower, which may lead to a lesser degree of involvement of I− in the regeneration/reduction of the oxidized dye. The low efficiency of non-volatile electrolyte-based solar cells is mainly caused by the poor mass-transport ability with respect to complex cobalt species. By further optimization, using a thinner TiO2 film and more mobile counterions, efficiencies up to at least 7.0% have been obtained in our lab using tandem redox electrolytes showing the overall necessity of cell optimization. Figure 4 shows the short-circuit current density of DSCs assembled with the CI0-MPN and the CI3-MPN electrolytes at different light intensities. From these results we can deduce that at illumination levels higher than 500 W m-2, DSCs with the commonly employed MPN- and cobalt-based electrolytes suffer from a significant current-density limitation. However, by using tandem redox electrolytes, the current-density limitations are partly overcome. Two possible reasons for the observed increase in the current density and fill factor can be identified. One is the enhanced charge transport in the tandem electrolytes, and the other is the lower charge-transfer resistance at the counter electrode surface. In order to test these possible explanations, DSCs containing the MPN-based electrolytes were exposed to further investigations (vide infra). Electrochemical impedance spectroscopy (EIS) was performed in order to study the interfacial charge-transfer processes in DSCs containing the electrolytes CI0-MPN and CI3-MPN. The Nyquist plots are shown in Figure 5. The small semicircle arc in the left (high-frequency) part of the Nyquist plot is typically attributed to the charge-transfer process (RCE) at the electrolyte/counter electrode interface. The semicircle arc in the middle is normally attributed to the recombination charge-transfer resistance (Rrec) at the TiO2/dye/electrolyte interface, and the semicircle arc in the right (low-frequency) part is ascribed to the diffusion process of ion transport in the electrolyte. The plot for DSCs containing the CI0-MPN electrolyte shows that the cells suffer from both mass-transport problems in the electrolyte and high change-transfer resistance at the counter electrode. When I2 is added, the resistance at the counter electrode is significantly decreased. A fitted electrochemical model shows that the charge-transfer resistance at the counter electrode decreases from 25 Ω to 1 Ω, but the recombination resistance at the TiO2/dye/electrolyte interface (≈40 Ω) and in the electrolyte (≈40 Ω) both remain unchanged. Based on the size of the semicircle arc to the right in the spectra, it can be concluded that the mass-transport abilities must be quite similar in both electrolytes. Therefore, it is unlikely that the current increase observed is caused by a decrease in electrolyte viscosity. These results show that the added I2 mainly has an effect on the reduction reaction at the counter electrode (the equivalent circuit used in the analyses is shown in the Supporting Information). Transient absorption spectroscopy (TAS) was performed in order to determine the regeneration rate constants of the dye LEG4 in contact with two different electrolytes (CI0-MPN 0.22 M Co(bpy)3(PF6)2, 0.05 M Co(bpy)3(PF6)3, 0.10 M LiClO4 and 0.20 M TBP in MPN; CI1-MPN 0.001 M I2 in CI0-MPN). The transient absorption kinetics of LEG4-sensitized TiO2 electrodes employing the two electrolytes are shown in Figure 6. The transient optical signal was observed at 700 nm, caused by the absorption by the oxidized dye LEG4, after the laser pulse excitation at 545 nm. As calculated from Equation 1, the regeneration efficiency is found to be 82% for the dye in contact with the electrolyte CI0-MPN, and 84% for the dye in contact with the electrolyte CI1-MPN. Details of the analysis of TAS results are given in the Supporting Information. The two electrolytes thus exhibit highly similar regeneration characteristics, both with respect to regeneration rates and efficiencies, which shows that differences in regeneration cannot be the main cause of the current increase in the cells. In this context, it should be noted that irrespective of the models used for the analysis of TAS data or numbers obtained therefrom, the main support for the above conclusion is obtained from the almost identical decay curves in Figure 6. A possible mechanism is suggested to describe the experimental results observed for the DSCs containing the tandem redox electrolytes in Figure 7. After optical excitation of dye, electrons are injected into the TiO2 conduction band. The dye turns into the oxidized state. The oxidized dye cations are expected to mainly react with Co2+ species in the electrolyte, which consequently form the Co3+ species. An important observation is the high voltage obtained from the DSCs containing the tandem redox electrolytes, since that strongly suggests the Co2+/Co3+ redox system to be the main tandem component involved in the dye cation reduction. The regeneration data also leads to the same conclusion. In a pure cobalt-based redox system, Co3+ would then be the species to become reduced at the counter electrode. However, in the tandem redox systems, a significantly lower charge-transfer resistance at counter electrode was observed. This clearly shows that the addition of I2 to the electrolyte facilitates the reduction process at the counter electrode. I2 is known to experience a faster reaction rate of reduction than Co(bpy)33+ on a platinized FTO surface. Therefore, the experimental results strongly suggest that I2 is the primary species reduced at the counter electrode. With the two observations mentioned above in mind, there must be a rapid electron exchange between Co3+ and I− in the electrolyte, which generates Co2+ and I3−. This makes the tandem redox system a regenerative system exploiting the advantages of the Co2+/Co3+ redox system at the working electrode and the I−/I3− couple at the counter electrode. The electron exchange process was simulated by adding I− into a cobalt redox solution. Raman spectra clearly showed the formation of I3− (Figure 8; the entire spectral range is given in the Supporting information). From the difference spectrum, the peak at about 115 cm−1 can be ascribed to the formation of polyiodides (viz. I3−) from I− and I2 (I2 is formed upon the reduction of Co3+). With respect to the cell function, I− will be generated at the counter electrode, and subsequently it will diffuse into the electrolyte towards the photoanode. In our tandem redox systems, the I− formed will on its way towards the photoanode experience a statistically large risk of encountering a Co3+ species on the way and thus be consumed in a redox reaction. Therefore, there will be a decrease in I− concentration going from the counter electrode towards the photoanode; a concentration gradient. The same effect can be expected for the Co3+ species formed upon dye regeneration at the photoanode, and a concentration gradient is expected to be generated from the photoanode towards the counter electrode. The mechanism of exchange (redox reaction between I− and Co3+) in the electrolyte is expected to generate the concentration gradients described in Figure 7 and it will persist as long as short-circuit conditions and illumination are maintained. It should be emphasized that the concentration gradients included in Figure 7 represents a reasonable hypothesis in the absence of detailed reaction dynamics between the tandem components. The turn-over time for a redox species in a cell under working conditions is in the range of seconds, and thus the steady-state gradient is expected to be generated very quickly. If too much I2 is added into the electrolyte, there is a larger probability (risk) that larger amounts of I− formed can diffuse to the TiO2 side and react with the oxidized dye. Moreover, the increase in I− concentration can also decrease the concentration of Co3+ on TiO2 side. From the simplified Nernst equation (Equation S1, Supporting information), it can be noted that an increase in Co3+ concentration (the oxidized species) makes the potential of the Co2+/3+ redox couple more negative. These two effects together are likely to cause the observed lower photovoltage at higher I2 concentrations in the electrolyte. However, by controlling the concentration of I2, the diffusion of I− can be limited and the high photovoltage characteristic of the Co2+/3+ redox couple can be maintained. This makes DSCs containing the tandem redox electrolyte much more efficient than those containing the pure cobalt-based redox electrolyte. The whole process is schematically visualized in Figure 7. Apparently, I−/I3− system is not the most ideal one to be used in a tandem redox system, because of the corrosive nature of I2. Therefore, the less corrosive redox couple ferrocene/ferrocenium cation (Fc/Fc+) was also investigated in an analogous tandem redox system. In this context it should be emphasized that the effective redox potential of a redox system can be significantly changed by the relative concentrations ratios employed for the reduced and oxidized components of a redox system. The photovoltaic characteristics of DSCs containing the Fc/Fc+-cobalt tandem redox electrolytes in MPN can be found in Table 3. According to the literature on Fc+ in DSC electrolytes, it is expected to experience serious recombination problems at the TiO2 surface. Therefore, in the presence of the same concentration of the Co2+/3+ redox system components, the decrease of the open-circuit voltage is more pronounced than for the corresponding I−/I3−-Co2+/3+ tandem electrolyte system. However, the DSCs containing the Fc0/+-Co2+/3+ tandem redox electrolyte also showed good results; an efficiency of 4.8% was achieved, which is similar to that of the I−/I3−-Co2+/3+ tandem one. In analogy to the iodine-based tandem systems, the ferrocene/ferrocenium redox system was introduced through the addition of the reduced form (Fc; immediately reacting with Co3+ in the electrolyte forming the oxidized form Fc+) of the redox couple to the Co2+/3+-containing electrolyte CF3-MPN. In conclusion, a tandem redox strategy has been investigated in cobalt-based electrolytes. Electrolytes with the I−/I3− or Fc/Fc+ redox coupled together with the Co(bpy)32+/Co(bpy)33+ system were designed and applied to DSCs. By optimizing the concentration of the redox species in the electrolytes, the DSCs with tandem redox electrolyte systems maintained the high open-circuit voltage characteristic for cobalt-based redox systems and, at the same time, the short-circuit current density, fill factor, and light-to-electricity conversion efficiency were improved. By further studies, the I−/I3− or Fc/Fc+ redox couples were concluded to mainly be active at the counter electrode, which significantly reduced the charge-transfer resistance lowering the over-all cell series resistance. The Co(bpy)32+/Co(bpy)33+ system was consequently concluded to mainly be active at the photoanode, thus retaining a high open-circuit voltage. By this strategy, the inefficient charge-transfer problem at the counter electrode repeatedly noted for cobalt-based electrolytes was successfully resolved. Moreover, the results suggest that redox couples that have suitable potentials, but display high counter electrode charge-transfer resistances can be utilized in DSCs using a tandem strategy by appropriate choice of the co-redox system. Broadening the spectrum of potentially useful redox couples that can be used in the DSCs may prove to become a strategy to also to solve the electrolyte mass-transport problem. Device fabrication procedures and analysis of electrochemical and spectroscopic data can be found in the Supporting Information. This work was supported by the Swedish Research Council, the Swedish Energy Agency and the Knut & Alice Wallenberg Foundation, and Natural Science Foundation of China (21120102036 and 91233201). As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.