ARTICLE pubs.acs.org/Langmuir
Nanocrystalline Electrodes Based on Nanoporous-Walled WO3 Nanotubes for Organic-Dye-Sensitized Solar Cells Kohjiro Hara,*,† Zhi-Gang Zhao,† Yan Cui,† Masahiro Miyauchi,† Masanori Miyashita,‡ and Shogo Mori*,‡ † ‡
National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan Shinshu University, 3-15-1 Tokida, Ueda, Nagano 386-8567, Japan ABSTRACT: Nanoporous-walled tungsten oxide (WO3) nanotubes (NTs), which had a more positive conduction band edge level compared to that of TiO2, were applied to various organic dyes for dye-sensitized solar cells (DSSCs). The dye-sensitized WO3 NTs displayed photosensitization for the organic dyes whose lowest unoccupied molecular orbital (LUMO) level was relatively positive to the conventional TiO2 electrode and, thus, not applicable for electron injection to the TiO2 electrode. Electron transport time and electron lifetime for the WO3 electrode in the DSSCs were investigated. In comparison to the DSSCs based on TiO2, SnO2, and In2O3, the WO3 DSSCs displayed the longest lifetime. On the other hand, non-diffusion-like electron transport may be an issue to apply WO3 for the DSSCs.
1. INTRODUCTION Over the past decade, dye-sensitized solar cells (DSSCs) have attracted much attention because these unconventional solar cells exhibit high performance and have the potential for low-cost production.16 Solar power-to-electricity conversion efficiencies higher than 11% under simulated AM 1.5 G irradiation have been attained with DSSCs based on Rucomplex sensitizers.7,8 One of approaches to improve the performance of DSSCs furthermore is to use photons of the near-infrared (IR) region. For example, phthalocyanines,9 naphthalocyanines,10 and organic dyes,1113 which absorb the near-IR region, have been developed as sensitizers for DSSCs. Generally, however, these dyes have relatively positive lowest unoccupied molecular orbital (LUMO) levels and/or negative highest occupied molecular orbital (HOMO) levels, causing potential mismatch with the conduction band edge level (Ecb) of the TiO2 electrode and/or iodine redox potential, resulting in lower solar-cell performance. To obtain a high current using near-IR dyes, for example, different semiconductor electrodes, whose Ecb is more positive than that of the conventional TiO2 electrode to ensure the potential difference needed for the charge injection from dye to semiconductor, are required. To increase the energy conversion efficiency, not only the current but also high voltage is needed. However, employing such more Ecb positive electrodes results in the decrease of open circuit voltage (VOC). This is because VOC is limited by the difference between the Ecb and redox potential of electrolytes. Therefore, replacing TiO2 with such an electrode in DSSCs simply decreases the energy conversion efficiency. A remedy of this issue is introducing a tandem r 2011 American Chemical Society
structure; that is, more than two sensitized solar cells are connected in series. Tungsten oxide (WO3), whose band gap (Eg) is ca. 2.8 eV, has been employed as material for the photocatalyst driven by visible light irradiation.1418 On the other hand, synthesizing small nanoparticles (NPs) of WO3 was not trivial, limiting its practical applications. Recently, several groups have published synthesis methods for WO3 NPs.19,20 Zhao and Miyauchi synthesized nanoporous-walled WO3 nanotubes (NTs) and investigated their photocatalytic activity for oxidation of acetoaldehyde under ultraviolet (UV) and visible light irradiation.21 The WO3 NTs showed higher photocatalytic activity compared to that of conventional WO3 particles. One of reasons for higher photocatalytic activity of WO3 NTs was suggested to be their larger surface area compared to that of conventional particles, where the NT walls consist of individual WO3 NPs, whose average diameter is ca. 50 nm. WO3 is also expected to be one of the candidates for DSSCs using near-IR dyes because its Ecb is about 0.5 V more positive than that of TiO2.22 However, only Ru-complex dyes have been applied to WO3.23,24 In this paper, we report photovoltaic performance of DSSCs based on WO3 NT electrodes and dyes having various LUMO levels and compare to conventional DSSCs based on TiO2 electrodes. Electron-transfer kinetics in the electrodes are also studied to assess WO3 for DSSCs. Another motivation is to compare the electron-transfer kinetics among various metal oxides. Received: May 3, 2011 Revised: August 4, 2011 Published: September 23, 2011 12730
dx.doi.org/10.1021/la201639f | Langmuir 2011, 27, 12730–12736
Langmuir
ARTICLE
Chart 1. Molecular Structures of the Organic Dyes
was loaded into a Teflon-lined autoclave, which was then sealed, maintained at 180 °C for 12 h, and then allowed to cool to room temperature. The resulting white precipitate was collected and rinsed several times with distilled water and absolute ethanol. It was then dried at 60 °C for 12 h. The solid products were heated from room temperature to 450 °C for 6 h and annealed at 450 °C for another 3 h to obtain WO3 NTs (yellow products). 2.3. Preparation of Photoelectrodes. Nanocrystalline photoelectrodes were prepared by a screen-printing technique. Organic WO3 pastes for screen printing were prepared from WO3 powder or WO3 NTs, with ethyl cellulose as a binder and α-terpineol as a solvent. The TiO2 paste or WO3 paste was printed on a glass substrate coated with transparent conducting oxide (TCO, F-doped SnO2, Asahi Glass Co.) and subsequently sintered at 500 °C in air for 1 h. The thickness of the thin films, measured with an Alpha-Step 300 profiler (Tencor Instruments), was 314 μm. Dyes were dissolved at a concentration of 0.3 mM in solvent, e.g., toluene for MK-2 and tert-butanol/AN (50:50 vol %) for NKX-2475 and NKX-2883. The semiconductor film electrodes were immersed in dye solution and then kept at 25 °C for at least 12 h to allow the dye to adsorb to the semiconductor surface.
2.4. Solar-Cell Fabrication and Photovoltaic Measurements. The sealed electrochemical cell used for photovoltaic measure-
Table 1. LUMO Level of the Dyes dye
LUMO (V) versus NHE
NKX-2813 NKX-2475
0.28 0.5725
NKX-2883
0.7226
MK-2
0.8928
2. EXPERIMENTAL SECTION 2.1. Materials. Reagent chemicals and materials, WCl6 (Kanto Chemicals), urea (Kanto), toluene (Kanto, dehydrated), acetonitrile (AN, Wako Pure Chemicals), tert-butanol (Tomiyama Pure Chemical Industries, Ltd.), 1,2-dimethyl-3-n-propylimidazolium iodide (DMPImI, Tomiyama), LiI (Tomiyama), I2 (Tomiyama), 4-tert-butylpyridine (TBP, Tomiyama), WO3 powder (High Purity Chemicals), and TiO2 paste for screen printing (T20/SP, Solaronix SA), were used without further purification. The molecular structures of organic dyes, NKX-2475, NKX-2883, NKX-2813, and MK-2, used in this work are shown in Chart 1. These dyes were selected to examine the effect of LUMO potential on the electron injection. The values are listed in Table 1. The detailed synthesis procedures for them are described elsewhere.2528 2.2. Prepareation of WO3 NTs. Detailed procedures of the preparation of nanoporous-walled WO3 NTs have been described elsewhere.21 Tungsten chloride (WCl6, 0.40 g, 1 mmol) and urea (0.6 g, 10 mmol) were added to 40 mL of absolute ethanol. The mixture
ments consisted of dye-sensitized WO3 NTs or TiO2 thin-film electrodes, a Pt-coated TCO electrode as a counter electrode, a spacer (Surlyn film, 30 μm thick), and an electrolyte. The counter electrode consisted of a Pt film (ca. 200 nm thick) sputtered onto a TCO-coated glass plate. We used two electrolytes for the solar cells composed of 0.6 M DMPImI, 0.1 M LiI, and 0.4 M I2, with and without 0.5 M TBP in AN. The photovoltaic performance of the solar cells was measured with a source meter (Advantest, R6243). We employed an AM 1.5 G solar simulator (Wacom Co., WXS-80C-3 with a 300 W Xe lamp and an AM filter) as the light source. The incident light intensity was calibrated using a standard solar cell composed of a crystalline silicon solar cell and an IR cutoff filter (Schott, KG-5), giving the photoresponse range of an amorphous silicon solar cell (produced and calibrated by the Japan Quality Assurance Organization). To avoid the penetration of diffuse light into the active dye-sensitized film, a black mask with an aperture area of 0.2354 cm2 was employed to measure the photovoltaic performance. Action spectra of the monochromatic incident photon-tocurrent conversion efficiency (IPCE) of the DSSC were measured with a CEP-99W system (Bunkoh-keiki). 2.5. Electron Lifetime and Diffusion Coefficient. Electron lifetimes (τ) and diffusion coefficient of the DSSCs were obtained by stepped light-induced transient measurements of photovoltage. Here, the lifetime is the average time spent in the semiconductor after the injection from the dyes and before the recombination with the dye cation or I3. Electron diffusion coefficients here are apparent values obtained from the time to travel in the semiconductor to reach the transparent conducting. Experimental procedures of the measurements are described in detail elsewhere.29 In short, for the lifetime, DSSCs were irradiated by a diode laser (Coherent, Lablaser, λ = 635 nm) and the decay of open-circuit voltage (Voc), caused by a stepwise decrease of a small fraction of the laser intensity, was recorded. The Voc drop was typically less than 1 mV. The measurement was repeated with various initial laser intensities, giving different electron densities in the DSSCs. The values of τ was obtained by fitting an exponential function, exp(t/τ), to the voltage decay. For diffusion coefficients, the same measurements are repeated at short-circuit conditions and current transients were fitted with a single exponential function. The electron density in the DSSCs was estimated by the charge extraction method introduced by Peter and co-workers under the same initial light intensity used for the lifetime measurements.30 12731
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
Langmuir
ARTICLE
Figure 1. SEM images of (a and b) WO3 NTs and (c and d) WO3 NPs.
3. RESULTS AND DISCUSSION 3.1. Scanning Electron Microscope (SEM) Analyses. SEM images of nanoporous-walled WO3 NTs and commercialized WO3 NPs are shown in Figure 1. WO3 NTs have outer diameters of 3001000 nm (inner diameter is 300600 nm) and lengths of 25 μm (Figure 2a). The particle size of the commercialized WO3 NPs is 50500 nm. As shown in Figure 2, the surface of the WO3 NTs is not smooth, indicating that the NTs are polycrystalline. As already reported, the WO3 NTs consist of individual WO3 NPs, whose diameter is 50100 nm, resulting in a nanoporous-walled structure.21 The BrunauerEmmettTeller (BET) surface area of WO3 NPs and WO3 NTs is 4.4 and 25 m2 g1, respectively.21 3.2. Sensitization of WO3 NT Electrodes. Figure 2 shows action spectra of IPCE for DSSCs based on the WO3 NP and NT electrodes (ca. 5.5 μm thickness) sensitized with MK-2, NKX2813, and NKX-2883. The IPCE, which corresponds to the external quantum efficiency, is given by
IPCE ð%Þ ¼
1240 ðeV nmÞJph 100 λΦ
ð1Þ
where Jph (mA cm2) is the short-circuit photocurrent density obtained under monochromatic irradiation and λ (nm) and Φ (mW cm2) are the wavelength and intensity of the monochromatic light, respectively. The electrodes based on WO3 NPs and WO3 NTs themselves showed photoresponse from 300 to 480 nm with a peak at 353 nm, which is derived from direct photoexcitation of the WO3 electrodes (Figure 2a). When organic dyes were adsorbed on the WO3 electrodes, photocurrent because of absorption by the organic dyes was observed from 400 to 850 nm, showing that electron transfer from the dyes to the WO3 electrodes occurs. It is remarkable that even the NKX-2813, whose LUMO is located at 0.28 V versus NHE, showed sensitization (Figure 2c). If dyes with a HOMO level of about 1 V and a LUMO level of 0.28 V are sensitized, the WO3 electrode is capable of collecting electrons generated by up to 1000 nm of light. The IPCE values of the DSSCs based on the WO3 NT electrode sensitized were higher than that of the DSSC based on the WO3 NP electrode (panels a and b of Figure 2), probably because of the larger surface area of WO3 NTs. On the other
Figure 2. IPCE spectra for DSSCs based on WO3 electrodes with electrolyte A: (a) (black - - -) WO3 NP electrode, (black —) WO3 NP electrode sensitized with MK-2, (red - - -) WO3 NT electrode, and (red —) WO3 NT electrode sensitized with MK-2, (b) (black —) WO3 NP electrode with NKX-2883 and (red —) WO3 NT electrode with NKX-2883, and (c) (black —) TiO2 NP electrode sensitized with NKX-2813 and (red —) WO3 NT electrode sensitized with NKX-2813.
hand, the degree of the increase was not as high as expected on the basis of the difference in the surface area. This is probably due to the relatively large portion of the pore consists of less than 4 nm size,21 where the organic dyes (∼3 nm) may hardly penetrate into the pore. In addition, WO3 NT electrodes appeared to be white. This suggests that a large portion of incoming light is scattered, resulting in shallow light penetration depth. The highest IPCE value reached 48% at 470 nm for the DSSC based on the WO3 NT electrode and MK-2 (Figure 2a). Light-harvesting efficiency (LHE) of the dye-sensitized electrode can be estimated by LHE ¼ 1 T ¼ 1 10A
ð2Þ
where T is transmittance and A is absorbance. The LHE of the MK-2-sensitized WO3 electrode at 470 nm was ca. 55% (data not 12732
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
Langmuir
ARTICLE
Table 2. Photovoltaic Performance of DSSCs Based on WO3 and TiO2 Nanoelectrodes and Organic-Dye Sensitizersa η
Jsc
Voc
(mA cm2)
(V)
FF
(%)
MK-2 MK-2
4.03 5.93
0.34 0.29
0.21 0.29
0.29 0.50
4.3
NKX-2883
5.39
0.33
0.43
0.76
7.4
NKX-2475
1.45
0.32
0.40
0.19
thickness electrode
(μm)
WO3 NP WO3 NT
5.3 4.8
WO3 NT WO3 NT
dye
WO3 NT
14
NKX-2813
1.52
0.30
0.39
0.15
TiO2 T20
14
NKX-2813
0.059
0.36
0.68
0.015
NKX-2475
0.39
0.41
0.69
0.11
TiO2 T20
5.6
Light, AM 1.5 G (100 mW cm2); electrolyte, 0.6 M DMPImI + 0.1 M LiI + 0.05 M I2 + 0.5 M TBP in AN. a
Figure 3. (a) Absorption spectrum of NKX-2475 adsorbed on a TiO2 film electrode (6 μm thickness) and (b) IPCE spectra for DSSCs based on NKX-2475: (blue —) TiO2 electrode (0 M TBP), (blue - - -) TiO2 electrode (0.5 M TBP), (red —) WO3 NT electrode (0 M TBP), and (red - - -) WO3 NT electrode (0.5 M TBP).
shown). The IPCE is also given by IPCE ¼ LHEϕinj ηc
ð3Þ
where ϕinj is the quantum yield of electron injection from the dye to the semiconductor electrode and ηc is the efficiency of the collection of the injected electrons at the back contact. The values of IPCE and LHE for the DSSC based on the MK-2sensitized WO3 NT electrode indicate high ϕinj and ηc in the system according to eq 3. If LHE can be increased more, a higher IPCE is expected. 3.3. Comparison to DSSCs Based on TiO2. For NKX-2813, its LUMO level (0.28 V versus NHE) is lower than the Ecb of TiO2 (about 0.5 V versus NHE). Thus, no sensitization is expected for TiO2.31 Actually, Figure 2c shows that, when NKX2813 was employed, little sensitization was observed with TiO2, while the WO3 NT showed up to 850 nm. For the case of dyes, NKX-2475, having a comparable LUMO level to TiO2 Ecb, TiO2 was sensitized for the half of the absorption spectrum, while the WO3 electrode used the full spectrum. Figure 3a shows an absorption spectrum of NKX-2475 adsorbed on a nanocrystalline TiO2 electrode (6 μm thickness), and Figure 3b shows IPCE spectra for DSSCs based on a NKX-2475-sensitized TiO2 electrode and a WO3 NT electrode (7 μm thickness). The NKX-2475 adsorbed on the TiO2 electrode showed two peaks at 440 and 580 nm. From the measured absorbance and eq 2, the LHE of the NKX-2475-sensitized TiO2 electrode at 440 and 580 nm was estimated to be 95 and 97%, respectively. The IPCE values at 440 and 580 nm for the DSSC based on the TiO2 electrode and NKX-2475 (electrolyte with 0 M TBP) were 12 and 2.8%, respectively (Figure 3b). On the basis of the high LHE
Figure 4. Effect of the TBP concentration on photovoltaic parameters for DSSC based on (a) MK-2 and (b) NKX-2883: (black b) η, (red b) Jsc, and (blue b) V oc. The electrolyte was 0.6 M DMPImI, 0.1 M LiI, and 0.05 M I2, with a changing TBP concentration in AN.
values of the NKX-2475-sensitized TiO2 electrode, lower IPCE values are probably due to lower ϕinj. When an electrolyte containing 0.5 M TBP shifts Ecb of the TiO2 electrode negatively, both IPCE values decreased to 6 and 0.4%, respectively (Figure 3b). This also indicates that electron transfer from NKX-2475 to the TiO2 electrode is the limiting factor. In contrast, for the DSSC based on the WO3 NT electrode and NKX-2475, the IPCE value higher than 10% was observed at 575 nm, whereas the LHE was 48% (data not shown). 3.4. Solar Cell Performance. The photovoltaic performance of DSSCs based on WO3 and TiO2 electrodes are summarized in Table 2. The low values of Voc for the DSSCs based on WO3 electrodes are due to the positive Ecb of WO3 but not due to faster 12733
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
Langmuir
ARTICLE
Figure 6. Double logarithmic plot of the electron lifetime (τ) in the DSSCs based on TiO2 and WO3 NP and NT electrodes sensitized with MK-2 as a function of the electron density in the electrodes. Data for the TiO2 are from ref 37. Figure 5. IV characteristics of dye-sensitized WO3 NT solar cells under 1 sun conditions. The dye was NKX-2883. The electrolyte was 0.6 M DMPImI, 0.1 M LiI, 0.05 M I2, and 0.7 M TBP in AN. The electrode thickness was 4.3 μm. The energy conversion efficiency was 0.8%.
recombination, as discussed later. In other words, the low Voc is the intrinsic property of WO3. The reason for low fill factor (FF) is not clear, but it seems to be related to the position of the conduction band edge, as discussed later. Among the dyes, coumarin dyes showed relatively high FF. In comparison to TiO2-based cells, the energy conversion efficiency of the WO3-based cells was low. This is mostly due to the intrinsic low Voc of the WO3-based cells. Our WO3 cells also showed relatively low Jsc. As mentioned above, this is probably a particular issue of our WO3 NT electrode, which has lightscattering property. Normally, the light-scattering layer is placed at the back side of the electrodes, so that penetrated light can scatter back to the electrode. For our WO3 electrode, light is scattered all over the place of the electrodes, so that some incident light is reflected back out of cell before absorption. This also suggested that Jsc was not increased by increasing the thickness of the WO3 electrode. Besides MK-2, the nonoptimized design of the organic dyes for electron injection, e.g., the location of CN group, would be another reason. 3.5. Effect of TBP on the Performance of DSSCs Based on WO3 NT Electrodes. Pyridine compounds, such as TBP, have been used as additives in electrolytes for DSSCs.2,3235 TBP can improve the open-circuit voltage of the solar cell, and the improvement is partially due to the negative shift of Ecb of the TiO2 electrode. TBP is useful to optimize the energy difference between LUMO and Ecb of the semiconductor. Here, we investigated the effect of TBP if there is room to improve the photovoltaic performance of DSSCs based on the WO3 NT electrode. This engineering method is important when the difference between the LUMO of dye and Ecb of the semiconductor is redundantly large.35 The change in the performance parameters for DSSCs with MK-2 and NKX-2883 by changing the TBP concentration in the electrolyte is shown in panels a and b of Figure 4, respectively. In the case of the DSSC with MK-2, the values of both Jsc and Voc increased with the increase of the TBP concentration (Figure 4a).36 Consequently, the total conversion efficiency increased from 0.1 to 0.6%. The value of Voc for the DSSC based on NKX-2883 also increased as the TBP concentration was increased (Figure 4b). In contrast, the value of Jsc for the DSSC with NKX-2883 decreased as the TBP concentration was increased. The decreased Jsc was most likely due to too much negative shift of Ecb for NKX-2833, which has a
Figure 7. Values of Voc as a function of the electron density for the DSSCs with TiO2 and WO3 NP and NT electrodes sensitized with MK-2. Data for the TiO2 are from ref 37.
more positive LUMO level (0.72 V versus NHE26) than that of MK-2 (0.89 V versus NHE28). For the DSSCs with NKX-2883, the concentration was optimized and the highest efficiency of the WO3 solar cell was obtained with 0.7 M TBP. Figure 5 shows the IV curves under 1 sun conditions. 3.6. Electron Lifetime for the DSSCs with the WO3 Electrode. Figure 6 shows a double logarithmic plot of the electron lifetime (τ) in the DSSCs based on the WO3 NP and NT electrodes sensitized with MK-2 as a function of the electron density in the electrodes. The electrode thickness was around 5 μm for both electrodes. Data for TiO2 were obtained from a previous paper.37 Both DSSCs showed the decrease of τ with the increase of the electron density. This has been partially modeled with transport/trap-limited recombination, where the recombination rate is limited by the probability of electrons hitting the interface of the semiconductor and electrolytes.38,39 In the model, most trapped electrons do not recombine because there is no hole in the semiconductor and tunneling probably most trapped electrons to the acceptor level in the electrolyte seems to be low. The DSSCs based on WO3 showed a 2 orders of magnitude longer lifetime than that with TiO2. This can be partially explained with the much smaller potential difference, that is, driving force, between the Ecb and potential of I3. Figure 7 shows the values of Voc as a function of the electron density for the DSSCs based on the WO3 NP and NT electrodes sensitized with MK-2. The Voc is determined by the difference between the Fermi level of semiconductor electrodes and the redox potential of I/I3. Between two electrodes at the matched electron density, the Fermi level of the semiconductor with respect to the conduction 12734
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
Langmuir
ARTICLE
advantages of WO3. To exploit the properties of WO3, one simple strategy is to increase the roughness factor. Another would be to add the dielectric layer to screen the WO3 electrode.
Figure 8. Apparent electron diffusion coefficients of the DSSCs based on TiO2 and WO3 NP and NT electrodes sensitized with MK-2. Data for the TiO2 are from ref 37. Note that D values were obtained from the time constant of current transients. The transport in WO3 is probably not due to diffusion; therefore, the values are plotted only to compare to TiO2, but they do not represent actual diffusion coefficients.
band edge is the same if the intraband trap density and distribution are the same. From Figure 7, the Ecb of the WO3 electrode is roughly estimated to be 0.4 V more positive in comparison to the TiO2 Ecb. By the same measurements for SnO2- and In2O3-based DSSCs, the Ecb of SnO2 and In2O3 were about 0.3 and 0.4 V versus TiO2 Ecb.40,41 Thus, the electron lifetime in the DSSCs based on WO3 could be similar to that in the DSSCs based on In2O3. However, the observed lifetime with DSSCs/WO3 was an order of magnitude longer. Because the two different WO3 electrodes (NT and particle) showed a longer lifetime, the behavior is probably related to the intrinsic characteristics of the material, which could be different dielectric constants, different mobilities, etc. Between the WO3 NT and NP, the NT electrode showed a longer electron lifetime than the WO3 NP electrode. This implies a higher density of the trap site in the WO3 NT electrode compared to the WO3 NP electrode because of its higher surface area. A similar relationship was seen among the DSSCs using different particle size TiO2 NPs.42 In Figure 7, the DSSC with the WO3 NP electrode showed a higher Voc of more than 0.1 V compared to that of the DSSC based on the WO3 NT electrode, at the matched electron density. The lower Voc for NT-based solar cells suggests higher trap density, in agreement with the interpretation for the lifetime difference. Note that the porosity is not taken into account to calculate the density. Because the NT appeared to have higher porosity (see Figure 1), the Voc difference at the matched electron density would become even larger. Transient current measurements were also performed for the DSSCs/WO3 showing light intensity independent behavior (Figure 8). The much longer time constant for NT, that is, lower values of D, appears to be consistent with a higher trap density and/or number of boundaries. On the other hand, the intensity independence is very different from TiO2, SnO2, and ZnO but similar to In2O3.41 At this moment, we consider that the transport is not totally governed by diffusion for the WO3 electrode. Thus, let us emphasize that the D values in Figure 8 for WO3 were to compare to that for TiO2 but do not show actual diffusion coefficients. We then suspect that the low FF would be related to the non-diffusion transport properties.41 Therefore, before employing the WO3 films for highly efficient DSSCs, these issues should be clarified. However, the high internal quantum efficiency for the low LUMO potential dyes and very long electron lifetime observed with WO3 are the superior characteristics that are not obtained by TiO2, displaying the
4. CONCLUSION For the dyes having very positive LUMO levels, sensitization was not seen with TiO2, while WO3 displayed in the configuration of DSSCs. This shows the clear advantage of WO3 when low LUMO level dyes, typically absorbing up to near IR, are needed. High IPCE performance of the DSSCs based on WO3 was obtained when a WO3 NT electrode was employed. The highest IPCE value reached 48% at 470 nm for the DSSC based on a WO3 NT electrode and an organic dye named MK-2. Electron injection from the NKX-2475 dye, which has a relatively positive LUMO level, to the WO3 NT electrode was suggested to occur effectively compared to the TiO2 electrode because of the positive Ecb of the WO3 NT electrode. The addtion of TBP in the electrolyte was effective to control the Ecb of WO3 to optimize the balance between the injection efficiency and Voc of the DSSCs, as well as in the case of TiO2-based DSSCs. The DSSC based on the WO3 NT electrode showed a much longer electron lifetime than that for the DSSC with TiO2. This is another advantage of WO3, while the electron transport in WO3 seems to be a problem to be solved for highly efficient WO3based DSSCs. ’ AUTHOR INFORMATION Corresponding Author
*Telephone: +81-29-861-4638. Fax: +81-29-861-4638. E-mail:
[email protected] (K.H.);
[email protected] (S.M.).
’ ACKNOWLEDGMENT We acknowledge financial support from the New Energy and Industrial Technology Development Organization (NEDO) of Japan. ’ REFERENCES (1) O’Regan, B.; Gr€atzel, M. Nature 1991, 353 (6346), 737. (2) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; M€uller, E.; Liska, P.; Vlachopoulos, N.; Gr€atzel, M. J. Am. Chem. Soc. 1993, 115 (14), 6382–6390. (3) Smestad, G.; Bignozzi, C.; Argazzi, R. Sol. Energy Mater. Sol. Cells 1994, 32 (3), 259–272. (4) Kay, A.; Gr€atzel, M. Sol. Energy Mater. Sol. Cells 1996, 44 (1), 99–117. (5) Hagfeldt, A.; Gr€atzel, M. Chem. Rev. 1995, 95, 49–68. (6) Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000, 33 (5), 269–277. (7) Nazeeruddin, M. K.; De Angelis, F.; Fantacci, S.; Selloni, A.; Viscardi, G.; Liska, P.; Ito, S.; Takeru, B.; G€rvzel, M. J. Am. Chem. Soc. 2005, 127 (48), 16835–16847. (8) Chiba, Y.; Islam, A.; Watanabe, Y.; Komiya, R.; Koide, N.; Han, L. Jpn. J. Appl. Phys. 2006, 45 (2428), L638–L640. (9) He, J.; Benk€o, G.; Korodi, F.; Polívka, T.; Lomoth, R.; Åkermark, B.; Sun, L.; Hagfeldt, A.; Sundstr€om, V. J. Am. Chem. Soc. 2002, 124 (17), 4922–4932. (10) Li, X. Y.; Long, N. J.; Clifford, J. N.; Campbell, C. J.; Durrant, J. R. New J. Chem. 2002, 26 (8), 1076–1080. (11) Matsui, M.; Hashimoto, Y.; Funabiki, K.; Jin, J. Y.; Yoshida, T.; Minoura, H. Synth. Met. 2005, 148 (2), 147–153. (12) Burke, A.; Schmidt-Mende, L.; Ito, S.; Gr€atzel, M. Chem. Commun. 2007, No. 3, 234–236. 12735
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
Langmuir (13) Otsuka, A.; Funabiki, K.; Sugiyama, N.; Mase, H.; Yoshida, T.; Minoura, H.; Matsui, M. Chem. Lett. 2008, 37 (2), 176–177. (14) Darwent, J. R.; Mills, A. J. Chem. Soc., Faraday Trans. 2 1982, 78, 359–367. (15) Erbs, W.; Desilvestro, J.; Borgarello, E.; Graetzel, M. J. Phys. Chem. 1984, 88 (19), 4001–4006. (16) Sclafani, A.; Palmisano, L.; Marci, G.; Venezia, A. M. Sol. Energy Mater. Sol. Cells 1998, 51 (2), 203–219. (17) Abe, R.; Sayama, K.; Sugihara, H. J. Phys. Chem. B 2005, 109 (33), 16052–16061. (18) Abe, R.; Takami, H.; Murakami, N.; Ohtani, B. J. Am. Chem. Soc. 2008, 130 (25), 7780–7781. (19) Hahn, R.; Macak, J. M.; Schmuki, P. Electrochem. Commun. 2007, 9 (5), 947–952. (20) Sadek, A. Z.; Zheng, H. D.; Breedon, M.; Bansal, V.; Bhargava, S. K.; Latham, K.; Zhu, J. M.; Yu, L. S.; Hu, Z.; Spizzirri, P. G.; Wlodarski, W.; Kalantar-zadeh, K. Langmuir 2009, 25 (16), 9545–9551. (21) Zhao, Z. G.; Miyauchi, M. Angew. Chem., Int. Ed. 2008, 47 (37), 7051–7055. (22) Scaife, D. E. Sol. Energy 1980, 25, 41–54. (23) Sayama, K.; Sugihara, H.; Arakawa, H. Chem. Mater. 1998, 10 (12), 3825–3832. (24) Zheng, H. D.; Tachibana, Y.; Kalantar-zadeh, K. Langmuir 2010, 26 (24), 19148–19152. (25) Hara, K.; Sato, T.; Katoh, R.; Furube, A.; Ohga, Y.; Shinpo, A.; Suga, S.; Sayama, K.; Sugihara, H.; Arakawa, H. J. Phys. Chem. B 2003, 107 (2), 597–606. (26) Wang, Z. S.; Cui, Y.; Dan-Oh, Y.; Kasada, C.; Shinpo, A.; Hara, K. J. Phys. Chem. C 2008, 112 (43), 17011–17017. (27) Koumura, N.; Wang, Z. S.; Mori, S.; Miyashita, M.; Suzuki, E.; Hara, K. J. Am. Chem. Soc. 2006, 128 (44), 14256–14257. (28) Wang, Z. S.; Koumura, N.; Cui, Y.; Takahashi, M.; Sekiguchi, H.; Mori, A.; Kubo, T.; Furube, A.; Hara, K. Chem. Mater. 2008, 20 (12), 3993–4003. (29) Nakade, S.; Kanzaki, T.; Wada, Y.; Yanagida, S. Langmuir 2005, 21 (23), 10803–10807. (30) Duffy, N. W.; Peter, L. M.; Rajapakse, R. M. G.; Wijayantha, K. G. U. Electrochem. Commun. 2000, 2 (9), 658–662. (31) When an electrolyte containing only LiI and I2 were used with TiO2, the TiO2 Ecb is shifted positively. However, the degree of the shift is expected up to 0.3 V (Supporting Information in ref 33), and thus, the shifted Ecb is not enough to have a 0.2 V difference between the LUMO of the dye and the TiO2 Ecb. The 0.2 V is the empirical value to obtain a high injection yield. This is also seen for the combination of NKX-2475 and TiO2 electrode in Figure 3. (32) Liu, Y.; Hagfeldt, A.; Xiao, X. R.; Lindquist, S. E. Sol. Energy Mater. Sol. Cells 1998, 55 (3), 267–281. (33) Kusama, H.; Konishi, Y.; Sugihara, H.; Arakawa, H. Sol. Energy Mater. Sol. Cells 2003, 80 (2), 167–179. (34) Hara, K.; Dan-Oh, Y.; Kasada, C.; Ohga, Y.; Shinpo, A.; Suga, S.; Sayama, K.; Arakawa, H. Langmuir 2004, 20 (10), 4205–4210. (35) Haque, S. A.; Palomares, E.; Cho, B. M.; Green, A. N. M.; Hirata, N.; Klug, D. R.; Durrant, J. R. J. Am. Chem. Soc. 2005, 127 (10), 3456–3462. (36) Interestingly, for the case of MK-2, Jsc was increased with an increasing TBP concentration. The Voc was also increased, suggesting the negative shift of Ecb of the electrode. Thus, this phenomenon cannot be explained by the potential difference dependent injection yield. Another possible effect of TBP would be to increase the dyecation reduction rate by changing the reorganization energy of the dye. However, substantiation of such speculation is beyond the scope of this paper. (37) Miyashita, M.; Sunahara, K.; Nishikawa, T.; Uemura, U.; Koumura, N.; Hara, K.; Mori, A.; Abe, T.; Suzuki, E.; Mori, S. J. Am. Chem. Soc. 2008, 130 (52), 17874–17881. (38) Nelson, J. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59 (23), 15374–15380. (39) Kopidakis, N.; Benkstein, K. D.; van de Lagemaat, J.; Frank, A. J. J. Phys. Chem. B 2003, 107 (41), 11307–11315.
ARTICLE
(40) Fukai, Y.; Kondo, Y.; Mori, S.; Suzuki, E. Electrochem. Commun. 2007, 9 (7), 1439–1443. (41) Mori, S.; Asano, A. J. Phys. Chem. C 2010, 114 (30), 13113–13117. (42) Tsekouras, G.; Miyashita, M.; Kho, Y. K.; Teoh, W. Y.; Mozer, A. J.; Amal, R.; Mori, S.; Wallace, G. G. IEEE J. Sel. Top. Quantum Electron. 2010, 16 (6), 1641–1648.
12736
dx.doi.org/10.1021/la201639f |Langmuir 2011, 27, 12730–12736
