TiO₂ surface engineering with multifunctional oligomeric polystyrene coadsorbent for dye-sensitized solar cells

Abstract

Oligomeric, hydrophobic coadsorbents based on polystyrene (o-PS, M_n = 2600) terminated by a carboxylic acid exhibit dual functions in dye-sensitized solar cells (DSCs): suppression of electron recombination at the TiO₂ surface, and enhanced concentration of the strongly-anchored dye, N719, which has two carboxylic acid groups. Engineering the TiO₂ surface via o-PS results in the concurrent and significant enhancement of photovoltage and photocurrent, consequently increasing the energy conversion efficiency of DSCs by as much as 28.7%. The electron recombination rate was largely reduced via the blockage of vacant sites with o-PS chains on the TiO₂ surface due to the physical hindrance to I₃⁻s in electrolyte. In addition, the formation of the o-PS:I₂ charge transfer complex at the photoanode/electrolyte interface lessened the effective concentrations of free I₃⁻ and/or I₂ for electron recombination. Upon sequential o-PS coadsorption, the concentration of the strongly-anchored dyes on the TiO₂ surface was increased via deprotonation of the weakly-anchored dyes, giving rise to an increase in the electron injection efficiency and, subsequently, the overall power conversion efficiency. The dual functions of the o-PS coadsorbent have been therefore demonstrated to increase the overall efficiency of DSCs.

---

Introduction

Dye-sensitized solar cells (DSCs) exhibit high efficiency in operation and manufacture, and show great potential in the commercialization of organic photovoltaics.^1,2 Many attempts have been made towards the improvement of photovoltaic performance in DSCs, such as the development of highly-efficient photoanode nanostructures,^3–7 dyes,^8–12 alternative redox electrolytes,^13–16 and the tuning of the electrode/electrolyte interface.^17–25 In operating DSCs, the dyes adsorbed on a semiconducting layer such as TiO₂ under solar irradiation become oxidized to inject electrons in the TiO₂ conduction band (CB) state, subsequently passing through the semiconducting layer to a transparent conducting oxide (TCO) substrate of the charge collector. Meanwhile, the oxidized dyes are regenerated by reduced species such as I⁻s in the redox electrolyte, which are regenerated following transport to the counter electrode. It is worth noting that individual electrons are intercepted by oxidized redox species such as I₃⁻s in the redox electrolyte, while transporting of electrons through the TiO₂ layer and subsequently to the TCO substrate. This electron loss, referred to as electron recombination, can significantly deteriorate the photovoltaic parameters of open-circuit voltage (V_OC), short-circuit current density (J_SC), and fill factor (FF), thus attenuating energy conversion efficiency (η). Electron recombination is therefore considered to be a key factor in determining the photovoltaic performance of DSCs.²

In principle, electron recombination with I₃⁻ in the electrolyte occurs at the interfaces of electrolytes with a TiO₂ nanostructure and TCO layer. Electron recombination through the TCO layer could be largely reduced via the formation of a thin blocking layer made of TiO₂. However, there has been a large body of research on the suppression of electron recombination through the surface of the TiO₂ photoanode. A typical approach is exemplified by use of nano-thin insulating oxides layers such as Nb₂O₅ (ref. 17) and Al₂O₃ (ref. 20) on the photoanode structure, or electrolyte additives such as 4-tert-butylpyridine (tBP),²⁶ guanidinium thiocyanate (GuSCN)²⁷ and nitrogen-containing derivatives.²⁸ More recently, it was reported that breakthroughs could be achieved by creating a barrier via an increase in steric bulk, such as alkyl chains of the dye.¹³ Another method of suppressing electron recombination is to incorporate coadsorbents that are adsorbable onto the TiO₂ surface for repairing leakage sites.^{18,19,21,22,25} The coadsorbents typically contain a carboxylic or phosphonic acid functional group for adsorption on the surface of the TiO₂ layer, and feature a rather bulky structure such as chenodeoxycholate¹⁸ to physically deactivate the TiO₂ empty sites. An additional function of coadsorbents is that intermolecular interaction between organic dye molecules could be attenuated, preventing dye aggregation.¹⁸ In some instances, the TiO₂ CB edge may also shift upward or downward with respect to the dipole moment of the coadsorbents, which in turn determines the electron injection from excited dyes, as well as the open circuit voltage.^18,19 Recently, our group has reported the application of poly(ethylene glycol) (PEG)-based oligomeric, hydrophilic coadsorbent (M_w = 2000) containing rather long ethylene oxide units to obtain energetically- and kinetically-favorable DSCs.²¹

In this work, we introduce oligomeric, hydrophobic polystyrene (o-PS, M_n = 2600) terminated by a carboxylic acid as a coadsorbent on a TiO₂ surface with N719 ruthenium dye (Scheme 1). The hydrophobic o-PS coadsorbent can be firmly attached to TiO₂ surface via an ester bond of the COOH group. Compared to the small molecular coadsorbents previously reported, our coadsorbent has rather long hydrophobic o-PS chains. Their incompatibility with the acetonitrile (ACN)-based electrolyte may lead to the formation of a bulky collapsed-chain structure which was closely tangled on the TiO₂ surface offering effective coverage for passivation. Physical hindrance by the collapsed chains can be given to access of I₃⁻s in electrolyte to TiO₂ empty surface. Furthermore, o-PS-containing benzene rings can chemically decompose I₃⁻ in the electrolyte, forming a charge-transfer complex of PS–I₂. The I₂s trapped in the collapsed o-PS chains can be markedly increased, thereby restricting the access of I₃⁻s to the TiO₂ surface. In DSCs, thus, such physical, chemical interruptions to the I₃⁻ approach were considerably effective for doubly-extended electron lifetime, and subsequently yielded a 28.7% rise in energy conversion efficiency. Furthermore, the hydrophobic, bulky chains in the o-PS coadsorbent may cause the water molecules to repel from the TiO₂ surface, which could in turn facilitate deprotonation of the carboxylic acid in the dye, as confirmed by our UV-vis spectra.

Scheme 1 (a) Synthetic route for monocarboxyl-terminated, oligomeric, o-PS coadsorbent (o-PS, M_n = 2400). (b) Schematic blockage by o-PS chains on empty sites of the TiO₂ surface and complex formation to reduce the effective concentrations of free I₃⁻ and I₂ for electron recombination.

Results and discussion

When hydrophilic TiO₂ nanoparticles are coated with o-PS coadsorbent containing one carboxylic group at one chain end to form TiO₂/o-PS core–shell microspheres, their surface becomes hydrophobic.²⁹ Fig. 1 presents the surface wetting property of a TiO₂/dye electrode passivated both with and without o-PS treatment. The results show that the TiO₂/dye/o-PS film exhibits less affinity toward water (25° → 44°), i.e. inducing a water repelling property in comparison to the neat o-PS free film. Thus, when the TiO₂ surface is modified with hydrophobic o-PS it becomes less probable for water to contact with the dyes adsorbed to it, thereby suppressing a possible breakage of the ester bond between the TiO₂ and dye. The detachment of the dyes adsorbed on the TiO₂ surface is suppressed when compared to the film without the o-PS coadsorbent treatment.

Fig. 1 Photographs of the water contact angle on (a) the TiO₂/dye electrode and (b) TiO₂/dye electrode passivated by o-PS coadsorbent. UV-vis spectra of various photoelectrodes in the absence of dye (c) to verify formation of o-PS:I₂ complex following solvent evaporation (inset: photographs of TiO₂ films both with and without o-PS after I₂ solution dropping and solvent evaporation under ambient condition) and (d) the blue shift of the dye band affected by o-PS coadsorbent (inset: enlarged spectra).

It is well known that the electron “donor–acceptor” complex, or the charge transfer complex, can be formed between a phenyl group (donor) of o-PS and iodine (acceptor), phenyl group + I₂ ↔ phenyl:I₂.³⁰ Likewise, formation of this o-PS:I₂ complex occurs in the current system, as confirmed by the UV-vis spectra shown in Fig. 1(c). The TiO₂/I₂ and TiO₂/o-PS/I₂ substrates were first prepared by dropping 0.3 mM I₂ solution in ACN onto the TiO₂ or TiO₂/o-PS substrate. The UV-vis spectra were obtained after allowing the solvent to evaporate for 10 min. The TiO₂/I₂ substrate did not produce any notable change in the measured spectra when compared to the neat TiO₂, suggesting no complex formation of I₂ with TiO₂ and its subsequent instant sublimation during the solvent evaporation. However, it is intriguing that a new, broad absorption band appeared in the TiO₂/o-PS/I₂ substrate, likely due to the formation of the o-PS:I₂ complex, suggesting the presence of I₂ even after evaporation of the solvent. The maximum peak was positioned at 360 nm and the absorption band was extended to 550 nm, whereas light absorption by TiO₂ or o-PS rarely occurs up to 400 nm. This finding is consistent with the color change revealed in the images in the inset of Fig. 1(c). Both substrates were initially stained by the 0.3 mM I₂ solution to exhibit a yellow color; however, this color became faint within a few minutes in the pristine TiO₂ film at ambient conditions, and eventually disappeared due to I₂ sublimation (left image). In contrast, the initial yellow color remained nearly unchanged in the TiO₂/o-PS film (right image). These results suggest the formation of the o-PS:I₂ complex between the electron donating phenyl rings in the o-PS coadsorbent and the electron-accepting iodine. Collectively, it is suggested that the concentrations of I₃⁻ and/or I₂ are lower at the TiO₂ surface because of the complex formation between o-PS and I₂.

Another characteristic feature is that deprotonation of the dyes attached to the TiO₂ film was greatly facilitated by the o-PS coadsorption. In principle, two COOH groups in the N719 dye can chemically react with OH groups on the TiO₂ surface resulting in a strongly-anchored dye. However, the N719 dye, which has one unbound COOH group and is a weakly-anchored dye, may also be formed in a TiO₂/dye film depending on factors such as the pH, types of solvent, and presence of water.^2,24 One representative property caused by converting a weakly- to a strongly-anchored dye is the band shift associated with π–π* and dπ–π* charge transfers, thus yielding the blue shift in UV-vis spectra.³¹ As seen in the UV-vis spectra of Fig. 1(d), the two absorption bands for the N719 dye were blue-shifted from 394 to 386, and from 539 to 532 nm via the inclusion of the o-PS coadsorbent to the TiO/N719 dye film. These results suggest that the blue-shift of the absorption bands is likely ascribed to the deprotonation of the dyes in the conversion of weakly to strongly-anchored dye.³¹ A slight detachment of the N719 dyes (∼5%) occurs during the o-PS passivation process, regardless of the o-PS concentration (see Table 1). Due to the slight solubility of N719 dyes in toluene, the weakly-anchored dye was partially removed.

Table 1 Characteristic parameters for relative dye loading amount, work function of the photoanodes, and photovoltaic performance parameters of DSCs

Concentration of coadsorbent [mM]	Relative dye load	Work function [eV]	VOC [V]	JSC [mA cm^−2]	JSC/(relative dye load) [mA cm^−2]	FF	η [%]
0	1	5.34	0.66	9.1	9.1	0.73	4.4
0.1	0.96	5.34	0.67	10.7	11.1	0.72	5.0
0.3	0.95	5.34	0.72	10.8	11.4	0.73	5.7
0.5	0.95	5.34	0.75	9.2	9.7	0.73	4.9
0.7	0.95	5.34	0.78	6.2	6.5	0.76	3.7

---

The o-PS coadsorption can give rise to a displacement in the CB edge of the TiO₂ nanoparticulate layer, chiefly impacting electron injection, charge collection, and the TiO₂ quasi-Fermi level under working DSC.^2,18 The upward or downward shift in the TiO₂ CB edge may be mirrored at a work function of the photoanode, as measured via photoelectron spectroscopy.²¹ The results listed in Table 1 indicate that the work function of the TiO₂ photoanode remains unchanged regardless of the addition of o-PS coadsorbent. This finding is unexpected, because the deprotonation of N719 dye tends to compel a downward shift of the TiO₂ CB edge, while the potential coordinative interactions³² between the TiO₂ surface and the electron donating phenyl rings in the o-PS may derive an upward shift of the TiO₂ CB edge. Therefore, it is plausible to suggest that the downward shift of the TiO₂ CB edge via deprotonation is compensated for by the upward shift due to the coordinative interaction.

DSCs were fabricated by varying the o-PS concentration in toluene within a range of 0.1 to 0.7 mM. The photocurrent–voltage (J–V) characteristic curves of DSCs shown in Fig. 2 were evaluated under 1 sun illumination (AM 1.5, 100 mW cm⁻² with shading mask). Even in the presence of the small detachment of N719 during the o-PS coadsorption, a J_SC of 10.7 mA cm⁻² was attainable at 0.1 mM o-PS, which is significantly enhanced relative to that of the pristine case (9.1 mA cm⁻²). By further increase in the o-PS concentration, the J_SC approached a peak of 10.8 mA cm⁻² at 0.3 mM o-PS, then dropped sharply to 6.2 mA cm⁻² at 0.7 mM o-PS. A similar, but more marked concentration dependency was observed in the η, while the V_OC steadily rose without fluctuation alongside the o-PS concentration; the FF remained unchanged. With optimal conditions (0.3 mM PS) DSC resulted in improvement in V_OC of 60 mV and J_SC of 1.7 mA cm⁻², which in turn causes an enhancement of 23% for η up to 5.7%, as summarized in Table 1. When compared at a fixed dye concentration, J_SC considerably enhanced by as much as 2.3 mA cm⁻² was yielded. Thus, once the dye desorption during o-PS coadsorption is prevented, there is further room for DSC efficiency improvement.

Fig. 2 Photovoltaic action spectra of DSCs incorporating o-PS coadsorbent. (a) Photovoltaic J–V curves measured under 1 sun illumination (AM 1.5G, 100 mW cm⁻² with light-shading mask). (b) Representative IPCE spectra.

The J_SC is integration of incident-photon-to-current conversion efficiency (IPCE) spectra measured over the entire solar light range. In the IPCE spectra shown in Fig. 2(b), the quantum efficiency at the 0.3 mM o-PS coadsorbent is increased over that of the reference. Intriguingly, the IPCE spectra are quite blue-shifted, corresponding to the deprotonation effects previously verified by the blue-shifted UV band (Fig. 1(d)). In fact, the quantum efficiency is closely associated with charge collection and electron injection, as well as light harvesting and dye regeneration, where the latter two seem to have marginal effects in this current case. The high IPCE may be due both to increased charge injection and collection efficiencies. It is worthy to note that the concentration increase in the strongly-anchored dyes due to the deprotonation effects may substantially reinforce electronic coupling between the dye and TiO₂, thereby promoting electron injection efficiency.²⁴ Charge collection efficiency, which can be characterized by interfacial charge transfer at the dye/electrolyte, is discussed in the next section.

The interfacial charge transfer occurring between dyes and electrolytes in DSCs can be evaluated in terms of the electron lifetime, which can be determined via the stepped light-induced measurement of photocurrent and voltage transient (SLIM-PCV) method.²¹ The electron lifetime shown in Fig. 3(a) offers an insight that the TiO₂ film modified by o-PS under the ACN-based electrolyte efficiently suppress the electron recombination reaction. Specifically, 2- and 5-fold elongations in the electron lifetime were caused by 0.3 and 0.5 mM o-PSs, respectively, indicating that the o-PS passivation layer becomes more effective in suppressing the electron recombination with increasing o-PS concentration. This effect was further confirmed via the impedance spectra (Fig. 3(b)) measured under dark condition, in which the electron lifetime was improved with results similar to those achieved via the SLIM-PCV method. One explanation is that o-PS adsorbed on the TiO₂ surface has two functions in suppressing the electron recombination: steric hindrance via the collapsed o-PS chains, and reduction in the I₃⁻ concentration via formation of the o-PS:I₂ complex. Thus, the o-PS passivation layer could lead to higher charge collection efficiency, consequently improving both J_SC and V_OC (see Table 1). However, the excessive passivation layer prepared at o-PS concentrations over 0.5 mM may provide a shield for the dyes to some extent, which can significantly restrict the charge transfer reaction, or dye regeneration, subsequently leading to a marked deterioration in J_SC with o-PS concentrations over 0.5 mM.

Fig. 3 Electron lifetime results of representative DSCs attained via (a) the SILM-PCV method, and (b) EIS spectra, respectively.

Experimental section

Oligomeric polystyrene containing one COOH group

Oligomeric polystyrene was synthesized via atomic transfer radical polymerization (ATRP) and terminated with a carboxylic acid at the chain end (o-PS). The number-average molecular weight (M_n) of the o-PS coadsorbent was evaluated to be 2.6k according to gel permeation chromatography (GPC), and the radius of gyration (R_g) of the o-PS chain in toluene was estimated at approximately ∼2 nm. Since the o-PS coadsorbent is not soluble in ACN/tert-butanol used for dye dissolution, toluene was used as a solvent. Concentration of the o-PS coadsorbent in toluene was varied from 0.3 to 0.7 mM.

DSC fabrication

A mesoporous TiO₂ layer (12 μm thick with an average TiO₂ particle size of 20 nm, Solaronix Ti-Nanoxide T) was prepared on FTO substrates (TEC 8, Pilkington) following the formation of a TiO₂ compact layer and was sintered at 450 °C for 30 min. The mesoporous TiO₂ layer was subsequently immersed in 0.3 mM bis tetrabutylammonium cis-dithiocyanatobis (4,4′-dicarboxylic acid-2,2′-bipyridine)ruthenium(II) (N719) dye solution in a mixture of ACN and tert-butanol (1 : 1 v/v) solution for 18 h at room temperature, and the solvent was evaporated to dryness at room temperature. The electrodes were then dipped in toluene-dissolving o-PS coadsorbents for 18 h at room temperature, and were subsequently washed with toluene to remove the unreacted o-PS.

The electrolyte was composed of 0.5 M 1-methyl-3-propyl imidazolium iodide (MPII), 0.05 M iodine (I₂) dissolved in ACN. Common additives such as lithium iodide (LiI), tBP, or GuSCN were not required to clearly characterize the coadsorbent effect. The Pt counter electrode was prepared via spin-coating 0.01 M of H₂PtCl₆ in isopropyl alcohol solution, followed by sintering at 450 °C for 30 min. DSCs were then produced by sandwiching a TiO₂/N719 dye photoanode and a Pt counter electrode with Surlyn (25 μm, Solaronix) as a spacer; an electrolyte solution was then injected into the predrilled hole of the Pt counter electrode via vacuum backfilling. Finally, the holes were heat-sealed with a small piece of Surlyn and a cover glass.

Characterization methods

The absorption behavior of the o-PS coadsorbent, I₂, and N719 dye in toluene and on the TiO₂/N719 photoanode was detected via UV-vis spectroscopy (V-670 UV-vis spectrophotometer, JASCO), and the relative dye loading amounts on TiO₂ photoanodes were calculated upon the addition of o-PS coadsorbent. The surface wetting properties of the TiO₂ layer passivated with the o-PS coadsorbent, which was characterized via water-contact angle tests (DSA100, KRÜSS). The TiO₂ film thickness and active area were measured with a surface profiler (α-step IQ, TENCOR) and optical microscope (EGTECH). Current–voltage characterization of DSCs was performed using a Keithley model 2400 source meter and solar simulator, with a 300 W Xenon arc-lamp (Newport) under 1 sun illumination (AM 1.5G, 100 mW cm⁻²), in which the light intensity was calibrated via a silicon solar cell (PV Measurements, Inc.). A light-shading mask was used on the front side of the FTO substrate, excluding the 0.16 cm² TiO₂ active area, for preventing overestimation of the power conversion efficiency. In addition, the quantum efficiency of the DSC was analyzed using an IPCE instrument (PV Measurements, Inc.). For analysis of CB-edge shifts, the work function of the TiO₂/dye photoanode both with and without the o-PS coadsorbent was evaluated via photoelectron spectroscopy using AC2 (RKI Instruments, Inc., Japan) under ambient conditions. The lifetime of electrodes on the TiO₂ photoanode was measured using the SLIM-PCV method and impedance spectroscopy.

Conclusions

Oligomeric, monocarboxyl-terminated polystyrene (o-PS, M_n = 2.6k) can serve as a coadsorbing agent to suppress electron recombination and increase charge injection efficiency, resulting in an increased energy conversion efficiency of as much as 28.7% due to a concurrent improvement in V_OC and J_SC. The suppressed electron recombination, verified by the elongated lifetime, is primarily attributable to the effective blockage of empty sites on the TiO₂ surface, which is a result of the collapsed o-PS chains and reduced concentration of both I₃⁻ and I₂, caused by complexing between o-PS and I₂. The increase in the charge injection efficiency may stem from the increased concentration of the strongly anchored dye by deprotonation effects occurring during the dipping process of the o-PS coadsorbent. We therefore believe that a good avenue for designing coadsorbent molecules for use in photovoltaic devices is presented herein, and with further optimization, state-of-the-art devices can be attainable.

Acknowledgements

This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korea Center for Artificial Photosynthesis (KCAP) (No. 2009-0093883), the National Creative Research Initiative Center for Intelligent Hybrids (No. 2010-0018290) and the BK21 Plus Program funded by the Ministry of Education, Science, and Technology (MEST) of Korea.

References

1. B. O'Regan and M. Gratzel, Nature, 1991, 353, 737–740.
2. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
3. M. Law, L. E. Greene, J. C. Johnson, R. Saykally and P. Yang, Nat. Mater., 2005, 4, 455–459.
4. G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese and C. A. Grimes, Nano Lett., 2005, 6, 215–218.
5. B. Liu and E. S. Aydil, J. Am. Chem. Soc., 2009, 131, 3985–3990.
6. A. Birkel, Y.-G. Lee, D. Koll, X. V. Meerbeek, S. Frank, M. J. Choi, Y. S. Kang, K. Char and W. Tremel, Energy Environ. Sci., 2012, 5, 5392–5400.
7. S. H. Ko, D. Lee, H. W. Kang, K. H. Nam, J. Y. Yeo, S. J. Hong, C. P. Grigoropoulos and H. J. Sung, Nano Lett., 2011, 11, 666–671.
8. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, F. E. CurchodBasile, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, K. NazeeruddinMd and M. Grätzel, Nat. Chem., 2014, 6, 242–247.
9. T. Bessho, S. M. Zakeeruddin, C.-Y. Yeh, E. W.-G. Diau and M. Grätzel, Angew. Chem., Int. Ed., 2010, 49, 6646–6649.
10. H. Choi, C. Baik, S. O. Kang, J. Ko, M.-S. Kang, M. K. Nazeeruddin and M. Grätzel, Angew. Chem., 2008, 120, 333–336.
11. J.-H. Yum, P. Walter, S. Huber, D. Rentsch, T. Geiger, F. Nüesch, F. De Angelis, M. Grätzel and M. K. Nazeeruddin, J. Am. Chem. Soc., 2007, 129, 10320–10321.
12. M. K. Nazeeruddin, F. De Angelis, S. Fantacci, A. Selloni, G. Viscardi, P. Liska, S. Ito, B. Takeru and M. Grätzel, J. Am. Chem. Soc., 2005, 127, 16835–16847.
13. S. M. Feldt, E. A. Gibson, E. Gabrielsson, L. Sun, G. Boschloo and A. Hagfeldt, J. Am. Chem. Soc., 2010, 132, 16714–16724.
14. M. Wang, N. Chamberland, L. Breau, J.-E. Moser, R. Humphry-Baker, B. Marsan, S. M. Zakeeruddin and M. Grätzel, Nat. Chem., 2010, 2, 385–389.
15. T. Daeneke, T.-H. Kwon, A. B. Holmes, N. W. Duffy, U. Bach and L. Spiccia, Nat. Chem., 2011, 3, 211–215.
16. D. Song, M.-S. Kang, Y.-G. Lee, W. Cho, J. H. Lee, T. Son, K. J. Lee, S. Nagarajan, P. Sudhagar and J.-H. Yum, Phys. Chem. Chem. Phys., 2012, 14, 469–472.
17. F. Lenzmann, J. Krueger, S. Burnside, K. Brooks, M. Grätzel, D. Gal, S. Rühle and D. Cahen, J. Phys. Chem. B, 2001, 105, 6347–6352.
18. N. R. Neale, N. Kopidakis, J. van de Lagemaat, M. Grätzel and A. J. Frank, J. Phys. Chem. B, 2005, 109, 23183–23189.
19. Z. Zhang, S. M. Zakeeruddin, B. C. O'Regan, R. Humphry-Baker and M. Grätzel, J. Phys. Chem. B, 2005, 109, 21818–21824.
20. C. Prasittichai and J. T. Hupp, J. Phys. Chem. Lett., 2010, 1, 1611–1615.
21. Y.-G. Lee, S. Park, W. Cho, T. Son, P. Sudhagar, J. H. Jung, S. Wooh, K. Char and Y. S. Kang, J. Phys. Chem. C, 2012, 116, 6770–6777.
22. S.-H. Park, J. Lim, I. Y. Song, N. Atmakuri, S. Song, Y. S. Kwon, J. M. Choi and T. Park, Adv. Energy Mater., 2012, 2, 219–224.
23. H.-J. Son, X. Wang, C. Prasittichai, N. C. Jeong, T. Aaltonen, R. G. Gordon and J. T. Hupp, J. Am. Chem. Soc., 2012, 134, 9537–9540.
24. H. An, D. Song, J. Lee, E.-M. Kang, J. Jaworski, J.-M. Kim and Y. S. Kang, J. Mater. Chem. A, 2014, 2, 2250–2255.
25. D. Song, H. An, J. H. Lee, J. Lee, H. Choi, I. S. Park, J.-M. Kim and Y. S. Kang, ACS Appl. Mater. Interfaces, 2014, 6, 12422–12428.
26. G. Boschloo, L. Häggman and A. Hagfeldt, J. Phys. Chem. B, 2006, 110, 13144–13150.
27. C. Zhang, Y. Huang, Z. Huo, S. Chen and S. Dai, J. Phys. Chem. C, 2009, 113, 21779–21783.
28. H. Kusama, Y. Konishi, H. Sugihara and H. Arakawa, Sol. Energy Mater. Sol. Cells, 2003, 80, 167–179.
29. R. Ghosh Chaudhuri and S. Paria, Chem. Rev., 2011, 112, 2373–2433.
30. N. J. Rose and R. S. Drago, J. Am. Chem. Soc., 1959, 81, 6138–6141.
31. M. K. Nazeeruddin, S. M. Zakeeruddin, R. Humphry-Baker, M. Jirousek, P. Liska, N. Vlachopoulos, V. Shklover, C.-H. Fischer and M. Grätzel, Inorg. Chem., 1999, 38, 6298–6305.
32. K. Hadjiivanov, D. Klissurski, G. Busca and V. Lorenzelli, J. Chem. Soc., Faraday Trans., 1991, 87, 175–178.

---

Footnote

† These authors contributed equally to this work.

---