Modulation of photovoltage in mesoscopic perovskite solar cell by controlled interfacial electron injection

Abstract

Controlled interfacial electron injection was investigated by surface modification of mesoporous TiO₂ film with insulating thin ZrO₂ layer in a mesoscopic CH₃NH₃PbI₃ perovskite solar cell. Open-circuit voltage (V_oc) was linearly increased with increasing the thickness of the surface ZrO₂ film from 863 mV for the bare TiO₂ to 988 mV for the surface modified TiO₂ with 2.2 nm-thick ZrO₂ layer (V_oc was 1030 mV for the mesoporous ZrO₂), which was due to decrease in electron injection by thickening the surface ZrO₂ layer. Change in electron injection was confirmed by photoluminescence spectra. Charge collection was diminished with the surface ZrO₂ layer, which might be related to the limited electron diffusion length in the solution-processed perovskite layer due to the reduced injection into TiO₂. Fill factor was reproducibly improved by surface modification, which is due to increased shunt resistance. By the controlled electron injection method, a power conversion efficiency was improved from 11.7% to 13.6% by introducing 1.1 nm-thick ZrO₂ layer on the TiO₂ surface.

---

1. Introduction

Perovskite solar cells have been proved to be a promising next-generation photovoltaic technology since the report on long-term durable perovskite solar cells in 2012.¹ Organometal halide perovskite materials with chemical formula of CH₃NH₃PbX₃ (X = Br or I) were first attempted as sensitizer in dye-sensitized solar cell structure in 2009,² which showed power conversion efficiency (PCE) of 3–4%. The low PCE in the sensitized structure was improved to 6.5% in 2011 by optimizing coating condition.³ Thanks to the pioneer works on perovskite solar cell, especially on the solid-state perovskite solar cells based on mesoporous TiO₂ (ref. 1) and Al₂O₃ (ref. 4) together with the confirmed long-term stability,^1,4 researches on perovskite solar cell has been intensively performed. Consequently, power conversion efficiencies (PCEs) of 18–20% were recently certified.^5,6 CH₃NH₃PbI₃ (MAPbI₃) material possess intriguing opto-electronic properties such as high absorption coefficient,^2,7–9 low reflectivity,^9–11 small exciton binding energy,¹² photo-induced giant dielectric constant,¹³ long charge diffusion length ranging from 1 μm for the solution-processed perovskite films^14,15 upto 10 μm for the millimeter scale single crystal,¹⁶ and high charge carrier mobility.¹⁷ Due to the balanced charge transport behavior and small exciton binding energy, MAPbI₃ can be applied to various types including mesoscopic structure and p–i–n or p–n junction planar structure.^18–21

Electron injection from MAPbI₃ to TiO₂ is obvious for the sensitization structure, where the MAPbI₃ nanodots were deposited locally on TiO₂ surface and the mesopores in the TiO₂ film were filled with hole transporting spiro-MeOTAD.¹ However, working mechanism may be different in case of the mesoscopic structure in which pores of the TiO₂ film is filled with MAPbI₃. Although charge separation was observed at both TiO₂ and spiro-MeOTAD junctions in the mesoscopic structure by transient laser spectroscopy and microwave photoconductivity measurements,²² it was argued that the dominant transport pathway would be the perovskite absorber even in the mesoscopic structure.²³ This indicates that electron injection is debatable in the mesoscopic structure with mesoporous TiO₂ film impregnated with MAPbI₃. We are thus motivated to study on the degree of electron injection in the mesoscopic structure by depositing the electron non-accepting ZrO₂ layer²⁴ on TiO₂ surface and controlling the thickness of ZrO₂ layer.

Here, we report on effect of the ZrO₂ layer deposited on TiO₂ surface on photovoltaic performance of the mesoscopic structured perovskite solar cell. In order to investigate electron injection behavior, the thickness of the ZrO₂ thin layer was varied from about 5 Å to 22 Å by changing the concentration of zirconium butoxide coating solution. Photovoltaic parameters were substantially changed with increasing the ZrO₂ layer thickness, where photovoltage increased with ZrO₂ layer thickness and eventually approached to that for the pure ZrO₂ mesoporous film. Our experimental observation proves that electron injection occurs at least in the interfacial MAPbI₃ being in contact with TiO₂ surface.

2. Experimental

2.1 Preparation of mesoporous TiO₂ films

Anatase TiO₂ nanoparticles (average diameter of about 50 nm) were hydrothermally synthesized by two step method.²⁵ The paste was prepared by mixing the TiO₂ nanoparticles with terpineol (Aldrich, 99.5%), ethyl cellulose (Aldrich, 46 cp) and lauric acid (Fluka, 96%). The nominal composition of TiO₂ : terpineol : ethylcellulose : lauric acid was about 1.25 : 6 : 0.9 : 0.1. Fluorine-doped tin oxide (FTO) glass (Pilkington, TEC-8, 8 Ω sq⁻¹) was cleaned by UV/ozone treatment and sonication with ethanol. FTO surface was covered with a thin compact TiO₂ layer prior to deposition of the mesoporous TiO₂ film by spin-coating the 0.15 M titanium diisopropoxide bis(acetylacetonate) (75 wt%, Aldrich) solution in 1-butanol at stepwise spin rate of 700 rpm for 8 s, 1000 rpm for 10 s, and 2000 rpm for 40 s, which was followed by drying at 125 °C for 5 min and then finally annealed at 500 °C for 15 min. The nanocrystalline TiO₂ paste was deposited on the compact TiO₂ layer using spin-coating technique at the rate of 2000 rpm for 20 s and subsequently annealed at 550 °C for 1 h in a muffle furnace to obtain mesoporous TiO₂ film. Thickness of the annealed mesoporous TiO₂ layer was measured to be 300 ± 20 nm by an alpha-step IQ surface profiler (KAL Tencor). The annealed TiO₂ film was immersed in the 20 mM TiCl₄ (Aldrich) aqueous solution at 70 °C for 10 min. The substrates are washed with distilled water and dried under air flow, which was followed by annealing at 500 °C for 30 min.

2.2 Formation of ZrO₂ thin layers on TiO₂ surface

A 2-propanol solutions of 0.01 M, 0.02 M, and 0.04 M zirconium(IV) butoxide (Sigma-Aldrich, 80 wt% in 1-butanol) were prepared in argon atmosphere to be used as the precursor solution. 20 μL of the precursor solution was spin-coated on the annealed mesoporous TiO₂ film at spin rate of 4000 rpm for 20 s. The film was then dried under N₂ flow. After drying at room temperature, the film was annealed at 500 °C for 20 min in air. The thickness was controlled by varying precursor concentration and the number of spin coating/annealing cycles.

2.3 Preparation of mesoporous ZrO₂ film

ZrO₂ nanoparticles were prepared according to the method reported elsewhere.²⁶ Zirconium(IV) propoxide solution (51.43 mL, 70% in propanol) was mixed with acetic acid (13.5 mL) under argon condition, to which H₂O (214.29 mL) was added under vigorous stirring at room temperature. Colorless precipitate was formed. After stirring for about 1 hour, the propanol was removed by evaporation using a rotatory evaporator at 80 °C for 30 min, to which HNO₃ (8.57 mL, assay 68–70%) was added and then the solution was aged for about 12 h under stirring at room temperature. A homogeneous and transparent yellowish liquid was observed. The solution was diluted with 19.24 mL of H₂O, which was transferred to a Teflon liner autoclave and heated at 230 °C for 12 h. The resultant was washed with EtOH several times. The prepared ZrO₂ nanoparticles were mixed with terpineol, ethyl cellulose and lauric acid to make a viscous paste with nominal composition of ZrO₂ : terpineol : ethyl cellulose : lauric acid = 1 : 6 : 0.3 : 0.1. Mesoporous ZrO₂ film was prepared on the compact TiO₂ layer by spin coating technique, where 2.5 g of the ZrO₂ paste was diluted with 10 mL of EtOH. 0.5 mL of the diluted solution was spin-coated at 2000 rpm for 20 s, which was heated at 550 °C for 1 h. The annealed ZrO₂ film was not post-treated with TiCl₄. The thickness of ZrO₂ film was adjusted to about 300 nm.

2.4 Fabrication of perovskite solar cells

MAPbI₃ was deposited by two-step costing method.²⁵ The 20 μL of 1 M solution of PbI₂ in N,N-dimethylformamide (DMF, >99% Sigma) was first spin-coated on the mesoporous TiO₂ films with and without surface modification and the ZrO₂ film at 3000 rpm for 5 s and 6000 rpm for 5 s. The film was dried consecutively at 40 °C for 3 min and 100 °C for 5 min. In the second step, a 200 μL CH₃NH₃I solution (7 mg of CH₃NH₃I in 1 mL 2-propanol) was spin-coated onto the dried PbI₂ film at 4000 rpm for 30 s and then dried at 100 °C for 5 min to form MAPbI₃, where a loading time of 20 s was required prior to spinning. Spiro-MeOTAD hole transporting layer (HTL) was formed on the perovskite capping layer by spin-coating 20 μL of spiro-MeOTAD solution at 4000 rpm for 30 s. The composition of the spiro-MeOTAD solution was composed of 0.06 M spiro-MeOTAD (72.3 mg of spiro-MeOTAD in 1 mL of chlorobenzene), 0.198 M 4-tert-butylpyridine (TBP) and 64 mM Li-TFSI (520 mg Li-TFSI in 1 mL acetonitrile (Sigma-Aldrich, 99.8%)). Au electrode was finally formed on the HTL by thermal evaporation of gold under 1.0 × 10⁻⁵ Torr at 0.7–1.0 Å s⁻¹. The thickness of the deposited Au layer was measured to be about 100 nm.

2.5 Characterizations

Photocurrent (J) and voltage (V) were measured using a solar simulator (Oriel Sol3A class AAA) equipped with 450 W xenon lamp (Newport 6279NS) and a Keithley 2400 source meter. The NREL-calibrated Si solar cell with KG-2 filter was used to adjust light intensity to one sun illumination (100 mW cm⁻²). While measuring photocurrent and voltage, the cell was covered with a metal aperture mask with area of 0.123 cm². Incident photon-to-electron conversion efficiency (IPCE) was measured at DC mode using the IPCE system (PV measurement Inc.). A 75 W xenon lamp (USHIO, Japan) was used as a white light source to generate a monochromatic beam. Calibration was conducted using NIST-calibrated photodiode G425 as a standard. The thin layers of ZrO₂ deposited on TiO₂ surfaces were investigated using a high-resolution transmission electron microscope (HR-TEM, JEOL JEM-2100F) at an acceleration voltage of 200 kV. Steady-state photoluminescence (PL) spectra were collected using a Cary Eclipse Fluorescence Spectrophotometer (Agilent Technologies) with a xenon flash lamp and a photomultiplier tube (PMT) detector. All samples were photoexcited at 530 nm.

3. Results and discussion

Fig. 1 shows HR-TEM images of TiO₂ nanoparticles with and without surface modification with thin ZrO₂ layers along with that of ZrO₂ nanoparticles. Compared to bare TiO₂ in Fig. 1(a), deposition of zirconium butoxide solution forms thin layer of ZrO₂ on the TiO₂ surface. Presence of ZrO₂ layer is also confirmed by elemental analysis based on energy dispersive X-ray, which is consistent with previous result.²⁷ Thickness of the surface formed ZrO₂ layer increases with increasing the precursor concentration or number of coatings, where ZrO₂ layers with average thickness of about 0.5 nm, 1.1 nm, 1.7 nm and 2.2 nm can be coated on the TiO₂ surface. Except for the ZrO₂ layers formed from single coating of 0.01 M and 0.02 M, conformal coating of ZrO₂ layer on TiO₂ surface is formed by dual coating. Compared to ∼50 nm in diameter for the TiO₂ particle, ZrO₂ nanoparticle shows smaller size of around 20 nm in diameter.

Fig. 1 High resolution transmission electron microscopy (HR-TEM) images of (a) bare TiO₂, surface modified TiO₂ by (b) single coating of 0.01 M Zr(OBu)₄ (about 0.5 nm), (c) 0.02 M (about 1.1 nm), (d) dual coating of 0.02 M (about 1.7 nm), (e) dual coating of 0.04 M (about 2.2 nm), and (f) bare ZrO₂.

Fig. 2(a) and (b) shows SEM images of bare nanocrystalline TiO₂ and ZrO₂ coated TiO₂. In Fig. 2(b), TiO₂ nanoparticles coated with ZrO₂ shows no significant difference in surface roughness, which indicates that conformal ZrO₂ layer might be formed on the surface of TiO₂ nanoparticles. Energy dispersive X-ray mapping is further studied to investigate elemental distribution for the ca. 1 nm-thick ZrO₂ layer deposited on the TiO₂ nanoparticle (Fig. 2(c)). In elemental distribution images, zirconium (Zr) is detected on the edge of the particle, while titanium (Ti) and oxygen (O) are distributed over the whole part of the particle, which confirms that conformal ZrO₂ layer is formed on the surface of the TiO₂ nanoparticle.

Fig. 2 Scanning electron microscopy (SEM) images of (a) mesoporous TiO₂ and (b) mesoporous TiO₂ coated with ZrO₂ (ca. 1 nm). (c) TEM images and energy dispersive X-ray mapping (element distribution images) of the TiO₂ nanoparticle coated with ZrO₂.

Photovoltaic performance is found to be significantly altered by the surface modification as can be seen in Fig. 3. Short-circuit photocurrent density (J_sc) decreases but open-circuit voltage (V_oc) increases predominantly with increasing the thickness of the surface coated ZrO₂ layer. Change in J_sc is well consistent with change in IPCE, where IPCE at long wavelength decreases for the 0.5 nm- and 1.1 nm-thick ZrO₂ layers. However, decrease in J_sc is observed in the entire wavelength when the ZrO₂ layer thickness increases to 1.7 nm and 2.2 nm. Average photovoltaic parameters are listed in Table 1. V_oc increases substantially from 0.863 V for bare TiO₂ to 0.988 V for 2.2 nm-thick ZrO₂ layer on TiO₂. In addition, surface modification improves fill factor (FF) by 9.1–11.8%. Increase in both V_oc and FF compensate the decreased J_sc, which consequently leads to improvement of PCE. Average PCE is enhanced from 11.7% for the bare TiO₂ to 13.6% for the 1.1 nm-thick ZrO₂ layer on TiO₂, corresponding to 16.2% increment.

Fig. 3 (a) Current (I)–voltage (V) curves and (b) incident photon-to-electron conversion efficiency (IPCE) spectra of the MAPbI₃ perovskite solar cells depending on concentration of ZrO₂ precursor solution for surface modification of mesoporous TiO₂ (mp-TiO₂). Mesoporous ZrO₂ based perovskite solar cell was prepared for comparison. The I–V data were collected by scanning from open-circuit to short circuit.

Table 1 Photovoltaic parameters of short-circuit photocurrent density (J_sc), open-circuit voltage (V_oc), fill factor (FF) and power conversion efficiency (PCE) of the MAPbI₃ perovskite solar cells depending on concentration of ZrO₂ precursor solution for surface modification of mesoporous TiO₂ (mp-TiO₂). The thickness of mesoporous TiO₂ layer was 300 ± 20 nm. Mesoporous ZrO₂ layer with thickness of about 300 nm was also compared. The data were obtained under AM 1.5G one sun illumination (100 mW cm⁻²)

Concentration of Zr(OBu)4	Average layer thickness	Jsc (mA cm^−2)	Voc (V)	FF	PCE (%)
Bare TiO2		21.0 ± 0.18	0.863 ± 0.01	0.646 ± 0.02	11.7 ± 0.49
0.01 M	∼0.5 nm	20.8 ± 0.64	0.896 ± 0.01	0.705 ± 0.02	13.1 ± 0.48
0.02 M	∼1.1 nm	20.6 ± 0.87	0.915 ± 0.01	0.718 ± 0.01	13.6 ± 0.38
Dual 0.02 M	∼1.7 nm	17.2 ± 0.39	0.962 ± 0.01	0.722 ± 0.01	12.0 ± 0.52
Dual 0.04 M	∼2.2 nm	15.9 ± 0.63	0.988 ± 0.01	0.711 ± 0.01	11.2 ± 0.40
Mp-ZrO2		10.9 ± 0.95	1.03 ± 0.03	0.593 ± 0.06	6.74 ± 1.4

---

In Fig. 4, J_scs calculated based on the IPCE data are compared with J_scs obtained from I–V curves. The calculated J_sc from the IPCE data has about 90% of the J_sc obtained from the I–V measurement, which indicates that the calculated J_sc is well matched with the observed J_sc with small deviation.

Fig. 4 Relationship between photocurrent densities measured by white xenon light (J_sc (I–V)) and calculated based on IPCE data.

We plot the photovoltaic parameters as a function of ZrO₂ layer thickness in Fig. 5. J_sc is almost invariant with ZrO₂ thickness upto ∼1 nm but starts to decrease upon increasing the ZrO₂ layer thickness. On the other hand, V_oc decreases continuously with ZrO₂ thickness. For the case of the mesoporous ZrO₂ film, injection from MAPbI₃ to ZrO₂ will be inhibited because of higher conduction band position and insulating property of ZrO₂.²⁸ Thus, free electrons are transported in the perovskite and collected to the photoanode. The highest V_oc observed for the mesoporous ZrO₂ film is related to Fermi energy level of MAPbI₃.²⁸ The large difference in V_oc between bare TiO₂ film (0.863 V) and bare ZrO₂ film (1.03 V) is indicative of evidence of electron injection from MAPbI₃ to TiO₂. The increasing rate of V_oc looks to be extrapolated to V_oc for the mesoporous ZrO₂ film, which indicates that the amount of electron injection into TiO₂ is diminished by increasing the surface ZrO₂ layer thickness. The controlled electron injection by the surface ZrO₂ layer is also confirmed by PL measurement.

Fig. 5 Dependence of J_sc, V_oc, FF and PCE on the thickness of ZrO₂ layer deposited on TiO₂ surface.

Fig. 6 compares PL intensity depending on the surface ZrO₂ layer thickness. All the PL peaks are located at around 780 nm, which is well consistent of emission peak of MAPbI₃.²⁹ The bare TiO₂ film shows lowest PL intensity, while PL intensity increases with increasing surface ZrO₂ layer thickness and eventually highest PL intensity is observed from the bare ZrO₂ film. When comparing PL intensities of the surface modified TiO₂ with those of both bare TiO₂ and ZrO₂, increase in PL intensity with increasing the surface ZrO₂ layer thickness is clearly indicative of reduction of electron injection by the surface coated ZrO₂. The tendency of PL intensity correlates with change of V_oc with the surface ZrO₂ layer.

Fig. 6 Photoluminescence (PL) spectra of the FTO/compact-TiO₂/mesoporous-TiO₂/thin ZrO₂ layer/MAPbI₃ films and the FTO/compact-TiO₂/mesoporous ZrO₂/MAPbI₃ film.

Fig. 7 shows dark J–V curve of devices incorporating bare TiO₂ and TiO₂ coated with ca. 1.1 nm-thick ZrO₂ layer. Shunt and series resistance are obtained from the inverse slope at zero bias and V_oc region, respectively.³⁰ Series resistance is increased by 26% (from 6.25 to 7.91 Ω cm⁻²) with ZrO₂ layer, while shunt resistance is increased by 65% (from 23 364.49 to 38 654.81 Ω cm⁻²). Therefore, the improved FF is mainly due to the increased shunt resistance. Ideality factor (n) and saturation current (J₀) were calculated by fitting the data using following eqn (1),

where J_D, V_b, q, k_B, and T represent current density, bias voltage, electron charge, Boltzmann constant, and temperature, respectively. Ideality factor is calculated to be 2.09 for bare TiO₂ and 2.32 for ZrO₂ coated sample. J₀ is estimated to be lowered by ZrO₂ coating (2.09 × 10⁻⁷ mA cm⁻²) compared to bare sample (3.03 × 10⁻⁷ mA cm⁻²), which correlates with higher V_oc by ZrO₂ coating.³¹

Fig. 7 Dark current density (J)–voltage (V) curves of MAPbI₃ perovskite solar cell incorporating bare TiO₂ and TiO₂ coated with 1.1 nm-thick ZrO₂ layer. Empty circles and solid lines represent the measured data and the fit results, respectively.

I–V hysteresis is investigated since planar structure without mesoporous TiO₂ film or mesosuperstructured perovskite with scaffold oxide layer such as Al₂O₃ shows more pronounced I–V hysteresis than the mesoscopic structure with mesoporous TiO₂ film.^32,33 Fig. 8 compares I–V hysteresis, where the I–V hysteresis of the bare TiO₂ is much less than that of the bare ZrO₂. Since ZrO₂ is non-injection oxide, the structure containing the mesoporous ZrO₂ film is similar to those with Al₂O₃ or without mesoporous oxide layer in terms of electron injection. Thus I–V hysteresis of the mesoporous ZrO₂ is expected to be severe, which is well consistent with the previous reports showing severs I–V hysteresis of planar structure or Al₂O₃ bearing perovskite.^32,33 Such I–V hysteresis is thus related to non-injection property in the mesoscopic structure. Compared to I–V hysteresis of the bare TiO₂ film, more developed I–V hysteresis of the surface modified TiO₂ with ZrO₂ is due to the reduced electron injection into TiO₂ by the surface ZrO₂ layer. Extent of I–V hysteresis provides indirect evidence of interfacial electron injection at TiO₂–perovskite junction.

Fig. 8 Scan direction dependent I–V curves for the perovskite solar cells based on (a) mesoporous TiO₂ (without surface modification), (b) surface modified TiO₂ with 1.1 nm-thick ZrO₂ layer, and (c) mesoporous ZrO₂.

Photo-stability of the device with and without ca. 1.1 nm-thick ZrO₂ layer is evaluated (Fig. 9). The measurement was conducted under one sun illumination without encapsulation. Both of the devices show severe degradation of PCE during the light soaking in which 90% of initial performance is degraded within an hour. The degradation of PCE looks faster with ZrO₂ coating. The poor stability of device with ZrO₂ coating is probably related to characteristics of the device similar to planar junction device without mesoporous TiO₂ layer (hysteresis, electron transport pathway).³⁴ Our result suggests that stability of perovskite solar cell might be closely related to electron transport pathway.

Fig. 9 Evolution of PCE of the devices incorporating bare TiO₂ and TiO₂ coated with 1.1 nm-thick ZrO₂ layer as a function of light soaking time.

4. Conclusion

We investigated interfacial electron injection by means of surface modification of the mesoporous TiO₂ film in the mesoscopic perovskite solar cell structure. V_oc was found to depend predominantly on thickness of ZrO₂ layer. For the non-injection system employing mesoporous ZrO₂ film, highest V_oc was demonstrated. When assuming that no interfacial injection occurs, V_oc is expected to be independent on surface modification. Thus, our experiments evidences obviously that electron injection occurs at TiO₂–perovskite interface. By controlling electron injection and thereby optimizing photovoltaic parameters under the given experimental condition, a power conversion efficiency (PCE) was improved by 16.2%, compared to the bare (untreated) TiO₂ device.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grants funded by the Ministry of Science, ICT & Future Planning (MSIP) of Korea under contracts no. NRF-2010-0014992, NRF-2012M1A2A2671721, NRF-2012M3A7B4049986 (Nano Material Technology Development Program) and NRF-2012M3A6A7054861 (Global Frontier R&D Program on Center for Multiscale Energy System).

References

1. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel and N.-G. Park, Sci. Rep., 2012, 2, 591–597.
2. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051.
3. J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park and N.-G. Park, Nanoscale, 2011, 3, 4088–4093.
4. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643–647.
5. http://www.nrel.gov/ncpv/images/efficiency_chart.jpg.
6. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476–480.
7. S. D. Wolf, J. Holovsky, S.-J. Moon, P. Löper, B. Niesen, M. Ledinsky, F.-J. Haug, J.-H. Yum and C. Ballif, J. Phys. Chem. Lett., 2014, 5, 1035–1039.
8. W.-J. Yin, T. Shi and Y. Yan, J. Phys. Chem. C, 2015, 119, 5253–5264.
9. Z. Xie, S. Liu, L. Qin, S. Pang, W. Wang, Y. Yan, L. Yao, Z. Chen, S. Wang, H. Du, M. Yu and G. G. Qin, Opt. Mater. Express, 2015, 5, 29–43.
10. P. Löper, M. Stuckelberger, B. Niesen, J. Werner, M. Filipič, S.-J. Moon, J.-H. Yum, M. Topič, S. D. Wolf and C. Ballif, J. Phys. Chem. Lett., 2015, 6, 66–71.
11. C.-W. Chen, S.-Y. Hsiao, C.-Y. Chen, H.-W. Kang, Z.-Y. Huang and H.-W. Lin, J. Mater. Chem. A, 2015, 3, 9152–9159.
12. Q. Lin, A. Armin, R. C. R. Nagiri, P. L. Burn and P. Meredith, Nat. Photonics, 2015, 9, 106–112.
13. E. J. Juarez-Perez, R. S. Sanchez, L. Badia, G. Garcia-Belmonte, Y. S. Kang, I. Mora-Sero and J. Bisquert, J. Phys. Chem. Lett., 2014, 5, 2390–2394.
14. G. Xing, N. Mathews, S. Sun, S. S. Lim, Y. M. Lam, M. Grätzel, S. Mhaislkar and T. C. Sum, Science, 2013, 342, 344–347.
15. S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013, 342, 341–344.
16. D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, A. Buin, Y. Chen, S. Hoogland, A. Rothenberger, K. Katsiev, Y. Losovyj, X. Zhang, P. A. Dowben, O. F. Mohammed, E. H. Sargent and O. M. Bakr, Science, 2015, 347, 519–522.
17. C. S. Ponseca Jr, T. J. Savenije, M. Abdellah, K. Zheng, A. Yartsev, T. Pascher, T. Harlang, P. Chabera, T. Pullerits, A. Stepanov, J.-P. Wolf and V. Sundstrom, J. Am. Chem. Soc., 2014, 136, 5189–5192.
18. N.-G. Park, J. Phys. Chem. Lett., 2013, 4, 2423–2429.
19. H. J. Snaith, J. Phys. Chem. Lett., 2013, 4, 3623–3630.
20. H.-S. Kim, S. H. Im and N.-G. Park, J. Phys. Chem. C, 2014, 118, 5615–5625.
21. M. Grätzel and N.-G. Park, Nano, 2014, 9, 1440002.
22. A. Marchioro, J. Teuscher, D. Friedrich, M. Kunst, R. van de Krol, T. Moehl, M. Grätzel and J.-E. Moser, Nat. Photonics, 2014, 8, 250–255.
23. V. Gonzalez-Pedro, E. J. Juarez-Perez, W.-S. Arsyad, E. M. Barea, F. Fabregat-Santiago, I. Mora-Sero and J. Bisquert, Nano Lett., 2014, 14, 888–893.
24. H.-S. Kim, I. Mora-Sero, V. Gonzalez-Pedro, F. Fabregat-Santiago, E. J. Juarez-Perez, N.-G. Park and J. Bisquert, Nat. Commun., 2013, 4, 2242.
25. J.-H. Im, I.-H. Jang, N. Pellet, M. Grätzel and N.-G. Park, Nat. Nanotechnol., 2014, 9, 927–932.
26. N. Papageorgiou, M. Grätzel, O. Enger, D. Bonifazi and F. Diederich, J. Phys. Chem. B, 2002, 106, 3813–3822.
27. H.-W. Kang, J.-W. Lee and N.-G. Park, Faraday Discuss., 2014, 176, 287–299.
28. D. Bi, S.-J. Moon, L. Häggman, G. Boschloo, L. Yang, E. M. J. Johansson, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, RSC Adv., 2013, 3, 18762–18766.
29. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorg. Chem., 2013, 52, 9019–9038.
30. S.-H. Oh, S.-I. Na, J. Jo, B. Lim, D. Vak and D.-Y. Kim, Adv. Funct. Mater., 2010, 20, 1977–1983.
31. D. Zhu, J. Xu, A. N. Noemaun, J. K. Kim, E. F. Schubert, M. H. Crawford and D. D. Koleske, Appl. Phys. Lett., 2009, 94, 081113.
32. H. J. Snaith, A. Abte, J. M. Ball, G. E. Eperon, T. Leijtens, N. K. Noel, S. D. Stranks, J. T.-W. Wang, K. Wojciechowski and W. Zhang, J. Phys. Chem. Lett., 2014, 5, 1511–1515.
33. H.-S. Kim and N.-G. Park, J. Phys. Chem. Lett., 2014, 5, 2927–2934.
34. E. L. Unger, E. T. Hoke, C. D. Bailie, W. H. Nguyen, A. R. Bowring, T. Heumüller, M. G. Christoforo and M. D. McGehee, Energy Environ. Sci., 2014, 7, 3690–3698.

---

Footnote

† These authors are contributed equally to this work.

---