Photovoltaic (PV) solar energy has the prospect to supply

Transcription

1 This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes. Letter pubs.acs.org/nanolett Highly Efficient Perovskite Solar Cells with Tunable Structural Color Wei Zhang,, Miguel Anaya,, Gabriel Lozano, Mauricio E. Calvo, Michael B. Johnston, Hernań Míguez,*, and Henry J. Snaith*, Department of Physics, University of Oxford, Clarendon Laboratory, Parks Road, Oxford OX1 3PU, United Kingdom Instituto de Ciencia de Materiales de Sevilla, Consejo Superior de Investigaciones, Científicas-Universidad de Sevilla, Calle Ameŕico Vespucio 49, Sevilla, Spain *S Supporting Information ABSTRACT: The performance of perovskite solar cells has been progressing over the past few years and efficiency is likely to continue to increase. However, a negative aspect for the integration of perovskite solar cells in the built environment is that the color gamut available in these materials is very limited and does not cover the green-to-blue region of the visible spectrum, which has been a big selling point for organic photovoltaics. Here, we integrate a porous photonic crystal (PC) scaffold within the photoactive layer of an opaque perovskite solar cell following a bottom-up approach employing inexpensive and scalable liquid processing techniques. The photovoltaic devices presented herein show high efficiency with tunable color across the visible spectrum. This now imbues the perovskite solar cells with highly desirable properties for cladding in the built environment and encourages design of sustainable colorful buildings and iridescent electric vehicles as future power generation sources. KEYWORDS: Structural color, perovskite solar cells, building integrated photovoltaic, porous photonic crystal, multilayers Photovoltaic (PV) solar energy has the prospect to supply our global energy demands hundreds of times over. A large fraction of the cost of solar energy is not the solar cells but the fixed cost of land, frames, supports and services. A smart move to slash these fixed costs is to build the PV directly into the infrastructure and supports already constructed for other purposes, such as building facades, fences, walls, awnings, and automotive skins. However, for widespread uptake of building integrated PV (BIPV), 1 the PV has to be aesthetically versatile. Organic inorganic halide perovskite solar cells have undergone rapid progress, 2 11 with efficiencies approaching those of existing thin film and silicon. 9 Perovskite solar cells are constructed by forming a polycrystalline layer of an organic inorganic halide perovskite, typically CH 3 NH 3 PbI 3 or the mixed halide variant CH 3 NH 3 PbI 3 x Cl x, sandwiched between an n-type and p-type charge collection layer. 6 In addition to this simple planar structure, the perovskite layer can be infiltrated into a mesoporous layer of n-type TiO 2 or insulating dielectric scaffolds. 3,4,7,12,13 Although significant advance has been made to achieve highly efficient perovskite solar cells, the negative aspect for perovskites in BIPV is that the color pallet available in these materials is very basic and, notably absent of greens and blues, 14,15 highly desirable in the built environment. There has been significant previous work developing semitransparent solar cells, which is another key feature to consider for window applications in BIPV. In this regard, thinner perovskite layers have been already demonstrated to achieve semitransparent cells with high power conversion efficiencies However, strong coloration, and specifically blues and greens, has not been achieved. A solution may be to follow the route often employed by nature: structural colors, in contrast to pigments or dyes, arise from the interference of coherently scattered light waves. With the aim of endowing this type of device with color, we propose a versatile approach based on photonics rather than on chemical management because the latter has demonstrated limited success in the blue-to-green region of the spectrum. 14,15 Although the multilayer approach has been proven to be compatible with semitransparent devices, 19,20 in this study, we will focus on opaque cells and develop a multilayered scaffold that exhibits photonic crystal (PC) properties, which simultaneously provides optical reflection at selected visible wavelength ranges and allows good electrical contact with the charge collecting layers. We successfully achieve a broad gamut of iridescent solar cells, which deliver up to 8.8% efficiency with a blue hue. The thickness of such a supporting porous structure is restricted by the charge transport properties of the perovskite, which typically imposes the scaffold to be no thicker than 400 nm in order to attain highly efficient devices, which are consequently nontransparent. 21 In order to integrate a photonic structure within a perovskite device, our initial attempts involved the substitution of the porous scaffold by a PC Received: November 13, 2014 Revised: January 31, 2015 Published: February 4, American Chemical Society 1698 DOI: /nl504349z Nano Lett. 2015, 15,

2 Letter Nano Letters Figure 1. Structural and aesthetic characterization. (a) Scheme of a photonic-crystal (PC)-based perovskite solar cell. (b) Secondary electron scanning electron microscope image of a cross section of a PC-based solar cell. The PC is made by the alternated deposition of SiO2 nanoparticles and TiO2 layers. Devices under investigation are comprised of gold as metallic contact, spiro-ometad as hole transport material, the PC is infiltrated by CH3NH3PbI3 xclx, and fluorinated tin oxide as transparent conducting oxide. The cell is deposited over a glass substrate. (c) Colors (from blue to red) displayed by devices integrating different PCs. Figure 2. Color analysis. (a) Experimental (solid lines) and theoretical reflectance (dashed lines) of the PC-based orange (orange lines) and bluegreen (dark cyan lines) perovskite solar cells. Orange curves are vertically shifted by ΔR = 0.3 for the sake of clarity. (b c) Calculated spatial (vertical axis) and spectral (horizontal axis) distribution of the square of the electric field enhancement ( E 2/ E0 2), that is, the square of the electric field normalized by the square of the electric field in the incoming medium, along a cross section of the solar cell, for the PC-based (b) orange and (c) blue-green cells. Calculations were carried out considering the multilayer parameters extracted from the fittings of the reflectance measurements: (b) dh = 58 nm, dl = 87 nm and (c), dh = 39 nm, dl = 90 nm. Horizontal dashed lines are guides to the eye to delimit the interfaces between layers in the solar cells. The bottom of the images is the glass side, and the top, the gold electrode side. (d) CIE 1931 chromaticity space, showing the color hues of the PC-based (R) red (0.41, 0.34), (O) orange (0.41, 0.40), (G) green (0.33, 0.41), (BG) blue-green (0.29, 0.35), and (B) blue (0.29, 0.30) cells. The dashed black line delimits the device color gamut. It encompasses all different hues that can be achieved tuning the lattice parameter from 20 to 300 nm, and the thickness ratio of the high and low index material of the PC. The porosities of the high and low index materials, prior to perovskite infiltration, are considered to be 4% and 50%, respectively. made by stacking nanoparticle layers of different kind,22 following a recipe that had permitted the successful integration of colored mirrors in dye sensitized solar cells.20 However, although we achieved efficient perovskite solar cells by this approach, the color they displayed was not different to that of nonstructured ones, since the refractive index contrast of the porous photonic crystal after perovskite infiltration was largely diminished. A potential solution to this problem is to increase the number of layers, so the reflections arising from the photonic crystal mirrors are stronger and color can be preserved after infiltration. Unfortunately, we found that such an approach rendered the device performance very poor, as we show in Supporting Information Figure S1, because the total scaffold thickness is double that of the optimum one. Thus, a balance between minimization of the photonic skeleton thickness and preservation of its high reflectance after infiltration is required. To meet this compromise, we devise multilayers in which the refractive index contrast between layers in the PC is maximized by grossly varying the porosity of each layer. In this way, the number of unit cells, and therefore the thickness, required to achieve a strong reflection peak can be reduced below the optimum scaffold size, but the perovskite precursors is still capable of soaking the photonic matrix. We employ layers of materials in different aggregation states that permitted us to tune the open porosity of each alternate layer. We make low and high refractive index layers by spin-coating SiO2 nanoparticles and TiO2 sol-gel precursors,23 respectively, onto a fluorinated tin oxide (FTO) coated glass substrate. During the deposition of the multilayer mirror, and in order to prevent infiltration of the TiO2 precursors into the porous 1699 DOI: /nl504349z Nano Lett. 2015, 15,

3 Nano Letters Letter Figure 3. Electrical characterization and device performance analysis. (a c) Spectral dependence of the external quantum efficiency of (a) a nonphotonic structured SiO 2 scaffold-based perovskite solar cell, which acts as a reference, (b) an orange PC-based cell, and (c) blue-green PC-based cell. The shaded areas correspond to the spectral regions in which the PC-based cells show high reflectance. (d g) Box-plot of the device performance indicators for the different PC-based cells: (d) short-circuit current density (J sc ), (e) open circuit voltage (V oc ), (f) power conversion efficiency (PCE), and (g) fill factor (FF). (h j) Digital camera pictures of the devices labeled as (h) blue, (i) green, and (j) orange. SiO 2 layer, the latter is filled with a polymer. In this way, we form relatively dense (4% open porosity) layers of TiO 2 alternately onto hybrid polymer-sio 2 layers. Subsequent sintering decomposes the polymer matrix and renders the SiO 2 layer porous again (50% open porosity). We obtained porosity and refractive index values (n L = 1.25 and n H = 2.38, at λ = 500 nm) from the optical analysis of the PCs. After depositing the perovskite precursors and annealing, we observe a strong red-shift of the reflectance peak of the multilayer (see Supporting Information Figure S2), which indicates that the mirror is being infiltrated by the perovskite semiconductor, resulting in a shift in the effective refractive index. In fact, from the fitting of the optical properties, we estimate that the perovskite completely fills the voids in both SiO 2 and TiO 2 layers, demonstrating their accessibility and interconnectivity, despite the very low porosity in the dense TiO 2 layers. We show a schematic of the structure of the perovskite solar cell incorporating a three-unit-cell one-dimensional PC in Figure 1a and a scanning electron microscopy (SEM) image of a cross section of such a device in Figure 1b. This image reveals that the first SiO 2 nanoparticle layer is capable of leveling the roughness, characteristic of the FTO coating, thus providing a flat surface for subsequent layer deposition. Furthermore, back scattered electron image analysis shows that the infiltrated perovskite percolates through the void network of the photonic structure, which can then act as an electronic bridging path between the solid perovskite overlayer and the FTO anode. A study of the characteristics of the perovskite infiltration is provided as Supporting Information (see Figure S2). As we show in the photographs of constructed solar cells in Figure 1c, the obtained devices show intense colors variable across the visible spectrum. We can precisely tune the reflectance peak of the PC by modifying the thickness of the individual layers. The vivid iridescences observed are achieved by the combination of reflection from the PC and strong absorption from the infiltrated and overlying perovskite material. This strategy, in which structural color and absorption are combined, is analogue to that found in beetles and may be considered, on these grounds, biomimetic. 24 We performed an in-depth experimental and theoretical analysis of the optical effects occurring in perovskite solar cells that integrate PCs. We developed a theoretical model using a method based on the transfer matrix formalism, which allows us to calculate the electric field intensity within the layered structure and, thus, to visualize the effect of integrating the PC on the spatial distribution of the radiation in the photovoltaic device. Full details on the model can be found elsewhere, as well as evidence of its suitability to describe the optics of perovskite solar cells. 25 We fitted the experimental reflectance spectra of PC based cells measured at normal incidence, using the thickness of each layer in the structure extracted from the SEM as input parameters. Small variations around these reference values were permitted in the code in order to attain a fine-tuning of the calculated reflectance and hence a proper fitting. Good agreement between theory and experiment is attained, as displayed in Figure 2a. From the calculation, the unit cells of the periodic stacks are estimated to be 129 nm (d H = 39 nm, d L = 90 nm, for the TiO 2 and SiO 2 layers comprising the unit cell, respectively) for the PC based blue-green cell, and 145 nm (d H = 58 nm, d L = 87 nm) for the orange cell. Please note that the notation for the ranges herein employed to identify the color of the cells is the one commonly employed in CIE (Comission Internationale de léćlairage) color diagrams. Such reflectance spectra can also be considered as the complementary of the absorptance spectra because light transmission through the cells is null. In that regard, each photonic design leads to a different field intensity distribution along the cell and hence to a different spatial and spectral distribution of optical absorption in the device. In Figure 2b c, we show the way in which the optical field is distributed within the multilayer for the two cases chosen as examples. For comparing, we note the strong difference of light field distribution for photons with wavelength around 700 nm, whose penetration is much enhanced in the blue-green case, in comparison to the orange structure. This allows us to estimate the amount of light that is captured by each absorbing layer in the perovskite cell, which is provided as Supporting Information (Figure S3), and hence to discriminate between 1700 DOI: /nl504349z Nano Lett. 2015, 15,

4 Nano Letters Letter Table 1. Solar Cell Performance Parameters device color J sc (ma/cm 2 ) PCE (%) V oc (V) FF blue best average 12.1 ± ± ± ± 0.02 blue-green best average 10.1 ± ± ± ± 0.03 green best average 9.4 ± ± ± ± 0.03 orange best average 8.9 ± ± ± ± 0.05 red best average 7.2 ± ± ± ± 0.01 productive (absorbed in the perovskite material) and parasitic (absorbed in other components such as FTO) absorption. 25 The device owes its color to the convolution of the reflectance spectrum that results from the interference of all beams partially transmitted and reflected at each interface in the device with the photopic response of the human eye. In order to characterize the color hue of the PC based photovoltaic device, we calculate their color coordinates and display the results as black symbols on the CIE 1931 chromaticity diagram in Figure 2d. By making use of our theoretical model, we have also extrapolated the full range of colors that can be eventually achieved by varying the thickness of the layers comprising the PCs infiltrated by perovskite. Such colors are encompassed by the dashed polygon in Figure 2d, which represents the PCbased cell tone gamut. Related to the appearance of the cells, structural colors are known to vary with the angle of observation, as they result from interference effects that are sensitive to the direction of illumination. However, in contrast to incorporating a chromatic filter in the surface of the outer glass, the integration of the PC within the solar cell strongly reduces the angular variability of the perceived color, as we described in Figure S4 of the Supporting Information. This could be advantageous where consistent coloration with angle is desired in a determined environment, such as buildings. In Figure 3, we show the photovoltaic performance of perovskite solar cells displaying different colors. If the light is being reflected, it is not being absorbed, and hence, we expect there to be a trade-off between color and photocurrent generation for these solar cells. In Figure 3a c, we show the different spectral response of the external quantum efficiency (EQE) measured from perovskite solar cells integrating orange and blue-green reflecting PCs as well as of a reference solar cell, made using a mesoporous SiO 2 scaffold with a thickness of ca. 370 nm. As expected, the EQE curves show clear depletions in those spectral regions for which the PC scaffold presents its reflectance maximum (shaded in gray in Figure 3b and c). Other potentially relevant effects such as photonic band gap edge absorption enhancement, 26 are not readily detectable in these measurements. In Figure 3d g, we present the performance parameters for a series of solar cells made using photonic scaffolds that display a range of optical reflectance peaks. Measurements were taken under simulated AM1.5 sun light, at 100 mw cm 2 irradiance, with the parameters summarized in Table 1. The first encouraging observation is that the solar cells operate well, despite the requirement for conduction through the PC, consistent with the perovskite absorber bridging the conduction through the photonic scaffold. The short circuit current density (J sc ) monotonically decreases with the increase of the PC lattice parameter, as the 1701 color shifts from blue to red. However, the open-circuit voltage (V oc ) and fill factor (FF) are only slightly varying, displaying independence of the PC lattice parameter and indicating that the observed differences in photovoltaic performance have predominantly a photonic origin. We also observe that for blue, blue-green, green and orange hues, the power conversion efficiency (PCE) of the best performing devices are all above 6%, with the peak of 8.8% being achieved for the blue cells, showing good potential for making efficient working devices with vivid and tunable colors. As a comparison, we show the performance of the reference cell in Supporting Information Figure S5, which exhibits a peak and average PCE of 10.5 and 9.5%, respectively, indicating reasonable retention of the expected performance for the PC perovskite solar cells. We note that a current widely adopted technology for BIPV is amorphous silicon, where the BIPV modules integrated with a red-orange hue are less than 5% efficiency, illustrating potential for much superior performance for the colorful perovskite modules. Our work represents another example of the successful integration of multilayer structures with photonic properties in PV devices aiming at modifying their appearance without being detrimental for their performance. 21,27,28 In Figure 3h j, we show photographs of the blue, green, and orange cells. The complete color palette is shown in the Supporting Information (see Figure S6). In addition, PC based solar cells show a good stabilized power output, which is the PCE measured under constant load and illumination. 29 As an example, a blue cell that showed a maximum JV determined efficiency of 7.9% and exhibited a stabilized power output of 6.6%. Results are also provided as Supporting Information (Figure S7). In summary, we have demonstrated that a porous photonic crystal, built from alternating liquid processed layers of dense TiO 2 and porous SiO 2, can be constructed and integrated as a substitute scaffold into the active layer of a perovskite solar cell. The resulting solar cells exhibit well-defined reflectance bands, with the remaining nonreflected light being channelled into the perovskite absorber to undergo photovoltaic conversion at high efficiency. Aesthetically attractive perovskite solar cells in blue, blue-green, and green hues are now achievable and offer a new avenue for efficient and colorful PV for integration into the built environment. Key advantages of this route, as opposed to employing a colorful absorber layer, are that the structural color is unlikely to bleach or fade with time, and in addition, a single panchromatic photovoltaic absorber optimized for efficiency and stability can be integrated with a broad range of color hues. An unexpected advantage of the colorful perovskite solar cells for BIPV applications may be that the additional reflected light will reduce the thermal gain within the solar cell and in the DOI: /nl504349z Nano Lett. 2015, 15,

5 Nano Letters Letter building, which may have added benefits of lower operating temperatures of the solar cells, and hence increasing efficiency under real world conditions, reducing the cooling demand in buildings. We expect future improvements to be made by understanding and enhancing the electronic properties of the photonic crystal scaffolds, narrowing the reflectance bands and further optimization of the perovskite solar cells. This structure will also lend itself to perovskite light emitting diodes and may prove to be important for the realization of perovskite injection lasers. This method may also be applicable to a broader range of solution processed photovoltaic or light emitting materials. ASSOCIATED CONTENT *S Supporting Information Method details, additional device performance, SEM image, absorptance and reflectance spectra, color hues analysis. This material is available free of charge via the Internet at pubs.acs.org. AUTHOR INFORMATION Corresponding Authors * h.miguez@csic.es. * h.snaith1@physics.ox.ac.uk. Author Contributions These authors contributed equally to this work. W.Z. and M.A. were in charge of the photonic crystal based perovskite solar cell design, preparation and characterization. Both of them contributed equally to this work. G.L. performed all the optical modeling of the cells and the color hue analysis. M.E.C. was in charge of optimizing the photonic crystal structure for integration in the solar cell and of SEM characterization. M.B.J. performed the EQE measurement. H.J.S. and H.M. supervised the work and wrote the manuscript. Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS The research leading to these results has received funding from the EPSRC Supergen, the European Research Council under the European Union s Seventh Framework Programme (FP7/ )/ERC grant agreement no (POLIGHT) and HYPER Project, the Spanish Ministry of Economy and Competitiveness under grant MAT and the Junta de Andaluciá under grant FQM5247. M.A. is grateful to La Caixa Foundation for its financial support. FESEM characterization was performed at CITIUS, and we are grateful for its support. W.Z. thanks Ms Mingzhen Liu for providing vapordeposited film sample. ABBREVIATIONS PV, photovoltaic; PCs, photonic crystals; BIPV, building integrated photovoltaics; spiro-ometad, 2,2,7,7 -tetrakis (N,N-di-p-methoxyphenylamine)-9,9 -spirobifluorene; FTO, fluorinated tin oxide; PCE, power conversion efficiency; J sc, short-circuit current density; FF, fill factor; V oc, open-circuit voltages; EQE, external quantum efficiency; SEM, scanning electron microscopy REFERENCES (1) Henemann, A. Renewable Energy Focus 2008, 9, (2) Kojima, A.; Teshima, K.; Shirai, Y.; Miyasaka, T. J. Am. Chem. Soc. 2009, 131, (3) Kim, H.-S.; Lee, C.-R.; Im, J.-H.; Lee, K.-B.; Moehl, T.; Marchioro, A.; Moon, S.-J.; Humphry-Baker, R.; Yum, J.-H.; Moser, J. E.; Graẗzel, M.; Park, N.-G. Sci. Rep. 2012, 2. (4) Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith, H. J. Science 2012, 338, (5) Burschka, J.; Pellet, N.; Moon, S.-J.; Humphry-Baker, R.; Gao, P.; Nazeeruddin, M. K.; Graẗzel, M. Nature 2013, 499, (6) Liu, M.; Johnston, M. B.; Snaith, H. J. Nature 2013, 501, (7) Heo, J. H.; Im, S. H.; Noh, J. H.; Mandal, T. N.; Lim, C.-S.; Chang, J. A.; Lee, Y. H.; Kim, H.-j.; Sarkar, A.; Nazeeruddin, M. K.; Graẗzel, M.; Seok, S. I. Nat. Photonics 2013, 7, (8) Liu, D.; Kelly, T. L. Nat. Photonics 2014, 8, (9) Zhou, H.; Chen, Q.; Li, G.; Luo, S.; Song, T.-b.; Duan, H.-S.; Hong, Z.; You, J.; Liu, Y.; Yang, Y. Science 2014, 345, (10) Jeon, N. J.; Noh, J. H.; Kim, Y. C.; Yang, W. S.; Ryu, S.; Seok, S. I. Nat. Mater. 2014, 13, (11) Im, J.-H.; Jang, I.-H.; Pellet, N.; Graẗzel, M.; Park, N.-G. Nat. Nanotechnol. 2014, 9, (12) Green, M. A.; Ho-Baillie, A.; Snaith, H. J. Nat. Photonics 2014, 8, (13) Snaith, H. J. J. Phys. Chem. Lett. 2013, 4, (14) Noh, J. H.; Im, S. H.; Heo, J. H.; Mandal, T. N.; Seok, S. I. Nano Lett. 2013, 13, (15) Eperon, G. E.; Stranks, S. D.; Menelaou, C.; Johnston, M. B.; Herz, L. M.; Snaith, H. J. Energy Environ. Sci. 2014, 7, (16) Eperon, G. E.; Burlakov, V. M.; Goriely, A.; Snaith, H. J. ACS Nano 2014, 8, (17) Roldan-Carmona, C.; Malinkiewicz, O.; Betancur, R.; Longo, G.; Momblona, C.; Jaramillo, F.; Camacho, L.; Bolink, H. J. Energy Environ. Sci. 2014, 7, (18) Ono, L. K.; Wang, S.; Kato, Y.; Raga, S. R.; Qi, Y. Energy Environ. Sci. 2014, 7, (19) Colodrero, S.; Mihi, A.; Ha ggman, L.; Ocana, M.; Boschloo, G.; Hagfeldt, A.; Miguez, H. Adv. Mater. 2009, 21, (20) Betancur, R.; Romero-Gomez, P.; Martinez-Otero, A.; Elias, X.; Maymo, M.; Martorell, J. Nat. Photonics 2013, 7, (21) Ball, J. M.; Lee, M. M.; Hey, A.; Snaith, H. J. Energy Environ. Sci. 2013, 6, (22) Colodrero, S.; Ocanã, M.; Míguez, H. Langmuir 2008, 24, (23) Brinker, C. J.; Frye, G. C.; Hurd, A. J.; Ward, K. J.; Ashley, C. S. J. Non-Cryst. Solids 1990, 121, (24) Kolle, M. Photonic structures inspired by nature. Springer: New York, (25) Anaya, M.; Lozano, G.; Calvo, M. E.; Zhang, W.; Johnston, M. B.; Snaith, H. J.; Miguez, H. J. Phys. Chem. Lett. 2015, 6, (26) Anaya, M.; Calvo, M. E.; Luque-Raigoń, J. M.; Míguez, H. J. Am. Chem. Soc. 2013, 135, (27) Colonna, D.; Colodrero, S.; Lindstro m, H.; Di Carlo, A.; Miguez, H. Energy Environ. Sci. 2012, 5, (28) Lee, K.-T.; Lee, J. Y.; Seo, S.; Guo, L. J. Light: Sci. Appl. 2014, 3, e215. (29) Snaith, H. J.; Abate, A.; Ball, J. M.; Eperon, G. E.; Leijtens, T.; Noel, N. K.; Stranks, S. D.; Wang, J. T.-W.; Wojciechowski, K.; Zhang, W. J. Phys. Chem. Lett. 2014, 5, DOI: /nl504349z Nano Lett. 2015, 15,

6 Supporting information for: Highly Efficient Perovskite Solar Cells with Tuneable Structural Color Wei Zhang 1, Miguel Anaya 2, Gabriel Lozano 2, Mauricio E. Calvo 2, Michael B. Johnston 1, Hernán Míguez 2*, Henry J. Snaith 1* 1 Department of Physics, University of Oxford, Clarendon Laboratory, Parks Road, Oxford, OX1 3PU, UK 2 Instituto de Ciencia de Materiales de Sevilla Consejo Superior de Investigaciones, Científicas-Universidad de Sevilla, Américo Vespucio 49, Sevilla, 41092, Spain W. Z. and M. A. contributed equally to this work. Corresponding Authors * h.snaith1@physics.ox.ac.uk. * h.miguez@csic.es Methods Preparation of photonic crystals (PCs) One dimensional PCs were prepared using a TiO 2 dispersion obtained by sol-gel techniques, a suspension of SiO 2 nanoparticles and a solution of polystyrene. TiO 2 dispersion was obtained by mixing titanium tetraisopropoxide in an ethanolic solution of hydrochloric acid. SiO 2 colloids (Ludox TMA, Aldrich) and polystyrene (M w 280,000 Aldrich) were diluted in methanol to 3 wt% and in toluene to 0.5 wt%, respectively. Fluorine-doped tin oxide (FTO) coated glass (TEC 7, 7Ω/sq) was used as substrate in this work. Initially, FTO was removed from regions under the anode contact by etching with a 2 M HCl solution and zinc powder. Substrate was then cleaned sequentially in 2% Hellmanex, acetone, 2-propanol and oxygen plasma. PCs were built by an alternated deposition of the three compounds over FTO glass using a spin coater (Laurell WS-400E-6NPP) in which both the acceleration ramp and the final rotation speed could be precisely determined. The first layer was deposited using 200 µl of TiO 2 dispersion and then the sample was accelerated up to different final speeds with nominal value between 2000 and 3000 rpm. The total spin-coating process (ramping-up and final speed) was completed in 60 s. Then, the sample was treated at a temperature of 500 ºC during 200 s in order to densify the TiO 2 film. After cooling the substrate in a metal surface at room temperature, 200 µl of SiO 2 1

7 nanoparticles solution were deposited using final speeds with nominal value between 3000 and 8000 r.p.m for 60 s. In order to fill the voids network of this last layer, 200 µl of polystyrene solution were spin-coated at 4500 r.p.m for 60 s. This last step allows us to deposit on top a layer of TiO 2 without penetration within the SiO 2 nanoparticulated layer. After that, the sample is heated again at 500 ºC during 200 s, in order to remove the polystyrene and to densify the TiO 2 layer.. This sequence was repeated until a total of 5 to 7 layers, depending on the cell, were deposited on top of FTO. Finally the sample is treated at 500 ºC during 30 min. Perovskite solution preparation. Methylammonium iodide (CH 3 NH 3 I) was prepared by reacting methylamine, 33 wt% in ethanol (Sigma-Aldrich), with hydroiodic acid (HI) 57 wt% in water (Sigma- Aldrich), at room temperature. Typical quantities were 24 ml methylamine, 10 ml hydroiodic acid and 100 ml ethanol. Upon drying at 100 o C, a white powder was formed, which was placed overnight in a vacuum oven and rinsed with ethanol before use. To generate the perovskite solution, CH 3 NH 3 I and PbCl 2 (Sigma-Aldrich) were dissolved in anhydrous N,N-Dimethylformamide (DMF) at a 3:1 molar ratio with final concentrations of ~ 40 wt%. Perovskite deposition For devices, the perovskites were deposited onto PC scaffolds by spin-coating a 40 wt% solution at 2000 r.p.m in a nitrogen-filled glovebox. After spin-coating, samples were annealed at 100 C for 2 h. The spiro-ometad hole-transporting layer was then deposited from a 80 mm chlorobenzene solution containing additives of lithium bis(trifluoromethanesulfonyl)imide and 4-tert-butylpyridine. Finally, 50 nm gold electrodes were thermally evaporated under vacuum of ~10-6 Torr, at a rate of ~0.1nm/s, to complete the devices. Structural, optical and photovoltaic characterization Cross-section field emission scanning electron microscopy (FESEM) images of PCs and corresponding full devices were taken by using a microscope Hitachi 5200 operating at 5kV. The spectral optical response of the PCs and full devices were obtained using a light source (Ocean Optics HL-2000) as the incident beam and an integrating sphere (Labsphere RTC-060-SF). The current density voltage (J-V) curves were measured (2400 Series SourceMeter, Keithley Instruments) under 2

8 simulated AM 1.5 sunlight at 100 mwcm -2 irradiance generated by an Abet Class AAB sun 2000 simulator, with the intensity calibrated with an NREL calibrated KG5 filtered Si reference cell. The mismatch factor was calculated to be less than 1%. The solar cells were masked with a metal aperture to define the active area, typically cm 2 (measured individually for each mask) and measured in a light-tight sample holder to minimize any edge effects and ensure that the reference cell and test cell are located in the same spot under the solar simulator during measurement. External quantum efficiency (EQE) was measured via Fourier-transform photocurrent spectroscopy using a Bruker Vertex 80v Fourier Transform Interferometer with tungsten lamp source, CaF 2 beam splitter and a Stanford Research SR570 current preamplifier. Samples were calibrated to a Newport-calibrated reference silicon solar cell with a known external quantum efficiency. The solar cells were masked with a metal aperture to define the active area, typically cm -2. Optical modelling of the solar cell The perovskite solar cell, that integrates PCs in which dense and porous layers are alternated, is herein modeled through a layered structure, as depicted in Figure 1a. The incoming and outgoing media are considered to be the glass substrate and the air. In our model (full details are provided in Reference 1), the optical effects due to the presence of all layers comprising the photovoltaic device (see SEM cross sectional image in Figure 1b) are accounted, i.e. glass, FTO, perovskite-infiltrated PC, perovskite overlayer, spiro-ometad, gold and air. This theoretical model serves to describe the optical reflectance of the PC based devices shown as solid lines in Figure 2a. As in the experiments, the calculation considers a plane wave that impinges perpendicularly to the electrode and therefore to the stacking of dielectric layers. The optical constants of the different constituents of the stack, i.e. glass, FTO, spiro- OMeTAD and Au, are taken from literature. 2-4 The complex refractive index of the perovskite is attained by the simultaneous fitting of the reflectance and transmittance of thin layers of the perovskite absorber deposited over a glass substrate. The effective complex dielectric permittivity of the layers infiltrated by perovskite, which comprise the PC, are approximated by the volume-weighted average of the permittivity of each component within the layer, i.e. SiO 2 or TiO 2 and perovskite. To obtain the reflectance of the cell, the amplitude of the electric field in the incident medium is calculated by solving the set of equations established by imposing the 3

9 continuity of the electric and magnetic field across each interface using a method based on the transfer matrix formalism. A linear least-squares method is employed to fit the experimental reflectance. We consider the thickness of the perovskite overlayer, the thickness of the spiro-ometad layer and the thickness of the layers comprising the unit cell of the PC as fitting parameters. Figure 2a shows the comparison between the reflectance spectra measured and calculated according to the method described herein. The color hue of a PC-based cell is obtained from its reflectance spectrum. The region delimited by a dashed line in the CIE 1931 chromaticity diagram shown in Figure 2d represents the PC gamut. It encompasses the different hues that can be achieved tuning the unit cell of the PC. Using the optical model previously described, we have calculated the reflectance spectrum of cells in which the thickness of each constituent layer of the PC was varied from 10 to 150 nm. As a consequence, the lattice parameter of the unit cell ranges from 20 to 300 nm. The porosities of the high and low index materials are considered to be 4% and 50%, respectively. Figure S1. Device performance of perovskite solar cells based on PCs comprising alternated both mesoporous nanoparticle-based TiO 2 and SiO 2 layers with lattice parameter of 180 nm and total multilayer thickness ca. 800 nm, for which performance is significantly poorer. 4

10 Figure S2. (a) Backscattered electrons field emission scanning electron microscope image of the photonic crystal (PC) scaffold infiltrated by the perovskite. Brighter areas within nanoparticle-based SiO 2 layers are due to perovskite infiltration. Perovskite accumulation over TiO 2 layers can be also observed. (b, c) Reflectance spectra of two PC scaffolds before (gray lines) and after the infiltration with perovskite, which lead to a (b) blue-green and (c) orange device. Figure S3. (a) Calculated absorptance and reflectance spectra of the PC-based orange (orange lines) and blue-green (dark cyan lines) perovskite solar cells. Solid lines represent total absorptance. Dashed lines represent the fraction of the incident light absorbed by the perovskite. (b) Spectral dependence of the fraction of light absorbed by the FTO (solid lines), spiro-ometad (short dashed lines), and gold (dashed lines) in PC-based orange (orange lines) and blue-green (dark cyan lines) perovskite solar cells. 5

11 Figure S4. CIE 1931 chromaticity space, showing the calculated color hues of the PC-based perovskite cells for the range of angles of light incidence comprised between 0º and 60º with respect to the surface normal. Two cases are evaluated, when the PC is integrated within the device (solid line) and when the light impinges directly on the photonic crystal (dashed line), a hypothetical case in which no FTO coated glass is considered. Please notice that the color change perceived from different angles is much larger in the latter case. Figure S5. Device performance of reference perovskite solar cells realized employing ~ 370 nm thick mesoporous SiO 2 as scaffold. 6

12 Figure S6. Digital camera pictures of the colorful devices integrating different PCs. Figure S7. (a-d) Current-density/voltage (J-V) curves of the typical device using (a) PC and (c) mesoporous SiO 2 as scaffold, measured under simulated AM 1.5 sunlight of 100 mw/cm 2 irradiance, and the corresponding stabilized current density and power output of the same device using (b) PC and (d) mesoporous SiO 2 as scaffold. All J-V scans were performed from forward bias to short circuit (FB-SC) and from short circuit to forward bias (SC-FB) at a scan rate of 0.15 V/s. 7

13 References (1) Anaya, M.; Lozano, G.; Calvo, M. E.; Zhang, W.; Johnston, M. B.; Snaith, H. J.; Miguez, H. J. Phys. Chem. Lett. 2015, 6, (2) Wenger, S.; Schmid, M.; Rothenberger, G.; Gentsch, A.; Gratzel, M.; Schumacher, J. O. J. Phys. Chem. C 2011, 115, (3) Moule, A. J.; Snaith, H. J.; Kaiser, M.; Klesper, H.; Huang, D. M.; Grätzel, M.; Meerholz, K. J. Appl. Phys. 2009, 106, (4) Palik, E. D., Handbook of optical constants of solids. Academic Press: Orlando,