The metal organic halide perovskites are attractive

Transcription

1 Letter pubs.acs.org/jpcl Surface Photovoltage Spectroscopy Study of Organo-Lead Perovskite Solar Cells Lee Barnea-Nehoshtan, Saar Kirmayer, Eran Edri, Gary Hodes,* and David Cahen* Department of Materials and Interfaces, Faculty of Chemistry, Weizmann Institute of Science, Rehovot, Israel *S Supporting Information ABSTRACT: The field of organo-lead perovskite absorbers for solar cells is developing rapidly, with open-circuit voltage of reported devices already approaching the maximal theoretical voltage. Obtaining such high voltages on spun-cast or evaporated thin films is intriguing and calls for detailed investigation of the source of photovoltage in those devices. We present here a study of the roles of the selective contacts to methylammonium lead iodide chloride (MAPbI 3 x Cl x ) using surface photovoltage spectroscopy. By depositing and characterizing each layer at a time, we show that the electron-extracting interface is more than twice as effective as the hole-extracting interface in generating photovoltage, for several combinations of electrode materials. We further observe the existence of an electron-injection related spectral feature at 1.1 ev, which might bear significance for the cell s operation. Our results illustrate the usefulness of SPV spectroscopy in highlighting gaps in cells efficiency and for deepening the understanding of charge injection processes in perovskite-based photovoltaics. SECTION: Spectroscopy, Photochemistry, and Excited States The metal organic halide perovskites are attractive candidates for absorber materials in a solar cell: 1 they have a sharp optical absorption edge and high absorption coefficients in the visible spectrum range, are chemically diverse (allowing tunability of the absorption edge), 2,3 are solutionprocessable, and form polycrystalline films composed of highquality crystallites. This high quality leads to high electronic quality, with relatively low defect density, which contributes to their long charge-carrier diffusion lengths. 4 6 Solar cells, based on these interesting materials were reported with different configurations to give efficiencies, now approaching 18%. 7 Methylammonium lead iodide, synthesized with or without lead chloride as one of the starting materials (MAPbI 3 x Cl x or MAPbI 3, respectively), has been used to form highly efficient solar cells through various deposition techniques and device configurations Although quite efficient cells were made also without a hole-transporting material, the most efficient cells to date were implemented with both electronselective (e.g., TiO 2, ZnO, C60-derivatives) and hole-selective (e.g., doped spiro-meotad, polytriarylamine) contacting materials. The role(s) of the hole-transporting material (HTM) and electron-transporting material (ETM) can be partially understood within the p-i-n model that was used to explain charge collection microscopic data of planar device cross sections. 14,15 Yet, the exact contributions of each selective contact to the photovoltaic action and, specifically to the open circuit voltage (V OC ), were not investigated directly. Because, in general, these devices can show small (E gap qv OC ), it is of special interest and importance to try and identify the V OC loss processes so as to further improve the device efficiencies. Surface photovoltage (SPV) spectroscopy, measured using a macroscopic Kelvin probe setup, provides information on the role of the two charge-extracting interfaces that is complementary to the information gleaned from the dynamic approaches such as transient optical absorption 16 and timeresolved photoluminescence. A surface photovoltage is defined as the difference between the surface potential of a sample in the dark and under illumination. A nonzero SPV indicates redistribution of photogenerated free charges. 17 In samples of inorganic semiconductors with crystalline domains on a scale larger than a few to tens of nanometers, SPV most commonly results from photoinduced decrease in a built-in electrical potential gradient and flattening of the bands (canceling of the band bending) in a space-charge region, whether near the free surface of a semiconductor or around a buried interface. 18,19 The macroscopic Kelvin-probe method is contactless, nondestructive, and does not impose constraints such as transparent substrate. (See the Supporting Information for detailed description of the technique.) This allows measurement of the photovoltage of partial device structures and analysis of the contribution of each contact to the voltage. The SPV of a complete solar cell is of particular interest, as it closely corresponds to the open-circuit voltage under the same illumination conditions. 20,21 Spectrally resolved SPV further extends the information that can be gleaned from these measurements and can provide detailed insight into the presence and effect of photoelectrically active intragap states. These could include interface states at the Received: June 8, 2014 Accepted: June 23, 2014 Published: June 23, American Chemical Society 2408 dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

2 The Journal of Physical Chemistry Letters absorber/htm or absorber/etm heterointerfaces, surface states at the absorber grain boundaries, or bulk defect states. Following the response of the SPV spectrum to changes in the interfacing electrodes or changes in the fabrication process can help identify the origin of specific charge-separation processes. 22 As part of our efforts to understand the mode of action of the perovskite-based solar cells, we report here results of our measurements of the SPV spectra of methylammonium lead iodide chloride (MAPbI 3 x Cl x ) perovskite absorber layers with several types of HTMs and ETMs. Figure 1 shows the SPV spectrum of a (by now) conventional flat spiro-ometad/mapbi 3 x Cl x /TiO 2 /FTO Figure 1. Surface photovoltage spectrum of spiro-ometad/ MAPbI 3 x Cl x /TiO 2 /FTO cell (black) and the derivative of the spectrum with respect to photon wavelength (red), showing the band-edge transition at 1.5 ev and a composite IR feature at 1.0 to 1.2 ev. cell, 14 illuminated from the spiro-ometad side. The spectrum shows two prominent features, viz., the optical band edge at 800 nm ( band-edge transition ) and a shallow composite feature at nm ( IR feature ). Because both features stem from light-induced redistribution of free charge carriers at high-intensity regions of the solar spectrum, they may be relevant to the cell s performance. The band-edge transition is sharp, with a full width at halfmaximum of the derivative of 40 nm. This width is comparable to the transition widths measured on high-quality Letter crystalline samples such as GaAs 23 and InP, 24 indicating a low density of gap states tailing from the band edges. This observation is in agreement with previous measurements on MAPbI 3 samples using other experimental methods. 25,26 The complete cell of the type that we studied involves two selective contacts: TiO 2 on the electron-extracting side and spiro-ometad on the hole-extracting side. Both interfaces may potentially induce band bending in the MAPbI 3 x Cl x and thus contribute to the SPV. The existence of band-bending was 27 indeed reported at the MAPbI 3 on mesoporous TiO 2 interface but was excluded for the MAPbI 3 x Cl x /spiro- OMeTAD interface. 28 However, band bending is determined by the Fermi-level alignment across the interface. Unlike what is the case for the spiro-ometad that was deposited in situ for the XPS studies in ref 27, in the current study, the cells were kept in air overnight. This allows doping of the spiro-ometad by oxygen from air, prior to measurement, as is done for fabricating efficient cells. 10,29 Such doping is expected to shift the Fermi level and hence the energy level alignment across the interface. We note in passing that the SPV can result from mechanisms other than band bending, as we discuss in the Supporting Information (SI). The signals due to the electron drift toward the ETM and of the hole drift toward the HTM are summed up in the SPV regardless of the order of layer stacking, as depicted in Figure 2a,b. To isolate the independent contributions of each of the two junctions to the photovoltage, we compare the SPV spectrum before and after the deposition of an ETM or HTM on top of the MAPbI 3 x Cl x. To decrease the sensitivity of the interpretation of the SPV measurements to possible cross-talk between the newly created top junction and the bottom junction, 30 we compare the case of ETM/MAPbI 3 x Cl x /HTM to that of HTM/MAPbI 3 x Cl x /ETM. Figure 3 shows the SPV of an inverted stack of MAPbI 3 x Cl x on an HTM, poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS) (Figure 3a) or spiro- OMeTAD, (Figure 3b) before and after deposition of a top TiO 2 layer for electron extraction. We first consider the incomplete inverted cells, that is, those with a free MAPbI 3 x Cl x Figure 2. Band diagrams showing the charge accumulation at the side of the sample that was illuminated (the side facing the Kelvin probe): (a) Complete cell in conventional geometry. Holes diffuse to the surface from both the partially shaded absorber/etm interface and the HTM/ absorber interface to yield positive SPV. (b) Complete cell in inverted geometry; electrons diffuse to the surface from both the partially shaded absorber/htm interface and ETM/absorber interface to yield negative SPV. (c) Incomplete cell in conventional geometry; holes diffuse to the surface from the partially shaded absorber/etm interface and the space-charge region near the free surface to yield positive SPV. (d) Incomplete cell in inverted geometry; electrons diffuse from the absorber/htm interface, and holes diffuse from the space-charge region near the free surface. The net SPV depends on the relative efficiency of charge separation at the surface versus that at the interface dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

3 The Journal of Physical Chemistry Letters Letter Figure 3. Surface photovoltage spectra of (a) MAPbI 3 x Cl x /PEDOT:PSS/FTO and (b) MAPbI 3 x Cl x /spiro-ometad/fto with (black) and without (red) a top layer of TiO 2. The photovoltage change across the band edge at 780 nm is 12 times stronger for the sample with TiO 2 layer compared with that of the sample with HTM layer only. The decrease in SPV at 1.1 ev is apparent only if the TiO 2 is present. surface (cf. Figure 2d). A negative SPV develops at the bandedge transition with both HTM types, indicating that the surface potential has decreased by illuminating the stack. The free surface of an n-type semiconductor is expected to bend upward, and, upon supra-band gap illumination, generate a positive SPV signal, as the photogenerated free electrons reduce the depletion. Previously we observed such upward band bending at the grain boundaries of MAPbI 3 using scanning Kelvin probe microscopy. 15 Therefore, the negative sign of the SPV of the incomplete inverted cells indicates a higher efficiency of charge separation at the buried HTM interface than at the free surface. The canceling out of free charges originating from the HTM/MAPbI 3 x Cl x interface and from the MAPbI 3 x Cl x surface reduces the amplitude of the absorption edge transition of the incomplete inverted cell to a few millivolts. The IR feature around 1100 nm is further reduced to below the instrumental sensitivity of 1 mv. When a TiO 2 layer is deposited on top of the MAPbI 3 x Cl x (cf. Figure 2b), the SPV at the band edge transition increases by an order of magnitude, and the IR feature is restored. This indicates that the ETM/MAPbI 3 x Cl x interface is the dominant one for photovoltage generation and therefore is the dominant contributor to the complete cell voltage under operation conditions. In addition, Figure 4 shows the SPV of MAPbI 3 x Cl x on TiO 2 ( conventional configuration), where deposition of p- doped spiro-ometad as top HTM contributes up to 30% of the overall photovoltage at the band-edge transition. The dominance in charge separation of the TiO 2 /MAPbI 3 x Cl x over the MAPbI 3 x Cl x /spiro-ometad interface is clear, especially considering that <30% of the supra-bandgap photons reach the back TiO 2 interface. (See Figure S2a in the SI for the MAPbI 3 x Cl x transmittance spectrum.) Consequently, the initial population of excited electrons in the immediate vicinity of the interface with TiO 2 is smaller than that near the interface with spiro-ometad. The dominance of the TiO 2 interface was also observed by transient absorption spectroscopy, 31 where 70% of the charge-separation signal was attributed to electron injection into TiO 2. Furthermore, our observation agrees with the higher charge-extracting efficiency at the ETM than at the Figure 4. Surface photovoltage spectra of MAPbI 3 x Cl x /TiO 2 /ITO with (red) and without (black) a top layer of spiro-ometad. The addition of the HTM contributes 30% of the overall photovoltage above the band gap. HTM interface, as was previously found to be the case in scanning electron beam-induced current experiments. 14 The red shift of the band-edge transition that follows the deposition of the HTM (cf. Figure 4) can be explained by field-assisted photoexcitation at the HTM/MAPbI 3 x Cl x interface, known as the Franz Keldysh effect. 32 The shift of the SPV cutoff indicates the existence of stronger internal field at the spiro- OMeTAD/MAPbI 3 x Cl x interface than at the free surface of the MAPbI 3 x Cl x, prior to HTM deposition. (See the SI for further discussion.) We note that a red shift may also occur due to the creation of shallow trap states upon top-layer deposition. However, if that were the case then we would expect a broadening of the transition, which is not observed. Photovoltaic efficiency surveys of CH 3 NH 3 PbI 3 -based devices with various HTM materials showed an increase of 20 40% in open-circuit voltage upon the introduction of an HTM dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

4 The Journal of Physical Chemistry Letters Letter The observation of relative contributions of ETM and HTM interfaces is not unique to TiO 2 ; C60 and its derivatives were reported as alternative ETMs in an efficient perovskite-based solar cell. 36,37 In Figure 5, C60 is used as ETM rather than TiO 2. The addition of C60 increases the SPV signal magnitude at the Figure 6. Surface photovoltage spectra of MAPbI 3 x Cl x with various electron-transport layers: TiO 2 as top (red) or bottom (blue) electrode, C60 (green), and SnO 2 (black). The IR feature has the same spectral position for all layers. Inset: absolute SPV spectra. The shift of SPV onset at the band edge is possibly due to Franz Keldysh effect or to morphological changes within the MAPbI 3 x Cl x film. Figure 5. Surface photovoltage spectra of MAPbI 3 x Cl x /PEDOT:PSS/ FTO with (black) and without (red) a top layer of C60. absorption edge by an order of magnitude (from <1 mv to >16 mv). Furthermore, the IR spectral feature that previously appeared only when TiO 2 is in contact with MAPbI 3 x Cl x,is observed with C60 as ETM as well. A similar broad feature at nm was also observed in photoinduced absorption spectroscopy of the MAPbI 3 /TiO 2 interface and was attributed to an electron-transfer process from the perovskite to the TiO Figure 6 shows, however, that this feature is not specific for TiO 2 and is also observed in junctions of MAPbI 3 x Cl x with SnO 2 and C60. The presence of the ETM enhances the IR feature not only in the SPV spectrum but also in the optical transmittance spectrum (Figure S2b in the SI), possibly due to field-assisted absorption. The spectral position of the 1100 nm feature is invariant under change of ETM type. Moreover, the IR features appear at the same position also if MAPbI 3 is used as absorber instead and appears to be insensitive to process parameters, such as the solvents used and the initial inclusion of Cl or HI additives. (See the SI for further details.) However, the magnitude of the feature depends on the absorber morphology (Figures S5 and S6 in the SI). We speculate that the feature is related to a reduced effective band gap at the grain boundaries, as suggested by the IR response of the grain boundaries when scanned with Kelvin-probe microscope (Figure S8 in the SI). Better understanding of the source of this feature and its relation to the performance of a working cell requires further investigation. To summarize, we have studied the contributions of the electron- and hole-transport materials to the photovoltage of a solar cell based on a MAPbI 3 x Cl x perovskite absorber. Using the Kelvin-probe technique, we show directly that the electronselective junction contributes most of the solar-illuminationinduced photovoltage. We therefore confirm that the electronselective junction is dominant in determining the open-circuit voltage. Furthermore, we show that this junction is related to a spectral feature at 1.1 ev, the origin and photovoltaic importance of which remains to be determined. Our results illustrate the usefulness of the relatively simple SPV spectroscopy in enhancing understanding of charge injection processes in perovskite-based photovoltaics. EXPERIMENTAL METHODS Sample Fabrication. The preparation of fluorine-doped tin oxide (FTO) and compact TiO 2 substrate as well as preparation and deposition of MAPbI 3 x Cl x were described elsewhere. 28 The film thickness was not uniform and varied in the range of nm. TiO 2 as a top layer (i.e., deposited onto MAPbI 3 x Cl x ) was spin-coated from a sol gel solution prepared according to a published literature procedure. 39 In short, titanium isopropoxide was mixed with a dilute HCl solution in isopropanol to yield a M concentration solution. The solution was filtered with a PTFE filter with 0.22 μm pore size before use. The resulting TiO 2 thickness was nm. A film of hole conductor was deposited onto MAPbI 3 x Cl x by spin-coating at 2000 rpm for 45 s an 80 mg/ml Spiro- MeOTAD solution in chlorobenzene to which 7 μl oftert- butylpyridine and 15 μl of LiTFSI solution (170 mg/ml in acetonitrile) were added. The resulting film thickness was nm, depending on the thickness of the underlying MAPbI 3 x Cl x. A 30 nm top layer of C60 was deposited onto the MAPbI 3 x Cl x layer by thermal evaporation. For the inverted cell configuration, glass coated with indium tin oxide (ITO) was used as substrate. It was cleaned by solvent sonication before spin-coating of the HTM layer. Spiro- MeOTAD was applied as previously described and left overnight in the dark in dry air (relative humidity <20%) to 2411 dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

5 The Journal of Physical Chemistry Letters Letter allow (p-)doping by oxygen from air. A film of 100 nm of PEDOT:PSS (Clevios PH-1000, Heraeus), doped with 5% ethylene glycol and 0.4% zonyl surfactant, was deposited by spin-coating and annealed for 30 min at 150 C in a dry N 2 environment. SPV Measurements. The contact potential difference (CPD) between the sample and a gold grid reference electrode was measured in a Kelvin-probe configuration using a Besocke Delta-Phi probe and controller. The sample was illuminated from a 600 W tungsten halogen lamp with wavelength selection by 1/4 m monochromator. The wavelength resolution was 0.6 nm. The SPV was defined as {CPD(dark) CPD(light)}. The samples were allowed to stabilize in the dark for 3 h prior to measurement. The dwelling time on each wavelength was a few minutes, until the signal stabilized (i.e., changed <0.3 mv over the time span of 1 min). The scan direction was from long to short wavelength to minimize possible effects of radiative recombination from long-lived trap states. To avoid degradation of the perovskite layer, we conducted the measurement in a dry N 2 environment in a glovebox. ASSOCIATED CONTENT *S Supporting Information Overview of the Kelvin probe technique; transmittance spectra; discussion on possible sources of SPV signal other than band bending; discussion of the source of red-shift of the band edge; discussion of the relations between morphology and IR feature. This material is available free of charge via the Internet at AUTHOR INFORMATION Corresponding Authors * gary.hodes@weizmann.ac.il. * david.cahen@weizmann.ac.il. Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS We thank Leeor Kronik (WIS) for illuminating discussions, Sabyasachi Mukhopadhyay (WIS) for conducting scanning Kelvin probe measurements, Arie Zaban (Bar-Ilan Univ., Israel) for providing the SnO 2 samples, and the Leona M. and Harry B. Helmsley Charitable Trust, the Israel Ministry of Science s Tashtiot program, the Israel National Nano-Initiative s Focused Technology Area program, the Nancy and Stephen Grand Center for Sensors and Security, and the Weizmann- U.K. Joint Research Program for support. D.C. holds the Sylvia and Rowland Schaefer Chair in Energy Research. REFERENCES (1) Snaith, H. J. Perovskites: The Emergence of a New Era for Low- Cost, High-Efficiency Solar Cells. J. Phys. Chem. Lett. 2013, 4, (2) Noh, J. H.; Im, S. H.; Heo, J. H.; Mandal, T. N.; Seok, S. I. Chemical Management for Colorful, Efficient, and Stable Inorganic Organic Hybrid Nanostructured Solar Cells. Nano Lett. 2013, 13, (3) Eperon, G. E.; Stranks, S. D.; Menelaou, C.; Johnston, M. B.; Herz, L. M.; Snaith, H. J. Formamidinium Lead Trihalide: A Broadly Tunable Perovskite for Efficient Planar Heterojunction Solar Cells. Energy Environ. Sci. 2014, 7, (4) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.; Alcocer, M. J. P.; Leijtens, T.; Herz, L. M.; Petrozza, A.; Snaith, H. J. Electron-Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber. Science 2013, 342, (5) Wehrenfennig, C.; Eperon, G. E.; Johnston, M. B.; Snaith, H. J.; Herz, L. M. High Charge Carrier Mobilities and Lifetimes in Organolead Trihalide Perovskites. Adv. Mater. 2014, 26, (6) Xing, G.; Mathews, N.; Sun, S.; Lim, S. S.; Lam, Y. M.; Gratzel, M.; Mhaisalkar, S.; Sum, T. C. Long-Range Balanced Electron- and Hole-Transport Lengths in Organic-Inorganic CH 3 NH 3 PbI 3. Science 2013, 342, (7) National Center for Photovoltaics. NREL Best Research-Cell Efficiencies Chart; National Renewable Energy Laboratory: Boulder, CO, (8) Liu, D.; Kelly, T. L. Perovskite Solar Cells with a Planar Heterojunction Structure Prepared Using Room-Temperature Solution Processing Techniques. Nat. Photonics 2013, 8, (9) Burschka, J.; Pellet, N.; Moon, S.-J.; Humphry-Baker, R.; Gao, P.; Nazeeruddin, M. K.; Graẗzel, M. Sequential Deposition as a Route to High-Performance Perovskite-Sensitized Solar Cells. Nature 2013, 499, (10) Liu, M.; Johnston, M. B.; Snaith, H. J. Efficient Planar Heterojunction Perovskite Solar Cells by Vapour Deposition. Nature 2013, 501, (11) Etgar, L.; Gao, P.; Xue, Z.; Peng, Q.; Chandiran, A. K.; Liu, B.; Nazeeruddin, M. K.; Graẗzel, M. Mesoscopic CH 3 NH 3 PbI 3 /TiO 2 Heterojunction Solar Cells. J. Am. Chem. Soc. 2012, 134, (12) Aharon, S.; Gamliel, S.; Cohen, B. E.; Etgar, L. Depletion Region Effect of Highly Efficient Hole Conductor Free CH 3 NH 3 PbI 3 Perovskite Solar Cells. Phys. Chem. Chem. Phys. 2014, 16, (13) Shi, J.; Luo, Y.; Wei, H.; Luo, J.; Dong, J.; Lv, S.; Xiao, J.; Xu, Y.; Zhu, L.; Xu, X.; Wu, H.; Li, D.; Meng, Q. Modified Two-Step Deposition Method for High-Efficiency TiO 2 /CH 3 NH 3 PbI 3 Heterojunction Solar Cells. ACS Appl. Mater. Interfaces 2014, DOI: / am502131t. (14) Edri, E.; Kirmayer, S.; Mukhopadhyay, S.; Gartsman, K.; Hodes, G.; Cahen, D. Elucidating the Charge Carrier Separation and Working Mechanism of CH 3 NH 3 PbI 3 x Cl x Perovskite Solar Cells. Nat. Commun. 2014, 5, (15) Edri, E.; Kirmayer, S.; Henning, A.; Mukhopadhyay, S.; Gartsman, K.; Rosenwaks, Y.; Hodes, G.; Cahen, D. Why Lead Methylammonium Tri-Iodide Perovskite-Based Solar Cells Require a Mesoporous Electron Transporting Scaffold (but Not Necessarily a Hole Conductor). Nano Lett. 2014, 14, (16) 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. Lead Iodide Perovskite Sensitized All- Solid-State Submicron Thin Film Mesoscopic Solar Cell with Efficiency Exceeding 9%. Sci. Rep. 2012, 2, 591. (17) Kronik, L.; Shapira, Y. Surface Photovoltage Phenomena: Theory, Experiment, and Applications. Surf. Sci. Rep. 1999, 37, (18) Gal, D.; Beier, J.; Moons, E.; Hodes, G.; Cahen, D.; Kronik, L.; Burstein, L.; Mishori, B.; Leibovitch, M.; Shapira, Y. Band Diagram and Band Line-up of the Polycrystalline CdS/Cu(In,Ga)Se2 Heterojunction and Its Response to Air Annealing. AIP Conf. Proc. 1996, 353, (19) Gross, D.; Mora-Sero, I.; Dittrich, T.; Belaidi, A.; Mauser, C.; Houtepen, A. J.; Como, E. D.; Rogach, A. L.; Feldmann, J. Charge Separation in Type II Tunneling Multilayered Structures of CdTe and CdSe Nanocrystals Directly Proven by Surface Photovoltage Spectroscopy. J. Am. Chem. Soc. 2010, 132, (20) Kronenberg, N. M.; Deppisch, M.; Wu rthner, F.; Lademann, H. W. A.; Deing, K.; Meerholz, K. Bulk Heterojunction Organic Solar Cells Based on Merocyanine Colorants. Chem. Commun. 2008, dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

6 The Journal of Physical Chemistry Letters Letter (21) Lee, Y.-J.; Wang, J.; Hsu, J. W. P. Surface Photovoltage Characterization of Organic Photovoltaic Devices. Appl. Phys. Lett. 2013, 103, (22) Osterloh, F. E.; Holmes, M. A.; Chang, L.; Moule, A. J.; Zhao, J. Photochemical Charge Separation in Poly(3-Hexylthiophene) (P3HT) Films Observed with Surface Photovoltage Spectroscopy. J. Phys. Chem. C 2013, 117, (23) Burstein, L.; Bregman, J.; Shapira, Y. Surface Photovoltage Spectroscopy of Gap States at GaAs and InP Metal Interfaces. Appl. Phys. Lett. 1990, 57, (24) Bastide, S.; Gal, D.; Cahen, D.; Kronik, L. Surface Photovoltage Measurements in Liquids. Rev. Sci. Instrum. 1999, 70, (25) De Wolf, S.; Holovsky, J.; Moon, S.-J.; Lo per, P.; Niesen, B.; Ledinsky, M.; Haug, F.-J.; Yum, J.-H.; Ballif, C. Organometallic Halide Perovskites: Sharp Optical Absorption Edge and Its Relation to Photovoltaic Performance. J. Phys. Chem. Lett. 2014, 5, (26) Yamada, Y.; Nakamura, T.; Endo, M.; Wakamiya, A.; Kanemitsu, Y. Near-Band-Edge Optical Responses of Solution-Processed Organic inorganic Hybrid Perovskite CH 3 NH 3 PbI 3 on Mesoporous TiO 2 Electrodes. Appl. Phys. Express 2014, 7, (27) Laban, W. A.; Etgar, L. Depleted Hole Conductor-Free Lead Halide Iodide Heterojunction Solar Cells. Energy Environ. Sci. 2013, 6, (28) Schulz, P.; Edri, E.; Kirmayer, S.; Hodes, G.; Cahen, D.; Kahn, A. Interface Energetics in Organo-Metal Halide Perovskite-Based Photovoltaic Cells. Energy Environ. Sci. 2014, 7, (29) Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith, H. J. Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites. Science 2012, 338, (30) Itzhaik, Y.; Hodes, G.; Cohen, H. Band Alignment and Internal Field Mapping in Solar Cells. J. Phys. Chem. Lett. 2011, 2, (31) Marchioro, A.; Teuscher, J.; Friedrich, D.; Kunst, M.; van de Krol, R.; Moehl, T.; Graẗzel, M.; Moser, J.-E. Unravelling the Mechanism of Photoinduced Charge Transfer Processes in Lead Iodide Perovskite Solar Cells. Nat. Photonics 2014, 8, (32) Seeger, K. Semiconductor Physics: An Introduction; Springer: Berlin, (33) Heo, J. H.; Im, S. H.; Noh, J. H.; Mandal, T. N.; Lim, C.-S.; Chang, J. A.; Lee, Y. H.; Kim, H.; Sarkar, A.; Nazeeruddin, M. K.; et al. Efficient Inorganic-Organic Hybrid Heterojunction Solar Cells Containing Perovskite Compound and Polymeric Hole Conductors. Nat. Photonics 2013, 7, (34) Jeon, N. J.; Lee, J.; Noh, J. H.; Nazeeruddin, M. K.; Graẗzel, M.; Seok, S. I. Efficient Inorganic Organic Hybrid Perovskite Solar Cells Based on Pyrene Arylamine Derivatives as Hole-Transporting Materials. J. Am. Chem. Soc. 2013, 135, (35) Juarez-Perez, E. J.; Wuβler, M.; Fabregat-Santiago, F.; Lakus- Wollny, K.; Mankel, E.; Mayer, T.; Jaegermann, W.; Mora-Sero, I. Role of the Selective Contacts in the Performance of Lead Halide Perovskite Solar Cells. J. Phys. Chem. Lett. 2014, 5, (36) Malinkiewicz, O.; Yella, A.; Lee, Y. H.; Espallargas, G. M.; Graetzel, M.; Nazeeruddin, M. K.; Bolink, H. J. Perovskite Solar Cells Employing Organic Charge-Transport Layers. Nat. Photonics 2014, 8, (37) You, J.; Hong, Z.; Yang, Y.; Chen, Q.; Cai, M.; Song, T.-B.; Chen, C.-C.; Lu, S.; Liu, Y.; Zhou, H.; Yang, Y. Low-Temperature Solution-Processed Perovskite Solar Cells with High Efficiency and Flexibility. ACS Nano 2014, 8, (38) Bi, D.; Yang, L.; Boschloo, G.; Hagfeldt, A.; Johansson, E. M. J. Effect of Different Hole Transport Materials on Recombination in CH 3 NH 3 PbI 3 Perovskite-Sensitized Mesoscopic Solar Cells. J. Phys. Chem. Lett. 2013, 4, (39) Docampo, P.; Ball, J. M.; Darwich, M.; Eperon, G. E.; Snaith, H. J. Efficient Organometal Trihalide Perovskite Planar-Heterojunction Solar Cells on Flexible Polymer Substrates. Nat. Commun. 2013, 4, dx.doi.org/ /jz501163r J. Phys. Chem. Lett. 2014, 5,

7 Supplementary Information for Surface Photovoltage Spectroscopy Study of Organo-Lead Perovskite Solar Cells Lee Barnea-Nehoshtan, Saar Kirmayer, Eran Edri, Gary Hodes*, David Cahen* Dept. of Materials and Interfaces, Weizmann Inst. of Science, Rehovot, 76100, Israel 1

8 1. Surface photovoltage measurement using the Kelvin probe technique The Kelvin probe technique 1 probes the relative work function of a sample by measuring the contact potential difference (CPD) between the sample and a probe with a known work function. A conducting probe is suspended near the surface of the sample and connected electrically to the other side of the sample (see scheme S1), such that a plate capacitor is created. The probe is vibrated mechanically so that the distance between the sample surface and the probe changes periodically, thus inducing an AC displacement current in the circuit, which is proportional to the CPD. Then, a DC bias is applied, and its magnitude is adjusted so that the AC current is nulled, which occurs if the applied bias equals the CPD. The surface photovoltage (SPV) is deduced from the change in CPD when the sample is illuminated. In our system, the probe is made of gold wire mesh to allow illumination of the sample through the probe. The area of the illumination spot is larger than the probe area, to avoid edge effects. Three conditions are required for obtaining a non-zero SPV: (1) the incident photon has to have an energy equal to an allowed optical transition, (2) the excited electron-hole pair are separated within the measurement period and (3) The concentration of free carriers and resulting field are high enough to be detectable with given instrumental sensitivity. Following these accumulative conditions, the SPV spectrum would present a knee at the edge of a transition between two continuum levels or a localized state and a continuum level (a sharp peak will appears in the case of quantized system, not discussed here). At photon energies higher than the absorption edge of a specific transition, the value of SPV may further change due to 2

9 changes in the photon flux or the absorption coefficient. Therefore, any supra-band gap spectral features should be compared to the photon flux spectrum feature to rule out instrumental artifacts.(see, for example, the decrease in the absolute value of SPV for E > 1.7 ev in figs. 1 and 5 of the main text, which is related to the decrease in photon flux in this spectral region). Scheme S1. Kelvin probe setup. A gold mesh probe is vibrating close (~ 1 mm) to the sample. The sample is illuminated through the mesh. A variable voltage source is used to apply a bias between the probe and the sample, while a lock-in amplifier is used to track the AC displacement current. 3

10 Transmittance [%] Transmittance [%] 2. Optical transmittance spectra (a) 400 Photon Wavelength [nm] Photon Energy [ev] (b) Wavelength [nm] Figure S1. Transmittance spectra of (a) MAPbI3 -x Cl x /TiO 2 /FTO, corrected for the FTO transmittance and (b) MAPbI 3-x Cl x /PEDOT:PSS/FTO with (black, top) and without (red, bottom) top layer of C60. The transmission is not corrected for reflection. 4

11 3. Source of SPV While SPV is most commonly related to band bending, there are alternative mechanisms that can cause SPV. One such mechanism to be considered for low-doped semiconductors is photodoping 2. In the presence of a selective contact (for holes or electrons) the photogenerated charges are separated, and an electric potential difference builds up across the stack. If the density of photocharges is significant with respect to the intrinsic density of charges in the MAPbI 3-x Cl x, a measurable shift of the work function would occur. Figure S2 shows a logarithmic dependence of the SPV of MAPbI 3- xcl x /TiO 2 junction on the illumination intensity (at wavelengths longer than the band edge of the TiO 2 ).This can be interpreted either as typical for band flattening of an n-type semiconductor, or as p-type like photodoping. However, the red shift of SPV onset that is observed upon moving from incomplete to complete cells (cf. Figures 3, 4 and 5 of the main text) indicates very strong internal electric fields in the MAPbI 3-x Cl x near the contacts, as were also identified by us in electron beam induced current microscopy experiments on cell cross- sections 3,4 (See section 4 below for further discussion). Such fields indicate band bending, rather than the field due to the electrochemical potential gradient that results from photodoping. Therefore, photodoping alone cannot explain the red-shift. Another possible mechanism is the creation of a Dember potential, a photo-induced imbalance of hole and electron populations 5 due to, for example, unequal diffusion rate towards selective contacts 6. In the case of MAPbI 3-x Cl x absorber, the diffusion lengths 7 of both holes (1200 nm) and electrons (1100nm) are larger than the layer thickness 4. Figure S3 shows a comparison of the SPV spectrum of MAPbI 3-x Cl x on mesoporous TiO 2 with that of the same layer over flat TiO 2. The required diffusion length for charge separation is shorter on the mesoporous compared to with the flat TiO 2. Nevertheless, the mesoporous TiO 2 based sample shows no enhancement of the SPV relative to the flat TiO 2, as would be expected in the case of a Dember potential. Moreover, the transparency of the absorber film of more than ~20% above the band gap (cf. Figure S1) enables photoexcitation of charges at the immediate vicinity of the back interface. 5

12 CPD [V] Therefore, a Dember potential is not a plausible mechanism under the low illumination intensity used. We therefore conclude that the SPV results from changes in band-bending within the absorber layer Lamp power [W] Figure S2 Photosaturation of the CPD signal of MAPbI 3-x Cl x /TiO 2 as the intensity of incident light is increased.halogen lamp with 530 nm long pass filter was used as a light source 6

13 Surface photovoltage [mv] meso-porous TiO 2 compact TiO Wavelength [nm] Figure S3. SPV spectra of MAPbI 3-x Cl x on a layer of compact (black) or mesoporous (red) TiO 2. Both samples have a top layer of spiro-ometad. 7

14 4. Spectral shift of band edge transition in complete cell, relative to incomplete cells The Franz Keldysh effect (FKE) describes a red shift of the absorption edge of a semiconductor under a large electric field, as a result of photon-assisted tunneling from the valence to the conduction band 8. If the HTM/MAPbI 3-x Cl x or MAPbI 3-x Cl x /ETM interfaces induce band bending in the MAPbI 3-x Cl x film, the built-in field at the depletion region may suffice for detectable band-edge shift due to the FKE. Figure S5 shows a plot of the wavelength of the SPV onset of complete and incomplete devices, as a function of the SPV magnitude at h >E g. The plot shows a change of ~50 mev in band edge position correlated with a change of ~400 mv in the measured SPV at the band edge. The correlation suggests that the red shifts that appear in figures 4, 5 and 6 of the main text are a result of the FKE, and hence, indicate the existence of very strong internal electric fields. Such fields, in turn, point to band bending as the most probable cause for the SPV at both the ETM and the HTM interfaces. Interestingly, the red shifts in the spiro- OMeTAD /MAPbI 3-x Cl x / TiO 2 /ITO and in the inverted TiO 2/ MAPbI 3-x Cl x /spiro- OMeTAD/FTO cell are of the same magnitude, ~30 mev, while we would expect, all other factors being equal, a stronger red shift for TiO 2/ MAPbI 3-x Cl x interface. As we have to use the inverted geometry for our measurements of the TiO 2/ MAPbI 3-x Cl x interface, this may be due to changes in the morphology and energetics of the MAPbI 3-x Cl x deposited on spiro-ometad compared to that deposited on TiO 2. While the effective reduction of band gap should decrease the open circuit voltage, such dependence was not observed, to the best of our knowledge. This demonstrates again the sensitivity of the SPV methods to details of the energetics in the MAPbI 3-x Cl x based devices. 8

15 SPV onset [nm] SPV at band edge [mv] Figure S4. Red shift of the band edge (in nm) vs. the absolute value of SPV at supra-gap illumination of the various complete and incomplete devices described in the main text. The dashed red line is a guide to the eye. 5. Morphology and IR feature Figure S5 shows the SPV spectra of MAPbI 3 films prepared from solutions of different solvents: N,N-Dimethylformamide (DMF), Dimethyl sulfoxide (DMSO), γ- butyrolactone (GBL) and Propylene Carbonate (PC). As seen in the right-hand side of the Figure, the resulting films have different morphologies. While the position of the IR feature remains constant, its magnitude relative to the magnitude of the SPV at the band edge depends on the solvent. Similarly, Figure S6 shows the SPV spectra of films for which the nucleation process was manipulated by adding HI to the solution as seeding 9

16 agent. Here again, the SPV(IR)/SPV(band edge) changes, while the spectral positions remain unaltered. We can speculate that the change in SPV(IR) is related to a change in volume fraction of the grain boundaries. Figure S7a shows the Kelvin probe scanning microcopy image of an HI-seeded MAPbI 3 film with and without and illumination in the nm range. In both scans the sample was exposed to background illumination at 680 nm from the AFM laser. The average SPV at the grain boundaries (the position of which was determined from the topographic scan, Figure S7b) is ~20 mv, while the average SPV at the grain interior is ~0 mv, as results from a pair-wise comparison of set of matching points. Figure S5. SPV spectra of MAPbI 3 films, prepared from different solvents: DMF, GBL and PC. Insets: SPV derivative with respect to the energy. Right: optical microscope images of the films. There is apparent difference in the relative intensity of the IR feature from ~20% in the case of DMF to below measuring threshold in the case of GBL. The feature between 1.4 and 1.5 ev is an instrumental artifact. 10

17 Surface Photovoltage [mv] 700 Photon Wavelength [nm] l 35 l 10 l Photon Energy [ev] 1 Figure S6. SPV spectra of MAPbI 3 films dispensed from solution with 2% (10 l), 7% (35 l) and 9% (45 l) v/v HI added. 11

18 Figure S7. Top-view scanning probe micrograph of HI-seeded MAPbI 3 /TiO 2 /FTO (a) in Kelvin probe mode in the dark (left) and under IR illumination (wavelength > 980 nm) and (b) in topographic (AFM) mode. The grain boundaries show larger response to the light than the grain interior. The color axis is given in V. References (1) Kronik, L.; Shapira, Y. Surface Photovoltage Phenomena: Theory, Experiment, and Applications. Surf. Sci. Rep. 1999, 37, (2) Lenzmann, F.; Krueger, J.; Burnside, S.; Brooks, K.; Grätzel, M.; Gal, D.; Rühle, S.; Cahen, D. Surface Photovoltage Spectroscopy of Dye-Sensitized Solar Cells with TiO2, Nb2O5, and SrTiO3 Nanocrystalline Photoanodes: Indication for Electron Injection from Higher Excited Dye States. J. Phys. Chem. B 2001, 105, (3) Edri, E.; Kirmayer, S.; Henning, A.; Mukhopadhyay, S.; Gartsman, K.; Rosenwaks, Y.; Hodes, G.; Cahen, D. Why Lead Methylammonium Tri-IODIDE Perovskite-Based Solar Cells Requires a Mesoporous Electron Transporting Scaffold (but Not Necessarily a Hole Conductor). (4) Edri, E.; Kirmayer, S.; Mukhopadhyay, S.; Gartsman, K.; Hodes, G.; Cahen, D. Elucidating the Charge Carrier Separation and Working Mechanism of CH3NH3PbI3 xclx Perovskite Solar Cells. Nat. Commun. 2014, 5. (5) Zhao, J.; Osterloh, F. E. Photochemical Charge Separation in Nanocrystal Photocatalyst Films: Insights from Surface Photovoltage Spectroscopy. J. Phys. Chem. Lett. 2014, 5, (6) Goodman, A. M. A Method for the Measurement of Short Minority Carrier Diffusion Lengths in Semiconductors. J. Appl. Phys. 1961, 32, (7) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.; Alcocer, M. J. P.; Leijtens, T.; Herz, L. M.; Petrozza, A.; Snaith, H. J. Electron-Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber. Science 2013, 342, (8) Seeger, K. Semiconductor Physics: An Introduction; Springer Berlin Heidelberg,