Supporting Information Metal to Halide Perovskite )HaP(: an Alternative Route to HaP Coating Directly from Pb (0) or Sn (0) films Yevgeny Rakita, Satyajit Gupta, David Cahen*, Gary Hodes* Department of Materials and Interfaces, Weizmann Institute of Science, Rehovot, 7610001, Israel *Corresponding authors: gary.hodes@weizmann.ac.il ; david.cahen@weizmann.ac.il Table of contents: Page 2 Figure S1 3 Figure S2 4 Figure S3 5 Figure S4 6 Figure S5 7 Figure S6 8 Figure S7 9 Figure S8 + Conversion of Ag/AgX reference electrode potentials to the SHE potential scale 10 Figure S9 11 Figure S10 12 Figure S11 13 Table S1 14 Comment S1 + References 1
Figure S1: Cross-sectional SEM image of (i) Pb on an FTO substrate used for solar cell fabrication; (ii) Pb on glass; (iii) Sn on FTO; (iv) Sn on glass. The scale is the same (100 nm scale bar) for all images. 2
Figure S2: XRD patterns of a Pb film (deposited on: (i) glass; (ii) FTO/d-TiO 2 substrates) reacted with: (i) 100 mm MABr for ~ 24 hr without bias and (ii) 80 mm FABr solutions in IPA for 1 hr with bias of 1 V vs. Ag/AgBr and another 2 hr without bias (total 3 hours). In both cases there are no obvious signs of Pb (cf. (i) top pattern). The FAPbBr 3 pattern includes FTO/TiO 2 peaks and some other unidentified peaks. The bottom spectrum in (ii) is a simulated pattern from a crystallographic information file (CIF) 1. At this point we have not been able to identify the impurities in the FAPbBr 3 thin film. Nevertheless, it is clear that FAPbBr 3 is the main product of this process and hopefully we will be able to identify and eliminate (if necessary) impurities in future work. 3
Figure S3: (i) Photographs of Pb films (~100 nm on FTO) treated with 50 mm MAI for ~2.5 hr in IPA, EtOH and MeOH solutions. (ii) (left) SEM image of the reacted film in (i) using IPA; (right) optical microscope image of the reacted film in (i) using EtOH. (iii) SEM images of similarly reacted Pb films in 80 mm MABr in (left) IPA or (right) EtOH) for 4 hr. For reactions in MeOH, the metallic film did not change color to that of the HaP but instead its opaqueness faded with time until it completely dissolved. In all images the scale is the same (including the optical microscope image). 4
Figure S4: X-ray diffraction of a ~100 nm Pb film on glass reacted for 4 hours in a solution of 80 mm CsBr in MeOH and containing ~50 mm of HBr (accompanied with ~50% by weight of water from the HBr). 5
Figure S5: Plan-view SEM image of (i) ~100 nm thick Pb film on glass and (ii) similar film after treatment in 50 mm MAI solution of IPA for 2 hr at RT. 6
Figure S6: (i) A series of 50 mm MAI solutions in IPA with different additives (shown above each example; % means molar percentage with respect to MAI). A darker yellow/brown color indicates a higher I 2 concentration. When adding HI and TFA to the solutions, the color of the solution became more yellow. KOH makes the solution colorless. Bubbling N 2 or O 2 does not change the solution color. (ii) XRD pattern of the films shown in Figure 2 (iv-vi). The counts of the transformed films were normalized to the MAPbI 3 (110) peak. 7
Figure S7: Electrochemically-assisted Pb (~100 nm on FTO) transformation to MAPbBr 3 in 200 mm MABr in IPA at room temperature. (i) A photograph of the reaction system ~ 1 min after applying 1.20 V between the reference (R) and the working (W) electrodes. Both counter (C) and R electrodes are Pt coils. W is an evaporated film of Pb on FTO glass. The yellow cloud next to the Pb electrode is elemental, Br 2 which is yellow in IPA. (ii) Plan-view (top) and cross-section (bottom) SEM images of the electrochemically-assisted reacted films after 1 hr. (iii) XRD diffraction pattern from reacted films under similar reaction conditions but with (red) and without (blue) applying 1.20 V anodic bias to W. The Pb-{111} peak disappears after applying this bias for 1 hr, indicating an accelerated reaction rate. The inset shows a visible difference between the two films, (in which the darker color indicates residual Pb 0. 8
Figure S8: I-V scans using Pt working and counter electrodes and Ag/AgX (X=Br or I) reference electrodes in 50 mm of MAI or 80 mm MABr both dissolved in IPA. Ag/AgI and Ag/AgBr potentials versus SHE are approximated to be -0.08 V and 0 V, respectively. The X-axis gives the potential vs. SHE. Conversion of Ag/AgX reference electrode potentials to the SHE potential scale: The standard potential, E 0, of the Ag/AgI electrode in water has been given as -0.15 V. 2 In a 0.05 M iodide solution, by the Nernst equation, this will be: 0.15 + 0.06 log 10 ( 1 ) = 0.07 V 0.05 In order to account for the density differences between water and IPA, 3 another 10 mv should be included, meaning that at 0.05 M of I -, the potential of Ag/AgI vs. SHE should be -0.08 mv. For the Ag/AgBr electrode, and from the same reference, the potential in 0.08 M bromide is calculated to be ~ 0 V on the SHE scale. In view of the very small effect of changing from water to IPA on the polyiodide redox potential (see main text), we make the simplifying assumption that this holds also for the other potentials used in this paper. 9
Figure S9: Demonstration of control over Pb transformation (acceleration, deceleration or reversal) as a function of the applied bias. Photographed samples of: (i) unreacted Pb film deposited on glass; (ii) reacted Pb films deposited on glass, after 5 min in 50 mm MAI in IPA at -0.58 V bias vs. SHE (left) and disconnected from electrodes (right) (potential in solution was measured to be ~ -0.25 V (iii) MAPbI 3 on FTO (obtained after transforming Pb), reacted in a similar solution as in (i) but at -1.08 V. All potentials were measured vs. Ag/AgI and then related to the SHE scale. 10
Figure S10: Time-resolved photoluminescence of Pb films treated with IPA solutions of (a) MABr at 70 ºC and (b) MAI at RT with different additives. Reaction times varied between the different reaction solutions reaction was terminated ~ 1 hr after the metallic shine of the Pb disappeared from the backside of the glass. 11
Figure S11: (i) Cross-section SEM images of cells in which the HaP is prepared in an electrochemically-assisted process. In both cases 1 V (vs. Ag/AgI) was applied to a FTO/d-TiO 2 /Pb substrate against a Pt electrode for 20 min in (left) 50 mm MAI and (right) 80 mm MABr solutions in IPA. (ii) Dark and light (solar simulated 100 mw/cm 2 ) I-V scans of MAPbI 3 and MAPbBr 3 cells where the perovskite was formed as in (i). 12
Table S1: Summary of the trends reported in this work regarding reaction speed, average crystallite size and excited carrier lifetime as a result of changing reaction conditions (50 mm MAI or 70 mm MABr in IPA at RT). Not all combinations of experiments were carried out; the triple dashed lines ( --- ) means no data were collected. Additive Reaction speed Average crystallite size Excited carrier lifetime (long decay) with respect to no additives Lighter halide salt (e.g., MAI vs. MABr) Elemental halide (I 2 or Br 2 ) Not relevant Not relevant 10% I 2-10% Br 2 - Acid - (HX or TFA) HI - 1%- ; 5% 1% HBr - Base - KOH --- --- O 2, N 2, H 2 O No visible effect No visible effect --- --- lower alcoholic solvent (e.g., EtOH) --- --- Higher concentration --- --- Higher temperature --- --- --- Longer reaction time Not relevant --- --- Electrochemical bias (above Pb oxidation potential) --- --- --- 13
Comment S1: The fact that the Sn(II) perovskite is formed with MA as the A cation, while the Sn(IV) form results if Cs is used as the A cation, is not yet understood. On the one hand, CsSnI 3 is expected to be more stable to Sn oxidation than MASnI 3 (see, e.g. ref. 4, where results are reported that show that (MA,Cs)(Sn,Pb)I 3 is more stable to Sn oxidation than pure MA(Sn,Pb)I 3 ). On the other hand, CsSnI 3, unlike MASnI 3, forms a phase that is more stable than the black perovskite phase, the yellow non-perovskite phase that is still very susceptible to oxidation and with a different structure (one dimensional, double chain). 5 Therefore, making a direct comparison of the transformation of Sn(II) to Sn(IV) for the materials with the two different A cations is difficult and requires further study. References: (1) Leppert, L.; Reyes-Lillo, S. E.; Neaton, J. B. Electric Field- and Strain-Induced Rashba Effect in Hybrid Halide Perovskites. J. Phys. Chem. Lett. 2016, 7, 3683 3689. (2) Kundu, K. K.; Mazumdar, K. Standard Potentials of Ag-AgBr and Ag-AgI Electrodes in Urea + Water Mixtures. Free Energies and Entropies of Transfer of the Hydrogen Halides. J. Chem. Soc., Faraday Trans. 1 1975, 71, 1422 1431. (3) Sen, U.; Kundu, K. K.; Das, M. N. Standard Potentials of the Silver-Silver Chloride Electrode in Ethylene Glycol and Its Aqueous Mixtures at Different Temperatures and Related Thermodynamic Quantities. J. Phys. Chem. 1967, 71, 3665 3671. (4) Liu, X.; Yang, Z.; Chueh, C.-C.; Rajagopal, A.; Williams, S. T.; Sun, Y.; Jen, A. K.- Y. Improved Efficiency and Stability of Pb Sn Binary Perovskite Solar Cells by Cs Substitution. J. Mater. Chem. A 2016, 4, 17939 17945. (5) Chung, I.; Song, J.-H.; Im, J.; Androulakis, J.; Malliakas, C. D.; Li, H.; Freeman, A. J.; Kenney, J. T.; Kanatzidis, M. G. CsSnI3: Semiconductor or Metal? High Electrical Conductivity and Strong Near-Infrared Photoluminescence from a Single Material. High Hole Mobility and Phase-Transitions. J. Am. Chem. Soc. 2012, 134, 8579 8587. 14