ABSTRACT Title of Dissertation: DYNAMIC ELECTRICAL RESPONSE AT THE NANOSCALE IN METAL HALIDE PEROVSKITES Richa Lahoti Doctor of Philosophy, 2022 Dissertation Directed by: Professor Peter Kofinas Professor Marina S. Leite Department of Chemical and Biomolecular Engineering Photovoltaic (PV) technology holds a promise that can change the energy dynamics globally. Next-generation materials, with the focus on two factors: high efficiency and economic feasibility, are being extensively explored to enable widespread implementation of solar cells. Metal halide perovskites (MHPs) provide the ideal combination of both these pressing factors, and also fall under thin film technology making their applications increasingly ubiquitous. In little over a decade, perovskites have reached a power conversion efficiencies of> 25%. Despite the prodigious advancements in efficiency, device and material instabilities have precluded their commercialization. While the origin of instabilities is multifaceted, instability under environmental factors (light, humidity, oxygen, temperature) is a central hub. Therefore, efforts are being directed towards understanding the behavior of photovoltaic properties under environmental conditions. Investigation at the material level is necessary to develop optimization strategies. My dissertation focuses on the electrical dynamics at length scales of grains and grain boundaries in MHP thin films. In the first part of my thesis, I present a comprehensive electrical analysis by probing surface voltage and photocurrent on Cs-containing dual-cation and Rb-containing quad- cation perovskite thin films. I measure surface voltage response using Kelvin probe force microscopy (KPFM) and map photocurrent via photoconductive atomic force microscopy (pc-AFM) under an inert environment. The Dark KPFM voltage maps indicate upward band bending at the grain boundaries for both chemical compositions. Using an illumination cycle (OFF-ON-OFF), I find a 55% larger post-illumination residual voltage drop in quad- cation perovskite. Photocurrent maps reveal highly photo-active grain boundaries in the quad-cation, while photo-inactivity is observed at grain boundaries in dual-cation perovskite. With the integrated knowledge about the upward band bending from KPFM and the electrical nature of the grain boundaries in the two chemical compositions, I infer defect passivation at the grain boundaries due to Rb+ cations and defect-assisted recombination at the grain boundaries of dual-cation perovskites. The highly conductive grain boundary network seen in quad-cation perovksite increases the overall photocurrent by 50%. The second part of my thesis demonstrates, for the first time, the ability of in situ humidity- dependent KPFM measurements to capture localized moisture-induced electrical dynamics in MHPs. I perform a controlled humidity cycle from 5 - 65% rH and back down from 65 - 5% rH. I observe an enhanced voltage response up to 45% relative humidity and an electrical failure at 65% rH. I capture a self-recovery value of over 90% post-humidity cycle and a recovery value of 99% 24 hours post-humidity cycle. Using XPS and PL before and after the humidity cycle, I confirm moisture-induced structural and chemical changes at the surface of the perovskite which are interconnected to the unstable electrical behavior seen during the humidity cycle. My comprehensive analytical approach on KPFM, and pc- AFM together with my in situ results showcase powerful methods for perovskite stability investigations. DYNAMIC ELECTRICAL RESPONSE AT THE NANOSCALE IN METAL HALIDE PEROVSKITES by Richa Lahoti Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2022 Advisory Committee: Professor Peter Kofinas, Chair/Advisor Dr. Marina S. Leite, Co-Advisor Professor Paul Albertus Professor Mario Dagenais Professor Dongxia Liu ? Copyright by Richa Lahoti 2022 Acknowledgments I would like to dedicate this dissertation to all my mentors, professors, family members, and friends, as my Ph.D. was only possible due to their guidance and immense support. First, I?d like to thank my advisor, Professor Peter Kofinas for believing in me and providing support through every stage of my Ph.D. journey. You have been very kind and accommodating throughout every challenging situation that came my way. You always offered help and went out of your way to make sure I was taken care of. It is remarkable how someone as busy as you always made time to meet with me. I admire your humbleness. Next, I?d like to thank my co-advisor, Professor Marina S. Leite. You gave me the opportunity to join your research group for which I am very grateful. You opened many collaborative doors for my research without which this thesis would not have been possible. Your example has made me want to be a better researcher and scientist. I appreciate all your efforts in making my Ph.D. a success. I also want to thank the members of my Ph.D. committee: Professor Mario Dagenais, Professor Paul Albertus, and Professor Dongxia Liu. Thank you for your time and insightful perspectives. I express my gratitude to Professor Jeffery B. Klauda. You were my first point of contact at the University of Maryland. You helped me through my application to the graduate program at UMD, welcomed me to the Chemical and Biomolecular Engineering Department, and remained a constant throughout my Ph.D. endeavors. Anytime I came ii across a difficult situation, you were the first person to whom I reached out. I cannot thank you enough for your support throughout my graduate school experience. To past and present Leite lab members: John M. Howard, Erica Lee, Tao Gong, Peifen Lyu, Margaret Duncan, Ece Deniz, and Meghna Srivastava. Special thank you to John M. Howard for the KPFM measurements for the first part of my work. I learned a lot from shadowing you. You are a great mentor and an outstanding researcher. Margaret Duncan, thank you for your guidance on the XPS measurements. To past and present Munday lab members: Raphaella Banholzer, Matt Corrado, Tristan Deppe, Calum Shelden, Professor Mariama Dias, and Professor Jeremy Munday. Thank you all for your time and endless help. You created a very welcoming and collaborative lab environment. To collaborators that provided samples for the research conducted in this dissertation: From Australia National University: Thomas P. White and Md Arafat Mahmud. From Georgia Institute of Technology: Yu-An, Sanggyun Kim, Juan-Pablo Correa-Baena. Thank you for all the insightful discussions. To the faculty and staff members who helped me through all the small unseen steps from research tasks and health insurance to everyday inquiries. From ChBE: Kathleen Ann Gardinier, Ruth H Yun, Karin Shortino, and Nina K Morris. From IREAP: Nancy Boone, Bryan Quinn, and Dottie Brosius. Special thanks to Nolan Alexander Ballew for going above and beyond during the pandemic to help me get remote access to our AFM computer, which prevented my research work from coming to a complete halt. From UCDavis: Gwendolyn Caramanica, Norma Andrade, Pia Flory, and Paula Lee. Special thank you to Dr. Bill Doering for your guidance in safety-associated research challenges. To members of Asylum Research, Oxford Research: Greg Koppel, Keith Jones, and iii Raphaella Banholzer. Thank you for your immense patience and intellectual discussions on troubleshooting system-related hurdles. I graciously acknowledge all the funding agencies and organizations that have provided financial support to make this dissertation possible: the NSF-EPMD (award 20-23974, 16- 10833), ARO (W911NF1810177), the James A. Clark School of Engineering, University of California Davis, and McAvoy Fellowship (2018-2019). I would also like to thank my friends who supported me on this rollercoaster-like journey. I could not have made it without them. In no particular order: Peifen Lyu, Mrinali Gupta, Arushi Sharma, Ece Deniz, Aarathi Vadapalli, Gaurav Iyer, Margaret Duncan, Sneha Gathani, Meghna Jha, Calum Shelden, Ishita Verma, Dhanya Vijayasarthy, Meghna Srivastava. Thank you for grabbing quick meals with me in between intense lab sessions, hour-long conversations about life, random hangouts, and trips. Thank you for listening to my rants and being my unconditional support system. As cheesy as it may sound, it comes from my heart - I don?t have words that would justly describe how much every one of you means to me. (I am sorry if I forgot someone, but I was on a rather tight time crunch when I wrote this.) To my partner, Ayush Bhat, whom I met in the middle of graduate school. Thank you for all the hot homemade meals that you would have ready for me when I would come home late, defeated, from a long day at the lab. Thank you for being my pillar. Thank you for taking care of me, for being your goofy self, and for making me laugh every day. I could not have survived these crazy years of graduate school without you by my side. I am very grateful to have found such a happy-go-lucky, hard-working, thoughtful, and loving person in my life. iv To my sister, Minnie Lahoti, thank you for trying to be your best self and trying your hardest to be there for me. Thank you for taking care of me when I was very sick during the second year of graduate school. Thank you for always pushing me to do better, for being the overachiever that you are, and for setting the bar for everything so high. I am very grateful that I had you to look up to. To my aunt, Angela Summers, thank you for being the strong person you are and for being such a successful example in my life. I am grateful for everything you have done for me. To my late uncle (Chachu), Sanjeev Lahoti, whom I lost during the third year of my Ph.D. You were like a second dad to me. I am heartbroken that you?re not here to see me graduate because I know of everyone in my life, you would be the proudest of me right now. You made the dream of studying abroad a reality for me. Thank you for always being there, for financially supporting me, helping me learn how to drive, helping me move across states, taking me grocery shopping, including me in the amazing life that you had built from scratch, and above all, thank you for showering me with your unconditional love. To my mother, Mukta Lahoti, and father, Vikas Lahoti, who have supported my education through endless hours of labor, I don?t know how to thank you. But I?ll try. I am extremely blessed to have you as my parents. Thank you for always believing in my abilities and supporting my dreams. Thank you for always putting me first, for not sparing a single opportunity to make my life better. Thank you for imbibing in me all your goodness and shaping me into who I am today. Richa Lahoti May 2022 College Park, MD v Table of Contents Acknowledgements ii Table of Contents vi List of Tables viii List of Figures ix Chapter 1: Nanoscale Imaging of Perovskite Optoelectronics 1 1.1 Perovskite Solar Cells: Promises and shortcomings . . . . . . . . . . . . . . 1 1.2 Atomic force microscopy (AFM) . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Kelvin probe force microscopy (KPFM) . . . . . . . . . . . . . . . . . . . . 4 1.4 Photoconductive atomic force microscopy (pc-AFM) . . . . . . . . . . . . . 5 Chapter 2: Research Objectives 8 Chapter 3: Rubidium-incorporated Perovskite Solar Cells Deliver Superior Electrical Response at the Nanoscale 11 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.3.1 Macroscopic solar cell performance . . . . . . . . . . . . . . . . . . 19 3.3.2 Nanoscale characterization . . . . . . . . . . . . . . . . . . . . . . . 21 3.3.3 Band bending insights through dark KPFM . . . . . . . . . . . . . 22 3.3.4 Imaging ion motion through illuminated KPFM . . . . . . . . . . . 27 3.3.5 Mapping local photocurrent through pc-AFM . . . . . . . . . . . . 34 3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Chapter 4: In Situ Humidity - Dependent KPFM on Metal Halide Perovskites 51 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.2 Experimental set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.2.1 Humidity enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.2.2 Humidity control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.2.3 Bottom Illumination . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.3 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 vi 4.4.1 Surface voltage under illumination cycle and inert environment . . . 65 4.4.2 Surface chemistry and structure . . . . . . . . . . . . . . . . . . . . 68 4.4.3 Nanoscale surface voltage response . . . . . . . . . . . . . . . . . . 75 4.4.4 Humidity- and illumination-dependent surface voltage behavior . . 89 4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Chapter 5: Conclusion and Outlook 96 Chapter A: Products of This Research 101 A.1 Awards and Honors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 A.2 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 A.3 Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Bibliography 103 vii List of Tables 3.1 Table 3.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 viii List of Figures 1.1 Traditional structure of perovskite . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 pc-AFM circuit diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.1 Macroscopic J-V Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 3.3 Dark KPFM on dual-cation perovskite . . . . . . . . . . . . . . . . . . . . 24 3.4 Dark KPFM on quad-cation perovskite . . . . . . . . . . . . . . . . . . . . 25 3.5 OFF-ON-OFF illumination cycle KPFM . . . . . . . . . . . . . . . . . . . 29 3.6 Full series of OFF-ON-OFF illumination cycle on quad-cation perovskite . 30 3.7 Post-illumination behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.8 Average surface voltage versus change in time . . . . . . . . . . . . . . . . 32 3.9 Surface voltage distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.10 pc-AFM on Dual-cation perovskite . . . . . . . . . . . . . . . . . . . . . . 36 3.11 pc-AFM on quad-cation perovskite . . . . . . . . . . . . . . . . . . . . . . 38 3.12 Photoinactivity statistics in dual and quad-cation perovskite . . . . . . . . 39 3.13 Grain statistics on topography maps obtained simultaneously with surface voltage maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3.14 Grain statistics on topography maps obtained simultaneously with photocurrent maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.15 Photocurrent maps at applied biases for dual-cation perovskite . . . . . . . 42 3.16 Photocurrent distribution at applied biases for dual-cation perovskite . . . 43 3.17 Photocurrent maps at applied biases for quad-cation perovskite . . . . . . 44 3.18 Photocurrent distribution at applied biases for quad-cation perovskite . . . 45 3.19 Average photocurrent versus applied bias . . . . . . . . . . . . . . . . . . . 46 3.20 Charge carrier mobility at grain boundaries . . . . . . . . . . . . . . . . . 48 4.1 Humidity enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.2 Experimental set up for in situ humidity dependent KPFM . . . . . . . . . 57 4.3 Humidity control functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.4 Real-time humidity for CdTe topography maps . . . . . . . . . . . . . . . . 59 4.5 Topography scans on p-doped CdTe under humidity . . . . . . . . . . . . . 60 4.6 Bottom illumination optics customization . . . . . . . . . . . . . . . . . . . 61 4.7 Customized bottom illumination setup . . . . . . . . . . . . . . . . . . . . 62 4.8 Broadband LED Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.9 KPFM light OFF-ON-OFF cycle . . . . . . . . . . . . . . . . . . . . . . . 66 ix 4.10 Stability test of surface voltage response under dark-and illuminated- KPFM on Cs0.33FA0.67PbI3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 4.11 Real-time humidity and temperature profiles during the in situ humidity- dependent KPFM experiment . . . . . . . . . . . . . . . . . . . . . . . . . 68 4.12 Core spectra analysis of I3d, Pd4f and Cs3d . . . . . . . . . . . . . . . . . 71 4.13 Core spectra analysis of C1s, N1s and O1s . . . . . . . . . . . . . . . . . . 73 4.14 Photoluminescence measurements before and after humidity cycle . . . . . 74 4.15 Topography scans on Cs0.33FA0.67PbI3 from 5% to 25% rH transition of the UP cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.16 Dynamic surface voltage response on Cs0.33FA0.67PbI3 at 5% and 25% rH of the UP cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.17 Topography scans on Cs0.33FA0.67PbI3 at 45% and 65% rH transition of the UP cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.18 Dynamic surface voltage response on Cs0.33FA0.67PbI3 at 45% and 65% rH of the UP cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 4.19 Topography scans on Cs0.33FA0.67PbI3 at 45% and 25% rH transition of the DOWN cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.20 Dynamic surface voltage response on Cs0.33FA0.67PbI3 at 45% and 25% rH of the DOWN cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.21 Topography scans on Cs0.33FA0.67PbI3 at 5% rH at the end of the humidity cycle and at ? 12 hours post humidity cycle capture recovery at 5% rH . . 83 4.22 Surface voltage response on Cs0.33FA0.67PbI3 at 5% rH at the end of the humidity cycle and at ? 12 hours post humidity cycle capture recovery at 5% rH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.23 Topography scans on Cs0.33FA0.67PbI3 at ? 24 hours post humidity cycle recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.24 Surface voltage maps on Cs0.33FA0.67PbI3 at ? 24 hours post humidity cycle recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 4.25 Drift analysis on topography and surface voltage maps . . . . . . . . . . . 87 4.26 Dynamic humidity-dependent surface voltage response versus time . . . . . 88 4.27 Electrical response recovery post humidity cycle on Cs0.33FA0.67PbI3 . . . . 89 4.28 Humidity-dependent illumination and post-illumination electrical behavior of Cs0.33FA0.67PbI3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 5.1 Experimental setup for in situ temperature-dependent KPFM and pc-AFM 100 x Chapter 1: Nanoscale Imaging of Perovskite Optoelectronics 1.1 Perovskite Solar Cells: Promises and shortcomings It has been a little over a decade since perovskite thin films were first introduced to the world of photovoltaics. When they were first applied as photoabsorbers, their power conversion efficiency (PCE) was 3.8% (2009). [1] However, today, perovskite solar cells have reached a PCE value of 25.7%, [2] making them one of the fastest growing advancements the photovoltaic industry has seen. The attraction towards perovskite absorbers increased due to their ease of fabrication [3] and tunable photovoltaic properties [4?6]. Perovskites have a chemical structure of the form ABX3 (Figure 1.1), where A is an organic or inorganic cation (usually methyammonium, formadinium, Cs+, Rb+ or a combination of these) in the center of the structure, B is a metal cation (either Pb2+ or Sn2+) in a 6-fold coordination surrounded by X, which is a halide (Cl-, B-, I-, or a combination of these) forming an octahedron. [7,8] Further, with perovksite absorbers, solar cells can be made with thin film technology, allowing flexibility of the devices. [9, 10] This has allowed endless possibilities of solar power application in a multitude of areas ranging from wearable technology to terrestrial applications. [11] To add to these promises that perovskites have brought to the solar cell industry, they are made from abundant and inexpensive materials, making them economically viable for large-scale production. 1 A is an organic cation at the center B is a metal cation in a 6-fold coordination X is a halide forming an octahedron Figure 1.1: Traditional structure of perovskite materials. Despite the above mentioned advantageous qualities, perovskite solar cells are far from commercialization. [12] Perovskites are infamous for their instabilities. [13?15] The high PCE values that have been achieved today are for unstable devices. There are multiple areas in which improvement and optimization of PSCs is necessary to attain the stability levels required to compete with the current life of silicon-based photovoltaic technology (> 25 years). [16] The degradation mechanisms and dynamic behavior under environmental factors are an ongoing area of study and optimization for PSCs. [17?20] The high sensitivity of PSCs to extrinsic factors such as oxygen [21?23], humidity [24?26], bias [27?29], illumination [30?32], heat [33?35] remains a pressing concern in their adoption on a large scale. All of the advances currently necessary for the commercialization of PSCs are at the material level and therefore require nanoscale probing of the perovskite thin films. 1.2 Atomic force microscopy (AFM) Atomic force microscopy is a widely used tool to probe various surface properties of materials at the nanoscale. For photovoltaic materials, the technique is useful for 2 simultaneously capturing morphological and electrical properties. AFM has been used to map chemical heterogeneity [36], ferroelectricity [37, 38], ferroelasticity [39, 40], surface passivation [41], and stability [42, 43] in perovskite thin films opening many pathways for material optimization. Besides, photovoltaics, its applications have been seen in numerous other fields such as microbiology [44], polymer science [45], photovoltaics [46], medicine [47], and forensic science [48]. The working principle of an atomic force microscope is based on the tip and surface interaction mechanism. The microscope consists of a sophisticated assembly of parts. The cantilever and nano-sized tip form a unit in which the cantilever acts as a spring. A diode laser is positioned on the cantilever. The tip is driven by applying an AC voltage at its resonance frequency. The tip interaction on the sample?s surface cause the cantilever to bend. This bending of the cantilever causes the laser to deflect. The deflected laser is caught by a quadrant diode detector. This information is then sent to the x, y, z piezo stage and a spatial map is produced. AFM provides < 50 nm spatial resolution. There are two main operating modes of the AFM: Dynamic mode and contact mode. In Dynamic mode, also known as tapping mode, the tip oscillates with a specific amplitude making the contact between the tip and sample surface intermittent. This prevents both the tip and sample from getting damaged. The tip experiences Van der Waals (attractive) and Coulombic (repulsive) forces, which govern the movement of the tip. The feedback loop maintains a constant force by keeping the amplitude constant. In contact mode, the tip is always in contact with the sample surface. Having the tip pressed on the sample does cause tip attenuation and can also damage the sample if the tip set point voltage is too high. The feedback loop keeps the force constant by keeping the deflection constant. 3 1.3 Kelvin probe force microscopy (KPFM) KPFM is a dynamic mode AFM technique that measures the contact potential difference (CPD) between the tip and the sample. The tip used in this measurement is made of a conductive metal. The tip is kept at a constant height away from the sample?s surface. The tip and the sample are electrically connected. This set-up of the tip and sample forms a parallel capacitor, and the local contact potential is quantified by detecting a capacitive force between them. The contact potential difference is calculated as follows: VCPD ? (?s ? ?t)/e (1.1) where e is the positive elementary charge, ?s and ?t are the work functions of the sample and tip, respectively. [49] KPFM can provide a variety of electrical information about the semiconductor such as work function, band bending, dopant density, surface charge, density of surface states. [49] KPFM measurements can be performed in two types of modes: single pass and dual pass mode. In the single-pass set up, the probe goes over the surface only once. The probe is kept at a constant height and never comes in contact with the sample surface. In the dual-pass set up, the probe goes over the sample?s surface twice. During the first pass, the probe maps the topography in an amplitude modulated mode in which the probe oscillates at its resonance frequency. The probe is then lifted up at a constant height above the sample. The second pass runs similar to the single-pass mode, however, uses nap mode imaging to acquire the surface voltage data. 4 KPFM has helped explore dynamics in ion migration [50, 51], open-circuit voltage (Voc) [52], charge accumulation [53], charge carrier generation and transport [54, 55] in PSCs. These electrical nanoscale maps provide insight into the heterogeneity of photovoltaic properties in the perovskite material. KPFM measurements are done under dark and light conditions. In the absence of light, KPFM helps resolve intrinsic charge carriers. When illuminated, light-induced charge carriers are collected and the difference between the dark and light and dark scans gives us the surface photovoltage (SPV) response of the sample. The relationship between SPV and work function of the sample is given as: SPV ? (?light dark light darks ? ?s )/e = VCPD ? VCPD (1.2) where, e is the elementary positive charge, ?s is the sample work function, V lightCPD is the difference in contact potential between the tip and the sample under illuminated conditions, and V darkCPD is the difference in contact potential between the tip and the sample under dark conditions. [49] 1.4 Photoconductive atomic force microscopy (pc-AFM) While macroscopic J-V curves provide information on device performance parameters and hysteresis behaviors, nanoscale photocurrent mapping allows exploration of inhomogeneities, charge carrier density, photoactivity of grain and grain boundaries, defect sites, and defect-assisted charge carrier recombination. [12,28,56,57] pc-AFM is a contact mode AFM technique that spatially maps the photocurrent 5 response of the sample?s surface. pc-AFM is performed with a conductive probe, an electrical circuit, and an illumination laser. The circuit is completed when the probe comes in contact with the samples surface allowing charges to flow. The circuit diagram of the pc-AFM is shown in Figure 1.2. Local I-V curves can also be extracted under this mode, at different morphological features of the surface, to capture bias-dependent current behavior. In addition, bias-dependent photocurrent maps can reveal information about the behavior of the donor and acceptor molecules in the material. When a bias is applied above the Voc value of the photovoltaic material, the holes move towards the probe and the electrons move towards the indium tin oxide (ITO) layer. The reverse behavior (i.e. electrons move towards the probe and holes towards ITO) is seen when the bias value applied is below the Voc value. Rgain ? ? #! ?" Probe ORCA Vout Cantilever Holder Rsample Vbias Figure 1.2: Circuit diagram for pc-AFM measurements.(Adapted from Asylum Research - Oxford Instruments) From Figure 1.1, applying Kirchhoff?s junction rule, the sum of current flowing into a node 6 is zero. This gives the following Equation: i1 + i2 + i3 = 0 (1.3) Now, we know i3 = 0 since the current flows only from the sample to the cantilever holder. This reduces Equation 1.3 to i1 = ?i2 (1.4) Applying Ohm?s law, Equation 1.4 can be written as VBias ? VOut= (1.5) RSample RGain Here, VBias is the voltage applied to the sample (this value is limited to ? 10 V), RSample is the resistance value of the sample, VOut is the output voltage from the cantilever impedance amplifier, and RGain is gain resistor Rearranging Equation 1.5, we can solve for the output voltage ?VBias ?RGainVOut = (1.6) RSample The magnitude of RGain is megaohms (M?). The magnitude of RSample is sample dependent, however, the magnitude of resistance for photovoltaic materials is in kiloohms (k?). The sensitivity of the current reading can be estimated by the reciprocal of the RGain value. 7 Chapter 2: Research Objectives In this dissertation, I accomplish the following research objectives by successfully developing and implementing nanoscale imaging techniques for probing electrical dynamics (surface voltage, surface photovoltage and photocurrent) at length scales of grains and grain boundaries in promising perovskite materials for photovoltaic applications. (i) Rubidium-incorporated Perovskite Solar Cells Deliver Superior Electrical Response at the Nanoscale. With the help of Kelvin probe force microscopy (KPFM) and photoconductive atomic force microscopy (pc-AFM), I spatially map the surface voltage and photocurrent, respectively, in dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and quadruple-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. Dark surface voltage maps, resolved at grains and grain boundaries, revealed higher surface voltage at grain boundaries (GBs) than the grains, establishing upward band bending at GBs for both chemical compositions. KPFM measurements under an illumination cycle of OFF- ON-OFF on quad-cation perovskite reveal a post-illumination voltage drop that is 55% greater than that seen in dual-cation perovskite, indicating reduced ion migration in the former. Photocurrent imaging on both samples, when correlated with the band-bending information obtained from the surface maps, revealed defect-assisted recombination at GBs in dual-cation perovskite and defect-passivation at GBs in quad-cation perovskite by Rb+. 8 50% higher photocurrent was seen in the quad-cation perovskite largely due to the highly photoactive network of GBs in the quad-cation perovskite. Comprehensive analysis of electrical behavior captured in dual- and quad-cation perovksite highlights the superior electrical performance of the Rb-incorporated quad-cation perovksite. (ii) In situ Humidity - Dependent KPFM on Metal Halide Perovskites. Using a customized atomic force microscopy (AFM) chamber, a custom-built humidity control set up and customized optics for movable bottom illumination, I captured for the first time, in-situ humidity-dependent KPFM response in Cs0.33FA0.67PbI3 perovksite thin film. Capturing surface voltage maps while cycling the relative humidity (rH) up from 5% ? 25% ? 45% ? 65% (UP cycle) and cycling it back down from 65% ? 45% ? 25% ? 5% rH (DOWN cycle) reveal real-time moisture-dependent dynamics in electrical behavior and moisture-stability of the perovskite thin film. In addition to cycling humidity, an illumination cycle (OFF-ON-OFF) at each level of humidity. During the UP cycle, we observe that the voltage response increases proportionally with relative humidity up to 45% rH. At 65% rH, the voltage response drops by 67%. During the DOWN cycle, as the humidity value is brought back down from 65% to 45%, the voltage response increases and starts recovering. Comparing the voltage scans at the beginning of the experiment (at 5% rH) and the voltage scans at the end of the humidity cycle (at 5% rH), we observe 94.34% recovery in the sample. Topography scans reveal no material degradation of the perovskite material. Through XPS measurements before and after the humidity cycle, we uncover prominent chemical changes in the surface chemistry, ion migration to the surface and structural differences in the perovskite post-humidity cycling. 50% Loss in PL signal post- humidity cycle indicates higher nonradiative recombination in the perovskite. Combining 9 the results, high self-recovery, moisture-induced electrical and structural instabilities are seen in Cs0.33FA0.67PbI3 thin film. 10 Chapter 3: Rubidium-incorporated Perovskite Solar Cells Deliver Superior Electrical Response at the Nanoscale Rb+ cations at the A-site of metal halide perovskite can considerably improve device performance. While the macroscopic properties are well established, a comprehensive understanding of electrical characteristics at the nanoscale remains incomplete. Here, we quantify the electrical performance of dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and quadruple-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) photovoltaics {FA = CH(NH2)2, MA = CH3NH3} at the nanoscale, by combining scanning probe microscopy methods that allow independent local measurements of voltage and photocurrent. The voltage maps indicate upward band bending at the grain boundaries (GBs) for both chemical compositions. We find a 34% reduction in voltage heterogeneity and a 55% larger post illumination residual voltage drop in quad-cation perovskite, combined with a 50% increase in the photocurrent response. Further, we uncover higher photocurrent at the quad-cation GBs compared to the grain interiors, a direct consequence of defect-passivation at the former due to Rb+ cations. Surprisingly, for dual-cation perovskites, the majority of GBs display photocurrent values comparable to dark current, indicating defect-assisted recombination at GBs. Our results demonstrate the benefits of incorporating Rb+ cations into perovskite, including improved photovoltage and photocurrent homogenization at the 11 ? Ec E EDF nanoscale. Our analytical approach can be impEvlemented as a general method to identify Grain 1 GB Grain 2 GB with defects the presence and passivation of defects at GBs for any crystalline optoelectric material. ? ?? Ec Ecc E E This chapter is adaptedEFE from R. L D ahoti, J.M. HEFEoward et Dal. (TBD) v 150 vvSurface voltage150 mV Photocurrent100100 pA 100 250 50 50 500Grain 15500 GB G20r0ain 2 Grraiin 1 4000 0 GB Grraiin 20 300 GB with de--f5e0 cts 150 GB wiitthh dReb-f50e+-5-c0ptsassivated defe20c0 ts -100 100 -100-100 100 --150 0 -150-150 ?? AFM probe ? Ecc Ec ? E ED E EcFF FEF E Evv 15 Sv urface voltage1151050 E0 Surface voltage mV Photocurrent 50 v 1m00V100 Photocurrent100100 ppAA 101000 i 225500 50 50 500 Grain 15500 GB Graii 50050 n 2 GGrairna 1in 1 GB ?GBGrain 2 50005500 0 22000 400 GB with R0b0+0-0passivated defects G40r0a0 in 2 0 300 GB wiitth dReb--f5-e+-05-0c0ptasssivated defe1c15t50s0 EcGB with Rb-5-05+-05-050p0 assivated defe 3c00220000t 0 0s -1-01000 11000 E -1-010-100 EF 0-0100 D 1100000 --1--5105500 -1E - 510-51-0515500 ? v AAFFM p prroobbee Ec Grain? ?1 GB Grain 2 E EccF EFF GB with defects Ev E150 Svv urface voltage150 mV Photocurrent100100 pA 100 ? 250 Grraainin 11 50050GB ? GGrarainin 2 2 5000Grain 15500 GB G2r0a0in 2 GBGB wiitthh Rbb0++-0-ppaassssivivaatteedd d deefefecctsts 4000E 0cGB with Rb-+-5-00paEssivated defec1t5s0 Ec -50-0 300050 2000E -100 D 100 E -10-0100F F 1000 --15E 5 00 -1505E Surface voltage -150v 115500 v mV PhotocurrentAFM pr1o00b1e00 pA 100 250 50 50 500Grain 15500 GB G20r0ain 2 Grain ?1 GB G4000 0 rain 20 E 300 GB with de--f5e0 cts 150 cEF GB with Rb-50+-5-0passivated defe20c0 ts -100 100 E -100-100v 100 --150 0 -150-150 ? Grain 1 AFM proGbeB Grain 2 GB with Rb+- passivated defects Ec ? E EcF EF E Ev 150 Sv urface voltage150 mV Photocurrent100100 pA 100 250 Grain 1 50050GB Grain 2 5000Grain 15500 GB G2r0a0in 2 GB with Rb0+-0passivated defects 4000 0 GB with Rb-+-5-00passivated defec1t5s0 3000 -50-050 2000 -100 100 -10-0100 1000 --15500 -15-05150 AFM probe ? Ec EF Ev Grain 1 GB Grain 2 GB with Rb+- passivated defects 12 3.1 Introduction With the traditional chemical structure of ABX3 (A = monovalent organic/inorganic cation, B = divalent metal cation, X = halide ion), halide perovskites are now multi- cation, multi-halide materials. In the realm of photovoltaics, hybrid organic-inorganic perovskites are a promising option for next-generation energy harvesting devices, with power conversion efficiency comparable to state-of-the-art Si solar cells. [58] Yet, the bottleneck in the commercialization of perovskite solar cells (PSCs) is instability. [59, 60] With the incorporation of multiple cations at the A-site, such as methylammonium (MA), formamidinium (FA), Cesium (Cs), and Rubidium (Rb), there has been significant improvement in device stability while minimally sacrificing their power conversion efficiency. [61?69] In 2020, the highest efficiency of a quadruple-cation PSC was recorded to be 22.81%. [70] A-site cations, such as Rb+, assist in the tuning of open-circuit voltage (V +oc). [64, 65, 71?73] In particular, the addition of Rb at the perovskite A-site has been shown to increase charge carrier mobility, material stability, and overall reproducibility by reducing current-voltage hysteresis. [64,72,74?77] Further, incorporation of Rb+ promotes an increase in efficiency and external quantum efficiency (EQE) of PSCs. [64,65,72,74,78] Additionally, perovskites with Rb+ have exhibited greater stability under moisture-rich environments [65, 79] and improved thermal stability without forgoing performance. [80] Another cation that has supported improvements in PSCs is Cesium (Cs+). Rb+ and Cs+ both belong to Group-1A of the periodic table, more commonly known as Alkali Metals. The two therefore share common physical and chemical properties. However, with the precedence of Rb+ over Cs+ in this group, a significant difference in 13 size, electronegativity, electrical conductivity, electron affinity, etc. The inclusion of Rb+ and Cs+, individually, to the perovskite layer, has led to advances in device performance, stability, and reproducibility. [72, 74, 81] Nevertheless, the extent of these advances is appreciably different for both types of devices. Together, their addition has led to a drastic reduction in phase segregation in halide perovskites. [82] The incorporation of Cs+ reduces trap density and second-order recombination rate of free mobile charges in the bulk of the perovskite layer. [74, 83] The mechanisms for the above-mentioned progress made in PSCs due to the addition of Rb+ and Cs+ are reasonably understood and explained in literature. [82] Optimizing photovoltaic performance requires a fundamental understanding of the electrical and optical behavior of the mesoscale constructs that form the light-absorbing perovskite layer, which requires spatially resolving the information at the nanoscale. [61] Kelvin probe force microscopy (KPFM) [18, 42, 75, 84], and photoconductive atomic force microscopy (pc-AFM) [12, 64, 85] are quantitative characterization methods suitable for spatially resolving the nanoscopic photo-response of PSCs. [61] These state-of-the-art characterization techniques have been successfully implemented to study various properties of halide perovskites and PSCs, including work function; [86] local surface potential and the real-time transient behavior of the Voc; [18] surface potential distribution; [87,88] local photocurrent using pc-AFM [12]; and to deconvolute the direction of charge conduction at grain boundaries through three-dimensional photocurrent maps using tomographic (T-) AFM. [85] Here, we quantify, with nanoscale spatial resolution, the key photovoltaic attributes of a dual-cation perovskite (Cs0.17FA0.83Pb(I0.83Br0.17)3) and a quad-cation perovskite 14 (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3), revealing the enhanced electrical performance lent by Rb+ at length scales relevant to the film morphology. These two chemical compositions were chosen for comparison for two main reasons: (i) both have tolerance factor values (? 0.96) that fall within the ideal range for structural stability, and (ii) the addition of cations with a small ionic radius to the A-site has been shown to provide valuable properties to perovskites. Dual-cation perovskite contains Cs+ at the A-site while quad-cation perovskite has both Cs+ and Rb+ as small cations, respectively, allowing the deconvolution of the effect of Rb+ on their electrical properties. By performing a comparative analysis between dual and quad-cation PSCs, we discover a 34% reduction in voltage heterogeneity and 55% larger drop in post-illumination relaxation of voltage, in the latter, by implementing illuminated-KPFM (1-sun illumination). This improved voltage response results from the increased uniformity in the halide distribution and the reduced ion migration due to the incorporation of Rb+ cations. We establish upward band bending and, therefore, the formation of an electron barrier at grain boundaries (GBs) for both samples. For quad-cation perovskite, photocurrent maps reveal that the GBs are more effective current collectors than the grain cores, due to defect-passivation by Rb+ cations. The resulting conductive network increases the overall photocurrent response by ? 50% in quad- cation perovskite. Strikingly, for dual-cation perovskite, the majority of the GBs present a photocurrent response similar to dark current values, revealing photoinactivity (under 1-sun illumination conditions). Here, the free holes are consumed through defect-assisted recombination at GBs in a dual-cation perovskite. KPFM and pc-AFM have been widely applied as individual techniques to extract information on band bending, photovoltage, and photocurrent response in optoelectrical materials. Here, we show that combining these 15 state-of-the-art techniques provides a tool for identifying defects and their passivation at GBs for optoelectronic materials. Our insights into the role of Rb+ in perovskites at the nanoscale will allow further optimization of the composition and process for applications in photovoltaics. 3.2 Experimental Methods Perovskite device fabrication: For device fabrication, FTO/glass substrates were sequentially cleaned with deionized (DI) water, detergent, acetone, isopropanol, and ethyl alcohol in an ultrasonic bath for a duration of 10 minutes for each step. The substrates were then treated in a UV-ozone-clean machine for 15 minutes. Then, a 70-nm thick TiO2 compact layer was deposited on the cleaned substrates by spin coating the titanium (IV) isopropoxide solution in isopropanol with a spin speed of 5000 rpm for 15 s with a ramp rate of 5000 rpm/s. The compact TiO2 film was then annealed at 500 ?C for 30 minutes in a furnace. Subsequently, a 100 nm thick mesoporous TiO2 layer was deposited on the substrate by spin coating the 30 NR-D titania paste solution in ethanol (1:12 weight ratio) at a spin speed of 5000 rpm for 15 s with a ramp rate of 5000 rpm/s. The samples were again annealed at 500 ?C for 30 minutes in the furnace. Once the substrates were cooled down to room temperature, a thin PMMA:PCBM passivation layer was deposited on the substrates at a spin rate of 4000 rpm for 15 s with a ramp rate of 4000 rpm/s. For dual-cation FA0.83Cs0.17Pb(I0.83Br0.17)3 [FA: formamidinium] perovskite composition, the precursor solution was prepared by mixing FAI (1.1 moles), PbI2 (1.1 moles), CsBr (0.2 moles) and PbBr2 (0.2 moles) in 1 ml of solvent mixture 16 of N,N -dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) with a volume ratio of the solvents being 4:1. For the composition of the quadruple cation Cs0.07Rb0.03FA0.765MA0.135PbI2.55Br0.45 [MA: methylammonium] perovskite, the precursor solution contained 1.1 mole FAI, 0.2 mole MABr, 0.2 mole PbBr2, 1.2 mole PbI2, 0.091 mole CsI and 0.039 mole RbI in 1 ml solvent mixture of DMF and DMSO with the volume ratio of these solvents being similar as a dual cation perovskite precursor. For the deposition of dual or quad-cation perovskite, a two-step spin-coating program (1st step: 1000 rpm, 10 s, 100 rpm/s, 2nd step: 4000 rpm, 25 s, 1000 rpm/s) was used to coat the perovskite film on the substrates. During the second step, ? 150 ?L of chlorobenzene was dropped into the center of the spinning substrate 8 s before the end of the spinning program. Immediately after spin-coating, the substrates were placed on a hot plate and annealed for 30 minutes at 100 ?C. For the hole transport layer (HTL), a solution of 2,2 ?, 7,7 ?-Tetrakis [N, N-di (4- methoxyphenyl) amino] -9,9 ?-spirobifluorene (Spiro-OMeTAD) was prepared by dissolving 73.5 mg of Spiro-OMeTAD in 1 ml of chlorobenzene and the solution was doped with 17.5 ?L of bis(trifluoromethane)sulfonimide lithium salt (Li-TFSI) (520 mg / ml in acetonitrile) and 28.5 ?L of 4-tert-butylpyridine. The spinning recipe for Spiro-MeOTAD was 3000 rpm for 30 seconds with a ramp rate of 3000 rpm/s. Finally, an 80-nm gold layer was deposited on the Spiro-OMeTAD coated substrates by thermal evaporation. The active area of the cell (0.165 cm2) was defined using a shadow mask during the evaporation process. To prepare the samples for KPFM or c-AFM, the whole process was repeated except for the final gold deposition step. J-V measurements: The J-V measurements were conducted using a solar simulator 17 system (Weblab Inc.) under 1-sun illumination (AM 1.5G, 1000 W/m2, 25 ?C). A certified Fraunhofer CalLab reference cell was used to calibrate the light intensity prior to measurements J-V. No preconditioning protocol was applied prior to cell measurement. Cells were tested in a custom built measurement jig under a flow of N2 gas. For the forward and reverse scans, the voltage range was maintained at -0.1 V?1.25 V and 1.25 V?-0.1 V, respectively. Unless otherwise stated, all the measurements were done at a scan rate of 50 mV/s with a voltage step of 0.01 V. A FEI Verios SEM machine was used to investigate the surface morphology. For cross-sectional SEM images, a Helios Nanolab 600 FIB system was utilized. Prior to the cross-sectional imaging, a ? 2 ?m protection layer of Pt was deposited on the substrate. Kelvin probe force microscopy: KPFM was implemented with an Asylum MFP-3D AFM using standard lift-mode AC, also known as dual-pass amplitude modulated KPFM. Topographic information was collected during the initial pass, after which the contact potential difference was recorded during the second pass. For all measurements, the second pass lift height (?H) was 20 nm, with the tip voltage and drive amplitude set to 3.0 V. The same Pt / Ir coated Si probe (f = 320 Hz, k = 42 N / m, radius of curvature < 25 nm) was used to measure both compositions. Both samples were grounded by connecting the FTO sample to the AFM scan plate using a 2.5 k? resistor. The samples were illuminated from the bottom using a broad light-emitting diode (P = 100 mW/cm2) with a peak centered ? 550 nm and measured under ambient conditions (23 ?C, 35% rH). The dark KPFM images displayed in Figure 3.3 were flattened and their mean values zeroed. For the KPFM maps shown in Figure 3.4, the mean of the initial dark image was subtracted from the entire series; no other transformations were applied. 18 Photoconductive-atomic force microscopy: pc-AFM was accomplished using an Asylum MFP-3D AFM using a dual gain ORCA cantilever holder. The layer of the sample being imaged was perovskite in all cases. One edge of the perovskite layer was scratched off, exposing the FTO layer. The circuit setup: A clip on the sample mount was placed on the FTO layer of the sample. A jumper wire connected the clip to the magnet on the sample mount. A bias wire connected to the cantilever holder was dropped down and connected to the magnet on the sample mount, completing the circuit. For each measurement on both compositions, the sample mount was sealed with a membrane using a metal disk with a steady flow of nitrogen, imitating a set-up of a fluid cell with the in and out fluid being nitrogen gas. Thus, making the experimental setup, in its entirety, inert (< 7% relative humidity). The samples were illuminated from the bottom using a broad light emitting diode (P = 100 mW/cm2) with a centered peak ? 550 nm and measured at ambient temperature (23 ?C). For every measurement, identical Pt/Ir coated Si probe (f = 13 Hz, k = 0.2 N/m) was employed. The noise level for pc-AFM was recorded as 45 pA for a 1 nA/V output. 3.3 Results 3.3.1 Macroscopic solar cell performance We present a quantitative, spatially resolved analysis of the electrical behavior mapped at grains and their boundaries, in dual- and quad-cation perovskite, through illuminated-KPFM and pc-AFM under 1-sun illumination and low humidity environment (< 7%), emulating device operating conditions. The details regarding device fabrication 19 and processing are provided in the Experimental methods section. The behavior of the macroscopic device J-V, as presented in Figure 3.1, shows that we use state-of-the-art photovoltaic devices to validate the relevance of our nanoscopic measurements. Overall, the presence of both Rb+ and MA+ suppresses J-V hysteresis by ? 90% in the quad-cation PSC (see Figure 3.1 a). Prior research has revealed that the addition of both Rb+ and Cs+ spatially homogenizes the halide concentrations. [69] We compute the hysteresis index (HI ) for the dual- and quad-cation devices as 0.279 and 0.029, respectively, to determine the gravity of the hysteresis seen in each device. Here, HI is defined as HI = (PCERS - PCEFS)/PCERS; where PCERS and PCEFS are the power conversion efficiencies for the reverse scan (RS) and forward scan (FS), respectively, measured at a scan rate of 50 mV/s. Despite the stark difference in hysteresis, the two devices display comparable maximum power conversion efficiency (?) based on their reverse J-V scans (20.01% and 20.81% for the dual- and quad-cation, respectively). The decrease in dual-cation ? from 20.01% to 14.43% for reverse and forward bias, respectively (Figure 3.1 a), is largely due to the decrease in fill factor (FF ) from 76.67% to 58.98%, which is likely induced by ionic migration within the perovskite absorber. The open-circuit voltage (VOC) shows a similar drop, lowering from 1.12 V to 1.05 V. By contrast, the quad-cation PSC (Figure 3.1 b) exhibits a < 3% change in FF and negligible reduction in VOC . Cross-sectional scanning electron microscopy (SEM) images (inset Figures 3.1 a and 3.1 b) reveal nearly identical layer structures. Thus, unraveling how and why virtually equal morphologies deliver distinct electrical behavior requires interrogation of the perovskites at the mesoscale. 20 Dual-cation perovskite Quad-cation perovskite Gold Gold Spiro-OMeTAD Spiro-OMeTAD Dual-Cation Perovskite Quad-Cation Perovskite PMMA:PCBM PMMA:PCBM Mesoporous TiO2 Mesoporous TiO2 Compact TiO2 300 nm Compact TiO2 300 nm FTO FTO 500 nm 500 nm Figure 3.1: Adding Rubidium to halide perovskite solar cells reduces J-V hysteresis. J- V behavior for (a - red) dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and (b - blue) quad- cation perovskites (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) from reverse (VOC to Jsc) and forward (Jsc to VOC) and scans at 50 mV/s. Insets: cross-sectional scanning electron microscopy (SEM) images of all active layers. The figures of merit for the dual-cation perovskite are: reverse direction (triangles) VOC = 1.12 V, JSC = 23.30 mA cm?2, FF = 76.67%, ? = 20.01%; forward direction (circles) V = 1.05 V, J = 23.30 mA cm?2OC SC , FF = 58.98%, ? = 14.43%; Hysteresis Index (HI ) = 0.279. The figures of merit for the quad-cation perovskite are: reverse direction (triangles) VOC = 1.16 V, JSC = 23.60 mA cm?2, FF = 76.02%, ? = 20.81%; forward direction (circles) VOC = 1.15 V, JSC = 23.50 mA cm?2, FF = 74.75%, ? = 20.20%; Hysteresis Index = 0.029. 3.3.2 Nanoscale characterization To resolve both photovoltage and photocurrent at the nanoscale, we applied combined illuminated-KPFM and pc-AFM (as shown in Figure 3.2 a) to half-processed devices with the following layers: titanium oxide (compact + mesoporous) as the electron transport layer, [6,6] -phenyl-C61-butyric acid methyl ester and polymethyl methacrylate mixture (PMMA: PCBM) as the passivation layer and halide perovskite as the light-harvesting layer; all deposited on a fluorine-doped tin oxide-coated glass substrate. Figures 3.2 b 21 and c display representative three-dimensional (3D) images for the perovskite morphology overlaid with both photovoltage and photocurrent obtained using KPFM and pc-AFM techniques, respectively. The perovskite layers for both the dual- and quad-cation samples have an energy band gap of 1.6 eV; their energy band diagram is presented in Figure 3.2 d. (b(b) )PhVootoltvaogleta mgaep (c)( Pc)hoCtoucrurerrnet nmt ap (a) (b(b()b )P )Phottovvooltage (c) tocurrentX, Y, Z piezo Voltagleta mgaep (c)( Pc)hoCtoucrurerrnet nmt ap(a) (a) X, Y, Z piezo AFM controller AFM controller Diode Detector Diode Laser ? m 2 Diode Detector Diode Laser ?m 2 Pt/Ir Tip -1050 -900 mV 75 655 pA Pt/Ir Tip -1050 -900 mV 75 655 pA ((dd)) (d) Pe ((dd))rovPsekritoevskite PMM PMA: MP AC :PB CM BM -3.8 -3.8 T Ti iO O2 -3.92 -3.9 P-3e.r8ovskite ProbeFT P-3e.r8ovskite Probe F OTO -3-.P39.C9BMGla -4.2 -P3.C9BPMerovPskeite AFM Tip -5.0 ss s Perovskirtoevskite ProbAeFM Tip -5.0Gla us bstrate -4.5s -4.2 -4.2 -5.4 -5.0 substrate -4.5-4.5-4.5 T-iO4.2 -5.0 2 -5.9 -5.4 -5.4 1 sun, 550 nm FTO TTiO2 TiO iO -25.9 -5.9 -5.42 1 sun, 5~5 100 n0m mW/cm2 Laser FFTTOOFTO TiO2 -5.9 ~ 100 mW/cm2 La -7.2ser FTO -7.2 -7.2 -7.2 Figure 3.2: Nanoscale imaging of photovoltage and photocurrent in multi-cation perovskite solar cells. (a) Schematic representation of illuminated-KPFM and pc-AFM operating principles. Examples of (b) photovoltage and (c) photocurrent maps in halide perovskites overlaid over 3D plots of their respective topography images. (d) Energy band diagram of multi-cation half-cell device, showing charge mobility and collection of holes during contact-mode pc-AFM imaging. 3.3.3 Band bending insights through dark KPFM We first resolve the surface potential for both dual- and quad-cation perovskites under dark KPFM conditions to determine the band bending distribution of GBs as a function 22 2.2 mm 2.2 mm EEneEnrengerygrg yLy eL LveeevveeEnEenrgeryg yL eLl ve( ll e((eeVVV) )evel l( ()eeVV)) PPCCBBMM 40 nm PCBM 40 nm 2 ?m 2 40 nm ?m 40 nm of the perovskite composition (see Figure 3.3 and Figure 3.4, respectively). We observe upward band bending in GBs for both compositions and note that the incorporation of Rb+ increases voltage homogeneity (Figure 3.4) in quad-cation perovskite. The topography images, shown in Figures 3.3 a and 3.4 a, reveal slightly larger grains in the dual-cation thin films compared to those in quad-cation. To measure the contact potential difference values of the films, dual-pass amplitude modulated KPFM (see Experimental Methods for details) was used. The contact potential difference (CPD) maps reveal an increased voltage signal in the GBs for both samples (Figures 3.3 b, c and 3.4 b, c). The line profiles (Figures 3.3 d - m and 3.4 d - m) for the representative GBs present higher CPD values compared to the adjoining grains for both samples, indicating upward band bending at the GBs. This forms an effective electron barrier at GBs in both perovskite compositions, analogous to previous reports for RbF- and CsF- treated CIGSe solar cells. [89] 23 Topography nm (a) d j l 40 (d) (i) e -700 -70020 f k 0 -800 -800 i 0 200 400 0 250 500 750 -20 (e) -600 (j) g h m -40 -700 -650 Topography Surface potential !CPD = 48.0 mVmV -800 -700CPD (a) (b) (b) (dc) l 150 j 0 200 400 0 200 400100 -700 (f) -700e (k) 50 -750 f k -7500 -800 i -800-50 -850 0 200 400 m 0 200 400g h -100 (g) (l) -150 -700 -750 Topography Surface potential (c) !!CPD = 4CPD= 382..00 mmVV ((ad)) Topography ((be)) Surface potential 2000 ((cf)) !CPD = 48.0 mV -750 -800 (a) (b) (c) 0 200 400 600 0 100 200 300 4000 (h) (m) -700 6000 -700 -800 0 -800 -200 -100 0 100 200 0 100 200 0 200 Voltage! (!mCVP=D)= 32.0 mV Distance (nm) Distance (nm) CPD 32.0 mV (d) (d) (e) (e) (f) (f) Figure 3.3: Upward band bending at GBs in dual-cation perovskites. (a) Topography and (b) contact potential difference maps for dual-cation perovskite (Cs0.17FA0.83Pb(I0.87Br0.17)3). (c) Histogram computed from the dark-KPFM image. (d-m) CPD line profiles across GBs (dashed lines in maps a and b) reveal upward band bending at GBs. The upward band bending at GBs occurs with the structural changes brought about by the addition of Cs+ and Rb+. A small cation such as Rb+ when added to the A- site causes the cubo-octahedral structure to tilt due to the distortion of the Pb-X frame, which changes the electronic assembly at the band edges. [90] The tolerance factor for dual- and quad-cation perovskite is 0.96, calculated using the values in Table 3.1. From the CPD maps it is evident that the voltage distribution at the GBs is consistently lower 24 Quad-cation QDuuaald--ccaattioionn Dual-cation perovskite ppeerroovvsskkiittee perovskite Quad-cation Dual-cation perovskite perovskite Counts CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) than at the grain. We quantify the difference in voltage heterogeneity between the two samples by deriving histograms (Figure 3.3 c and 3.4 c) from CPD maps (same bin widths and positions). The quad-cation perovskite presents a 16 mV (34%) reduction in voltage standard deviation (?CPD) (Figure 3.4 c), ultimately displaying a more uniform voltage response. This superior voltage performance may be due to the ability of Rb+ to increase halide homogeneity, as previously indicated by nano-X-ray fluorescence measurements. [77] Topography nm (a) -900 (d) (i) d i 40j -920 -900m h 20 -940 -960 -1000 k 0 l 0 100 200 0 100 200e g -20 (e) f -850 -900 (j) -40 -875 Topography Surface potential !CPD = 48.0 mV -950 (a) (b) (c) CPD mV -900 (b) 100 ! 0 200 0 100 200Topography Surface potential d i CPD = 48.0 mVj 50 -825 (f)(a) (b) (c) m -850 (k) h 0 -850 -875 k -875 -900 e l -50 -900g 0 100 200 300 0 50 100 150 f -100 -840 (g) (l)-850 !CPD= 32.0 mV -150(c) -860(d) (e) (f) !CPD= 32.0 mV -880 -900 (d) (e) 8000 (f) -900 6000 0 100 200 0 100 200 300 (h) (m) 4000 -900 -900 2000 -950 -950 0 -200 -100 0 100 200 0 100 200 0 100 200 300 Voltage (mV) Distance (nm) Distance (nm) Figure 3.4: Increased voltage homogeneity in quad-cation perovskites. (a) Topography and (b) contact potential difference maps for quad-cation perovskite (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskites. (c) Histograms computed from the dark-KPFM images reveal a 34% increase in voltage homogeneity. (d-m) CPD line profiles across GBs (dashed lines in maps a and b) reveal upward band bending at GBs. 25 Quad-cation Dual-cation perovskite perovskite Quad-cation Dual-cation perovskite perovskite Counts CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) CPD (mV) Tolerance Factor Calculations: Species Ionic radius (?) Cs+ 1.74 Rb+ 1.64 MA+ 2.17 FA+ 2.53 Pb2+ 1.19 I- 2.2 Br- 1.96 Table 3.1: Values of ionic radii for difference species in the perovskite structure ABX3. [91?93] Equations: rDual_Cation_eff = xrA1 + (1? x)rA2 (3.1) ranion_eff = xrX1 + (1? x)rX2 (3.2) rQuad_cation_eff = xrA1 + yrA2 + zrA3 + (1? (x+ y + z))rA4 (3.3) r?cation_eff + ranion_effteff = (3.4) 2(rPb2+ + ranion_eff ) Here, rDual_cation_eff is the effective ionic radius calculated using the contribution of the two cations (Cs+ and FA+) present at the A-site of the dual-cation perovskite. In equation (3.1), rA1 and rA2 are the ionic radii values of the two cations and x and 1-x are the fractions of the cations, respectively. In Equation (3.2), ranion_eff is the effective ionic radius of the two anions (I?, Br?) at the X site of the dual- and quad-cation perovskite. In Equation (3.3), rQuad_cation_eff is the effective ionic radius calculated by the contribution of the four cations (Rb+, MA+, Cs+ and FA+) present at the A-site of the quadcation perovskite. x,y,z and 1 ? (x+y+z) are the fractions of the cations, respectively. In Equation 26 (3.4), teff is the effcetive tolerance factor value and rPb2+ is the ionic radius of Pb2+ in the B-site of the perovskite. Using Equations (3.1), (3.2) and (3.4), plugging in values from Table 3.1, tolerance factor for dual-cation perovskite was calculated to be 0.9616. Similarly, using Equations (3.2), (3.3) and (3.4), tolerance factor for quad-cation perovskite was calculated to be 0.9617. For perovskite, tolerance factor values between 0.9 and 1.0 are considered structurally stable. 3.3.4 Imaging ion motion through illuminated KPFM We unravel the role of Rb+ in the local voltage response of perovskites by performing KPFM measurements throughout an illumination cycle (light OFF-ON-OFF) and find substantially different voltage decay profiles for the two compositions, indicating that the presence of Rb+ suppresses post-illumination residual voltage (Figure 3.5). Overall, under illumination (light ON), reduced band bending is observed in comparison to dark (light OFF) conditions. The surface voltage measurements for the dual-cation perovskite (Figure 3.5 a-g) reveal that the spatial characteristics under dark and illuminated conditions remain unchanged, with grain boundaries presenting higher voltage values relative to their cores. Surface voltage maps for the quad-cation perovskite show similar, unchanged spatial characteristics (Figure 3.5 h-n). Note that a black dashed arrow next to each scan denotes its direction (top to bottom/bottom to top), and the color-coded timescale indicates the ending time for each scan (different for each sample to highlight the distinct timeframe of their transient electrical responses). To make the voltage response comparable, the mean 27 of the first dark scan for dual-cation was set as a reference value and subtracted from itself and the subsequent scans (Figure 3.5 a-g), resulting in the mean of Figure (a) as zero. Correspondingly, for quad-cation perovskite, the mean of the first dark scan was set as the reference value and the scans shown in Figure 3.5 h-n are relative to this mean. The mean for Figure 3.5 h is centered at zero. Two surface photovoltage scans for dual-cation perovskite and quad-cation perovskite are shown in Figure 3.5 b, c and Figure 3.5 i, j, respectively. The total duration of illumination for the quad-cation perovskite is 16 min (the full series is presented in Figure 3.6), while the total duration of illumination for the dual-cation perovskite is 8 min. After the light is turned OFF (Figure 3.5 d), a residual voltage emerges through the gradient in the scan direction (bottom to top). Even after the extended illumination (16 minutes instead of 8 minutes), quad-cation perovskite displays a greater residual voltage drop as evidenced by a 55% decrease in the mean voltage of the first post-illumination scan (Figures 3.5 k, 3.7). After four consecutive dark scans (images (k-n), the voltage equilibrates and essentially reaches the levels for the initial dark map (shown in Figure 3.5 h). All illuminated-KPFM imaging takes place while subjecting the devices to a broad-spectrum illumination [94] (centered around 550 nm ? see Experimental Methods) from the glass side, as illustrated in Figure 3.2, with the power calibrated to 100 mW/cm2 to approximate 1-sun illumination. [95,96] 28 Dual-cation Quad-cation (a) (h) (o) Time (min) 4 4 4 4 (b) (i) (r) 8 8 8 8 ? (c) (j) (s) 12 20 12 20 (d) (k) (t) 16 24 16 24 Time (e) (l) (u) 20 28 20 28 (f) (m) (v) 24 32 24 32 (g) (n) (w) 28 36 28 36 - 100 + 700 Surface voltage (mV) Surface voltage (mV) Figure 3.5: Rb suppresses post-illumination residual voltage in halide perovskites. KPFM images across an OFF-ON-OFF illumination cycle for the (a-g) dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and (h-n) quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. The black dashed arrows indicate the scan direction. (o-w) Histograms for the KPFM images showing lower surface voltage values and a gradient of voltage relaxation for the dual-cation perovskite. The black dashed line marks the mean voltage value for the initial dark scans. The inset in (t) presents the voltage drop distribution for the dual-cation perovskite as a function of time for the scan (d). The color-coded time scale indicates the ending time for each scan. Both samples are bottom-illuminated with an LED with an emission spectrum centered at 550 nm at a power density of 100 mW cm?2. Each scan was completed in 4 min. 29 Time (min) Time (min) Counts (norm.) Light OFF (4 min) Light ON (8 min) Light ON (12 min) Light ON (16 min) Light ON (20 min) Light OFF (24 min) Light OFF (28 min) Light OFF (32 min) Light OFF (36 min) Figure 3.6: Surface potential measurements of the quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite across an OFF-ON-OFF light cycle. The time in minutes represents the start time for each scan. The sample were bottom-illuminated with an LED with an emission spectrum centered at 550 nm at a power density of 100 mW cm?2. A subset of the data was used in Figure 3.5 h-n. 30 (a) (b) Figure 3.7: Highlighting the larger photovoltage drop, post-illumination, exhibited by quad- cation perovskite. (a) Last illuminated scan for dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite (Figure 3.5 c) and quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite (Figure 3.5 j) concatenated with the first post-illumination scan for dual-cation perovskite (Figure 3.5 d) and quad-cation perovskite (Figure 3.5 k) resolved as a function of time. (b) Increased focus on the photovoltage drop post-illumination (shown by dashed rectangle in (a)). 31 Dark conditions Period of illumination for Dual-cation perovskite Period of illumination for Quad-cation perovskite Figure 3.8: The quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) film presents a faster residual voltage decay despite longer illumination. Average surface voltage versus change in time (?t) for the dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) (red) and quad-cation (blue) perovskites. The image averages are extracted from the series shown in Figure 3.5 a-g and 3.6, respectively. The error bar displays the scan standard deviation. The shaded regions represent the period of illumination and dark scans for each composition. The photovoltage behavior of both compositions is quantified by extracting histograms from the images (Figure 3.5 o-u), along with the average values of the scan (Figure 3.8) and the standard deviation (Figure 3.9). Throughout the illumination series, the quad-cation perovskite presents 25% less voltage variance (computed from the standard deviation of the histograms shown in Figure 3.5 p and q) despite having a larger surface photovoltage response. Furthermore, the histograms in Figure 3.5 t underscore the difference in the drop in residual voltage, which is further highlighted in Figure 3.7. Rb+ suppresses the residual voltage, possibly due to reduced ion migration and/or interfacial 32 recombination. We stress that the quad-cation sample has a shorter residual voltage even after being subjected to illumination for twice the duration of the dual-cation sample as shown in Figure 3.8. There are two possible situations that have been foreshadowed by the extended duration of quad-cation perovskite: (1) the residual voltage for the quad-cation material plateaus after a certain illumination dosage has been reached or (2) the residual voltage decay would occur faster under illumination conditions of equal durations. Both (1) and (2) would additionally support the conclusion that Rb+ reduces ion migration and accelerates the post-illumination residual voltage drop that is commonly observed in metal halide perovskites. [18,42,97] 33 (a) Dual-cation (b) Quad-cation (c) Surface voltage (mV) (d) Surface voltage (mV) Figure 3.9: Tracking the surface voltage distribution for both dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. (a, b) Histograms computed from the KPFM images displayed in Figure 3.6 and 3.7 reveal that, despite the longer light soaking, the quad-cation material exhibits more rapid equilibration when returned to dark conditions. The SPV variance values, denoted by ?SPV , (c, d) indicate that the dual-cation shows greater voltage heterogeneity across all light conditions (average SPV variance of 9.32 versus 5.88 mV, for dual- and quad-cation, respectively). 3.3.5 Mapping local photocurrent through pc-AFM While the macroscopic photocurrent response of the devices is similar (see J-V measurements in Figure 3.1), at the mesoscale the current collection from the perovskite layers is quite distinct. Figures 3.10 a and 3.10 b show topography and photocurrent 34 maps for the dual-cation sample, which are uncorrelated (Figure 3.10 c). We extract representative line profiles (Figure 3.10 d-i) through grain interiors and GBs that reveal uniform photocurrent at the grain cores. From Figure 3.10 b, the majority of GBs display photocurrent values similar to dark current, ? 72 pA. Although this observation seems in contradiction with the upward band bending established at the GBs from the CPD maps in Figure 3.3, the absence of holes at the GBs is probably due to their consumption at the defects. In turn, defect-assisted recombination occurs at these GBs (see some examples marked with a black dashed circle in Figures 3.10 d and e). However, we find a minority of GBs that present photocurrent values ? 250 pA, a consequence of the large number of free holes, which are collected at the probe end during the pc-AFM measurements. The presence of these free holes at the GBs can be related to the passivation of defects by Cs+ or FA+ (see Figures 3.10 g and i, marked with a green dashed rectangle). Defects at GBs for perovskites can be both negatively (n-type) and positively charged (p-type) [81]; Cs+ and FA+ neutralize the negatively charged defects (n-type) instead of holes. In turn, the holes are free to participate in conduction. The reason why only a minority of GBs are passivated is due to the scarcity of free A-site cations, which arises in dual-cation structures because Cs+ and FA+ form an alloy, as has been previously confirmed through NMR measurements. [90,98] Further, evaluating the overall photocurrent response, dual-cation devices present ? 20% photo-inactive grains and GBs (Figure 3.12). The average grain diameter, surface area, and surface roughness (found using a multi-step watershed algorithm comprising grain detection, segregation, and distribution) are 275 nm, 40,695 nm2 and 16.5 nm, respectively, for dual-cation perovskite (Figure 3.13, 3.14). 35 (a) (a) Topography mapf f h g h (b) (b) Photocurrent map (c) (c) g (a) (a) f (a) f h (b)g h (b) (b) (c) (c) (c) g f f e i i ei i i e d e g i d g e e d d h hd d (d) ((dd)) (d) (d)(e) (e(e)) (e) (e)(f) (f()f) (f) (f) (g) ((gg)) (g) (g)(h) (h(h)) (h) (h) (i) (i()i) (i) (i) Vertical charge carVrGiTeeyrBrtp iG cewaB Ali tchh adVregrfete iLccaTatylr e rcpriheaearl s BrcGsghBieva rcagateriVor ciTeneayrrrtp riGLcieeaB rtAl ecGrhaBal rcghea LcraGTgatyereBrp ricee awrl r BcrGiihteBhar r dGgeB fceacrrtLiseart eGraBl charge carrier GB Figure 3.10: Resolving photocurrent in dual-cation perovskites at the nanoscale. (a) Topography and (b) photocurrent maps of dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite. (c) Scatter plot and histograms of height and photocurrent confirm no correlation between the two. (d-i) Line profiles for height and photocurrent for six representative regions of the sample, as indicated by the dashed lines in (a) and (b). To unravel the role of Rb+, we perform a similar pc-AFM investigation on the quad- cation perovskite. Figures 3.11 a and 3.11 b show the topography and photocurrent maps, respectively. The photocurrent distribution presented in Figure 3.11 c expresses an almost Gaussian trend, with no correlation between topography and photocurrent. The dark regions in the quad-cation photocurrent map also express uniformly saturated current values similar to the average dark current (see Figure 3.11 d). Here, a majority of GBs in the quad-cation perovskite have higher photocurrent in comparison to their adjacent grains (Figure 3.11 e, g, h, i, marked by the green dashed rectangle). The 36 reason behind these highly photoactive GBs is defect-passivation by Rb+. [98] We ponder three main reasons for the defect passivation at GBs by the addition of Rb+ to the perovskite ? (1) Unlike Cs+, Rb+ does not form an alloy with FA+ [90, 98] making Rb+ available to participate in the passivation of n-type defects at GBs. (2) A recent investigation of depth-dependent chemical composition on MAFA, CsMAFA, RbMAFA, and RbCsMAFA perovskite revealed that the presence of both Rb+ and Cs+ together in quadruple cation perovskite (CsRbMAFA) makes the distribution of Cs+ and Rb+ cations highly homogeneous within the surface and bulk of the sample. [99] (3) Due to the small ionic radius of Rb+, it is more easily released to the surface from the bulk of the material. [99] We observe very few GBs that still have defects and show photocurrent values similar to dark current (Figure 3.11 i). This highly conductive GB network enhances the overall photocurrent response by ? 50% (Figure 3.10 b, 3.11 b). We find ? 11% photo- inactivity in the quad-cation perovskite (Figure 3.12). The average grain diameter, surface area, and surface roughness for quad-cation perovskite, found by the watershed algorithm, are 230 nm, 36,245 nm2, and 14.6 nm, respectively (Figure 3.13, 3.14). 37 (a) f (a) f Tohpography maph (b) (b) Photocurrent map(c) (c)g (a) f (a) gf (a) f h (b) h (b) (b) (c) (c) (c)g g f g h g h i i i i e ee e i e e i d d d d d d (d) ((dd)) (d) (d)(e) ((ee)) (e) (e)(f) ((ff)) (f) (f) (g) ((gg)) (g) (g)(h) ((hh)) (h) (h) (i) ((ii)) (i) (i) Vertical charge carVrGiTeeyrBrtp iG cewaB Ali tchh adVregrfete iLccaTatylr e rcpriheaearl s BrcGsghBieva rcagateriVor ciTeneayrrrtp riGLcieeaB rtAl ecGrhaBal rcghea LcraGTgatyereBrp ricee awrl r BcrGiihteBhar r dGgeB fceacrrtLiseart eGraBl charge carrier GB Figure 3.11: Enhanced photocurrent in quad-cation perovskites due to Rb+. (a) Topography and (b) photocurrent maps of quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. (c) Joint-plot of histograms and scatter data points of height and photocurrent data show the distribution of each and no correlation between the two, respectively. (d-i) Line profiles for height and photocurrent for six representative regions of the sample, as indicated by the dashed lines in (a) and (b). 38 Avg. inactivity = 21.1% Avg. inactivity = 10.6% Figure 3.12: (a-b) Dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and (g-h) quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite flattened topography map. (b-c) Dual-cation and (i-j) quad-cation (6 ?m x 6 ?m) current maps measured at 1 sun illumination from a light source centered at 550 nm. Current distribution histograms for (e-f) dual-cation and (k-l) quad-cation with fraction of inactive grains (< 72 pA) reported on the top right corner. 39 Quad-cation perovskite Dual-cation perovskite Watershed Mask Watershed Mask 40 60 (a) (b) 0 0 - 60 - 60 (c) (d) (e) Figure 3.13: Grain statistics using watershed algorithm on topography maps shown in Figure 3.3 a and 3.4 a. Grain detection on topography map using 40 steps and 0.11% drop size for grain location, 700 steps and 1.3% drop size for segmentation for (a) dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite and (b) quad- cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. (c) The trend between surface area of grains and grain diameter is nearly identical for dual and quad-cation perovskites. Similar (d) grain size distribution and (e) surface area of grains distribution is observed for both dual and quad-cation perovskite. 40 nm nm Watershed Mask Watershed Mask 100 40 (a) (b) 0 0 - 120 - 60 (c) (d) (e) Figure 3.14: Grain statistics using watershed algorithm on topography maps shown in Figure 3.10 a and 3.11 a. Grain detection on topography map using 40 steps and 0.11% drop size for grain location, 700 steps and 1.3% drop size for segmentation for (a) dual-cation perovskite (Cs0.17FA0.83Pb(I0.83Br0.17)3) and (b) quad- cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite. (c) The trend between surface area of grains and grain diameter is nearly identical for dual and quad-cation perovskites. Similar (d) grain size distribution and (e) surface area of grains distribution is observed for both dual and quad-cation perovskite. For understanding the voltage dependency of photocurrent at the nanoscale, we apply bias to each sample from 0 V to 500 mV at 30 mV steps and map photocurrent values. Plotting the average photocurrent from each map (Figure 3.15, 3.17), against the applied bias we uncover the trend shown in Figure 3.19. For dual- and quad-cation perovskite, the photocurrent dependence on voltage shows parallel trends. From the current maps for the dual-cation perovskite in figure 3.15 and the quad-cation perovskite in figure 3.16, it can be inferred that photocurrent decreases to the average dark current value throughout the sample for applied bias at and beyond 160 mV. This is further confirmed by photocurrent distribution analysis, for each photocurrent map at the applied bias, shown in Figure 3.16 41 nm nm and Figure 3.18 for dual-cation and quad-cation perovskite, respectively. When comparing the short-circuit photocurrent values for dual- and quad-cation perovskite, we see a ? 50% increase in photocurrent due to the addition of Rb+ to quad-cation perovskite. In addition, a higher photocurrent value in quad-cation perovskite is seen up to an applied bias of 70 mV. At and beyond 100 mV, the photocurrent, for both samples, converges to similar values. Figure 3.15: Dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite current maps at biases applied with a 30 mV step between each scan. Light source: 550 nm, 1 sun illumination, ? 100 mW/cm2. 42 Figure 3.16: Dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite current distribution histograms at biases applied with a 30 mV step between each scan. Saturated current observed at and beyond 160 mV applied bias. 43 Figure 3.17: Quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite current maps at biases applied with a 30 mV step between each scan. Light source: 550 nm, 1 sun illumination, ? 100 mW/cm2 44 Figure 3.18: Quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite current distribution histograms at biases applied with a 30 mV step between each scan. Saturated current observed at and beyond 160 mV applied bias. 45 1601 m60V mV160 mV 160 mV Figure 3.19: The addition of Rb shows higher photocurrent and a negligible influence on the voltage dependency of photocurrent at the nanoscale. Avergae photocurrent versus applied bias (30 mV steps). Quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite displays a higher short-circuit photocurrent yet converges to a similar level for voltages ? 100 mV. For applied voltage ? 160 mV, the current values across the map are saturated for both dual and quad cation (Figure 3.15 and 3.17). The error bars represent standard deviation values from the mean of each photocurrent scan imaged at the respective applied bias. 3.4 Discussion The juxtaposition of the electrical behavior at the nanoscale for the two chemical compositions reveals superior electrical performance by quad-cation perovskite. As stated earlier, for both dual- and quad-cation samples photocurrent is independent of topography. From Figures 3.10 b and 3.11 b, the latter perovskite has a photocurrent response 50% higher compared to the former. In both compositions, two types of boundary are present: Type I, i.e. GBs with defects, and Type II, i.e. GBs with defect-passivation. However, 46 the dominant type in each chemical composition is different: the majority of GBs in the dual-cation film are Type I, whereas they are mostly Type II in the quad-cation film. The interaction of upward band bending, hole accumulation, and defect recombination at GBs is illustrated as a schematic in Figure 3.20. The schematic highlights the type of GB that is primarily found for each chemical composition and the movement of charges based on our KPFM and pc-AFM measurements. Figure 3.20 a shows the charge flow in the dominant Type 1 (no passivation) GBs in dual-cation perovskite. The energy levels for the conduction band (EC), valence band (EV ), Fermi energy (EF ) and defect energy (ED) are indicated using solid and dashed lines. Qualitatively, the upward band bending at GBs, derived from the KPFM results, is translated to the energy bands displayed in the schematic. The holes from the valence band of Grain 1 accumulate at the valence band of the GB, however, due to upward band bending, the formation of an electron barrier depletes electrons at the GB (as illustrated in Figure 3.20 a). Because of the presence of defects, the electrons from the conduction band and holes from the valence band cross into the defect energy level, where they recombine. This incident is termed defect-assisted recombination and is shown by white arrows in Figure 3.20 a. Because this process consumes the free charge carriers when the AFM probe makes contact with the GB, there are no free charge carriers to participate in conduction, and the photocurrent is similar to the dark current. Interestingly, for quad-cation perovskite, with the exception of a few, we note the dominance of GBs where defects have been passivated by Rb+. In Figure 3.20 b, we illustrate the movement of charges at the grains and GBs in quad-cation perovskite. The holes accumulate at the valence band of the GB (because of upward band bending). The electrons from the conduction band of Grain 1 47 do not move toward the conduction band of the GB (due to electron barrier formation), and this further prevents the recombination of electrons and holes at the boundary. The elimination of recombination of free charge carriers at the GBs allows the charge carriers to spontaneously participate in electrical conduction and therefore are collected by the probe during photocurrent mapping (Figure 3.20 b). As a result, the majority of the boundaries exhibit higher current values than the grain cores in the photocurrent scan, as displayed in Figure 3.11 b. (a) AFM probe ? No free-charge carriers due to defect-assisted Ec recombination E E DF Ev Grain 1 GB Grain 2 GB with defects (b) AFM probe ? Free charge carriers participate in electrical Ec conduction EF Ev Grain 1 GB Grain 2 GB with Rb+-passivated defects Figure 3.20: Charge carrier mobility at GBs. Schematic of upward band bending at GBs in (a) dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) perovskite, where defect-assisted recombination consumes free carriers (shown by white arrows), and in (b) quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite, where defect-passivation allows free charge carriers to participate in electrical conduction. (Schematic not to scale for clarity) 48 Quad-cation perovskite Dual-cation perovskite Overall, the addition of Rb+ notably improves nanoscale photovoltaic characteristics in quad-cation perovskite (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3). Quad-cation perovskite compositions have offered some of the highest PSC efficiencies to date, yet macroscale device performance can only be fully understood by considering the nanoscale analysis of optoelectronic properties at length scales of individual grains in the perovskite material. Thus, recording the spatial and local responses of photovoltage and photocurrent at the nanoscale provides insights into how the heterogeneities within grains and GBs affect charge carrier conduction and collection. Here, we analyzed and compared the nanoscale photoactivity of two half-cell device samples with different chemical compositions, but otherwise identical device structures. At the macroscale, we find similar photocurrent performance for both dual and quad-cation devices, while at the nanoscale we measure a 50% increase in photocurrent in the latter perovskite. This apparent discrepancy in the macroscopic and nanoscopic analyses arises because the macroscopic J-V measurements are performed on the full device and nanoscale photocurrent maps are acquired by contacting the perovskite layer in the half-cell. The local photocurrent response of the grains and GBs reveal electrical performance comparisons of the perovskite material that remain hidden at the macroscopic level. Our nanoscopic measurements provide a quantitative comparison between two state-of-the-art halide compositions with noticeable promise for photovoltaics. 3.5 Conclusions In summary, we demonstrated the superior electrical performance of quad-cation perovskites at the nanoscale, which stems from the presence of Rb+ at the A-site, using 49 complementary data obtained from KPFM and pc-AFM. We performed a comparative analysis between dual-cation and quad-cation compositions and quantified the beneficial effects of the addition of Rb+ on the electrical properties of perovskites. Our findings demonstrate 34% increased voltage homogeneity, 25% lower photovoltage variance, and 55% greater post-illumination voltage drop for quad-cation perovskite. In particular, the voltage maps revealed upward band bending at GBs, which in conjunction with photocurrent maps led to the revelation of defect-assisted recombination of free charges at the majority of GBs in dual-cation perovskite. On the contrary, the GBs in the quad-cation perovskite are highly conductive due to a large accumulation of holes, resulting in a 50% increase in the overall photocurrent response. Our comprehensive electrical behavior analysis to identify defects and their passivation at GBs could be readily applied to study the nanoscale electrical characteristics of other perovskite chemical compositions. Our methodology can be expanded by performing a complete illumination assisted in-situ and ex-situ probing of electrical characteristics, i.e., photovoltage and photocurrent, such as the one presented in this paper, under different environmental conditions (humidity, temperature, oxygen) to gain knowledge about electrical response of individual grains and GBs. 50 Chapter 4: In Situ Humidity - Dependent KPFM on Metal Halide Perovskites Metal halide perovskites have become one of the leading materials for next-generation photovoltaics due to their promising optoelectrical properties. However, the instability of perovksite materials under environmental factors, especially humidity, limits device operating lifetime. Nanoscale mapping of photovoltaic properties allows detection of local electrical behavior which can be directly correlated with morphology and linked to overall device performance. In this work, we perform, for the first time, in situ humidity-dependent Kelvin probe force microscopy (KPFM) on Cs0.33FA0.67PbI3 and capture real-time surface voltage behavior resolved at length scales of grains and grain boundaries. We observe an enhanced voltage response up at 25% and 45% relative humidity and a 67% drop in the voltage response at 65% rH indicating electrical failure. While the voltage response recovered by 94.34% post humidity-cycle, significant moisture-induced electrical instability is observed during the humidity cycle which we infer originates from structural and chemical changes at the surface of the perovskite as observed through XPS measurements on the pristine and humidity-cycled sample. Our results provide insights of moisture-induced variations on the local photoelectrical properties of metal halide perovskites to help further their development by improving moisture stability. 51 This chapter is adapted from R. Lahoti, et al. (TBD) Surface Voltage oC (KPFM) N2 Inert Temperature Sensor % 1 Sun Illumination Humidity Sensor 52 4.1 Introduction Hybrid organic-inorganic metal halide perovskites have shown great potential for applications in photovoltaics, having reached a power conversion efficiency (PCE) of 25.7% comparable to commercial Si-based solar cells. [2] However, perovskite solar cells (PSCs) are far from operational stability. [100?103] Efforts are being directed toward a fundamental understanding of the perovskite material that are related to long-term stability of the devices. [104?106] One direction toward stability is the composition and structural engineering of the traditional perovskite chemical structure, ABX3. [107] Although the incorporation of mixed cations such as Cs+ and FA+ at the A-site has improved device stability, [83, 108?110] perovskite instability under environmental stressors [111] such as oxygen [21, 22], temperature [35, 112], light [113?115] and humidity [115, 116] remains an obstacle in their applications. In situ measurements under varying extrinsic factors are a powerful method to capture real-time dynamics in material behavior and point toward stability. There have been macroscopic in situ analysis on temperature dependent photoluminescence [117], bias dependent oxygen diffusion [118], material degradation [119, 120], photo- and moisture- dependent phase evolution [121] and humidity-induced photoluminescence hysteresis [116] behavior in perovskite thin films. While the effects of humidity during the fabrication [122?125] of the perovksite thin films and the macroscopic effects [116, 121] are well represented in literature, the effect of humidity on local photovoltaic properties at the nanoscale has not been investigated in 53 metal halide perovskites. While macroscopic analysis of photovoltaic parameters provides valuable insights at the device level [126, 127], nanoscale investigation through functional imaging techniques allows mapping of optoelectrical properties at the material level. Atomic force microscopy is a state-of-the-art characterization tool that maps material properties at microareas, capturing morphological and functional data simultaneously. AFM has been utilized in exploring many areas of perovskite material optimization for photovoltaic applications. Kelvin probe force microscopy (KPFM) is an AFM-based technique that maps the contact potential difference in semiconductors. [49,54] Here, we present, for the first time, in situ humidity-dependent KPFM on Cs0.33FA0.67PbI3 by cycling the humidity from 5 - 65% and down to 65 - 5% in steps of 20% relative humidity (rH). We explore real-time spatially resolved surface voltage response that provides information about microscale electrical stability. The voltage response is observed to be unstable under humidity. Enhanced electrical characteristics are observed up to 45% and we note an electrical failure at 65% rH. This failure is attributed to structural and surface chemistry changes seen post humidity cycling through XPS measurements. Furthermore, the intensity of the bulk PL spectra after the humidity cycle decreased by 50%, pointing to losses due to nonradiative recombination. Capturing the sample voltage response at the end of the humidity cycle, at 12 hours and 24 hours after the humidity cycle (under 5% rH) revealed 96.34%, 98% and 99.4% recovery, respectively. 54 4.2 Experimental set up 4.2.1 Humidity enclosure A humidity stage accessory from Asylum Research Oxford Instruments (shown in Figure 4.1 a) was used for sensing real-time humidity in this work. However, the original stage was not designed for electrical measurements. The stage was customized by creating screw holes around the stage along with the addition of a metal block with a screw hole to screw in a metal clip that makes an electrical contact with the sample (Figure 4.1 b). A schematic of the humidity enclosure is shown in Figure 4.1 c. The enclosure comprises of the following parts: a magnetic stage with an attached humidity sensor, a metal clip that makes an electrical contact on the conducting fluorine-doped tin oxide (FTO) layer of the half-cell perovskite sample, a bias wire that connects the metal clip to the cantilever holder, a flexible membrane around the cantilever that makes a seal around the sample stage and a metal disc that keeps the membrane sealed to the magnetic stage. 55 (a) Before (c) Metal disk Cantilever holder (b) After customization Flexible membrane Humidity sensor FTO Perovskite Magnetic sample stage Figure 4.1: Humidity enclosure. (a) Humidity sensing sample stage by Asylum Research Oxford Instruments. (b) Humidity sensing sample stage with electrical customization. (c) Schematic of enclosure for in situ humidity-dependent electrical measurements. 4.2.2 Humidity control Relative humidity (rH) is a ratio of the amount of water present in the air at a given temperature to the greatest amount possible at that temperature. Relative humidity control can be achieved through a regulated mixture of dry air and wet air. In this work, relative humidity was controlled using the experimental setup shown in Figure 4.2. The schematic from left to right: a nitrogen line is subdivided into two lines - dry and wet. The mass flow controller regulates the flow of N2 through the water bubbler based on the set point voltage transmitted by the control computer (not shown). The dry and wet lines are merged, and then the combined flow is introduced into the humidity sensing enclosure. The line exiting the enclosure on the right is passed out of the AFM hood. The samples are 56 illuminated through the bottom of the sample stage. The minimum and maximum relative humidity value that can be achieved using this set up are 5% and 100% respectively. Nitrogen Purge Line N Humidity Sensing I T Enclosure R (c) PhotocurrentO G E N Out Mass Flow Controller Water Bubbler (d()d)(d) 1 Sun Illumination -3-.-383 .8.8 M Tip .-- .9 933.9 PPeerroovvsskit e AAFMFTMipTip -3 AFPerovskitekite -5-.05.0 --44..2 -5.0 - -4.5 -4.2 2 -5.4 -4.54.5 9 -5. -45.4 TTiOiO2 -5. FTO TiO2 2 -5.-95.9 FTFOTO - 2 -7.-27 7..2 Figure 4.2: Schematic of experimental set up for in situ humidity-dependent KPFM measurements. The experimental setup shown in Figure 4.2 is calibrated by modulating the dry line pressure and the mass flow controller set point voltage. In the first step of calibration, the chamber is purged with nitrogen at the highest value of the dry line flowmeter output (150 mm). This purge brings the enclosure environment to 5% rH. An investigation on the effect of gas flow pressure into the chamber on scan quality revealed that gas flow greater than 70 mm of inlet pressure causes scan artifacts during AFM measurements. After the initial purge, the dry line is brought down to 70 mm to ensure scan quality is not compromised and the relative humidity value at 70 mm dry line maintains 5% rH. Modulating the dry line alone from 70 mm to 20 mm creates a humidity variation of 5% rH to 9%, respectively. To achieve a relative humidity value beyond 9%, we capture relative humidity as a function of the Bronkhorst set point value. While keeping the dry line value constant at 20 mm, the set point voltage is varied from 0.0 to 1.0 V with a step size of 0.1 V to achieve high humidity 57 Energy Level (eV) Energy Level (eV) Energy Level (eV) PCBM PPCCBBMM values. For simplicity, the calibration data was collected using the set-point voltage with a single decimal. To enable fine tune of the humidity control and achieve values in between the calibrated values, two decimal values for the set point can be used. Although it is possible to achieve 100% rH with the set up, to avoid any damage to electrical components in the AFM head and to prevent a water-layer formation on the perovskite surface from condensation, the maximum operation relative humidity value was set to 70%. Figure 4.3 a shows the calibration values of relative humidity as a function of the Bronskhorst set point and the interpolated values in between that can be used to achieve any relative humidity value between 5% and 70% rH in the chamber. To reduce reliability on the automated mass flow controller, the relative humidity as a function of the wet line pressure was also recorded for manual operation (Figure 4.3 b). (a) (b) Figure 4.3: Humidity control functions (a) Variation of relative humidity as a function of Bronkhorst setpoint (V) and (b) relative humidity as a function of wet line pressure (mm). The quality of the scan at the calibrated values shown in Figure 4.3 was tested by acquiring topography maps on CdTe. The real-time humidity captured while acquiring the images is shown in Figure 4.4. The test shows the excellent humidity control that can be achieved from 5% to 70% in real time while scanning. Temperature during these 58 measurements was 22.8 ?C. The CdTe solar cell measured has the following device structure: a 4.0 mm of glass substrate, 550 nm of a bilayer indium tin oxide, 50 nm of n-type CdS, and 3.5 m of p-type CdTe. Based on the order of layers and the geometry of the device, the topography measurements were performed on the p-doped CdTe layer. Figure 4.5 (a - l) shows the topography scans at each relative humidity from 5% to 70%. The maps were acquired under standard AC mode using a Si probe with 300 kHz drive frequency, 26 N/m spring. Each scan took 2 minutes 41 seconds, and each humidity value was stabilized for 5 minutes. Total time for each step is 6 minutes 41 seconds. The scans confirm that the calibration values do not interfere with the acquisition of nanoscale spatial data. Figure 4.4: Real-time humidity for CdTe topography maps. Constant temperature 22.8 ?C. 59 Figure 4.5: Topography scans on p-doped CdTe under varying humidity from 5% to 70%. (a-l) Spatial maps at calibration values shown in Figure 4.3 4.2.3 Bottom Illumination An important component of the experimental setup for KPFM measurements is an illumination setup to acquire illumination-assisted KPFM responses that provide information about ion migration in the perovskite material. [128, 129] In this work, we illuminate the sample from the bottom. The bottom illumination set-up from Asylum Research Oxford Instruments shoots a broad-spectrum LED light with peak centered at 550 nm to the sample stage, and the illumination spot location on the sample is fixed. However, to capture real-time immediate response to illumination, it is imperative that the beam spot location coincides with the location where the probe lands on the sample. To achieve this, the bottom illumination pathway was customized and the beam spot made 60 movable. The right-angle prism cage (shown in Figure 4.6 a) was replaced with a custom- built cage (shown in Figure 4.6 b) in which a right-angle prism mirror sits on a rotating mount (Figure 4.6 c) to change the angle of incidence of the LED. This enables the beam spot a movement range of 10 mm. Figure 4.6 d shows the top view of the cage with an adjustable lens tube for focusing the beam spot on the sample. A 25 mm focusing lens is used. The samples are of different thicknesses and focusing the beam on the sample surface is necessary to obtain an accurate illumination power density. (a) Before (b) After (c) Inside (d) Top Figure 4.6: Bottom illumination optics customization. (a) A snapshot of the fixed right angle prism from Asylum Research. (b) A snapshot of the customized mirror cage. (c) Right angled prism mirror mounted on a rotating platform. (d) Top view of the custom cage with an adjustable lens tube for focusing the beam spot on the sample. Figure 4.7 shows the optical pathway for the customized bottom illumination optics. The light enters a 1.5" lens tube and passes through a collimating lens. The beam then goes through an optional band pass filter and is incident on the right-angled prism mirror and shot upward towards the sample stage. The prism on the rotating mount is rotated clockwise or anticlockwise to change the location of the illumination spot on the sample. 61 Sample Sample Stage Focusing Lens Laser Fiber Collimating Lens Turning Filter Mount Figure 4.7: Customized bottom illumination setup. Optical pathway for movable beam spot location on the sample. (Asylum Research MFP3D AFM). The broadband LED spectrum was measured using a Thorlabs compact CCD spectrometer and Thorlabs OSA software, which reveals that the spectrum has two peaks. The smaller of the two is at 440 nm and the main peak of the spectrum is centered at 550 nm. The LED is controlled using the igor pro software. 62 Figure 4.8: Broadband LED spectrum centered at 550 nm used as bottom illumination source for KPFM measurements. 4.3 Experimental Methods Perovskite thin film fabrication: The thin films studied in this work were fabricated by our collaborators, the Correa-Baena group, at the Georgia Institute of Technology. [130] 1.0 M perovskite precursor solution is prepared for thin film fabrication. The precursor solution is prepared by combining PbI2, FAI, and CsI with 5% excess stoichiometry of PbI2 in a mixed solvent of DMF and DMSO [DMF (v) : DMSO (v) = 4:1]. The precursor solution is produced under 70 ?C heating for 1 hour. All solutions are prepared in a N2- filled glovebox with <2 ppm of O2 and H2O to ensure inert environment. The 1.0 M perovskite precursor solution is spin-coated onto fluorine-tin-oxide (FTO) coated substrate pre-heated to 65 ?C using a two-step spin-coating process. The first spin- coating step is performed at an acceleration rate of 1000 rpm s?1 for 10 s at 1000 rpm. 63 The second step acceleration is 2000 rpm s?1 for 20 seconds at 6000 rpm. An antisolvent, CB (250 ?L), is dripped onto the sample 5 seconds prior to the end of the second step. Following the spin-coating, the film is annealed for 10 minutes at 150 ?C. X-ray photoelectron spectroscopy: XPS measurements were taken at the UC Davis Advanced materials Charactrization and Testing Labotatory (AMCaT). Acquisition of the Axis Supra was funded by NSF-MRI 182838. The Axis Supra manufactured by Kratos Analytical is utilized for all XPS measurements. A monochromated Al K? radiation (1486.6 eV) at 7 mA emission current and an analysis chamber at a base pressure of 3.2 ?10?8 Torr is used. Measurements were acquired with the charge neutralizer with a filament current of 0.45 A and a filament bias of 1.2 V. The scan area is 450 ? 900 ?m. The angle between the incoming X-ray and the detector is 54.7 degrees. The angle of the photoelectron is 90 degrees to the sample. The spectra are calibrated with respect to the C 1s peak centered at 284.8 eV. Photoluminescence: A Princeton Instruments HRS-300 spectrometer equipped with a CCD camera detector was used for all measurements. A Vortron laser of wavelength 532 nm is used as the excitation source. A 7 mW input laser power is used, and after optical losses, the power of the laser incident on the sample is 4 mW. The laser reaches the sample through an optical set up. The emitted light from the sample passes through two lenses before entering the spectrometer. The first lens is a collimating lens that collects the emitted light. Next, a focusing lens directs the light into the entrance slit of the spectrometer. A 630 nm long pass filter is used at the entrance of the spectrometer to block stray reflected light from the laser. A 532 nm band pass filter (10 nm FWHM) is used to clean the excitation light. A grating, with groove density 300 g/mm, disperses the collected light to generate PL spectra. 64 Kelvin probe force microscopy: KPFM was performed using an Asylum Research Oxford Instruments MFP-3D infinity AFM using dual-pass amplitude modulated KPFM. In the first pass, the morphology of the sample is measured in dynamic mode. The second pass is run in nap-mode which records the contact potential difference. A Ti / Ir coated Si probe (f = 300 kHz, k = 42 N / m, radius of curvature 28 ? 10 nm) was used for all measurements. The scan parameters for the first pass are- set point of 205 mV, 0.6 Hz scan rate, 20 mV drive amplitude, 310 kHz drive frequency, and 0 V tip voltage. Scan parameters for the second pass are- 205 mV set point, 0.6 Hz scan rate, 1 V drive amplitude, 310 kHz drive frequency and 4 V tip voltage. The second pass lift height (?H) was 30 nm. The sample was grounded by connecting the FTO sample to the AFM scan plate. The sample was illuminated from the bottom using a broad light-emitting diode (P = 100 mW/cm2) with a peak centered ? 550 nm and measured under ambient conditions (23 ?C, 35% rH). The topography images were flattened and no other transformations were applied. 4.4 Results 4.4.1 Surface voltage under illumination cycle and inert environment Prior to subjecting the perovskite to humidity, we perform KPFM measurements on Cs0.33FA0.67PbI3 under an inert environment (5% rH). The sample surface voltage is measured under two illumination cycles of OFF-ON-OFF conditions to capture the stability of the material under light. The topography of the area mapped is shown in Figure 4.9 (0). Surface voltage maps acquired during the two illumination cycles are shown in Figure 4.9 (1-47). An average of each of the surface voltage maps is plotted in Figure 4.10. Observing 65 the trend, when light is turned ON, the voltage response increases and stabilizes at the fifth illuminated scan. When the light is turned OFF, the voltage decays and stabilizes at the fifth dark scan. In the second cycle, the illuminated voltage response stabilizes faster than in the first cycle. The dark scans post illumination show identical trends in both cycles. Comparing the initial dark voltage response to the surface voltage values under dark conditions post illumination, we note a 94% recovery in the sample after illumination cycling. 1750 1110 Figure 4.9: KPFM light OFF-ON-OFF cycle on Cs0.33FA0.67PbI3 under inert environment (< 5% rH), Scan (0) shows the topography of the area on which the KPFM scans were acquired. Scans (1-47) are the Light OFF-ON-OFF cycle. The arrows to the left of the scans show scan direction (top to bottom and vice versa) For the illuminated scans, a laser at 680 nm and power density of 100 mW/cm2 was used. Each scan took ? 7 min. Scan area is 6 ?m and scale bar is 1 ?m. 66 Surface voltage (mV) Figure 4.10: Stability test of surface voltage response under dark-and illuminated- KPFM. Two illumination cycles (light OFF-ON-OFF) performed on Cs0.33FA0.67PbI3. Average values extracted from KPFM scans shown in Figure 4.9 (1 - 47) vs. time. We perform in situ humidity-dependent dark and illuminated KPFMmeasurements to capture the nanoscale electrical behavior of Cs0.33FA0.67PbI3. Here, we observe the dynamic changes in surface voltage as humidity is cycled up from 5% rH to 65% rH in steps of 20% rH and cycled back down with the same step size. Figure 4.11 shows the real-time temperature and humidity during the experiment. The temperature remained constant at ? 23.8 ?C. Relative humidity varied from 5% ? 25% ? 45% ? 65% (UP cycle) and cycled back down from 65% ? 45% ? 25% ? 5% rH (DOWN cycle). Humidity control was well controlled during the 50+ hour experiment, with one humidity spike of 100% rH during the 45% DOWN cycle. The spike was brought back down in less than 4 minutes. Prior to starting the humidity cycle, the sample was kept in the chamber under inert environment (5% rH) for about 12 hours. After the humidity cycle, the sample was kept in the chamber for an additional 24 hours to capture recovery of the electrical response. 67 Figure 4.11: Real-time humidity and temperature profiles during the in situ humidity- dependent KPFM experiment. Humidity was maintained at 5% for ? 12 hours and then cycled up from 5% to 65% with steps of 20% and then cycled back down. Post cycling, the sample was kept at 5% for ? 24 hours. T = 23.8 ?C 4.4.2 Surface chemistry and structure We utilize XPS to track changes in surface chemistry and structure of Cs0.33FA0.67PbI3 thin film before and after humidity exposure. Here, we observe structure stability in the perovskite before and after humidity by exploring the chemical state of the elements. Figure 4.12 shows the XPS spectra for the I3d, Pb4f, and Cs3d core levels before and after the humidity cycle. The I3d spectra before the cycle (Figure 4.12 a) show two peaks at 630.5 eV and 619 eV and the I3d spectra after the humidity cycle (Figure 4.12 b) show two peaks at 630.3 eV and 618.8 eV. The positions of the remain the same; however, the intensity of the peaks is 30% stronger after the humidity cycle. The two peak values correspond to the 3d5/2 and 3d3/2 spin-orbital split. Before and after humidity cycling, the peak separation 68 value of 11.5 eV and the 1.5 peak intensity ratio of I3d5/2:I3d3/2 remains unchanged. These peak locations are for the ionic bonded I? as has been seen earlier. [131,132] The deconvoluted Pb4f spectra before and after exposure to the humidity cycle, respectively, are shown in Figure 4.12 c and d. The spin-orbital split for Pb4f is Pb4f7/2 and 4f5/2 which is attributed to the +4 and +2 ionic states of Pb, respectively. The peak locations of Pb4f7/2 are 138.2 eV and 138 eV before and after the humidity cycle. The peak location for Pb4f7/2 are 143.1 eV and 142.9 eV before and after humidity cycle, respectively. The peak intensity ratio under both situations is approximately 1.3. The peak separation is 4.9 eV for both. There is no significant structural variation based on Pb4f spectra, although the overall intensity of the peaks is 25% higher after the humidity cycle, suggesting a more dominant presence of Pb on the surface after the humidity cycle. Up to this point, based on the I3d and the Pb 4f spectra, the XPS data were pointing towards no structural changes and bonding of the elements in the perovskite structure. Strikingly, the Cs3d spectral profile before (Figure 4.12 e) and after humidity cycling (Figure 4.12 f) unveil a considerable change in the structural integrity of the perovskite after humidity cycling. Cs3d has two spin orbitals Cs3d5/2 and Cs3d3/2 which correspond to its ionic state in the bond formed with I?. First, we understand the chemical state of Cs in the pristine thin film. From the profile, the peak locations for Cs3d5/2 and Cs3d3/2 are 724.1 eV and 738 eV, respectively. In addition to these two expected peaks, we observe a third peak at 716.7 eV. This third peak has been previously seen in literature and can be attributed to an altered chemical state of Cs by change in its interaction with FA at the A-site and its electrostatic interlinkage with Iodide at the X-site . [108] It has also been identified previously, through NMR measurements, that Cs and FA form an alloy. [98] The 69 alloying of Cs and FA could lead to the third Cs3d peak at 716.7 eV. The modified chemical interactions of Cs post humidity cycling can be inferred from the disappearance of the third Cs3d peak at 716.4 eV. Interestingly, along with the disappearance of the third Cs3d peak, the intensity of the Cs3d5/2 and Cs3d3/2 peak after humidity cycling almost tripled (increased by 140%) at the surface. This large increase in the presence of Cs at the surface can be explained by the high mobility of the Cs ion due to its small ionic radius. The 1.4 intensity ratio of Cs3d5/2:Cs3d3/2 and the peak separation value of 13.8 eV remained the same before and after exposure to moisture. The crucial role of Cs in stabilizing the lattice stems from its small ionic radius and strong electronic interactions with halide ions. Therefore, the significant change in its chemical state before and after humidity shows the structural instability of Cs0.33FA0.67PbI3 under moisture. 70 Figure 4.12: Core spectra analysis of I3d, Pd4f and Cs3d. (a) I3d before humidity cycle. (b) I3d after humidity cycle. (c) Pb4f before humidity cycle. (d) Pb4f after humidity cycle. (e) Cs3d before humidity cycle (f) Cs3d before humidity cycle. FA has a chemical formula of [CH(NH2)2, CH3NH3]. The structure constitutes of the following carbon bonds : C-C, C-N, and C=N. Figures 4.13 a and b show the C1s spectra before and after the humidity cycle, respectively. In the pristine thin film, from right to left, the peak at 284.5 eV corresponds to the single C-C bond, the peak at 285.2 eV denotes a single C-N bond and the peak at 288.1 eV is the single bond C=N. These peaks positions are consistent with values previously seen in the literature. [133] After humidity 71 cycling, the C-C peak shifts to 285.1 eV and the overall peak intensity increases three times (150%), the C-N peak shifts to 286.8 eV, likely due to the emergence of the C-O bond and the C=N peak shifts to 288.5 eV due to the presence of C=O bond. These changes in the C1s peak can be directly related to the altered chemical state of FA under the presence of humidity (H2O). This further confirms moisture-induced surface chemistry modification and structural instability of Cs0.33FA0.67PbI3. The N1s spectra signifies the C-N and C=N bond in FA and its peak before (Figure 4.13 c) and after the humidity cycle (Figure 4.13 d) remains at ? 400.3 eV. The O1s spectra changes substantially before and humidity cycle. Observing the peaks in the profile in Figure 4.13 e, the overall peak is made of two peaks located at: 530.7 eV and 531.9 eV. The peak at 530.7 eV signifies surface bonded O2? and the peak at 531.9 eV corresponds to oxygen vacancies. [134?136] After the surface is exposed to humidity, the oxygen spectra shows bonded oxygen at the surface (Figure 4.13 f). The peak at 531.1 eV corresponds to C-O bond and the 532.7 eV peak corresponds to C-O-H. [137] 72 Figure 4.13: Core spectra analysis of C1s, N1s and O1s. (a) C1s before humidity cycle. (b) C1s after humidity cycle. (c) N1s before humidity cycle. (d) N1s after humidity cycle. (e) O1s before humidity cycle (f) O1s before humidity cycle. The overall analysis of the XPS data before and after the humidity cycle, indicate considerable changes in Cs0.33FA0.67PbI3 after the humidity cycle. While the material did not decompose, the surface chemistry variations suggest structure instability under humidity. To expand our exploration of the effects of humidity beyond the surface, we employ bulk-photoluminescence. Figure 4.14 shows the PL measurements acquired on the perovskite before and after the humidity cycle. Measurements were acquired under 73 identical conditions. We note that the peak location does not change. The PL intensity after humidity cycling decreases to half that of the pristine sample. This decrement in PL signal can be attributed to an increase in nonradiative recombination. Our XPS measurements indicated a higher concentration of elements on the surface post humidity cycle. The movement of ions to the surface creates charged defects in the bulk of the perovskite that foster recombination of charge carriers. In addition, the alterations seen in the surface chemistry after the humidity cycle be linked to creation of surface defects that further promote recombination of electrons and holes and reduce the photoactivity of the perovskite. Figure 4.14: Photoluminescence measurements before and after humidity cycle. Exposure time 2500 ms. A 532 nm laser was used as the excitation source (100 mW/cm2 power density). 74 4.4.3 Nanoscale surface voltage response The in situ humidity- and illumination- dependent KPFM on Cs0.33FA0.67PbI3 captures the dynamic local electrical behavior at length scales of grains and grain boundaries. Mapping both topography and surface voltage response simultaneously under humidity and illumination, provide insights into material and electrical stability. At every second of the humidity cycle from 5% ? 25% ? 45% ? 65% (UP cycle) and cycled back down from 65% ? 45% ? 25% ? 5% rH (DOWN cycle), spatial and functional information was mapped. At each humidity value during the UP and DOWN cycle, 15 maps were acquired (5 light OFF scans, 5 light ON scans, followed by 5 light OFF scans). Maps during the UP cycle: Figure 4.15 (1-15) and Figure 4.16 (1-15) show the topography and contact potential difference (CPD) maps, respectively, under 5% rH. Figure 4.15 (16-30) and Figure 4.16 (16-30) show the topography and CPD maps, respectively, under 25% rH. Figure 4.17 (31-45) and Figure 4.18 (31-45) show the topography and CPD maps, respectively, below 45% rH. Figure 4.17 (46-60) and Figure 4.18 (46-60) show the topography and CPD maps, respectively, under 65% rH. Maps during the DOWN cycle: Figure 4.19 (61-75) and Figure 4.20 (61-75) show the topography and CPD maps, respectively, under 45% rH. Figure 4.19 (76-90) and Figure 4.20 (76-90) show the topography and CPD maps, respectively, under 25% rH. Figure 4.21 (91-105) and Figure 4.22 (91-105) show the topography and CPD maps, respectively, under 5% rH. Figure 4.21 (106-120) and Figure 4.22 (106-120) show the topography and CPD maps, respectively, under 5% rH at approximately 12 hours post humidity cycle. Figure 4.23 (121-135) and Figure 4.24 (121-135) show the topography and CPD maps, respectively, 75 under 5% rH at approximately 24 hours post humidity cycle. From the topography images at each step of the humidity cycle and comparing the initial and final maps, we confirm that there is no degradation of the material at length scales of grains and grains boundaries. The CPD maps show the transient behavior of surface voltage response in the Cs0.33FA0.67PbI3 thin film as a function of both light and humidity. 76 Figure 4.15: Topography scans on Cs0.33FA0.67PbI3 from 5% to 25% rH transition reveal no degradation.(1-15) Topography maps at 5% rH. (16-30) topography maps at 25% rH. The arrows to the right of the scans show scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 77 25 % rH 5 % rH Figure 4.16: Dynamic surface voltage response on Cs0.33FA0.67PbI3 from 5% to 25% rH transition of the UP cycle. (1-15) Surface voltage scans at 5%. (16-30) Surface voltage scans at 25% rH reveal enhanced electrical response at 25% rH. The arrows to the right of the scans show scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. 78 25 % rH 5 % rH Figure 4.17: Topography scans on Cs0.33FA0.67PbI3 at 45% and 65% rH of the UP cycle reveal no degradation. (31-45) Topography scans at 45% rH. (46-60) Topography scans at 65% rH. The arrows to the right of the scans show scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 79 65 % rH 45 % rH Figure 4.18: Dynamic surface voltage response on Cs0.33FA0.67PbI3 at 45% and 65% rH of the UP cycle. (31-45) Surface voltage scans at 45% reveal enhanced electrical response. (46-60) Surface voltage scans at 65% rH reveal failure of electrical response at 65% rH. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. 80 65 % rH 45 % rH Figure 4.19: Topography scans on Cs0.33FA0.67PbI3 at 45% and 65% rH of the UP cycle reveal no degradation. (61-75) Topography scans at 45% rH. (76-90) Topography scans at 25% rH. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 81 25 % rH 45 % rH Figure 4.20: Dynamic surface voltage response on Cs0.33FA0.67PbI3 at 45% and 25% rH of the DOWN cycle. (61-75) Surface voltage scans at 45% reveal recovery of electrical response post failure at 65% rH. (76-90) Surface voltage scans at 25% rH uncover additional recovery of electrical response. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. 82 25 % rH 45 % rH Figure 4.21: Topography scans on Cs0.33FA0.67PbI3 at 5% rH at the end of the humidity cycle and at ? 12 hours post humidity cycle capture recovery at 5% rH. (91-105) Topography scans at 5% rH at the end of the humidity cycle reveal no material degradation. (106-120) Topography scans at ? 12 hours post humidity cycle. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 83 5 % rH (~12 hrs) 5 % rH Figure 4.22: Surface voltage response on Cs0.33FA0.67PbI3 at 5% rH at the end of the humidity cycle and at ? 12 hours post humidity cycle capture recovery at 5% rH (91- 105) Surface voltage maps at 5% rH at the end of the humidity cycle reveal no material degradation. (106-120) Surface voltage maps at ? 12 hours post humidity cycle. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. 84 5 % rH (~12 hrs) 5 % rH Figure 4.23: (121-135) Topography scans on Cs0.33FA0.67PbI3 at ? 24 hours post humidity cycle recovery at 5% rH. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 85 5 % rH (~24 hrs) Figure 4.24: (121-135) Surface voltage maps on Cs0.33FA0.67PbI3 at ? 24 hours post humidity cycle recovery at 5% rH. The arrows to the right of the scans indicate scan direction. Each scan was ? 7 min. Scale bar 1 ?m. At each humidity value, an illumination cycle of OFF-ON-OFF is indicated by the light bulb next to the scans. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. The raster scan mechanism, used to capture the nanoscale maps of topography and surface voltage, causes drift. This drift is caused by interactive forces (such as Van der Waals and Coulomb force) between the probe and the sample surface. Figure 4.25 shows a spatial comparison between the first scan (1) and the last scan (105). The scan area for all maps is 6 ?m ? 6 ?m. We note that the sample drifts approximately 0.8 ?m in the -x direction. The common area between the two maps is highlighted by a dashed red rectangle. 86 5 % rH (~24 hrs) Figure 4.25: Drift analysis on topography and surface voltage maps. ? 0.8 ?m drift in the -x direction is observed by comparing scan areas before and after humidity cycles. The overlapping areas are highlighted with a dashed-rectangle. Total scan area is 6 ?m ? 6 ?m. Average values extracted from the surface voltage scans in Figure 4.16 (1-30), Figure 4.18 (31-60), Figure 4.20 (61-90) and Figure 4.22 (91-105) are plotted in Figure 4.26. Comparing the average value of the last illuminated map at each humidity value consecutively, we quantify the following changes in surface voltage behavior during the UP cycle: from 5 ? 25% rH the voltage response increases by 11.4%; from 25 ? 45% rH the voltage response is further enhanced and increases by 17.9%; from 45 ? 65% rH electrical failure is observed based on the 67% decrease in voltage response. Similarly, during the down cycle: from 65? 45% rH the voltage response recovers and increases 17.6%; from 45 ? 25% rH an additional increase of 7.4% is seen; from 25 ? 5% rH a decrease of 13.9% is noted. The surface voltage behavior is unstable under humidity due to the chemical and structural instability of the perovskite as seen from the XPS measurements. We speculate that the enhanced voltage response upto 45% is due moisture-induced passivation of defects [138] and the electrical failure at 65% rH is due to changes in the chemical states 87 of Cs and FA in the perovskite structure. Figure 4.26: Dynamic humidity-dependent surface voltage response on Cs0.33FA0.67PbI3. Relative humidity is cycled from 5% ? 25% ? 45% ? 65% (UP cycle) and cycled back down from 65% ? 45% ? 25% ? 5% rH (DOWN cycle). The averages are extracted from the surface voltage images shown in Figure 4.16 (1-30), 4.18 (31-60), 4.20 (61-90), 4.22 (91-105), respectively. 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. We capture recovery of the electrical response in Cs0.33FA0.67PbI3 up to approximately 24 hours post humidity cycling (Figure 4.27). The sample was kept in the humidity sensing enclosure under an inert environment (5% rH) with the probe engaged to ensure the area was not lost. The illumination cycle performed at the beginning of the humidity cycle at 5% rH was replicated at the end of the humidity cycle, at 12 hours post humidity and at 24 hours post humidity cycle. Comparing the average values from the last illuminated scan from each time stamp to the average value of the last illuminated scan at the beginning of the humidity cycle, we observe a 94.34% recovery at the end of the humidity cycle, a 98% recovery 12 hours post humidity cycling and a 99.4% recovery 24 hours post humidity 88 cycling. Recovery in perovskites after degradation due to extrinsic factors has been seen in the field and has also been termed as self-healing. [139?141] This is a property that is unique to the perovskite material, showing promise toward photovoltaic applications. Figure 4.27: Electrical response recovery post humidity cycle on Cs0.33FA0.67PbI3. The light OFF-ON-OFF cycle at 5% rH before the humidity cycle, at the end of the humidity cycle, ? 12 hours post humidity cycle and ? 12 hours post humidity cycle are shown. The average values are extracted from images shown in Figure 4.16 (1-15), 4.22 (91-120) and 4.24 (121-135). 1 sun illumination from a light source centered at 550 nm, 100 mW2 power density. 4.4.4 Humidity- and illumination-dependent surface voltage behavior To understand the post illumination behavior under humidity, we resolve illuminated and post illumination surface voltage maps as a function of time as shown in Figure 4.28. Each scan has 128 ? 128 pixel density. Therefore, for each scan, we have a data frame of 16384 voltage values. It takes ? 7.2 min/432 seconds to measure each scan. The system measures ? 38 points in 1 second. We collapse the voltage data frame to 432 points by 89 averaging every 38 points to correspond each point to 1 second in time. For all humidity values, under the light OFF-ON-OFF cycle, 5 scans were imaged under each condition. From the 5 illuminated scans imaged, the last two illuminated scans and the following 2 dark scans post-illumination were used to analyze the voltage behavior under humidity as a function of time. In an inert environment (5% rH) shown in Figure 4.28 a, the voltage behavior under illumination is stable. When the light is turned OFF, the voltage response decays as a second-order polynomial function of time (Equation 4.1). V5% = (1.25? 10?4)t2 ? 0.38t? 457.22 (4.1) Where V5% is the surface voltage (mV) and t is time (seconds). At 25% rH (Figure 4.28 b) the illumination response is unstable and decreases over time. This decrease in voltage response can be attributed to structural instability under humidity that causes an increase in the number of surface defect sites and nonradiative recombination, which can greatly affect the photovoltaic performance of the perovskite. [142] When the light is turned OFF, the function of voltage decay as a function of time is given by Equation 4.2. At 45% rH (Figure 4.28 c) we see a similar trend under light ON and OFF as seen under 25% rH. Equation 4.3 shows the post-illumination voltage response at 45% rH as a function of time. V ?4 225% = (3.08? 10 )t ? 0.93t? 93.37 (4.2) 90 V45% = (1.55? 10?4)t2 ? 0.53t? 257.29 (4.3) At 65% rH, we observed voltage response failure that we ascribe to changes in the perovskite structure based on the XPS measurements. Figure 4.28 d shows the dynamic voltage response at 65% rH under light ON and OFF conditions as a function of time. Under illumination, we observe a sharp drop followed by an increase in the voltage response. Once the light is turned OFF, the voltage decays initially and then starts to increase and reaches values similar to those of illuminated scans (Equation 4.4). Once the DOWN cycle is started and the humidity value is brought down from 65% to 45% rH, the voltage response starts to recover. The upward recovery can be seen in Figure 4.28 e. Under light ON condition, the voltage response increases with time, and even after the light is turned off (Equation 4.5), the voltage response continues to increase as the material recovers. V65% = (4.82? 10?4)t2 ? 1.21t? 113.41 (4.4) V = (3.19? 10?445% )t2 ? 0.76t? 271.11 (4.5) Continuing the DOWN cycle, the voltage response trend at 25% rH under light ON and OFF conditions (Figure 4.28 f) is similar to that seen during the UP cycle at 25% rH. The voltage decay as a function of time at 25% rH is shown in Equation 4.6. Upon ending the cycle at 5% rH, the voltage response under illumination is constant, similar to the response observed before the humidity cycle. Interestingly, the voltage decay seems to be instantaneous, displaying another dynamic behavior. Though after 24 hours under 5% 91 rH the voltage response under the light ON-OFF condition is nearly identical to that seen at the beginning. The voltage decay post-illumination as a function of time for 5% rH at the end of the humidity cycle and after 24 hours under 5% rH is given by Equations 4.7 and 4.8 respectively. V25% = (2.86? 10?4)t2 ? 0.84t? 141.24 (4.6) V 25% = (1.66)t ? 0.05t? 753.17 (4.7) V5% = (1.55? 10?4)t2 ? 0.46t? 435.97 (4.8) 4.5 Discussion The in-situ humidity dependent KPFM on Cs0.33FA0.67PbI3 reveals electrical instability in the perovskite. XPS analysis on the film before and after exposure to humidity cycle, uncovers migration of ions to the surface, structural changes and and altered chemical states in Cs and FA. An increase in migration of ions is observed post humidity cycle by the increase in peak intensities of the I3d, Pb4f, and Cs3d core level spectra profiles. The movement of ions within the perovskite causes charged defects; especially, iodide vacancies have been seen to play a role in device performance losses. [143] In addition to charged defects, surface defects are also expected based on the surface chemistry changes observed through the C1s and O1s spectra. The significantly lower PL signal observed post humidity 92 93 Figure 4.28: Humidity-dependent post-illumination electrical behavior of Cs0.33FA0.67PbI3 (a) at 5% rH, fitted curve. cycle points towards nonradiative charge carrier recombination at charged defects in the bulk and at surface defects. Since both humidity and light were varied during the in situ measurements, the possibility of photo-induced changes was ruled out based on the high photo-stability of the sample seen during two illumination cycle (OFF-ON-OFF)- KPFM experiment performed prior to the humidity cycle (Figure 4.10). From the combination of these measurements, it can be inferred that exposure to humidity up to 65% induces ion migration, the formation of charged and surface defects, nonradiative recombination, structural and electrical instability in Cs0.33FA0.67PbI3. There are many types of defects in metal halide perovskites (intrinsic and extrinsic). Defects are highly detrimental to the photovoltaic performance and their localization can be helpful in material optimization. To expand this work and capture the moisture-induced local electrical behavior, pc-AFM can be employed, and the in situ behavior of the photocurrent can be mapped to provide insight into the photovoltaic performance of the perovskite under humidity. 4.6 Conclusions In summary, we have captured for the first time, in situ humidity-dependent local surface voltage response in Cs0.33FA0.67PbI3. We observe dynamic changes in electrical response as a function of humidity and light. During the humidity UP cycle, from 5 - 65% rH, we see an increase in voltage up to 45% rH likely due to moisture-induced passivation of defects. At 65% rH, the voltage drops by 67% indicating electrical failure. We observe recovery from this failure as the voltage response increases during the DOWN 94 cycle from 65 - 5% rH. We note that the voltage response recovers 94.34% at the end of the humidity cycle. Further, voltage scans at ? 12 and 24 hours post humidity cycling show 98% and 99.4% recovery, respectively. From the topography maps during the UP and DOWN cycle, we conclude no degradation. We observe a third peak for Cs3d in the pristine film which disappears after the humidity cycle. This peak is attributed to electrostatic interactions between Cs, FA, and halide ions that help to stabilize the lattice. Furthermore, the altered chemical state of C1s and O1s peaks post humidity cycling indicate C-O and C-O-H bonds at the surface. Changes in C1s also suggest changes in FA interactions. Lower PL intensity after the humidity cycle indicate a higher nonradiative recombination. Our results show the real-time moisture-induced dynamic electrical response that stem from structural instabilities in Cs0.33FA0.67PbI3. 95 Chapter 5: Conclusion and Outlook In conclusion, we probe the optoelectronic stability at the nanoscale of exemplar metal halide perovskite compositions pertinent to photovoltaic applications. We identified superior photovoltaic performance in Rb-based perovskite by performing comparative electrical measurements on dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite under an inert environment. For the first time, we performed in situ humidity-dependent KPFM measurements on metal halide perovskites and captured moisture-induced dynamics in the electrical response at the length scales of grains and grain boundaries in Cs0.33FA0.67PbI3. We investigated the effect of humidity on the structure, surface chemistry and bulk of the perovskite through XPS and PL measurements before and after the in situ humidity cycling to provide insight into moisture-induced degradation, passivation, ion migration and structural changes. First, we performed a comprehensive electrical analysis (voltage and current) at the nanoscale on dual-cation (Cs0.17FA0.83Pb(I0.83Br0.17)3) and quad-cation (Cs0.07Rb0.03FA0.76MA0.14Pb(I0.85Br0.15)3) perovskite to understand the effects of A-site cations on the photovoltaic properties of the perovskite. Mixed-cation and mixed-halide perovskites have been shown to impart stability to perovskites, which is consistent with our observations. We utilize the electrical capabilities of the AFM and probe the 96 surface voltage and photocurrent spatially through KPFM and pc-AFM, respectively, under an inert environment (5% rH) to eliminate effects of environmental stressors on electrical performance. Macroscopic J-V measurements on full devices for both chemical compositions revealed ? 90% reduced hysteresis in quad-cation perovskite. We accredit the reduced hysteresis to an increase in halide homogeneity and a decrease in ion migration in the structure due to Rb+ cation at the A-site. Although, under reverse-bias conditions, we note comparable maximum power conversion efficiency (PCE) of 20.01% and 20.81% for the dual- and quad-cation, respectively. To understand this global device behavior, we step down to the material level and scrutinize the electrical characteristics at the mesoscale. With KPFM, we measure the contact potential difference (CPD) between the probe and the sample and find a 34% higher surface voltage homogeneity in the quad-cation perovskite. A Dark-Light-Dark series of KPFM measurements on both samples revealed a 55% greater voltage decay in quad-cation perovskite. We infer that the suppression of post-illumination residual voltage is due to reduced ion migration, and interfacial recombination in the quad-cation perovskite due to the incorporation of of Rb+. Dark KPFM maps revealed higher voltage values at the grain boundaries compared to their adjacent grains indicating upward band bending at grain boundaries for both chemical compositions. Upward band bending creates an electron barrier which leads to an accumulation of holes. Next, we utilize pc-AFM to capture the photocurrent response at grains and grain boundaries for both dual- and quad-cation perovskite. On the basis of the energy band diagram of the stacked layers in the sample and the material of the probe used for pc-AFM measurements, we establish that the charge carriers collected as photocurrent are holes. From our KPFM measurements, we expected to see highly photoconductive 97 grain boundaries due to the accumulation of holes. From spatially resolved photocurrent maps, surprisingly, we observe that majority of grain boundaries in the dual-cation perovskite are photo-inactive (displaying values of photocurrent similar to dark current values measured prior to photoexcitation). On the other hand, the photocurrent response at grain boundaries in quad-cation perovskite was in agreement with the upward band bending observed through CPD maps, with the majority of grain boundaries being highly photo-active (displaying values of photocurrent higher than their adjacent grains). We quantified a 50% higher overall photocurrent response in quad-cation perovskite likely due to the highly conductive network of grain boundaries. We deduce from the combined knowledge of the KPFM and pc-AFM response, defect-assisted recombination at grain boundaries in dual-cation perovskite, and defect-passivation at grain boundaries by Rb+ cations. After measuring the electrical response under an inert environment, we shifted focus and explored humidity-induced effects on the electrical characteristics of metal halide perovskites. We cycle humidity up from 5 - 65% rH and down from 65 - 5% rH in steps of 20% rH and measure surface voltage and topography maps in situ. Along with varying humidity, we perform an illumination cycle of (OFF-ON-OFF) acquiring 15 maps (5 scans at each light condition) for each humidity value. We observe an elevated voltage response upto 45% rH which we attribute to moisture-induced passivation of the defects. Once the humidity increases to 65% rH, a 67% drop in the voltage response points to electrical failure of the material. We note increase in the electrical response once the humidity is brought back down to 45% rH and further increase at 25%. The voltage response at 5% rH at the end of the humidity cycle shows a 94.34% recovery. We also measured voltage scans at 98 12 and 24 hours post-humidty cycle and quantified remarkable recovery values of 98% and 99.4%, respectively at 5% rH. To understand the dynamic electrical response seen during the in situ measurements at the structure level, we performed XPS measurements on the sample before and after the humidity cycle. The location of the I3d and Pb4f peaks did not change, however, the intensity of the peaks increased by 30% and 25%, respectively, suggesting higher concentrations of these ion at the surface post-humidity cycle. Notable changes were seen in the Cs3d spectra. The third Cs3d peak, which denotes the interactions between Cs, FA, and halide ions, disappeared after the humidity cycle. The presence of Cs ions on the surface increased three folds based on the peak intensities before and after exposure to high humidity. Altered C1s and O1s peaks suggested the presence of C-O and C=O at the surface after humidity cycling. The N1s peak indicating the C-N and C=N bonds of FA did not show change. To explore changes in the bulk of the material, PL measurements were taken before and after the in situ experiment. The PL peak did not shift, however, the intensity of the peak decreased by 50%. We infer nonradiative recombination due to charged defects in the bulk and surface defects. Overall, our results provide insight into the self-recovery and real-time moisture-based instabilities in the electrical behavior of the perovskite. Future work will use in situ humidity-dependent pc-AFM measurements to map photocurrent behavior and expand the knowledge of localized moisture-induced electrical response to a relationship between material and device performance. In addition to humidity, another environmental stressor that needs to be better understood at the nanoscale is temperature. Temperature elevations during operating conditions can reach 99 up to 45 - 60 oC above room temperature as a result of extended illumination-induced thermal-losses and poor thermal conductivity of the material. Therefore, it is imperative that the next-generation perovskites be thermally stable. There is limited information in literature on the effects of temperature on electrical properties of PSCs and the studies that exist are performed under ambient conditions and fail to account for the effects of relative humidity and oxygen with the variation of temperature. Nitrogen Purge Line 1 Sun Top Illumination AFM Controller (c) Photocurrent mp OutN Pu I T R O Coolant Reservoirs G Temperature E (d()d)(d)N Modula.8 ting Stage -3-.-383.8 -3.9 kite AAF MFTM TipipTip -3.-93.9 PPeerovs F M Perovrsokvistekite A -5.0 -4.2 -5.0 -5.0 -4-4 -4.2 .5.5 -4.2 --5.4-5.45.4 Environmental Controller -4.5 TTiOiO2 --2 55..99 FFTTOO TiO2 -5.9 FTO -7.- - 27 7..22 Figure 5.1: Experimental setup for 25 - 85 oC in situ temperature-dependent KPFM and pc-AFM under inert environment. Utilizing the controlled humidity set up to maintain an inert environment, local temperature-driven morphological and electrical changes will be measured. Figure 5.1 shows a schematic for the experimental setup which includes the AFM controller to control the scanning probe, a programmable environmental controller to eliminate human error, two coolant reservoirs, and an enclosed nitrogen-purged temperature stage. Due to temperature stage limitations, the device will need to be illuminated from the top. The target temperature range for this experiment would be 25 - 85 oC based on the operating temperature conditions of the PSCs. 100 Energy Level (eV) Energy Level (eV) Energy Level (eV) PCBM PPCCBBMM Appendix: A Products of This Research A.1 Awards and Honors 1. UMD Goldhaber Travel Award 09/2020 2. UMD McAvory Family Fellowship 2018 - 2019 A.2 Publications 1. R. Lahoti, E. Deniz, S. Kim, J.-P. Correa-Baena, M.S. Leite. ?In situ humidity- dependent KPFM on Cs0.33FA0.67PbI3?. - In preparation. (expected 2022) 2. R. Lahoti, J.M. Howard, M.A. Mahmud, T. White, M.S. Leite. ?Rb-incorporated perovskite solar cells deliver superior nanoscale electrical response?. - In Review (expected 2022) 3. J.M. Howard, R. Lahoti, M.S. Leite. ?Imaging metal halide perovskites material and device properties at the nanoscale.? Adv. Energy Mater., 10.1002/aenm.201903161, 2019, 10 (26), 1903161. - Front Cover 101 A.3 Presentations 1. R. Lahoti, J.M. Howard, et.al. ?Imaging the exceptional electrical performance of Rb-based halide perovskites at the nanoscale? SPIE Photonics West, San Francisco, CA (01/2022) - Talk 2. R. Lahoti, J.M. Howard, M.A. Mahmud, T. White, M.S. Leite. ?Nonpareil electrical performance of Rb-based halide perovskite at the nanoscale? MRS Fall Meeting, Boston, MA (09/2021) - Poster 3. R. Lahoti, J.M. Howard, M.A. Mahmud, T. White, M.S. Leite. ?Mapping Rb- Perovskite Photovoltaics: Photocurrent and Photovoltage at the Nanoscale? APS March Meeting, Virtual (03/2021) - Talk 102 Bibliography [1] A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka. Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J Am Chem Soc, 131(17):6050?1, 2009. [2] National renewable energy laboratory, 2022. [3] W. A. Dunlap-Shohl, Y. Zhou, N. P. Padture, and D. B. Mitzi. Synthetic approaches for halide perovskite thin films. Chem Rev, 119(5):3193?3295, 2019. [4] Qing Liao, Xue Jin, and Hongbing Fu. Tunable halide perovskites for miniaturized solid-state laser applications. Advanced Optical Materials, 7(17):1900099, 2019. [5] H. J. Feng, T. R. Paudel, E. Y. Tsymbal, and X. C. Zeng. Tunable optical properties and charge separation in ch3nh3sn(x)pb(1-x)i3/tio2-based planar perovskites cells. J Am Chem Soc, 137(25):8227?36, 2015. [6] Xianxiong He, Peng Liu, Haihua Zhang, Qing Liao, Jiannian Yao, and Hongbing Fu. Patterning multicolored microdisk laser arrays of cesium lead halide perovskite. Advanced Materials, 29(12):1604510, 2017. [7] Chonghea Li, Xionggang Lu, Weizhong Ding, Liming Feng, Yonghui Gao, and Ziming Guo. Formability of abx3 (x = f, cl, br, i) halide perovskites. Acta Crystallographica Section B, 64(6):702?707, 2008. [8] The ABX3 Perovskite Structure, chapter 1, pages 1?41. John Wiley Sons, Ltd, 2016. [9] Hyun Suk Jung, Gill Sang Han, Nam-Gyu Park, and Min Jae Ko. Flexible perovskite solar cells. Joule, 3(8):1850?1880, 2019. [10] Y. Li, L. Meng, Y. M. Yang, G. Xu, Z. Hong, Q. Chen, J. You, G. Li, Y. Yang, and Y. Li. High-efficiency robust perovskite solar cells on ultrathin flexible substrates. Nat Commun, 7:10214, 2016. 103 [11] X. Meng, Z. Cai, Y. Zhang, X. Hu, Z. Xing, Z. Huang, Z. Huang, Y. Cui, T. Hu, M. Su, X. Liao, L. Zhang, F. Wang, Y. Song, and Y. Chen. Bio-inspired vertebral design for scalable and flexible perovskite solar cells. Nat Commun, 11(1):3016, 2020. [12] Pachaiyappan Murugan, Ting Hu, Xiaotian Hu, and Yiwang Chen. Current development toward commercialization of metal-halide perovskite photovoltaics. Advanced Optical Materials, 9(17):2100390, 2021. [13] S. Tan, T. Huang, I. Yavuz, R. Wang, T. W. Yoon, M. Xu, Q. Xing, K. Park, D. K. Lee, C. H. Chen, R. Zheng, T. Yoon, Y. Zhao, H. C. Wang, D. Meng, J. Xue, Y. J. Song, X. Pan, N. G. Park, J. W. Lee, and Y. Yang. Stability-limiting heterointerfaces of perovskite photovoltaics. Nature, 2022. [14] Tomas Leijtens, Giles E. Eperon, Nakita K. Noel, Severin N. Habisreutinger, Annamaria Petrozza, and Henry J. Snaith. Stability of metal halide perovskite solar cells. Advanced Energy Materials, 5(20):1500963, 2015. [15] Yingping Fan, Hongguang Meng, Li Wang, and Shuping Pang. Review of stability enhancement for formamidinium-based perovskites. Solar RRL, 3(9):1900215, 2019. [16] Zhaoning Song, Chad L. McElvany, Adam B. Phillips, Ilke Celik, Patrick W. Krantz, Suneth C. Watthage, Geethika K. Liyanage, Defne Apul, and Michael J. Heben. A technoeconomic analysis of perovskite solar module manufacturing with low-cost materials and techniques. Energy Environ. Sci., 10:1297?1305, 2017. [17] Ioannis Spanopoulos, Weijun Ke, and Mercouri G. Kanatzidis. In quest of environmentally stable perovskite solar cells: A perspective. Helvetica Chimica Acta, 104(1):e2000173, 2021. [18] Rui Wang, Muhammad Mujahid, Yu Duan, Zhao-Kui Wang, Jingjing Xue, and Yang Yang. A review of perovskites solar cell stability. Advanced Functional Materials, 29(47):1808843, 2019. [19] Hong-Hua Fang, Sampson Adjokatse, Haotong Wei, Jie Yang, Graeme R. Blake, Jinsong Huang, Jacky Even, and Maria Antonietta Loi. Ultrahigh sensitivity of methylammonium lead tribromide perovskite single crystals to environmental gases. Science Advances, 2(7), 2016. [20] Tomas Leijtens, Kevin Bush, Rongrong Cheacharoen, Rachel Beal, Andrea Bowring, and Michael D. McGehee. Towards enabling stable lead halide perovskite solar cells; interplay between structural, environmental, and thermal stability. J. Mater. Chem. A, 5(23):11483?11500, 2017. [21] Alessandro Senocrate, Tolga Acart?rk, Gee Yeong Kim, Rotraut Merkle, Ulrich Starke, Michael Gr?tzel, and Joachim Maier. Interaction of oxygen with halide perovskites. J. Mater. Chem. A, 6:10847?10855, 2018. 104 [22] M. Anaya, J. F. Galisteo-Lopez, M. E. Calvo, J. P. Espinos, and H. Miguez. Origin of light-induced photophysical effects in organic metal halide perovskites in the presence of oxygen. J Phys Chem Lett, 9(14):3891?3896, 2018. [23] N. Aristidou, C. Eames, I. Sanchez-Molina, X. Bu, J. Kosco, M. S. Islam, and S. A. Haque. Fast oxygen diffusion and iodide defects mediate oxygen-induced degradation of perovskite solar cells. Nature Communications, 8:15218, 2017. [24] Shangjun Cheng and Haizheng Zhong. What happens when halide perovskites meet with water? The journal of physical chemistry letters, pages 2281?2290, 2022. [25] Un-Gi Jong, Chol-Jun Yu, Gum-Chol Ri, Andrew P. McMahon, Nicholas M Harrison, Piers R. F. Barnes, and Aron Walsh. Influence of water intercalation and hydration on chemical decomposition and ion transport in methylammonium lead halide perovskites. Journal of Materials Chemistry A, 6(3):1067?1074, 2018. [26] Yun-Hyok Kye, Chol-Jun Yu, Un-Gi Jong, Yue Chen, and Aron Walsh. Critical role of water in defect aggregation and chemical degradation of perovskite solar cells. The journal of physical chemistry letters, 9(9):2196?2201, 2018. [27] M. Prete, M. V. Khenkin, D. Glowienka, B. R. Patil, J. S. Lissau, I. Dogan, J. L. Hansen, T. Lei?ner, J. Fiutowski, H.-G. Rubahn, B. Julsgaard, P. Balling, V. Turkovic, Y. Galagan, E. A. Katz, and M. Madsen. Bias-dependent dynamics of degradation and recovery in perovskite solar cells. ACS Applied Energy Materials, 4(7):6562?6573, 2021. [28] Yong Hui, Yan-Yan Tan, Liang Chen, Zi-Ang Nan, Jian-Zhang Zhou, Jia-Wei Yan, and Bing-Wei Mao. Stability of perovskite thin films under working condition: Bias- dependent degradation and grain boundary effects. Advanced Functional Materials, 31(36):2103894, 2021. [29] Mark V. Khenkin, Anoop K. M., Eugene A. Katz, and Iris Visoly-Fisher. Bias- dependent degradation of various solar cells: lessons for stability of perovskite photovoltaics. Energy Environ. Sci., 12:550?558, 2019. [30] Wolfgang Tress, Konrad Domanski, Brian Carlsen, Anand Agarwalla, Essa A Alharbi, Michael Gr?tzel, and Anders Hagfeldt. Performance of perovskite solar cells under simulated temperature-illumination real-world operating conditions. Nature Energy, 4(7):568?574, 2019. [31] Shuai Ruan, Maciej-Adam Surmiak, Yinlan Ruan, David P. McMeekin, Heike Ebendorff-Heidepriem, Yi-Bing Cheng, Jianfeng Lu, and Christopher R. McNeill. Light induced degradation in mixed-halide perovskites. J. Mater. Chem. C, 7:9326? 9334, 2019. [32] Xiaolian Liu, Qiang Han, Yufei Liu, Chengyi Xie, Chenggang Yang, Dongmei Niu, Youzhen Li, Huan You Wang, Lixin Xia, Yongbo Yuan, and Yongli Gao. Light- induced degradation and self-healing inside ch3nh3pbi3-based solar cells. Applied Physics Letters, 116:253303, 2020. 105 [33] G. Divitini, S. Cacovich, F. Matteocci, L. Cina, A. Di Carlo, and C. Ducati. In situ observation of heat-induced degradation of perovskite solar cells. Nature Energy, 1(2):15012, 2016. [34] M. S. Tumusange, B. Subedi, C. Chen, M. M. Junda, Z. Song, Y. Yan, and N. J. Podraza. Impact of humidity and temperature on the stability of the optical properties and structure of mapbi3, ma0.7fa0.3pbi3 and (fapbi3)0.95(mapbbr3)0.05 perovskite thin films. Materials (Basel), 14(14), 2021. [35] Lin Ma, Deqiang Guo, Mengting Li, Cheng Wang, Zilin Zhou, Xin Zhao, Fangteng Zhang, Zhimin Ao, and Zhaogang Nie. Temperature-dependent thermal decomposition pathway of organic?inorganic halide perovskite materials. Chemistry of Materials, 31(20):8515?8522, 2019. [36] Le Wang, Haomin Wang, Martin Wagner, Yong Yan, Devon S. Jakob, and Xiaoji G. Xu. Nanoscale simultaneous chemical and mechanical imaging via peak force infrared microscopy. Science Advances, 3(6):e1700255, 2017. [37] Nina Balke, Petro Maksymovych, Stephen Jesse, Andreas Herklotz, Alexander Tselev, Chang-Beom Eom, Ivan I Kravchenko, Pu Yu, and Sergei V Kalinin. Differentiating ferroelectric and nonferroelectric electromechanical effects with scanning probe microscopy. ACS Nano, 9(6):6484?6492, 2015. [38] A. D. Schulz, H. Rohm, T. Leonhard, S. Wagner, M. J. Hoffmann, and A. Colsmann. On the ferroelectricity of CH3NH3PbI3 perovskites. Nature Materials, 18(10):1050, 2019. [39] Ilka M. Hermes, Simon A. Bretschneider, Victor W. Bergmann, Dan Li, Alexander Klasen, Julian Mars, Wolfgang Tremel, Fr?d?ric Laquai, Hans-J?rgen Butt, Markus Mezger, R?diger Berger, Brian J. Rodriguez, and Stefan A. L. Weber. Ferroelastic fingerprints in methylammonium lead iodide perovskite. The Journal of Physical Chemistry C, 120(10):5724?5731, 2016. [40] Y. Liu, L. Collins, R. Proksch, S. Kim, B. R. Watson, B. Doughty, T. R. Calhoun, M. Ahmadi, A. V. Ievlev, S. Jesse, S. T. Retterer, A. Belianinov, K. Xiao, J. Huang, B. G. Sumpter, S. V. Kalinin, B. Hu, and O. S. Ovchinnikova. Chemical nature of ferroelastic twin domains in CH3NH3PbI3 perovskite. Nature Materials, 17(11):1013? 1019, 2018. [41] Sawanta S. Mali, Hyungjin Kim, Jyoti V. Patil, and Chang Kook Hong. Bio- inspired carbon hole transporting layer derived from aloe vera plant for cost-effective fully printable mesoscopic carbon perovskite solar cells. ACS Applied Materials & Interfaces, 10(37):31280?31290, 2018. PMID: 30130386. [42] Elizabeth M. Tennyson, John M. Howard, Bart Roose, Joseph L. Garrett, Jeremy N. Munday, Antonio Abate, and Marina S. Leite. The effects of incident photon energy on the time-dependent voltage response of lead halide perovskites. Chemistry of Materials, 31(21):8969?8976, 2019. 106 [43] Bert Conings, Jeroen Drijkoningen, Nicolas Gauquelin, Aslihan Babayigit, Jan D?Haen, Lien D?Olieslaeger, Anitha Ethirajan, Jo Verbeeck, Jean Manca, Edoardo Mosconi, Filippo De Angelis, and Hans-Gerd Boyen. Intrinsic thermal instability of methylammonium lead trihalide perovskite. Advanced Energy Materials, 5(15):1500477, 2015. [44] Shaoyang Liu and Yifen Wang. Application of afm in microbiology: a review. Scanning, 32(2):61?73, 2010. [45] Phuong Nguyen-Tri, Payman Ghassemi, Pascal Carriere, Sonil Nanda, Aymen Amine Assadi, and Dinh Duc Nguyen. Recent applications of advanced atomic force microscopy in polymer science: A review. Polymers, 12(5), 2020. [46] John M. Howard, Richa Lahoti, and Marina S. Leite. Imaging metal halide perovskites material and properties at the nanoscale. Advanced Energy Materials, page 10.1002/aenm.201903161, 2019. [47] Xiangying Deng, Fang Xiong, Xiayu Li, Bo Xiang, Zheng Li, Xu Wu, Can Guo, Xiaoling Li, Yong Li, Guiyuan Li, Wei Xiong, and Zhaoyang Zeng. Application of atomic force microscopy in cancer research. Journal of Nanobiotechnology, 16(1):102, 2018. [48] Gaurav Pandey, Maithri Tharmavaram, Deepak Rawtani, Sumit Kumar, and Y. Agrawal. Multifarious applications of atomic force microscopy in forensic science investigations. Forensic Science International, 273:53?63, 2017. [49] Leo Polak and Rinke J. Wijngaarden. Quantitative Analysis of Kelvin Probe Force Microscopy on Semiconductors, book section Chapter 9, pages 227?247. Springer Series in Surface Sciences. 2018. [50] Xue Zheng, Xiao Wang, Weimin Li, Zhenghao Liu, Wenjie Ming, Huan Wang, Han Wang, Da Li, Bo Liu, and Chunlei Yang. Correlations of ionic migration and deep- level traps leads to surface defect formation in perovskite solar cells. The Journal of Physical Chemistry C, 125(35):19551?19559, 2021. [51] Yao Zhao, WenKe Zhou, Zhengyuan Han, Dapeng Yu, and Qing Zhao. Effects of ion migration and improvement strategies for the operational stability of perovskite solar cells. Phys. Chem. Chem. Phys., 23:94?106, 2021. [52] J. L. Garrett, E. M. Tennyson, M. Hu, J. Huang, J. N. Munday, and M. S. Leite. Real-time nanoscale open-circuit voltage dynamics of perovskite solar cells. Nano Letters, 17(4):2554?2560, 2017. [53] David Toth, Bekele Hailegnaw, Filipe Richheimer, Fernando A. Castro, Ferry Kienberger, Markus C. Scharber, Sebastian Wood, and Georg Gramse. Nanoscale charge accumulation and its effect on carrier dynamics in tri-cation perovskite structures. ACS Applied Materials & Interfaces, 12(42):48057?48066, 2020. PMID: 32969644. 107 [54] Jingjing Chang, Juanxiu Xiao, Zhenhua Lin, Hai Zhu, Qing-Hua Xu, Kaiyang Zeng, Yue Hao, and Jianyong Ouyang. Elucidating the charge carrier transport and extraction in planar heterojunction perovskite solar cells by kelvin probe force microscopy. J. Mater. Chem. A, 4:17464?17472, 2016. [55] C. S. Jiang, M. Yang, Y. Zhou, B. To, S. U. Nanayakkara, J. M. Luther, W. Zhou, J. J. Berry, J. van de Lagemaat, N. P. Padture, K. Zhu, and M. M. Al-Jassim. Carrier separation and transport in perovskite solar cells studied by nanometre-scale profiling of electrical potential. Nat Commun, 6:8397, 2015. [56] Haonan Si, Suicai Zhang, Shuangfei Ma, Zhaozhao Xiong, Ammarah Kausar, Qingliang Liao, Zheng Zhang, Abdul Sattar, Zhuo Kang, and Yue Zhang. Emerging conductive atomic force microscopy for metal halide perovskite materials and solar cells. Advanced Energy Materials, 10(10):1903922, 2020. [57] Avigail Stern, Sigalit Aharon, Tal Binyamin, Abeer Karmi, Dvir Rotem, Lioz Etgar, and Danny Porath. Electrical characterization of individual cesium lead halide perovskite nanowires using conductive afm. Advanced Materials, 32(12):1907812. [58] Martin Green, Ewan Dunlop, Jochen Hohl-Ebinger, Masahiro Yoshita, Nikos Kopidakis, and Xiaojing Hao. Solar cell efficiency tables (version 57). Progress in Photovoltaics: Research and Applications, 29(1):3?15, 2020. [59] Mark V. Khenkin, Eugene A. Katz, Antonio Abate, Giorgio Bardizza, Joseph J. Berry, Christoph Brabec, Francesca Brunetti, Vladimir Bulovi?, Quinn Burlingame, Aldo Di Carlo, Rongrong Cheacharoen, Yi-Bing Cheng, Alexander Colsmann, Stephane Cros, Konrad Domanski, Micha? Dusza, Christopher J. Fell, Stephen R. Forrest, Yulia Galagan, Diego Di Girolamo, Michael Gr?tzel, Anders Hagfeldt, Elizabeth von Hauff, Harald Hoppe, Jeff Kettle, Hans K?bler, Marina S. Leite, Shengzhong (Frank) Liu, Yueh-Lin Loo, Joseph M. Luther, Chang-Qi Ma, Morten Madsen, Matthieu Manceau, Muriel Matheron, Michael McGehee, Rico Meitzner, Mohammad Khaja Nazeeruddin, Ana Flavia Nogueira, ?a?la Odaba??, Anna Osherov, Nam-Gyu Park, Matthew O. Reese, Francesca De Rossi, Michael Saliba, Ulrich S. Schubert, Henry J. Snaith, Samuel D. Stranks, Wolfgang Tress, Pavel A. Troshin, Vida Turkovic, Sjoerd Veenstra, Iris Visoly-Fisher, Aron Walsh, Trystan Watson, Haibing Xie, Ramazan Y?ld?r?m, Shaik Mohammed Zakeeruddin, Kai Zhu, and Monica Lira-Cantu. Consensus statement for stability assessment and reporting for perovskite photovoltaics based on isos procedures. Nature Energy, 5(1):35?49, 2020. [60] John M Howard, Elizabeth M Tennyson, Bernardo RA Neves, and Marina S Leite. Machine learning for perovskites? reap-rest-recovery cycle. Joule, 3:325?337, 2019. [61] T. Duong, H. K. Mulmudi, Y. Wu, X. Fu, H. Shen, J. Peng, N. Wu, H. T. Nguyen, D. Macdonald, M. Lockrey, T. P. White, K. Weber, and K. Catchpole. Light and electrically induced phase segregation and its impact on the stability of quadruple cation high bandgap perovskite solar cells. ACS Appl Mater Interfaces, 9(32):26859? 26866, 2017. 108 [62] G. Grancini, C. Roldan-Carmona, I. Zimmermann, E. Mosconi, X. Lee, D. Martineau, S. Narbey, F. Oswald, F. De Angelis, M. Graetzel, and M. K. Nazeeruddin. One-year stable perovskite solar cells by 2d/3d interface engineering. Nat Commun, 8:15684, 2017. [63] Mei Lyu and Nam-Gyu Park. Effect of additives ax (a=fa, ma, cs, rb, nh, x=cl, br, i) in fapbi on photovoltaic parameters of perovskite solar cells. Solar RRL, 4(10):2000331, 2020. [64] I. J. Park, S. Seo, M. A. Park, S. Lee, D. H. Kim, K. Zhu, H. Shin, and J. Y. Kim. Effect of rubidium incorporation on the structural, electrical, and photovoltaic properties of methylammonium lead iodide-based perovskite solar cells. ACS Appl Mater Interfaces, 9(48):41898?41905, 2017. [65] Yun Hee Park, Inyoung Jeong, Seunghwan Bae, Hae Jung Son, Phillip Lee, Jinwoo Lee, Chul-Ho Lee, and Min Jae Ko. Inorganic rubidium cation as an enhancer for photovoltaic performance and moisture stability of hc(nh2)2pbi3 perovskite solar cells. Advanced Functional Materials, 27(16):1605988, 2017. [66] N. Pellet, P. Gao, G. Gregori, T. Y. Yang, M. K. Nazeeruddin, J. Maier, and M. Gratzel. Mixed-organic-cation perovskite photovoltaics for enhanced solar-light harvesting. Angew Chem Int Ed Engl, 53(12):3151?7, 2014. [67] L. Q. Xie, L. Chen, Z. A. Nan, H. X. Lin, T. Wang, D. P. Zhan, J. W. Yan, B. W. Mao, and Z. Q. Tian. Understanding the cubic phase stabilization and crystallization kinetics in mixed cations and halides perovskite single crystals. J Am Chem Soc, 139(9):3320?3323, 2017. [68] Shaun Tan, Ilhan Yavuz, Nicholas De Marco, Tianyi Huang, Sung-Joon Lee, Christopher S. Choi, Minhuan Wang, Selbi Nuryyeva, Rui Wang, Yepin Zhao, Hao- Cheng Wang, Tae-Hee Han, Bruce Dunn, Yu Huang, Jin-Wook Lee, and Yang Yang. Steric impediment of ion migration contributes to improved operational stability of perovskite solar cells. Advanced Materials, 32(11):1906995, 2020. [69] John M. Howard, Richa Lahoti, and Marina S. Leite. Imaging metal halide perovskites material and properties at the nanoscale. Advanced Energy Materials, 10(26):1903161, 2019. [70] Bowen Gao and Jing Meng. Rbcs(mafa)pbi3 perovskite solar cell with 22.81precise ions cascade regulation. Applied Surface Science, 530:147240, 2020. [71] Daniel Prochowicz, Mohammad Mahdi Tavakoli, Abul Kalam, Rohit D. Chavan, Suverna Trivedi, Manoj Kumar, and Pankaj Yadav. Influence of a-site cations on the open-circuit voltage of efficient perovskite solar cells: a case of rubidium and guanidinium additives. Journal of Materials Chemistry A, 7(14):8218?8225, 2019. [72] P. Yadav, M. I. Dar, N. Arora, E. A. Alharbi, F. Giordano, S. M. Zakeeruddin, and M. Gratzel. The role of rubidium in multiple-cation-based high-efficiency perovskite solar cells. Adv Mater, 29(40):1701077, 2017. 109 [73] Martin Stolterfoht, Pietro Caprioglio, Christian M Wolff, Jos? A M?rquez, Joleik Nordmann, Shanshan Zhang, Daniel Rothhardt, Ulrich H?rmann, Yohai Amir, Alex Redinger, Lukas Kegelmann, Fengshuo Zu, Steve Albrecht, Norbert Koch, Thomas Kirchartz, Michael Saliba, Thomas Unold, and Dieter Neher. The impact of energy alignment and interfacial recombination on the internal and external open-circuit voltage of perovskite solar cells. Energy Environmental Science, 12(9):2778?2788, 2019. [74] Yinghong Hu, Eline M. Hutter, Philipp Rieder, Irene Grill, Jonas Hanisch, Meltem F. Ayg?ler, Alexander G. Hufnagel, Matthias Handloser, Thomas Bein, Achim Hartschuh, Kristofer Tvingstedt, Vladimir Dyakonov, Andreas Baumann, Tom J. Savenije, Michiel L. Petrus, and Pablo Docampo. Understanding the role of cesium and rubidium additives in perovskite solar cells: Trap states, charge transport, and recombination. Advanced Energy Materials, 8(16):1703057, 2018. [75] V. W. Bergmann, Y. Guo, H. Tanaka, I. M. Hermes, D. Li, A. Klasen, S. A. Bretschneider, E. Nakamura, R. Berger, and S. A. Weber. Local time-dependent charging in a perovskite solar cell. ACS Appl Mater Interfaces, 8(30):19402?9, 2016. [76] Ankur Solanki, Pankaj Yadav, Silver-Hamill Turren-Cruz, Swee Sien Lim, Michael Saliba, and Tze Chien Sum. Cation influence on carrier dynamics in perovskite solar cells. Nano Energy, 58:604?611, 2019. [77] J. P. Correa-Baena, Y. Luo, T. M. Brenner, J. Snaider, S. Sun, X. Li, M. A. Jensen, N. T. P. Hartono, L. Nienhaus, S. Wieghold, J. R. Poindexter, S. Wang, Y. S. Meng, T. Wang, B. Lai, M. V. Holt, Z. Cai, M. G. Bawendi, L. Huang, T. Buonassisi, and D. P. Fenning. Homogenized halides and alkali cation segregation in alloyed organic-inorganic perovskites. Science, 363(6427):627?631, 2019. [78] Meng Zhang, Jae S. Yun, Qingshan Ma, Jianghui Zheng, Cho Fai Jonathan Lau, Xiaofan Deng, Jincheol Kim, Dohyung Kim, Jan Seidel, Martin A. Green, Shujuan Huang, and Anita W. Y. Ho-Baillie. High-efficiency rubidium-incorporated perovskite solar cells by gas quenching. ACS Energy Letters, 2(2):438?444, 2017. [79] M. Abdi-Jalebi, Z. Andaji-Garmaroudi, A. J. Pearson, G. Divitini, S. Cacovich, B. Philippe, H. Rensmo, C. Ducati, R. H. Friend, and S. D. Stranks. Potassium- and rubidium-passivated alloyed perovskite films: Optoelectronic properties and moisture stability. ACS Energy Lett, 3(11):2671?2678, 2018. [80] Taisuke Matsui, Tomoyasu Yokoyama, Takayuki Negami, Takashi Sekiguchi, Michael Saliba, Michael Gr?tzel, and Hiroshi Segawa. Effect of rubidium for thermal stability of triple-cation perovskite solar cells. Chemistry Letters, 47(6):814?816, 2018. [81] The Duong, YiLiang Wu, Heping Shen, Jun Peng, Xiao Fu, Daniel Jacobs, Er-Chien Wang, Teng Choon Kho, Kean Chern Fong, Matthew Stocks, Evan Franklin, Andrew Blakers, Ngwe Zin, Keith McIntosh, Wei Li, Yi-Bing Cheng, Thomas P. White, Klaus Weber, and Kylie Catchpole. Rubidium multication perovskite with optimized 110 bandgap for perovskite-silicon tandem with over 26Advanced Energy Materials, 7(14), 2017. [82] Hoang X. Dang, Kai Wang, Masoud Ghasemi, Ming-Chun Tang, Michele De Bastiani, Erkan Aydin, Emilie Dauzon, Dounya Barrit, Jun Peng, Detlef-M. Smilgies, Stefaan De Wolf, and Aram Amassian. Multi-cation synergy suppresses phase segregation in mixed-halide perovskites. Joule, 3(7):1746?1764, 2019. [83] M. Saliba, T. Matsui, J. Y. Seo, K. Domanski, J. P. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt, and M. Gratzel. Cesium-containing triple cation perovskite solar cells: Improved stability, reproducibility and high efficiency. Energy Environ Sci, 9(6):1989?1997, 2016. [84] Andr?s-Felipe Castro-M?ndez, Juanita Hidalgo, and Juan-Pablo Correa-Baena. The role of grain boundaries in perovskite solar cells. Advanced Energy Materials, 9(38):1901489, 2019. [85] J. Song, Y. Zhou, N. P. Padture, and B. D. Huey. Anomalous 3d nanoscale photoconduction in hybrid perovskite semiconductors revealed by tomographic atomic force microscopy. Nat Commun, 11(1):3308, 2020. [86] F. Zhang, F. Ullrich, S. Silver, R. A. Kerner, B. P. Rand, and A. Kahn. Complexities of contact potential difference measurements on metal halide perovskite surfaces. J Phys Chem Lett, 10(4):890?896, 2019. [87] Stefan A. L. Weber, Ilka M. Hermes, Silver-Hamill Turren-Cruz, Christopher Gort, Victor W. Bergmann, Laurent Gilson, Anders Hagfeldt, Michael Graetzel, Wolfgang Tress, and R?diger Berger. How the formation of interfacial charge causes hysteresis in perovskite solar cells. Energy Environmental Science, 11(9):2404?2413, 2018. [88] Dohyung Kim, Rama K. Vasudevan, Kate Higgins, Anna Morozovska, Eugene A. Eliseev, Maxim Ziatdinov, Sergei V. Kalinin, and Mahshid Ahmadi. Exploring responses of contact kelvin probe force microscopy in triple-cation double-halide perovskites. The Journal of Physical Chemistry C, 125(22):12355?12365, 2021. [89] N. Nicoara, R. Manaligod, P. Jackson, D. Hariskos, W. Witte, G. Sozzi, R. Menozzi, and S. Sadewasser. Direct evidence for grain boundary passivation in cu(in,ga)se2 solar cells through alkali-fluoride post-deposition treatments. Nat Commun, 10(1):3980, 2019. [90] Feray ?nl?, Eunhwan Jung, Jinane Haddad, Ashish Kulkarni, Senol ?z, Heechae Choi, Thomas Fischer, Sudip Chakraborty, Thomas Kirchartz, and Sanjay Mathur. Understanding the interplay of stability and efficiency in a-site engineered lead halide perovskites. APL Materials, 8(7):070901, 2020. [91] Y. Sun, J. Peng, Y. Chen, Y. Yao, and Z. Liang. Triple-cation mixed-halide perovskites: towards efficient, annealing-free and air-stable solar cells enabled by pb(scn)2 additive. Sci Rep, 7:46193, 2017. 111 [92] W. Travis, E. N. K. Glover, H. Bronstein, D. O. Scanlon, and R. G. Palgrave. On the application of the tolerance factor to inorganic and hybrid halide perovskites: a revised system. Chem Sci, 7(7):4548?4556, 2016. [93] G. P. Nagabhushana, R. Shivaramaiah, and A. Navrotsky. Direct calorimetric verification of thermodynamic instability of lead halide hybrid perovskites. Proc Natl Acad Sci U S A, 113(28):7717?21, 2016. [94] Wolf-Alexander Quitsch, Dane W. deQuilettes, Oliver Pfingsten, Alexander Schmitz, Stevan Ognjanovic, Sarthak Jariwala, Susanne Koch, Markus Winterer, David S. Ginger, and Gerd Bacher. The role of excitation energy in photobrightening and photodegradation of halide perovskite thin films. The Journal of Physical Chemistry Letters, 9(8):2062?2069, 2018. [95] Konrad Domanski, Essa A. Alharbi, Anders Hagfeldt, Michael Gr?tzel, and Wolfgang Tress. Systematic investigation of the impact of operation conditions on the degradation behaviour of perovskite solar cells. Nature Energy, 3(1):61?67, 2018. [96] Fukashi Matsumoto, Sarah M. Vorpahl, Jannel Q. Banks, Esha Sengupta, and David S. Ginger. Photodecomposition and morphology evolution of organometal halide perovskite solar cells. The Journal of Physical Chemistry C, 119(36):20810? 20816, 2015. [97] J. G. Hu, R. Gottesman, L. Gouda, A. Kama, M. Priel, S. Tirosh, J. Bisquert, and A. Zaban. Photovoltage behavior in perovskite solar cells under light-soaking showing photoinduced interfacial changes. Acs Energy Letters, 2(5):950?956, 2017. [98] D. J. Kubicki, D. Prochowicz, A. Hofstetter, S. M. Zakeeruddin, M. Gratzel, and L. Emsley. Phase segregation in cs-, rb- and k-doped mixed-cation (ma)x(fa)1-xpbi3 hybrid perovskites from solid-state nmr. J Am Chem Soc, 139(40):14173?14180, 2017. [99] Bertrand Philippe, Michael Saliba, Juan-Pablo Correa-Baena, Ute B. Cappel, Silver-Hamill Turren-Cruz, Michael Gr?tzel, Anders Hagfeldt, and H?kan Rensmo. Chemical distribution of multiple cation (rb+, cs+, ma+, and fa+) perovskite materials by photoelectron spectroscopy. Chemistry of Materials, 29(8):3589?3596, 2017. [100] Laura T. Schelhas, Zhen Li, Jeffrey A. Christians, Anuj Goyal, Paul Kairys, Steven P. Harvey, Dong Hoe Kim, Kevin H. Stone, Joseph M. Luther, Kai Zhu, Vladan Stevanovic, and Joseph J. Berry. Insights into operational stability and processing of halide perovskite active layers. Energy Environ. Sci., 12:1341?1348, 2019. [101] Nengxu Li, Xiuxiu Niu, Qi Chen, and Huanping Zhou. Towards commercialization: the operational stability of perovskite solar cells. Chem. Soc. Rev., 49:8235?8286, 2020. [102] Sikai Mei, Zhipeng Yin, Pengcheng Gu, Hai-Qiao Wang, Jianhao Wang, and Weijie Song. Improved operational stability of perovskite solar cells via au barrier layer incorporation. ACS Applied Energy Materials, 4(10):11062?11068, 2021. 112 [103] Taame Abraha Berhe, Wei-Nien Su, Ching-Hsiang Chen, Chun-Jern Pan, Ju-Hsiang Cheng, Hung-Ming Chen, Meng-Che Tsai, Liang-Yih Chen, Amare Aregahegn Dubale, and Bing-Joe Hwang. Organometal halide perovskite solar cells: degradation and stability. Energy Environ. Sci., 9:323?356, 2016. [104] Jiangyu Li, Boyuan Huang, Ehsan Nasr Esfahani, Linlin Wei, Jianjun Yao, Jinjin Zhao, and Wei Chen. Touching is believing: interrogating halide perovskite solar cells at the nanoscale via scanning probe microscopy. npj Quantum Materials, 2(1):56, 2017. [105] Haonan Si, Suicai Zhang, Shuangfei Ma, Zhaozhao Xiong, Ammarah Kausar, Qingliang Liao, Zheng Zhang, Abdul Sattar, Zhuo Kang, and Yue Zhang. Emerging conductive atomic force microscopy for metal halide perovskite materials and solar cells. Advanced Energy Materials, 10(10):1903922. [106] Silver-Hamill Turren-Cruz, Anders Hagfeldt, and Michael Saliba. Methylammonium- free, high-performance, and stable perovskite solar cells on a planar architecture. Science, 362(6413):449?453, 2018. [107] Nam Joong Jeon, Jun Hong Noh, Woon Seok Yang, Young Chan Kim, Seungchan Ryu, Jangwon Seo, and Sang Il Seok. Compositional engineering of perovskite materials for high-performance solar cells. Nature, 517(7535):476?480, 2015. [108] Melepurath Deepa, Manuel Salado, Laura Calio, Samrana Kazim, S. M. Shivaprasad, and Shahzada Ahmad. Cesium power: low cs+ levels impart stability to perovskite solar cells. Phys. Chem. Chem. Phys., 19:4069?4077, 2017. [109] Xingtao Wang, Yuetian Chen, Taiyang Zhang, Xiaoyu Wang, Yong Wang, Miao Kan, Yanfeng Miao, Haoran Chen, Xiaomin Liu, Xinjiang Wang, Jielin Shi, Lijun Zhang, and Yixin Zhao. Stable cesium-rich formamidinium/cesium pure-iodide perovskites for efficient photovoltaics. ACS Energy Letters, 6(8):2735?2741, 2021. [110] Ning Zhou, Yiheng Shen, Yu Zhang, Ziqi Xu, Guanhaojie Zheng, Liang Li, Qi Chen, and Huanping Zhou. Csi pre-intercalation in the inorganic framework for efficient and stable fa1x csxpbi3(cl) perovskite solar cells. Small, 13(23):1700484. [111] Caleb C. Boyd, Rongrong Cheacharoen, Tomas Leijtens, and Michael D. McGehee. Understanding degradation mechanisms and improving stability of perovskite photovoltaics. Chemical Reviews, 119(5):3418?3451, 2019. [112] Craig H. Swartz, Nischal Khakurel, Selena R. Najar, Mohammed I. Hossain, and Alex Zakhidov. Temperature- and bias-dependent degradation and regeneration of perovskite solar cells with organic and inorganic hole transport layers. physica status solidi (a), 218(7), 2021. [113] Craig H. Swartz, Nischal Khakurel, Selena R. Najar, Mohammed I. Hossain, and Alex Zakhidov. Temperature- and bias-dependent degradation and regeneration of perovskite solar cells with organic and inorganic hole transport layers. physica status solidi (a), 218(7), 2021. 113 [114] Leiping Duan and Ashraf Uddin. Defects and stability of perovskite solar cells: a critical analysis. Mater. Chem. Front., 6:400?417, 2022. [115] Kevin Ho, Mingyang Wei, Edward H. Sargent, and Gilbert C. Walker. Grain transformation and degradation mechanism of formamidinium and cesium lead iodide perovskite under humidity and light. ACS Energy Letters, 6(3):934?940, 2021. [116] J. M. Howard, E. M. Tennyson, S. Barik, R. Szostak, E. Waks, M. F. Toney, A. F. Nogueira, B. R. A. Neves, and M. S. Leite. Humidity-induced photoluminescence hysteresis in variable Cs/Br ratio hybrid perovskites. The Journal of Physical Chemistry Letters, 9(12):3463?3469, 2018. [117] Benjamin T. Diroll, Georgian Nedelcu, Maksym V. Kovalenko, and Richard D. Schaller. High-temperature photoluminescence of cspbx3 (x = cl, br, i) nanocrystals. Advanced Functional Materials, 27(21):1606750, 2017. [118] Hee Joon Jung, Daehan Kim, Sungkyu Kim, Joonsuk Park, Vinayak P. Dravid, and Byungha Shin. Stability of halide perovskite solar cell devices: In situ observation of oxygen diffusion under biasing. Advanced Materials, 30(39):1802769, 2018. [119] Xin Meng, Xueying Tian, Shasha Zhang, Jing Zhou, Yiqiang Zhang, Zonghao Liu, and Wei Chen. In-situ characterization for understanding the degradation in perovskite solar cells. Solar RRL, n/a(n/a). [120] Soumya Kundu and Timothy L. Kelly. In situ studies of the degradation mechanisms of perovskite solar cells. EcoMat, 2(2):e12025, 2020. [121] Bo-An Chen, Jin-Tai Lin, Nian-Tzu Suen, Che-Wei Tsao, Tzu-Chi Chu, Ying-Ya Hsu, Ting-Shan Chan, Yi-Tsu Chan, Jye-Shane Yang, Ching-Wen Chiu, and Hao Ming Chen. In situ identification of photo- and moisture-dependent phase evolution of perovskite solar cells. ACS Energy Letters, 2(2):342?348, 2017. [122] Lidia Contreras-Bernal, Clara Aranda, Marta Valles-Pelarda, Thi Tuyen Ngo, Susana Ramos-Terr?n, Juan Jes?s Gallardo, Javier Navas, Antonio Guerrero, Ivan Mora- Sero, Jesus Idigoras, et al. Homeopathic perovskite solar cells: effect of humidity during fabrication on the performance and stability of the device. The Journal of Physical Chemistry C, 122(10):5341?5348, 2018. [123] Isabel Mesquita, Luisa Andrade, and Ad?lio Mendes. Effect of relative humidity during the preparation of perovskite solar cells: Performance and stability. Solar Energy, 199:474?483, 2020. [124] Joel Troughton, Katherine Hooper, and Trystan M Watson. Humidity resistant fabrication of ch3nh3pbi3 perovskite solar cells and modules. Nano Energy, 39:60?68, 2017. [125] Jian Wu, Jing-Jing Dong, Si-Xuan Chen, Hui-Ying Hao, Jie Xing, and Hao Liu. Fabrication of efficient organic-inorganic perovskite solar cells in ambient air. Nanoscale research letters, 13(1):1?7, 2018. 114 [126] Didac Pitarch-Tena, Thi Tuyen Ngo, Marta Vall?s-Pelarda, Thierry Pauport?, and Iv?n Mora-Ser?. Impedance spectroscopy measurements in perovskite solar cells: Device stability and noise reduction. ACS Energy Letters, 3(4):1044?1048, 2018. [127] Yanbo Wang, Xiao Liu, Zhongmin Zhou, Pengbin Ru, Han Chen, Xudong Yang, and Liyuan Han. Reliable measurement of perovskite solar cells. Advanced Materials, 31(47):1803231. [128] Y. Yuan and J. Huang. Ion migration in organometal trihalide perovskite and its impact on photovoltaic efficiency and stability. Accounts of Chemical Research, 49(2):286?93, 2016. [129] C. Li, A. Guerrero, Y. Zhong, A. Graser, C. A. M. Luna, J. Kohler, J. Bisquert, R. Hildner, and S. Huettner. Real-time observation of iodide ion migration in methylammonium lead halide perovskites. Small, 13(42):1701711, 2017. [130] Yu An, Carlo Andrea Riccardo Perini, Juanita Hidalgo, Andr?s-Felipe Castro- M?ndez, Jacob N. Vagott, Ruipeng Li, Wissam A. Saidi, Shirong Wang, Xianggao Li, and Juan-Pablo Correa-Baena. Identifying high-performance and durable methylammonium-free lead halide perovskites via high-throughput synthesis and characterization. Energy Environmental Science, 14(12):6638?6654, 2021. [131] Tianyue Li, Yue Hu, Carole A. Morrison, Wenjun Wu, Hongwei Han, and Neil Robertson. Lead-free pseudo-three-dimensional organic?inorganic iodobismuthates for photovoltaic applications. Sustainable Energy Fuels, 1:308?316, 2017. [132] Conor Rocks, Vladimir Svrcek, Paul Maguire, and Davide Mariotti. Understanding surface chemistry during mapbi3 spray deposition and its effect on photovoltaic performance. J. Mater. Chem. C, 5:902?916, 2017. [133] Qin Zhou, Lusheng Liang, Junjie Hu, Bingbing Cao, Longkai Yang, Tingjun Wu, Xin Li, Bao Zhang, and Peng Gao. High-performance perovskite solar cells with enhanced environmental stability based on a (p-fc6h4c2h4nh3)2[pbi4] capping layer. Advanced Energy Materials, 02 2019. [134] Youngin Goh, Sung Hyun Cho, Sang-Hee Ko Park, and Sanghun Jeon. Oxygen vacancy control as a strategy to achieve highly reliable hafnia ferroelectrics using oxide electrode. Nanoscale, 12:9024?9031, 2020. [135] Shipra Jain, Jyoti Shah, N. Negi, Chhemendra Sharma, and R.K. Kotnala. Significance of interface barrier at electrode of hematite hydroelectric cell for generating ecopower by water splitting. International Journal of Energy Research, 43, 06 2019. [136] Alexander Klasen, Philipp Baumli, Qu Sheng, Ewald Johannes, Simon A. Bretschneider, Ilka M. Hermes, Victor W. Bergmann, Christopher Gort, Amelie Axt, Stefan A. L. Weber, Heejae Kim, Hans-J?rgen Butt, Wolfgang Tremel, and R?diger Berger. Removal of surface oxygen vacancies increases conductance through tio2 thin 115 films for perovskite solar cells. The Journal of Physical Chemistry C, 123(22):13458? 13466, 2019. PMID: 31205577. [137] Ye Wu, Yingcheng Lin, and Jian Xu. Synthesis of ag?ho, ag?sm, ag?zn, ag?cu, ag?cs, ag?zr, ag?er, ag?y and ag?co metal organic nanoparticles for uv-vis-nir wide-range bio-tissue imaging. Photochem. Photobiol. Sci., 18:1081?1091, 2019. [138] Fan Zhang, Qingxun Huang, Jun Song, Shuzi Hayase, Junle Qu, and Qing Shen. A new strategy for increasing the efficiency of inverted perovskite solar cells to more than 21%: High-humidity induced self-passivation of perovskite films. Solar RRL, 4(9):2000149, 2020. [139] M. Prete, M. V. Khenkin, D. Glowienka, B. R. Patil, J. S. Lissau, I. Dogan, J. L. Hansen, T. Lei?ner, J. Fiutowski, H.-G. Rubahn, B. Julsgaard, P. Balling, V. Turkovic, Y. Galagan, E. A. Katz, and M. Madsen. Bias-dependent dynamics of degradation and recovery in perovskite solar cells. ACS Applied Energy Materials, 4(7):6562?6573, 2021. [140] Yicheng Zhao, Jiyun Zhang, Zhengwei Xu, Shijing Sun, Stefan Langner, Noor Titan Putri Hartono, Thomas Heumueller, Yi Hou, Jack Elia, Ning Li, et al. Discovery of temperature-induced stability reversal in perovskites using high-throughput robotic learning. Nature communications, 12(1):1?9, 2021. [141] Sang-Won Lee, Seongtak Kim, Soohyun Bae, Kyungjin Cho, Taewon Chung, Laura E Mundt, Seunghun Lee, Sungeun Park, Hyomin Park, Martin C Schubert, et al. Uv degradation and recovery of perovskite solar cells. Scientific reports, 6(1):1?10, 2016. [142] Hiroki Uratani and Koichi Yamashita. Charge carrier trapping at surface defects of perovskite solar cell absorbers: A first-principles study. The Journal of Physical Chemistry Letters, 8(4):742?746, 2017. PMID: 28129504. [143] Handong Jin, Elke Debroye, Masoumeh Keshavarz, Ivan G. Scheblykin, Maarten B. J. Roeffaers, Johan Hofkens, and Julian A. Steele. It?s a trap! on the nature of localised states and charge trapping in lead halide perovskites. Mater. Horiz., 7:397?410, 2020. 116