DEFECT CHEMISTRY AND PHOTO-IONIC EFFECTS IN BROMIDE AND IODIDE PEROVSKITES Von der Fakultät Chemie der Universität Stuttgart zur Erlangung der Würde eines Doktors der Naturwissenschaften (Dr. rer. nat.) genehmigte Abhandlung vorgelegt von YARU WANG (Ya-Ru Wang) aus Hebei, China Hauptberichter: Prof. Dr. Joachim Maier Mitberichter: Prof. Dr. Thomas Schleid Prüfungsvorsitzender: Prof. Dr. Joris van Slageren Tag der Einreichung: 22.05.2023 Tag der mündlichen Prüfung: 26.07.2023 Physikalische Festkörperchemie Max-Planck Institut für Festkörperforschung Stuttgart, 2023 苟于行之不迷 虽颠沛何其伤 -韩愈 i Erklärung über die Eigenständigkeit der Dissertation Ich versichere, dass ich die vorliegende Arbeit mit dem Titel “Defektchemie und Photoionische Effekte in Bromid- und Iodidperowskiten” selbständig verfasst und keine anderen als die angegebenen Quellen und Hilfsmittel benutzt habe; aus fremden Quellen entnommene Passagen und Gedanken sind als solche kenntlich gemacht. Declaration of Authorship I hereby certify that the dissertation entitled “Defect Chemistry and Photo-ionic Effects in Bromide and Iodide Perovskites” is entirely my own work except where otherwise indicated. Passages and ideas from other sources have been clearly indicated. Name/Name: Ya-Ru Wang Unterschrift/Signed: Datum/Date: ii iii Acknowledgments This journey of working as a Ph.D. student in Max Planck Institute in Solid State Research is such a rewarding and enjoyable experience for me. A very important part of it are those people who had accompanied me through all these years. I would like to express my greatest gratitude to all of you. I would like to thank Prof. Joachim Maier for welcoming me to his group. I appreciate it so much for the insightful discussions, great support and the freedom for me to explore. In addition, I am grateful for the very friendly and international group atmosphere that you had created, which made me feeling as home, especially at the difficult COVID time. Many thanks to my daily supervisor Dr. Davide Moia for being always supportive and helpful. Thank you for the continuous effort on listening to and trusting on me, without which I cannot go so far. Thank you also for a lot of training and mentorship, those are the very precious experiences and memories that I will value a lot. A lot of thanks to Dr. Alessandro Senocrate and Dr. Gee Yeong Kim, who I had worked with at the early stage of my PhD. Thank you Ale for always being inspiring and helpful, showing me around in the lab, brainstorming a lot, missing a lot on your contagious laughs Thank you Gee Yeong for a lot of insightful discussions and collaborative work together. I enjoyed so much working and sharing life with you. Many thanks to Prof. Michael Grätzel and Prof. Jovana V. Milic for their continuous support and fruitful collaborations on this project. I’m also grateful to Prof. Ursula Röthlisberger and Dr. Marko Mladenovic on the collaborative work on theoretical calculations. I’m also grateful and indebted to Prof. Bettina Lotsch. Thank you also being a great role model to me. Special thanks for taking the time to talk to me and support me to take the next step of my research. I owe great thanks to the people who work in nanostructuring lab in MPI-FKF. I would like to thank the head of the lab Jürgen Weis for his commitment on being my external supervisor of my Ph.D. committee and his great and continuous support on my research project. Especially I would like to express the greatest thanks to Bernhard Fenk and Ulrike Waizmann. To Bernhard Fenk: It’s always been so much fun and excitement to work with you. Your passion and expertise on the work had inspired me so much. We joked around with Ulli that you are the “nano-master”, but I mean it. To Ulrike Waizmann: thank you so much for your dedication to this project, hope that halide perovskites will not appear in your dream anymore. I’m also grateful to Achim Güth for making electrode, Thomas Reindl for training me to use SEM, Marion Hagel for some test on ALD growth. I’m deeply grateful to the people who I work with in StEM group. To head of the group Peter van Aken: thank you for your support on my project. To Kersten Hahn: thank you so iv much for your dedication to the project and your saying about “let’s check our calendar”. To Wilfried Sigle, thank you for always being available for me to ask questions and always asking how is it going when you encounter me in the computer room. To Peter Kopold: thank you for the assistance of the TEM measurements. To Vesna Srot, thank you for training me using the Plasma cleaning instrument in the group. Special thanks goes to Florian Kaiser for helping me building home-made measurements cells, set ups. I had learnt so much with you on technical stuff. A lot of thanks to Dr. Rotraut Merkle for always being available for any help or support whenever is needed. It truly astounds me how you manage it for everyone. Many thanks to Udo Klock for electronics und auch für Ihre ständige Ermutigung, sich auf Deutsch zu unterhalten. To Uwe Traub, thanks a lot for your support on IT stuff and always updating me some news in Germany. To Armin Sorg and Helga Hoier, thank you so much for help on the numerous XRD measurements during these years. To Dr. Klaus-Dieter Kreuer, thank you for taking the time for reviewing the Zusammenfassung of this thesis, loved a lot for the yearly “Waldmeister” event. Thanks to Prof. Eugene Kotomin for a lot of stimulating discussions and the greetings on Chinese new year almost every year. Thanks to Dr. Igor Moudrakovski, who I had some discussion on NMR. Thanks Armin Schulz for Raman measurements on looking for vibration modes may or may not exist. Sincerely many thanks go to our secretaries, Madeleine Burkhardt and Sofia Weiglein, who had done tremendous support on the administrative stuff behind the scene. I’m also grateful for Birgit King, who had done a lot of work for my residency in Germany. Special thank also to your kind heart when I’m dealing with housing issues. Furthermore, I would like to thank Dr. Hans-Georg Libuda for his kind help in the IMPRS-CMS program, meeting you in Shanghai and getting to know this program has made my life change in another direction, in a good way. Special thanks to Eva Benckiser, for her kind help in the IMPRS-CMS program and mentor program. Many thanks to my fellow students and scientists who I shared a lot of same values and experiences with, making my time in MPI extremely joyful. To Mina Jung, thank you very much for sharing and talking both work and life related stuff with me. I like so much of the resonation between us. To Chulian Xiao, I enjoyed so much being as xiaoshimei of you, enjoyed a lot talking with you. To Dr. Algirdas Dučinskas, sharing perspective on books is so much fun! I also like us always on the same page of “5km? Let’s walk home”. To Dr. Yuanye Huang and Dr. Markus Joos, thanks for a lot of activities and laughs together. To Dr. Giulia Raimondi, Sicily is definitely one of the most beautiful place. To Dr. Torben Saatkamp, I like so much of your contagious smile and kind heart, thank you for helping me through some difficult time. To Dr. Andreas Münchinger, thank you for sharing experiences and thoughts v with me. To Dr. Maximilian Hödl, thank you for taking my ‘red face’ pictures in Frühlingsfest 2019 and some laughs. To Dr. Christian Berger, thanks for always being kind to me. To Dr. Dr. Yue Zhu, I liked our talk when encountering in the lab in late evenings. To Dr. Chuanhai Gan and Liyu shi, thanks a lot for inviting me for a lot of delicious food which made me feeling like home. To Dr. Kyungmi Lim, I liked our after-lunch walk and talk. To Dr. Xiaolan Kang, we need to hike together on Alps. To Prof. Jelena Popovic-Neuber, thanks for sharing your opinions and thoughts with me. To some other friends I met in and out of MPI, who had accompanied me and provided a lot of support. To Liwen Feng, thank you for being around almost my whole period of my PhD, I like your altitude of let’s do it now. To Yu-Jung Wu, so much surprise for the gifts that you brought to me, thanks for always being so sweet to me. To Gabriele Domaine, I enjoyed so much having photography activities with you. To Niklas Windbacher, thank you for a lot of talks and support and the shared laughs. To Dr. Giniyat Khaliullin, thanks for continuous invitation to paly table tennis, that made my life much healthier. To Yuanshan Zhang, thanks for being a nice football coach to me. Büsnau Baozi restaurant is a solid idea. To Jihuan Gu, thanks for some shared activities with you. To Prof. Feixiang Wu, the mental picture of you looking back and smiling in Christmas Market 2018 feels like yesterday. To Dr. Qing yu He, I will never forget the email that you replied to me. To Prof. Laifa shen, being a shimei of you is such a joy, thank you for helping me looking for apartments in Germany. To the friends that I met in IS institute: Fan Wang, Dr. Zhen Yin, Dr. Wenbin Kang, Dr. Minchao Zhang and Xianglong Lyu, enjoyed the Friday dinner and supermarket shopping with you. Many thanks to Prof. Thomas Schleid and Prof. Joris van Slageren for their commitment to be my thesis committee members and to read this thesis. Lastly, I would like to express my heartfelt gratitude to my parents, my brother for being always supportive and proud of me. Your unconditional love for me has made me being courageous to explore and to adventure. Love you as always. Stuttgart, May 18th, 2023 Ya-Ru Wang (王亚茹) vi vii Abstract Bromide and iodide perovskites, especially their mixtures, hold great potential for opto- electronic application due to their optical absorption properties in the visible range. While it is established that these materials are mixed ionic-electronic conductors, their ionic transport properties both in the dark and under light are poorly understood. The present work deals with the defect chemistry and photo-ionic effects in halide perovskites, including iodide and bromide perovskites with special focus on the photo induced phase separation (photo de-mixing) in mixed bromide - iodide perovskites. The first part covers the defect chemical study of bromide perovskites, including 3D MAPbBr3 and 2D Dion-Jacobson (PDMA)PbBr4. The results reveal that both 3D MAPbBr3 and 2D (PDMA)PbBr4 are mixed ionic-electronic conductors. 2D (PDMA)PbBr4 show three orders of magnitude lower ionic conductivity compared with 3D MAPbBr3. This implies that dimensionality reduction is an effective strategy for reducing ion migration in these systems. From a bromine partial pressure dependent study, it is concluded that MAPbBr3 and 2D(PDMA)PbBr4 are both P-type conductor and that the surface reaction is the limiting process for the incorporation and exporation of the Br2 gas. A non-monotonic dependence of the electronic conductivity on bromine partial pressure is detected for both 2D and 3D bromide perovskites. It can be attributed to the reversible formation and dissociation of AuBrx on the gold electrode and perovskites interface. The second part covers the investigation of the thermodynamic properties of the 2D mixed halide perovskites under light. It has been shown that light can be used as a knob for inducing photo de-mixing from single phase 2D mixed halide perovskites to to I-rich and Br- rich phases. In the dark, the photo de-mixed phases re-mix with complete reversibility of both their optical and structural properties, demonstrating the full miscibility of mixed bromide- iodide perosvkites in the dark. The temperature-dependence of absorption spectra for the photo de-mixed phases gave clear evidence for a miscibility gap under light, from which photo de- mixed phases’ compositions are extracted. The photo-miscibility-gap is mapped and confirmed by various methods. The shape of the photo-miscibility-gap shows limited variation in the 0.01 – 0.1 sun illumination intensity range. The non-encapsulation of surface, however, demonstrated a widening of the photo-miscibility gap. The third part covers the kinetic analysis and mechanistic investigation of photo de- mixing in 2D mixed halide perovskites. Simultaneous monitoring of the electrical conductivity and optical absorption allows for a local probe of electronic and ionic charge carriers, and the composition evolution. Furthermore, time dependent phase distribution is investigated with the aid of top view SEM, showing that I-rich nanodomains forming along the grain boundaries at early times after light exposure with further formation of such domains also within the grain viii at longer times. Local elementary distribution is probed with TEM. From the temperature dependent de-mixing half-time, an activation energy for photo de-mixing of 0.39 eV is obtained. Finally, together with DFT calculation on defect formation energy of the mixture of different defect type, a mechanistic description for photo de-mixing, both from molecular and microscopic level is proposed. The last part deal with the phase stability study of mixed halide perovskites in other dimensionalities, including 3D and nanocrystal based thin films. 3D mixed halide perovskites show that similar to 2D mixed halide perovskites, photo de-mixing occur in two stage. Different from the full reversibility of 2D, photo degradation of 3D perovskites into PbI2 in the dark over long time scales is observed. The nanocrystalline mixed halide perovskite (BA- MAPb(I0.5Br0.5)3) thin films show no de-mixing in contrast to 2D and 3D under same measurement conditions. Nanocrystals mixtures show superior phase stability under the same illumination condition, with neither degradation nor de-mixing. This thesis contributes to the understanding of the defect chemistry and ion transport properties of bromide and iodide perovskites, with specific focus on photo de-mixing in mixed halide perovskites. These findings will aid compositional engineering related to halide mixtures to enable optimization of optoelectronic devices as well as the development of other emerging systems exploiting photo-ionic effects. ix Zusammenfassung Bromid- und Iodidperowskite, insbesondere ihre Mischungen, bergen aufgrund ihrer optischen Absorptionseigenschaften im sichtbaren Spektrum großes Potenzial für optoelektronische Anwendungen. Es steht zwar fest, dass diese Materialien gemischt ionisch- elektronische Leiter sind, jedoch sind ihre Ionentransporteigenschaften sowohl im Dunkeln als auch bei Licht nur unzureichend bekannt. Die vorliegende Arbeit befasst sich mit der Defektchemie und den photoionischen Effekten in Halogenidperowskiten, einschließlich Iodid- und Bromidperowskiten, mit besonderem Schwerpunkt auf der photoinduzierten Phasentrennung (Photoentmischung) in gemischten Bromid-Iodid-Perowskiten. Der erste Teil behandelt die defektchemische Untersuchung von Bromidperowskiten, einschließlich 3D MAPbBr3 und 2D Dion-Jacobson (PDMA)PbBr4. Die Ergebnisse zeigen, dass sowohl 3D MAPbBr3 als auch 2D (PDMA)PbBr4 Mischleiter sind. 2D (PDMA)PbBr4 zeigt eine um drei Größenordnungen geringere Ionenleitfähigkeit im Vergleich zu 3D MAPbBr3. Dies deutet darauf hin, dass die Reduzierung der Dimension eine wirksame Strategie zur Verringerung der Ionenwanderung in diesen Systemen ist. Aus einer Brom-Partialdruck abhängigen Studie wird gefolgert, dass MAPbBr3 und 2D(PDMA)PbBr4 beide P-Leiter sind und dass die Oberflächenreaktion der einschränkende Prozess für den Ein- und Austritt des Br2-Gases ist. Sowohl für 2D- als auch für 3D-Bromidperowskite wird eine nicht-monotone Abhängigkeit der elektronischen Leitfähigkeit vom Brom-Partialdruck festgestellt. Sie lässt sich auf die reversible Bildung und Dissoziation von AuBrx an der Grenzfläche zwischen Goldelektrode und Perowskiten zurückführen. Der zweite Teil befasst sich mit der Untersuchung der thermodynamischen Eigenschaften der 2D-Mischhalogenidperowskite unter Lichteinwirkung. Es konnte gezeigt werden, dass Licht als Mittel zur Einleitung der Photoentmischung von einphasigen 2D- Mischhalogenidperowskiten zu I-reichen und Br-reichen Phasen verwendet werden kann. In der Dunkelheit vermischen sich die durch Licht entmischten Phasen wieder mit vollständiger Reversibilität ihrer optischen und strukturellen Eigenschaften, was die vollständige Mischbarkeit von gemischten Bromid-Iodid-Perowskiten in der Dunkelheit zeigt. Die Temperaturabhängigkeit der Absorptionsspektren für die photoentmischten Phasen lieferte eindeutige Hinweise auf eine Mischungslücke unter Licht, aus der die Zusammensetzung der photoentmischten Phasen abgeleitet werden kann. Die Fotomischbarkeitslücke wird mit verschiedenen Methoden kartiert und bestätigt. Die Form dieser Lücke zeigt wenig Veränderung im Bereich der Beleuchtungsintensität von 0,01 bis 0,1 Strahlungsleistungen der Sonne. Die Nichtverkapselung der Oberfläche führte jedoch zu einer Vergrößerung der Fotomischbarkeitslücke. x Der dritte Teil befasst sich mit der kinetischen Analyse und der mechanistischen Untersuchung der Photoentmischung in 2D-Perowskiten mit gemischten Halogeniden. Die gleichzeitige Überwachung der elektrischen Leitfähigkeit und der optischen Absorption ermöglicht eine lokale Untersuchung der elektronischen und ionischen Ladungsträger sowie der Entwicklung der Zusammensetzung. Darüber hinaus wird die zeitabhängige Phasenverteilung mit Hilfe von REM-Aufnahmen untersucht, die zeigen, dass sich I-reiche Nanodomänen entlang der Korngrenzen zu frühen Zeiten nach der Lichtexposition bilden, wobei sich solche Domänen nach längerer Zeit auch innerhalb des Korns bilden. Die lokale Elementarverteilung wird mit TEM untersucht. Aus der temperaturabhängigen Halbwertszeit der Entmischung ergibt sich eine Aktivierungsenergie für die Photoentmischung von 0,39 eV. Zusammen mit DFT- Berechnungen zur Defektbildungsenergie der Mischung verschiedener Defekttypen, wird schließlich eine mechanistische Beschreibung der Photoentmischung sowohl auf molekularer als auch auf mikroskopischer Ebene vorgestellt. Der letzte Teil befasst sich mit der Untersuchung der Phasenstabilität von gemischten Halogenidperowskiten in anderen Dimensionen, einschließlich 3D- und nanokristallbasierten Dünnschichten. 3D-Perowskite mit gemischten Halogeniden zeigen, dass ähnlich wie bei 2D- Perowskiten mit gemischten Halogeniden eine Entmischung in zwei Phasen stattfindet. Im Gegensatz zur vollständigen Reversibilität der 2D-Perowskite wird bei den 3D-Perowskiten im Dunkeln über lange Zeiträume ein Photoabbau zu PbI2 beobachtet. Die nanokristallinen Mischhalogenid-Perowskite (BA-MAPb(I0,5Br0,5)3) zeigen im Gegensatz zu 2D und 3D unter gleichen Messbedingungen keine Entmischung. Nanokristallmischungen zeigen unter den gleichen Beleuchtungsbedingungen eine bessere Phasenstabilität, ohne dass es zu einer Degradation oder Entmischung kommt. Diese Arbeit trägt zum Verständnis der Defektchemie und der Ionentransporteigenschaften von Bromid- und Iodidperowskiten bei, wobei der Schwerpunkt auf der Photoentmischung in gemischten Halogenidperowskiten liegt. Diese Erkenntnisse werden bei der Zusammensetzungsentwicklung im Zusammenhang mit Halogenidmischungen helfen, um die Optimierung optoelektronischer Bauelemente sowie die Entwicklung anderer neuer Systeme zu ermöglichen, die opto-ionische Effekte nutzen. xi Contents Acknowledgments ................................................................................................... iii Abstract ..................................................................................................................vii Zusammenfassung ....................................................................................................ix Contents ..................................................................................................................xi I INTRODUCTORY REMARKS .............................................................................. 13 1 Introduction ........................................................................................................... 1 1.1 Motivation ....................................................................................................... 1 1.2 Thesis outline .................................................................................................. 1 II STATE OF RESEARCH ......................................................................................... 3 2.1 Halide perovskites and their application ............................................................. 5 2.1.1 Dimensionality engineering ........................................................................... 6 2.1.2 Halide engineering and mixed halide perovskites ......................................... 7 2.2 Defect chemistry and ion transport in halide perovskites ................................... 8 2.3 Photo induced ionic effects in halide perovskites ................................................ 9 2.4 De-mixing and photo de-mixing ........................................................................ 10 2.4.1 Experimental signature of photo de-mixing ................................................ 11 2.4.2 Photo de-mixing and device performance ................................................... 12 2.4.3 Mitigation strategies ................................................................................... 13 2.4.4 Photo de-mixing models ............................................................................. 15 2.4.5 Open questions ........................................................................................... 18 III THEORETICAL FRAMEWORK......................................................................... 21 3.1 Mixture thermodynamics .................................................................................. 23 IV MATERIALS AND METHODS ........................................................................... 27 4 Materials and methods ......................................................................................... 29 4.1 Materials and synthesis ................................................................................. 29 4.2 Experimental apparatus and techniques ........................................................ 30 4.3 Evaluation of the partial conductivities ........................................................ 35 V RESULTS AND DISCUSSIONS ............................................................................ 37 5 Defect chemistry equilibria in bromide and iodide perovskites ............................ 39 5.1 Introduction ................................................................................................... 39 5.2 Results and discussion ................................................................................... 40 5.3 Conclusion ..................................................................................................... 50 6 Photo de-mixing in 2D mixed halide perovskites: thermodynamics ..................... 51 xii 6.1 Introduction ................................................................................................... 51 6.2 Results and discussion ................................................................................... 52 6.3 Conclusion ..................................................................................................... 61 7 Photo de-mixing in 2D mixed halide perovskites: kinetics and mechanism ......... 63 7.1 Introduction ................................................................................................... 63 7.2 Results and discussion ................................................................................... 64 7.4 Conclusion ..................................................................................................... 77 8 Photo de-mixing in mixed halide perovskites: 3D and nanocrystals .................... 79 8.1 Introduction ................................................................................................... 79 8.2 Results and discussion ................................................................................... 79 8.3 Conclusion ..................................................................................................... 87 9 Photo de-mixing in mixed halide perovskites: mechanism ................................... 89 9.1 Introductory remarks..................................................................................... 89 9.2 Evaluating the energetics of photo-ionic effects in mixed halide perovskites 90 VI CONCLUSIONS .................................................................................................... 97 10 Concluding remarks ........................................................................................... 99 VII APPENDIX ....................................................................................................... 103 A Supporting Material .......................................................................................... 105 A.1 supporting material for Chapter 5 .............................................................. 105 A.2 supporting material for Chapter 6 .............................................................. 111 A.3 supporting material for Chapter 7 .............................................................. 123 A.4 supporting material for Chapter 8 .............................................................. 135 B Bibliography ...................................................................................................... 137 C Curriculum Vitae .............................................................................................. 149 I INTRODUCTORY REMARKS 1 1 Introduction 1.1 Motivation Organic-inorganic hybrid halide perovskites are regarded as promising light absorbers for a wide range of photovoltaic applications owing to their low cost and easy processibility.1- 3 However, these materials are susceptible to external stimuli, such as humidity, light, and heat, which limits their long-term stability. This remains to be one of the pressing concerns for the future utilization of these photovoltaic materials.4-7 Unlike traditional semiconductors which conduct electrons and holes only, this class of materials belongs to the category of mixed ionic-electronic conductors with both electron and ion conduction intrinsically involved. In addition, under light, MAPbI3 (an architype of halide perovskites) showed huge enhancement in the ionic conductivity.8 Such photo enhanced ion transport is linked to degradation, influencing long-term stability.9, 10 It is therefore critical to better understand these materials and the photo-enhanced ionic transport in order to improve their stability and expand their potential. In addition to its application importance, this novel phenomenon of ion transport out of equilibrium is of great interest fundamentally, from both a molecular and microscopic perspective. Another striking example a photo-induced ionic process is the photo-induced phase separation occurring in mixed halide perovskites.11, 12 This process concerns the change in miscibility of perovskites with two different halides on application of light. Several steps are expected to give rise to this photo induced separation, including photo-induced defect formation, ionic defects transport, and phase transformation. As such, it constitutes a complex problem to unravel as well as a convenient model phenomenon for investigating photo-ionic effects and defect behavior under light. 1.2 Thesis outline The present work deals with the defect chemistry and photo-ionic effects in halide perovskites, specifically in iodide and bromide perovskites, with special focus on the photo- induced phase separation (photo de-mixing) occurring in mixed bromide-iodide perovskites. The thesis has the following structure: the introductory sections summarize the state of research that is related to defect chemistry and photo-ionic effects in iodide and bromide perovskites (Part Ⅱ, chapter 2) and the theoretical framework of mixture thermodynamics (Part Ⅲ, chapter 3). Part Ⅳ (chapter 4) summarizes the materials and methods as well as the experimental apparatus and measurement techniques. The results and discussion chapters are presented in Part Ⅴ: Chapter 5 covers the defect chemical study of bromide perovskites. Specifically, methylammonium lead bromide (MAPbBr3) and 2D Dion-Jacobson 1,4- phenylenedimethanammonium lead bromide ((PDMA)PbBr4) are investigated. Their mixed 2 conducting properties are measured as a function of varying stoichiometry (controlled via a fixed bromine partial pressure) and temperature. Their properties are discussed in relation to previously investigated iodide perovskites (3D methylammonium lead iodide (MAPbI3) and 2D Dion-Jacobson 1,4-phenylenedimethanammonium lead iodide (PDMA)PbI4). Furthermore, this study highlights the range of the bromine partial pressure in which reversible behavior of the bromide perovskites is observed. Chapter 6 covers the investigation of the phase stability, reversibility and the thermodynamic properties of the 2D Dion-Jacobson mixed halide perovskites ((PDMA)Pb(I0.5Br0.5)4) under light. Temperature dependent absorption measurements are used to identify the photo de-mixed compositions and the photo-miscibility-gap of the mixture. Various experimental methods are developed and compared. Lastly, the effect of encapsulation and light intensity on the shape of the photo-miscibility-gap are also investigated. Chapter 7 covers the kinetic analysis and mechanistic investigation of photo de-mixing in 2D mixed halide perovskites, by analyzing the time dependent in-situ optical absorption and conductivity changes during illumination. The effect of surface encapsulation is discussed. Furthermore, SEM and TEM are utilized to investigate the morphological changes and the nature of the iodide rich and bromide rich phases resulting from phase segregation. Finally, a mechanistic description for photo de-mixing is proposed. Chapter 8 covers the investigation on photo de-mixing in mixed bromide-iodide perovskites with different dimensionality: 3D (MAPb(I0.5Br0.5)3) and nanocrystals (BA- MAPb(I0.5Br0.5)3. Their phase properties under light were investigated by optical absorption and XRD. The simultaneous measurements of conductivities are discussed in terms of understanding the electronic and ionic charge carriers’ properties during photo de-mixing and dark-remixing. Chapter 9 covers the mechanism and driving force of photo de-mixing. In particular, various energetic contributions have been considered to drive photo de-mixing in mixed halide perovskites. Finally, the thesis ends with conclusions (Part Ⅵ, chapter 10) and the appendix section (Ⅶ). This thesis contributes to the understanding of the defect chemistry and ion transport properties of bromide and iodide perovskites, with specific focus on photo de-mixing in mixed halide perovskites. These findings will aid compositional engineering related to halide mixtures to enable optimization of optoelectronic devices as well as the development of other emerging systems exploiting photo-ionic effects. II STATE OF RESEARCH 5 2.1 Halide perovskites and their application The term “Perovskite” was named after the Russian scientist Lev A. Perovski by Gustav Rose,13 who discovered first the perovskite mineral calcium titanate CaTiO3 in 1839. Since then, a variety of oxide perovskites ABO3 were discovered and investigated, stimulating the technological development of sensors, fuel cells, etc.14-17 Halide perovskites, with halogen anion instead of oxygen, exhibit the same crystallographic structure. They have the general formula ABX3 where BX6 halide octahedra form a corner sharing network and A site cations sit in the cavities formed by 12 nearest neighboring X atoms (X = I, Br, Cl) (Figure 2.1). The first successful validation of the crystallization of halide perovskites (CsPbX3, X = I, Br, Cl) was reported by Wells and his collaborators in 1893.18 In 1978, D. Weber19 further successfully replaced A site cation Cs+ with organic (A = MA+, methylammonium ion), opening the era of the organic-inorganic hybrid halide perovskites. Figure 2.1. 3D halide perovskites structure T. Miyasaka et al. were the first to apply these material as “sensitizers” in dye - sensitized photo-electrochemical cells and developed a device with 3.8% solar energy conversion efficiency.20 After this demonstration, the efficiency of solar cell using halide perovskites as photo-absorbers underwent a significant increase. During the last decade, extensive research in this field led to a record efficiency of 25.7%21 in 2022. Several strategies have been used to increase the power conversion efficiency (PCE) of halide perovskite based solar cells. These include altering the intrinsic optical and electronic properties of the perovskite active layer and defect passivation / engineering of the interfaces. For the former, tuning the band gap of the halide perovskite has been a pressing requirement for high performance solar cell devices, with the objective of obtaining compounds with a band gap close to the optimal value of 1.1 eV for single junction solar cells under the ideal case that the radiative recombination is the only recombination mechanism.22 Halide perovskites are also interesting candidates for the development of tandem solar cells.23, 24 Of particular interest is their use as top solar cell in combination with traditional silicon based devices.25, 26 6 2.1.1 Dimensionality engineering Figure 2.2. Halide perovskite structures with different dimensionalities: from 3D to 2D and nanocrystals. As an additional knob for material design, dimensionality control serves as an effective way to tune the electronic and optical properties, further expanding the halide perovskite family. Compounds with reduced dimensionality are particularly promising for optoelectronic applications.27 The electrons' and holes' wave functions are confined within these materials, giving rise to quantum confinement effects and, in many cases, to enhanced emission quantum efficiencies.28-30 By introducing large organic spacer, this confinement can be achieved in 1 dimension (2D halide perovskites, A'BX3) and 3 dimensions (“0D” nanocrystals, A''-ABX3). In both cases large organic cations act as insulating barrier separating semiconducting lead halide sheets (Figure 2.2). Depending on the monovalent and bivalent nature of the organic spacer, 2D perovskites can be further classified as Ruddlesden–Popper (RP) and Dion- Jacobson (DJ) (Figure 2.3). Figure 2.3. 2D Halide perovskite structures: Ruddlesden-Popper(RP) and Dion-Jacobson (DJ). 7 2.1.2 Halide engineering and mixed halide perovskites Halide engineering (mixing and substitution) is used as a facile way for tuning the bandgap of the material. Noh and co-workers have demonstrated the band gap increase linearly with the increase of the Br content. By changing I/Br mixing ratio in MAPb(I1−xBrx)3, (MA = methyl ammonium), almost the entire visible spectrum can be covered.31 Eperson et al replaced methylammonium cation(MA+) with formamidinium(FA+) cation, by which a tunable but also slightly red shifted bandgap range between 1.48 and 2.23 eV was achieved.32 Halide engineering in low dimentional (quasi 2D and nanocrystal system shown in Figure 2.2) further expand this range to higher band gap energies33-35. In addition to stoichiometry synthesis, 31-33, 35-39, various other methods, such as halogen gas exchange,40, 41 halide anion exchange via halide salts,34, 42, 43 dihalomethane44 or with physical contact of the two films with pure perovskites are also used for halide mixing.45-47 Figure 2.4 summarizes the tunable optical properties achieved by halide engineering and dimensionality control. Figure 2.4. Tunable optical properties achieved by halide engineering and dimensionality control. Data extracted and replotted with permission, 3D: MAPb(I1-xBrx)3 from Ref31, the energy here refers to the bang gap energy extracted from Tauc plot of the UV-Vis absorption spectra; 2D, n=1: PEAPb(I1-xBrx)4, PEAPb(Br1-xClx)4 from Ref33, the energy refers to the low energy excitonic absorption energy ; NC: CsPb(I1-xBrx)3 from Ref 34 ; 3D: FAPb(I1-xBrx)3 from Ref 32,the energy here refers to the bang gap energy extracted from Tauc plot of the UV-Vis absorption spectra; 2D,n=1 and n=2: L2Pb(I1-xBrx)4, L2Pb(Br1-xClx)4 from Ref35 8 2.2 Defect chemistry and ion transport in halide perovskites While the mixed ionic-electronic conducting properties of fully inorganic halide perovskites have been investigated as early as 1983,48 hybrid halide perovskites, such as MAPbI3, used in solar cells have initially been treated as traditional electronic semiconductors. In the early years of perovskite solar cells, an anomalous hysteresis behavior in the current– voltage curves was reported for these devices.49 This hysteresis refers to the fact that the current – voltage curve is significantly different when voltage is scanned in either the forward or the reverse direction (Figure 2.5). This effect was found to depend on scan rate, measurement delay time, solar cell architecture/contact.49-51 In addition, a very large capacitance at low frequency from impedance measurements was also observed.52-54 The explanation for these observations became clearer only after it was found that, unlike other purely electronic semiconductors, ion conduction in halide perovskites is significant even at room temperatures.55 Figure 2.5. I–V hysteresis of perovskite solar cell and their Frequency (f)-dependent capacitance (C) properties. TiO2/MAPbI3/spiro-MeOTAD configuration showed significant I– V hysteresis along with highest capacitance (10–2 F/cm2), see black curve. Substantial reduction in capacitance to 10–3 F/cm2 was observed upon replacing TiO2 with PCBM (Red). Figure adapted with permission from Ref 50. Yang et al.55 showed for the first time that MAPbI3 is a mixed ionic-electronic conductor, with 4 times higher ionic conductivity compared with its electronic conductivity in the dark. The transient voltage curve measured with a galvanostatic DC polarization experiment fits a square root of time law at short times and an exponential behaviour for longer time, suggesting that the observed long time scale behaviour is due to stoichiometric polarization (Figure 2.6a) with chemical diffusion coefficient of 2.4 E-08 cm2 s-1. Further EMF measurements (Figure 2.6b), using MAPbI3 as electrolyte under an iodine chemical potential gradient, show substantial open circuit voltage, further indicating that MAPbI3 is a mixed conductor.55, 56 Doping experiments57 and atomic modelling58 revealed that iodine vacancies are the majority ionic charge carriers. The dominant moving ion in this compound is I-, as confirmed via Faradaic reaction cell experiments (Figure 2.6c), where formation of the metal iodide at the metal/MAPbI3 contact is observed.59 Solid state NMR measurements of 14N, 1H 9 and 207Pb in MAPbI3 rule out significant MA long range motion at least in its tetragonal phase (T < 327 K). Figure 2.6. (a) DC polarization curve for a C|MAPbI3|C cell measured by applying a current of 2 nA. Voltage versus square root of time and semi-log plot of voltage versus time at longer time scale was plotted on the right. Figure adapted with permission from Ref 55. (b) Voltage of an emf cell obtained applying two different P(I2) (1.7 and 1.5 bar). MAPbI3 pellet kept at 403 K. Inset shows the schematics of the experiment. (c) Schematic of a faradaic reaction cell (+Cu|MAPbI3|AgI|Ag-) and XRD patterns of surfaces A and B of Cu foil after the measurement. Figure adapted with permission from Ref56. 2.3 Photo induced ionic effects in halide perovskites Kim et al. showed that with above bandgap illumination, the electrical response of MAPbI3 is consistent with a significant increase in not only the electronic but also the ionic conductivity (two orders of magnitude in σion and three orders of magnitude in σeon, even at a light intensity of only 1 mW cm–2).8 Such light induced apparent enhancement in the ionic conductivity increases with the increase of the light intensity (Figure2.7a). In addition, such enhancement in ionic conductivity is further confirmed in a permeation cell when applying p(I2) on one side of the MAPbI3 film as I2 source, while using Cu on the other side as I2 sink under dark and light conditions (Figure 2.7b). Compared to the dark situation where the CuI signal is almost at noise level, clear formation of CuI was found at the illumination side. This indicates that transport of iodine flux from one side to another side of MAPbI3 is enhanced by light. The ionic conductivities under light were also measured in MAPbBr3. Strikingly, a 10 different behaviour was observed, whereby in MAPbBr3 little or no ionic conductivity enhancement under light is detected.60 This difference is invoked below in the discussion of the driving force for photo de-mixing in mixed halide perovskites (See Chapter 9). Figure 2.7. (a) Electronic and apparent ionic conductivities under different light intensities (extracted fromd.c. galvanostatic polarization at 40 °C under an Ar atmosphere). Figure adapted with permission from Ref8. (b) Permeation cell in the dark and under light: an iodine flux through a MAPI film is induced by applying a defined iodine partial pressure (3.6 × 10−6 bar) on one side and Cu as the iodine sink on the other side.(c) Electronic and ionic conductivity of in MAPb(I1−xBrx)3 in the dark and under light with light intensity of 1 mW cm−2. σion,light refers to apparent values calculated from the polarization measurements. 2.4 De-mixing and photo de-mixing De-mixing refers to the process by which two or more components or phases are separated when they are mixed together. This can occur in a variety of contexts, including in chemical mixtures, and particulate systems.61, 62 In terms of chemical mixtures, de-mixing has been extensively investigated in multicomponent systems including polymer blend, metallic alloys and ionic solid mixtures.61, 63-65 According to thermodynamics, the stability of the phases relative to one another is determined by the free enthalpy of the phases involved.66, 67 The free enthalpy is result of the counter-play between enthalpy and entropy of mixing, which depends on pressure (P), temperature (T) and composition(x) (see detailed introduction in Part Ⅲ). Recently, Hoke et al. 68reported another de-mixing phenomenon in mixed halide perovskites, whose occurrence depends on the ON/OFF state of the light that shines on the 11 material, suggesting that light can potentially be another important variable for changing the thermodynamic properties of the mixture. This finding indicates that there might be a link between light and phase thermodynamic properties, which presents great opportunities for studying the out-of-equilibrium thermodynamics with photo-generated charge carriers present in the system.11 Specifically, mixed halide perovskites such as MAPb(IxBr1-x)3 de-mix into I-rich and Br- rich phases under illumination (photo de-mixing) and re-mix to the initial composition in the dark (dark re-mixing). Anionic mixtures with different A-site cation materials CsPb(BrxI1-x)3 43, 69, 70 and FAPb(BrxI1-x)3 71-73 have also shown a similar de-mixing behavior. Halide perovskites with multiple mixed A cations and anions MAzCsyFA(1-y-z)Pb(BrxI(1-x))74-76 show a lower tendency to phase segregate. No segregation of cations are observed. Furthermore, de-mixing in these mixtures was found not only to be induced by light but also by voltage bias77 and high dose electron beam78. 2.4.1 Experimental signature of photo de-mixing Experimentally, the occurrence of photo de-mixing was observed in photoluminescence(PL) emission at low photon energy upon illuminating the mixed halide perovskite accompanied by a simultaneous decrease in intensity of the high photon energy feature, attributed to the bandgap of the mixture (Figure 2.8 a).11, 12 In addition, a recent study from Eva Unger’s group shows the complexity of the evolution in the photoluminescence during phase segregation. A short-lived intense red-shifted band which is close to the pure iodide perovskite’s emission formed within the first few seconds of light exposure, then disappeared followed by emission of I-rich phase with higher energy.79 UV-Vis absorption measurements shown in Figure 2.8b indicate the flattening of the absorption edges after de- mixing, consistent with the formation of phases presenting larger and smaller absorption edge energies. From a structural point of view, XRD measurements show the splitting of a single diffraction peak into two peaks at different diffraction angles indicating Br- rich and I-rich domains (Figure 2.8c);11 Other studies in the literature also report the broadening of these diffraction peaks, evidencing a decrease in crystallinity in the material on de-mixing.80 Electroluminescence mapping, combined with SEM have also aided the investigation of the phase separation process in terms of spatial distribution of the de-mixed phases (Figure 2.8d)81, 82 Additionally, the long stabilization time of optoelectronic devices using mixed halide compositions compared with its end members has been interpreted as an “electrical signature” for photo de-mixing (Figure 2.8e).68, 83, 84 Various reviews on this topic can be found in the literature.80, 85-90 12 Figure 2.8. Key experimental signature for photo de-mixing based on different techniques: (a) Change in photoluminescence of MAPb(I0.5Br0.5)3 under λexc = 405 nm CW excitation (Iexc =20 mWcm−2, 3s); Inset show the emission spectra from 475 nm to 600 nm, indicating the formation of the Br-rich domains. Figure adapted with permission from Ref91. (b) Change in absorption intensity of MAPb(I0.5Br0.5)3 under λexc = 405 nm CW excitation (Iexc =20 mWcm−2 time sequence: blue, 0 min; red, 1min; and green 30 min). Figure adapted with permission from Ref91; (c) X-ray diffraction of the 200 XRD peak of an MAPb(I0.4Br0.6)3 film before (black) and after (red) under white-light LED(∼50 mW cm−2 , 5min). Figure adapted with permission from11 (d) Scanning electron microscope (SEM) and cathodoluminescence (CL) imaging overlay under at 405 nm LED (100 mW/cm2, 5min). The scale bar is 2 μm. Figure adapted with permission from Ref81. (e) Stabilization times of anion mixtures MAPb(I1-xBrx)3 in the dark (open symbols) and under light (closed symbols with hollow). Figure adapted with permission from Ref60. 2.4.2 Photo de-mixing and device performance Photo de-mixing causes an undesirable drop in power conversion efficiency of solar cells based on mixed halide perovskites under operation.68 Early work on this subject shows that, within a period as short as fifteen minutes under continuous visible light, both open circuit voltage (OCV) and short circuit current (SCI) of a mixed halide perovskite based solar cell decrease significantly. As a result, the photo-conversion efficiency of the device decreases (Figure 2.9a).68, 84 By simultaneously recording changes in photoluminescence and solar cell performance under prolonged illumination, it was also demonstrated that, compared with the open-circuit voltage, the loss in short-circuit current density shows a more significant drop, with concurrent red shift, increased lifetime, and higher quantum yield of photoluminescence.92 This can be attributed to the I-rich phase forming in the active layer, which favors charge carrier trapping and recombination. The funneling of electronic charge carriers into the I-rich domains93 is enabled by the large diffusion length of the carriers94 and the favorable valence band offset between the mixture and the I-rich phase (Figure 2.9b)95. 13 Figure 2.9. (a) Left: Schematic illustration of the cell design employed for evaluating photovoltaic performance. Center and Right: J–V characteristics of PSCs with MAPb(Br1.5I1.5)3 films under 1 Sun continuous illumination for 25 min. Figure adapted with permission from Ref84. 2.4.3 Mitigation strategies A non-uniform distribution of the halides within the perovskite layer due to photo de- mixing can significantly hinder the potential use of mixed-halide perovskites in tandem solar cells and adjustable LED technology. There have been several reported methods in the literature that aim to reduce the segregation of halides in mixed-halide perovskite materials, potentially overcoming this limitation. Stoichiometric engineering: Partially replacing A cation of methylammonium (MA, CH3NH3 +) with caesium (Cs+) or formamidinium (FA, CH(NH2)2 +) show decrease in de-mixing rate.75, 96-98Substituting MA+ with FA+ shows a decrease in the rate constant of de-mixing on increasing of the FA+ content. Significant suppression of such rate constant was observed in the case of 90% substitution.96 Substituting MA+ with Cs+ suppresses phase separation to some extent, although detection of I-rich clusters was confirmed in this case too.97 Substituting partially or completely MA cation with a mixture of FA and Cs cations also shows a less prominent photo de-mixing, but it does not completely prevent it. The I-rich phases for the mixed (FACs)-perovskites have a similar emission peak as of the MA-perovskites, centered around 740−750 nm.75. Cs incorporation in (FA0.83MA0.17)Pb(I0.83Br0.17)3 greatly enhances the overall performance of the solar cell99.100, 101 Dang et al. reported that the addition of Cs+ and Rb+ could drastically suppress the segregation of halides.98 14 Crystallinity control: Rehman et al. demonstrated that the CsyFA(1−y)Pb(Br0.4I0.6)3 thin films with increased crystallinity (indicated by the width of X-ray diffraction peaks) show improved photostability against photo de-mixing.74, 77 The grains size of the mixed halide perovskites has also been shown to influence the occurrence of photo de-mixing. CsPb(I0.5Br0.5)3 nanocrystal-based films show no redshift of the PL peak under illumination, indicating the absence of photo de-mixing in this system.91 According to Zhang et al., CsPbBr1.2I1.8 nanocrystal with a lateral size of ~23 nm shows no de-mixing. 102 Andrés et al. demonstrated the gradual transition from de-mixing free to de-mixing by gradual increase the crystallites’ sizes in CsPbBr1.56I1.44 nanocube thin films. The threshold grain size of the occurrence of photo de-mixing is 46 ± 7 nm.103 Similar conclusion have been drawn by Hu et al. who showed that a threshold size of phase segregation of approximately 43 nm is relevant to CsPbBr1.5I1.5 nanocrystalline films. Consistently, the device performance of solar cells with crystal sizes lower than the threshold size is stable, with the highest PCE achieved at 150 °C annealed films with an average grain size of 106 nm.104 Other reports have commented on the beneficial effects of an increased grain size in suppressing photo de-mixing.105, 106 Hu et al. showed that MAPb(Br0.27I0.73)3 perovskite films deposited on poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA, hole-transport layer, HTL) is more stable against photo de-mixing compared with the same film deposited on poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS). The authors ascribe this effect to the larger grains formed on PTAA when compared to PEDOT:PSS.105 This study is complicated by the unclear role of the hole transport layer, which influences the charge carrier density in the film (see below). Charge carrier manipulation: The manipulation of the charge carrier recombination, extraction and injection rates is found to influence the occurrence of photo de-mixing. Hole accumulation achieved by contacting mixed halide perovskites with electron transport layers107, 108 and through electrochemical anodic bias109 also provide direct experimental evidence for the important role of the hole concentration on the photo de-mixing behaviour in 3D mixed halide perovskites. Belisle et al. demonstrated the effect of selective electron/hole extraction by depositing mixed halide perovskites on C60 (eletron transporting layer, ETL)/ poly (triaryl amine) (PTAA, hole-transport layer, HTL). The initial PL of the perovskite is greatly quenched due to its contact with the transport layers. Strikingly, while the perovskite capped with the ETL shows substantial PL from a lower-bandgap phase, the PL of the HTL capped perovskite is largely unaffected.108 Similarly, Dubose et al. discovered that segregation is observed when mixed halide perovskites deposited on an electron transporting layer such as TiO2 or insulating ZrO2 substrate. By depositing hole transporting layer spiro-OMeTAD on mixed halide perovskites/TiO2 film, the halide ion segregation is suppressed.107 These results 15 point towards the importance of the photo-generated holes in the mixed halide perovskite layer and indicate that efficient hole extraction greatly suppresses the occurrence of photo de-mixing. Defect passivation: Stranks et al. proposed mitigating photo-induced phase segregation by decorating the surfaces and grain boundaries with passivating potassium halide (e.g. KI) layers.110 They proposed that the excess iodide ions from KI could be able to fill the iodine vacancies in the material and thus suppress halide migration. Belisle et al. showed that by coating a CH3NH3PbI2Br surface with the electron-donating ligand trioctylphosphine oxide (TOPO)108, 111, the rate of halide segregation as monitored by PL of the lower-bandgap iodide- rich domains in the mixed-halide perovskite, is substantially reduced. Based on a similar idea, Kamat et al. treated a perovskites solution with I2 solution in IPA before spin coating the film to reduce the iodine vacancy concentration. This approach resulted in the mitigation of segregation under light with quick recovery in the dark.84 These results suggest that carrier trapping and charge accumulation at perovskite surfaces are important components determining the driving force of photo induced halide segregation, and efficiently passivating and treating surfaces are potential pathways towards stabilized wide-bandgap perovskites.108 2.4.4 Photo de-mixing models The mechanism and driving force for photo de-mixing are still under debate. Here is a summary of the models that have been proposed to explain this phenomenon and of how each of the, is (in)consistent with experimental observations: i) Dark miscibility gap model (Brivio et al.)112 Brivio et al. 112 performed a thermodynamic analysis of MAPb(BrxI1-x)3 alloys with DFT calculation. Their work shows a large miscibility gap in the intermediate composition range, predicting the spinodal (0.3 < x < 0.6 at 300 K) and bimodal (x =0.2, 0.7 at 300 K) decomposition regions and a critical temperature of 343K (Figure 2.10a). This model proposes that, thermodynamically, the studied mixture is metastable in the dark and that light simply catalyzes the de-mixing into I-rich and Br-rich phases. In other words, it is the de-mixing kinetics which under light is accelerated, enabling a faster occurrence of de-mixing, which would occur under dark over very long time scales. This model can explain the independence of the de-mixed I-rich composition after illumination (measured by PL peak) on the initial composition.11, 113 However, the fact that de- mixing process occurs reversibly, at least partially indicates that the miscibility gap must be influenced by light thermodynamically.11, 12 ii) Bandgap reduction (Draguta et al.) 91, 114 Draguta et al. proposed a model whereby the valence band edge difference between the mixture and the demixed phases introduces an energetically favourable path for photogenerated holes. This energy offset counterplays the configurational entropy of mixing 16 and serves as the driving force of de-mixing. The mixing free energy is calculated between a photoexcited hole in MAPb(I1−x Brx)3 and a hole localized in an iodide-rich phase(Figure 2.10c).91, 115, 116 ΔF∗(x,n)=n(𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚∗ (𝑥𝑥) -xFBr−𝐹𝐹𝐼𝐼∗)=nΔFGS(x)+ΔEg(x) Eq.2.4.1 𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚∗ (𝑥𝑥), FBr, and 𝐹𝐹𝐼𝐼∗denote photoexcited MAPb(I1−x Brx)3, MAPbBr3, and MAPbI3 free energies and asterisks denote the presence of a photogenerated hole. ΔEg(x) denotes the band gap differences between the MAPb(I1−x Brx )3 and MAPbI3. According to their calculations, de-mixing is favored thermodynamically and will result in de-mixed compositions xI-rich ∼ 0. Experimentally, the xI-rich is estimated to be ∼ 0.18− 0.2.114 The authors argue that kinetic limitations and, specifically, a temperature-independent percolation threshold leads to finite xI-rich values. They support their argument by the observation of a limited redshifts ∼10 meV of the emission arising from parent mixed halide perovskites when de-mixed at a low temperature range between 295 and 185 K. However, this model leaves the observed strong temperature dependence of the threshold light intensity unexplained.117 iii) Polaron and strain (Bischak et al.)81, 97, 118 Bischak et al. proposed another thermodynamic model where they considered strain energy as an important component determining miscibility of the iodide and bromide perovskites. Concentrating on the interaction between photo generated electronic charges and the ionic lattice through electron−phonon coupling, they predicted de-mixing based on the standard solid state solution phase transition theory(Figure 2.10b).81 Molecular simulations suggest polaronic strain locally changes the free energy of halide mixing and favors de-mixing, leading to stabilized regions enriched in Iodide. Compared with chemical interaction, the relative energetic contributions to the heat of mixing due to elastic effects from lattice strain are much larger, which leads to a de-mixing transition that depends strongly on strain. This model is consistent with the observation that reducing the electron−phonon coupling in the system by replacing MA+ with less polar Cs+ reduces the tendency to phase separation.97 This model rationalizes the observation that the small clusters enriched in I-rich localize near grain boundaries (in polycrystalline films)81 or of the edges of the crystals(in single crystals)2 by releasing strain. The formation of the I-rich domains is stochastic in space is also consistent with the stochastic formation of polarons in space.119 In addition, this model can explain that a continued presence of photo generated carriers is required for cluster stability as this is also the necessity for polaron formation.81 Polarons have also been invoked to rationalize re-mixing at high light intensity due to the merged polarons that leads to the reduced strain gradient that favors mixing thermodynamically.120 However, it is still not clear if such “photo re-mixing” is not a 17 illumination induced temperature effect rather than a light induced polaron effect given the high light intensity (200 W cm-2) used (the fact that a controlled sample show photo de- mixing at 70 °C does not exclude temperature effects). Figure 2.10. Photo de-mixing models (a) Dark miscibility gap model(i); (b) polaron and strain model(ii); (c) Bandgap reduction model(iii); (d) Triple phase model(iv); (e) Surface trapping model(v). Figures are adapted with permission. iv) Triple phase model (Chen et al.)121-123 Chen et al. proposed a model considering the minimization of the total free energy expressed as the sum of the compositional and photocarrier free energy (Figure 2.10d). three- phase coexistence is predicted in this model. The phase diagram of the mixture in the dark is calculated by considering the compositional Helmholtz free energy, within the quasi-chemical approximation. Compared with the miscibility gap model in (i), the symmetry lowering by the specific orientation of the MA cations is additionally considered, resulting a Tc of 266 K, compared with the value of 343 K in (i). The energy contribution to the free energy from the photo generated electronic charge carriers are further considered by assuming a Boltzmann distribution of photocarriers on different phases. Eq.2.4.2 where n1 and n2 are the photocarrier densities in the two phases. With ϕ1 and ϕ2 the corresponding volume fractions of the two phases, the mixing free energy ΔF⋆ per formula unit under illumination becomes Eq.2.4.3 18 In steady state, the rate of generation of photocarriers in the system should be equal to the sum of the rates of photocarrier annihilation by monomolecular and bimolecular recombination in the different phases: Eq.2.4.4 Based on these equations, a phase diagram under different light intensities in 5 different mixed halide perovskites was constructed. While spinodals change marginally with increasing the photo carrier density, two different bimodals are predicted, The first type (full blue lines) can be viewed as a modification of the dark binodals by illumination, which are refered to as ‘compositional binodals’. Under illumination, a new type of binodals appears (full green lines), named ‘light induced binodals’. A distinct feature of this model is the prediction that triple points exist where two different phases with different halide composition can be nucleated from the parent phase, including one nearly iodine-pure nucleated phase. Furthermore, two types of three-phase coexistence, with the nearly I-pure nuclei being present in (i) both the I-rich and B-rich phase or (ii) only the I-rich phase are discussed122. Such prediction has not been confirmed experimentally to date. v) Surface trapping model (Belisle et al.)108 Belisle et al. propose a model stating that carrier trapping at surface states induces electric fields which impact the movement or accumulation of ionic defects. Such field drives the occurrence of the photo de-mixing (Figure 2.10e). Surface treatments with electron donating materials reduce non-radiative recombination or charge accumulation, thus inhibiting halide segregation. The model assumes that mobile halide vacancies are uniformly distributed and are charge compensated by the immobile ions or other ionized point defects and considers a density of surface traps that favor electron trapping at the top surface intrinsically. The model proposes the following: (1) trapping of the electrons at the surface, inducing the formation of an electric field pointing towards the surface of the film; (2) Vacancies migrate towards the surface to shield the trap-induced field; (3) redistribution of the ions and precipitation of I-rich and Br-rich phases occurs. The low activation energy for the formation of bromide vacancy compared with iodide vacancy, combined with the drift of the positively charged vacancies towards the surface and the expulsion of bromide ions away from the surface can result in the formation of the I-rich phase. 2.4.5 Open questions 1) Is the origin of photo de-mixing purely thermodynamic in nature or are kinetic effects also involved? Clarifying the (im)miscibility in mixed halide perovskites in the dark (prior to any illumination) is critical for understanding whether photo de-mixing is purely thermodynamically driven or kinetic effects are also involved. Early studies using normal thin 19 film X-ray diffraction and synchrotron X-ray powder diffraction showed conflicting results on whether CH3NH3Pb(I1–xBrx)3 exhibits full miscibility or not.31, 36 Another approach to clarify this question is to investigate whether the process of photo-mixing is completely reversible. The experimental situation is complicated by concomitant photo induced decomposition of the mixture after exposure to light.124 Decoupling photo de-mixing from light induced decomposition is important and necessary, however, it remains a challenging and unresolved question. 2) What are the properties of the miscibility gap in mixed halide perovskites under light? It is still unclear whether the lack of miscibility observed under light for mixed halide perovskites follows the traditional properties of a miscibility gap for mixtures at thermodynamic equilibrium. Among the open question in this context the following are of interest: 1) do the de-mixed compositions vary with temperature and can a critical temperature be defined? 2) Are the de-mixed compositions fixed for any starting composition of the pristine mixture? Nandi et al. observed that temperature affects the occurrence of photo de-mixing in MAPb(I1−xBrx)3 (x = 0.15, 0.24, 0.27) differently with different initial composition, where only x=0.15 remains mixed in the temperature range of 77 K – 300 K. For x = 0.24 and 0.27, photo de-mixing only occurs in the intermediate temperature window (200 – 300K).125 This result implies the possibility to measure such miscibility gap under light. Very small variations in the de-mixed compositions were also observed in MAPb(I0.5Br0.5)3 between 185K and 295K.114 Therefore, it remains unclear if temperature influences the de-mixed compositions, and what is the shape of such miscibility gap under light. In addition, methods that allows one to measure such miscibility gap are currently missing, also due to the complication from photo- induced decomposition. 3) What is the driving force for photo de-mixing? Concerning the driving force inducing phase instability in mixed halides perovskites, strain effects and electronic effects are discussed in the literature: a)The interaction between photo generated charges and the ionic lattice through electron−phonon coupling induces strain effects that changes the free energy of de-mixing.81, 97, 118 b)Valence band edge variation and energy gain associated with holes stabilization in I-rich domains influences the free energy of de-mixing.12, 91 Fundamentally, the fact that photo de-mixing occurs, indicates that such process involves the formation of (electronic or also ionic) defects and their transport, as well as the formation of new phases. The contribution of ionic effects on the driving force of the photo de- mixing is missing. In addition, what is the dimensionality effect on such contribution? 20 4) Why do MAPbBr3 and MAPbI3 show significant differences in ionic conductivities under light? MAPbI3 shows an enormous increase in the apparent ionic conductivity under light (two orders of magnitude in σion even at a light intensities of only 1 mW cm–2).8 Under the same conditions, MAPbBr3 shows little or no ionic conductivity enhancement.60 This effect can potentially influence how halide ions are differently distributed and therefore contribute to the driving force of the photo de-mixing. The reason MAPbBr3 and MAPbI3 show significant difference on ionic conductivities under light remains unclear. In addition, other than electrical measurements, which are indirect in nature, can these effects be probed directly, e.g. via spectroscopy techniques? 5) What are the differences occur in the defect chemistry of bromide and iodide perovskites? The defect chemistry of MAPbI3 has been extensively investigated,55-57, 126 revealing iodide vacancies to be the majority ionic charge carrier for ion conduction and electron holes for electronic conduction. By substituting iodide with bromide, whether such behavior varies significanly or remains largely unvaried is currently an unanswered question. Given the difference on ionic conductivities under light mentioned in last point, such investigation is even more crucial. III THEORETICAL FRAMEWORK 23 3.1 Mixture thermodynamics This section starts with a brief description of the thermodynamics of mixing and de- mixing (for further reading one finds treatments in the literature, e.g. Guggenheim’s book127). Let us consider a binary mixture C≡(A, B) that can be thought to be formed according to 𝑥𝑥A + (1 − 𝑥𝑥)𝐵𝐵 ⇌ 𝐶𝐶(𝑥𝑥) Eq. 3.1 For the discussion on mixed halide perovskites, it could be MAPb(I1-xBrx)3 ≡(MAPbBr3， MAPbI3). Figure 3.1. Mixing free enthalpy (ΔmG), mixing enthalpy (ΔmH) and mixing entropy term (TΔmS) for mixture AxB1-x for the case of (a) an ideal mixture, (b) a regular mixture at high temperature (T >Tc) where mixing is favored and (c) a regular mixture at low temperature (T < Tc) where de-mixing is favoured. (d) Free enthalpy of mixing of a sub-regular mixture. The two regions in (d) indicated with green dashed squares refer to the (I) spinodal and the (II) binodal regions. By constructing a common tangent to the free enthalpy curve (orange solid line), the de-mixed compositions of x1 and x2 can be obtained. The contributions of the enthalpy (∆mH) and entropy (∆mS) of mixing to the free enthalpy (∆mG) determine if mixing or de-mixing is favored. In ideal mixtures (Figure 3.1a), the enthalpy of mixing is zero and mixing is always thermodynamically favorable owing to the contribution of the entropy (if the vibrational contribution to the entropy is neglected). The same applies for a negative mixing enthalpy. If the mixing enthalpy is positive, temperature is a decisive parameter with respect to potential de-mixing. At sufficiently high temperatures, mixing is favored (Figure 3.1b). At lower temperatures, where the mixing entropy does not dominate, the free enthalpy of mixing presents a double-minimum profile (Figure 3.1c). Such curve will look rather symmetrical (regular mixture) if A and B are similar. Figure 3.1d shows the general case where A and B are dissimilar. In either case, the double minimum profile 24 identifies a region of compositions, referred to as miscibility gap, where the mixed situation is not thermodynamically stable. A simplified treatment uses a random distribution of A and B atoms for determining the entropy of mixing irrespective of the energetics. Strictly speaking this is only correct for an ideal mixture for which ∆mH = 0. In the general case (∆mH≠0), the energetics may be described by an interaction parameter 𝛺𝛺. Its meaning is obvious when the transition from A and B to C is approximated by resorting to the reaction (a-a)+(b-b)⇌2(a-b) Eq. 3. 2 Here one assumes that A is constituted by a-a pairs, B by b-b pairs and of that C exhibits also mixed pairs (there may be other structure elements that are common to A and B but not varied by the mixing process). Casting the reaction enthalpy of Eq. 2 and hence ΔmH in the form ( ) (1 )mH W x x x∆ = = − Ω Eq. 3.3 the free enthalpy curve follows as ( ln (1 ) ln(1 )) (1 )mG RT x x x x x x∆ = + − − + − Ω Eq. 3.4 The last term disappears for an ideal mixture (Ω=0, i.e., it does not matter whether a is connected with a or b). If there are asymmetrical interactions, but a and b are similar, Ω is non-zero but approximately constant and mG∆ is symmetrical about x=1/2. It is straightforward to show that the critical temperature below which the miscibility gap closes and thus the dome-like form of the free enthalpy of mixing develops into a double minimum form is given by / 2cT R= Ω Eq. 3.5 More realistic approximations assume Ω to be a function of x typically making ( )mH x∆ and ( )mG x∆ asymmetrical (sub-regular mixtures) (Figure 3.1d). For temperatures T< Tc, the coexistence compositions x1 and x2 can be obtained from constructing a double tangent to the ∆mG curve. The fact that it is not the minimum values that correspond to the stable compositions but the points of tangency is due to mass conservation which has to be fulfilled. When considering the kinetics of de-mixing, it is relevant to distinguish between regions of positive and negative curvature. In the middle of the phase diagram (Figure 3.1d, region Ⅰ, region between the two inflection points), the curvature is negative such that an infinitesimal compositional fluctuation leads to a decrease in mG∆ and hence to spontaneous de-mixing (this region is called regime of spinodal decomposition). In the two neighboring regions (Figure 3.1d, region Ⅱ, region between x1 and x2 where the curvature is positive), a sufficiently large variation in composition is needed to drive de-mixing. In other words, decomposition requires 25 nucleation and transport of longer range to bring the system into the thermodynamically favored compositional range. It follows that Tc can be evaluated as the temperature at which the two inflection points coincide (meaning that at the critical composition the third derivative is zero). The above treatment (Eq. 3.3, 3.4) assumes random distribution of atoms, which is inconsistent with assuming a non-zero interaction energy. A better approach is to only assume random distribution of pairs, and to partially account for the non-zero interaction energy in the evaluation of the enthalpy and entropy of mixing by applying a mass action law for Eq.3.1. The latter approach is termed quasi-chemical approximation and results in a more complex, but more accurate, formulation, in which complete randomness is not assumed. Accurate statistical treatments are very involved. A particular shortcoming of the above treatment is its referring to pair potentials which is not sufficient for coulombic systems. A further shortcoming is the neglect of gradient effects. The latter point has been addressed by the Cahn-Hilliard treatment. Here the free enthalpy exhibits, beyond the homogeneous term (no gradients), a correction term being proportional to ∫(∇c)2dV (where ∇c is effectively the gradient in local composition, which is integrated in the volume of the system). In terms of pair considerations, this is essentially referring to the fact that, in a non-zero compositional gradient, the pair interactions to the “left” are different from the pair interactions to the “right”, as already pointed out by Becker128. This theory is more precise in describing the correct transitions between spinodal and binodal behavior. It is also able to predict the influence of the “domain size” in mixtures (see Figure 3.2), whereby an additional contribution to the free enthalpy of mixing due to the surface energy associated with the formation of domain boundaries appears in the balance. Figure 3.2. Schematic of the free energy of a mixture with composition that lies within the miscibility gap for various domain sizes. The influence of domain size is qualitatively shown. (a) Mixed single phase state (atomic “domain size”); (b) de-mixed two-phase state with nanoscopic domain size; (c) de-mixed two-phase state with macroscopic domain size. 26 IV MATERIALS AND METHODS 29 4 Materials and methods 4.1 Materials and synthesis Materials: Lead iodide (PbI2, 99.9985%) and lead bromide (PbBr2, 99.999%) were purchased from Alfa Aesar. Methylammonium iodide (MAI) and Methylammonium bromide (MABr) were synthesized as reported by Im et al.129 Butylamine Hydroiodide (C4H11N·HI, 97%) was purchased from TCI. 1,4-phenylenedimethanammonium iodide ((PDMA)I2) spacer, 1,4-phenylenedimethanammonium bromide ((PDMA)Br2) spacer were synthesized following the procedure reported for the (PDMA)I2 spacer130 and the one described below. Dimethylsulfoxide (DMSO, 99.9%), Dimethylformamide (DMF, 99.8%) and Poly(methyl methacrylate (PMMA, average Mw ~120000) were purchased from Sigma-Aldrich. Chlorobenzene (CB, 99.9%) was purchased in Acros organics. Single crystal sapphire (orientation: (0001), C-plane; single polished and double polished) and quartz (molten isotropic quartz) were purchased from CrysTec. Synthesis of 2D (PDMA)Pb(I1-xBrx)4 precursor solutions: The precursor solutions with Br content of 0%, 10%, 15%, 20%, 30%, 40%, 50% 60%, 70%, 80%, 90% and 100% were prepared following the relative stoichiometry of the halides. Specifically, 0.33 M (PDMA)PbI4 or (PDMA)PbBr4 solutions were prepared by dissolving 0.33 mmol (PDMA)I2 ((PDMA)Br2) and 0.33 mmol PbI2 (PbBr2) in the solvent mixture of DMF and DMSO (3:2, (v:v), 200 μL). The precursor solutions for x = 0.1, 0.15, 0.2, 0.3, 0.4, 0.5 were prepared by dissolving (PDMA)I2, PbI2, and PbBr2 with stoichiometry of 1: (1-2x): 2x. For x = 0.6, 0.7, 0.8, 0.9, (PDMA)Br2, PbI2, and PbBr2 with stoichiometry of 1: (2-2x): (2x-1) were used. Preparation of 2D (PDMA)Pb(I1-xBrx)4 films: The film preparation procedure was conducted in Ar-filled glovebox with controlled atmosphere (O2 and H2O < 0.1 ppm). The (PDMA)Pb(I1-xBrx)4 films were deposited on quartz substrate by spin coating the precursor solution at 9 rps and 66 rps for 2s and 48s, respectively. The films were annealed at 150 °C for 10 minutes. A PMMA encapsulation layer was deposited on perovskite films, unless stated otherwise, by drop casting a PMMA solution (10 mg/ml, dissolved in chlorobenzene). The coated sample was dried at 40 °C for 2 hours in the glovebox. Synthesis of 3D MAPb(BrxI1-x)3 precursor solutions: The mixed iodide-bromide perovskites precursors were prepared by dissolving (1-x) mmol MAI and (1-x) mmol PbI2 and x mmol MABr and x mmol PbBr2 (x=0.0 , 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0) in 1 ml DMSO. After that, the as prepared MAPb(BrxI1-x)3 precursor solution was filtered by using a PTFE filter (0.45 μm，Whatman) for preparation of the thin film. Preparation of 3D MAPb(BrxI1-x)3 film : All procedures in this part were conducted in Ar-filled glovebox with well controlled atmosphere (O2 and H2O < 0.1 ppm). The MAPb(BrxI1-x)3 precusor solution are deposited on the Al2O3 (001) substrate by spin coating 30 the as synthesized precursor solution at 65 rps and 150 rps for 2s and 180s on Al2O3 substrates. During the spin coating, a 650 μL chlorobenzene drop was dropped on the substrate to induce a quick crystallization. Lastly, the films are annealed at 373 K for 2 minutes. PbI2 and PbBr2 thin film are prepared using the same procedure. A PMMA encapsulation layer was deposited on perovskite films for encapsulation experiments by drop casting a PMMA solution (10 mg/ml, dissolved in chlorobenzene). The coated sample was dried at 40 °C for 2 hours in the glovebox. Synthesis of nanocrystalline BA-MAPb(Br0.5I0.5)3 precursor solutions: BA based mixed Iodide and bromide halide nanocrystal perovskites precursor were prepared by dissolving 2 mmol MAI, 0.5 mmol PbI2, 1.5 mmol PbBr2 and 0.4 mmol BAI in 2 ml DMF. Synthesis of nanocrystalline BA-MAPb(Br0.5I0.5)3 thin films: The film preparation procedure was conducted in Ar-filled glovebox with controlled atmosphere (O2 and H2O < 0.1 ppm). The BA-MAPb(Br0.5I0.5)3 nanocrystal thin films were deposited on sapphire substrate by spin coating the precursor solution at 100 rps for 60 s. a 300 μL chlorobenzene drop was dropped at the time of being spinned for 6s to induce a quick crystallization.The films were annealed at 65 °C for 5 minutes. 4.2 Experimental apparatus and techniques Optical microscope measurements: The OLYMPUS DSX510 digital microscope in glovebox is used for imaging the morphology of the thin film before illumination, after illumination and after SEM measurements. Dark field imaging mode is used. SEM measurements: The samples are transferred via vacuum shuttle from argon glove box after optical microscope imaging to a Zeiss Merlin scanning electron microscope, which is used for imaging the change in morphology down to the nanometer scale. Four independent electron detectors are used to obtain comprehensive information of the perovskites film: 1) Chamber secondary electron detector (Everhardt-Thornley), 2) In-lens secondary electron (SE) detector (available for < 20 keV), 3) In-lens backscattered electron (BSE) detector (available for < 20keV), 4) 4- segment low-angle scattered electron detector. Accelerating voltage of 1.5 kV was used for the measurement. Focus Ion Beam (FIB) Processing: Focused Ga Ion Beam processing was used to prepare thin lamella for TEM investigation. Beam-parameter for preparation the blank (cutting of the lamella from substrate): 30kV10nA. Beam-parameter for preparation the rough lamella: 30kV 2nA. Final polishing with 30kV 200pA/ and 30kV 50 pA. TEM-EDX measurements: 31 The TEM-EDX measurements are conducted with JEOL ARM 200CF: Scanning transmission electron microscope equipped with a cold field emission electron source, a DCOR probe corrector (CEOS GmbH), a 100 mm2 JEOL Centurio EDX Detector, and a Gatan GIF Quantum ERS electron energy-loss spectrometer. Ultraviolet–Visible (UV-Vis) spectroscopy experiment under controlled conditions: UV-Vis experiments were conducted using a Shimadzu UV-2600 spectrometer. The setup was modified to achieve control of the sample temperature (1 °C accuracy) and atmosphere (< 30 ppm O2). A LED lamp (MCWHF2, Thorlabs) coupled to an optical fiber was used for the illumination during the experiments. The light intensity was measured directly at the exit of the optical fiber using a thermal power meter (PM100 D, Thorlabs). Figure 4.1. Schematic of UV-Vis set up under controlled condition. To estimate the exact light intensity shined on the sample during the experiments, we conducted the calibration as follows (Figure 4.2). I = �𝐷𝐷12 � 2 �𝑎𝑎∗tan (𝛼𝛼+𝛽𝛽)−𝑎𝑎∗sin (𝛼𝛼−𝛽𝛽1) 2 � 2 I0. Eq. 4.1 where: I0: incident light intensity measured directly after the optical fiber; I: calibrated light intensity hitting the sample; D1: diameter of the optical fiber; a: vertical distance of the optical fiber and the sample; α: Acceptance angel of the optical fiber ( 22.95° in this case); β: illumination angel of the optical fiber ( 45° in this case); 32 Figure 4.2. Calibration of light intensity Modulation of the bias light to allow for UV-Vis measurements was performed to show in Figure 4.3. For each cycle, light was switched on for 900 s followed by a 300 s dark window, during which the UV-Vis measurements were taken. Figure 4.3. Representation of the bias light modulation during photo de-mixing experiments (three complete cycles per hour). The black dashed lines indicate the times where the UV-Vis measurements were taken. The blue, yellow and purple data points correspond to the changes in absorbance evaluated by means of the UV-Vis measurements. X-Ray Diffraction (XRD): All the XRD patterns were acquired by a PANalytical diffractometer of Empyrean Series 2 (Cu Kα radiation, 40 kV, 40 mA) equipped with a parallel beam mirror and a PIXcel 3D detector. All measurements were conducted in Bragg-Brentano or grazing incidence configuration modules with programmable divergence slits. The Anti scatter slits were used and recorded with a PIXcel 3D detector. The samples that were not encapsulated were mounted in a polycarbonate domed sample holder for protection from ambient atmosphere during the measurements (Panalytical). ELECTRICAL MEASUREMENTS D.C. galvanostatic polarisation measurements were carried out with a Source Meter(Keithley 6430). A.C. impedance spectroscopy data were acquired with Novocontrol Alpha A with electrochemical interface. The impedance data are analyzed in Z view. Two 33 types of different in-house built measurmenet cells for chaterizing the electrical properties of the materials have been set up. Measurement cells for different p(Br2) measurements (Relevant to Chapter 5): Br2 gas have high vapor pressure. The temperature at which I2 and Br2 present vapor pressure of 10-7 bar is -50 °C and -130 °C, respectively.131 To obtain low partial pressure suitable for Br2 partial pressure dependent electrical measurements in Chapter 5, low temperature and gas diluting strategies are used. For the former, a home-made cooling system was implemented. A spiral-shaped cooler element (Thermo Scientific EK 45) immersed in a coolant is used to obtain a temperature of -80 °C. To improve the temperature stability problem, a low power heating element is applied to the system, obtaining a temperature range of 80 ± 1°C. For the latter, different ratios of Ar carrier gas containing Br2 and pure Ar gas are mixed to gain an extra 2 order of magnitude decrease of the Br2 partial pressure (See Figure 4.4 for the design of the Br2 set up). Via these two strategies, Br2 partial pressure range of 1×10-6 bar to 1×10-4 bar can be obtained. One should also note it is important to a constant flow of the gas mixture of 100 SCCM used in the measurments cell is maintained for all experiments. Figure 4.4. Schematic for the Br2 set up. The gas mixture with 5SCCM Br2 and 95 SCCM Ar was used to get a p2(Br2) of 5×10-6 bar. Br2 gas also has very high reactivity even at very low partial pressure, which makes corrosion of all other materials being exposed to it a critical issure. This presents challenges for desigining the apporiate cells for this measurement. Figure 4.5 show the the design of such measurement cell with good control of the gas atmosphere and temperature. The main body of the measurement cell (in grey) is constcuted with stainless steel, which had been proved to be resistant to Br2 gas the most. Quartz are used for covering the remaining part of the cell, ensuring a gas tight atmosphere. 34 For electrical measurements, platinum thin wires and copper beryllium coated(addtionally coated with Au) contact are utilized. A thermocouple close to the sample is used to contantly motoring the temperature of the sample. However, even given the best effort to avoid the reaction of Br2 gas with other materials of the cell, possible reaction cannot be fully ruled out. Therefore, it is very important to check the ‘health condition’ of the sample by evaluating reversibility of the measurements. Figure 4.5. Schematic for the measurement cell for Br2 partial pressure measurements. Measurement cell for Photo de-mixing measurements(Relevant to Chapters7&8): The measuement cell that could allow for characterizing optical and electrical propertiers of the materials at controlled temperature, atmosphere and light intensity is shown in Figure 4.6. Figure 4.6. Schematic for the measurement cell for UV-Vis & Electrical measurements Interdigitated electrode for electrical measurements: Interdigitated electrode are used for the eletrical measurement with the electrode geometry shown in Figure 4.7. For 3D halide perovskites, including MAPbBr3 and MAPb(0.5Br0.5)3, eletrode geometry of L = 10 μm, Lc = 5 μm and W = 80 cm is used. For 2D 35 halide perovskites, including (PDMA)PbBr4 and (PDMA)Pb(I0.5Br0.5)4, the eletrode geometry of L= 5 μm, Lc = 5 μm and W = 120 cm is used. Figure 4.7. Schematic for the interdigitated electrode for electrical measurements. L: channel length; Lc: finger width; t: thickness of the Au fingers; W: overlapping lengths of the fingers; N: number of the fingers. 4.3 Evaluation of the partial conductivities Chemical diffusion in ionic compounds refers to a charge neutral ambipolar diffusion of at least two chemically different charge species.132-134 A driving force for such a diffusion can be a voltage difference or a chemical potential difference applied over the material. An internal gradient in stoichiometry is built up over the time, which involves a combination of ionic and electronic transport. Therefore, measuring diffusion process allows us to disclose important information on the mixed ionic – electronic nature of the conductivity in a mixed conductor.132, 135 DC galvanostatic polarization measurements are helpful for such evaluation in time domain, while impedance characterization allows for such analysis in the frequency domain. I now discuss how the partial conductivities are evaluated with taking MAPbBr3 as an example (see Chapter 5) and using the electrode geometry discussed above (horizontal device, Figure 4.7).134, 136, 137 Figure 4.8 shows the DC galvanostatic polarization measurements of MAPbBr3 thin film measured with ionic blocking electrode. Immediately after applying a constant current, both electrons and ions contributes to the conduction according to their transference number. The initial IR drop (voltage increase, U1 = IRtot) corresponds to the total resistance where both electronic and ionic charge carriers are contributing. With increasing of the time, the partial current of the ionic charge carriers decreases until (ion blocking electrode used) it is completely vanished. Once the steady state is reached (U2 = IReon), the current is only contributed by the electronic charge carriers.134, 138 Therefore, ionic and electronic conductivities can be calculated based on ( 1 𝑅𝑅𝑡𝑡𝑡𝑡𝑡𝑡 = 1 𝑅𝑅𝑒𝑒𝑡𝑡𝑒𝑒 + 1 𝑅𝑅𝑖𝑖𝑡𝑡𝑒𝑒 ,𝜎𝜎 = 𝐿𝐿 𝑅𝑅×𝑡𝑡×𝑊𝑊 ). The Rtol extracted from the initial jump in the DC 36 measurement depends on time resolution of the instrument used, therefore, one needs to resort to frequency resolved impedance to obtain accurate Rtol by fitting the high frequency semicircle to the equivalent circuit model discussed below.136 The impedance shown in Figure 4.8b involves different processes that can be modeled based on the equivalent circuit that includes bulk and interfacial processes. 1) in the high frequency regime (1E06 Hz - 400 Hz), equivalent circuit = Par (Reon, Rion, Cbulk), which results in the first semicircle in the Nyquist plot (Par: parallel, Reon: electronic resistance, Rion: ionic resistance, Cbulk: bulk capacitance). The dielectric constant of 280 is calculated based on Cbulk (2.48 E-08 F/cm2). Such value is higher than the typical value of the halide perovskites, indicating that substrate have a dominant contribution to Cbulk. 2) In the mid-frequency range (15 Hz – 0.03 Hz), equivalent circuit = Ser(Par(Rinf, Cinf), Par(Rbulk, Cbulk)) (Ser: series, Rinf: interfacial resistance, Cinf: interfacial capacitance ). In such case, an interfacial capacitance of 1.13E-04 F/cm2 is obtained. 3) At low frequencies (0.03 Hz -0.001 Hz), equivalent circuit = Ser(Par(Rinf, Cinf), Par(Par(Reon, Ser(Rion, Cion)), Cbulk))). In this case, a chemical capacitance Cion = 1.25E-04 F/cm2 is obtained. The electronic resistance obtained from DC polarization and impedance is similar, therefore, the equivalent circuit model mentioned above can well describe such electrochemical response. For situations where the interfacial capacitance at electrode / perovskite interface dominates, the same equivalent circuit can be used. And Cion can be interpreted as an interfacial capacitance.136 Figure 4.8. Electrochemical response of MAPbBr3 thin film measured with ionic blocking electrode (Au, see section 4 for electrode geometry) under p1(Br2) = 1E-06 bar in the dark at 60 °C. (a) DC galvanostatic polarization measurement when applying current (80 pA) for 3000 s (b) Nyquist plot of impedance spectra (frequency range: 1E06 – 1E-03) after reaching equilibration under such bromine partial pressure and the corresponding equivalent circuits. The read dashed line indicate the fitting of the data. V RESULTS AND DISCUSSIONS 39 5 Defect chemistry equilibria in bromide and iodide perovskites 5.1 Introduction Ion migration is a key material property of hybrid halide perovskites.55 It influences the response of devices based on these materials and it is critical for their stability against illumination, humidity, or heat.8, 139-142 Understanding the equilibrium charge carrier chemistry of organic-inorganic halide perovskites is a prerequisite for understanding the ion transport behavior of these materials and finding the potential ways to modify it. The defect chemistry of MAPbI3 has been extensively investigated.55-57, 126 In this compound, the transport of iodide defects is the dominant factor regarding ionic conduction as discussed in Section 2.2. Halide substitution33-35, 45-47 and dimensionality control27, 143-146 serve as effective methods to adjust the electrical and optical properties of the materials, and to optimize the performance of opto-electronic devices. Especially, mixing iodide and bromide perovskites is of interest, due to the favourable band gap range that becomes accessible through this approach, making this strategy relevant to several device applications (See Section 2.1 for detailed introduction). In addition, halide substitution and dimensionality tuning can be used to vary the mixed ionic-electronic conducting properties of hybrid perovskites. Previous reports have shown that moving from iodide to bromide perovskite compositions leads to decrease in electronic conductivity and an increase in ionic conductivity.60 Regarding dimensionality, a number of studies have pointed to a significant decrease in ion conduction in two-dimensional iodide perovskites compared to their three-dimensional counterparts.45, 147 While these preliminary work showed that both iodide and bromide perovskites show generally similar properties, and that dimensionality reduction is a promising strategy to reduce ion conduction, a detailed investigation of the defect chemistry of bromide perovskites, in the 3D or 2D form, is currently missing. In this chapter, I present the investigation of the mixed ionic-electronic conducting properties of 3D methylammonium lead bromide (MAPbBr3) and 2D Dion-Jacobson 1,4- phenylenedimethanammonium lead bromide ((PDMA)PbBr4). Their electronic and ionic conductivities are measured as a function of varying stoichiometry (controlled via a fixed bromine partial pressure) and temperature. The incorporation and ex-coporation kinectics are investigated. Their properties are discussed in relation to iodide perovskites (3D methylammonium lead iodide (MAPbI3) and 2D Dion-Jacobson 1,4- phenylenedimethanammonium lead iodide (PDMA)PbI4). Furthermore, this study highlights the range of the bromine partial pressure in which a reversible behavior of conductivities of bromide perovskites is observed. 40 5.2 Results and discussion 5.2.1 Bromine incorporation / ex-corporation and Defect equilibria in MAPbBr3 The measurements that I will discuss in the following text are conducted using specially designed cells to prevent Br2 corrosion (see Section 4.2 and Figure 4.5, for detailed description). By using such setup, temperature and the gas flow and the components partial pressures are controlled (Figure 4.4), in order to measure the electrochemical properties of the material. Interdigitated gold electrodes that are ion blocking are used for electrical measurements (Figure 4.7, see detail electrode geometry in Section 4.2). In the following discussion, the transport properties of the material are evaluated by both DC galvanostatic polarization and impedance measurements. Figure 5.1. Electrochemical response of MAPbBr3 thin film measured with ionic blocking electrode (Au, see section 4 for electrode geometry) under Ar in the dark at 60 °C. (a) Nyquist plot of impedance spectra (frequency range: 1E06 Hz – 1E-03 Hz) after reaching equilibration under Ar. The top left inset figure show the impedance in very low frequency range (0.1 Hz – 0.001 Hz). The bottom right inset figure show the corresponding equivalent circuits that is used to fit the impedance in frequency range of 1E06 – 0.1 Hz.136 The red dashed line indicates the fitting of using such equivalent circuit to the data. (b) DC galvanostatic measurement: left, polarization (10pA is applied for 10000 s); right, de-polarization (160000s). The time resolution is 1 s. A baseline in voltage of 6 mV is subtracted from the raw data. (c) Short time scale polarization (Voltage vs the square root of time) and (d) long time scale polarization ( -(US - U)/U) vs time). US is a fitted value representing the saturation voltage when steady state of the polarization is reached. Let us start with the electrochemical response of MAPbBr3 thin film in Argon atmosphere (Figure 5.1). Impedance characterization allows one to measure the electrical 41 response in the frequency domain. Figure 5.1a shows that such impedance can be modelled by a bulk contribution at high frequencies followed by a low frequency process. DC galvanostatic polarization measurements are used to obtain such response in the time domain. Especially, from the long time polarization behaviour of the material, one can evaluate the electronic charge carriers’ properties (ion blocking electrode used). Over the time scale range probed with this measurement (1s – 10000 s) at the available resolution (1 s), the response on current application shows firstly a very tiny IR drop (inset, Figure 5.1b) followed by a large polarization with a steady state (not completely reached in this case) at which the ionic current can be assumed to be vanishing.134, 138 This suggests a very small contribution from electronic charge carriers to the total conductivity, indicating that MAPbBr3 is very ionic in nature. After switching off the current, a long de-polarization is observed, with a time constant of 40 ks estimated from the plot in Figure 5.1c. Figure 5.2. Electrochemical response of MAPbBr3 thin film measured with ionic blocking electrode (Au, see section 4 for electrode geometry) under p1(Br2) = 1E-06 bar in the dark at 60 °C. (a) Nyquist plot of impedance spectra (frequency range: 1E06 – 1E-03) after reaching equilibration under such bromine partial pressure and the corresponding equivalent circuits. The read dashed line indicate the fitting of the data. (b) DC galvanostatic polarization me