Design and Optimization of Luminescent Semiconductor Nanocrystals for Optoelectronic Applications

Design and optimization of luminescent semiconductor nanocrystals for optoelectronic applications

Erstellung und Optimierung von lumineszierenden Halbleiter-Nanokristallen für optoelektronische Anwendungen

Der Technischen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg zur Erlangung des Doktorgrades Dr.-Ing.

vorgelegt von Ievgen Levchuk

aus Myhove (Ukraine)

Als Dissertation genehmigt von der Technischen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg Tag der mündlichen Prüfung: 20 Juni 2017

Vorsitzender des Promotionsorgans: Prof. Dr. Reinhard Lerch Gutachter: Prof. Dr. Christoph J. Brabec Prof. Dr. Rainer Hock

To my parents Volodymyr and Halyna Levchuk

iii

Acknowledgments First of all I would like to thank Prof. Dr. Christoph J. Brabec for the great opportunity to perform a Ph.D. in his group. It was a pleasure to work under his supervision, because his always the open-minded look to all of my research, perfect suggestions, new ideas, really great inspiration and guidance, encourage me to make a nice and very interesting scientific work. I highly appreciate that Prof. Brabec always has a time for my results discussion and other topics. My special thanks I would like to address to my group leader Priv.-Doz. Dr. Miroslaw Batentschuk who opened for me beautiful and colorful world of luminescent materials. He always was positive and ready to help me with any issue even beyond scientific life. Most importantly Dr. Batentschuk all the time support my ideas and giving me a lot of scientific freedom resulting in right choice of thesis direction. His role in making of this work is invaluable. I am very grateful to Dr. Andres Osvet for all of his help with optical characterization, generation and implementation of new ideas, fast correction and improvement of my manuscripts as well as for very nice environments in the group and always interesting daily discussion. This thesis would not be complete without successful collaboration. I would like to thank Patrick Herre for almost all TEM, HRTEM and SEM measurements, without which work in nanochemistry is almost impossible. I also would like to acknowledge Prof. Spiecker for his support in electron microscopy. I want to express my gratitude to Prof. Rik Tykwinski and Marco Gruber for their help in NMR spectroscopy and results evaluation. My next thanks I would like to address to Prof. Rainer Hock and his Ph.D. student Marco Brandl for performed XRD measurements. Dr. Christian Würth and Prof. Ute Resch-Genger thank you for the great collaboration and photoluminescence quantum yield measurements. I want to thank also Claudia Kolbeck and Prof. Hans-Peter Steinrück for XPS measurements. I would love to acknowledge i-MEET peoples for the warm atmosphere in the last 4 years. My special thanks going to my officemates Nicola. Cesar, Lili, Carina, Chaohong, Xie, Chao and always part of our screw Shreetu who brought to my scientific life not only great collaborations but also a lot of fun; they encourage me to be a better than I am. I also would like to thank Yi Hou and Gebhard J. Matt for many crazy ideas and fantastic realization thereof. Also, I am grateful to my former master and bachelor students, who not only contributed to my projects but also became my friends. Dr. Anastasiia Solodovnyk, thank you for more than 2 years of great collaboration and your colourful and bright LDS layers. Also my warmest thanks to Liudmyla for her great friendly support.

Special thanks goes to Winfried Habel, Leonid Kuper, Uli Knerr, Claudia Koch and Corina Winkler for their unlimited support in administrative matters and my German language improvement. Especially, I want to thank warmly Prof. Maksym V. Kovalenko, who played a significant role in my decision to become a scientist and to join the group of Prof. Dr. Christoph J. Brabec. I am also thankful to my friend Anatolii Polovitsyn who convinced me to try the life of Ph.D. student and his support in the field of nanochemistry. Most importantly I would like to thank my parents Volodymyr and Halyna as well as my dear sisters Adriana and Natasha for unconditional support during my entire life regardless of the path on which I decided to go. Thousands of times I thank my girlfriend Nathalie for her patience, invaluable support and love. Last but not least, I would like to thank the members of the examination committee for their efforts and time and also to you, the reader, for considering my work

Summary

Luminescent colloidal semiconductor nanocrystals have attracted prominent attention for the last three decades since their size-dependent optical properties were discovered. Numerous applications in fields of light conversion such as light-emitting diodes (LED), photovoltaics, medicine, lasers and TV displays were developed. Despite the strong and rapid expansion of this field in the scope of material quality reflected by narrow size distribution and photoluminescence quantum yield, simplification of the production technology, fabrication in industrial scale and their cost reduction are still under development. Furthermore, reduction of toxicity and design of new Cd-free highly luminescent nanocrystalline materials is the last hurdle before commercial application. The primary goal of this thesis was to develop new cost-effective luminescent colloidal semiconductor materials as spectral convertor with tunable optical properties and application thereof. Being a chemist in and having taken my master’s degree in the field of bulk material growth (CdTe), during my Ph.D. study I delved into several new fields like synthesis and surface modification of colloidal nanocrystals, device processing and photophysics of the semiconductor materials . The thesis is subdivided into two introductory chapters (Chapter 1, Chapter 2), and 4 chapters on scientific results. Chapter 3 describes the large-scale and one-pot synthesis of highly luminescent core-shell ZnCdS:Mn/ZnS colloidal nanocrystals (NCs). The key task of this chapter was designing highly luminescent colloidal nanocrystals with zero reabsorption for down-conversion of UV-blue light to the visible region in Si solar cells. Before starting the successful study on highly luminescent zero-reabsorption ZnCdS:Mn/ZnS NCs, several NC systems such as La-doped ZnO, YVO4:Eu,Bi, Carbon Quantum Dots and highly luminescent PMMA-ZnO NCs were tested. Further study on the synthesis of ZnCdS:Mn/ZnS NC system and the optimization of all the reaction parameters affecting the photoluminescence properties lead to a record photoluminescence quantum yield (PLQY) of 70%. These highly fluorescent nanocrystals were efficiently employed as down shifting layers for the ultraviolet (UV) to yellow wavelength region to improve the efficiency of monocrystalline silicon (mc-Si) solar cells by nearly 0.5 percentage points. The resulting power conversion efficiency (PCE) of conventional solar modules with a 14.6 % energy yield, which are coated with the ZnCdS:Mn/ZnS light converter, will enable a cost reduction for the solar electricity production by 2.1 %. Chapter 4 devoted to the synthesis of hybrid organic-inorganic metal halide perovskite colloidal nanocrystals. In the first part of the chapter, a simple ligand-assisted re- precipitation approach allows to fabricate highly luminescent (PLQY 30-90%) and nearly

vi monodisperse CH3NH3PbX3 (X=Br, I) colloidal nanoplatelets with tailored thicknesses. Broadly tunable emission wavelengths (450–730 nm) are achieved via the pronounced quantum size effect without anion–halide mixing. However, pure chemical and colloidal stability was a main motivation to find a more stable analog. Therefore, in the second part of this chapter, we employed the same approach to synthesize novel formamidinium

(CH(NH3)2PbX3, X=Cl, Br, I) based perovskite nanocrystals, which is an analog to + methylammonium (CH3NH3 ) perovskite. The cubic and platelet-like nanocrystals with their emission in the range of 415-740 nm, full width at half maximum (FWHM) of 20-44 nm and lifetimes of 5−166 ns, enable band gap tuning by halide composition as well as by their thickness tailoring; they have a high photoluminescence quantum yield (up to 85%), colloidal and thermodynamic stability. Further optoelectronic measurements verify that the photodetector based on FAPbI3 nanocrystals paves the way for perovskite quantum dot photovoltaics.

Despite of high photovoltaic performance of the bulk perovskites, a FAPbI3 nanocrystalline photodetector prototype has shown low photoresponse mainly due to the existence of insulating long-chain organic ligands on the NCs surface. Therefore, in Chapter 5 a new ligand-free and shape controlled synthesis of submicron perovskite

CH3NH3PbX3 (X=Br, I) crystallites and their mixed-halide analogs was designed. This allowed fabricating stable perovskite inks for the large-scale printing. Photodetector devices bladed out of this ink have shown remarkably high photoresponse, indicating the possibility of precursor-free and large area perovskite solar cell module manufacturing. Chapter 6 is devoted to the study of significant importance of the chemical purity of the perovskite precursors and their impact on the semiconductor quality. The findings presented in this chapter show that certain amount of impurities formed during the CH3NH3I synthesis promote PbHPO3 nanoparticle formation in the perovskite precursor. These particles play the role of seeds for high quality large grain growth, which led finally to high efficiency of the solar cells. This study demonstrates that the reproducibility problems, which are observed and discussed in the perovskite solar cell community, can be mainly attributed to the unexpected impurities in CH3NH3I. Therefore, purity control for the self-synthesized or commercially available CH3NH3I are required. In the last Chapter 7, conclusions as well as outlook for further investigations on the topic of the thesis are presented.

vii

Zusammenfassung Durch die Entdeckung der Größenabhängigkeit der optischen Eigenschaften von kolloidalen Halbleiter-Nanokristallen hält das Interesse zur Entwicklung und Untersuchung von lumineszierenden Materialen dieser Klasse in den letzten drei Jahrzehnten ununterbrochen an. Kristalline Nanopartikel für zahlreiche Anwendungen wie Leuchtdioden (LEDs), Photovoltaik, Medizin, Laser und TV- Displays wurden entwickelt. Trotz starker und rasanter Weiterentwicklung dieses Feldes im Rahmen der Materialqualität, wie z.B. schmale Größenverteilung und Photolumineszenzquantenausbeute (PLQY), bleibt der Bedarf nach einer wesentlichenVereinfachung der Produktionstechnologie, der Fertigung im industriellen Maßstab und deren Kostensenkung noch hoch-aktuell. Darüber hinaus ist die Beschränkung durch die Materialtoxizität von Cd-haltigen Nanopartikeln und eine ziemlich begrenzte Palette von den neuen Cd- freien hochlumineszierenden nanokristallinen Materialien die letzte Hürde vor der Auftragsfertigung für den Markt. Das primäre Ziel dieser Arbeit war, neue kostengünstige lumineszierende kolloidale Halbleitermaterialien als Spektralwandler mit einer Anpassung von optischen Eigenschaften für bestimmte Anwendungen zu entwickeln. Ich bin Chemiker und habe einen Master-Abschluss auf dem Gebiet des Wachstums von Volumenkristallen (CdTe). Während meiner Promotion habe ich mich in mehrere neue Felder, wie die Synthese und Oberflächenmodifikation von kolloidalen Nanokristallen, Geräteverarbeitung und Photophysik der Halbleitermaterialien eingearbeitet. Die Dr.- Arbeit besteht aus einem anleitenden Teil (Kapitel 1, Kapitel 2) und vier Kapiteln über die erzielten wissenschaftlichen Ergebnisse. Das Kapitel 3 beschreibt die entwickelte Eintopf-Synthese von hochlumineszierenden Kern-Schale-ZnCdS:Mn/ZnS- Kolloid Nanokristallen (NCs), welche auch für eine Hochskalierung d.h. eine Fertigung im industriellen Maßstab geeignet ist. Die Hauptaufgabe dieses Kapitels war die Entwicklung von lumineszierenden kolloidalen Nanokristallen, die auf eine Si-Solarzelle aufgebracht werden könnten. Die Partikel sollten keine re-Absorption bei der Umwandlung des UV- und blauen Teils des Sonnenlichtes ins Licht mit einer größeren Wellenlänge aufweisen. Vor Beginn der erfolgreichen Studie über die hochlumineszierenden mit Null-re-Absorption ZnCdS: Mn / ZnS NCs wurden mehrere NC -Systeme wie z.B. La-dotiertes ZnO, YVO4: Eu, Bi, Carbon Quantum Dots und hochlumineszierende PMMA-ZnO NCs zu diesem Zweck getestet. Eine systematische Untersuchung und Optimierung aller Parameter, welche die Photolumineszenzeigenschaften dieser NCs beeinflussen, wie z.B. die

viii

Materialzusammensetzung einschließlich der Dotierungskonzentration sowie die Schalendicke, führen zu einer Rekord- PLQY von 70%. Diese hochfluoreszierenden Nanokristalle wurden als eine der Hauptkomponenten der Schichten zum Down-Schift vom ultravioletten (UV) zum gelben Spektralbereich eingesetzt. Dadurch wurde die Effizienz von monokristallinen Silizium (mc-Si) - Solarzellen um nahezu 0.5 Prozentpunkte verbessert. Die daraus resultierende Leistungsumwandlungseffizienz (PCE) von konventionellen Solarmodulen mit einer Energieausbeute 14,6%, die mit dem ZnCdS:Mn/ZnS- Lichtumwandler beschichtet sind, ermöglicht eine Kostenreduzierung von 2.1 % für die Solarstromproduktion. Das Kapitel 4 ist dem Synthese-Design von hybriden organisch-anorganischen Metallhalogenid-Perovskit kolloidalen Nanokristallen gewidmet. Die im ersten Teil des Kapitels beschriebenen Ansätze ermöglichen eine durch die Liganden unterstützte Wiederabscheidung und dadurch eine einfache Herstellung von stark lumineszierenden

(PLQY 30-90%) und nahezu monodispersen CH3NH3PbX3 (X = Br, I) kolloidalen Nanoplättchen mit maßgeschneiderten Dicken. Weitgehend abstimmbare Emissionswellenlängen (450-730 nm) wurden allein durch den ausgeprägten Quantengrößeneffekt ohne Anionenhalogenidmischung erreicht. Die rein chemische und kolloidale Stabilität hatte Priorität, um stabilere Analoga zu entwickeln. Daher haben wir im zweiten Teil dieses Kapitels den gleichen Ansatz verwendet, um neuartige Formamidinium

(CH(NH3)2PbX3, X = Cl, Br, I) basierende Perovskit-Nanokristalle zu synthetisieren, die ein + Analogon von Methylammonium (CH3NH3 ) Perovskit ist. Die kubischen und plättchenartigen Nanokristalle, mit ihrer Emission im Bereich von 415 nm bis 740 nm, einer Halbwerstbreite (FWHM) von 20 nm bis 44 nm und Lumineszenzabklingzeiten von 5 ns bis 166 ns, haben durch die Halogenidzusammensetzung sowie durch die Dickenänderung der Nanokristalle eine Bandlückenanpassung erlaubt. Darüber hinaus wurden hohe Photolumineszenzquantenausbeuten (bis zu 85%), sowie kolloidale und thermodynamische Stabilität der Partikel erreicht. Weitere optoelektronische Messungen verifizieren, dass ein Photodetektor auf der Basis von FAPbI3-Nanokristallen den Weg für die Perovskit- Quantenpunkt-Photovoltaik ebnet. Trotz der hohen Photovoltaik-Leistung der Bulk-

Perovskite zeigte ein Photodetektor-Prototyp auf der Basis von FAPbI 3-Nano-Kristallinen eine geringe Photoempfindlichkeit, hauptsächlich aufgrund des Vorhandenseins eines isolierenden langkettigen organischen Liganden auf der NC- Oberfläche. Daher wurde in Kapitel 5 eine neue ligandenfreie und mit einer Steuerung der Morphologie Synthese von CH3NH3PbX3 (X = Br, I) – Kristalliten und deren Mischhalogenid-Analoga in Submikronen -Skala entwickelt. Die beschriebene Entwicklung erlaubte die Herstellung von stabilen Perovskit-Tinten fürs Drucken im industriellen Maßstab. Photodetektor-Geräte, die

ix aus diesen Tinten hergestellt wurden, zeigten eine bemerkenswert hohe Photoempfindlichkeit. Dieses Ergebnis führt ohne Zweifel zu einer prekursorenfreien Herstellung von großflächigen Perovskit-Solarmodulen. Das letzte Ergebnis - Kapitel 6 widmet sich der Untersuchung der signifikanten Bedeutung der chemischen Reinheit von Perovskit-Prekursoren und ihrer Auswirkungen auf die Halbleiter-Qualität. Die in diesem Kapitel präsentierten neuen Erkenntnisse zeigten, dass eine gewisse Menge an Nebenproduktverunreinigungen während einer CH3NH3I-Synthese durch die Spuren von Hypophosphorsäure gebildet wird. Die entstandenen Nebenprodukte fördern dann die Bildung von PbHPO3-Nanopartikeln in Perovskit-Prekursoren. Diese Nanopartikel wirken weiter wie ein Keim fürs Wachstum von großen Körnern mit einer guten Kristallqualität und führen schließlich zu einer hocheffizienten Leistung in den Perovskit- Solarzellen. Die Studie zeigt, dass die beobachteten Reproduzierbarkeit - Probleme, die in der Perovskit-Solarzellengemeinschaft beobachtet und diskutiert werden, hauptsächlich auf unerwartete Verunreinigungen in CH3NH3I zurückzuführen sind. Daher sind die

Reinheitskontrollen für das selbstsynthetisierte oder kommerzielle CH3NH3I erforderlich. Im letzten Kapitel 7 werden Schlussfolgerungen sowie Ausblick auf weitere Untersuchungen zum Thema der Arbeit vorgestellt.

List of Figures

Figure 1.1. Area of luminescent semiconductor NCs application…………………….….3 Figure 1.2. Examples of different semiconductor NCs synthesized via “hot injection” approach. Reproduced from ref. 22. Copyright 2016 American Chemical Society…………...5 Figure 1.3. (a) Volume free energy (red) and surface free energy (blue) contribution to the total free energy transformation (black) as a function of nucleus size r (reprinted from 37). (b) LaMer diagram: prenucleation stage (I) nucleation (II) and growth (III) phases as a function of the reaction time. Reproduced with permission from ref. 37. Copyright 2016 American Chemical Society. ………………………………………………………………….7 Figure 1.4. Illustration of the focusing and defocussing regime during NCs growth. Reproduced with permission from ref. 37. Copyright 2016 American Chemical Society……………………………………………………………………………..…………...8 Figure 1.5. In situ synchrotron SAXS/WAXS monitoring of the formation of CdSe NCs. Top left: Progress in time of SAXS (A) and WAXS (B) patterns. Top right: Transient progress of the imposed temperature and of different characteristic parameters obtained from Monte Carlo fitting of the X-ray data. Bottom: Scathe of step by step CdSe NC formation. Reproduced with permission from ref. 37. Copyright 2016 American Chemical Society………………………………………………………………………………………….9 Figure 1.6. a) Evolution of the characteristic reactions in the “heat-up” synthesis of NCs. b) Effect of heating rate on the final particle size, size distribution, and concentration. Reproduced with permission from ref. 32. Copyright 2016 American Chemical Society…...11 Figure 1.7. Examples of different semiconductor NCs synthesized via “heat-up” approach. Cu2ZnSnS4, Cu2-xS, CdSe and CdSe/CdS NCs are adapted with permission from ref. 46, 44, 48 respectively. Copyright 2016, 2015, 2003 American Chemical Society. PbS and ZnS NCs are adapted with permission from ref. 47. Copyright 2015 AAAS www.sciencemag.org...... 11 Figure 1.8. Variation of the size and morphology control of CH3NH3PbBr3 perovskite synthesized via solvent-antisolvent crystallization approach; this work……………………..12 Figure 1.9. Typical example of size-dependence optical properties of colloidal CdSe QDs in solutions. Picture taken from http://nanocluster.mit.edu/research.php...... ……...... …13 Figure 1.10. The electronic structure of the semiconductor material in bulk, quantum dots and molecule system in relation to the size……………………………………………...……14 Figure 1.11. a) Absorption spectra of 4.1 nm CdSe NCs with well-resolved transitions contain states the 1S and 1P electron states. Reproduced and adapted with permission from ref. 52. Copyright 2000 American Chemical Society. b) Closer look into the CdSe NCs band gap that shows more complex structure of quantized states of valence subbands……………14 Figure 1.12. a) Perovskite metal halide structure. b) Applications based on metal halide perovskites…………………………………………………………………………………....16 Figure 1.13: a) Calculated octahedral and tolerance factors different organic-inorganic hybrid perovskites. Reproduced from ref. 65. b) Orientational disorder in a organic-inorganic perovskites. Reproduced from ref. 67…………………………………………………...... …18 Figure 1.14. Various preparation techniques for perovskite films, single crystals and nanocrystals. Reproduced and adapted from ref. 69…………………………………….……19 Figure 1.15. a) Illustration of ligand-assisted re-precipitation (LARP) technique applied to synthesize MAPbX3 NCs. b) Representative TEM and HRTEM image of the MAPbBr3 NCs. c) Photoluminescence spectra of MAPbX3 NCs toluene solutions and corresponding optical images under room light as well as UV lamp. (d) Photoluminescence spectra and image of size-tailoring emission of brightly emissive MAPbX3 NCs, where the size was controlled by temperature of reaction. Reproduced from ref 79…………………………...…21

Figure 1.16. (a) Structure of the cubic CsPbBr3 perovskite lattice. b,c) HRTEM and TEM images of self-assembled monolayer of CsPbBr3 NCs. d,e) Optical image of CsPbX3 NC solutions under UV lamp, and their corresponding size- and composition-tunable Photoluminescence spectra. (d) Absorption and Photoluminescence spectra of CsPbX3 NCs with different halide composition (e) Time-resolved PL decays of CsPbBr3 NCs. Reproduced from ref. 84……………………………………………………………………………………22 Figure 17. a,b) Sketch of anion-exchange raeactions within CsPbX3 crystal lattice and suitable reagents. c,d) Photographs of ion-exchanged, mixed-halide MAPb(Br/Cl)3 and MAPb(Br/I)3 NCs under room light. (d) UV-visible and PL spectra of MAPb(Br/Cl) and MAPb(Br/I)3 NC films. Reproduced from ref. 79……………………………………..……..23 Figure 1.18. Quantum-size effect in perovskite metal halide NCs. a,b) Illustration of MAPbBr3 nanoplatelets with different number of unit cell numbers (n=1,2,3 and ∞) and their optical PL and absorbance spectra. The number of unit cell monolayers layers decreased with increasing octylamine content in the reaction mixture. c,d,e) Absorption and Photoluminescence spectra of CsPbBr3 nanocubes and nanoplatelets made of 1–5 unit cells monolayers with their correspondence TEM micrographs for the NCs synthesized at higher (150°C, green PL) and lower (90°C, blue PL) temperature. Reproduced from ref. 79…...….24 Figure 1.19. (a) Calculated defect charge-transition levels for CsPbBr3. (b) Schematic representation of shallow levels formation . Reproduced with permission from ref. 104. Copyright 2017 American Chemical Society……………………………………………...…25 Figure 2.1. Solvent - antisolvent extraction technique………………………..…....……40 Figure 2.2. Schematic representation of perovskite ink preparation………,……...... ….42 Figure 3.1. Schematic illustration of the one pot, two step synthesis of ZnxCd1-xS:Mn/ZnS core-shell NCs…………………………………………………………...51 Figure 3.2. TEM and HRTM images of (a) Zn0.5Cd0.5S:Mn and (b) ZnS-coated Zn0.5Cd0.5S:MnNCs; the corresponding size distributions obtained by TEM are given as insets. (c) XRD patterns of the two NC samples. The standard data for ZnS zinc blende bulk material (top, ICSD 00-005-0566) and zinc blende bulk CdS (bottom, ICSD pattern 01-080-0019) are shown as reference. (d) FTIR spectra of Zn0.5Cd0.5S:Mn(5%)/ZnS NCs, and (e) an enlarged view of the same spectra in the wavelength region of 2380-2880 cm-1 for the NCs and DDT……………………………………………………………………………...….53 Figure 3.3. (a) Dependence of the intensity of the Mn2+ emission at 598 nm on reaction time and temperature. The PL intensities are normalized to the maximal Mn2+ PL intensity of the NCs taken from the reaction mixture directly after reaching 230 °C. (b) Thermal and temporal evolution of absorption (left axis) and emission spectra (right axis) of (unshelled) 2+ Zn0.5Cd0.5S NCs doped with 5 mol % of Mn using excitation wavelengths (λex) of 375 and 390 nm for the sample taken at 150°C and 175 °C, respectively, and 400 nm for all other samples to always excite each sample at the first excitonic absorption maximum. (c) PL intensity at different Cd/Zn precursor ratios and (d) the corresponding absorption (left axis) and normalized emission (right axis, excitation at the first excitonic absorption maximum of each sample) spectra of Zn1-xCdxS:Mn/ZnS NCs. For all core compositions in panels c, d, e, and f, the NCs were coated with one layer of ZnS. (e) PL intensity of undoped 2+ Zn0.5Cd0.5S/ZnS and Zn0.5Cd0.5S/ZnS doped with different Mn concentrations (λex = 400 nm); (f) absorption (left axis) and normalized emission spectra (right axis) of undoped 2+ Zn0.5Cd0.5S/ZnS and Mn -doped NCs. (g) Dependence of the PL intensity on the number of ZnS coating layers, and (h) evolution of the corresponding absorption (left axis) and normalized emission spectra (right axis, λex = 400 nm) of the Zn0.5Cd0.5S:Mn core (black 2+ curve) and Zn0.5Cd0.5S:Mn/ZnS core-shell particles, doped with 5 mol % Mn . All absorption spectra were recorded with samples having the same absorbance at the first excitonic absorption band/transition. The inset in panel (h) illustrates the light absorption associated with the valence band-conduction band transition, energy transfer from the accordingly formed exciton to the Mn2+ ions, and the nonradiative relaxation initiated by surface defects

xii

(SD). All measurements were performed with NCs dispersed in chloroform………….…….55 Figure 3.4. EDX analysis of the elemental composition of the Zn0.5Cd0.5S:5% Mn (a) and Zn0.5Cd0.5S:5% Mn/ZnS (b) d-dots. Survey XPS spectrum of Zn0.5Cd0.5S:Mn NCs with 10% Mn (c) and Zn0.5Cd0.5S:Mn/ZnS (c). In the inset, the detail spectra of the Mn 2p region is shown; the red line is added as guide to the eye…………………………………….……...... 57 Figure 3.5. a) PL decay curves of the Zn0.5Cd0.5S:Mn and Zn0.5Cd0.5S:Mn/ZnS NCs with six Zn precursor injection. Chloroform solutions of the NCs exited with Xe-lamp at 320 nm, and detection wavelength at 600 nm for both samples. b) Upscaling of the synthesis of Zn0.5Cd0.5S:Mn(5%)/ZnS core shell NCs. …………………………………………..………..60 Figure 3.6. Performance of mono-Si solar cells coated with Zn0.5Cd0.5S:Mn(5%)/ZnS NCs, using a volume 200 μl of the differently concentrated NCs dispersion. (a) Scheme of the down-shifting mechanism of the Zn0.5Cd0.5S:Mn(5%)/ZnS NC converter material and design of proof-of-concept solar cells. (b) EQE of the best mono-Si solar cell (black curve) obtained with a Zn0.5Cd0.5S:Mn(5%)/ZnS NC dispersion containing 10 mg/ml NCs (red curve); the EQE in the wavelength region of 300 to 400 nm is magnified in the inset. (c) J–V curves of the corresponding solar cells. . (d) SEM picture of the mono-Si solar cell surface structure coated with 5 mg/ml NCs solution. (e) cross-section of the same mono-Si solar cell with 260 nm NCs thickness…………………………………………………………….…….62 Figure 3.7. Selected SEM picture of the mono-Si solar cell surface structure cross- section of the same mono-Si solar coated with 0.5 mg/ml (a), 2.5 mg/ml (b), 5 mg/ml (c), 10 mg/ml (d) and 25 mg/ml (e) NCs solution. The obtained NCs thicknesses varies from <50 to 750 nm…………………………………………………………………….……….………….63 Figure 3.8. Enhancement ratio of EQE of mono-Si solar cells as function of wavelength for the coating with differently concentrated Zn0.5Cd0.5S:Mn(5%)/ZnS NCs dispersions…...………………………………………………………………………………..64 Figure 4.1. MAPbBr3 NPLs with different thicknesses in toluene solutions. a) Digital photo of the NPLs under UV-365 nm illumination. b) PL decay. c) Absorbance and PL spectra. d) Typical bright-field TEM image of the NPLs with a PL peak at 447 nm obtained with a ligand ratio of OA/OAm = 200 μl/30 μl. e) XRD pattern of NPLs with characteristic PL (green, cyan and blue)……………………...…………………..73 Figure 4.2. PL spectra of the blue (447 nm) emitting NPLs with two different lateral size (5 nm and 11 nm) but same thickness of 1.6 nm…………………………………………….74 Figure 4.3. Electron diffraction analysis of MAPbBr3 NPLs: a) exemplary diffraction pattern for NPLs with PL peak at 514 nm. b) Line profile obtained by radial averaging of the diffraction pattern and expected peak intensities for random orientation of the NPLs (red lines). The analysis confirms the perovskite crystal structure and points to a preferred <001> orientation of the NPLs on the carbon support grid………………………..…………………75 Figure 4.4. Bright-field TEM characterization of the ligand-assisted layer control of MAPbBr3 NPLs. Lateral particle size and thickness determination are derived from the upper and lower images respectively, of a) blue (447 nm), b) cyan (488 nm) and c) green (514 nm) emitting NPLs……………………………………………………………………….………..76 Figure 4.5. Self-assembled superstructure of blue emitting (447 nm) MAPbBr3 NPLs. a) Bright-field TEM image. The marked areas (red and blue squares) and corresponding Fourier transformations indicate a 2D platelet superstructure b) Sketch of a possible OAm ligand overlapping. c) Proposed sketch of NPLs assembling with average distances between individual NPL measured from the TEM image……………………………………..……….77 Figure 4.6. Radiative (krad ) and non-radiative (knon-rad) rate for MAPbBr3 NPLs (a) and MAPbI3 NPLs with different PL emission peaks (b)…………………………..…………….78 Figure 4.7. a) MAPbI3 NPLs solutions under UV-365 nm illumination. b) PL decay. c) Absorbance and PL spectra. d) Typical bright-field TEM image of the NPLs with a PL peak at 722 nm obtained with a ligand ratio OA/AOm = 200 μl/50 μl e) XRD pattern of purified NPLs with PL of 722 nm (ICSD card №250739)………………………...…………79

xiii

Figure 4.8. a) Quantum size effect in CdTe quantum dots (QDs)50. b) Quantum size effect in MAPbI3 nanoplatelets (NPLs)…………………………...... ……80 Figure 4.9. TEM picture of the MAPbI3 NPLs. a) Red NPLs with PL peak at 683 nm; white dashed square indicated non-destroyed NPLs; yellow circles marks destroyed NPLs under high energy beam irradiation b) NIR NPLs with PL peak 722 nm. Differences in contrast of the NPLs display different thicknesses obtained nanocrystals………………...…81 + Figure 4.10. a) Representative perovskite FAPbX3 (X=Cl, Br, I) unit cell with FA cation 4- in the centre of eight [PbX6] octahedra. b) HRTEM of single FAPbI3 NC. c) TEM micrograph of the monolayer FAPbI3 NCs on a grid; σ - size distribution. d) XRD curves for single and mix-halide FAPbX3 NCs………………………………………………………….82 Figure 4.11 TEM micrograph of mainly lean FAPbCl3 (a-b) and FAPbBr3 (c-d) NPLs..82 Figure 4.12. a) XRD patterns of FAPbI3 NCs colloidal solution stored for a period of 150 days. b) Black to yellow phase transition of unpurified FAPbI3 NCs as result of agglomerates formation of poorly passivated NCs upon centrifugation washing process. References for cubic “black” α- and hexagonal “yellow” δ-phase were taken from CIF file reported by Stoumpos et al.61………………………………………………………………………...……83 Figure 4.13. Optical properties of the colloidal solution of FAPbX3 nanocrystals in toluene: a) Representative optical absorption and PL spectra; b) PL decay dynamics of the pure and mixed-halide NCs c) Compositional PL tuning diagram; d) Digital picture of colloidal solution in toluene taken under UV-light (λ=365 nm). e) Compositional tuning of the PL spectra of the mixed halide NCs………….……………………………………………….85 Figure 4.14. Photostability of the FAPbBr3 NCs film under different light irradiation. a) PL intensity upon 445 nm (7.5W/cm2) laser illumination. b) PL spectra after different exposure times. Note that the PL peak energy and the FWHM do not change with exposure to blue light. c) PL intensity upon 375 nm (3W/cm2) laser illumination. d) PL spectra after different exposure times. The PL peak shows a red shift of 5 nm………………………..…..85 Figure 4.15. Comprehensive characterization of purified FAPbBr3 NPLs toluene solutions with different thicknesses. a) Band gap and PL tuning by changing the OAm/OA ratio with corresponding PLQY values; b–d) Bright-field TEM characterization of the vertically stacked FAPbBr3 NPLs synthesized with different OAm/OA volume ratio: b) 200 μl /150 μl blue (438 nm) emitting NPLs with lateral size 36.6±9 nm and thickness 1.4±0.1 nm; c) 200 μl /80 μl cyan (486 nm) emitting NPLs with lateral size 22.7±3 nm and thickness 2±0.1nm; d) 200 μl /40 μl green (533 nm) emitting NPLs with lateral size 21.5±4 nm and thickness 2.6±0.2 nm; σ - thicknesses distribution. e) Theoretical (EMA) versus experimental band gap of the FAPbBr3 NPLs as a function of the thicknesses (d). Inset shows unit cell number (n) for single NPL with different thickness; f) PL decay………………………….…87 Figure 4.16. a) Illustration of FAPbX3 NCs coating with Mercaptopropylisobutyl-POSS cascade molecule resulting bright emitted powder and toluene solution. b) - d) PL spectra of FAPbCl2Br, FAPbBr3 and FAPbBrI2 NCs powder before and after dipping into the water...89 Figure 4.17. a) The current–voltage (I–V) curves of FAPbI3 drop casted NCs film on patterned ITO finger substrates (inset figure) in the dark (black) and under AM 1.5 illumination (red). b) On/Off switching properties (at a bias of 3 V)…………………...……89 Figure. 4.18. a) Schematic illistration of anion-exchange raeactions within FAPbX3; b) Optical absorption and PL spectra of anion-exchanged FAPbBr3 NCs to their Cl and I pedants by adding TBAX (X=Cl, I) to FAPbBr3 NCs crude solution. c) XRD of anion exchanged NCs. d) PL spectra evolution upon anion-exchange of purified FAPbBr3 NCs by OAmI toluene solution adding. e) Interparticle anion-exchange realized by blending purified FAPbBr3 and FAPbI3 NCs toluene solutions in different ratio. Inset show result of this reaction under UV-365 nm lamp………………………………………………………………………..……91 Figure 5.1. a) Pallets pressed from dried perovskite powder. b) Absorption mixed halide perovskites powders. c) XRD………………………………………………………….……..99

xiv

Figure 5.2. a-f) SEM overview of various perovskite powders with different morphology…………………………………………………………………………….……100 Figure 5.3. Rheological stability in time of inks based on MAPbB3(DMF)T, MAPbI3(GBL)C and MAPbBr3(NMF)T employing a 30:7.5 concentration ratio…………102 Figure 5.4. Different sizes of glass substrates, coated with MAPbI3(GBL)C and MAPbBr3(DMF)T inks……………………………………………………………...………103 Figure 5.5. Dependence of the film thickness over blade gap, with varying viscosities for (a) MAPbBr3(NMF)T, (b) MAPbBr3(DMF)T and (c) MAPbI3(GBL)C. Each Material features the following three ink ratios: 30:5 - 30%(w) perovskite and 5% (w) EC in toluene, 30:7.5 - 30%(w) perovskite & 7.5% (w) in toluene; 30:10 - 30%(w) perovskite & 10%(w) EC in toluene. b,d,f) Confocal microscopy results for the 30:5 series of all three materials at a constant blading gap of 80 µm………………………………………………………...…….104 Figure 5.6. Absorption and PL spectra of thin films bladed out of (a) MAPbBr3(NMF)T, (b) MAPbBr3(DMF)T and (c) MAPbI3(GBL)C inks. Dashed lines corresponds to PL of pristine powder samples…………………………………………...……………………..….107 Figure 5.7. a,b,c, right) SEM study of coated film (blade gap 80 µm) out of MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C inks with in different magnification; a,b,c, left) XRD patterns for powder, film coated from fresh prepared and two weeks aged ink………………………………………………………………………………108 Figure 5.8. Sketch of the ITO- substrate……………………………………….………109 Figure 5.9: a, d, g) Photo of produced MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C photodetectors; b, e, h) Corresponding microscopic view on ITO-substrate and (c, f, i) dark and light photocurrent over the voltage…………………………………....110 Figure 5.10. Reponsivity and Specific Detectivity of MAPbBr3(DMF)T (left) and MAPbI3(GBL)C (right) based photodetector over wavelength……………………………..111 Figure 5.11. a) Schematic representation of the recrystallization process. b) SEM pictures of a MAPbI3(GBL)C film before and after MA gas treatment at three different magnifications. c) XRD patterns of a film of MAPbI3(GBL)C before and after MA gas treatment d) I-V measurements of a MAPbI3(GBL)C photodetector before and after MA gas treatment…………….……………………………………………………………………….112 Figure 6.1. a) XRD diffraction patternof the Pb(H2PO2)2 . Reference diffraction data was taken from Kuratieva et al. 12 . b) XRD diffraction pattern of the precipitate is in excelent agreement with the diffraction data for PbHPO3 taken from ICSD database (card №. 00-020-0580)…………………………………………………………………..…..119 Figure 6.2. Comparison of 1H NMR spectra of NH3 protons signal for purified, non-purified (a) and commercially available MAI from different providers (b). All of them contain MAH2PO2 and MAH2PO3 in varying amounts, except purified MAI…………...….119 Figure 6.3. Compositional investigation of the precipitate and impurities nature. a) SEM image of the white precipitate from perovskite precursor solution precursor based on i-MAI. b) XRD diffraction pattern of this precipitate shows agreement with diffraction data 1 31 for PbHPO3 taken from ICDD database (PDF №. 00-020-0580). c) H and P NMR spectra 1 31 for с-MAI. The inset shows the NH3 (blue) and CH3 (cyan) signal. d) H and P NMR spectra of i-MAI. The inset shows the broadening of the NH3 (blue) signal, compared to pure с-MAI, with additional peaks stemming from contamination. The CH3 (cyan) signal also becomes broader compared to с-MAI. e) XRD diffraction pattern of the c-MAI and i-MAI. Inset photo of the holder for XRD after i-MAI measurements. The samples seem to be radiation damaged during the measurements, which is clearly shown on XRD diffractogram (red curve) - some of the higher angle reflexes are weaker or missing entirely. f) DSC measurement of the c-MAI and i-MAI. In the case of i-MAI, all endothermic and exothermic peaks (melting, crystallization and sublimation) is shifted to lower temperature due to the presence of impurities………………………………………………………………………………...….121

Figure 6.4. NMR study of MAIs and its impurities. a) 1H NMR spectra of possible impurities in perovskite precursor solution and their mixtures with pure c-MAI. Dashed lines show the position 1 of the impurities in i-MAI. Ideal fits (dashed lines) as found for MAH2PO2 and MAH2PO3. b) H NMR spectra of i-MAI as compared to c-MAI with randomly added MAH2PO2 and MAH2PO3. The same position of the sharp peaks confirms the presence of both MAH2PO2 and MAH2PO3 in i-MAI. c) Quantitative analysis c-MAI and MAH2PO2 mixture at different concentrations. Inset images show an expansion of the signal growth after MAH2PO2 addition. d) Quantitative analysis of c-MAI and MAH2PO3 mixture in different concentrations. Inset images show an expansion of the signal growth after MAH2PO3 addition. e) Effect of MAH2PO2: increasing the MAH2PO2 concentration shifts the main peak of NH3 group signal in c-MAI to lower field (higher frequency), but the FHWM remains constant. f) Effect of MAH2PO3: FHWM of NH3 group signal in c-MAI becomes broader with increasing MAH2PO3 concentration, but the main peak signal remains nearly unchanged. All measurements were performed in DMSO-d6…………………………………………..……………..123 Figure 6.5. DSC study of MAH2PO3. Melting start around 60°C with further decomposition of the materials at higher temperature………………………...…………….124 Figure 6.6. Tyndall effect in DMF perovskite precursor. a) Green laser light passed through the bottles under background illumination and in the dark. From left to right: pure DMF, perovskite precursor solution based on C-MAI and filtered through 450 nm filter, precursor solution based on i-MAI filtered through 200 nm filter; the same solution filtered through 450 nm filter.b) DMF solutions of the c-MAI (A) and i-MAI (B) illuminated with green laser.No Tyndall effect is observed. c) The same solutions after Pb(CH3COO)2 addition, dissolved in DMF. The Tyndall effect in solution (B) is observed due to due to PbHPO3 colloid formation. d) Tyndall effect of the filtered (450 nm filter) diluted c-MAI based perovskite precursor after adding one drop 1mg/ml DMF solution of MAH2PO3………….125 Figure 6.7. Schematic interpretation of the perovskite thin film growth based on different precursor solution with and without PbHPO3 NPs………………………………………….126 Figure 6.8. Morphological and crystallographic analysis of perovskite thin film with c-MAI and i-MAI. Top view SEM images of the perovskite film prepared on top of the LT- NiO by using c-MAI (a) and i-MAI (b), respectively. The insets show a higher resolution picture captured at a high electron beam power of 20 kV and the average grain size distributions; AFM and KPFM measurements were performed on perovskite/LT- NiO/ITO/glass samples. Topographic images of perovskite with c-MAI (c) and perovskite i-MAI (f), phase-contrast imaging of perovskite with c-MAI (d) and perovskite i-MAI (g), potential images of the corresponding topography (e,h). The GB regions can be easily separated from all the above three images; (i), X-ray diffraction spectra of two perovskite thin films employing differentMAI…………………………………………………….……..….127 Figure 6.9. Photophysical properties of perovskite films and their solar cell performances.(a) steady-state PL spectra for the perovskite films; (b) time-resolved PL decay for the perovskite films; (c) I-V characteristics measured under 100 mW cm-2 AM1.5G illumination for the highest-performing devices employing c-MAI and i-MAI; (d) EQE spectrum of the highest-performing devices employing c-MAI and i-MAI……………..….128 Figure 6.10. I-V curves of device employing perovksite with i-MAI (a) and c-MAI (b) measured under forward and reverse directions. c) I-V curves of device employing perovksite with i-MAI measured under different I-V scan rate…………………………………………129

xvi

List of Tables

Table 2.1. Parameters of synthesis of CH3NH3PbBr3 NPLs with different thicknesses ...... 36

Table 2.2. Parameters of synthesis of CH3NH3PbI3 NPLs with different thicknesses ...... 36

Table 2.3. Parameters of synthesis FAPbCl3 NCs with different thicknesses ...... 39

Table 2.4. Parameters of synthesis FAPbBr3 NCs with different thicknesses ...... 39

Table 2.5. Parameters of synthesis FAPbI3 NCs with different thicknesses ...... 39 Table 2.6. Solvent-Antisolvent extraction process parameters ...... 40 Table 2.7. Blend table for mixed halide perovskites ...... 40 Table 2.8. Applicable solvent for the precursor materials ...... 41 Table 3.1. Previously reported Mn2+ doped semiconductor nanocrystals and their properties ...... 61 Table 3.2. Photovoltaic properties of mono-Si solar cells coated with differently concentrated dispersions of Zn0.5Cd0.5S:Mn(5%)/ZnS NCs. Open-circuit voltage (UOC) and fill factor (FF; FF = max. power/JSC UOC) are taken from the J-V measurements under the solar simulator, the short circuit current (JSC) and the power conversion efficiency (PCE) are calculated from the EQE spectra...... 65 Table 4.1. PLQY, PL decay time, radiative (krad ) and non-radiative (knon-rad) rate for MAPbBr3 NPLs with different PL emission peaks...... 78 Table 4.2. PLQY, PL decay time, radiative (krad ) and non-radiative (knon-rad) rate for MAPbI3 NPLs with different PL emission peaks...... 78 Table 5.1. Square root mean values for surface roughness of films of the three perovskite materials for varying ink consistencies at a constant blading gap of 80 µm...... 106 Table 6.1.Summary of photovoltaic parameters of the investigated perovskite solar cells prepared by c-MAI and i-MAI based perovskite precursor solution...... 129

xvii

Table of contents Acknowledgments ...... iv Summury ...... vi Zusammenfassung ...... viii List of Figures ...... xi List of Tables ...... xvii Table of Contents ...... xviii Chapter 1. Introduction ...... 1 1.1. Motivation ...... 2 1.2. Solution-processed synthesis of semiconductor nanocrystal ...... 4 1.2.1. Fundamentals of Colloidal Synthesis: Nucleation and Growth ...... 6 1.2.2. Optical properties of semiconductor nanocrystals ...... 13 1.3. Metal halide perovskites ...... 16 1.3.1. Synthesis of hybrid organic-inorganic perovskites ...... 19 1.3.2. Synthesis of hybrid all-inorganic and metal halide colloidal nanocrystals ...... 21 1.3.3. Band-gap tuning of perovskite NCs via anion exchange reactions ...... 23 1.3.4. Quantum-size effect in perovskite nanocrystals ...... 24 1.3.5. Origin of the high photoluminescence quantum yield in perovskites NCs ...... 25 1.4. Bibliography ...... 26 Chapter 2. Materials and Methods ...... 31 2.1. Materials ...... 32 2.2. Nanocrystal synthesis ...... 34 2.3. Perovskite ink preparation ...... 42 2.4. Device fabrication ...... 42 2.5. Characterization techniques ...... 43 2.6. Bibliography ...... 47 Chapter 3. Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells ...... 48 3.1. Motivation and State of the art ...... 49 3.2. One-pot synthesis of Mn-doped ZnxCd1-xS:Mn/ZnS core/shell NCs ...... 51 3.3. Analysis of NCs Surface...... 53 3.4. Optimization of the NCs growth synthesis ...... 54 3.5. Applying NCs as downshifting layer for Si solar cell ...... 60 3.6. Summary...... 66 3.7. Bibliography ...... 67 Chapter 4. RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals ...... 70 4.1. Motivation and State of the art ...... 71 4.2. Quantum size-confined MAPbX3 (X=Br, I) colloidal NPLs ...... 72 4.3. Brightly luminescent and stable FAPbX3 (X=Cl, Br, I) colloidal NCs ...... 81 4.4. Summary ...... 92 4.5. Bibliography ...... 93 Chapter 5. Perovskite ink: from nanocrystaline powder to device application ...... 97 5.1. Motivation and State of the art ...... 98 5.2. Perovskite nanocrystalline powder ...... 98 5.3. Perovskite ink formulation ...... 101 5.4. Summary ...... 114 5.5. Bibliography ...... 115 Chapter 6. Purity control of the precursor materials for perovskite synthesis ...... 116 6.1. Motivation and State of the art ...... 117 6.2. Precursor and starting materials analysis ...... 118 6.3. Summary ...... 130 6.4. Bibliography ...... 131 Chapter 7. Conclusions and outlook ...... 132 7.1. Conclusions...... 132 7.2. Outlook ...... 133 7.3. Bibliography ...... 135 Curriculum vitae ...... 136

xviii

Motivation and Introduction Chapter 1

Chapter I

Motivation and Introduction

This chapter gives a motivation to the thesis and highlights the importance of optimization of state of the art nanocrystals with respect to designing new luminescent semiconductor colloids for optoelectronic applications. The basics of nanocrystal synthesis are further introduced to the fundamentals of controlling the colloidal properties of the nanocrystal are outlined.

- 1 -

Motivation and Introduction Chapter 1

1.1. Motivation

The global market of luminescence materials is steadily growing. For instance, world leading lighting manufacturer OSRAM Licht AG (Germany) is showing a high annual growth rate of 27% (http://www.tradingeconomics.com). However, most of the material still contains expensive rare-earth elements such as Ce and Eu, reserves of which are continuously decreasing. At the same time, rapidly growing interest in cheap photovoltaics, has shown strong progress of solution process semiconductors such as CdTe, PbS,

CuInSe2,Cu2ZnSnS4 as well as perovskites. Solution processing technique is the best alternative to the expensive rather vacuum processing. Furthermore, in the form of colloidal NCs solutions, semiconductors might be used in the form of inks for solar cell printing, which is well developed for organic photovoltaic (OPV). Therefore, design and developing new low-cost luminescent and photovoltaics-attractive semiconductor NCs with the potential of production in large scale are strongly desirable for both fields. Luminescence material development in the last 30 years was basically focused on colloidal semiconductor nanocrystals, which are also named quantum dots (QDs). Their unique properties were applied in the field of numerous applications, mostly due to their unique properties and performance (Figure 1.1). The most successful approach to achieve high quality material is hot-injection synthesis (e.g. CdSe). However, for hot-injection-based methods, several kinetic and thermodynamic factors such as temperature instability, mixture heterogeneity, and overall high reaction rates and limited mass-transfer rates, often hinder the reproducible preparation of high quality nanocrystals with high PLQY on a large scale. In typical QDs (CdE, E=S, Se, Te) luminescence is occurring via band-band transition yielding comparably high PLQY (up to 50%) and close to the unity when properly passivated with wide bandgap shell (CdS or ZnS). Despite good optical properties, insignificant overlap between absorption and emission spectra of QD, resulting in self-absorption losses, limits the usage in spectral convertors such as luminescent concentrator (LSC), luminescent down- shifting (LDS), light emitting diodes (LED) and other light harvesting applications. Therefore, zero-reabsorption luminescent NCs with high PLQY are major in a target. Additionally, novel less toxic and cost effective preparation techniques for highly luminescent semiconductor colloidal nanocrystal are strongly desirable, considering the always strong interest into this class of materials.

- 2 -

Motivation and Introduction Chapter 1

Figure 1.1. Area of luminescent semiconductor NCs application

To bring the luminescent NCs to the industrial level with low cost fabrication practical, the main aim of this thesis is to optimize current luminescence semiconductor state of the art NCs and design new highly luminescent colloidal NCs. This thesis describes the design and 2+ optimization of highly luminescent and zero-reabsorption Mn doped core/shell ZnxCd1-xS + + NCs, perovskite APbX3 (A= CH3NH3 or MA; CH(NH2)2 or FA; X= Cl, Br, I) NCs colloids and their ligand-free and printable inks. Additionally and most importantly, quality control methods to test the precursor purity are established as a key factor to achieve high quality perovskite semiconductors and solar cell devices thereof. Efficient LDS layers for monocrystalline Si solar cells and photodetectors were further presented.

- 3 -

Motivation and Introduction Chapter 1

1.2. Solution-processed synthesis of semiconductor nanocrystal

Nanocrystals (NCs) are small crystals of different classes of crystalline materials such as metals, semiconductors or magnetic material and contain hundreds or few thousands atoms with size range 1-50 nm. Their unique quantum mechanical properties paved the way to produce new materials with tailored optical, magnetic and semiconductor properties. Size-tuning of that NCs leading to new phenomena in comparison to bulk such as surface plasmon resonance in Ag and Au nanocrystals, supermagnetism in magnetic NCs as well as quantum size-effect in semiconductor NCs. Colloidal wet-chemistry synthesis are new branch for growing a wide range of nanostrucrures types, offering control over particle size and shape, and also minimizing particle polydispersity. The pioneer work of the wet-chemical synthesis of colloids was conducted over 150 years ago by Faraday1 to produce nanosized colloidal gold. In aqueous solution of AuCl3, phosphorous as a strong reducing agent was added and Au3+ reduced to Au0 in form of a nanocrystal. This was the first example of a methodology that has evolved into a class of reactions in which colloid formation is induced by the rapid combination (injection one into another one) of two or more critical reagents. The first work2 that describe synthesis of highly monodisperse Cd-chalcogenide (CdE; E=S,Se,Te) colloidal semiconductor nanocrystals (also known as quantum dots (QDs)) was published in 1993 by Murray, Norris and Bawendi which following up the pioneering work of Brus3 from USA, Ekimov and Onushenko from USSR4 and Henglein from Germany5 on Cu-halide doped glass matrixes and CdS colloids. They show that their optical properties such as photoluminescence (PL) and band gap energy can be modulated by NCs size. In the next 3 decades, colloidal CdSe QDs was the work horse for numerous studies and applications6, 7. This work opened the possibility for the synthesis of various semiconductor 8 9 10 11 12 13 QDs and their multicomponent alloys such as InP , InAs , PbS , PbTe , Cu2-xS , GaAs , 14 15 16 17 In2Se3 , Cu2ZnSnS4 , ZnxCd1-xS , CdSexS1-x as well as their core/shell structures (e.g. 18 19 20 21 CdSe/ZnS , PbS/CdS , HgTe/CdTe , CdSe/CdS/ZnxCd1-xS/ZnS ) with enhanced optical properties caused via electronic passivation of the surfaces (Figure 1.2).

- 4 -

Motivation and Introduction Chapter 1

Various shapes in certain crystallographic dimensions were developed: nanoplatelets23, nanowires24, nanodiscs25, tetrapods26, octopods27, multipods28, pyramids28, nanorings29 or nanobelts30. As consequence, NCs do find commercial use as building blocks in large-area and low-cost electronic and optoelectronic devices31 via simple solution processing such as spin-coating, blading, inject-printing, spraying, roll-to-roll printing etc. However, the well established hot-injection method is less attractive from practical point of view, mainly due to reproducibility risks and in scaling-up32. An alternative to “hot-injection” is the “heat-up” (or non-injection) technique32, were all reagents are loaded together and controllably heating initiates nucleation and growth. In the last 3 years due to the extremely growing interest in perovskite NCs, was well developed the method of ligand assisted re-precipitation (LARP)33, inspired from organic nanoparticle or polymer dots synthesys34, 35. In that case, all blended precursors and ligands injected into “bad-solvent” causing nucleation and nanocrystal formation. All processes occur mainly at room temperature and can be easily scaled-up. These techniques are discussed in further detail next.

- 5 -

Motivation and Introduction Chapter 1

1.2.1. Fundamentals of Colloidal Synthesis: Nucleation and Growth

Colloidal NCs are typically synthesized via appropriate molecular precursors that are the salts of fatty acid, metal-amino complex or inorganic salt. Typical “bottom-up” approach of NCs growth includes the following stages: nucleation from free monomer, growth of nuclei, nanocrystal formation and growth, isolation and purification, post-synthesis treatment. For growing of monodisperce NCs, nucleation and growth should be separated22. Surface capping ligands (tri-n-octylphosphine, oleyilamine, oleic acid, dodecanthiol etc.) provides colloidal stability, keep the obtained nanocrystals in solution and avoid their agglomeration even after isolation and purification process. Further in detail nucleation and growth process are described for “hot-injection”, “heat-up” and solvent-antisolvent NCs synthesis approach. (i) Hot-injection. The first work that described the nucleation phenomena was published by LaMer36 in 1950 for precipitation reaction of the sulfur hydrosol and so-called as “Theory of nucleation”. This model is widely applied for describing solution NCs growth with narrow size distribution. The nucleation process is most crucial in the synthesis and should be as fast as possible with no further nuclei formation during the growth process, which leads to monodisperse NCs growth. Nucleation process can be divided into the heterogeneous and homogeneous type. Heterogeneous nucleation requires lower energy and monomer growth carried out onto preferential sites of the surface, while homogeneous nucleation occurs via the formation of a new nucleation phase from dissolved monomer37. The magnitude of this energy depends on binding precursor’s constants, the level of supersaturation, temperature and surfactant concentration in solution. Both process can be described as thermodynamic systems which tend to minimize the Gibbs free energy for the nucleus with a spherical shape37. During nucleation, the Gibbs free energy change is the sum of the negative bond formation and the positive Gibbs free energy surface32: 4 ∆퐺 = − 휋푟3|∆퐺 | + 4휋푟2훾 (1.1) 3 푉

푑∆퐺 2훾 for = 0 → 푟푐 = (1.2) 푑푟 |∆퐺푉 where r is nucleation radii, GV the Gibbs bulk free energy per unit volume, and γ the surface free energy per unit area. Nuclei smaller than critical radius rc redissolve while larger enter the growth stage. This process is depicted in Figure 1.3a, where the maximum of ΔG with r is corresponding to the energy barrier and ΔGc activation energy.

- 6 -

Motivation and Introduction Chapter 1

Figure 1.3. (a) Volume free energy (red) and surface free energy (blue) contribution to the total free energy transformation (black) as a function of nucleus size r (reprinted from 37). (b) LaMer diagram: prenucleation stage (I) nucleation (II) and growth (III) phases as a function of the reaction time. Reproduced with permission from ref. 37. Copyright 2016 American Chemical Society.

To link nucleation and growth, the LaMer model was proposed and adapted for different NCs reaction model depending on the reactivity of the precursor-to-monomer conversion dynamics (Figure 1.3b). For example, metal carboxylates may simply form metal chalcogenides in presence of chalcogens at temperatures of 100-150 °C (PbS10, PbSe38, PbTe11, NCs). However, to obtain their metal oxides higher temperatures39, 40 are required (300-380°C) and high boiling solvents such as 1-Octadecene or Oleayilamin, which also can act as surfactant, are ideal for the NCs synthesis. Within the LaMer model there should be a well-defined time stage in which a lot of nuclei are formed. When the concentration of monomers rapidly grows upon addition of the precursors (Stage I) monomer supersaturation is achieved of the critical level C* and nucleation start37. This stage continues until the monomers for the growing nuclei exceed the monomer generation via precursor transformation (Cmon

- 7 -

Motivation and Introduction Chapter 1

of the seeds. Nonetheless, this model excludes the NCs aggregation process resulting in a broadening of the NCs size distribution. Therefore, Sugimoto proposed a more comprehensive model for the size-dependent growth rate42. This model suggests that under low supersaturation (when r/r*<2; r - NCs radius; r* - particle radius in equilibrium with the bulk solution) the growth rate increases resulting broader size-distribution37. In another case, when r/r*>2, the growth rate decreases and the size-distribution becomes narrower. In the framework of this model, size-focusing can be reached under condition where the supersaturation is high enough for NCs with mean size (Figure 1.4). This model was well performed and proofed experimentally in number of reports on CdSe and InAs NCs synthesis performed via “hot-injection” approach.

- 8 -

Motivation and Introduction Chapter 1

Figure 1.5. In situ synchrotron SAXS/WAXS monitoring of the formation of CdSe NCs. Top left: Progress in time of SAXS (A) and WAXS (B) patterns. Top right: Transient progress of the imposed temperature and of different characteristic parameters obtained from Monte Carlo fitting of the X-ray data. Bottom: Scathe of step by step CdSe NC formation. Reproduced with permission from ref. 37. Copyright 2016 American Chemical Society.

Despite the outstanding and fundamental insight of the LaMer nucleation theory, many experimental findings do not completely fellow the prediction. Impressive results were obtained by Abécassis and co-workers43 for in situ synchrotron SAXS/WAXS study on CdSe NCs synthesis with detailed analysis on nucleation and further NC growth (Figure 1.5).

- 9 -

Motivation and Introduction Chapter 1

Obtained results show that the nucleation stage was longer than theoretically presume (10 s vs 1 s) and proposed a generally low energetic barrier for nuclei formation. As a consequence lowering the Gibbs free energy leads to: dropping of the solid/liquid surface tension in the presence of surface ligands and quite low solubility of monomers in nonpolar organic solvents like 1-octadecene43. Conclusion of that experiments were that the growth rate was on 8 order of magnitude lower than previously expected in theoretical studies, and that NC growth is restrained by the precursor transformation kinetics37. (ii) Heat-up. Another fashion of NC synthesis in organic solvents is the “heat-up” strategy32. This approach has become more attractive than conventional “hot-injection” due to the higher reproducibility rate and easiness of up-scaling. The main stages of NC growth in that system include: monomer formation (yellow region), nucleation (red region), growth (blue region) and equilibrium (Figure 1.6a). Here, the monomer generation occurs until a critical temperature and concentration for homogeneous nucleation is reached. In comparison to the LaMer model for “hot-injection” approach, the nucleation stage is prolonged, which arises from the progressive monomer production accelerated upon temperature increasing 32. As a result, supersaturation is preserved for a prolonged period as well. In that system nucleation and growth are overlapping and that leads to broader NCs size distribution. Nonetheless, the continuous monomer formation and nuclei high grow rate are consistent with size focusing condition r>2r*. Similarly to “hot-injection” synthesis, the precursor’s reactivity here is a very important factor, especially at lower temperatures where all precursors, solvent and surfactants are loaded together in reaction flask and must be unreactive at preparation stage to avoid uncontrolled monomer and nuclei formation. Therefore, precise and right choice of the precursor's reactivity does yield high-quality NCs in the end. Another important parameter in this system is the heating rate. Figure 1.6b depicts the dependence between heating rate and NCs diameter and their concentration in solution. With higher heating rate, semiconductor NCs yield a narrower size-distribution as well as full half width maximum (FHWM) in PL32. This underlines that comparable semiconductor NCs quality can be obtained via “heat-up” approach and enabled the synthesis of monodisperse metal sulphide 44 45 46 47 47 48 48 NCs such as Cu2-xS , CuInS2 , Cu2ZnSnS4 , PbS , ZnS , CdSe , CdSe/CdS etc.

- 10 -

Motivation and Introduction Chapter 1

Figure 1.6. a) Evolution of the characteristic reactions in the “heat-up” synthesis of NCs. b) Effect of heating rate on the final particle size, size distribution, and concentration. Reproduced with permission from ref. 32. Copyright 2016 American Chemical Society.

Figure 1.7. Examples of different semiconductor NCs synthesized via “heat-up” approach. Cu2ZnSnS4, Cu2-xS, CdSe and CdSe/CdS NCs are adapted with permission from ref. 46, 44, 48 respectively. Copyright 2016, 2015, 2003 American Chemical Society. PbS and ZnS NCs are adapted with permission from ref. 47. Copyright 2015 AAAS www.sciencemag.org.

- 11 -

Motivation and Introduction Chapter 1

(iii) Solvent-antisolvent crystallization. As discussed above, supersaturation may be generated by changing the solubility of the system. An alternative way to reach supersaturation is the addition of an antisolvent – a miscible solvent which reduces solute solubility in the mixed solvent49. The addition of an antisolvent reduces the precursor solubility in the mixed medium (precursor solvent and antisolvent). This phenomena is well known in organic chemistry as material purification/solvent re-crystallisation process. In theory, nucleation occurs due to the reduction of Gibbs free energy, which is depends on the critical size of a cluster. This cluster is unstable as long as the surface part of Gibbs free energy is larger than the volume part. If the critical size is reached, the critical Gibbs energy is overcome and therefore the crystal starts to grow. Gibbs critical energy can be interpreted as an energy barrier for the nucleus growth. Equation 1.3 shows the critical Gibbs energy of this process, where ѵ is the molecular volume, γ is the surface energy, S supersaturation and kT the product of Boltzmann constant and temperature.

16 π γ3ѵ2 ΔGcrit = 2 (1.3) 3 (k T ln(S)) One significant advantage of the solvent-antisolvent extraction technique is that synthesis can be carried out at ambient conditions, thus protecting sensitive materials from heat. However, problems can occur if the antisolvent reduces the solubility of the precursor salts too strongly, resulting in un-reacted precursor crystal formation rather than the desired material phase formation. Solvent-antisolvent crystallization approach was successfully applied to perovskite NCs, microcrystal and single crystal synthesis (Figure 1.8) and will be discussed im more details in Chapter 4.

Figure 1.8. Variation of the size and morphology control of CH3NH3PbBr3 perovskite synthesized via solvent-antisolvent crystallization approach; this work.

- 12 -

Motivation and Introduction Chapter 1

1.2.2. Optical properties of semiconductor nanocrystals

Optical properties of colloidal semiconductor NCs (or quantum dots - QDs) are dominated by their size-dependence of the electronic structure which is reflected in absorption and photoluminescence (PL). In QDs electron and hole states are discrete and can be precisely adjusted by size. As a result, PL can cover a very broad spectral region as a function of QD size (Figure 1.9).

Figure 1.9. Typical example of size-dependence optical properties of colloidal CdSe QDs in solutions. Picture taken from http://nanocluster.mit.edu/research.php.

This unique properties make this class of materials highly attractive for numerous optoelectronic applications (Figure 1.1). As was already mentioned in Section 1.2, their outstanding optical properties were discovered by Ekimov4 in 1981 in form of emended NCs in glass matrixes, and later Brus3 performed size-controlled synthesis of CdS QDs in solution. The term “quantum dots” was firstly proposed by Reed50 in 1988 for “zero-dimensional semiconductor nanostructures”. In 1993 the monodisperse solution synthesis was performed by Murray, Norris, and Bawendi2. In a typical inorganic bulk semiconductor electron can be exited across the bandgap by light excitation. However, unlike to bulk materials, QDs create more complex valence and conduction bands that are similar to discrete electronic states in single atoms resulting in a wider band gap for smaller QDs51. Larger QDs demonstrate a more continuous electronic structure closer to their bulk analogues and the band gap can be easily tuned as a function of size (Figure 1.10). For example, the band gap of CdSe QDs can be adjusted from 1.74 (bulk) to 3 eV via size modulation. This change in band gap is also known as quantum confinement effect. In pioneering work of Lois Brus3 this phenomena was described by a “particle-in-a-box” model and can explain the absorption features of semiconductor NCs (Figure 1.11a-b). Photoluminescence properties of typical 4.1 nm QDs with well resolved absorption spectra can only be understood by taking into account the fine structure of the

- 13 -

Motivation and Introduction Chapter 1

band-edge 1S(e)–1S3/2(h) transition (e-electron; h-hole; 3/2 – subband level) (Figure 1.11a) which is also an indicator of high material quality and monodispersity of QDs solution52. Other peaks in the high energy region are consistent with subbands transitions (Figure 1.11b). Time-resolved spectroscopy showed peculiar features in recombination dynamics of CdSe QDs related to single-exciton radiative lifetimes51.

Figure 1.10. The electronic structure of the semiconductor material in bulk, quantum dots and molecule system in relation to the size.

Figure 1.11. a) Absorption spectra of 4.1 nm CdSe NCs with well-resolved transitions contain states the 1S and 1P electron states. Reproduced and adapted with permission from ref. 52. Copyright 2000 American Chemical Society. b) Closer look into the CdSe NCs band gap that shows more complex structure of quantized states of valence subbands.

- 14 -

Motivation and Introduction Chapter 1

The “particle-in-a-box” model (or effective mass approximation) quantum-confinement effect can be described as a confinement of the exciton (electron-hole pair). In QDs confinement occurs if the exciton Bohr radius is smaller than the size of quantum dots. For example the exciton Bohr radius for CdSe is 5.9 nm. Below this value, weak and strong confinement can be observed in absorption or PL spectra via a characteristic blue shift. The confinement energy can be calculated as follow53: ℏ2휋2 1 1 ℏ2휋2 퐸푐표푛푓푖푛푒푚푒푛푡 = 2 ( + ) = 2 (1.4) 2푎 푚푒 푚ℎ 2휇푎 where me is the effective mass of the electron, mh is the effective mass of the hole, µ is the reduced mass of the exciton system, and a is the radius of the QD. If the QD size is similar to the Bohr exciton radius, the QD is in the “weak confinement regime” resulting in insignificant changes in band gap and optical properties54. In the “strong confinement regime” where the QD size is smaller than the Bohr exciton radius, confinement is more influential and energy levels do no longer form a continuous spectrum as seen in Figure 1.9 and Figure 1.11a (Bohr radius of 5.9 nm vs. QD size of 4.1 nm)54. In that case the Bohr exciton radius is calculated by:53 푚 푎∗ = 푎 휀 ( ) (1.5) 푏 푏 푟 휇 * where ab , ab, εr is the is Bohr exciton radius, Bohr radius (about 0.53Å), and the dielectric constant of the semiconductor respectively. Interaction between electron and hole of the exciton can be described as Coulombic attraction:53 1 µ 퐸푒푥푐푖푡표푛 = 2 푅푦 (1.6) 휀푟 푚푒 where εr is the size-dependent dielectric constant of the semiconductor and Ry is the Rydberg energy (~13.6 eV). Then, the band gap energy of a QD (EQD) is calculated by taking into 53 account the band gap of the bulk material (EBulk) : ℏ2휋2 1 µ 퐸푄퐷 = 퐸퐵푢푙푘 + 2 + 2 푅푦 (1.7) 2휇푎 휀푟 푚푒

- 15 -

Motivation and Introduction Chapter 1

1.3. Metal halide perovskites

Three dimensional (3D) organic-inorganic metal halide perovskites evolved to top promising materials for optoelectronic applications in the past decade due to their direct bandgap, high absorption coefficient, low exciton binding energies and cost effective solution fabrication55. The potential of solar cells based on this semiconductor material is proven to reach certified power conversion efficiencies of 22%.56 Perovskites were discovered in 1839 by the German mineralogist Gustave Rose57, the name was given in honor to Russian mineralogist Lev Alekseevich von Perovski. The original 58 discovered mineral was calcium titanium oxide (CaTiO3). Later in 1958, Møller has published a paper reporting photoconductive properties of CsPbX3 (X=Cl, Br, I) perovskites. After 20 years, Weber synthesized hybrid organic inorganic perovskites with the formula

CH3NH3PbX3 (X=Cl, Br, I) and describe their structural as well as photoconductive properties59. Significant investigations were made by Mitzi in the 90s and 2000s60. Low interest into that class of materials for device application continued until Kojima et al.61 incorporate this material in dye sensitized solar cell with of PCE 3.8% in 2009. Henry Snaith in 201262 occurred a breakthrough to a 10% solar cell starting the new era of efficient and cost effective perovskite solar cells. Despite the success in photovoltaic, this newcomer material was quickly identified as a very good semiconductor for multiple optoelectronic applications63 due to the easiness of synthesis as well as band gap engineering (Figure 1.12b).

Figure 1.12. a) Perovskite metal halide structure. b) Applications based on metal halide perovskites.

- 16 -

Motivation and Introduction Chapter 1

The three-dimensional (3D) crystal structure of lead halide perovskites is shown in Figure 12a, where A-site cations are methylammonium (MA), formamidinium (FA), or inorganic 4- Cesium (Cs) which create together with BX6 a neutral charge balance in the perovskite material. The A-site ions occupy defined states in the crystal lattice as shown and contribute a strong interaction between the organic and inorganic compounds. These A cations are essential to determine the tolerance factor (t) and with it the crystal symmetry. T is defined by the geometric constraints imposed by a rigid sphere model for the perovskite structure. The tolerance factor first developed by Goldschmidt64, describes the maximal size of an A cation by given radii for the B and X ions. For cubic perovskite the tolerance factor is defined by Equation 1.8. Empirically the tolerance factor for cubic perovskite is between 0.8 ≤ t ≤ 0.955. R + R t = A x (1.8) √2 (RB + RX) Perovskite with t > 1 obtain a high degree of symmetry and a more stable hexagonal structure as a result of larger A-sites and smaller B-sites. However, if t >> 1 the large increase of A-sites generates a reduction in dimensionality resulting in 2D, 1D or quantum dots perovskite structures55. If t < 1, the crystal is affected by compression of the B-X bond and tension on the 4- A-X bond due to stress compensation inside the crystal resulting in a tilting of BX6 octahedra and reduced symmetry to tetragonal structure65. The tolerance factor exhibits a trend of the band gap, if t reaches 1 the crystal structure has approached the highest packing symmetry, this also corresponds to a reduction in band gap. In structures with X-site = Iodine

(RX = 2.2 Å) and B-site = Lead (RB = 1.19 Å) the largest possible A - sphere for a cubic perovskite (t=1) can only be a small organic molecule consisting out of 2-3 atoms excluding hydrogen, owing to the C-C and C-N bonds length are in the order of 1.4 Å. Therefore A-site molecule MA exhibits cubic 3D perovskite crystals in combination with the Metal anions lead or tin and halide anions (Cl-, Br-, or I-) according to a report by Mitzi et al.60, 65 Fluoride as X-site obtains an undesirable tolerance factor and due to its small ionic radius is causing strain in the lattice structure.65 Furthermore and to establish a structure map of perovskites the octahedral factor (µ) was added by Li et al.66 to the theoretical model of the tolerance factor. The octahedral factor (µ) is defined by the ratio of the ionic radius of the B cation and the X anion: R µ = B (1.9) RX 4- The octahedral factor is strictly correlated to the BX6 octahedron and therefore crucial for perovskite formation. At µ > 0.442 perovskite formation occurs. However, the

- 17 -

Motivation and Introduction Chapter 1

perovskite phase will not form for values less than 0.442.66 Figure 1.13a shows the calculated values for t and µ of different organic-inorganic hybrid perovskite65. The crystal structure can not only be influenced by the tolerance factor and octahedral factor, also the temperature can cause phase transition. The phase transition depends on the orientational disorder of the organic molecule inside the perovskite network67. This is defined as order-disorder type (Figure 1.13b). The influence of the organic molecule is due to the hydrogen bond interaction with the eight halogens of the surrounding corner-sharing + 4- octahedras. For example in MAPbI3 perovskite the MA can tilt the BX6 octahedra depending on the temperature creating two different crystal structures due to thermal elongated bonds. Above 330 K the octahedra align along the same direction resulting in a perfect cubic structure, below 330 K the octahedra tilt toward each other along z-direction, generating a tetragonal crystal structure.60, 68 Perovskites change their crystal phase in general upon cooling from cubic to tetragonal to orthorhombic.65

Figure 1. 13: a) Calculated octahedral and tolerance factors different organic-inorganic hybrid perovskites. Reproduced from ref. 65. b) Orientational disorder in a organic-inorganic perovskites. Reproduced from ref. 67.

Also the dimensionality of the perovskite structure can simply be changed by the ratio of organic and inorganic precursor from zero dimensional ((CH3NH3)4PbI6·2H2O) to three 60 dimensional (CH3NH3PbX3) structures.

- 18 -

Motivation and Introduction Chapter 1

1.3.1. Synthesis of hybrid organic-inorganic perovskites

Different ways of synthesizing organic-inorganic metal halide perovskites have been reported69 as shown in Figure 1.14. The synthesis can be categorized in three different processes: vacuum, solution and hybrid processing. The vacuum process is, in general, accomplished by a co-evaporation (Fig. 4a). The organic and inorganic precursor salt is thermally evaporated simultaneously from two different sources, therefore a complex evaporation ratio of the two sources has to be chosen to create perovskite structures. With this process extreme uniform surfaces can be produced resulting in a high PCE of 15.4% for perovskite thin film solar cells. However, the co-evaporation process requires precursor salts, which are thermally stable for the evaporation process, a high vacuum and slow evaporation rates (Ås-1), which makes the process less attractive for large scale industrial application.60, 70

Figure 1.14. Various preparation techniques for perovskite films, single crystals and nanocrystals. Reproduced and adapted from ref. 69.

The most attractive synthesis for low cost applications is solution based processing. The general process protocol is based on a spin coating process: mixing the precursors in one solution, deposition of precursors, spinning and subsequent annealing (one-step process) or by a two-step precursor solution deposition followed by an annealing step (Figure. 1.14, top row). PCE of more than 15% can be reached by this technique as well. Nevertheless, the major disadvantage for large scale production of solution process is the spin coating step.71

- 19 -

Motivation and Introduction Chapter 1

An alternative to spin coating is the doctor blade approach. This was accomplished by Deng et al.72; his group was able to produce a thin film solar cell with a PCE of 15.1%. Doctor blade technique could replace the spin coating process step and made a major step to the roll-to-roll processing. Hybrid processes, for example the vapor-assisted solution process, involve a solution step for inorganic metal salt film formation and a subsequent annealing step with the presence of organic precursor vapor. This method guarantees a well-defined grain structure and a full surface coverage. However, it also involves high temperature and rather complex annealing.73 The sequential deposition process also requires a deposition of metal salt precursors on the substrate by blade coating or spin coating, before the conversion to perovskite in an organic salt solution can take place (Fig. 4d). Also grinding techniques in air with mortar and pestle of the salt precursors have been reported by Constantinos et al.74 Furthermore, the perovskite obtained by this method was not pure and showed un-reacted precursor phases. Brute force synthesis by heating up the precursor salt under vacuum showed also perovskite formation.74 Also perovskite single crystals can be produced by a simple solvent-assisted precipitation technique performed by Shi et al.75 High quality crystals of millimeter size with astonishing transport properties and carrier diffusion length of 10 µm have been produced with this technique. The biggest down side of this procedure however is the slow growth of such a crystal, which is approximately one week for a 5 mm cubic crystal. For the same purpose, Saidaminov et al.76 have proposed a new approach for perovskite metal halide single crystal growth so called “inverse crystallization” which is based on the retrograde solubility of perovskite upon heating. The main advantage of that method in comparison to solvent- assisted precipitation technique is that single crystals grow within few hours only (Figure 1.14, middle row). This technique allows to grow perovskite single crystals with different 77 77 76, 78 composition such as FAPbI3, FAPbBr3, MAPbX3 (X = Cl, Br, I) . The high luminescence of multicrystaline perovskite thin films motivated researcher to synthesize NCs colloids with outstanding optical properties especially in terms of very high PLQY.79 Various methods such as ligand assisted re-precipitation reaction and hot injection were used to obtain stable NCs solution (Figure 1.14, bottom row) and are further discussed in the next section.

- 20 -

Motivation and Introduction Chapter 1

1.3.2. Synthesis of hybrid all-inorganic and metal halide colloidal nanocrystals.

The first evidence on the synthesis of metal halide perovskite NCs was performed in 2009 by Kojima et al. as a sensitizers on of mesoporous titania in dye-sensitized solar cell. The NCs were formed after precursor spin-coating and further annealing yielded NCs with 2-3 nm in size.61 Then in 2014, Schmidt et al.80 reported wet-chemistry processing of colloidal 81 MAPbB3 NCs with high PLQY of 20% and later improved to the of 83%. This work inspired many groups to use perovskite NCs in luminescent application as well as to study their physical and chemical properties. Another fashion that was successfully applied for hybrid perovskite NC synthesis is the ligand-assisteed re-precipitation (LARP) reaction proposed by Zhang et al (Figure 1.15 a- c).33 A precursor solution is produced by dissolving the Pb-halide and MA-halide salts with organic ligands such as oleic acid (OA) or oleyilamine (OAm) in a polar solvent (DMF - N, N-dimethylformamide, DMSO - Dimethyl sulfoxide).

Figure 1.15. a) Illustration of ligand-assisted re-precipitation (LARP) technique applied to synthesize MAPbX3 NCs. b) Representative TEM and HRTEM image of the MAPbBr3 NCs. c) Photoluminescence spectra of MAPbX 3 NCs toluene solutions and corresponding optical images under room light as well as UV lamp. (d) Photoluminescence spectra and image of size-tailoring emission of brightly emissive MAPbX3 NCs, where the size was controlled by temperature of reaction. Reproduced from ref 79.

This solution was injected into antisolvents such as toluene o r chloroform under constant stirring and resulted in highly bright (PLQY 50-70%) and PL color emission tunable (400 – 750 nm) perovskite NCs33. The crystallization process of perovskite nanoparticles is controlled by the supersaturation induced by the solubility change upon solvent mixing. The kinetics of the crystallization is mainly controlled by the amines, thus influencing the size of the NCs, whereas the oleic acid suppresses NC aggregation and contributes to their colloidal stability.

- 21 -

Motivation and Introduction Chapter 1

Huang et. al.82 demonstrated the size-tunability of the bandgap of narrow size dispersion of

MAPbB3 perovskite NCs by using temperature to exert control over the LARP process (Figure 1.15d). Very high PLQY values ranging from 74% up to 93% with the corresponding emission peaks covering the range from 475 to 520 nm were found. For CH3NH3PbX3 (X=Br, I) perovskite NCs, hot injection technique was developed as well.83 However, this hybrid perovskite is less attractive as their all-inorganic pedant, mainly due to their lower humid stability causing the dissociation into of MA gas and Pb-halides or the formation zero-dimensional CH3NH3)4PbI6·2H2O. In 2015 the synthesis of all-inorganic

CsPbX3 (X=Cl, Br, I) monodisperse colloidal NCs was reported by Kovalenko group via hot-injection approach84 (Figure 1.16a-c). This NCs exhibit high PLQY of 50-90% and FHWM in range of 12-42 nm that makes them an ideal candidate for luminiscent application such as TV screens as they could cover over 140% of NTSC standard (Figure 1.16d-g).84

Later was shown that size-confined CsPbI3 NCs also can successfully perform as solar cell active material with a PCE of 10.77%.85

- 22 -

Motivation and Introduction Chapter 1

1.3.3. Band-gap tuning of perovskite NCs via anion exchange reactions

An alternative fashion to obtain tunable perovskite NCs is the post-preparative anion exchange reaction with freeze shape and NCs size upon the reaction, decreasing number of the synthesis and only one sample (CsPbBr3 for example) can be used to obtain mix-halide

NCs samples. This technique was successfully performed for CsPbX3 (X = Cl, Br, I) NC systems by Kovalenko86 and Manna87 (Figure 1.17a-b). Their PL emission spectra can be adjusted in the range of 410-700 nm and high PLQY of 20-80% was found. As halide anion source, multiple halide compounds can be used even in their solid form such as inorganic Zn- halide or Pb-halide salts.86-89 Most recently it was shown that photoinduced anion exchange occurs in halide-containing solvent dispersions (CH2I2, CH2Br2 etc.) of CsPbX3 NCs solution via free halide ion formation upon light irradiation.90 The simple and fast anion exchange was used for chemical reaction monitoring, where halide ions were produced in intermediate 91 stages of the reaction. As consequence, the PL of CsPbX3 NCs was changed by the postreatement due to anion exchange with the intermediate halide components. Post-preparative anion exchange was also reported for hybrid organic inorganic perovskite 92 NCs such as MAPbX3 (X = Cl, Br, I) and thin films thereof (Figure 1.17c-d). Optical emission and absorption spectra were changed by MA-halide solution adding. Also FAPbBr3 nanorods were converted via anion exchange to FAPbI3 nanorods preserving their size and morphology with lasing performance in near-infrared region.93

Figure 17. a,b) Sketch of anion-exchange raeactions within CsPbX3 crystal lattice and suitable reagents. c,d) Photographs of ion-exchanged, mixed-halide MAPb(Br/Cl)3 and MAPb(Br/I)3 NCs under room light. (d) UV-visible and PL spectra of MAPb(Br/Cl) and MAPb(Br/I)3 NC films. Reproduced from ref. 79.

- 23 -

Motivation and Introduction Chapter 1

1.3.4. Quantum-size effect in perovskite nanocrystals

Metal halide perovskites as semiconductor have shown quantum-size confinement similarly to conventional Cd-chalcogenide QDs.2 In form of colloidal NCs, a first example was reported by Protesescu for CsPbBr3 NCs in which the PL peak gradually shifted from 512 to 460 nm as function of size.84 For the same composition but with a nanoplatelets geometry, the thicknesses can be tuned from 1 single unit cell monolayer to 5, resulting PL tailoring from 400 nm to 520 nm by changing the temperature of hot-injection reaction which is based on the Protesescu et al. protocol (Figure 1.18c-e).94 Sichert et al.95 reported similar features for MAPbBr3 nanoplatelet suspensions, and monolayer thicknesses modulation was achieved via varying the ratio of two organic cations in the synthesis reaction (Figure 1.18a-b). More 96 precise and high quality nanoplatelet synthesis of MAPbX3 (X=Br, I) and FAPbBr3 NCs with very high PLQY (up to 90%) was performed by our group and will be further described in Chapter 4. Alternatively, size-dependent properties of metal-halide perovskites can be achieved by simple incorporation of perovskite precursors into mesostructured systems with different pore size and further solvent extraction via annealing.97-101

Figure 1.18. Quantum-size effect in perovskite metal halide NCs. a,b) Illustration of MAPbBr3 nanoplatelets with different number of unit cell numbers (n=1,2,3 and ∞) and their optical PL and absorbance spectra. The number of unit cell monolayers layers decreased with increasing octylamine content in the reaction mixture. c,d,e) Absorption and Photoluminescence spectra of CsPbBr3 nanocubes and nanoplatelets made of 1–5 unit cells monolayers with their correspondence TEM micrographs for the NCs synthesized at higher (150°C, green PL) and lower (90°C, blue PL) temperature. Reproduced from ref. 79.

- 24 -

Motivation and Introduction Chapter 1

1.3.5. Origin of the high photoluminescence quantum yield in perovskites NCs

Lead halide perovskites NC surprised researcher by its surprisingly bright luminescence and high PLQY (over 90%) in colloidal solutions without the need wide bad-gap shell (ZnS, CdS, ZnO) protecting of NCs core.84 In conventional QD systems such as CdSe, ZnSe, PbS or transitional metal (Mn and Cu) doped host lattice,102, 103 such shell is imperative to obtaining high PLQY. The low PLQY in metal-chalcogenide NCs system originated from intrinsic and surface defects (dandling bonds) which quench the luminescence emission and are mostly located within band gap. In metal halide perovskites, however, despite their high density of defect crystal states, they perform as a material with low trap density of and are not most insensetive to photophysical losses affecting such as carrier mobility, lifetime or recombination rates. This feature make this material class unique and explains the very high interest into that semiconductor durring the last 5 years. Very recently, 104 Kang and Wang reported on the intrinsic properties of point defects in CsPbBr3 from first- principles calculations. They show that most of the intrinsic defects induce shallow transition levels and only a minor defects with high formation energy can create deep transition levels (Figure 1.19a-b). This defect tolerance feature may caused by bonding-antibonding interaction between the conduction bands and valence bands. As consequence, CsPbBr3 perovskite can preserve excellent electronic quality despite the presence of defects and in form of NCs can exhibit very high PLQY.82, 84, 96, 105-107

Figure 1.19. (a) Calculated defect charge-transition levels for CsPbBr3. (b) Schematic representation of shallow levels formation . Reproduced with permission from ref. 104. Copyright 2017 American Chemical Society.

- 25 -

Motivation and Introduction Chapter 1

1.4. Bibliography

1. M. Faraday, Philosophical Transactions of the Royal Society of London, 1857, 147, 145-181. 2. C. B. Murray, D. J. Norris and M. G. Bawendi, Journal of the American Chemical Society, 1993, 115, 8706-8715. 3. L. E. Brus, The Journal of chemical physics, 1984, 80, 4403-4409. 4. A. I. Ekimov and A. A. Onushchenko, Jetp Lett, 1981, 34, 345-349. 5. H. Weller, U. Koch, M. Gutierrez and A. Henglein, Berichte der Bunsengesellschaft für physikalische Chemie, 1984, 88, 649-656. 6. J. M. Pietryga, Y.-S. Park, J. Lim, A. F. Fidler, W. K. Bae, S. Brovelli and V. I. Klimov, Chemical reviews, 2016, 116, 10513-10622. 7. A. J. Nozik, M. C. Beard, J. M. Luther, M. Law, R. J. Ellingson and J. C. Johnson, Chemical reviews, 2010, 110, 6873-6890. 8. D. V. Talapin, A. L. Rogach, I. Mekis, S. Haubold, A. Kornowski, M. Haase and H. Weller, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2002, 202, 145-154. 9. Y. Cao and U. Banin, Journal of the American Chemical Society, 2000, 122, 9692-9702. 10. M. A. Hines and G. D. Scholes, Advanced Materials, 2003, 15, 1844-1849. 11. J. E. Murphy, M. C. Beard, A. G. Norman, S. P. Ahrenkiel, J. C. Johnson, P. Yu, O. I. Mićić, R. J. Ellingson and A. J. Nozik, Journal of the American Chemical Society, 2006, 128, 3241-3247. 12. X. Liu, X. Wang, B. Zhou, W. C. Law, A. N. Cartwright and M. T. Swihart, Advanced Functional Materials, 2013, 23, 1256-1264. 13. V. Srivastava, W. Liu, E. M. Janke, V. Kamysbayev, A. S. Filatov, C.-J. Sun, B. Lee, T. Rajh, R. D. Schaller and D. V. Talapin, Nano Letters, 2017. 14. J. Lauth, F. E. Gorris, M. Samadi Khoshkhoo, T. Chassé, W. Friedrich, V. Lebedeva, A. Meyer, C. Klinke, A. Kornowski and M. Scheele, Chemistry of Materials, 2016, 28, 1728-1736. 15. S. C. Riha, B. A. Parkinson and A. L. Prieto, Journal of the American Chemical Society, 2009, 131, 12054-12055. 16. Y. Li, M. Ye, C. Yang, X. Li and Y. Li, Advanced Functional Materials, 2005, 15, 433- 441. 17. E. Jang, S. Jun and L. Pu, Chemical Communications, 2003, 2964-2965. 18. D. V. Talapin, A. L. Rogach, A. Kornowski, M. Haase and H. Weller, Nano letters, 2001, 1, 207-211. 19. Y. Justo, P. Geiregat, K. V. Hoecke, F. Vanhaecke, C. De Mello Donega and Z. Hens, The Journal of Physical Chemistry C, 2013, 117, 20171-20177. 20. M. Harrison, S. Kershaw, M. Burt, A. Eychmüller, H. Weller and A. Rogach, Materials Science and Engineering: B, 2000, 69, 355-360. 21. H. Shen, S. Wang, H. Wang, J. Niu, L. Qian, Y. Yang, A. Titov, J. Hyvonen, Y. Zheng and L. S. Li, ACS applied materials & interfaces, 2013, 5, 4260-4265. 22. D. V. Talapin, J.-S. Lee, M. V. Kovalenko and E. V. Shevchenko, Chemical reviews, 2009, 110, 389-458. 23. B. Abécassis, M. D. Tessier, P. Davidson and B. Dubertret, Nano letters, 2014, 14, 710- 715. 24. J. W. Grebinski, K. L. Hull, J. Zhang, T. H. Kosel and M. Kuno, Chemistry of Materials, 2004, 16, 5260-5272. 25. J.-t. Jang, S. Jeong, J.-w. Seo, M.-C. Kim, E. Sim, Y. Oh, S. Nam, B. Park and J. Cheon, Journal of the American Chemical Society, 2011, 133, 7636-7639.

- 26 -

Motivation and Introduction Chapter 1

26. Q. Pang, L. Zhao, Y. Cai, D. P. Nguyen, N. Regnault, N. Wang, S. Yang, W. Ge, R. Ferreira and G. Bastard, Chemistry of materials, 2005, 17, 5263-5267. 27. K. Miszta, J. De Graaf, G. Bertoni, D. Dorfs, R. Brescia, S. Marras, L. Ceseracciu, R. Cingolani, R. Van Roij and M. Dijkstra, Nature materials, 2011, 10, 872-876. 28. S. Ghosh, R. Gaspari, G. Bertoni, M. C. Spadaro, M. Prato, S. Turner, A. Cavalli, L. Manna and R. Brescia, ACS Nano, 2015, 9, 8537-8546. 29. I. Fedin and D. V. Talapin, J. Am. Chem. Soc, 2016, 138, 9771-9774. 30. C. Liu, P. Wu, T. Sun, L. Dai, Y. Ye, R. Ma and G. Qin, The Journal of Physical Chemistry C, 2009, 113, 14478-14481. 31. J.-H. Choi, H. Wang, S. J. Oh, T. Paik, P. Sung, J. Sung, X. Ye, T. Zhao, B. T. Diroll and C. B. Murray, Science, 2016, 352, 205-208. 32. J. van Embden, A. S. Chesman and J. J. Jasieniak, Chemistry of Materials, 2015, 27, 2246-2285. 33. F. Zhang, H. Zhong, C. Chen, X.-g. Wu, X. Hu, H. Huang, J. Han, B. Zou and Y. Dong, ACS nano, 2015, 9, 4533-4542. 34. Y. S. Zhao, H. Fu, A. Peng, Y. Ma, D. Xiao and J. Yao, Advanced Materials, 2008, 20, 2859-2876. 35. H. Kasai, H. S. Nalwa, H. Oikawa, S. Okada, H. Matsuda, N. Minami, A. Kakuta, K. Ono, A. Mukoh and H. Nakanishi, Japanese Journal of Applied Physics, 1992, 31, L1132. 36. V. K. LaMer and R. H. Dinegar, Journal of the American Chemical Society, 1950, 72, 4847-4854. 37. P. Reiss, M. Carrière, C. Lincheneau, L. Vaure and S. Tamang, Chemical Reviews, 2016, 116, 10731-10819. 38. H. Du, C. Chen, R. Krishnan, T. D. Krauss, J. M. Harbold, F. W. Wise, M. G. Thomas and J. Silcox, Nano Letters, 2002, 2, 1321-1324. 39. J. Park, K. An, Y. Hwang, J.-G. Park, H.-J. Noh, J.-Y. Kim, J.-H. Park, N.-M. Hwang and T. Hyeon, Nature materials, 2004, 3, 891-895. 40. N. R. Jana, Y. Chen and X. Peng, Chemistry of materials, 2004, 16, 3931-3935. 41. X. Peng, J. Wickham and A. Alivisatos, Journal of the American Chemical Society, 1998, 120, 5343-5344. 42. T. Sugimoto, Advances in Colloid and Interface Science, 1987, 28, 65-108. 43. B. Abécassis, C. Bouet, C. Garnero, D. Constantin, N. Lequeux, S. Ithurria, B. Dubertret, B. R. Pauw and D. Pontoni, Nano letters, 2015, 15, 2620-2626. 44. H. Shen, H. Wang, H. Yuan, L. Ma and L. S. Li, CrystEngComm, 2012, 14, 555-560. 45. D.-E. Nam, W.-S. Song and H. Yang, Journal of colloid and interface science, 2011, 361, 491-496. 46. A. Shavel, M. Ibáñez, Z. Luo, J. De Roo, A. Carrete, M. Dimitrievska, A. Genç , M. Meyns, A. Pérez-Rodrı́guez and M. V. Kovalenko, Chemistry of Materials, 2016, 28, 720-726. 47. M. P. Hendricks, M. P. Campos, G. T. Cleveland, I. Jen-La Plante and J. S. Owen, Science, 2015, 348, 1226-1230. 48. I. Mekis, D. V. Talapin, A. Kornowski, M. Haase and H. Weller, The Journal of Physical Chemistry B, 2003, 107, 7454-7462. 49. M. Giulietti and A. Bernardo, Crystallization by antisolvent addition and cooling, INTECH Open Access Publisher, 2012. 50. M. Reed, J. Randall, R. Aggarwal, R. Matyi, T. Moore and A. Wetsel, Physical Review Letters, 1988, 60, 535. 51. V. Klimov, in Nanocrystal Quantum Dots, Second Edition, CRC Press, 2010, DOI: 10.1201/9781420079272-c5, pp. 147-213. 52. V. I. Klimov, The Journal of Physical Chemistry B, 2000, 104, 6112-6123.

- 27 -

Motivation and Introduction Chapter 1

53. P. YU and M. Cardona, Fundamentals of Semiconductors: Physics and Materials Properties, Springer Berlin Heidelberg, 2013. 54. P. Guyot-Sionnest and M. Hines, Applied Physics Letters, 1998, 72, 686-688. 55. N.-G. Park, T. Miyasaka and M. Grätzel, 2016. 56. NREL, Best Research-Cell Efficiencies, http://www.nrel.gov/ncpv/images/efficiency_chart.jpg 2016. 57. G. Rose, Annalen der Physik, 1839, 124, 551-573. 58. C. K. MØLLER, Nature, 1958, 182, 1436-1436. 59. D. Weber, Zeitschrift für Naturforschung B, 1978, 33, 1443-1445. 60. D. B. Mitzi, Progress in Inorganic Chemistry, Volume 48, 2007, 1-121. 61. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, Journal of the American Chemical Society, 2009, 131, 6050-6051. 62. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-647. 63. J. Chen, S. Zhou, S. Jin, H. Li and T. Zhai, Journal of Materials Chemistry C, 2016, 4, 11-27. 64. V. M. Goldschmidt, Naturwissenschaften, 1926, 14, 477-485. 65. Q. Chen, N. De Marco, Y. M. Yang, T.-B. Song, C.-C. Chen, H. Zhao, Z. Hong, H. Zhou and Y. Yang, Nano Today, 2015, 10, 355-396. 66. C. Li, X. Lu, W. Ding, L. Feng, Y. Gao and Z. Guo, Acta Crystallographica Section B: Structural Science, 2008, 64, 702-707. 67. F. Brivio, A. B. Walker and A. Walsh, Apl Materials, 2013, 1, 042111. 68. A. Glazer, Acta Crystallographica Section B: Structural Crystallography and Crystal Chemistry, 1972, 28, 3384-3392. 69. Y. Chen, M. He, J. Peng, Y. Sun and Z. Liang, Advanced Science, 2016, 3. 70. M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395-398. 71. J.-H. Im, H.-S. Kim and N.-G. Park, Apl Materials, 2014, 2, 081510. 72. Y. Deng, E. Peng, Y. Shao, Z. Xiao, Q. Dong and J. Huang, Energy & Environmental Science, 2015, 8, 1544-1550. 73. Q. Chen, H. Zhou, Z. Hong, S. Luo, H.-S. Duan, H.-H. Wang, Y. Liu, G. Li and Y. Yang, Journal of the American Chemical Society, 2013, 136, 622-625. 74. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorganic chemistry, 2013, 52, 9019-9038. 75. D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, A. Buin, Y. Chen, S. Hoogland, A. Rothenberger and K. Katsiev, Science, 2015, 347, 519-522. 76. M. I. Saidaminov, A. L. Abdelhady, B. Murali, E. Alarousu, V. M. Burlakov, W. Peng, I. Dursun, L. Wang, Y. He and G. Maculan, Nature communications, 2015, 6. 77. M. I. Saidaminov, A. L. Abdelhady, G. Maculan and O. M. Bakr, Chemical Communications, 2015, 51, 17658-17661. 78. G. Maculan, A. D. Sheikh, A. L. Abdelhady, M. I. Saidaminov, M. A. Haque, B. Murali, E. Alarousu, O. F. Mohammed, T. Wu and O. M. Bakr, The journal of physical chemistry letters, 2015, 6, 3781-3786. 79. H. Huang, L. Polavarapu, J. A. Sichert, A. S. Susha, A. S. Urban and A. L. Rogach, NPG Asia Materials, 2016, 8. 80. L. C. Schmidt, A. Pertegás, S. González-Carrero, O. Malinkiewicz, S. Agouram, G. Minguez Espallargas, H. J. Bolink, R. E. Galian and J. Pérez-Prieto, Journal of the American Chemical Society, 2014, 136, 850-853. 81. S. Gonzalez-Carrero, R. E. Galian and J. Pérez-Prieto, Journal of Materials Chemistry A, 2015, 3, 9187-9193. 82. H. Huang, A. S. Susha, S. V. Kershaw, T. F. Hung and A. L. Rogach, Advanced Science, 2015, 2.

- 28 -

Motivation and Introduction Chapter 1

83. O. Vybornyi, S. Yakunin and M. V. Kovalenko, Nanoscale, 2016, 8, 6278-6283. 84. L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Krieg, R. Caputo, C. H. Hendon, R. X. Yang, A. Walsh and M. V. Kovalenko, Nano letters, 2015, 15, 3692-3696. 85. A. Swarnkar, A. R. Marshall, E. M. Sanehira, B. D. Chernomordik, D. T. Moore, J. A. Christians, T. Chakrabarti and J. M. Luther, Science, 2016, 354, 92-95. 86. G. Nedelcu, L. Protesescu, S. Yakunin, M. I. Bodnarchuk, M. J. Grotevent and M. V. Kovalenko, Nano letters, 2015, 15, 5635-5640. 87. Q. A. Akkerman, V. D’Innocenzo, S. Accornero, A. Scarpellini, A. Petrozza, M. Prato and L. Manna, Journal of the American Chemical Society, 2015, 137, 10276-10281. 88. P. Ramasamy, D.-H. Lim, B. Kim, S.-H. Lee, M.-S. Lee and J.-S. Lee, Chemical communications, 2016, 52, 2067-2070. 89. T. Zhang, G. Li, Y. Chang, X. Wang, B. Zhang, H. Mou and Y. Jiang, CrystEngComm, 2017. 90. D. Parobek, Y. Dong, T. Qiao, D. Rossi and D. Son, Journal of the American Chemical Society, 2017. 91. T. L. Doane, K. L. Ryan, L. Pathade, K. J. Cruz, H. Zang, M. Cotlet and M. M. Maye, ACS Nano, 2016, 10, 5864-5872. 92. D. M. Jang, K. Park, D. H. Kim, J. Park, F. Shojaei, H. S. Kang, J.-P. Ahn, J. W. Lee and J. K. Song, Nano letters, 2015, 15, 5191-5199. 93. Y. Fu, H. Zhu, A. W. Schrader, D. Liang, Q. Ding, P. Joshi, L. Hwang, X. Zhu and S. Jin, Nano letters, 2016, 16, 1000-1008. 94. Y. Bekenstein, B. A. Koscher, S. W. Eaton, P. Yang and A. P. Alivisatos, J. Am. Chem. Soc, 2015, 137, 16008-16011. 95. J. A. Sichert, Y. Tong, N. Mutz, M. Vollmer, S. Fischer, K. Z. Milowska, R. Garcı́a Cortadella, B. Nickel, C. Cardenas-Daw and J. K. Stolarczyk, Nano letters, 2015, 15, 6521-6527. 96. I. Levchuk, P. Herre, M. Brandl, A. Osvet, R. Hock, W. Peukert, P. Schweizer, E. Spiecker, M. Batentschuk and C. J. Brabec, Chemical Communications, 2017, 53, 244- 247. 97. D. N. Dirin, L. Protesescu, D. Trummer, I. V. Kochetygov, S. Yakunin, F. Krumeich, N. P. Stadie and M. V. Kovalenko, Nano Letters, 2016, 16, 5866-5874. 98. M. Anaya, A. Rubino, T. C. Rojas, J. F. Galisteo‐López, M. E. Calvo and H. Míguez, Advanced Optical Materials, 2017. 99. V. Malgras, S. Tominaka, J. W. Ryan, J. Henzie, T. Takei, K. Ohara and Y. Yamauchi, Journal of the American Chemical Society, 2016, 138, 13874-13881. 100. V. Malgras, J. Henzie, T. Takei and Y. Yamauchi, Chemical Communications, 2017, 53, 2359-2362. 101. S. Demchyshyn, J. Roemer, H. Groiss, H. Heilbrunner, C. Ulbricht, D. Apaydin, U. Rütt, F. Bertram, G. Hesser and M. Scharber, arXiv preprint arXiv:1607.04661, 2016. 102. W. Zhang, X. Zhou and X. Zhong, Inorganic chemistry, 2012, 51, 3579-3587. 103. I. Levchuk, C. Würth, F. Krause, A. Osvet, M. Batentschuk, U. Resch-Genger, C. Kolbeck, P. Herre, H. Steinrück and W. Peukert, Energy & Environmental Science, 2016, 9, 1083-1094. 104. J. Kang and L.-W. Wang, The Journal of Physical Chemistry Letters, 2017, 8, 489-493. 105. I. Levchuk, A. Osvet, X. Tang, M. Brandl, J. D. Perea, F. Hoegl, G. J. Matt, R. Hock, M. Batentschuk and C. J. Brabec, Submitted, 2017. 106. L. Protesescu, S. Yakunin, S. Kumar, J. Bär, F. Bertolotti, N. Masciocchi, A. Guagliardi, M. Grotevent, I. Shorubalko, M. I. Bodnarchuk, C.-J. Shih and M. V. Kovalenko, ACS Nano, 2017, DOI: 10.1021/acsnano.7b00116.

- 29 -

Motivation and Introduction Chapter 1

107. L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Bertolotti, N. Masciocchi, A. Guagliardi and M. V. Kovalenko, Journal of the American Chemical Society, 2016, 138, 14202-14205.

- 30 -

Materials and Methods Chapter 2

Chapter II

Materials and Methods

This chapter introduce to the materials characterizations, nanocrystals synthesis technique, device fabrication and characterization of thereof which were used in all experiments of this thesis. The most important for nanocrystals characterization TEM and HRTEM measurements was conducted with great financial support by the DFG through the research training group 1896 ‘‘In situ microscopy with electrons, X-rays and scanning probes’’. Electron microscopy resources have been kindly provided by the Center for Nanoanalysis and Electron Microscopy (CENEM) and all TEM and HRTEM measurements were performed by Patrick Herre (LFG).

- 31 -

Materials and Methods Chapter 2

2.1. Materials

In this section all materials which were used in NCs synthesis or device fabrication are separated according to the chapter number. All chemicals were used as received without further purification.

Zn1−xCdxS /ZnS core-shell NCs (Chapter 3).

Zinc acetate (Zn(OAc)2, 99.99%), cadmium oxide (CdO, 99.99+%), manganese acetate

(Mn(OAc)2, 99.99%), sulfur powder (99.99%), dodecanethiol (DDT, 99.9%), oleic acid (OA, 90%), oleylamine (OAm, 70%, technical grade), and 1-octadecene (ODE, 90%) were purchased from Sigma-Aldrich. Monocrystaline Si solar cell (41 x 19.5 x 0.3 mm) for down- shifting experiments was purchased from LEMO-SOLAR GmbH.

Perovskite NCs and Ink (Chapter 4, 5).

Methylamine (CH3NH2, 40% in methanol) was purchased from ABCR GmbH & Co KG.

Formamidine acetate salt (HN=CHNH2×CH3COOH, 99%), lead (II) iodide

(PbI2, 99.999% trace metals basis), lead (II) bromide (PbBr2, 99.999% trace metals basis),

(PbCl2, 99.999% trace metals basis), Germanium (II) bromide (GeBr2, 97% ), Germanium

(II) iodide (GeI2, ≥ 99.8% trace metals basis), Manganese (II) chloride (MnCl2, ≥ 98%),

Tin (II) chloride (SnCl2, 99.999% trace metals basis), Tin (II) Bbromide (SnBr2), Tin (II) iodide (SnI2, 99.999% trace metals basis), mercaptopropylisobutyl-POSS (99%), Ethylcellulose, Terpineol (95%), N,N-Dimethylformamide (DMF, anhydrous, 99.8%), γ- Butyrolactone (GBL, 99%), N-methylformamide (NMF, 99%) hydroiodic acid (57% in water), hydrobromic acid (48% in water, 99.99%), acetonitrile anhydrous (CH₃CN, 99.8%) and hexane anhydrous (95%) were purchased from Sigma-Aldrich. Chloroform (>99%), toluene (>99%), tetrabutylammonium iodide (TBAI, 98.0%), tetrabutylammonium chloride (TBACl, 95.0%) and tetrabutylammonium bromide (TOABr, 98%) were purchased from Merck Millipore. Oleylamine (approximate C18-content 80-90%) was purchased from Acros Organics. Oleic acid (97%) was purchased from VWR Chemicals.

Methylammonium iodide was synthesized in our lab. Briefly, in a 500 ml flask, 27.86 ml

CH3NH2, 40% in methanol, was mixed with 100 ml of ethanol. Then, at room temperature, 30 ml 57% water solution of the HI was added dropwise with continuous stirring. Obtained solution was placed in rotary evaporator at 60°C for removing all solvents. Then after several time washing with diethyl ether, MAI was dissolved in ethanol and precipitated with diethyl ether twice.

- 32 -

Materials and Methods Chapter 2

Then, for obtaining highly pure crystals, MAI was dissolved in 80 ml of hot ethanol and placed in refrigerator at -3°C for re-crystallization.

Methylammonium bromide was synthesized in our lab. In 500 ml flask, 27.86 ml

CH3NH2, 40% in methanol, was mixed with 100 ml of ethanol. Then8.5 ml 48% water solution of the HBr was added dropwise with continuous stirring at room temperature. Obtained solution was placed in rotary evaporator at 60°C for removing all solvents. After several times of washing with diethyl ether, MABr was dissolved in ethanol and precipitated with diethyl ether twice.

Formamidinium halides were synthesized according to method by Eperon et. al.1 Formamidinium iodide (FAI), formamidinium bromide (FABr) and formamidinium chloride (FACl) were synthesized by reaction of formamidinium acetate powder in a 2x molar excess of 57% w/w hydroiodic acid (for FAI), 48%w/w hydrobromic acid (for FABr) or 37%w/w hydrochloric acid (for FACl) at 0°C and was left stirring for 1h. Upon drying at 100°C, a yellow-white (FAI) or white (FABr, FACl) powder is formed. This was then washed with diethyl ether and recrystallized twice with ethanol, to form white crystals. Before use, it was dried overnight in a vacuum oven. Perovskite precursor purity control (Chapter 6).

Methylamine (CH3NH2, 40% in methanol) was purchased from ABCR GmbH & Co KG.

Lead acetate trihydrate (Pb(CH3COO)2×3H2O, 99.999% trace metals basis), lead (II) iodide

(PbI2,99.999% trace metals basis), hypophosphorous acid solution (H3PO2,50 wt% in

water),phosphorous acid (H3PO3, 99%),N,N-Dimethylformamide (DMF, anhydrous, 99.8%), Dimethyl sulfoxide (DMSO, 99.9%), ethanol (absolute, 99.8%), hydroiodic acid (57% in water, contains <1.5% hypophosphorous acid as stabilizer) were purchased from Sigma-

Aldrich. Commercially avalibleCH3NH3I was purchased from LumTec. Dimethyl sulfoxide- d6 (DMSO, 99.9 atom % D) was purchased from Deutero GmBH.

Methylammonium hypophosphite (MAH2PO2). 8,7 ml of 50% H3PO2 in H2O was mixed with 100 ml of ethanol. Then 29 ml CH3NH2 (40% in methanol) was slowly added under continuous stirring. After 10 min, solvents were extracted by a rotary evaporator at 60°C. Obtained highly viscous liquid was diluted with ethanol, and separated from ethanol by acetone. Small drops of CH3NH3H2PO2 were precipitated on the bottom of the flask, forms non-mixable biphasic mixture . Washing process by ethanol/acetone mixture was repeated twice. The bottom phase was separated, and dried in vacuum 60° for 1h.

- 33 -

Materials and Methods Chapter 2

Methylammoniumphosphite synthesis (MAH2PO3). In a 500 ml flask 6,3 g of H3PO3 was dissolved in 20 ml of water and further mixed with 100 ml of ethanol. Then, 29 ml CH3NH2 (40% in methanol) was slowly added under continuous stirring. After 10 min, solvents were extracted by a rotary evaporator at 60°C. Highly viscous liquid was diluted with methanol, and small white crystals were precipitated by acetone. Purification was repeated twice, and highly crystalline white powder was dried in vacuum oven at 30° for 5h

2.2. Nanocrystals synthesis

This section describe in details all steps of nanocrystals synthesis presented in this thesis.

Zn1−xCdxS /ZnS core-shell NCs (Chapter 3) Preparation of stock solutions. The Zn, Cd, and Mn stock solutions were prepared in an air-free glovebox. The. 0.1 M Zn stock solution was obtained by dissolving 0.220 g (1 mmol) of Zn(OAc)2 in 0.8 mL of OAm, and 9.2 mL of ODE at 160 °C. For the 0.1 M Cd stock solution, 0.128 g (1 mmol) of CdO were dissolved in 2.0 mL of OA, and 8.0 mL of ODE at 160 °C. The 0.01 M Mn stock solution was prepared by dissolving 0.017 g (0.05 mmol) of

Mn(OAc)2 in 1 mL of OAm, and 9 mL of ODE at 80 °C. For the preparation of the ZnS shell, we used a 0.4 M Zn stock solution, made by dissolving 0.878 g (4 mmol) of Zn(Ac)2 in 3.0 mL of OAm and 7 mL of ODE at 160 °C. The 0.4 M ODE-S solution was obtained by dissolving 0.128 g of sulfur (4.0 mmol) in 10.0 mL of ODE at 160 °C. All stock solutions were stored at 50 °C in a glovebox before use.

Synthesis of ZnxCd1‑xS:Mn NCs. All syntheses were carried out using Schlenk line under nitrogen atmosphere; isolation and purification of the NCs were carried out in air. In a typical procedure, 1 ml DDT and 2 ml ODE were filled into a 50-mL three-neck flask in a heating mantle, heated to 40 °C under vacuum, and kept under these conditions for 1 hour to remove oxygen and water. Then, the reaction mixture was cooled to room temperature and purged with nitrogen. Subsequently, 1.0 mL of a 0.1 M Zn stock solution (0.1 mmol), 1.0 mL of a 0.1 M Cd stock solution (0.1 mmol), 1 mL of a 0.01 M Mn stock solution (0.01 mmol), and 1.0 mL of a 0.4 M ODE-S solution (0.4 mmol) were added. The reaction mixture was heated to 230 °C with a heating rate of 12 °C/min under nitrogen flux and kept at this temperature to grow ZnxCd1‑xS:Mn NCs. For the monitoring of the reaction and the optimization of the reaction parameters with respect to the optical properties of our NCs, aliquots of the NCs dispersion were taken at different reaction times, injected into cold toluene to terminate the NC growth, and then diluted with chloroform for the spectroscopic measurements. After

- 34 -

Materials and Methods Chapter 2

completion of NC growth, the reaction mixture was cooled to 60 °C and 10 ml of toluene or chloroform was added. The NCs were precipitated by addition of an excess of ethanol, removed from the dispersion by centrifugation (8000 rpm, 10 min), and purified four times by a precipitation/re-dispersion method using a solvent mixture of chloroform and ethanol (2/4 volume mixture). The purified NCs were then re-dispersed in chloroform. To obtain

ZnxCd1‑xS:Mn NCs with different Zn/Cd ratios, the Zn/Cd precursor ratio was varied while all other variables were kept constant.

Deposition of ZnS shell. Shelling of the ZnxCd1‑xS:Mn cores with different layers of ZnS

was performed with the crude Mn:ZnxCd1‑xS reaction mixture after reaching temperature to 230°C when the emission of the core NCs was maximum. 3.0 ml of a 0.4M stock solution of the Zn precursor were injected into the reaction mixture in 0.5 ml portions drop by drop at time intervals of 15 min. For the monitoring and optimization of the shelling procedure with respect to maximum NCs brightness, NCs aliquots were taken before injection of a new portion of Zn stock solution and spectroscopically studied in chloroform. Purification of

ZnxCd1‑xS:Mn/ZnS NCs was performed similarly to that of the ZnxCd1‑xS:Mn NCs.

Scaling up of the NCs synthesis. The amount of all components used for the previously described synthesis of ZnxCd1‑xS:Mn were increased 40 times. Briefly, 40 ml DDT and 80 ml ODE were filled into a 500-ml three-neck flask and heated to 40°C under vacuum. At room temperature, 40 m of a 0.1 M Zn stock solution (4 mmol), 40 ml of a 0.1 M Cd stock solution (4 mmol), 40 ml of 0.01 M Mn stock solution (0.4 mmol), and 40 ml of a 0.4 M ODE-S solution (16 mmol) were added. The reaction mixture was heated to 230 °C to obtain ZnxCd1‑ xS:Mn NCs. For the coating with a ZnS Shell, the entire stock solution of the 0.4M Zn precursor (120 ml) was injected in six 10 ml portions drop by drop into the reaction vessel at intervals of 15 min at 230 °C. After completion of particle growth, the reaction mixture was cooled to 60 °C and 250 ml of toluene or chloroform were added. Purification was performed as previously described, taking into account the upscaling factor.

CH3NH3PbX3 (X=Br, I) perovskite nanoplatelets (NPLs) (Chapter 4). The synthesis perovskite NPLs was performed by ligand-assisted re-precipitation technique. All syntheses were carried at room temperature in air with average room humidity ~45%.

CH3NH3PbBr3 NPLs. CH3NH3Br (0.0112g, 0.1 mmol) and PbBr2 (0.0367, 0.1 mmol) were dissolved in 1 ml DMF forming 0.1 mM solution. Then, 200 μl of the Oleic acid and different amount of the Oleylamine for different thicknesses of NCs were added. Next, 100 μl

- 35 -

Materials and Methods Chapter 2

of this mixture was injected into 3ml toluene or chloroform. Bright emitting (PL = 447-520 nm) NPLs were formed within seconds. For purification, obtained NPL solution was precipitated once with acetonitrile/toluene (1:1) mixture followed by centrifugation at 9000 rpm for 2 min. Thus obtained NCs were fully dispersible in nonpolar solvents such as hexane, toluene, chloroform etc. Detailed synthesis procedure presented in Table 2.1:

Table 2.1. Parameters of synthesis of CH3NH3PbBr3 NPLs with different thicknesses

DMF, Oleic Oleylamine, Perovskite Toluene, PL peak, Volume of Thickness, Lateral ml acid, μl μl precursor, μl ml nm Acetonitrile/Toluen nm size, nm e mixture for washing, ml

1 200 16 100 3 514 2 2.6±0.2 13.8±1.8 1 200 18 100 3 511 2 - - 1 200 21 100 3 488 2.5 2±0.1 11.6±2.4 1 200 24 100 3 473 2.5 - - 1 200 28 100 3 452 3 - - 1 200 30 100 3 447 3.5 1.6±0.1 10.7±1.5 1 200 60 100 3 447 3.5 1.6±0.1 5.4±0.8

CH3NH3PbI3 NPLs. CH3NH3I (0.0159g, 0.1 mmol) and PbI2 (0.0461, 0.1 mmol) were dissolved in 1 ml of DMF forming 0.1 mM solution. Then, 200 μl of the Oleic acid and corresponding amounts of the Oleylamine for different thicknesses of NPLs were added. Next, 100 μl of this mixture was injected into anti-solvent (chloroform, toluene and chloroform-toluene mixture) media. Obtained NC solution (only for NCs with 722 nm 630 nm PL) and was precipitated by centrifugation with 15400 rpm at -10° for purification. Obtained precipitate was re-dissolved in hexane. Detailed synthesis procedure presented in Table 2.2:

Table 2.2. Parameters of synthesis of CH3NH3PbI3 NPLs with different thicknesses

DMF, Oleic acid, Oleylamine, μl Perovskite Chloroform, Toluene, Chlorofor PL peak, Lateral ml μl precursor, ml ml m nm size, nm μl + Toluene, ml 1 200 50 100 3 - - 722 22.3±6 1 200 100 100 3 - - 683 Wide distribution 1 200 150 100 3 - - 628 - 1 200 200 100 3 - - 592 - 1 200 250 10 - - 2 + 1 578 - 1 200 250 10 - 3 - 549 -

- 36 -

Materials and Methods Chapter 2

FAPbX3 (X=Cl, Br, I) Colloidal NCs (Chapter 4). The synthesis perovskite NPLs was performed by ligand-assisted re-precipitation technique. All syntheses were carried at room temperature in air with average room humidity ~45%.

FAPbCl3 NCs (416 nm PL). FACl (0.008 g, 0.1 mmol) and PbCl2 (0.028g, 0.1 mmol) were dissolved in 1 ml DMF forming 0.1 mM solution. Then, 200 μl of the Oleic acid and 20 μl Oleyilamine were added. Next, 100 μl of this mixture was injected into 3ml of chloroform. NCs were formed within seconds as a colorless solution. For purification, obtained NC solution was precipitated once with 1.5 ml of acetonitrile/toluene (1:1) mixture followed by centrifugation at 9000 rpm for 2 min. Thus obtained NCs were fully dispersible in nonpolar solvents such as hexane, toluene, chloroform etc.

FAPbBr3 NCs (533 nm PL). FABr (0.0112g, 0.1 mmol) and PbBr2 (0.0367, 0.1 mmol) were dissolved in 1 ml DMF forming 0.1 mM solution. Then, 200 μl of the Oleic acid and 40 μl Oleyilamine were added. Next, 100 μl of this mixture was injected into 3ml of chloroform. Bright green-emitting NCs were formed within seconds. For purification, obtained NC solution was precipitated once with 1.5 ml of acetonitrile/toluene (1:1) mixture followed by centrifugation at 9000 rpm for 2 min. Thus obtained NCs were fully dispersible in nonpolar solvents such as hexane, toluene, chloroform etc.

FAPbI3 NCs (737 nm PL). FAI (0.0159g, 0.1 mmol) and PbI2 (0.0461, 0.1 mmol) were dissolved in 1 ml of DMF forming 0.1 mM solution. Then, 200 μl of the Oleic acid and 150 μl Oleyilamine were added. Next, 100 μl of this mixture was injected into anti-solvent (chloroform) media (Video S1). For purification, obtained NC solution was precipitated once with 1 ml of acetonitrile/toluene (1:1) mixture followed by centrifugation at 9000 rpm for 2 min. Thus obtained NCs were fully dispersible in nonpolar solvents such as hexane, toluene, chloroform etc. Mixed-halide FAPb(Cl/Br)3 and FAPb(Br/I)3 nanocrystals were produced by precursor blending in desired ratio. A remark. The square shaped and/or sheet-like NCs, may build self-assembled super structure which cost colloidal stability2, 3. Therefore, additional steps was performed to improve colloidal stability: 20 μl of 1M TBACl (for FAPbCl3), 20 μl of 1M TOABr (for

FAPbBr3) or 20 μl of 1M TBAI (for FAPbI3) chloroform/toluene solution was added to NCs solution immediately after perovskite precursor injection. The washing process was performed accordingly.

Water resistance Silsesquioxane coating of FAPbX3 NCs by Mercaptopropylisobutyl- POSS was performed according to procedure published by Rogach group4.

- 37 -

Materials and Methods Chapter 2

Anion Exchange reaction were performed based on modified protocol of Akkerman et. al5 - for CsPbX3 (X=Cl, Br, I) nanocrystals. For I precursor, 0.370 g of TBAI was dissolved in 1 - ml chloroform and diluted by 2 ml of toluene. For I precursor, 0.546 g of TOABr was - dissolved in 3 ml of toluene. For Cl precursor, 0.278 g of TBACl was dissolved in 1 ml chloroform and diluted by 2 ml of toluene. Under continuous stirring, different quantities (normally ranging from 10 to 150 μL) of

halide precursors solution were swiftly injected into as sensitized 3 ml of FAPbX3 (X=Cl, I, Br) NCs crude solution. The exchange reaction finished rapidly within few seconds with changing colour of solution as well as luminescence under UV-light excitation. For purification, obtained NCs solution was precipitated once with acetonitrile/toluene (1:1) mixture followed by centrifugation at 9000 rpm for 2 min. Thus obtained NCs were fully dispersible in nonpolar solvents such as hexane, toluene, chloroform etc. For purified NCs, adding TBAX (X= Cl, Br,I) solution destroy NCs. Therefore, OAmX (X=Cl, Br, I) were used for washed NCs solution according to previously reported method5.

- 38 -

Materials and Methods Chapter 2

Protocol for FAPbX3 (Cl, Br, I) NPLs synthesis with tunable PL by thickness control

Table 2.3. Parameters of synthesis FAPbCl3 NCs with different thicknesses Volume of DMF, ml Oleic acid, μl Oleyilamine, Perovskite Chloroform, Acetonitrile/ PL peak, μl precursor ml Toluene mixture nm for injection, for washing, ml μl 1 200 20 100 3 1.5 416 1 200 60 100 3 1.5 398 1 200 110 100 3 1.5 388

Table 2.4. Parameters of synthesis FAPbBr3 NCs with different thicknesses Volume of Thicknesses, DMF, Oleic acid, Oleyilamine, Perovskite Chloroform, PL Acetonitrile/ nm ml μl μl precursor ml peak, Toluene for nm mixture for injection, μl washing, ml 1 200 40 100 3 533 1.5 2.6±0.2 1 200 50 100 3 526 1.5 - 1 200 60 100 3 496 2 - 1 200 80 100 3 486 2 2±0.1 1 200 100 100 3 483 3 - 1 200 120 100 3 476 3.5 - 1 200 140 100 3 470 3.5 - 1 200 150 100 3 438 3.5 1.4±0.1

Table 2.5. Parameters of synthesis FAPbI3 NCs with different thicknesses Volume of DMF, ml Oleic acid, Oleyilamine, Perovskite Chloroform, Acetonitrile/ PL peak, μl μl precursor ml Toluene mixture nm for injection, for washing, ml μl 1 200 150 100 3 1 737 1 200 180 50 3 1 698 1 200 220 20 3 1 668

- 39 -

Materials and Methods Chapter 2

Surfactant-free perovskite ABX3 (A= FA, MA; B=Pb, Sn, Ge, Mn, Eu; X=Cl, Br, I) nanocrystaline powder (Chapter 5). For the nanocrystaline powder synthesis a solvent-antisolvent extraction technique was performed, for all combinations of organic-inorganic metal halide perovskites. Therefore, the salt precursors had to be dissolved in an applicable solvent (Table 2.6 – 2.7) and corresponding antisolvent were added (Chloroform or Toluene) inducing nucleation and crystallization process (Figure 2.1). In our case only three different solvents were used to perform a generalized and up-scalable synthesis of perovskite nanocrystaline powder (NMF, DMF, and GBL) (Table 2.8). In all experiments 10-15 ml of Chloroform for each sample was used as antisolvent.

Figure 2.1. Solvent - antisolvent extraction technique

Table 2.6. Solvent-Antisolvent extraction process parameters

Perovskite Organic Metal precursor Precursor solvent Temperature Stirring time precursor MAPbI3 MAI 0.051 g PbI2 0.15 g GBL 3 ml 80 °C 3 h MAPbBr3 MABr 0.045 g PbBr2 0.15 g NMF 3 ml 60 °C 10 min MAPbCl3 MACl 0.035 g PbCl2 0.15 g NMF 3 ml 60 °C 10 min FAPbI3 FAI 0.055 g PbI2 0.15 g GBL 3 ml 80 °C 45 min FAPbBr3 FABr 0.050 g PbBr2 0.15 g NMF 3 ml 60 °C 20 min FAPbCl3 FACl 0.043 g PbCl2 0.15 g NMF 3 ml 60 °C 20 min MASnI3 MAI 0.063 g SnI2 0.15 g GBL 3 ml 60 °C 10 min MASnBr3 MABr 0.058 g SnBr2 0.15 g GBL 3 ml 60 °C 10 min MASnCl3 MACl 0.051 g SnCl2 0.15 g GBL 3 ml 60 °C 20 min FASnI3 FAI 0.069 g SnI2 0.15 g GBL 3 ml 60 °C 20 min FASnBr3 FABr 0.066 g SnBr2 0.15 g GBL 3 ml 60 °C 20 min FASnCl3 FACl 0.063 g SnCl2 0.15 g GBL 3 ml 60 °C 20 min MAGeI3 MAI 0.048 g GeI2 0.10 g GBL 3 ml 60 °C 20 min MAEuI3 MAI 0.038 g EuI2 0.10 g GBL 2 ml 60 °C 30 min MAMnCl3 MACl 0.052 g MnI2 0.10 g NMF 2 ml 60 °C 10 min

Table 2.7. Blend table for mixed halide perovskites

Mixed Precursor solution X Mol/L Blend-Volume Precursor solution Mol/L Blend-Volume of 2 Perovskite of 1 (mL) (mL)

MAPbCl2Br MACl + PbCl2 + NMF 0.18 1.35 MABr + PbBr2 + NMF 0.18 0.45

MAPbCl1.5Br1.5 MACl + PbCl2 + NMF 0.18 0.9 MABr + PbBr2 + NMF 0.18 0.9

MAPbClBr2 MACl + PbCl2 + NMF 0.18 0.45 MABr + PbBr2 + NMF 0.18 1.35

MAPbBr2I MABr + PbBr2 + NMF 0.18 1.35 MAI + PbI2 + GBL 0.18 0.45

MAPbBr1.5I1.5 MABr + PbBr2 + NMF 0.18 0.9 MAI + PbI2 + GBL 0.18 0.9

MAPbBrI2 MABr + PbBr2 + NMF 0.18 0.45 MAI + PbI2 + GBL 0.18 1.35

- 40 -

Materials and Methods Chapter 2

Table 2.8. Applicable solvent for the precursor materials Red: no perovskite precipitation; Green: perovskite forms

Metal salt+ organic salt NMF GBL

PbI2 + MAI

PbBr2 + MABr

(PbBr2 + MABr) + (PbI2 + MAI)

PbCl2 + MACl

(PbBr2 + MABr) + (PbCl2 + MACl)

PbI2 +FAI

PbBr2 +FABr

PbCl2 + FACl

SnI2 + MAI

SnBr2 + MABr

SnCl2 + MACl

SnI2 + FAI

SnBr2 + FABr Not stable

SnCl2 + FACl Not stable

GeI2 + MAI

GeI4 + MAI

GeI2 + FAI

GeBr2 + MABr

MnCl2 + MACl

EuI2 + MAI

EuCl2 + MACl

PbI2 + MAI + FAI

PbCl2 + MACl + FACl

- 41 -

Materials and Methods Chapter 2

LT-NiO synthesis (Chapter 6). The dry NiO nano powder was prepared by flame spray synthesis using a precursor solution of nickel acetate in 2-ethylhexanoic acid; the solution was diluted with 30 wt% xylene. The precursor then was fed (7 mlmin-1to a spray nozzle, dispersed by oxygen (9 l min-1) and ignited by a premixed methane-oxygen flame (CH4: 1.2 l min-1, O2: 2.2 l min-1). The off-gas was filtered through a steel mesh filter (20 micron mesh size) by a vacuum pump at about 20 m3 h-1. The obtained NiO nano powder was collected from the filter mesh. A stable suspension was prepared by dispersing 5 wt% NiO nano powder in methanol by using an undisclosed dispersant (proprietary information of Nanograde Ltd.).

2.3. Perovskite ink preparation (Chapter 5).

The retained perovskite powder was grinded in a mortar, to ensure that agglomerates are broken down. The powder was hand-grinded there for a time of about 8 minutes, until everything was homogenous (Figure 2.2). Subsequently, the powder was placed with toluene as dispersion media containing additives. Afterwards, a magnetic stirrer was given to the flask and stirred for several hours (15 – 20 hrs.) in order to make sure the particles were homogenously dispersed in the toluene solvent ( Figure 2.2).

Figure 2.2. Schematic representation of perovskite ink preparation.

2.4. Device fabrication

Deposition of ZnxCd1‑xS:Mn/ZnS NCs as luminescent down-shifting layer (Chapter 3).

ZnxCd1‑xS:Mn/ZnS NCs with PLQY of 70% were deposited directly onto the surface of monocrystaline Si solar cells (LEMO-SOLAR GmbH), previously cleaned with ethanol and dried, using a doctor blade technique. For the preparation of the coating solutions from

- 42 -

Materials and Methods Chapter 2

different concentrations of ZnxCd1‑xS:Mn/ZnS NCs, the NCs were precipitated from chloroform dispersions by addition of ethanol. The separated NCs were dried overnight in a vacuum oven and subsequently dispersed in toluene with different amounts, here 0.5-25 mg/ml. For the NC deposition, the temperature of the doctor blade hot plate was set to 40 °C, the blade gap was set to 400 μm, and 200 μl of each solution were coated with a speed of 10 mm/s. After NC coating, all solar cells were annealed at 80 °C for five min. to remove traces of organic solvent.

FAPbI3 and FAPbBr3 photodetectors fabrication (Chapter 4). Photodetector devices were fabricated by drop casting as-synthesized nanoparticles on a patterned ITO substrate with electrode distance of 20 um.

Perovskite solar cell fabrication (Chapter 6). First, The patterned ITO substrates were ultrasonic cleaned with acetone and isopropanol for 10 minutes each. On cleaned ITO substrate, the LT-NiO were deposited by spin coating at a speed of 4000 r.p.m and followed by annealing at 140 ◦C for 15min in air. Then, PbI2 and CH3NH3I mixed with mole ratio of 1:1 with concentration of 50% were stirred in a mixture of DMF and DMSO (2:1 v/v). After the coating of perovskite layer, a compact ∼50 nm thick layer of PCBM is spin coated by using a 2 wt% solution of PCBM in CB, then 5-10nm PrCMA is deposited by spinning a 0.15wt% solution in Methanol. Finally, a 100-nm-thick Ag electrode was deposited on top of PrCMA through a shadow mask by thermal evaporation.

2.5. Characterization techniques

Energy-Dispersive X-ray spectroscopy (EDX). Elemental analysis of the NCs was done with a Philips ESEM XL30 scanning electron microscope equipped with a field emission gun and operated at 10 kV. Transmission electron microscopy (TEM) and high resolution electron microscopy (HRTEM). TEM and HRTEM images were recorded using a PHILIPS CM 300 UT high- resolution transmission electron microscope (300 kV acceleration voltage, 0.17 nm point resolution at Scherzer defocus), equipped with a LaB6 filament and a CCD camera with an image size of 2048×2048 pixels. Samples for TEM were prepared by casting one drop of the NCs solution in hexane or toluene onto a standard copper grid coated with a continuous amorphous carbon film. The size distribution and thicknesses of NCs were obtained from the TEM image with ImageJ software. Fourier transform infrared spectroscopy (FTIR). FTIR measurements were done with a VERTEX-70 spectrometer (Bruker Optik GmbH).

- 43 -

Materials and Methods Chapter 2

X-ray Powder Diffraction (XRD). X-ray diffraction analysis was performed by classical ex situ Bragg – Brentano geometry using a PanalyticalX'pert powder diffractometer with filtered Cu-Kα radiation and an X'Celerator solid-state stripe detector. Concentrated chloroform/toluene solution of nanocrystals was drop casted on a zero background silicon wafer. X-ray photoelectron spectroscopy (XPS). The XP spectra were collected using a non-

monochromated Mg Kα anode operated at a power of 238 W, and an Omicron Argus electron analyzer; the pass energy was set to 200 eV for survey spectra and to 35 eV for detail spectra. The binding energy axis is referenced to the C 1s signal at 285.0 eV. The samples investigated by XPS were prepared by dripping the NC dispersion on a Mo-based sample holder, which was transferred to the XPS chamber using a load-lock system. To remove residual solvent, the sample was heated to 60°C in situ prior to the measurement.

Optical measurements. Absorption spectra of the NC dispersions were taken on a Perkin- Elmer Lambda 950 spectrometer and emission spectra on a Jasco Spectrofluorometer FP- 8500. The emission spectra shown were corrected for the wavelength-dependent spectral responsivity of the fluorometer. Lifetime measurements in the Chapter 3. The photoluminescence decay curves were measured by exciting the dispersed nanoparticle samples by a pulsed Xe lamp (pulse width 4 µs) through a 360 nm bandpass filter. The emission was dipersed by an H10 monochromator (Jobin-Yvon), detected with a 9816 photomultiplier (EMI) and recorded with a digital oscilloscope TDS540 (Tektronix).

Measurement of PLQY (pl. pl) in the Chapter 3 were measured absolutely with a calibrated integrating sphere setup at BAM previously reported.54,55This involves the

abs em determination of the absorbed photon flux ( qpem() ) and the emitted photon flux ( qpem() ) by a sample (see eq. 1) with an integrating sphere setup.54

 em2 II()()  x em b em d  em em pl  s()em q  em1 p pl ex  II()()  qabs b ex x ex d p  ex ex s()ex ex  (2.1)

The emitted photon flux follows from the blank and spectrally corrected emission spectrum

of the sample (Ix(em)) integrated over the emission band. The absorbed photon flux is

- 44 -

Materials and Methods Chapter 2

calculated from the difference between the transmitted excitation light with the blank (Ib(ex)) and the sample (Ix(ex)), within the wavelength region of the excitation bandwidth. s(em) and

s(ex) represent the spectral responsivity of the detection system in the spectral regions of emission and excitation. Measurement of PLQY in the Chapter 4 was carried out according to the method described by de Mello et al6. The samples were mounted in an integrating sphere, connected to an iHR 320 monochromator (Jobin-Yvon) with an optical fiber. The spectra were recorded with a Si CCD (Syncerity, Jobin-Yvon) camera and corrected for the spectral sensitivity of the setup, determined with the help of a calibrated Xe lamp (Hamamatsu). Lifetime measurements in the Chapter 4. The photoluminescence decay curves were measured on a time-resolved luminescence spectrometer Fluotime 300 (PicoQuant GmbH) under excitation with a pulsed diode laser 402 nm. The energy flux density was 0.5 μJ cm−2. Photoindused phase segregation and photostability experiments were performed with CW laser diodes working at 375 nm (Stradus-375-60) and 445 nm (Oxxius LBX-445). The samples were irradiated with the focused laser beam at intensities of 20 W/cm2 (for phase segregation experiment) and 3-7 W/cm2 (for UV-blue photostability experiment) The evolution of the photoluminescence spectra was monitored with a Si CCD camera attached to an iHR320 monochromator.

Photo-sensing of FAPbI3 and FAPbBr3 photodetectors All the electrical measurements were carried out using Keithley 236 source meter which was controlled by a self-made Labview program. A white LED was used for illumination purposes and the LED intensity was adjusted in a way that the generated photocurrent matched the one under a solar simulator. 1H, 13C and 31P NMR spectroscopy. 1H, 13C[1H], and 31P[1H] NMR spectra were recorded on a Bruker Avance 300 spectrometer operating at 300 MHz (1H NMR), 75 MHz (13C[1H] NMR), and 121 MHz (31P[1H] NMR). NMR spectra were referenced to the residual solvent signal (1H: DMSO, 2.49 ppm; 13C: DMSO, 39.5 ppm;  in parts per million (ppm); coupling constants are reported as observed) and recorded at ambient probe temperature (rt) unless otherwise noted. DMSO-d6 (99.5%, Deutero GmbH) was stored over 4 Å molecular sieves. Scanning Kelvin Probe Microscopy (KPFM). Scanning Kelvin Probe Microscopy was performed by using an AFM (Solver Nano, NT-MDT) with an AC voltage of 0.5V in air and in the dark. A gold-coated silicon probe (NSG30/Au, NT-MDT) with a force constant of

- 45 -

Materials and Methods Chapter 2

~22−100 N/m and a tip radius of ~35 nm was used for KPFM measurement. Samples were measured immediately after they were removed from nitrogen storage.

Monocrystaline Silicon Solar cell characterization in Chapter 3. A solar simulator (Oriel Sol 1A, Newport) with an AM1.5G spectrum operated at 100 mW cm−2 was used for illumination of the solar cells. A source measurement unit (Botest LIV) was used to record the current-voltage curves. EQE measurements were carried out with a quantum efficiency (QE) measurement system QE-R from Enlitech. Perovskite solar cell charachterization in Chapter 6. J-V characteristics of all the devices were measured using a source measurement unit from BoTest. The area of all measured devices were 0.104 cm2. Illumination was provided by a Newport Sol1A solar simulator with AM1.5G spectrum and light intensity of 100mW cm-2, which was determined by a calibrated crystalline Si-cell. During device characterization, a shadow mask with an opening of 10.4 mm2 was used. The EQE spectra were recorded with by an Enli Technology (Taiwan) EQE measurement system (QE-R), and the light intensity at each wavelength was calibrated with a standard single-crystal Si photovoltaic cell.

- 46 -

Materials and Methods Chapter 2

2.6. Bibliography

1. Eperon, G. E.; Stranks, S. D.; Menelaou, C.; Johnston, M. B.; Herz, L. M.; Snaith, H. J., Formamidinium lead trihalide: a broadly tunable perovskite for efficient planar heterojunction solar cells. Energy Environ. Sci. 2014, 7, 982-988. 2. Vybornyi, O.; Yakunin, S.; Kovalenko, M. V., Polar-solvent-free colloidal synthesis of highly luminescent alkylammonium lead halide perovskite nanocrystals. Nanoscale 2016, 8, 6278-6283. 3. Yang, J.; Fainblat, R.; Kwon, S. G.; Muckel, F.; Yu, J. H.; Terlinden, H.; Kim, B. H.; Iavarone, D.; Choi, M. K.; Kim, I. Y.; Park, I.; Hong, H.-K.; Lee, J.; Son, J. S.; Lee, Z.; Kang, K.; Hwang, S.-J.; Bacher, G.; Hyeon, T., Route to the Smallest Doped Semiconductor: Mn2+-Doped (CdSe)13 Clusters. J. Am. Chem. Soc. 2015, 137, 12776- 12779. 4. Huang, H.; Chen, B.; Wang, Z.; Hung, T. F.; Susha, A. S.; Zhong, H.; Rogach, A. L., Water resistant CsPbX3 nanocrystals coated with polyhedral oligomeric silsesquioxane and their use as solid state luminophores in all-perovskite white light-emitting devices. Chem. Sci. 2016, 7, 5699-5703. 5. Akkerman, Q. A.; Motti, S. G.; Srimath Kandada, A. R.; Mosconi, E.; D’Innocenzo, V.; Bertoni, G.; Marras, S.; Kamino, B. A.; Miranda, L.; De Angelis, F.; Petrozza, A.; Prato, M.; Manna, L., Solution Synthesis Approach to Colloidal Cesium Lead Halide Perovskite Nanoplatelets with Monolayer-Level Thickness Control. Journal of the American Chemical Society 2016, 138, 1010-1016. 6. de Mello, J. C.; Wittmann, H. F.; Friend, R. H., An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater. 1997, 9, 230-232.

- 47 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Chapter III

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells

The low efficiencies of inorganic solar cells (Si, CdTe, CuInSe2 etc.) in the ultraviolet (UV) and blue spectral region are key factors which restrict their power conversion efficiency (PCE). A simple and reliable technological step to overcome these limitations is the coating of the solar cells by UV-light conversion materials. This Chapter describe a simple one-pot synthesis of highly luminescence Zn1-xCdxS:Mn/ZnS core-shell (NCs) with record quantum yield of 70%. With our NCs and simple coating technique, the efficiency of a commercial Si solar cell was improved up to 12 % in the UV spectral region, which led to an enhancement of PCE by nearly 0.5 percentage points. The research presented in this chapter was supported by funding through the “Bavaria on the Move” initiative of the state of Bavaria and the project 1006-11 of the Bavaria Research Foundation (BFS). Electron microscopy resources was kindly provided by CENEM. Author gratefully acknowledge financial support from the Cluster of Excellence “Engineering of Advanced Materials” at the University Erlangen-Nürnberg and from the Federal Ministry for Economic Affairs and Energy (MNPQ program BMWI 11/12). I express my gratitude to Alfons Stiegelschmitt for XRD measurements.

 Parts of this chapter have been adapted or reproduced with permission from: I. Levchuk, C. Würth, F. Krause, A. Osvet, M. Batentschuk, U. Resch-Genger, C. Kolbeck, P. Herre, H. Steinrück and W. Peukert, Energy & Environmental Science, 2016, 9, 1083-1094.

- 48 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

3.1. Motivation and State of the art

During the past decades, doped semiconductor quantum dots (d-dots) have been increasingly studied since they retain nearly all intrinsic advantages of quantum dots (QDs) and benefit from larger Stokes shifts, minimum reabsorption, and longer excited state lifetimes related to dopant emission.1 Especially Mn2+-doped chalcogenide nanocrystals (NCs)2,3 together with Cu+-doped materials of the same composition4,5 have the potential to become a new class of emissive nanomaterials with high application potential as optical reporters in bioanalysis and active components in optical devices like solar cells. This, however, correlates with advances in synthetic routes that yield high quality d-dots with a photoluminescence quantum yield (PLQY) of at least 50 % with a high reproducibility and can be upscaled. One of the best studied d-dots are Mn2+-doped Zn and Cd chalcogenide NCs and their 1 4 6 2+ alloys. The luminescence of these NCs is governed by the T1− A1 emission of Mn in the orange wavelength region of 580-600 nm, which is nearly independent of the host material and crystal size. These d-d transitions are spin and parity forbidden, which leads to small absorption cross sections, hindering direct optical excitation, and a long intrinsic luminescence lifetime in the order of milliseconds. The absorption properties of these doped NCs are largely determined by the host matrix, which - together with an efficient energy transfer from the host to the Mn2+ ions - ensures bright photoluminescence. As photon absorption occurs mainly by the host lattice, the luminescence properties of doped QDs are similarly sensitive to surface defects like conventional semiconductor NCs, rendering the passivation of surface states, which favor non- radiative relaxation of the charge carriers formed upon light absorption, the main factor controlling PLQY of these d-dots. Surface state passivation can be achieved by a crystalline shell of a wide bandgap semiconductor like ZnS.6 Moreover, the shell material can enable band gap engineering to control the location of the charge carriers photogenerated in the core.7 Yang et al.8 recently presented a detailed study on factors governing the PL of Mn2+-doped CdS/ZnS core-shell NCs including the influence of the spatial localization of Mn2+ ions inside core-shell QDs as well as the role of surface trap states and energy transfer mechanisms. In this respect, they also reported a surface modification procedure to eliminate surface states, which yielded Mn2+-doped NCs with a PLQY in the order of 50 %. This is the first example of d-dots with a PLQY comparable to common undoped AIIBVI core–shell QDs.8,9,10 The synthesis of highly luminescent doped NCs requires reproducible and eventually up-scalable procedures which enable control of all parameters affecting crystal growth and incorporation of doping ions as well as proper surface passivation. Presently, high-quality

- 49 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

doped QDs are efficiently synthesized via high temperature synthetic routes to minimize the formation of surface defects.1 It is, however, challenging to incorporate dopant ions into a host QD lattice without simultaneously reducing the PLQY due to the size mismatch between host cations and dopant ions (e.g., the ion radius of Mn2+, 0.85 Å, differs from that of Zn2+, i.e., 0.75 Å, and Cd2+, i.e., 0.95 Å)11 and the ‘‘self-purification’’ effect.12 A key factor for obtaining high- quality doped QDs is the timing of dopant addition.13 There are three strategies to introduce dopant ions into host QDs: dopant introduction together with the host precursors,14-16 before nucleation of the host material,2,3 and during shelling of the host QD.17 Methods for the direct introduction of Mn2+ ions into a growing core or shell are very sensitive to different synthetic parameters and hence, frequently suffer from reproducibility problems. As a new approach to the synthesis of Mn2+-doped QDs, here exemplarily for ZnSe systems, Pradhan et al. utilized the initial formation of MnSe seeds followed by chemical substitution of Mn by Zn at a higher temperature in organic solvents.2 For Mn2+-doped ZnSe core NCs with a ZnS or ZnSe shell, with this method PLQY up to values 60 % could be realized,18,19 yet this required a complicated multistep hot injection phosphine strategy. In the case of injection-based methods, several kinetic and thermodynamic factors such as temperature instability, mixture heterogeneity, and overall high reaction rates and limited mass-transfer rates, however, often hinder the reproducible preparation of high quality nanocrystals with high PLQY on a large scale. Injection-based shelling methods involving the gradual injection of shell precursors at elongated temperatures, can allow the formation of gradient core–shell QDs with a smooth potential barrier for electrons and holes, without strains or interfacial defects, and, as a consequence, a PLQY approaching 100% as recently shown for undoped II-VI semiconductor QDs.20 This can be also utilized to introduce dopant ions during shelling of the host QD. Alternatives present non-injection syntheses.21 Such methods have been reported also for the preparation of Mn2+-doped CdS/ZnS8, CuInS/ZnS22 and MnS/ZnS/CdS23 core-shell QDs, yielding brightly emitting doped QDs with PLQY of 56 to 68 %. These non-injection methods require, however, several steps to obtain the final high quality products, which hampers upscaling. The application of very promising Mn-doped QDs as fluorescent reporters and active material in white light emitting diodes and spectral converters in photovoltaics (PV) 24,25,26 calls for simple and reproducible synthetic procedures using inexpensive precursors that yield high amounts of strongly luminescent NCs. This can be best met with a one-pot approach. The PLQY of Mn2+-doped QDs synthesized via one pot reactions are presently, however, still comparatively low, with maximum PLQY values of 30 %.27,24 This is ascribed to the general

- 50 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

challenge to incorporate Mn2+ into the host lattice and insufficient passivation of surface defects.11 This Chapter describe an industrially scalable and cost-effective one-pot, two-step non- injection synthesis for highly luminescent ZnxCd1-xS:Mn/ZnS NCs. A consequent optimization of the synthetic route for the NC core synthesis as well as for the surface passivation with a ZnS shell allowed to achieve the highest PLQY of 70% reported to date for this material. This NCs were subsequently used for the fabrication of codt-effectivedown-shifting layers and coupled to monocrystalline silicon (mono-Si) solar cells, resulting in an enhanced external quantum efficiency (EQE) in the UV- blue spectral range and an overall higher AM1.5 power conversion efficiency (PCE).

/ 3.2. One-pot synthesis of Mn-doped ZnxCd1-xS:Mn ZnS core/shell NCs

/ The phosphine-free synthesis of ZnxCd1-xS:Mn ZnS core/shell NCs was performed via a non-injection route based on a doping technique utilizing nucleation, where the starting materials for the core of the NC and the dopant precursor are loaded together into the reaction vessel at room temperature. The main steps of this synthesis, which is based on a modified 4 method by Zhang et al. for ZnxCd1-xS:Cu, are shown in Figure 3.1 and subsequently described; The PL mechanism of Mn-doped NCs is completely different from the one of Cu-doped NCs. In the case of Cu doping, electrons from the conduction band of the host material recombine with holes in the Cu-mediated state. Different to this, Mn doping leads to a recombination process within d-d levels of Mn2+. Therefore, a deep and systematic revision of the previously 4 reported synthetic approach for Cu-doped ZnxCd1-xS NCs was required. We fabricated highly 2+ regular Mn doped ZnxCd1-xS NCs which were further effectively passivated with a ZnS shell and report here for the first time the synthesis methods and the extended characterization.

Figure 3.1. Schematic illustration of the one pot, two step synthesis of ZnxCd1-xS:Mn/ZnS core-shell NCs.

- 51 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

A mixture of the solvents dodecanethiol (DDT) and octadecene (ODE) was degassed, stock solutions of the Zn, Cd, and Mn-precursors as well as an ODE-sulfur solution were added, and the reaction mixture was heated to 230 °C under nitrogen flow and continuous stirring. After completion of the particle growth, the reaction mixture was cooled to 60 °C. The resulting doped NCs were precipitated by addition of ethanol and purified by a precipitation/re- dispersion method. For control of the synthesis and optimization of the reaction conditions, aliquots of the reaction mixture were taken at different temperatures and times and spectroscopically studied. The growth experiments were made mainly with a Zn/Cd/Mn precursor ratio of 50/50/5, which was the first parameter optimized with this procedure.

The results of the structural characterization of the ZnxCd1‑xS:Mn NCs with transmission electron microscopy (TEM), X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR) are summarized in Figures 3.2a-e. All NCs reveal a near- spherical shape and possess high size uniformity (Figures 3.2a and 3.2b). The size of the nanoparticles was estimated from transmission electron microscopy (TEM) to be 2.88±0.4 nm with a size distribution of only 14% and 3.23±0.42 nm with size distribution of 13% for

Zn0.5Cd0.5S:Mn and Zn0.5Cd0.5S:Mn/ZnS, respectively. Interplanar spacing in the single nanoparticles was obtained by high resolution transmission electron microscopy (HRTEM) analysis. In both, core and core/shell NCs the interplanar spacing is 0.32 nm, which is in between the spacing for CdS (0.33544 nm) and ZnS (0.31261) for the 111 plane of the cubic structure. Wide-angle XRD patterns summarized in Figure 2c reveal the phase and elemental composition of representatives Zn0.5Cd0.5S:Mn core NCs (lower spectrum) and

Zn0.5Cd0.5S:Mn/ZnS core/shell NCs (upper spectrum) samples, which display the characteristic peaks of the zinc blende (cubic) structure, located between those of cubic ZnS and CdS materials, respectively. This confirms the Zn0.5Cd0.5S alloy structure. Coating the ZnS shell over Zn0.5Cd0.5S core also slightly changed the lattice constant from 5.57 to 5.50 Å. Due to the similar lattice parameter of the ZnS and Zn0.5Cd0.5S and a likely atomic diffusion at the core/shell interface, it was not possible to differentiate the core from the shell by HRTEM. This phenomena is well known for different kinds of core/shell structures, such as Ag/Au57, 58 5 NaYbF4:Er/NaYF4 or CdSe-Core CdS/Zn0.5Cd0.5S/ZnS multishell nanocrystals. The crystallite size for ZnxCd1‑xS:Mn and ZnxCd1‑xS:Mn/ZnS thin powder layers was determined from the XRD pattern to be 3.05 and to 3.5 nm by a Gaussian fit and the Debye-Scherrer formula, respectively. This is in excellent agreement with the results from HRTEM.

- 52 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Figure 3.2. TEM and HRTM images of (a) Zn0.5Cd0.5S:Mn and (b) ZnS-coated Zn0.5Cd0.5S:MnNCs; the corresponding size distributions obtained by TEM are given as insets. (c) XRD patterns of the two NC samples. The standard data for ZnS zinc blende bulk material (top, ICSD 00-005-0566) and zinc blende bulk CdS (bottom, ICSD pattern 01-080-0019) are shown as reference. (d) FTIR spectra of Zn0.5Cd0.5S:Mn(5%)/ZnS NCs, and (e) an enlarged view of the same spectra in the wavelength region of 2380-2880 cm-1 for the NCs and DDT. 3.3. Analysis of NCs Surface via FTIR spectroscopy.

FTIR spectra of purified Zn0.5Cd0.5S:Mn(5%)/ZnS NCs shown in Figure 3.2d and 3.2e provide information on the surface chemistry of the doped QDs. The bands in the range of 3000 to 2800 cm-1 are attributed to C-H stretching vibrations of dodecanethiol (DDT) and oleic acid -1 (OA) and to the asymmetric vibration of CH3 (2955 cm ) as well as the asymmetric -1 -1 (2922 cm ) and symmetric stretching of CH2 (2853 cm ), respectively. The peak at 1708 cm-1 is ascribed to the C=O stretch vibrations of OA. The band at 1560 cm-1 corresponds to carboxylate groups of the OA ligands and provides evidence of a coordinative bond between this group and Zn2+ or Cd2+ cations on the NC surface.28 Comparison of the IR spectra of pure DDT and the NCs samples (Figure 2, panel e) shows that the weak S-H vibrational band at 2576 cm-1 found in DDT is absent in the NCs. This confirms the broken S-H bonds due to chemisorption of thiolates onto the surface of the NCs.29,30 Hence, DDT serves not only as

capping ligand, but also as sulfur source for shell growth. Moreover, the CH2 scissoring mode

- 53 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

-1 -1 (1464 cm ) and the CH3 symmetric bending vibration (1377 cm ) of DDT and the in-plane OH bending vibrations of OA30 are observed.

3.4. Optimization of the NCs growth synthesis.

To optimize the PL properties of our NCs, we assessed different reaction parameters including temperature, ratio of the metal ion precusors, dopant ion concentration, and shell thickness. Crystal formation and NC growth due to the decomposition of the metalloorganic complexes starts at about 125 °C as indicated by a change in color of the solution from colorless to yellow and then to light orange.31 This was confirmed by absorption and emission measurements summarized in Figure 3.3a-h. The broad shoulder in the absorption spectrum at ca. 400 nm (Figure 3b, left axis), which undergoes a small red shift from 400 to 404 nm with

increasing reaction time, is assigned to the first excitonic transition of the ZnxCd1‑xS NCs. Such a poorly resolved excitonic absorption peak is characteristic for ternary and quaternary alloyed NCs. Its width points to a wide size distribution and the irregular composition distribution.4 The more or less constant spectral shape during a growth period of 3 h indicates no significant changes in alloy composition. The NCs extracted at 150 °C showed already a bright emission (Figure 3a). The highest emission intensity was revealed by samples taken at 230 °C (time “0” at the y-axis in Figure 3.3a). Longer reaction times and hence, further NC growth resulted in a rapid decrease in emission intensity. Figure 3.3b (Figure 3.3b, right axis) summarizes the corresponding emission spectra, which are not normalized to better visualize changes in spectral shape and PL intensity with temperature and reaction time. These spectra 4 6 2+ contain the characteristic T1− A1 emission of Mn (see inset in Figure 3.3h) with its maximum at 598 nm, matching the values reported for ZnCdS:Mn materials with zinc-blende23 as well as wurtzite crystal structure,11 and a broad red PL band located at about 710 nm. The latter is attributed to emission from defect states or Mn2+ ions at or near the surface in ZnCdS:Mn.32,33 Due to the “self-purification” effect,12 a relatively large number of Mn2+ ions can be located in the vicinity of the surface in these NCs after long annealing times at 230 °C, which may account also for the reduction of the Mn2+ PL intensity in Figure 3a. The intensity of the long wavelength band at 150 °C is rather high for doped NCs as a homogeneous Mn2+ distribution can be expected under these reaction conditions. Thus, an additional contribution from the emission of Cd2+ species due to local inhomogeneity cannot be excluded. For example, an extrinsic deep-level emission at about 710 nm has been also reported for CdS,23,33 which disappeared after coating of the NCs with a ZnS shell.

- 54 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Figure 3.3. (a) Dependence of the intensity of the Mn2+ emission at 598 nm on reaction time and temperature. The PL intensities are normalized to the maximal Mn2+ PL intensity of the NCs taken from the reaction mixture directly after reaching 230 °C. (b) Thermal and temporal evolution of absorption (left axis) and emission spectra (right axis) of (unshelled) Zn0.5Cd0.5S 2+ NCs doped with 5 mol % of Mn using excitation wavelengths (λex) of 375 and 390 nm for the sample taken at 150°C and 175 °C, respectively, and 400 nm for all other samples to always excite each sample at the first excitonic absorption maximum. (c) PL intensity at different Cd/Zn precursor ratios and (d) the corresponding absorption (left axis) and normalized emission (right axis, excitation at the first excitonic absorption maximum of each sample) spectra of Zn1-xCdxS:Mn/ZnS NCs. For all core compositions in panels c, d, e, and f, the NCs were coated with one layer of ZnS. (e) PL intensity of undoped Zn0.5Cd0.5S/ZnS and 2+ Zn0.5Cd0.5S/ZnS doped with different Mn concentrations (λex = 400 nm); (f) absorption (left 2+ axis) and normalized emission spectra (right axis) of undoped Zn0.5Cd0.5S/ZnS and Mn -doped NCs. (g) Dependence of the PL intensity on the number of ZnS coating layers, and (h) evolution of the corresponding absorption (left axis) and normalized emission spectra (right axis, λex = 400 nm) of the Zn0.5Cd0.5S:Mn core (black curve) and Zn0.5Cd0.5S:Mn/ZnS core-shell particles, doped with 5 mol % Mn2+. All absorption spectra were recorded with samples having the same absorbance at the first excitonic absorption band/transition. The inset in panel (h) illustrates the light absorption associated with the valence band-conduction band transition, energy transfer from the accordingly formed exciton to the Mn2+ ions, and the nonradiative relaxation initiated by surface defects (SD). All measurements were performed with NCs dispersed in chloroform.

- 55 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Impact of Cd and Zn ration on optical properties. To study the effect of the Cd/Zn molar ratio on the incorporation of Mn2+ into the crystal lattice and the optical properties of the 2+ resulting ZnxCd1-xS:Mn /ZnS NCs, we prepared samples with Cd/Zn ratios of 0.5, 1, and 2, respectively and kept the concentration of the Mn2+ precursor in the reaction mixture constant at 5 mol %. Subsequently, the NCs were coated with a single layer of ZnS by addition of the Zn precursor to the reaction mixture at 230°C. As follows e.g., from Figure 3.3d, the absorption spectra measured at Cd/Zn ratios of 0.5, 1, and 2, respectively, reflect the change of the NC composition from Zn-rich to the Cd-rich samples by a red shift of the first excitonic maximum from ca. 395 to ca. 420 nm. The most intriguing change in emission is the suppression of the red PL at ca. 710 nm by the ZnS shell (Figure 3.3d, right axis). This underlines its origin from Mn2+ ions at or near the surface. Although the spectral position of the emission maximum of the Mn2+ (d5) ion depends on crystal field,34 the spectral shape and position of the emission band at 598 nm (Figure 3.3d, right axis), which arises from the 4 6 2+ T1– A1 transition of Mn ions, is only slightly affected by the Cd/Zn ratio, with Zn-rich NCs (Cd/Zn ratio of 0.5) revealing a small red shift compared to Cd-rich NCs. This is in agreement with the data published by Chen et al.27, Xua at al.,35 and Kim et al.24 The intensity of the Mn2+ emission is maximum for a Cd/Zn ratio of one (Figure 3.3c). This is consistent with the results reported by Nag et al.11, who studied the influence of the Mn2+ concentration as a function of

the lattice mismatch between MnS and ZnxCd1-xS NCs and found an efficient incorporation of 2+ Mn ions into the wurtzite host lattice of ZnxCd1-xS for a Cd/Zn ratio of one. This seems to be 36 also true for ZnxCd1-xS NCs with a zinc blende structure as predicted by Klimov et al., because the enthalpy of Mn2+ ion doping is minimized at this NC composition. Study of Mn-doping. The dependence of the PL intensity of the Mn-doped 2+ ZnxCd1-xS/ZnS(x = 0.5) NCs on Mn concentration of is summarized in Figure 3.3e for concentrations of the Mn-precursor of 0.5 % to 10 % and the corresponding absorption and normalized emission spectra are shown in Figure 3.3f. The presence and the oxidation state of Mn2+ was confirmed by EDX and XPS measurements (Figure 3.4a-c). Doping of 2+ ZnxCd1-xS/ZnS (x = 0.5) NCs with increasing concentrations of Mn barely affects the absorption spectra of these materials, which are dominated by optical transitions of the host matrix (first excitonic transition and band-band transitions) in contrast to the emission spectra and especially the PL intensity. As shown in Figure 3f, incorporation of Mn2+ shifts the initially

broad emission of the undoped ZnxCd1-xS QDs, originating from a broad distribution of shallow trapped states,37,33 to 598 nm, i.e., to the typical Mn2+ emission band, and leads to a narrowing

- 56 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

of the PL band with increasing Mn2+ concentration. The highest PL intensity is reached for samples grown from stock solutions containing 5 mol. % of the Mn precursor (Figure 24e).

Figure 3.4. EDX analysis of the elemental composition of the Zn0.5Cd0.5S:5% Mn (a) and Zn0.5Cd0.5S:5% Mn/ZnS (b) d-dots. Survey XPS spectrum of Zn0.5Cd0.5S:Mn NCs with 10% Mn (c) and Zn0.5Cd0.5S:Mn/ZnS (c). In the inset, the detail spectra of the Mn 2p region is shown; the red line is added as guide to the eye. Higher dopant ion concentrations lead to a decrease in PL intensity. This reflects two processes, the incorporation of the Mn2+ ions into the crystal lattice of the NCs, acting as luminescence centers, and a concentration quenching at higher Mn2+ doping concentrations due to energy transfer and energy migration between neighboring Mn2+ ions, with Mn2+ ions located at or near the surface acting as energy sinks. Concentration quenching in d-dots has been reported also by other groups.11, 38 For example, Nag et al.11 observed maximum PL 2+ 38 intensities for unshelled Zn0.49Cd0.51S:Mn doped with about 1 % Mn and Kole et al. realized maximum PL intensities of unshelled ZnS d-dots with a Mn2+ doping concentration of 1.1 %. We ascribe our observation of Mn2+ concentration quenching at considerably higher dopant ion concentrations to the protecting ZnS shell of our NCs, which shields surface or near surface Mn2+ ions. Also, Mn2+ ions may partially diffuse into the shell during its formation, which affects Mn2+-Mn2+ distances and hence, energy transfer and energy migration between neighboring Mn2+ ions.

- 57 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Surface passivation. A common approach to increase PLQY of binary or ternary semiconductor 39 40 41,42,43 27 NCs such as CdSe, CdTe, CuInS2 and d-dots like ZnCdS:Mn, 7 44 Zn1-x-yCdxMnyS and ZnSe:Mn presents the shelling of their surface with a higher band gap semiconductor (e.g., ZnS, CdS) or with more sophisticated multi-shells like ZnS/CdS23 or 5 CdS/Zn0.5Cd0.5S/ZnS to passivate dangling bonds and surface states favoring radiationless decay or emission with low intensity from trapped states. The beneficial influence of a surface shell follows also from Figures 3.3c and 3.3d depicting the influence of a single shell on the 45,46 PL properties of our ZnxCd1-xS:Mn NCs. For maximum PLQY,, the shell thickness needs to be optimized material-specifically. Hence, we studied the influence of the ZnS shell

thickness on the PL intensity of already optimized d-dots, here Zn0.5Cd0.5S:Mn cores (Zn/Cd ratio of 1:1), doped with 5 mol % Mn. For this purpose, shelling with ZnS was performed in a simple one pot reaction at 230 °C by stepwise injection of identical amounts of a stock solution of the Zn precursor in time intervals of 15 min. At this temperature the alloying effect is absent. Zhong and co-workers59 defined 270 oC as the “alloying point” analogous to melting or boiling points, for the CdSe/ ZnSe and CdSe/ZnS systems. Below this temperature the alloying process between core and shell is energetically disadvantageous. The ZnS shell grown in these 15 min is considered as one layer. The corresponding absorption and normalized emission spectra and the resulting PL intensities are summarized in Figures 3g and 3h, respectively. The growth and increasing thickness of the ZnS shell is confirmed by the increasing absorption in the UV region of 320 to 370 nm (Figure 3.3h) and the disappearance of the red emission at 710 nm (Figures 3.3b and 3.3h). Additional proof for the ZnS shell is the increased Zn content in the core/shell NCs when compared to the uncoated core. Both, EDX and XPS spectra confirm an enhancement of the Zn/Cd ratio for the core/shell particles (Figure 3.4a-c), and the prolongation of the photoluminescence lifetime from 0.89 to 2.85 ms, does not allow any other conclusion than the formation of a ZnS shell acting as effective surface passivation layer (Figure 3.5a). The spectral position of the Mn2+ emission is independent of the number of ZnS layers, whereas its intensity strongly increases with shell thickness. A single ZnS layer enhances the PL intensity of the Zn0,5Cd0,5S core, showing a PLQY below 4%, already by a factor of 14 (Figure 3g). The photoluminescence decay curves of Zn0.5Cd0.5S:Mn and

Zn0.5Cd0.5S:Mn/ZnS nanocrystals are shown in Figure 3.5a. The intensity average lifetime τ, given by:

∫ 푡퐼(푡)푑푡 휏 = , (3.1) ∫ 퐼(푡)푑푡

- 58 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

0.89 ms is for Zn0.5Cd0.5S:Mn particles and 2.85 ms for the Zn0.5Cd0.5S:Mn/ZnS particles, giving evidence of the effective passivation of surface defects for the ZnS-coated sample. To the best of our knowledge, such a long excited-state carrier lifetime is highest reported for the ZnS and Cd based Mn2+ doped NCs (Table 3.1). The highest PLQY achieved for our core-shell NCs with this strategy amounts to 70 %. This PLQY is the highest values reported for Mn2+-doped sulfide-semiconductor NCs, 6,9,17, 21, 24,27,35,36 (Table 3.1) that were, however, prepared using much more sophisticated synthetic schemes as our simple one pot reaction.

XPS study. NCs Zn0.5Cd0.5S with nominal 10% mol. Mn and without the ZnS shell (Figure 3.4c) are measured by X-ray photoelectron spectroscopy (XPS) using Mg Kα-radiation (1253.6 eV); note that Al Kα-radiation cannot be used due to an overlap of the Mn 2p levels

with the intense Zn LMM Auger lines. In case of Zn0.5Cd0.5S:Mn/ZnS (Figure 3.4d), Mn 2p signal is strongly qunched by ZnS shell, which prevents precise definition of the Mn state. Figure 3.4c depicts the survey spectrum showing the contributions of all expected elements. The Mn 2p region falls within the spin-orbit split Cd 3p1/2 and 3p3/2 levels at 652.4

and 618.4 eV, respectively. While the Mn 2p1/2 level is hidden under the Cd 3p1/2 peak, the Mn

2p3/2 peak is clearly separated in energy and is shown as inset in the figure. The peak has its maximum at 641.8 eV, with a broad shoulder at higher binding energy. This peak shape is due to a complex multiplet structure73, which cannot be resolved due to the low signal intensity.

Within the error bars, the binding energy and peak shape of the Mn 2p3/2 core level is consistent with an oxidation state of either II, III or IV73. From the comparison to the spectra of Biesinger et al.73, who measured several manganese oxides with different oxidation levels, the large width of the multiplet structure is a hint towards Mn(II) species. Furthermore, we compared ratio between Zn3p/Cd4p and Zn3s/Cd4s signal intensities for Zn0.5Cd0.5S:Mn and

Zn0.5Cd0.5S:Mn/ZnS. In both cases the ratio increases, confirming the ZnS shell growth.

Scaling-up synthesis. Use of doped NCs e.g., for the fabrication of more efficient solar cells or other opto-electronic devices requires synthetic procedures that efficiently and reproducibly yield a large amount of high quality NC material. We subsequently studied the

upscaling of our simple one pot, two step synthesis of Zn0.5Cd0.5S:Mn/ZnS with optimum luminescence properties. Hence, the amount of all components was increased by a factor of 40 compared to the initially used synthesis. This allowed for the preparation of 3.5 g of

Zn0.5Cd0.5S:Mn/ZnS NCs in identical quality (Figure 3.5b).

- 59 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Figure 3.5. a) PL decay curves of the Zn0.5Cd0.5S:Mn and Zn0.5Cd0.5S:Mn/ZnS NCs with six Zn precursor injection. Chloroform solutions of the NCs exited with Xe-lamp at 320 nm, and detection wavelength at 600 nm for both samples. b) Upscaling of the synthesis of Zn0.5Cd0.5S:Mn(5%)/ZnS core shell NCs.

2+ 3.5. Applying Mn -doped ZnxCd1‑xS/ZnS NCs as down-shifting layer in Si solar cells.

2+ To demonstrate the application potential of our optimized Mn -doped ZnxCd1‑xS/ZnS NCs, we employed this material as UV-to-visible light converter in monocrystalline silicon (mono-Si) solar cells to improve the otherwise low sensitivity of mono-Si solar cells in the UV and blue spectral region by shifting the energy of these otherwise unused photons to a spectral region of higher sensitivity of the mono- Si cell47. Semiconductor NCs like CdSe/CdS, ZnSe,

CdSe/ZnSe, Zn0.5Cd0.5S/ZnS have been already used as effective UV-light converter materials for downshifting layers48,49,50,51, yet not ternary Mn-doped NCs. The scheme of the down- shifting mechanism and a schematic representation of our proof-of-concept converter film - mono-Si-solar cell system are shown in Figure 3.6a.

For this purpose, we dispersed Zn0.5Cd0.5S:Mn(5%)/ZnS NCs with PLQY of 70% in toluene in different concentrations of 0.5, 1, 2.5, 5, 10, 15, 20, 25 mg/ml and coated a thin film on a commercial mono- Si solar cell with a SiNx coated texture by the doctor blade technique. Subsequently, the resulting three solar cells were dried on a hot plate at 80 °C for five minutes. Then, the external quantum efficiency (EQE), i.e. the number of charge carriers to the number of incident photons ratio, and the current density (J)–voltage (V) curves (J-V diagrams) of the three solar cells depicted in Figure 5a (right panel) were measured. The open-circuit voltage

(UOC) and fill factor (FF; FF = max. power/JSC UOC) are taken from the J-V measurements under the solar simulator, the short circuit current (JSC) was calculated from the EQE spectra, and after that the power conversion efficiency (PCE) was calculated.

- 60 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Table 3.1. Previously reported Mn2+ doped semiconductor nanocrystals and their properties

QDs PL Peak/(nm) Lifetime/(ms) Method of Ref. QY/(%) synthesis

Mn:Zn-In-S 56 600 4.2 Multi-step 60 hot-injection method Mn:ZnS > 50 585 0.37 One-pot 61 two-steps

Mn:CdS not 580-620, not mentioned Low- 62 mentioned tunable temperature injection Mn:ZnSe 35 590 0.38 Hot-injection 63

Mn:ZnSSe ~60 ~595 not mentioned Hot-injection 64

MnS/ZnS/CdS 68 580 0.68 Multi-step 65 hot-injection method Mn:ZnCdS 30 580 not mentioned Reverse 66 micelle Mn:CuInZnS 45 600 2.12 One-pot 67 two-steps

Mn:CdInS 17 630 1.1 Hydrothernal 68

Mn:CdZnSe ~25 580 ~0.6 Multi-step 69 hot-injection method Mn:CuInS/ZnS 66 610 3.78 One-pot 70 two-steps Mn:ZnS > 50 590 1.71 Multi-step 71 hot-injection method ZnCdS:Mn/ZnS 70 598 2.85 One-pot This work two-steps

- 61 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

This is necessary, because due to the deviation of the spectrum of the solar simulator from the standard AM1.5G spectrum (much weaker emission at wavelengths below 400 nm), it is not possible to use the solar simulator for the characterization of the downshifting layers covered solar cells

Figure 3.6. Performance of mono-Si solar cells coated with Zn0.5Cd0.5S:Mn(5%)/ZnS NCs, using a volume 200 μl of the differently concentrated NCs dispersion. (a) Scheme of the down- shifting mechanism of the Zn0.5Cd0.5S:Mn(5%)/ZnS NC converter material and design of proof-of-concept solar cells. (b) EQE of the best mono- Si solar cell (black curve) obtained with a Zn0.5Cd0.5S:Mn(5%)/ZnS NC dispersion containing 10 mg/ml NCs (red curve); the EQE in the wavelength region of 300 to 400 nm is magnified in the inset. (c) J–V curves of the corresponding solar cells. . (d) SEM picture of the mono-Si solar cell surface structure coated with 5 mg/ml NCs solution. (e) cross-section of the same mono-Si solar cell with 260 nm NCs thickness. Measurement of the EQE curves of the mono-Si solar cells coated with different concentrations of Zn0.5Cd0.5S:Mn(5%)/ZnS NCs (Figures 3.6b, best sample) reveal an increased EQE in the range of 0.5-15 mg/ml and a significantly decreased EQE for higher concentrations of 20-25 mg/ml due to parasitic absorption by the thicker NC layer. The optimized performance with an EQE enhancement of 12 % is found for the layers coated form a 10 mg/ml concentrated solution (Figure 3.6b).

- 62 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Figure 3.7. Selected SEM picture of the mono-Si solar cell surface structure cross-section of the same mono-Si solar coated with 0.5 mg/ml (a), 2.5 mg/ml (b), 5 mg/ml (c), 10 mg/ml (d) and 25 mg/ml (e) NCs solution. The obtained NCs thicknesses varies from <50 to 750 nm.

- 63 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

The average thickness of the NCs layer was determined by SEM cross section measurements and yielded values between <50 nm to 750 nm (Figure 3.7a-e) for the various concentrations. The best performing coating from the 10 mg/ml concentrated solution showed a layer thickness about 300 nm (Figure 3.7c) at an EQE enhancement of app. 12 % as shown in Figure 3.8. We also observed an enhancement of the EQE in the visible and in the NIR region, here

mostly due to the NCs acting as antireflection coating on SiNx, as the refractive indexes of 52 ZnCdS and SiNx equal 1.95 and 2.0, respectively.

15 1 mg/ml 15 mg/ml 10 0.5 mg/ml 20mg/ml 2.5 mg/ml 25 mg/ml 5 mg/ml NCs Absorbance 5 10 mg/ml

-5

-10

-15 Enhancement % of ratio EQE, 300 400 500 600 700 800 900 1000

Wavelength, nm Figure 3.8. Enhancement ratio of EQE of mono-Si solar cells as function of wavelength for

the coating with differently concentrated Zn0.5Cd0.5S:Mn(5%)/ZnS NCs dispersions.

The electrical characteristics of the solar cells coated with Zn0.5Cd0.5S:Mn(5%)/ZnS NCs are summarized in Table 3.2. Lower concentrated solution with 0.5-15 mg/ml (Figures 3.6c, 5 mg/ml sample) led to an increase in power conversion efficiency (PCE) and

in short circuit current (JSC), demonstrating the down-shifting effect of our NCs. For the best sample of the mono-Si solar cell coated with 10 mg/ml, PCE led to an enhancement by nearly 0.5 percentage points, which is close to theoretical limit (0.6%) of the down-shifting layer for Si solar cell 47 .

- 64 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

Table 3.2. Photovoltaic properties of mono-Si solar cells coated with differently concentrated dispersions of Zn0.5Cd0.5S:Mn(5%)/ZnS NCs. Open-circuit voltage (UOC) and fill factor (FF; FF = max. power/JSC UOC) are taken from the J-V measurements under the solar simulator, the short circuit current (JSC) and the power conversion efficiency (PCE) are calculated from the EQE spectra.

Sample Uoc (V) FF, % Jsc (mA/cm²) PCE, % Δ PCE, % 1 Solar cell 0,522±0,0114 64,69±0,2504 35,979±0,0184 12.149±0,0448 + 0,132 Solar cell + 0.5mg/ml QDs 0,528±0,0116 64,70±0,8163 35,950±0,0184 12.281±0,0175

2 Solar cell 0,536±0,0117 63,63±0,8346 35,209±0,0180 12.008±0,1669 + 0,302 Solar cell + 0,537±0,0118 64,52±0,6900 35,531±0,0182 12.310±0,1083 1 mg/ml QDs

3 Solar cell 0,538±0,0118 64.59±0,9580 38,061±0,0195 13.225±0,1910 + 0,434 Solar cell + 2.5mg/ml QDs 0,545±0,0119 65.37±0,5120 38,341±0,0196 13.659±0,2490

4 Solar cell 0,538±0,0118 65.14±1,151 38,414±0,0197 13.463±0,224 + 0,458 Solar cell + 0,552±0,0121 65.14±0,605 38,716±0,0198 13.921±0,149 5 mg/ml QDs 5 Solar cell 0,546±0,0120 65.60±0,622 38,499±0,0197 13.809±0,109 + 0,462 Solar cell + 0,558±0,0122 65.81±0,754 38,861±0,0199 14.271±0,268 10 mg/ml QDs 6 Solar cell 0,546±0,0120 65,47±0,7355 36,033±0,0184 12.880±0,3141 + 0.02 Solar cell + 0,549±0,0120 65,61±0,7580 36,013±0,0184 12.900±0,1433 15 mg/ml QDs 7 Solar cell 0,534±0,0117 65,86±0,1542 35,866±0,0183 12.613±0,2162 - 0,025 Solar cell + 0,534±0,0117 35,795±0,0183 12.588±0,1913 20 mg/ml QDs 66,01±0,5872 8 Solar cell 0,538±0,0118 65,66±0,1443 35,857±0,0183 12.667±0,059 - 1,025 Solar cell + 0,540±0,0118 65,65±0,1443 32,570±0,0167 11.546±0,0427 25 mg/ml QDs

Although due to many research activities dedicated to new solar cells with improved efficiencies like perovskite, solar cells with an efficiency of more than 20%,56 the broad use of photovoltaics requires strategies for the simple and inexpensive upgrading of the currently used Si photovoltaic modules to increase their power conversion efficiency without costly replacement. In this respect, and to underline the application potential of these novel down- shifting layers, we performed a simple BoM (bill of materials) and BoS (bill of system) cost calculation for the fabrication of photovoltaic (PV) modules based on our down-shifting design. Based upon this estimation, we conclude that application of a thin layer of highly luminescent

Zn0.5Cd0.5S:Mn/ZnS NCs (use of 0.2 ml of a dispersion of 2.5 mg/ml NC distributed on an area of 8 cm2), can decrease the production costs of a one Watt energy for mono-Si solar module with power 117W by about 1.5 $ (3.3 %) due to the enhanced efficiency in parallel to the unprecedented low costs of this novel converter material.

- 65 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

3.6. Summary In summary, we developed a simple one-pot two-step synthesis for high quality

ZnxCd1-xS:Mn/ZnS NCs with a very high reproducibility. Very high photoluminescence quantum yields (PLQY) of up to 70 % and the ease of upscaling this industrially compatible fabrication process make this doped semiconductor material a very promising candidate especially for photovoltaic and optoelectronic applications including light-emitting devices. The potential of these NCs follows from the exemplarily shown use of the

Zn0.5Cd0.5S:Mn(5%)/ZnS NCs as down-shifting layer for the ultraviolet (UV) to orange wavelength region for monocrystalline silicon (mono-Si) solar cells. With our NCs and a simple coating technique, the efficiency of a commercial Si solar cell could be improved up to ca. 12 % in the UV spectral region which led to an enhancement of PCE by nearly 0.5 percentage points. This allows a reduction of the production costs of mono-Si solar cell modules by around 3.33 %. Future work on these doped ternary alloys will focus on the further improvement of their PLQY and their use as down-shifting layer for different types of solar cells such as thin film, e.g. CuInS2 or organic solar cells with a low sensitivity in the blue and UV spectral region, thereby paving the road for the next generation of more efficient and less costly photovoltaic modules for the broad use of solar energy.

- 66 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

3.7. Bibliography

1. P. Wu and X.-P. Yan, Chemical Society Reviews, 2013, 42, 5489-5521. 2. N. Pradhan, D. Goorskey, J. Thessing and X. Peng, Journal of the American Chemical Society, 2005, 127, 17586-17587. 3. N. Pradhan and X. Peng, Journal of the American Chemical Society, 2007, 129, 3339- 3347. 4. W. Zhang, X. Zhou and X. Zhong, Inorganic Chemistry, 2012, 51, 3579-3587. 5. R. Xie, U. Kolb, J. Li, T. Basché and A. Mews, Journal of the American Chemical Society, 2005, 127, 7480-7488. 6. R. W. Knoss, Nova Science, New York, 2009, 691. 7. L. R. Bradshaw, K. E. Knowles, S. McDowall and D. R. Gamelin, Nano Letters, 2015, 15, 1315-1323. 8. Y. Yang, O. Chen, A. Angerhofer and Y. C. Cao, Journal of the American Chemical Society, 2006, 128, 12428-12429. 9. U. Resch-Genger, M. Grabolle, S. Cavaliere-Jaricot, R. Nitschke and T. Nann, Nat Meth, 2008, 5, 763-775. 10. M. Grabolle, M. Spieles, V. Lesnyak, N. Gaponik, A. Eychmüller and U. Resch-Genger, Analytical Chemistry, 2009, 81, 6285-6294. 11. A. Nag, S. Chakraborty and D. D. Sarma, Journal of the American Chemical Society, 2008, 130, 10605-10611. 12. G. M. Dalpian and J. R. Chelikowsky, Physical Review Letters, 2006, 96, 226802. 13. P. I. Archer, S. A. Santangelo and D. R. Gamelin, Journal of the American Chemical Society, 2007, 129, 9808-9818. 14. F. V. Mikulec, M. Kuno, M. Bennati, D. A. Hall, R. G. Griffin and M. G. Bawendi, Journal of the American Chemical Society, 2000, 122, 2532-2540. 15. J. F. Suyver, S. F. Wuister, J. J. Kelly and A. Meijerink, Physical Chemistry Chemical Physics, 2000, 2, 5445-5448. 16. D. J. Norris, N. Yao, F. T. Charnock and T. A. Kennedy, Nano Letters, 2001, 1, 3-7. 17. P. V. Radovanovic and D. R. Gamelin, Journal of the American Chemical Society, 2001, 123, 12207-12214. 18. R. Zeng, T. Zhang, G. Dai and B. Zou, The Journal of Physical Chemistry C, 2011, 115, 3005-3010. 19. R. Zeng, M. Rutherford, R. Xie, B. Zou and X. Peng, Chemistry of Materials, 2010, 22, 2107-2113. 20. K.-H. Lee, J.-H. Lee, W.-S. Song, H. Ko, C. Lee, J.-H. Lee and H. Yang, ACS Nano, 2013, 7, 7295-7302. 21. P. Samokhvalov, M. Artemyev and I. Nabiev, Chemistry – A European Journal, 2013, 19, 1534-1546. 22. S. Cao, C. Li, L. Wang, M. Shang, G. Wei, J. Zheng and W. Yang, Sci. Rep., 2014, 4. 23. S. Cao, J. Zheng, J. Zhao, L. Wang, F. Gao, G. Wei, R. Zeng, L. Tian and W. Yang, Journal of Materials Chemistry C, 2013, 1, 2540-2547. 24. K. Jong-Uk, L. Myung-Hyun and Y. Heesun, Nanotechnology, 2008, 19, 465605. 25. C. S. Erickson, L. R. Bradshaw, S. McDowall, J. D. Gilbertson, D. R. Gamelin and D. L. Patrick, ACS Nano, 2014, 8, 3461-3467. 26. F. Meinardi, A. Colombo, K. A. Velizhanin, R. Simonutti, M. Lorenzon, L. Beverina, R. Viswanatha, V. I. Klimov and S. Brovelli, Nat Photon, 2014, 8, 392-399. 27. Z.-Q. Chen, C. Lian, D. Zhou, Y. Xiang, M. Wang, M. Ke, L.-B. Liang and X.-F. Yu, Chemical Physics Letters, 2010, 488, 73-76. 28. M. Luo, G. K. Olivier and J. Frechette, Soft Matter, 2012, 8, 11923-11932.

- 67 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

29. D. Xu, Z. Liu, J. Liang and Y. Qian, The Journal of Physical Chemistry B, 2005, 109, 14344-14349. 30. J. Yang, T. Ling, W.-T. Wu, H. Liu, M.-R. Gao, C. Ling, L. Li and X.-W. Du, Nat Commun, 2013, 4, 1695. 31. Y. Vahidshad, M. Nawaz Tahir, A. Iraji Zad, S. M. Mirkazemi, R. Ghasemzadeh, H. Huesmann and W. Tremel, The Journal of Physical Chemistry C, 2014, 118, 24670- 24679. 32. Hazarika, A. Layek, S. De, A. Nag, S. Debnath, P. Mahadevan, A. Chowdhury and D. D. Sarma, Physical Review Letters, 2013, 110, 267401. 33. Y. Chen, X. Zhang, C. Jia, Y. Su and Q. Li, The Journal of Physical Chemistry C, 2009, 113, 2263-2266. 34. G. Blasse and B. C. Grabmaier, in Luminescent Materials, Springer Berlin Heidelberg, 1994, DOI: 10.1007/978-3-642-79017-1_3, ch. 3, pp. 33-70. 35. W. Xu, X. Meng, W. Ji, P. Jing, J. Zheng, X. Liu, J. Zhao and H. Li, Chemical Physics Letters, 2012, 532, 72-76. 36. B. Rémi, T. O. Stefan and R. G. Daniel, in Nanocrystal Quantum Dots, Second Edition, CRC Press,2010, DOI: 10.1201/9781420079272-c11, pp. 397-453. 37. J. Ouyang, C. I. Ratcliffe, D. Kingston, B. Wilkinson, J. Kuijper, X. Wu, J. A. Ripmeester and K. Yu, The Journal of Physical Chemistry C, 2008, 112, 4908-4919. 38. K. Kole, C. S. Tiwary and P. Kumbhakar, Journal of Applied Physics, 2013, 113, 114308. 39. D. V. Talapin, R. Koeppe, S. Götzinger, A. Kornowski, J. M. Lupton, A. L. Rogach, O. Benson, J. Feldmann and H. Weller, Nano Letters, 2003, 3, 1677-1681. 40. W. Zhang, G. Chen, J. Wang, B.-C. Ye and X. Zhong, Inorganic Chemistry, 2009, 48, 9723-9731. 41. W. Zhang and X. Zhong, Inorganic Chemistry, 2011, 50, 4065-4072. 42. R. Xie, M. Rutherford and X. Peng, Journal of the American Chemical Society, 2009, 131, 5691-5697. 43. L. Li, T. J. Daou, I. Texier, T. T. Kim Chi, N. Q. Liem and P. Reiss, Chemistry of Materials, 2009, 21, 44. F. Zheng, W. Ping, Z. Xinhua and Y. Yong-Ji, Nanotechnology, 2010, 21, 305604. 45. Coropceanu and M. G. Bawendi, Nano Letters, 2014, 14, 4097-4101. 46. B. O. Dabbousi, J. Rodriguez-Viejo, F. V. Mikulec, J. R. Heine, H. Mattoussi, R. Ober, K. F. Jensen and M. G. Bawendi, The Journal of Physical Chemistry B, 1997, 101, 9463- 9475. 47. C. P. Thomas, A. B. Wedding and S. O. Martin, Solar Energy Materials and Solar Cells, 2012, 98, 455-464. 48. S. Kalytchuk, S. Gupta, O. Zhovtiuk, A. Vaneski, S. V. Kershaw, H. Fu, Z. Fan, E. C. H. Kwok, C.-F. Wang, W. Y. Teoh and A. L. Rogach, The Journal of Physical Chemistry C, 2014, 118, 16393-16400. 49. J.-Y. Jung, K. Zhou, J. H. Bang and J.-H. Lee, The Journal of Physical Chemistry C, 2012, 116, 12409-12414. 50. S.-W. Baek, J.-H. Shim, H.-M. Seung, G.-S. Lee, J.-P. Hong, K.-S. Lee and J.-G. Park, Nanoscale, 2014, 6, 12524-12531. 51. S.-W. Baek, J.-H. Shim and J.-G. Park, Physical Chemistry Chemical Physics, 2014, 16, 18205-18210. 52. S. Singhal, A. K. Chawla, H. O. Gupta and R. Chandra, Thin Solid Films, 2009, 518, 1402-1406. 53. C. K. Huang, Y. C. Chen, W. B. Hung, T. M. Chen, K. W. Sun and W. L. Chang, Progress in Photovoltaics: Research and Applications, 2013, 21, 1507-1513.

- 68 -

Mn-doped colloidal nanocrystals as down-shifting layer for Si solar cells Chapter 3

54. C. Würth, J. Pauli, C. Lochmann, M. Spieles and U. Resch-Genger, Analytical Chemistry, 2012, 84, 1345-1352. 55. C. Würth, M. Grabolle, J. Pauli, M. Spieles and U. Resch-Genger, Nat. Protocols, 2013, 8, 1535-1550. 56. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234-1237. 57. M. Tsuji, D. Yamaguchi, M. Matsunaga and M. J. Alam, Crystal Growth & Design, 2010, 10, 5129-5135. 58. X. Li, D. Shen, J. Yang, C. Yao, R. Che, F. Zhang and D. Zhao, Chemistry of Materials, 2013, 25, 106-112. 59. X. Zhong, M. Han, Z. Dong, T. J. White and W. Knoll, Journal of the American Chemical Society, 2003, 125, 8589-8594. 60. S. Cao, J. Zhao, W. Yang, C. Li and J. Zheng, Journal of Materials Chemistry C, 2015, 3, 8844-8851 61. B. B. Srivastava, S. Jana, N. S. Karan, S. Paria, N. R. Jana, D. D. Sarma and N. Pradhan, J. Phys. Chem. Lett., 2010, 1, 1454-1458. 62. Nag, R. Cherian, P. Mahadevan, A. V. Gopal, A. Hazarika, A. Mohan, A. S. Vengurlekar and D. D. Sarma, J. Phys. Chem. C 2010, 114, 18323-18329. 63. Z. Fang, P. Wu, X. Zhong and Y.-J. Yang, Nanotechnology 2010, 21, 305604. 64. R. Zeng, T. Zhang, G. Dai and B. Zou, J. Phys. Chem. C 2011, 115, 3005-3010. 65. S. Cao, J. Zheng, J. Zhao, L. Wang, F. Gao, G. Wei, R. Zeng, L. Tian and W. Yang, J. Mater. Chem. C 2013, 1, 2540-2547. 66. K. Jong-Uk, L. Myung-Hyun and Y. Heesun, Nanotechnology 2008, 19, 465605. 67. G. Manna, S. Jana, R. Bose and N. Pradhan, J. Phys. Chem. Lett. 2012, 3, 2528-2534. 68. J. Lin, Q. Zhang, L. Wang, X. Liu, W. Yan, T. Wu, X. Bu and P. Feng, J. Am. Chem. Soc. 2014, 136, 4769-4779. 69. Hazarika, A. Pandey and D. D. Sarma, J. Phys. Chem. Lett. 2014, 5, 2208-2213. 70. S. Cao, C. Li, L. Wang, M. Shang, G. Wei, J. Zheng and W. Yang, Sci. Rep. 2014, 4, 7510. 71. J. Zheng, W. Ji, X. Wang, M. Ikezawa, P. Jing, X. Liu, H. Li, J. Zhao and Y. Masumoto, J. Phys. Chem. C 2010, 114, 15331-15336. 72. M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Applied Surface Science, 2011, 257, 2717-2730.

- 69 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Chapter IV

RT Synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals

Chapter 4 describe a facile and rapid room temperature synthesis of cubic and platelet-like colloidal nanocrystals (NCs) of Formamidinium (FA) and Methylammonium (MA) Lead

Halide Perovskite MAPbX3 (X=Br, I) and FAPbX3 (X=Cl, Br, I, or mixed Cl/Br and Br/I) by ligand-assisted re-precipitation method (LARP). The obtained NCs are 5-25 nm in size and exhibit a remarkably high photoluminescence quantum yield (PLQY) of up to 90%. In contrast

to MAPbX3, FAPbX3 NCs demonstrate colloidal and chemical stability. The cubic and platelet- like NCs with their emission in the range of 415-740 nm, full width at half maximum (FWHM) of 20-44 nm and radiative lifetimes of 5−166 ns, allow precise band gap tuning by halide composition as well as by tailoring their dimensions. The study presented here was supported from the Joint Project Helmholtz-Institute Erlangen Nürnberg (HI-ERN) under the project number DBF01253, DFG supported training group 1896 ‘‘In situ microscopy with electrons, X-rays and scanning probes’’ and Electron microscopy resources which was kindly provided by the Center for Nanoanalysis and Electron Microscopy (CENEM). Special thanks to Patrick Herre for TEM and HRTEM measurements.

 Parts of this chapter have been adapted or reproduced with permission from: I. Levchuk, P. Herre, M. Brandl, A. Osvet, R. Hock, W. Peukert, P. Schweizer, E. Spiecker, M. Batentschuk and C. J. Brabec, Chemical Communications, 2017, 53, 244-247.  I. Levchuk, A. Osvet, X. Tang, M. Brandl, J. D. Perea, F. Hoegl, G. J. Matt, R. Hock, M. Batentschuk and C. J. Brabec, Nano Letters, 2017, DOI: 10.1021/acs.nanolett.6b04781.

- 70 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

4.1. Motivation and State of the art

The hybrid organic–inorganic and all-inorganic metal halide perovskite (ABX3, + + + 2+ 2+ − − − where A= Cs , CH(NH2)2 (or FA), CH3NH3 (or MA), B=Sn , Pb , X= Cl , Br , I ) thin films, single crystals films and their colloidal nanocrystals (NCs) offer a wide variability and multiple opportunities for fine tuning with relevance for optoelectronic applications.1-7 Ease of size control and compositional mixing, band-gap and emission tuning inspired researchers to successfully utilize this material class for efficient solar cells8, sensitive photodetectors,9-13 low threshold lasers14, 15, laser diodes,16 advanced photonics,16 light-emitting diodes4, 16, 17 as well as for chemical reactions´ monitoring.18 Furthermore, high photoluminescence quantum yield (PLQY) of these colloidal nanocrystals reaches up to over 90% without additional surface passivation, demonstrating that dangling bonds do not lead to non-radiative relaxation.1, 19 Various methods and approaches have been proposed for the synthesis of organic and inorganic − − − metal halide perovskite colloidal nanocrystals (CsPb(Sn)X3, MAPbX3, X= Cl , Br , I ) including hot-injection,1, 5, 20 ligand-assisted re-precipitation (LARP)3 or structural conversion of lead halide nanocrystals to perovskite by methylammonium halide incorporation.21 Recently, the synthesis of quasi-2D cesium lead halide nanoplatelets (NPLs) with layered structure has been performed by colloidal methods varying the starting ligand ratio, temperature of the synthesis, HBr amount, as well as the length of alkylammonium cations and carboxylic acids.22-27 Concerning the hybrid nanoplatelets (NPLs), only a few papers have demonstrated their thickness tailoring and quantum size effect in suspensions with PLQY <45%.24, 28, 29 For example, Cho et al.29 utilized alkylamines with different lengths to achieve thickness modulation. Sichert et al.24 tuned the ratio of the octylammonium bromide and methylammonium bromide for the same purpose. Furthermore, due to the lower stability and sensitivity to moisture of the organolead iodide perovskite colloids3, 20, the synthesis of the

MAPbI3 NPLs with different thicknesses is more challenging. By this author of dissertation was developed LARP synthesis for highly luminescent (PLQY up to 90%) and nearly

monodisperce MAPbBr3 and MAPbI3 NPLs. Despite their high optical performance, their

colloidal and chemical stability (especially for MAPbI3) still pure and their usage in long-term device applications is premature. Therefore, alternative an analogue to that system with similar properties was strongly desirable. For this purpose, author as an alternative organic action

‐ 71 ‐

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

+ (Formamidinium - CH(NH2)2 (or FA)), bulk pedant of which already performed in record perovskite solar cells with certified efficiency of 22%.30

Formamidinium lead halide perovskites (CH(NH2)2PbX3 or FAPbX3, X=Cl, Br, I) are an advanced class of direct bandgap semiconductors for opto-electronic devices and they have established themselves as a promising alternative for the thermodynamically + 31-35 less stable methyl ammonium (MA, CH3NH3 ) perovskites. Compared to the organic-inorganic MA or all-inorganic Cs perovskite analogues, the pristine formamidinium perovskites have a couple of attractive features like higher thermal, moisture and chemical stability.6, 15, 35-39 Nonetheless, compared to the MA perovskite analogues, the preparation technology of thin films,36 single crystals40, 41 or microcrystalline 32 powders is not well developed due to the following features: (i) FAPbI3 thin films or crystallites typically crystallize in a “yellow” nonperovskite phase36, 40 after a few hours of storage. (ii) Differences in the ionic radii (2.17Å for MA+ and 2.53 for FA+) may affect the growth kinetics.42 The unique properties of the pristine bulk material inspired author to synthesize and investigate the properties of corresponding colloidal nanocrystals, and to establish a new protocol for the synthesis of highly luminescent and stable colloidal perovskite nanocrystals with tuneable optical as well as structural properties.

As synthesized The brightly emitting FAPbX3 NCs are uniform in size, have a high quantum yield of up to 85% and their optical properties can be tuned by band gap engineering via simple halide mixing or size control, resulting in luminescence in the visible range between 415 nm and 740 nm. Furthermore, they shown excellent photostability under prolonged UV-blue light irradiation serving practical attractivity in real application. Additional NC surface modification allows protecting the material against water and storing in solution as well

as in the form of powder for a long time. In the end of this chapter, FAPbBr3 and FAPbI3 simple photodetectors are successfully demonstrated.

4.2. Quantum size-confined MAPbX3 (X=Br, I) colloidal NPLs

The synthesis of colloidal MAPbX3 (X=Br, I) NPLs was performed by controllable precipitation at room temperature, following a modified approach of Zhang et al.3 The main difference to the method presented in the paper of Zhang is using chloroform as

anti-solvent for the MAPbI3 NPLs, and tailoring the size and thickness by varying the oleic acid/oleylamine ligand ratio only for both bromide and iodide NPLs. Injection of a DMF solution of the MAPbBr3 precursor into toluene rapidly forms highly luminescent NPLs. By changing the ratio of oleic acid (OA) and oleylamine (OAm) as

- 72 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

ligands, different band gaps were obtained as well as luminescence wavelengths of the colloids (Figure 4.1a). Upon fixation of the amount of OA in the precursor solution, a certain volume of OAm was added (Table 2.1). This enables a precise tuning of the absorption edge and photoluminescence (PL) of the NPLs from green (514 nm) to blue (447 nm) color.

Figure 4.1. MAPbBr3 NPLs with different thicknesses in toluene solutions. a) Digital photo of the NPLs under UV-365 nm illumination. b) PL decay. c) Absorbance and PL spectra. d) Typical bright-field TEM image of the NPLs with a PL peak at 447 nm obtained with a ligand ratio of OA/OAm = 200 μl/30 μl. e) XRD pattern of NPLs with characteristic PL (green, cyan and blue). However, beginning from the OA/OAm ratio of 200 μl/30 μl, no further PL shift is observed (peak wavelength at 447 nm, Figure 4.2). Only the lateral size of the NPLs was further reduced from 10 to 5 nm at OA/OAm = 200 μl/60 μl (shown in Figure 4.1d and Figure 4.3a). The most challenging task in the synthesis was the cleaning process. Since the organometallic halide perovskite material is highly sensitive to polar solvents and especially water, the conventional alcohol/non-polar solvent (e.g. ethanol/toluene) washing approach which is common for quantum dots (QDs) such as CdSe,43 destroys the perovskite particles. The author found that a certain ratio of acetonitrile/toluene (Table 2.1) allows to precipitate NPLs without decomposition, that was already .44 successfully used for CsPbBr3 NCs Other mixtures such as isopropanol/toluene, acetone/toluene or ethanol/toluene completely destroyed and/or dissolved the perovskite NPLs. Washed NPLs can be re-dissolved in non-polar solvents such as toluene or hexane and form a stable colloidal solution at least for one week under ambient

- 73 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

conditions. Figure 4.1c shows the optical absorption and the PL spectra of a toluene solution of MAPbBr3 NPLs of various sizes that clearly demonstrate the quantum size confinement of the optical band gap. The change from 2.36 to 2.74 eV is in good agreement with the theoretical calculations of Sapori et al.45 Since the reaction kinetics is so fast due to the ionic nature of the metathesis,23 some of the PL spectra reveal a double peak. This is explained by the formation of NPLs with two different thicknesses at the same time, confirmed by electron microscopy analysis discussed below. Figure 4.1d shows a representative bright-field transmission electron microscopy (TEM) micrograph of the square particles having a mean side length of 10.7 nm ± 1.5 nm.

Figure 4.2. PL spectra of the blue (447 nm) emitting NPLs with two different lateral size (5 nm and 11 nm) but same thickness of 1.6 nm. Measurements of the photoluminescence quantum yield revealed an exceptionally high efficiency (89%) for the thickest platelets with a band gap of 2.36 eV (green PL, 514 nm). This efficiency is gradually decreasing with the thickness of the plates, reaching 50% for NPLs with a band gap of 2.59 eV (cyan PL, 488 nm) and 32 % for the platelets with the largest band gap of 2.74 eV (blue PL, 447 nm). To the best of 20, 24, 28 our knowledge, these are the highest reported values for the MAPbBr3 NPLs. The decrease in the quantum yield is accompanied with the reduction of the luminescence decay times from 12 to 5 ns (Figure 4.1 b). While the radiative decay rate remains as a rule almost unchanged, the non-radiative rate, calculated from the quantum yield and PL decay time, increases from 0.9×107 s-1 to 11×107 s-1 (Fig. 4.6a, Table 4.1). Thus, the faster decay in thinner platelets is caused by additional nonradiative relaxation related to the influence of the surface of the platelets. A typical selected electron diffraction pattern (SAED) of the NPLs shown is depicted in Figure

4.3a and nicely confirming the perovskite crystal structure of MAPbBr3.

- 74 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

The corresponding line profile obtained by rotational averaging of the ring pattern is shown in Figure 4.3b. The theoretical peak intensities for random orientation of the NPLs are indicated as red bars. Comparison with the experimentally observed peak intensities indicates a preferred (001) orientation of the NPLs on the carbon support grid consistent with the situation where most NPLs lie flat on carbon support film. This can be seen from the almost complete absence of the (111) peak and the reduced intensity of the (210) peak.

Figure 4.3. Electron diffraction analysis of MAPbBr3 NPLs: a) exemplary diffraction pattern for NPLs with PL peak at 514 nm. b) Line profile obtained by radial averaging of the diffraction pattern and expected peak intensities for random orientation of the NPLs (red lines). The analysis confirms the perovskite crystal structure and points to a preferred <001> orientation of the NPLs on the carbon support grid.

TEM studies of colloidal MAPbBr3 NPLs with blue, cyan and green PL clearly link the observed optical emission to the platelet thickness. Consistent with the preferred crystal orientation deduced from electron diffraction patterns, the NPLs in all of the three samples exhibit a square shape (Figure 4.4a-c, upper row). The average lateral sizes for blue, cyan and green emitting NPLs are 5.4±0.8 nm, 11.6±2.4 nm and 13.8±1.8 nm, respectively. In order to evaluate the thickness of the NPLs, the platelets have to be imaged in edge-on orientation. It turned out that by using diluted hexane (~0.5 mg/ml) instead of concentrated toluene (~10 mg/ml) solution during TEM sample preparation, a considerable number of standing NPLs are observed on the carbon support film.

- 75 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.4. Bright-field TEM characterization of the ligand-assisted layer control of MAPbBr3 NPLs. Lateral particle size and thickness determination are derived from the upper and lower images respectively, of a) blue (447 nm), b) cyan (488 nm) and c) green (514 nm) emitting NPLs.

In particular, the NPLs show upright stacked assemblies on the TEM grid allowing even for a statistical thickness determination directly from the micrographs (Figure 4.4a-c, bottom row). The average thicknesses for blue, cyan and green emitting NPLs are 1.6±0.1nm, 2±0.1 nm and 2.6±0.2 nm, respectively. According to the

MAPbBr3 unit cell size (5.93Å) and the measured thickness in TEM, the number of unit cell layers were estimated by n = 3, 4 and 5 for blue, cyan and green emitting NPLs respectively. The thickness of these platelets is comparable, or less than the 3D Bohr 46 radius in the MAPbBr3 perovskites (~2.2 nm ), below which quantum confinement occurs. Our data on the measured NPL thickness is in excellent agreement with the Bohr 24 radii values and with the data previously reported for suspension type MAPbBr3 NPLs as well as quantum dots.47 Furthermore, the 5 nm NPLs form a self-assembled superstructure on the grid (Figure 4.5a). The distances between the vertically stacked NPLs (2.82 nm) are related to the overlap of two OAm molecules with total length of 2.85 nm (Figure 4.5b-c).

- 76 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.5. Self-assembled superstructure of blue emitting (447 nm) MAPbBr3 NPLs. a) Bright-field TEM image. The marked areas (red and blue squares) and corresponding Fourier transformations indicate a 2D platelet superstructure b) Sketch of a possible OAm ligand overlapping. c) Proposed sketch of NPLs assembling with average distances between individual NPL measured from the TEM image.

In addition to the electron diffraction analysis (Figure 4.3a-b), X-ray diffraction (XRD) was used to determine the crystal structure of all NPLs (Figure 4.1e). The blue, cyan and green emitting NPLs were crystalized in the cubic perovskite structure according to the data obtained from a single crystal.48 In contrast to the samples prepared for electron diffraction (on carbon support film), the samples prepared for XRD (on glass substrate) show a rather random orientation of the NPLs, as can be seen from the relative peak intensities in the θ-2θ profile which nicely correspond to the theoretical values (Figure 4.1e). Since the NPLs are highly anisotropic, the Scherrer equation cannot be employed for an independent determination of the platelet dimensions from the observed peak widths. Therefore, concerning the dimension (thickness and lateral size) of the NPLs author rely on the direct measurement by TEM (Figure 4.4 a-c, bottom row).

In case of MAPbI3 NPLs, only chloroform worked as re-precipitation medium for thicker NPLs with their PL in the orange spectral range, and toluene for the thinner NPLs with yellow-green emission (Table 2.2). Nanocrystals were formed within a second. Washing process was performed for the red and NIR emitting NPLs only by rapid centrifugation at -10°C. NPLs with their PL in the green-orange range were destroyed during the centrifugation due to the higher sensitivity and instability of the

- 77 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

material compared to MAPbBr3 NPLs, as well as due to the presence of DMF in the crude solution.

Figure 4.6. Radiative (krad ) and non-radiative (knon-rad) rate for MAPbBr3 NPLs (a) and MAPbI3 NPLs with different PL emission peaks (b).

Table 4.1. PLQY, PL decay time, radiative (krad ) and non-radiative (knon-rad) rate for MAPbBr3 NPLs with different PL emission peaks. PL Peak , nm 513 509 488 474 452 447 PLQY,% 89 - 50 - - 35 Decay time T1/e, ns 11.9 8.4 5.8 6.6 5.7 5.6 k , s-1 7 7 7 rad 7.5×10 - 8.6×10 - - 6.3×10 k , s-1 7 7 7 non-rad 0.9×10 - 8.6×10 - - 11×10

Table 4.2. PLQY, PL decay time, radiative (krad ) and non-radiative (knon-rad) rate for MAPbI3 NPLs with different PL emission peaks. PL Peak , nm 722 683 628 592 578 549 PLQY,% 31 47 36 - 21 19 Decay time T1/e, ns 19.7 14.2 15.6 8.4 6.8 7.2 k , s-1 7 7 7 7 7 rad 1.5×10 3.3×10 2.3×10 - 3.1×10 2.6×10 k , s-1 7 7 7 7 7 non-rad 3.5×10 3.7×10 4.1×10 - 11×10 11.3×10

Figure 4.7a shows the visual appearance of the pure chloroform and chloroform/toluene solutions of MAPbI3 NPLs under UV-365 nm light; the thickness, band gap and emission wavelength were tailored by varying the OA/OAm ratio

(Table 2.2) in the same manner as for MAPbBr3 NPLs. The PL maximum and absorption edge gradually shift to shorter wavelengths, depending on the NPLs thickness (Fig. 4.7c). This phenomenon is demonstrated for the first time experimentally for MAPbI3 colloids. The PL peaks are in the range of 547 - 722 nm (optical band gap: 2.5-1.70 eV), being in good agreement with theoretical calculations performed by

- 78 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Sapori et al.45 with the estimated monolayer number n between 1 and 8. Colloidal stability of the NIR emitting NPLs from crude solution is higher compared to the washed NPLs due to the partial ligand washing after purification. Non-washed NPLs start to

agglomerate within one day after preparation. Tailoring the MAPbI3 NPLs thickness is slightly different compared to MAPbBr3 NPLs. The ligand-assisted approach in chloroform worked properly in the case of thicker platelets with PL between NIR and orange wavelengths. For the thin yellow and green emitting NPLs author used chloroform/toluene or toluene only for precipitation to obtain stable colloids. This phenomena can be explained by the difference between the dielectric constants of chloroform (ε=4.81), toluene (ε=2.38) and DMF (ε=37.8). In this way, chloroform better and faster blends with DMF if compared to toluene, and forms red and NIR emitting stabilized NPLs only, due to the fast nucleation kinetics of the reagents. In the case of toluene, reaction kinetics is comparably slow, and this is enough to stabilize the just-formed yellow or green emitting NPLs. The quantum yield of these NPLs lies between 20% and 50% with the highest value observed for the red emitting (683 nm) sample. This is the highest reported 3, 20 value for colloidal MAPbI3 perovskite according to the best of our knowledge.

Figure 4.7. a) MAPbI3 NPLs solutions under UV-365 nm illumination. b) PL decay. c) Absorbance and PL spectra. d) Typical bright-field TEM image of the NPLs with a PL peak at 722 nm obtained with a ligand ratio OA/AOm = 200 μl/50 μl e) XRD pattern of purified NPLs with PL of 722 nm (ICSD card №250739).

- 79 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Along with the decreasing PLQY in thinner platelets, the photoluminescence decay time

τ1/e decreases (Figure 4.7b) as well, explained by enhanced non-radiative relaxation. The rate of the latter increased from 3.5×107 s-1 in the platelets with 722 nm emission peak, to 11.3×107 s-1 in the platelets with 546 nm peak wavelength (Figure 4.6b, Table 4.2). The minimum thickness and the largest band gap achieved in our experiments is very similar to the quantum size effect for CdTe QDs,49, 50 where the band gap of the bulk material is almost similar (1.5 vs 1.48 eV) to the value of the 51 bulk MAPbI3 (Figure 4.8a-b). NIR-emitting NPLs have square-like shape in TEM micrograph (Figure 4.7d) with lateral size of 22.3 ± 6 nm. Since it was possible to extract only the NIR and red emitting NPLs, decreased contrast in the red emitting sample confirms its reduced NPL thickness (Figure 4.9a-b) in agreement with previous reports on perovskite NPLs.24, 52 In Figure 4.9a, it clearly shown square plates as well as circle dots, which is a result of NPL degradation of last one due to the higher sensitivity of the thinner red emitting NPLs to high voltage electron irradiation53 compared with NIR emitting NPLs. Importantly, the XRD analysis of the NIR emitting

NPLs sample demonstrates pure tetragonal phase of the MAPbI3 perovskite (the pattern from the Structure database (ICSD card № 250739) is shown in Figure 4.7e.

Figure 4.8. a) Quantum size effect in CdTe quantum dots (QDs)50. b) Quantum size effect in MAPbI3 nanoplates (NPLs).

- 80 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.9. TEM picture of the MAPbI3 NPLs. a) Red NPLs with PL peak at 683 nm; white dashed square indicated non-destroyed NPLs; yellow circles marks destroyed NPLs under high energy beam irradiation b) NIR NPLs with PL peak 722 nm. Differences in contrast of the NPLs display different thicknesses obtained nanocrystals.

4.3. Brightly luminescent and stable FAPbX3 (X=Cl, Br, I) colloidal NCs

Synthesis of colloidal FAPbX3 (X=Cl, Br, I) nanocrystals was performed via LARP technique similar to the one for MAPbX3 (X=Br, I) NPLs. As the anti-solvent media author utilized chloroform in all our experiments. The choice of the anti-solvent plays a decisive role in the LARP method. Toluene, which is commonly used for perovskite NC LARP synthesis,3, 54 did not result in the formation of FAPbI3 NCs, while FAPbBr3 or FAPbCl3 immediately formed a turbid suspension of large particles and agglomerates. In taat case, rapidly injected a

DMF solution of PbX2 and MAX (X=Cl, Br, I) precursors, containing oleic acid (OA) and oleylamine (OAm) as ligands, into vigorously stirred chloroform at room temperature and observed the immediate formation of NCs. Due to the ionic nature of the metathesis, nucleation and growth kinetics are very fast and the NCs appear within seconds. Blending appropriate

FAPbX3 precursor solutions with different aspect ratios is the usual strategy to produce mixed- halide FAPb(Cl/Br)3 and FAPb(Br/I)3 nanocrystals. However, author couldn’t obtain NCs with 1, 55, 56 FAPb(Cl/I)3 composition, mainly due to the large difference in the ionic radii.

The resulting OAm/OA ligand shell passivated FAPbX3 NCs were subsequently washed by precipitation/re-dissolving with an acetonitrile/toluene solvent composition. Afterwards, NCs could be easily re-dissolved in non-polar solvents such as chloroform, toluene and hexane.

The washed FAPbI3 NCs, re-dispersed in toluene, formed a colloidal solution which was stable for at least one week in glovebox under nitrogen atmosphere. This is a significant improvement

- 81 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

3, 20, 57 compared to the MAPbI3 NC dispersions also presented in previous Section 3.2. Not

surprisingly, FAPbBr3 NCs display even higher colloidal and chemical stability, being stable for several weeks under ambient conditions, while FAPbCl3 NCs precipitate within few hours.

+ Figure 4.10. a) Representative perovskite FAPbX3 (X=Cl, Br, I) unit cell with FA cation in 4- the centre of eight [PbX6] octahedra. b) HRTEM of single FAPbI3 NC. c) TEM micrograph of the monolayer FAPbI3 NCs on a grid; σ - size distribution. d) XRD curves for single and mix-halide FAPbX3 NCs.

Figure 4.11 TEM micrograph of mainly lean FAPbCl3 (a-b) and FAPbBr3 (c-d) NPLs.

As formed colloidal FAPbI3 NCs where 14.4 ± 3.4 nm in size with nearly cubic shape

(Figure 4.10c), while FAPbCl3 (Figure 4.11a-b) and FAPbBr3 (Figure 4.11c-d) crystallized as platelets with lateral sizes of 21.5±4 nm and 22±3 nm, respectively. High-resolution

transmission electron microscopy (HRTEM) displays high crystallinity of a single FAPbI3 NC as confirmed by the characteristic lattice spacing of 0.32 nm (Figure 4.10b) related to the (112) plane of the cubic perovskite structure. The measured X-ray diffraction (XRD) patterns display

pure cubic perovskite phase for all three FAPbX3 (X=Cl, Br, I) NCs and their mixed halide

FAPbCl1.5Br1.5 and FAPbBr1.5I1.5 NCs samples (Figure 4.10d). Interestingly, FAPbI3 NCs crystallize in cubic structure with trigonal symmetry which is also known as black α-phase.40, 41 It is well known that the black α-FAPbI3 polycrystalline films as well as single crystals slowly transform to non-perovskite hexagonal yellow δ-phase at room temperature.36, 38-42 Further temperature annealing over 154°C (phase transformation temperature) promotes a reverse

- 82 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

32, 40, 41 transformation from the yellow to the black phase. However, in the FAPbI3 NC colloidal solution investigated in this study no transition from the black to the yellow phase was observed after 150 days of storage (Figure 4.12a). Even after 30 days the drop-casted NC films remained in the black phase upon storage under ambient conditions. A similar phenomenon of reduced-

size stabilization of the cubic polymorph has been observed for CsPbI3 perovskite nanocrystals.1, 8 It is well known that nanosized materials can display different phase transition behaviour as compared to bulk crystals due to the surface energy difference between the polymorphs.58, 59 As the size decreases, the surface-to-volume ratio increases, thus resulting in a lower transition temperature for the nanosized materials. Author found that obtained results 60 are in excellent agreement with the reports by Li et. al on MAPbI3 system, where the temperature of tetragonal (330 K) to orthorhombic (161 K) phase transition strongly depended on the thickness of the microplates.

Figure 4.12. a) XRD patterns of FAPbI3 NCs colloidal solution stored for a period of 150 days. b) Black to yellow phase transition of unpurified FAPbI3 NCs as result of agglomerates formation of poorly passivated NCs upon centrifugation washing process. References for cubic “black” α- and hexagonal “yellow” δ-phase were taken from CIF file reported by Stoumpos et al.61.

Furthermore, purification of those NCs has a strong impact on the black α-phase

stabilisation. Unpurified or badly washed by high-speed centrifugation α-FAPbI3 NCs agglomerate within a few days causing a partial conversion to the yellow phase (Figure 4.12b). On the contrary, washing with acetonitrile/toluene mixtures isolates NCs without considerable ligand removal. A similar effect has been reported for size-confined CsPbI3 quantum dots (QDs).8 However, and to underline the main findings of this investigation, author do not aware

of any previous reports on the synthesis and size-reduced stabilization of α-FAPbI3 NCs perovskites.

- 83 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

The optical UV-Vis absorption and photoluminescence (PL) spectra of the obtained NCs with varying halide composition display a single peak at the first excitonic transition and PL emission (Figure 4.13a). Compared to bulk36, 41, 62 optical absorption edge and the emission

bands of the single halide FAPbX3 (X=Cl, Br, I) NCs are blue-shifted due to the quantum size

effect which will be discussed further. Purified and re-dispersed in toluene FAPbI3 and

FAPbBr3 nanocrystals exhibit a high PLQY of 55% and 84%, respectively, in contrast to

FAPbCl3 which has a PLQY of less than 1%. The mixed halide FAPbCl1.5Br1.5 or FAPbBr1.5I1.5 NC samples have 21% and 23% efficiency, respectively. Lower PLQY of mixed halide 56, 63-65 compositions and that of FAPbCl3 is an intrinsic property of the halide perovskites. Also,

Figure 4.14a-d shows that the PL from colloidal FAPbBr3 NCs film is stable under prolonged exposure to highly intense blue (445 nm, 7.5w/cm2) and UV (375 nm, 3W/cm2) light. In case of blue light, PL peak and FHWM remains the same, while under UV light, 5 nm red-shift was observed (Figure 4.14d), indicating partial NCs sintering at irradiated point.66 This observation suggests that blue light is more applicable for these NCs in luminescent application.

Precise mixing of various halide starting solutions results in a series FAPb(Cl/Br)3 and

FAPb(Br/I)3 NC samples with a seamlessly tuned PL peak wavelength covering (Figure 4.13c) the visible spectral range from 415 nm to 740 nm (2.98-1.68 eV). Furthermore, relatively narrow FHWMs ranging from 20 to 44 nm are observed (Fig. 2d-e). The bandgap undergoes a linear shift with Br/I and Br/Cl ratio following the Vegards law for lattice constants of alloys67-70

Time-resolved PL spectroscopy of FAPbX3 (Cl, Br, I) NCs shows nonexponential decay

traces with average lifetimes of τaver = 15-116 ns with a faster emission for wider band gap NC samples (Figure 4.13b). While the PL decay time in the mixed Br/I - samples increases gradually from 20 to 116 ns with the increase of the content of I ions, the Cl/Br alloys

(FAPbCl1.5Br1.5) display a slower decay dynamics, compared to pure FAPbBr3 (24 ns vs. 20 ns)

and FAPbCl3. This observation may be discussed in context with observations that the addition 71 of Cl improves charge carrier lifetime in FAPbCl1.5Br1.5 NCs. Similar results were reported 72 for CsPb(Cl/Br)3 perovskites.

- 84 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.13. Optical properties of the colloidal solution of FAPbX3 nanocrystals in toluene: a) Representative optical absorption and PL spectra; b) PL decay dynamics of the pure and mixed-halide NCs c) Compositional PL tuning diagram; d) Digital picture of colloidal solution in toluene taken under UV-light (λ=365 nm). e) Compositional tuning of the PL spectra of the mixed halide NCs. Observation of Quantum-size effect. As was shown above (Figure 4.13a), the luminescence bands of single halide nanocrystalline samples are blueshifted compared to the bulk due to quantum size effect. In colloidal semiconductor NCs (e.g. CdSe) this phenomenon occurs if the size of the crystallites becomes comparable to or smaller than the exciton Bohr radius in the bulk material. Decreasing the size of NCs down to the size of the exciton Bohr radius leads to a blueshift of their absorption edge and the PL peak. Size tuning in hybrid organic-inorganic or all-inorganic perovskite NCs22-27 was recently reported by several groups.

Figure 4.14. Photostability of the FAPbBr3 NCs film under different light irradiation. a) PL intensity upon 445 nm (7.5W/cm2) laser illumination. b) PL spectra after different exposure times. Note that the PL peak energy and the FWHM do not change with exposure to blue light. c) PL intensity upon 375 nm (3W/cm2) laser illumination. d) PL spectra after different exposure times. The PL peak shows a red shift of 5 nm.

- 85 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

In previous Section 3.2 the thicknesses tailoring in MAPbX3 (X=Br, I) NPLs was

demonstrated, which lead to a one dimensional quantum confinement. Since FAPbBr3 NCs crystallize as NPLs, author applied the same technique here. The ligand ratio control allowed

to tailor the thickness of the platelet-like FAPbBr3 NCs, leading to shift of the narrow PL band (Figure 4.15a) and the band-band absorption edge. The measured PL spectra are blueshifted from 533 to 438 nm with the increasing OAm content. Such a gradual blueshift indicates quantum confinement in FAPbBr3 NPLs which is supported by TEM image analysis

(Figure 4.15b-d). Similar features was observed for FAPbI3 and FAPbCl3 (not shown here). The observed NPL emission peak at 438 nm and its PL shape are in excellent agreement with data of Weidman et al.73 for two-layer perovskite NPL. Furthermore, PLQY value presented in our work for these NPLs (21%) is almost similar to the data of Weidman et al. (22%). From the evaluation of the TEM images was found the thickness of blue (438 nm), cyan (486 nm) and green (533 nm) emitting FAPbBr3 NPLs is 1.4±0.1 nm, 2±0.1 nm and 2.6±0.2 nm respectively (Figure 4.15b-d). This data is in excellent agreement with the calculated thickness of NPLs with n=2 (1.2 nm), n=3 (1.8 nm) and n=4 (2.4 nm), where n is the number of FAPbBr3 monolayers with the unit cell thickness of 5.9944 Å. The average lateral size of blue, cyan and green emitting NPLs is 36.6±9 nm, 22.7±3 nm and 21.5±4 nm respectively. The absolute PLQY increases from 21% (NPLs with 438 nm PL peak wavelength) to a remarkably high value of 84% (500 nm peak wavelength) (Figure 4.15a). For the thickest, more bulky NPL (2.6 nm), the PLQY drops to 74%. Previous studies discussed similar findings in the context of reduced exciton stabilization3, 47. Identical trends were also observed for the structurally 3, 24 similar MAPbBr3 micro- or single crystals which exhibit significantly lower PLQY (<1%) 57, 74 than their nanocrystalline pendants (>90%) . Our PLQY data for size-confined FAPbBr3 6 NPLs is in excellent agreement with the PLQY values of cube-like FAPbBr3 NCs . To estimate the Bohr radius in FA-perovskites, author used the values of reduced 62 effective mass (μ) and effective dielectric constant (εeff) calculated by Galkowski et al. for

bulk FAPbBr3 and FAPbI3 perovskites.

The Bohr radius of the three-dimensional exciton for FAPbBr3 is defined as:

휖 푚0 푚0 푎퐵 = 푎퐻 = 휖푒푓푓 푎퐻 = 3.50071푛푚 (1.1) 휖0 휇 휇

where 휖0 and 휖 are the dielectric constants of vacuum and the material 휖푒푓푓 = 8.6 ), 휇 and 푚0

62 are the effective masses of the exciton (휇=0.13푚0) and the free electron, and 푎퐻 is the Bohr radius (0.052917721nm).

- 86 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.15. Comprehensive characterization of purified FAPbBr3 NPLs toluene solutions with different thicknesses. a) Band gap and PL tuning by changing the OAm/OA ratio with corresponding PLQY values; b–d) Bright-field TEM characterization of the vertically stacked FAPbBr3 NPLs synthesized with different OAm/OA volume ratio: b) 200 μl /150 μl blue (438 nm) emitting NPLs with lateral size 36.6±9 nm and thickness 1.4±0.1 nm; c) 200 μl /80 μl cyan (486 nm) emitting NPLs with lateral size 22.7±3 nm and thickness 2±0.1nm; d) 200 μl /40 μl green (533 nm) emitting NPLs with lateral size 21.5±4 nm and thickness 2.6±0.2 nm; σ - thicknesses distribution. e) Theoretical (EMA) versus experimental band gap of the FAPbBr3 NPLs as a function of the thicknesses (d). Inset shows unit cell number (n) for single NPL with different thickness; f) PL decay.

For FAPbI3 :

휖 푚0 푚0 푎퐵 = 푎퐻 = 휖푒푓푓 푎퐻 = 6.35013푛푚 (1.2) 휖0 휇 휇

where are 휖푒푓푓 = 11.4 is the effective dielectric constant of the material, 휇 and 푚0 are the effective masses of the exciton (휇=0.095푚0) and the free electron, and 푎퐻 is the Bohr radius (0.052917721nm).

The size dependent band gap evolution of FAPbBr3 platelet-like NCs was carried out by effective mass approximation (EMA)75 which was successfully employed for various Quantum Dots (QDs) (CdSe, CdS, ZnS etc.). The effective mass approximation comes from solving the Schrödinger equation for an isolated electron and then for an isolated hole in a sphere, and assuming that the effective masses of carriers in a bulk semiconductor is76:

- 87 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

2 2 1 푚0 푚0 2 ℎ 1.8푒 퐸푔(푅) = 퐸푔,퐵푢푙푘 + 2 ( − ) − (1.3) 푅 푚푒 푚ℎ 8푚0 휖푒푓푓푅

where the first term is the band gap of the bulk materials, which was 2.13eV41. The second term approximate the kinetic energy contribution to the electron-hole pair energy. It is well

known that the product hc has a value of 1239.77 eV nm, and then the term h2/(8mo) is equal

to 0.376036 eV nm. The effective reduced mass 휇 of the electron and the hole (1/ 휇=1/me+1/mh) 62 of FAPbBr3 has a value of 0.13mo . The third term is the Coulomb attraction well known in 77 62 the literature , where R, 휖푒푓푓 (8.6) and e are the radii, the effective dielectric constant of the semiconductor and the free electron charge constant respectively. Then:

2 2 0.376036푚0 3.572푒 0.376036 1.8푒 퐸푔(푅) = 2.13푒푉 + 2 − = 2.13푒푉 + 2 − (1.4) 푅 휇 휖푒푓푓푅 0.13푅 휖푒푓푓푅

The dependence of the band gap of the FAPbBr3 NPLs on their thickness is shown in Figure 4.15e. For comparison was calculated the size-dependent bandgap of quantum dots with the same characteristic size (solid line in Figure 4.15e), using the EMA, and the literature values for the effective masses and dielectric constant.62 The calculated bandgap follows the trend in experimental values of the bandgap, although overestimates it especially in case of smaller particles. As expected, the PL decay time gradually increases with the FAPbBr3 NPLs thickness from 5 to 23 ns (Figure 4.15f). This is in excellent agreement with the PLQY enhancement with increasing thickness. The faster and nonexponential PL decay curves and lower PLQY in thinner NPLs may be explained by additional nonradiative relaxation caused by Br-rich surface 47 causing the quenching of the luminescence. Similar results were observed for MAPbBr3 NPLs in previous Section 3.2.

Such a high PLQY of FAPbX3 NCs in the visible spectral region and the remarkable stability of the FA-perovskites compared to other perovskites (MA or Cs) make this material an ideal candidate for optoelectronic devices or for application as a phosphor material. However, moisture resistance is still a problem for organo-lead halide perovskites.78 Only special surface coatings or the incorporation in a polymer matrix may protect the material from dissolution in water.79-81 To increase the stability of the NPLs, author applied a post- synthesis surface modification by polyhedral oligomeric silsesquioxane, previously proposed 79 by Rogach et al. This allows to protect the brightly luminescent nanocrystalline FAPbX3 from decomposition by water. The NC colloids and nanocrystalline powders with various halide compositions are stable in toluene solution as well as in water for at least 2 months without losing the brightness of their emission (Figure 4.16a-d).

- 88 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Figure 4.16. a) Illustration of FAPbX3 NCs coating with Mercaptopropylisobutyl-POSS cascade molecule resulting bright emitted powder and toluene solution. b) - d) PL spectra of FAPbCl2Br, FAPbBr3 and FAPbBrI2 NCs powder before and after dipping into the water.

Figure 4.17. a) The current–voltage (I–V) curves of FAPbI3 drop casted NCs film on patterned ITO finger substrates (inset figure) in the dark (black) and under AM 1.5 illumination (red). b) On/Off switching properties (at a bias of 3 V).

- 89 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

Photodetector. To examine the photoconductivity of the FAPbBr3 and FAPbI3 perovskite NCs, widely used in highly efficient perovskite photovoltaics, the photocurrent and photoresponse properties presented. A photodetector is fabricated by drop-casting of OAm/OA capped NCs on In2O3:Sn (ITO) patterned finger substrate (ITO). The I-V curves in the dark and under AM 1.5 illumination were measured (Figure 4.17a,c). The calculated light/dark ratio

of the as-synthesized FAPbBr3 and FAPbI3 NCs are ca.1.3 and ~2 respectively. Figure 4.17b,d, presents the on-off switching characteristics (at a bias of 3V) of photoconductors based on the

purified FAPbBr3 and FAPbI3 film NCs. The ON and OFF time durations are both 100 seconds. It is clearly seen that all the devices can be switched on and off repeatedly. The above-described

photo-sensing measurements show that FAPbBr3 and FAPbI3 NCs with further surface treatment for ligand removal, are very promising as materials for QDs photovoltaic and LEDs applications. Fast anion-exchange. It is already known that perovskite NCs bandgap as well as PL can be tuned via post reaction treatment anion-exchange reaction by adding halide-contain precursors to synthesized or purified NCs solutions (Figure 4.18a). This technique was

successfully performed for CsPbX3 and MAPbX (X = Cl, Br, I) NC systems and thin film of thereof by Kovalenko82, Manna63 and Song11 groups. Their PL emission spectra can be adjusted

in the range of 400-730 nm and high PLQY of 20-80%. FAPbX3 system was also successfully exanimated for anion-exchange reaction and summary results depicted in Figure 4.18a-e.

To crude FAPbBr3 NCs solution certain amount of 0.5M Tetra-n-butylammonium halide (TBAX (X= Cl, I)) chloroform/toluene solution were added resulting in fast and gradual changing of PL emission color as well as optical absorption to correspondence pure or mix-halide perovskite NCs (Figure 4.18b). Similar changes was observed in XRD spectra that confirm successful and homogenous halide exchange at structural level preserving size and morphology. Diffractograms for pure halide samples are in excellent agreement with separately synthesized samples (Figure 4.10d). It worth to notice that this anion exchange reactions are

reversible for each NCs pairs (FAPbCl3⇆FAPbBr3 and FAPbBr3⇆FAPbI3).

However, after purification the pristine FAPbBr3 NCs and further re-dissolving in toluene, anion-exchange did not work and solution becomes colorless and non-luminescent which indicates rapid NCs degradation upon TBAX adding. Therefore alternative halide source such as Oleylammonium halides (OAmX, X=Cl, Br, I) were used.63 Figure 4.18d depict the gradual

PL red-shift upon adding OAmI to toluene solution of purified FAPbBr3. Final converted

FAPbI3 NCs shown appreciable blue-shift in comparison to separately synthesized sample

(Figure 4.18d), which is result of the quantum-size effect in FAPbI3 NCs (initial thickness of

- 90 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

FAPbBr3 NCs are 2.6±0.2 nm and Bohr radius for FAPbI3 is 6.35 nm). Similar to CsPbX3 perovskite NCs, anion exchange also occurs while NCs are blended (Figure. 4.18d). PL and absorption spectra can be finely tuned by mixing of desirable ratio of NCs solutions.63, 82

Figure. 4.18. a) Schematic illistration of anion-exchange raeactions within FAPbX3; b) Optical absorption and PL spectra of anion-exchanged FAPbBr3 NCs to their Cl and I pedants by adding TBAX (X=Cl, I) to FAPbBr3 NCs crude solution. c) XRD of anion exchanged NCs. d) PL spectra evolution upon anion-exchange of purified FAPbBr3 NCs by OAmI toluene solution adding. e) Interparticle anion-exchange realized by blending purified FAPbBr3 and FAPbI3 NCs toluene solutions in different ratio. Inset show result of this reaction under UV-365 nm lamp.

- 91 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

4.4. Summary

In this chapter room-temperature, fast reliable and scalable synthesis of colloidal hybrid

organometallic halide perovskite MAPbX3 (X=Br, I) and FAPbX3 (X=Cl, Br, I, or mixed Cl/Br and Br/I) was successfully performed via a ligand-assisted re-precepitation method enabling a precise size and composition control as well as bangap of thereof. Thicknesses of the NPL can be tailored very precisely between 2-5 unit cell monolayers demonstrating size-dependent

optical properties. As synthesized colloidal MAPbBr3 NPLs have ideal square shape enable to form 2D self-assembled superstructures and superlattices, underlining the high quality of

nanocrystals. Both MAPbX3 and FAPbX3 NCs exhibit remarkably high PLQY (up to 90% ) and are highest reported values for size-confined NPLs of thereof. FAPbX3 cubic and platelet- like NCs with their emission in the range of 415-740 nm, full width at half maximum (FWHM) of 20-44 nm and radiative lifetimes of 5−166 ns. In contrast to MAPbX3 NCs, FAPbX3 exhibit a higher colloidal, chemical stability and photostability that makes this material almost ideal for various applications.

Notably, for the first time we demonstrate thermodynamically stable FAPbI3 NCs in the black cubic α-phase without transition to the yellow hexagonal δ-phase even after 150 days of storage. This is in strong contrast to polycrystalline films and single crystals which convert within hours. This fact paves the way to highly efficient perovskite based quantum dots (QDs) photovoltaics, which is underpinned by demonstrating FAPbBr3 and FAPbI3 NCs based photodetector. To highlight the potential of FAPbX3 NCs as a promising candidate for optoelectronic and luminescent applications, we modified the surface with polyhedral oligomeric silsesquioxane. This modification protects the brightly luminescent FAPbX3 NCs from decomposition even after storage in water for more than 2 months.

- 92 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

4.5. Bibliography

1. L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Krieg, R. Caputo, C. H. Hendon, R. X. Yang, A. Walsh and M. V. Kovalenko, Nano Lett., 2015, 15, 3692-3696. 2. L. C. Schmidt, A. Pertegás, S. González-Carrero, O. Malinkiewicz, S. Agouram, G. Mínguez Espallargas, H. J. Bolink, R. E. Galian and J. Pérez-Prieto, J. Am. Chem. Soc., 2014, 136, 850-853. 3. F. Zhang, H. Zhong, C. Chen, X.-g. Wu, X. Hu, H. Huang, J. Han, B. Zou and Y. Dong, ACS Nano, 2015, 9, 4533-4542. 4. M. F. Aygüler, M. D. Weber, B. M. D. Puscher, D. D. Medina, P. Docampo and R. D. Costa, J. Phys. Chem. C, 2015, 119, 12047-12054. 5. T. C. Jellicoe, J. M. Richter, H. F. J. Glass, M. Tabachnyk, R. Brady, S. E. Dutton, A. Rao, R. H. Friend, D. Credgington, N. C. Greenham and M. L. Böhm, J. Am. Chem. Soc., 2016, 138, 2941-2944. 6. L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Bertolotti, N. Masciocchi, A. Guagliardi and M. V. Kovalenko, J. Am. Chem. Soc., 2016, DOI: 10.1021/jacs.6b08900. 7. A. Perumal, S. Shendre, M. Li, Y. K. E. Tay, V. K. Sharma, S. Chen, Z. Wei, Q. Liu, Y. Gao, P. J. S. Buenconsejo, S. T. Tan, C. L. Gan, Q. Xiong, T. C. Sum and H. V. Demir, Sci. Rep., 2016, 6, 36733. 8. A. Swarnkar, A. R. Marshall, E. M. Sanehira, B. D. Chernomordik, D. T. Moore, J. A. Christians, T. Chakrabarti and J. M. Luther, Science, 2016, 354, 92-95. 9. P. Ramasamy, D.-H. Lim, B. Kim, S.-H. Lee, M.-S. Lee and J.-S. Lee, Chemical communications, 2016, 52, 2067-2070. 10. L. Lv, Y. Xu, H. Fang, W. Luo, F. Xu, L. Liu, B. Wang, X. Zhang, D. Yang, W. Hu and A. Dong, Nanoscale, 2016, 8, 13589-13596. 11. D. M. Jang, K. Park, D. H. Kim, J. Park, F. Shojaei, H. S. Kang, J.-P. Ahn, J. W. Lee and J. K. Song, Nano letters, 2015, 15, 5191-5199. 12. M. I. Saidaminov, M. A. Haque, M. Savoie, A. L. Abdelhady, N. Cho, I. Dursun, U. Buttner, E. Alarousu, T. Wu and O. M. Bakr, Adv. Mater., 2016, 28, 8144-8149. 13. M. I. Saidaminov, V. Adinolfi, R. Comin, A. L. Abdelhady, W. Peng, I. Dursun, M. Yuan, S. Hoogland, E. H. Sargent and O. M. Bakr, Nat. Commun., 2015, 6, 8724. 14. S. Yakunin, L. Protesescu, F. Krieg, M. I. Bodnarchuk, G. Nedelcu, M. Humer, G. De Luca, M. Fiebig, W. Heiss and M. V. Kovalenko, Nat. Commun., 2015, 6, 8056. 15. Y. Fu, H. Zhu, A. W. Schrader, D. Liang, Q. Ding, P. Joshi, L. Hwang, X. Y. Zhu and S. Jin, Nano Lett., 2016, 16, 1000-1008. 16. S. Colella, M. Mazzeo, A. Rizzo, G. Gigli and A. Listorti, J. Phys. Chem. Lett., 2016, 7, 4322-4334. 17. S. G. R. Bade, J. Li, X. Shan, Y. Ling, Y. Tian, T. Dilbeck, T. Besara, T. Geske, H. Gao, B. Ma, K. Hanson, T. Siegrist, C. Xu and Z. Yu, ACS Nano, 2016, 10, 1795-1801. 18. T. L. Doane, K. L. Ryan, L. Pathade, K. J. Cruz, H. Zang, M. Cotlet and M. M. Maye, ACS Nano, 2016, 10, 5864-5872. 19. H. Huang, A. S. Susha, S. V. Kershaw, T. F. Hung and A. L. Rogach, Advanced Science, 2015, 2, n/a-n/a. 20. O. Vybornyi, S. Yakunin and M. V. Kovalenko, Nanoscale, 2016, 8, 6278-6283. 21. Y. Hassan, Y. Song, R. D. Pensack, A. I. Abdelrahman, Y. Kobayashi, M. A. Winnik and G. D. Scholes, Adv. Mater., 2016, 28, 566-573. 22. Q. A. Akkerman, S. G. Motti, A. R. Srimath Kandada, E. Mosconi, V. D’Innocenzo, G. Bertoni, S. Marras, B. A. Kamino, L. Miranda, F. De Angelis, A. Petrozza, M. Prato and L. Manna, J. Am. Chem. Soc., 2016, 138, 1010-1016.

- 93 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

23. Y. Bekenstein, B. A. Koscher, S. W. Eaton, P. Yang and A. P. Alivisatos, Journal of the American Chemical Society, 2015, 137, 16008-16011. 24. J. A. Sichert, Y. Tong, N. Mutz, M. Vollmer, S. Fischer, K. Z. Milowska, R. García Cortadella, B. Nickel, C. Cardenas-Daw, J. K. Stolarczyk, A. S. Urban and J. Feldmann, Nano Letters, 2015, 15, 6521-6527. 25. A. Pan, B. He, X. Fan, Z. Liu, J. J. Urban, A. P. Alivisatos, L. He and Y. Liu, ACS Nano, 2016, 10, 7943-7954. 26. S. Sun, D. Yuan, Y. Xu, A. Wang and Z. Deng, ACS Nano, 2016, DOI: 10.1021/acsnano.5b08193. 27. P. Tyagi, S. M. Arveson and W. A. Tisdale, J. Phys. Chem. Lett., 2015, 6, 1911-1916. 28. Z. Yuan, Y. Shu, Y. Xin and B. Ma, Chemical Communications, 2016, 52, 3887-3890. 29. J. Cho, Y.-H. Choi, T. E. O’Loughlin, L. De Jesus and S. Banerjee, Chemistry of Materials, 2016, 28, 6909-6916. 30. NREL, Best Research-Cell Efficiencies, http://www.nrel.gov/ncpv/images/efficiency_chart.jpg 2016. 31. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J.-P. Correa Baena, J.-D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, Sci. Adv., 2016, 2, e1501170. 32. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476-480. 33. M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt and M. Gratzel, Energy Environ. Sci., 2016, 9, 1989-1997. 34. T. Jesper Jacobsson, J.-P. Correa-Baena, M. Pazoki, M. Saliba, K. Schenk, M. Gratzel and A. Hagfeldt, Energy Environ. Sci., 2016, 9, 1706-1724. 35. J. S. Manser, J. A. Christians and P. V. Kamat, Chem. Rev., 2016, 116, 12956-13008. 36. G. E. Eperon, S. D. Stranks, C. Menelaou, M. B. Johnston, L. M. Herz and H. J. Snaith, Energy Environ. Sci., 2014, 7, 982-988. 37. J. Song, W. Hu, X.-F. Wang, G. Chen, W. Tian and T. Miyasaka, J. Mater. Chem. A, 2016, 4, 8435-8443. 38. E. Smecca, Y. Numata, I. Deretzis, G. Pellegrino, S. Boninelli, T. Miyasaka, A. La Magna and A. Alberti, PCCP, 2016, 18, 13413-13422. 39. A. Binek, F. C. Hanusch, P. Docampo and T. Bein, J. Phys. Chem. Lett., 2015, 6, 1249- 1253. 40. A. A. Zhumekenov, M. I. Saidaminov, M. A. Haque, E. Alarousu, S. P. Sarmah, B. Murali, I. Dursun, X.-H. Miao, A. L. Abdelhady, T. Wu, O. F. Mohammed and O. M. Bakr, ACS Energy Lett., 2016, DOI: 10.1021/acsenergylett.6b00002, 32-37. 41. M. I. Saidaminov, A. L. Abdelhady, G. Maculan and O. M. Bakr, Chem. Commun., 2015, 51, 17658-17661. 42. Y. Zhou, M. Yang, S. Pang, K. Zhu and N. P. Padture, J. Am. Chem. Soc., 2016, 138, 5535-5538. 43. O. Chen, X. Chen, Y. Yang, J. Lynch, H. Wu, J. Zhuang and Y. C. Cao, Angewandte Chemie International Edition, 2008, 47, 8638-8641. 44. S. Yakunin, L. Protesescu, F. Krieg, M. I. Bodnarchuk, G. Nedelcu, M. Humer, G. De Luca, M. Fiebig, W. Heiss and M. V. Kovalenko, Nat Commun, 2015, 6. 45. D. Sapori, M. Kepenekian, L. Pedesseau, C. Katan and J. Even, Nanoscale, 2016, DOI: 10.1039/C5NR07175E. 46. K. Tanaka, T. Takahashi, T. Ban, T. Kondo, K. Uchida and N. Miura, Solid State Communications, 2003, 127, 619-623. 47. L. Peng, A. Tang, C. Yang and F. Teng, J. Alloys Compd., 2016, 687, 506-513.

- 94 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

48. D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, A. Buin, Y. Chen, S. Hoogland, A. Rothenberger, K. Katsiev, Y. Losovyj, X. Zhang, P. A. Dowben, O. F. Mohammed, E. H. Sargent and O. M. Bakr, Science, 2015, 347, 519-522. 49. A. L. Rogach, T. Franzl, T. A. Klar, J. Feldmann, N. Gaponik, V. Lesnyak, A. Shavel, A. Eychmüller, Y. P. Rakovich and J. F. Donegan, The Journal of Physical Chemistry C, 2007, 111, 14628-14637. 50. G. B. Sukhorukov, A. L. Rogach, M. Garstka, S. Springer, W. J. Parak, A. Muñoz-Javier, O. Kreft, A. G. Skirtach, A. S. Susha, Y. Ramaye, R. Palankar and M. Winterhalter, Small, 2007, 3, 944-955. 51. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-647. 52. V. A. Hintermayr, A. F. Richter, F. Ehrat, M. Döblinger, W. Vanderlinden, J. A. Sichert, Y. Tong, L. Polavarapu, J. Feldmann and A. S. Urban, Advanced Materials, 2016, DOI: 10.1002/adma.201602897, n/a-n/a. 53. H. Yuan, E. Debroye, K. Janssen, H. Naiki, C. Steuwe, G. Lu, M. Moris, E. Orgiu, H. Uji-i, F. De Schryver, P. Samorì, J. Hofkens and M. Roeffaers, The Journal of Physical Chemistry Letters, 2016, 7, 561-566. 54. X. Li, Y. Wu, S. Zhang, B. Cai, Y. Gu, J. Song and H. Zeng, Adv. Funct. Mater., 2016, 26, 2435-2445. 55. S. Sharma, N. Weiden and A. Weiss, Z. Phys. Chem., 1992, 175, 63-80. 56. B. A. Rosales, L. Men, S. D. Cady, M. P. Hanrahan, A. J. Rossini and J. Vela, Chem. Mater., 2016, 28, 6848-6859. 57. I. Levchuk, P. Herre, M. Brandl, A. Osvet, R. Hock, W. Peukert, P. Schweizer, E. Spiecker, M. Batentschuk and C. J. Brabec, Chem. Commun., 2016, 53, 244-247. 58. M. Mayo, A. Suresh and W. Porter, Reviews on Advanced Materials Science, 2003, 5, 100-109. 59. J. B. Rivest, L.-K. Fong, P. K. Jain, M. F. Toney and A. P. Alivisatos, J. Phys. Chem. Lett., 2011, 2, 2402-2406. 60. D. Li, G. Wang, H.-C. Cheng, C.-Y. Chen, H. Wu, Y. Liu, Y. Huang and X. Duan, Nat. Commun., 2016, 7, 11330. 61. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorg. Chem., 2013, 52, 9019- 9038. 62. K. Galkowski, A. Mitioglu, A. Miyata, P. Plochocka, O. Portugall, G. E. Eperon, J. T.- W. Wang, T. Stergiopoulos, S. D. Stranks, H. J. Snaith and R. J. Nicholas, Energy Environ. Sci., 2016, 9, 962-970. 63. Q. A. Akkerman, V. D’Innocenzo, S. Accornero, A. Scarpellini, A. Petrozza, M. Prato and L. Manna, Journal of the American Chemical Society, 2015, 137, 10276-10281. 64. N. Pellet, J. Teuscher, J. Maier and M. Grätzel, Chem. Mater., 2015, 27, 2181-2188. 65. S. D. Stranks, V. M. Burlakov, T. Leijtens, J. M. Ball, A. Goriely and H. J. Snaith, Phys. Rev. Appl., 2014, 2, 034007. 66. J. Chen, D. Liu, M. J. Al-Marri, L. Nuuttila, H. Lehtivuori and K. Zheng, Sci. China Mater., 2016, 59, 719-727. 67. R. J. Sutton, G. E. Eperon, L. Miranda, E. S. Parrott, B. A. Kamino, J. B. Patel, M. T. Hörantner, M. B. Johnston, A. A. Haghighirad, D. T. Moore and H. J. Snaith, Adv. Energy Mater., 2016, 6, 1502458. 68. R. K. Misra, L. Ciammaruchi, S. Aharon, D. Mogilyansky, L. Etgar, I. Visoly-Fisher and E. A. Katz, ChemSusChem, 2016, 9, 2572-2577. 69. U.-G. Jong, C.-J. Yu, J.-S. Ri, N.-H. Kim and G.-C. Ri, Phys. Rev. B, 2016, 94, 125139. 70. J. H. Noh, S. H. Im, J. H. Heo, T. N. Mandal and S. I. Seok, Nano Lett., 2013, 13, 1764- 1769.

- 95 -

RT synthesis of Highly Luminescent Perovskite Colloidal Nanocrystals Chapter 4

71. M. Zhang, H. Yu, M. Lyu, Q. Wang, J.-H. Yun and L. Wang, Chem. Commun., 2014, 50, 11727-11730. 72. D. Zhang, Y. Yang, Y. Bekenstein, Y. Yu, N. A. Gibson, A. B. Wong, S. W. Eaton, N. Kornienko, Q. Kong, M. Lai, A. P. Alivisatos, S. R. Leone and P. Yang, J. Am. Chem. Soc., 2016, 138, 7236-7239. 73. M. C. Weidman, M. Seitz, S. D. Stranks and W. A. Tisdale, ACS Nano, 2016, 10, 7830- 7839. 74. H. Huang, A. S. Susha, S. V. Kershaw, T. F. Hung and A. L. Rogach, Adv. Sci., 2015, 2, 1500194. 75. P. E. Lippens and M. Lannoo, Phys. Rev. B, 1989, 39, 10935-10942. 76. S. V. Gaponenko, Optical Properties of Semiconductor Nanocrystals, Cambridge University Press, 1998. 77. T. Takagahara and K. Takeda, Phys. Rev. B, 1992, 46, 15578-15581. 78. T. A. Berhe, W.-N. Su, C.-H. Chen, C.-J. Pan, J.-H. Cheng, H.-M. Chen, M.-C. Tsai, L.- Y. Chen, A. A. Dubale and B.-J. Hwang, Energy Environ. Sci., 2016, 9, 323-356. 79. H. Huang, B. Chen, Z. Wang, T. F. Hung, A. S. Susha, H. Zhong and A. L. Rogach, Chem. Sci., 2016, 7, 5699-5703. 80. Y. Wang, J. He, H. Chen, J. Chen, R. Zhu, P. Ma, A. Towers, Y. Lin, A. J. Gesquiere, S.-T. Wu and Y. Dong, Adv. Mater., 2016, 28, 10710-10717. 81. S. Huang, Z. Li, L. Kong, N. Zhu, A. Shan and L. Li, J. Am. Chem. Soc., 2016, 138, 5749-5752. 82. G. Nedelcu, L. Protesescu, S. Yakunin, M. I. Bodnarchuk, M. J. Grotevent and M. V. Kovalenko, Nano letters, 2015, 15, 5635-5640.

- 96 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Chapter V

Perovskite ink: from nanocrystaline powder to device application

Following the work presented in the Chapter 4 where perovskite colloidal NCs were synthesized via LARP technique, this Chapter presents a similar antisolvent-induced crystallisation of ligand-free nanosized perovskite crystallites with various elements content

(ABX3; A= FA, MA; B=Pb, Sn, Ge, Mn, Eu; X=Cl, Br, I). A corresponding precursor-free ink based on these nanocrystaline powder results in high photodetector performance opening a venue for large scale production of perovskite thin film devices. Moreover, all coating processes were conducted at low temperatures (<80°C) via simple blading technique. The quality of the differently films can be improved or healed via ease introduction of gaseous methylamine (CH3NH2) increasing crystallinity, healing defects and significantly improving photodetector performance. The easy and low-cost perovskite ink processing shows big potential for industrial application and allows straightforward fabrication of optoelectronic devices via precursor-free fashion. The author gratefully acknowledge ﬁnancial support from the Joint Project Helmholtz- Institute Erlangen Nürnberg (HI-ERN) under the project number DBF01253. Special thanks to Patrick Herre (LFG), Bastian Weisenseel (WW3) and Mykhailo Syntyk (EnCN) for SEM measurements of perovskite powder and film as well as Amir Abbas Yousefi Amin (EnCN) for photoresponsivity measurements. Additional thanks to Gebhard J. Matt (i-MEET) for useful idea and photodetector characterizations.

 Parts of this chapter have been prepared in cooperation with Florian Högl and Johannes Bergmann (former students at i-MEET, supervised by the author),

- 97 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

5.1. Motivation and State of the art

The production of defect-free perovskite film reference systems is needed to study the fundamental photo-physical processes. Furthermore, to commercially use and industrialize this material a scalable and low-temperature synthesis is beneficial, which can be performed via a simple synthesis setup. Current state of the art of lead-halide perovskite solar cell usually requires multistep

annealing processes of the wet perovskite precursor containing dissolved PbX2 and AX (A=Cs, MA, FA; X=Cl, Br, I) in polar solvents.1 As a result, the complexity of the thin film crystallization upon highly boiling solvent (DMF - N,N-Dimethylformamide, DMSO - Dimethyl sulfoxide, GBL - γ-Butyrolactone) evaporation impact on the device performance. Therefore, design of precursor-free perovskite inks is the way towards low cost and low temperature perovskite film manufacturing, which can revolutionize the market for light harvesting devices. This Chapter describe a study on low temperature, generalized, full solution based process to fabricate organic-inorganic metal halide perovskites and inks thereof. It was possible to successfully synthesize organometallic halide perovskites material with + + 2+ 2+ 2+ − − − general formula ABX3 (A= NH2CHNH2 , CH3NH3 , B=Pb Sn , Ge ; X= Cl , Br , I ). Further stable inks based on the synthesised perovskite particles were accomplished. These inks were successfully applied onto various substrates via doctor blading, thus enabling the production of photo- and optoelectronic devices via simple coating processes, forming a major step for large scale production. 5.2. Perovskite nanocrystalline powder

Antisolvent-assisted perovskite powder and NCs synthesis was developed by this author2, 3 independently in 2013 from already published work by Zhu et al.4 The system of

MAPbX3 perovskite was taken as already well-known material for model system in solvent- antisolvent crystallization technique. Figure 5.1a shows the pressed pellets from the precipitated perovskite powder. The gradual color change from MAPbCl3 to MAPbI3 indicates the difference in band gap and with this an absorption change due to the different halide ions obtained in the crystal structure5 which can by clearly seen in UV-Vis absorption measurements (Figure 5.1b).

- 98 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.1. a) Pallets pressed from dried perovskite powder. b) Absorption mixed halide perovskites powders. c) XRD. Powder X-Ray diffraction patterns where measured to investigate the crystal symmetry of the bulk material. Figure 5.1c show the full range XRD scan of all the MAPbX3 mixed halide perovskites. There are five characteristic peaks for MAPbI3 perovskite phase at around 14° corresponding to the miller indices (110) for iodine, at 20° for (112), at 25° for (202), at 6 28° for (220) and at 32° for the (310) symmetry. MAPbI3 exhibits a tetragonal symmetry, upon proceeding down to the MAPbBr3 phase a gradual change in the peak position can be observed for the mixed phases indicating a change in crystal symmetry from tetragonal to cubic state of

MAPbBr3 as well as MAPb(Br/Cl)3 and pure MAPbCl3. We further successfully applied solvent-antisolvent crystallization technique for a + + broad number of divalent metal with general formula ABX3 (A= NH2CHNH2 , CH3NH3 , B=Pb2+ Sn2+, Ge2+; X= Cl−, Br−, I−) which was described by Mitzi.5 To verify size and morphology, Scanning Electron Microscopy (SEM) measurements were employed. Figure 5.2a-f depicts SEM images for the perovskite powder with different cations and anions. Various shapes such as cubes (MAPbBr3, MASnBr3) wires (FAPbI3) or discs (MAMnCl3) were obtained in the size range of 0.1-20 µm depending on the solvent and antisolvent system (Table 2.6). For example, changing the solvent polarity from DMF to NMF

(Toluene as antisolvent) leads to changing the shape of MAPbBr3 crystalites from normal cubic to chamfer cubic (Figure 5.2c,d) which can be explained by Wulff facets theory.7 The shape of an equilibrium crystal is obtained by minimizing the total surface free energy associated to the crystal-medium interface, according to Gibb´s thermodynamic principle.8

- 99 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

When one of the substances involved is anisotropic, such as a crystal, the contribution to the total free energy of each element depends on its orientation. Wulff states that: “The minimum surface energy for a given volume of a polyhedron will be achieved if the distances of its faces from one given point are proportional to their capillary constants”.9 This material offers exemplary depiction of anisotropic behaviour for specific crystallographic directions having especial low surface energies and thus resulting in chamfered cubic symmetry. Therefore combination of the different applicable solvent-antisolvent system can yield multiple morphologies for the perovskite system with identical chemical composition.

Figure 5.2. a-f) SEM overview of various perovskite powders with different morphology.

- 100 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

5.3. Perovskite ink formulation

To investigate the possibility of ink formulation from crystalline perovskite particles,

MAPbBr3 and MAPbI3 were chosen as material which already was shown the high 10, 11 performance perovskite solar cells. For MAPbBr3 crystallites two shape variants were used

cubic (MAPbBr3 (DMF)T) and chamfer cubic MAPbBr3 (NMF)T with average size of 0.53± 0.25 µm and 11.92±3.89 µm respectively (Figure 5.2c,d). The size of as synthesized

MAPbI3 crystallites (MAPbI3 (GBL)C) (Figure 5.2a) was 0.3±0.12 µm. The abbreviation (DMF)T, (NMF)T and (GBL)C describe the solvent- antisolvent pair used in powder synthesis (DMF - N,N-Dimethylformamide; NMF - N-Mmethylformamide; GBL - γ-Butyrolactone; T - Toluene; C - Chloroform) First of all, a suitable solvent, functioning as the fundament of the ink, had to be provided. Several solvents including xylene, α-terpineol and toluene were tested. α-Terpineol is broadly use solvent in inject-printing technologies. Multiple stable ink 12 13 14 15 16 formulations based on silver nanoparticles, graphene, SiO2, MoS2, Ni nanoparticles and many more were developed and successfully applied for printing. The universality of the α-terpineol motivated us to utilize this dispersant for our perovskite powder as well. However, despite the low polarity α-terpineol (dielectric constant ε=2.8) strongly protonates the perovskite material that leads to material degradation. Nonetheless, inks based

on MAPbBr3 crystallites shown good chemical stability in time, while MAPbI3 particles in

α-terpineol degraded after 24h. Even mixd-halide composition (MAPbBrxI3-x) and additives/binders involving such as ethyl cellulose could not significantly avoid the degradation process of iodide based perovskite, which is more attractive for solar cell fabrication. Therefore, toluene was chosen as a “safe” solvent due to its low polarity (dielectric constant ε=2.38), absence of protonating groups and relatively high boiling point (111°C) which is favourable for printing process. Furthermore, toluene was previously used in our group as an efficient antisolvent for perovskite thin-film crystallisation17 as well as in perovskite NCs synthesis.2, 3 Since our perovskite particles are ligand-free and on the surface no surfactant molecules which can help to disperse them in the toluene, additives/binders were used to increase viscosity (EC - ethyl cellulose). Details of inks preparation are described in Section 2.3 (Chapter 2).

Several inks with different weight (w) concentrations of MAPbBr3(DMF)T and

MAPbI3 (GBL)C perovskite powder were tested, including: 10, 15, 20, 25, 30 and 35 %(w). After a few drop tests it became apparent that 30%(w) proved best ink formulation of all

- 101 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

samples. Furthermore an increase in the EC ratio was beneficial as well, and thus three different inks were fabricated: 30:5 - featuring 30%(w) of perovskite solute and 5%(w) EC in the solvent; 30:7.5 - incorporating 30%(w) of perovskite and 7.5%(w) of EC; 30:10 – where 30% (w) of perovskite powder and 10%(w) of EC were mixed in toluene. The Figure 5.3 shows a long term stability observation of the three inks in a 30:7.5 concentration.

The samples MAPbI3(GBL)C and MAPbBr3(DMF)T show stability up to almost two weeks, whereas MAPbBr3(NMF)T (right orange sample) sediments within the first 2 days.

The high sedimentation rate of MAPbBr3(NMF)T can be ascribed to the ~20x larger size of the particles. Moreover, there is no visible sign of chemical decomposition for any of the three samples in comparison to α-terpineol based inks. The inks proved high stabilityagainst decomposition up to two weeks after production, which could be later confirmed by XRD measurements. The occurred sedimentation presented no impediment, since the samples could be stirred up and thus be reused. This is comparable to reports of long term stability of silver nanowires, which can be stored up to one month in an aqueous solution without any signs of decomposition.18

Figure 5.3. Rheological stability in time of inks based on MAPbB3(DMF)T, MAPbI3(GBL)C and MAPbBr3(NMF)T employing a 30:7.5 concentration ratio.

- 102 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Film processing and characterization. The processed inks were applied via doctor blading onto glass substrates. The manifoldness and simplicity of this procedure allows industrial large scale applications as well. The Figure 5.4 shows that a variety of substrates in different sizes can be coated with the above prepared inks without any problems using simple blading techniques. Since the produced perovskite inks, do fulfil the desired requirements, such as long- term stability, mediocre viscosity and homogenous dispersion, further optical and elrctrical investigations were carried out of the corresponding thin films.

Figure 5.4. Different sizes of glass substrates, coated with MAPbI3(GBL)C and MAPbBr3(DMF)T inks.

Hence, three inks for each perovskite sample incorporating various additive/binder ratios were produced and coated onto glass substrates (1.25 x 1.25 and 10x10 cm) via doctor blading. Afterwards, the thickness of the measured samples was investigated by profilometer of type Kla Tencor D100. Figure 5.5a,c,e shows the influence of increasing viscosity and increased blading gaps onto the produced film thicknesses.

- 103 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.5. Dependence of the film thickness over blade gap, with varying viscosities for (a) MAPbBr3(NMF)T, (b) MAPbBr3(DMF)T and (c) MAPbI3(GBL)C. Each Material features the following three ink ratios: 30:5 - 30%(w) perovskite and 5% (w) EC in toluene, 30:7.5 - 30%(w) perovskite & 7.5% (w) in toluene; 30:10 - 30%(w) perovskite & 10%(w) EC in toluene. b,d,f) Confocal microscopy results for the 30:5 series of all three materials at a constant blading gap of 80 µm.

- 104 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Three inks with differing viscosities due to differing EC content were compared. When applying a 30:5 ink, featuring 30%(w) of perovskite and 5%(w) of EC, it can be observed, that only a small increase with rising blade gap can be achieved. For a blade gap of 1000 µm only a maximum thickness of 2.3 µm can be attained. When increasing the amount of EC in the ink, a steep increase in film thickness was recorded. The starting values of 30:5 and 30:7.5 differ greatly, whereas the values for 30:7.5 and 30:10 are quite similar at small blading gaps. While 30:7.5 shows a probable saturation at a blading gap of 200 µm and a resulting thickness of 13.1 µm, the 30:10 ink offers a continuous increase up to a value of 20.5 µm. With increasing amount of EC, film thickness can be increased while keeping the blading parameters.

MAPbBr3 (DMF)T inks show a similar behaviour to MAPbBr3(NMF)T based inks. With higher EC ratio and thus increasing viscosity higher values for the film thickness are recorded. However, in this example it appears that for the 30:5 series the maximum thickness values exceed the (NMF)T maximum point of 2.8 µm by far. An exponential increase for the film thickness is shown for the 30:5 series and a maximum film thickness of 12.8 µm is observed for a blading gap of 200 µm. The 30:7.5 series features a rapid increase of film thickness, followed by an approximate saturation and an anew rise in thickness. The 30:10 inks show film thickness incensement followed by saturation at a blade gap of about 160 µm. The maximum thickness values for the 30:10 series are in the range of 16 – 17 µm. We conclude that with increasing amount of EC, the density of the material in the ink increases thus enhancing viscosity, resulting in an increase in film thickness at constant blading parameters. The roughness of the doctor bladed perovskite films was investigated by confocal microscopy. This optical imaging technique enables the reconstruction of three-dimensional structures by collecting a set of pictures at different depths. The roughness values are presented in Sq, which represents the root mean square value of the ordinate value within the defined area. Hence it is equivalent to the standard deviation of heights. The investigated area of the samples was 1.5 x 1.6 mm². The surface roughness of the films of each perovskite material for a constant blading gap of 80 µm and the three different inks were examined (Figure 5.5b,d,f). The results of the surface roughness investigation for each sample with different composition are collected in Table 5.1.

- 105 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Table 5.1. Square root mean values for surface roughness of films of the three perovskite materials for varying ink consistencies at a constant blading gap of 80 µm.

Powder to EC ratio 30:5 30:7.5 30:10

Sample

MAPbBr3(NMF)T 0.83µm 2.43µm 3.45µm

MAPbBr3(DMF)T 2.46µm 1.99µm 3.21µm

MAPbI3(GBL)C 1.33µm 1.15µm 1.24µm

Not surprisingly, MAPbBr3(NMF)T is featuring the largest average crystal size of these three materials, shows a rougher surface with increasing thickness and EC amount..

MAPbBr3(DMF)T shows no specific tendency towards an increase or decrease in roughness depending on film thickness or ink compositions. It might be argued, that the cubic crystal structure does not align perfectly on the substrate and thus creates inhomogeneous layers with

high roughness. MAPbI3(GBL)C on the other hand offers a more constant value, regardless film thickness and ink concentration. It also offers the lowest value for the 30:7.5 and 30:10 series. The lower roughness for MAPbI3(GBL)C films is consistent with their smaller crystallites size (0.3±0.12 µm) which forms less local particles agglomerates upon blading. Figure 5.6a-c represent absorption and PL spectra of bladed films from

MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C inks. PL spectra of pristine powder shown as dashed line. The PL curves demonstrate peak intensity at a wavelength of 554 nm, 548 nm and ~780 nm respectively. Nevertheless, it is apparent that the film offers a narrower FWHM than the recorded PL of Powder. The difference in the two PL regimes is attributed to possible surface defects and crystal orientation of the powder. When applying the ink to the substrate, the perovskite particles will be more orientated on the surface along the doctor blading direction and therefore the PL intensity is expected to be higher as well due to the less scattering from the surface. The UV-Vis and PL measurements give a first impression how the perovskite properties are changed when transformed to their films.

- 106 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.6. Absorption and PL spectra of thin films bladed out of (a) MAPbBr3(NMF)T, (b) MAPbBr3(DMF)T and (c) MAPbI3(GBL)C inks. Dashed lines corresponds to PL of pristine powder samples. SEM pictures of films bladed from the three different inks (blade gap 80µm) are depicted in Figure 5.7a-c, right side at different magnification. The top view with the highest magnification shows minor particle damage after powders grinding during ink preparation.

For MAPbBr3(NMF)T film (Figure 5.7b, right side) the 250-fold magnification shows a surface which appears to be bristled with several holes or strong surface irregularities, while

MAPbBr3(DMF)T (Figure 5.7a, right side) and MAPbI3(GBL)C (Figure 5.7c, right side) films are pin-hole free and more smooth even at the highest magnification. Roughness measurements with confocal microscopy support these results.

The XRD analysis for MAPbBr3(NMF)T and MAPbBr3(DMF)T films shown no differences between pristine powder and before and after aging an ink for 2 weeks.

(Figure 5.7a,b left side). In case of MAPbI3(GBL)C films, there is no trace of free PbI2 in the XRD of the pristine pervoskite powder, while in a fresh prepared film sample the appearance

of PbI2 is indicating partial ink decomposition in ambient condition upon preparation.

After for two weeks of aging the ink does shows even more intense PbI2 peaks indicating further decomposition of the perovskite particles (Figure 5.7c, left side).

- 107 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.7. a,b,c, right) SEM study of coated film (blade gap 80 µm) out of MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C inks with in different magnification; a,b,c, left) XRD patterns for powder, film coated from fresh prepared and two weeks aged ink.

- 108 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Photodetector application. The fabricated inks were applied to special ITO- finger substrates via doctor blading. Beneath is a schematic blue print of the layout of such a substrate. The used substrates show finger structures made of indium tin oxide (ITO) with the following scheme (Figure 5.8): demonstrated dark-square are 23 fingers, which exhibit a length of 2.93 mm and a width of 0.07 mm. Hence, one contact measures a total of 137.9 mm².

Figure 5.8. Sketch of the ITO- substrate

The application of the perovskite material onto the above shown substrate enables measurements of a photoelectric response by comparing the dark current to the measured current, when the substrate is illuminated by a solar simulator with one sun intensity. The absorbed light creates free charge carriers which are increasing current density. The photocurrent is given by: 푞퐺 퐼 = µ휏퐸, (6.1) 퐿 where the µ휏 product is the most important parameter for a photo-conductor, here µ represents the mobility and τ stands for the lifetime of the free charge carriers which is also referred the “figure of merit” for a photo-conductor.19 The measurements were conducted as follows: dark current measurements were performed without any external source of light, and only the current in the range of -5 V – 5 V was recorded. Subsequently, the samples were exposed to the full range of AM 1.5 of the solar simulator and another current measurement in the same voltage range was completed. With the comparison of dark and illuminated current densities curves, it is possible to express the potential, held by the perovskite material, for optoelectronic applications and an indication for its efficiencies to harvest the power of the range of the sun. The lifetime τ over the transit time 푇푅is called the gain. A good photoconductor employs a gain of 1000 or more. In the following will be a short overview over the most effective devices that were tested.19

- 109 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.9: a, d, g) Photo of produced MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C photodetectors; b, e, h) Corresponding microscopic view on ITO-substrate and (c, f, i) dark and light photocurrent over the voltage.

- 110 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Figure 5.9a,d,g shows a photo of the produced films on ITO finger-substrate from

MAPbBr3(NMF)T, MAPbBr3(DMF)T and MAPbI3(GBL)C inks. The presented photodetector prototypes are the best performing devices studied in this chapter. The film thicknesses for each film was measured as 4.5µm, 5.5 µm and 5.2 µm respectively. Figure 5.9b,e,h shows a microscopic image of the coated films. It can be seen that the surface is almost homogenously

covered, but MAPbBr3(NMF)T film offering some spots without particle coverage due to larger size particles. In Figure 5.9c,f,i the dark and photoconductive current are calculated by the ratio of the current when illuminated and the dark current and offers a values of 7925, 12995 and 326 which is very high as for simple produced photodetector.

Further electrical characterisation of the produced photodetectors based on

MAPbBr3(DMF)T and MAPbI3(GBL)C inks was performed. The responsivity and specific detectivity for two devices were investigated (see Figure 5.10).

Figure 5.10. Reponsivity and Specific Detectivity of MAPbBr3(DMF)T (left) and MAPbI3(GBL)C (right) based photodetector over wavelength.

The responsivity for both a device based on MAPbBr3(DMF)T and MAPbI3(GBL)C represent a spectrally narrow regime. The responsivity and specific detectivity display a narrow

FHWM, which shows that the devices operate in a narrowband regime. The MAPbBr3(DMF)T device displays a peak of 550 nm, similar to previous devices. The iodide based device exhibits peak intensity at 780 nm. It is worth mentioning that the overall photocurrent of narrowband photodetectors decreases with the increase of the perovskite film thickness, which is a result of the suppressed absorption at low wavelength and larger resistance associated with thicker films.20

- 111 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

Perovskite film healing. Post-treatment methods of perovskite thin films via MA gas introduction were based on the research of Zhou et al.21 Recrystallization was performed by exposing a bladed film to methylamine gas, followed by applying heat (80 °C) to the sample for a minute (Figure 5.11a). The recrystallization process can be observed by eye.

When exposed to the gas, the black film of MAPBI3(GBL)C transforms into a yellow wet film, and then almost turns transparent. When exposed to heat, the reappearance of the black film is immediate, indicating exclusion of excessive MA. The surface however offers now a shiny grey colour and smooth appearance.

Figure 5.11. a) Schematic representation of the recrystallization process. b) SEM pictures of a MAPbI3(GBL)C film before and after MA gas treatment at three different magnifications. c) XRD patterns of a film of MAPbI3(GBL)C before and after MA gas treatment d) I-V measurements of a MAPbI3(GBL)C photodetector before and after MA gas treatment. Morphology of the film before and after MA treatement was studied by SEM measurements (Figure 5.11b). The SEM pictures show that after exposing the film to methylamine gas, the particles appearance clearly indicate recrystallization process like after melting. Hovewer, the islands formed after recrystalizations are not monocrystaline, but more like dense packed small crystallites. By repeating the post treatment process, we believe it is

possible to create almost quasi single crystalline layer of MAPbI3(GBL)C.

- 112 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

To determine any structural changes after MA treatment, XRD measurements were conducted

(Figure 5.11c). As earlier described, iodide-based perovskite films already show free PbI2 peaks in XRD coated from fresh and two weeks aged inks (Figure 5.7c, left side). After MA gas treatment, peaks indicating free lead iodide do not appear, suggesting that these were converted with the help of the methylamine and better balancing the PbI2:MA ratio equal. Moreover, the intensity of other diffraction peaks became higher after recrystallization resulting in higher crystallinity as well as film quality. As a consequence this approach could improve photoconductive behaviour tremendously in terms of better crystal quality. The I-V measurement displayed in Figure 5.11b reflect photoconductive behaviour of the same device, once before and after MA gas treatment. The dashed lines represent the illuminated and dark curves of the device before treatment. The potential of this device amounts to 104, which is not good enough reprehensive value for a photo detector. Post treatment this value increased up to 1298 which is 12 fold higher than pristine one. This is a tremendous increase, presenting huge potential for film and device improvement.

- 113 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

5.4. Summary

In summary, we demonstratet the successful processing of perovskite ink, created from perovskite powder and further employment for device application. Organic-inorganic halide perovskite powder was produced via solvent-antisolvent extraction and processed to an ink, by introducing it to a suitable solvent and adjusted ratios of additives. The perovskite ink offers the opportunity for optoelectronic device fabrication via straightforward coating techniques and has huge potential for up-scalable and industrial usage. The ink fabrication underwent several long term stability investigations, until a satisfactory stable product was achieved. The obtained ink was applied to various kinds of substrates via doctor blading and a film control for a thickness range of 500nm to 21µm was achieved. The coated substrates were investigated and characterized by the same means as the powder, allowing a direct comparison between the properties of the powder and a completed film. The electrical potential, held by the perovskite material, was investigated through the fabrication of photodetectors. These devices shown outstanding photoconductive response. Post treatment methods via MA gas introduction, showing great improvement of the film crystallinity, smoothness and I-V characteristics. The successful generalized synthesis of organic- inorganic metal halide perovskite inks open up new windows for a low cost and low temperature semiconductor perovskite film manufacturing, which can revolutionize the market for light harvesting devices

- 114 -

Perovskite ink :from nanocrystaline powder to device application Chapter 5

5.5. Bibliography

1. Ye, M.; Hong, X.; Zhang, F.; Liu, X., Recent advancements in perovskite solar cells: flexibility, stability and large scale. Journal of Materials Chemistry A 2016, 4, 6755-6771. 2. Levchuk, I.; Osvet, A.; Tang, X.; Brandl, M.; Perea, J. D.; Hoegl, F.; Matt, G. J.; Hock, R.; Batentschuk, M.; Brabec, C. J., Brightly Luminescent and Color-Tunable Formamidinium Lead Halide Perovskite FAPbX3 (X=Cl, Br, I) Colloidal Nanocrystals. Nano Letters 2017. 3. Levchuk, I.; Herre, P.; Brandl, M.; Osvet, A.; Hock, R.; Peukert, W.; Schweizer, P.; Spiecker, E.; Batentschuk, M.; Brabec, C. J., Ligand-assisted thickness tailoring of highly luminescent colloidal CH3NH3PbX3 (X = Br and I) perovskite nanoplatelets. Chemical Communications 2017, 53, 244-247. 4. Zhu, F.; Men, L.; Guo, Y.; Zhu, Q.; Bhattacharjee, U.; Goodwin, P. M.; Petrich, J. W.; Smith, E. A.; Vela, J., Shape evolution and single particle luminescence of organometal halide perovskite nanocrystals. Acs Nano 2015, 9, 2948-2959. 5. Mitzi, D. B., Synthesis, structure, and properties of organic‐inorganic perovskites and related materials. Progress in Inorganic Chemistry, Volume 48 2007, 1-121. 6. Jang, D. M.; Park, K.; Kim, D. H.; Park, J.; Shojaei, F.; Kang, H. S.; Ahn, J.-P.; Lee, J. W.; Song, J. K., Reversible halide exchange reaction of organometal trihalide perovskite colloidal nanocrystals for full-range band gap tuning. Nano letters 2015, 15, 5191-5199. 7. Wulff, G., Zur frage der geschwindigkeit des wachsthums und der auflösung der krystallflächen. Zeitschrift für Kristallographie-Crystalline Materials 1901, 34, 449-530. 8. Cohen, B.-E.; Etgar, L., Parameters that control and influence the organo-metal halide perovskite crystallization and morphology. Frontiers of Optoelectronics 2016, 9, 44-52. 9. Miracle-Sole, S., Wulff shape of crystals. Scholarpedia 2013, 8, 31266. 10. Yang, M.; Zhang, T.; Schulz, P.; Li, Z.; Li, G.; Kim, D. H.; Guo, N.; Berry, J. J.; Zhu, K.; Zhao, Y., Facile fabrication of large-grain CH3NH3PbI3-xBrx films for high-efficiency solar cells via CH3NH3Br-selective Ostwald ripening. Nature Communications 2016, 7. 11. Mali, S. S.; Shim, C. S.; Hong, C. K., Highly stable and efficient solid-state solar cells based on methylammonium lead bromide (CH3NH3PbBr3) perovskite quantum dots. NPG Asia Materials 2015, 7, e208. 12. Ma, R.; Suh, D.; Kim, J.; Chung, J.; Baik, S., A drastic reduction in silver concentration of metallic ink by the use of single-walled carbon nanotubes decorated with silver nanoparticles. Journal of Materials Chemistry 2011, 21, 7070-7073. 13. Gao, Y.; Shi, W.; Wang, W.; Leng, Y.; Zhao, Y., Inkjet printing patterns of highly conductive pristine graphene on flexible substrates. Industrial & Engineering Chemistry Research 2014, 53, 16777-16784. 14. Currle, U.; Wassmer, M.; Krueger, K. Particle Inks for Inkjet Printing of Electronic Components. In NIP & Digital Fabrication Conference, 2008; Society for Imaging Science and Technology: 2008; Vol. 2008; pp 702-706. 15. Li, J.; Naiini, M. M.; Vaziri, S.; Lemme, M. C.; Östling, M., Inkjet Printing of MoS2. Advanced Functional Materials 2014, 24, 6524-6531. 16. Tseng, W. J.; Chen, C.-N., Dispersion and rheology of nickel nanoparticle inks. Journal of materials science 2006, 41, 1213-1219. 17. Quiroz, C. O. R.; Levchuk, I.; Bronnbauer, C.; Salvador, M.; Forberich, K.; Heumüller, T.; Hou, Y.; Schweizer, P.; Spiecker, E.; Brabec, C. J., Pushing efficiency limits for semitransparent perovskite solar cells. Journal of Materials Chemistry A 2015, 3, 24071-24081. 18. Hong, B. H.; Bae, S. C.; Lee, C.-W.; Jeong, S.; Kim, K. S., Ultrathin single-crystalline silver nanowire arrays formed in an ambient solution phase. Science 2001, 294, 348-351. 19. Bube, R. H., Photoconductivity of solids. RE Krieger Pub. Co.: 1978. 20. Saidaminov, M. I.; Haque, M.; Savoie, M.; Abdelhady, A. L.; Cho, N.; Dursun, I.; Buttner, U.; Alarousu, E.; Wu, T.; Bakr, O. M., Perovskite photodetectors operating in both narrowband and broadband regimes. Advanced Materials 2016, 28, 8144-8149. 21. Zhou, Z.; Wang, Z.; Zhou, Y.; Pang, S.; Wang, D.; Xu, H.; Liu, Z.; Padture, N. P.; Cui, G., Methylamine‐Gas‐Induced Defect‐Healing Behavior of CH3NH3PbI3 Thin Films for Perovskite Solar Cells. Angewandte Chemie International Edition 2015, 54, 9705-9709.

- 115 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Chapter VI

Purity control of the precursor materials for perovskite synthesis

This Chapter represent systematic study of the chemical purity of a perovskite precursors to

the morphology of thin film perovskite layers. NMR spectroscopy of CH3NH3I reveals the

chemical nature of the impurities. These impurities initiate PbHPO3 nanoparticles (NPs) formation in the perovskite precursor solution, with further seed growth during thin film

formation. The side products that form during CH3NH3I synthesis in the presence of hypophosphorous acid as a stabilizer for HI, namely methylammonium hypophosphite

(CH3NH3H2PO2) and methylammonium phosphite (CH3NH3H2PO3), were independently synthesized and uniquely identified as effective anion source for PbHPO3 NPs. As consequence, increased perovskite grain size, the reduced defect density, and the enhanced intrinsic perovskite stability of the thin films is achieved by using perovskite precursor solution that contains PbHPO3 NPs. Finally, by using a simple planar inverted (p-i-n) structure, the performance of perovskite devices is shown so very from 12.2% to 17.2% as a function of impurity presence. The study presented here was supported from the Joint Project Helmholtz-Institute Erlangen Nürnberg (HI-ERN) under the project number DBF01253, DFG supported training group 1896 ‘‘In situ microscopy with electrons, X-rays and scanning probes’’ and Electron microscopy resources which was kindly provided by the Center for Nanoanalysis and Electron Microscopy (CENEM).

 Parts of this chapter have been adapted or reproduced with permission from: I. Levchuk, Y. Hou, M. Gruber, M. Brandl, P. Herre, X. Tang, F. Hoegl, M. Batentschuk, A. Osvet, R. Hock, W. Peukert, R. R. Tykwinski and C. J. Brabec, Advanced Materials Interfaces, 2016, 3, 1600593

- 116 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

6.1. Motivation and State of the art

Organometallic halide perovskites CH3NH3PbX3 (X = I, Br, Cl) have quickly become promising semiconductors for solar cell applications, reaching power conversion efficiencies up to 20%1,2,3. Development of the technology of perovskite-based solar cells requires precise control purity of the raw material and full understanding of the internal chemical process that occur during precursor formation. A precise stoichiometric ratio of methylammonium iodide

(CH3NH3I) and lead (II) iodide (PbI2 ) plays an important role in the final perovskite film formation, thermal stability, morphology, and final solar cell performance4, 5,6. Non- stoichiometric ratios with extra excess PbI2 favor formation of larger perovskite domains that provide higher performance and stability of the final device3. In additional to stoichiometry, “quasi” colloidal particles7 or reducing agent additives such as hypophosphorous acid4 (reduce – the oxidized I2 back into I ) were also shown to enhance optoelectronic quality of perovskite thin films.

It is well known, that CH3NH3 and HI are raw materials for the synthesis of CH3NH3I. Commercially available hydroiodic acid usually contains 1.5-2% hypophosphorous acid as stabilizer and is suspected to provoke side product during CH3NH3I formation. 8 Deng et al. claimed that purification of CH3NH3I is not necessary for spin-coated perovskite solar cells. However, it played a crucial role for doctor-bladed fabrication, because the perovskite precursor based on non-purified CH3NH3I formed rough films and consequently led 8 to poor or non-functional devices. Deng further hypothesized the formation of CH3NH3H2PO2 as a side product of the reaction between CH3NH3 and H3PO2 during CH3NH3I synthesis, which favor the formation of insoluble Pb(H2PO2)2 NPs during precursor preparation. It was assumed that such a kind of NPs could work as nucleation centers during perovskite thin film crystallization and provide larger domain formation. This approach based on precursors that contain NPs as seeds for bigger grain growth was already used for CZTS thin film growth9. In 3 the case of perovskite solar cells, Bi et al. claimed that excess of the PbI2 in the perovskite precursor solution forms nano-sized crystallites within the mesoporous TiO2 layer, with further bigger perovskite domain growth. Such devices demonstrate higher efficiency compared to the ones produced from stoichiometric precursor, because films with larger perovskite domains display better crystal quality due to reduced area of grain boundaries and therefore less defects. Li et al.10 have also demonstrated that direct incorporation of the PbS nanoparticles with the size of 5 nm creates nucleation sites for larger perovskite crystal domain. This data inspired us

- 117 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

for a deeper study of the perovskite precursor system to ultimately understand the nature of the impurities and their impact on perovskite film formation. This Chapter present an in-depth chemical study of the possible impurities formed

during CH3NH3I preparation and reveal their relevance on solar cell processing. A primary consideration is the chemical transformation of hypophosphorous acid, which plays the role of the stabilizer for HI

6.2. Precursor and starting materials analysis

Pure methylammonium iodide (c-MAI) was synthesized in our lab by mixing HI and

CH3NH2 in ethanol at room temperature. Compared to the standard diethyl ether/ethanol purification process, which is commonly used,1,2, 4, 7 we recrystallized our material three times from hot solutions of ethanol . Finally, we obtained fine and big flakes of highly crystalline c-MAI that was used as a standard high purity material. The synthesis of standard methylammonium iodide, which contains impurities (i-MAI) as discussed later, is described elsewhere.1,2, 4, 7 Commercially available MAI with various concentrations of impurities were employed for reference device fabrication. After perovskite precursor preparation by mixing

PbI2 and different MAI in DMF/DMSO at 60 °C, the impurity containing i-MAI based solution turned cloudy after cooling down to room temperature. In case of c-MAI, the solution remained completely transparent for a long time. For analysis of the particle suspensions formed in the i-MAI based perovskite precursor, we centrifuged the solution, and analyzed the precipitate by scanning electron microscopy (SEM), and x-ray diffraction (XRD) analysis. As shown in SEM Figure 6.3a, the obtained precipitate from i-MAI consists of nanoparticles (NPs) with an average size of 30±12 nm. Such a large size distribution is a result of the uncontrolled and surfactant-free NP formation, as compared to the standard colloidal synthesis of NPs with high monodispersity (e.g. CdSe quantum dots11). In analogy to the 8 previous work by Deng et al. , we expected the formation of Pb(H2PO2)2 NPs. However our XRD analysis found another type of phosphorene salt. Figure 6.3b demonstrates that the

diffraction pattern of the NPs is perfectly fit with the PbHPO3 phase of XRD data taken from ICDD database (PDF №. 00-020-0580). For comparison, we separately synthesized

Pb(H2PO2)2 by mixing lead acetate (Pb(CH2COO)2) with hypophosphorous acid (H3PO2). The

measured XRD data are completely different from the PbHPO3 XRD pattern (Figure 6.1a-b).

- 118 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Figure 6.1. a) XRD diffraction patternof the Pb(H2PO2)2 . Reference diffraction data was taken from Kuratieva et al. 12 . b) XRD diffraction pattern of the precipitate is in excelent agreement with the diffraction data for PbHPO3 taken from ICSD database (card №. 00-020-0580).

This unexpected result motivated us to investigate and understand more deeply the chemical processes occurring during the i-MAI based perovskite precursor solution. First, we analyzed both c-MAI and i-MAI by 1H NMR spectroscopy in order to detect any proton-containing impurities within the materials. The 1H NMR spectrum of c-MAI

(Figure 6.2c) shows signals for the CH3 group at 2.36 ppm and for the NH3 group at 7.48 ppm, respectively and 31P NMR spectrum confirm that c-MAI is phosphorus-free. In the case of i-MAI, we observe four additional sharp signals in the 1H NMR spectrum located at 7.81, 7.63, 6.22, and 5.71 ppm. The same peaks are observed in commercially available precursors as well (Figure 6.2a-b).

Figure 6.2. Comparison of 1H NMR spectra of NH3 protons signal for purified, non-purified (a) and commercially available MAI from different providers (b). All of them contain MAH2PO2 and MAH2PO3 in varying amounts, except purified MAI.

- 119 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

The NH3 proton signal of i-MAI at 7.49 ppm is broader than the sharp NH3 signals in c-MAI (Figure 6.3d, inset blue graph). The 31P NMR measurements for i-MAI further establish the existence of phosphorus-containing contaminants within the i-MAI material. Additional XRD measurements of the с-MAI and i-MAI did not show any difference in their structural properties (Figure 6.3e). Both of them are in the α phase at room temperature and have a tetragonal structure with space group P4/nmm13. However, after re-measuring i-MAI some of the higher angle reflexes show lower intensity or are missing entirely as compare to с-MAI. The digital photo of the sample holder with the i-MAI powder (Figure 6.3e, inset picture) proofs that the material was radiation damaged during the measurements. The brown color of the powder after X-ray radiation may indicate phosphoric acid (H3PO4) formation. Pure phosphoric acid is a colorless compound, turning brown in media with organic compounds (in our case methylamine group) .14 Differential scanning calorimetry (DSC) was utilized to compare melting, crystallization and sublimation points between c-MAI and i-MAI (Figure 6.3f ) and proves the presence of impurities. c-MAI, undergoes the phase transition from the room temperature stable α phase to the high temperature ε phase at 141 °C. At 234 °C the sublimation process of the solid phase starts. This data is in excellent agreement with DTA and DSC reported in literature15, 16. In case of i-MAI, an unknown endothermic process occurs at 65 °C that may correspond with the presence of impurities. Phase transitions from α to ε occur at temperature as low as 111 °C and with a broader shoulder which indicates the presence of extraneous components compared to c-MAI. The sublimation of i-MAI occurs at the same temperature (234 °C) as c-MAI. Crystallization of both c-MAI and i-MAI may occur in a multistep manner within the range between 20–150 °C, where the crystallization of i-MAI is 4 °C lower than for c-MAI. An additional exothermic peak around 75 °C was observed during i-MAI crystallization, that again may correspond to the presence of an impurity. Since crystallization temperature is higher than the melting point, we assume that the impurities partly decompose during heating. This effect is further explained by NMR spectroscopy (vide infra).

- 120 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Figure 6.3. Compositional investigation of the precipitate and impurities nature. a) SEM image of the white precipitate from perovskite precursor solution precursor based on i-MAI. b) XRD diffraction pattern of this precipitate shows agreement with diffraction data for 1 31 PbHPO3 taken from ICDD database (PDF №. 00-020-0580). c) H and P NMR spectra for 1 31 с-MAI. The inset shows the NH3 (blue) and CH3 (cyan) signal. d) H and P NMR spectra of i-MAI. The inset shows the broadening of the NH3 (blue) signal, compared to pure с-MAI, with additional peaks stemming from contamination. The CH3 (cyan) signal also becomes broader compared to с-MAI. e) XRD diffraction pattern of the c-MAI and i-MAI. Inset photo of the holder for XRD after i-MAI measurements. The samples seem to be radiation damaged during the measurements, which is clearly shown on XRD diffractogram (red curve) - some of the higher angle reflexes are weaker or missing entirely. f) DSC measurement of the c-MAI and i- MAI. In the case of i-MAI, all endothermic and exothermic peaks (melting, crystallization and sublimation) is shifted to lower temperature due to the presence of impurities.

- 121 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Description of chemical processes involved in impurity formation. To determine and understand the nature of the impurities, we next describe possible side reactions and product

formation occurring during MAI synthesis in the presence of H3PO2, which is used as a

stabilizer for HI. Hypophosphorous acid (H3PO2) and its salts are strong reducing agents and therefore can be oxidized to phosphorous acids or finally to phosphoric acids and their salts. A similar reaction proceeds during reduction of HI or I2:

3H3PO2 + HI –> PH4I + 2H3PO3

H3PO2 + I2 + H2O –> H3PO3 + 2HI (as a result of the stabilizer in HI solution)

In both cases H3PO3 is formed. In the presence of H2O (from HI solution) hydrolysis of PH4I occurs:

PH4I –> PH3 + HI

Next, H3PO3 and traces of non-oxidized H3PO2 can react with CH3NH2:

H3PO3 + CH3NH2 –> CH3NH3H2PO3

H3PO2 + CH3NH2 –> CH3NH3H2PO2

In perovskite precursor media, oxidation of CH3NH3H2PO2 to CH3NH3H2PO3 can occur as well.

The absence of Pb(H2PO2)2 was confirmed by XRD measurements of the obtained precipitate. Then, CH3NH3H2PO3 reacts with PbI2 in DMF:

PbI2 + CH3NH3H2PO3 –> PbHPO3↓ + CH3NH3I +HI Underlining the results and investigations described above, we decided the separately mix c-MAI with H3PO2 as well as with all other possible impurities that can be formed during the preparation of i-MAI (H3PO2, CH3NH3H2PO2 and CH3NH3H2PO3) and compare the corresponding 1H NMR spectra with that of i-MAI.

NMR studies. Since H3PO3 and H3PO2 are commercially available acids,

CH3NH3H2PO2 (MAH2PO2) and CH3NH3H2PO3 (MAH2PO3) could be easily synthesized within our laboratories. The pure MAH2PO2 was obtained as a highly viscous liquid, whereas

MAH2PO3 was obtained as a white crystalline powder. DSC measurements for MAH2PO3 showed a melting point around 65 °C. Further sharp peaks indicated the start of decomposition17 at around 125 °C (Figure 6.5), which documents the previously unknown

endothermic process observed in i-MAI. We found that MAH2PO3 showed very low solubility

in DMF (max. 1 mg/mL). In contrast, MAH2PO2 was well miscible in DMF. Interestingly, the

solubility of MAH2PO3 was significantly enhanced in the presence of MAH2PO2.

- 122 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

- 123 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Figure 6.5. DSC study of MAH2PO3. Melting start around 60°C with further decomposition of the materials at higher temperature.

The 1H NMR spectra of the pure additives as well as their mixed composites in c-MAI

are compared to that of i-MAI (Figure 6.4a). As demonstrated in Figure 6.4a, pure H3PO3 and 1 H3PO2 acids do not explain the H NMR spectrum of i-MAI. In both cases, the influence of the acidic protons was reflected in the chemical shift and the shape of the NH3 proton signal at 7.48 ppm. The proton signals of the impurities at 7.81, 7.63, 6.21, and 5.71 ppm, do not correspond to the proton peaks of the acids. A completely different picture is seen when c-MAI is mixed with MAH2PO2 and MAH2PO3. Blends of c-MAI + MAH2PO2 and c-MAI + MAH2PO3 give respective good fits with a single pair of impurities: 7.81 and 6.21 ppm corresponding to c-MAI + MAH2PO2 and 7.63 and 5.71 ppm for c-MAI+MAH2PO3, respectively. Most interestingly, pristine MAH2PO2 and MAH2PO3 fail to describe the impurities observed in i-MAI. This most likely originates from hydrogen bonding between 18 c-MAI and MAH2PO2 or MAH2PO3 . Randomly mixed c-MAI+MAH2PO2+MAH2PO3 perfectly describe all 1H NMR features of i-MAI (Figure 6.4b).

The NH3 proton signals for c-MAI + MAH2PO2 (Figure 6.4c) are shifted to higher

frequency and remain sharp, in contrast to c-MAI + MAH2PO3 in which the peak position is not shifted but is much broader, compared to pure c-MAI. This behavior may occur due to different hydrogen bonding mechanisms between MAH2PO2 or MAH2PO3 and c-MAI, according to different geometrical structures18 (Figure 6.4c and 6.4d, inset sketch). The impact

of different hydrogen bonding mechanisms on NH3 group protons in c-MAI and their effect on signal peak position is shown in Figure 6.4c and 6.4d. After adding a quantitative amount of

MAH2PO2 to c-MAI with concentration up to 0.1 wt.%, we observed a gradual peak shift to higher frequency, but no change in full width at half maximum (FHWM). The opposite

- 124 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

behavior was obtained in case of adding a quantitative amount of MAH2PO3 to c-MAI. These results are displayed in Figure 2e and 2f for MAH2PO2 and MAH2PO3, respectively. Please note, that both graphs have the same range on y-axis to better visualize the different coupling effects. All in all, 1H NMR spectroscopy documents the dual compositional nature of the impurities in i-MAI. Non-colloidal nature of the perovskite precursor. Recently it was reported that perovskite precursor solutions contain colloidal dispersions of soft lead polyhalide frameworks 7 [Pb-I]n, although these species were not extracted or chemically charactarezed . Since organolead halide perovskite precursors have an ionic nature19 that contain polyiodide 2− 3− 4− − complexes ([PbI4] , [PbI5] or [PbI6] ) in the surrounding of DMF desolvated CH3NH3 ions, we would not expect polyhalide frameworks to dominate the nature of the solution.

Figure 6.6. Tyndall effect in DMF perovskite precursor. a) Green laser light passed through the bottles under background illumination and in the dark. From left to right: pure DMF, perovskite precursor solution based on C-MAI and filtered through 450 nm filter, precursor solution based on i-MAI filtered through 200 nm filter; the same solution filtered through 450 nm filter.b) DMF solutions of the c-MAI (A) and i-MAI (B) illuminated with green laser.

No Tyndall effect is observed. c) The same solutions after Pb(CH3COO)2 addition, dissolved

in DMF. The Tyndall effect in solution (B) is observed due to due to PbHPO3 colloid formation. d) Tyndall effect of the filtered (450 nm filter) diluted c-MAI based perovskite precursor after adding one drop 1mg/ml DMF solution of MAH2PO3.

We used the Tyndall effect to detect the presence of any colloidal phase in the perovskite precursor solution. As shown in Figure 6.6a, green laser light doesn’t exhibit a Tyndall effect in pristine DMF as well as in the highly purified c-MAI solution (Figure 6.6a). However, perovskite solutions based on i-MAI display a significant Tyndall effect. Freshly

- 125 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

prepared suspensions like i-MAI based precursor solutions, were filtered with porous size of 200 nm and 450 nm respectively (Figure 6.6a). The Tyndall effect is significantly reduced upon filtering with a 200 nm porous size filter and shell reduced after 450 nm filtering.

We next verified the formation of PbHPO3 NPs by using a different Pb source namely lead acetate (Pb(CH2COO)2) instead of PbI2 and observed the identical trend. No NPs were

observed for c-MAI, while i-MAI again formed the PbHPO3 NPs (Figure 6.6b-c).

Furthermore, after adding one drop of DMF solution of MAH2PO3 to c-MAI based perovskite

precursor we observed PbHPO3 NPs formation confirmed by Tyndall effect (Figure 6.6d). Perovskite thin film and solar cell fabrication. Finally, it was most interesting to study

and understand the impact of PbHPO3 NP impurities on the device performance. Two perovskite precursor solutions based on c-MAI and i-MAI were prepared. Before spin coating, both precursor were passed through 450 nm PTFE filter. Due to the existence of numerous

PbHPO3 NPs, as nucleation centers in the i-MAI based solution, a dramatically different microstructure can be observed in SEM top view pictures (Figure 6.8a-b). The possible film growth route is depicted in Figure 6.7. The c-MAI based precursor without NPs gives a smaller average grain size of 123 ± 53 nm (Figure 6.8a) due to the absence of additional nucleation centers (PbHPO3 NPs) in the precursor. We observed that both of c-MAI and i-MAI based perovskite film degrade under a high energy electron beam (20 kV) due to ion migration within perovskites20,21.

Figure 6.7. Schematic interpretation of the perovskite thin film growth based on different precursor solution with and without PbHPO3 NPs. In strong contrast, the perovskite layers processed from i-MAI exhibit a much larger

grain size (234 ± 79 nm, Figure 6.8b) resulting from the PbHPO3 NPs as nucleation centers in the precursor. As one can see from the inset of Figure 6.8b , the i-MAI based perovskite film shows a very dense and large crystal arrangement that is beneficial for device performance as well as for stability.

- 126 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

In order to further investigate the local structures and optoelectronic properties, atomic force microscopy (AFM) and Kelvin probe force microscopy (KPFM) were combined to study the surface properties of the two different perovskite thin films. Perovskite films (300 nm thick)

were deposited on top of LT-NiO substrates (CH3NH3PbI3/LT-NiO/ITO/glass) facilitating as well as KPFM probing. The AFM topography of the two samples confirms the difference in grain sizes, which is in excellent agreement to previous SEM data. KPFM analysis (Figure 6.8e and 6.8h ) detects a potential difference between the grains and of the grain boundaries (GB), with a lower contact potential at the GB and a higher contact potential at grain bulk. One generally accepted explanation is the presence of more charged defects around the charged GBs. At the GBs, a larger density of defects like interstitials and vacancies is expected as compared to the bulk of the grain.22 The X-ray diffraction spectra (Figure 6.8i) show an identical tetragonal crystal structure for both precursors. However, the [110] peak of

i-MAPbI3 films is considerably increased, narrowed and display the higher crystallinity of i-

MAI based film. PbI2 impurities (12.6°) are absent in both samples.

Figure 6.8. Morphological and crystallographic analysis of perovskite thin film with c- MAI and i-MAI. Top view SEM images of the perovskite film prepared on top of the LT-NiO by using c-MAI (a) and i-MAI (b), respectively. The insets show a higher resolution picture captured at a high electron beam power of 20 kV and the average grain size distributions; AFM and KPFM measurements were performed on perovskite/LT-NiO/ITO/glass samples. Topographic images of perovskite with c-MAI (c) and perovskite i-MAI (f), phase-contrast imaging of perovskite with c-MAI (d) and perovskite i-MAI (g), potential images of the corresponding topography (e,h). The GB regions can be easily separated from all the above three images; (i), X-ray diffraction spectra of two perovskite thin films employing different MAI.

- 127 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

Figure 6.9. Photophysical properties of perovskite films and their solar cell performances. (a) steady-state PL spectra for the perovskite films; (b) time-resolved PL decay for the perovskite films; (c) I-V characteristics measured under 100 mW cm-2 AM1.5G illumination for the highest-performing devices employing c-MAI and i-MAI; (d) EQE spectrum of the highest-performing devices employing c-MAI and i-MAI.

Photophysical study of thin films. Steady-state PL and time-resolved PL decay studies investigated the electronic quality of the perovskite films prepared by the two different MAI precursors. The steady-state PL intensity of the perovskite film based on i-MAI is almost an order of magnitude larger than that of the c-MAI sample (Figure 6.9a). That enhancement is in agreement with the significantly enhanced charge-carrier lifetime found in time-resolved PL studies (Figure 6.9b).. The luminescence decay of the i-MAI sample is very close to a single exponential, with the intensity average lifetime t= 28 ns. In contrast, the nonexponential decay of the c-MAI sample with an average lifetime of 9.2 ns reflects strong quenching. Finally, we verified the solar cell performance of the two different MAI precursors in a simple planar

- 128 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

architecture (p-i-n) ITO/LT NiO/Perovskite/PCBM/PrCMA/Ag. As expected from SEM and PL studies, i-MAI based devices show superior performance compare to c-MAI. Under

standard AM1.5G illumination, Jsc, Voc, FF and PCEs were improved from 15.7 to 19.5 mA/cm2, 1.05 to 1.10 V, 74 to 80%, and 12.2 to 17.2%, respectively (Table 6.1).

Table 6.1.Summary of photovoltaic parameters of the investigated perovskite solar cells prepared by c-MAI and i-MAI based perovskite precursor solution. Precursor solution 2 V [V] FF[%] PCE[%] Jsc[mA/cm ] oc c-MAI 15.7 1.05 74 12.2 i-MAI 19.5 1.10 80 17.2

Figure 6.10. I-V curves of device employing perovksite with i-MAI (a) and c-MAI (b) measured under forward and reverse directions. c) I-V curves of device employing perovksite with i-MAI measured under different I-V scan rate.

In particular, the considerable improvement of Jsc, can be attributed to the enhanced charge carrier life time of i-MAI based perovskite thin film (Figure 6.9c). This is consistent

with the Jsc certified by integrating the external quantum efficiency (EQE) curves from Figure 6.9d , which deliver the values of 14.2 mA/cm2 and 19.2 mA/cm2, respectively. In addition, no strong anomalous hysteresis is observed for the two different scan directions and different scan rates (Figure 6.10a-c). Although there is no consensus on the origin of hysteresis, our work agrees well with the results of other groups, in which inverted structure perovskite employing PCBM are typically hysteresis free.23, 24

- 129 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

6.3. Summary In summary, the formation of impurities in MAI precursors in the presence of

hypophosphorous acid as a stabilizer for HI. MAH2PO2 and MAH2PO3 are uniquely identified as the dominant products from side reaction. 1H NMR spectroscopy reveals that these defects are only formed in MAI that has not been sufficiently purified. MAH2PO2 and MAH2PO3 are further shown to form PbHPO3 NPs as a precipitate in DMF. We could demonstrate that even tiny amounts of these impurities lead to the colloidal formation in the precursor solution. Most interestingly, PbHPO3 plays an essential role as nucleation centers, effective by controlling the perovskite grain size and the defect density as well as intrinsic perovskite quality material supported by photophysical studies. Moreover, efficiency of a planar inverted (p-i-n)

perovskite solar cell were greatly enhanced from 12.2% to 17.2 % in case of PbHPO3 NPs contain precursor. Our study demonstrates that the observed reproducibility problems in the perovskite solar cell community can be mainly attributed to unexpected impurities in MAI. Therefore, purity controls for self-synthesized or commercially available MAI are required. Finally, this work sheds light on the crystallization mechanism of perovskite thin films based on complex solutions of precursor, which rationalizes the performances variation from groups to groups and batches to batches.

- 130 -

Purity control of the precursor materials for perovskite synthesis Chapter 6

6.4. Bibliography 1. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, Journal of the American Chemical Society, 2009, 131, 6050-6051. 2. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-647. 3. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J.-P. Correa Baena, J.-D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, Science Advances, 2016, 2. 4. W. Zhang, S. Pathak, N. Sakai, T. Stergiopoulos, P. K. Nayak, N. K. Noel, A. A. Haghighirad, V. M. Burlakov, D. W. deQuilettes, A. Sadhanala, W. Li, L. Wang, D. S. Ginger, R. H. Friend and H. J. Snaith, Nat Commun, 2015, 6. 5. D. Prochowicz, M. Franckevicius, A. M. Cieslak, S. M. Zakeeruddin, M. Gratzel and J. Lewinski, Journal of Materials Chemistry A, 2015, 3, 20772-20777. 6. Q. Wang, Y. Shao, Q. Dong, Z. Xiao, Y. Yuan and J. Huang, Energy & Environmental Science, 2014, 7, 2359-2365. 7. K. Yan, M. Long, T. Zhang, Z. Wei, H. Chen, S. Yang and J. Xu, Journal of the American Chemical Society, 2015, 137, 4460-4468. 8. Y. Deng, E. Peng, Y. Shao, Z. Xiao, Q. Dong and J. Huang, Energy & Environmental Science, 2015, 8, 1544-1550. 9. C. Jiang, W. Liu and D. V. Talapin, Chemistry of Materials, 2014, 26, 4038-4043. 10. S.-S. Li, C.-H. Chang, Y.-C. Wang, C.-W. Lin, D.-Y. Wang, J.-C. Lin, C.-C. Chen, H.- S. Sheu, H.-C. Chia, W.-R. Wu, U. S. Jeng, C.-T. Liang, R. Sankar, F.-C. Chou and C.- W. Chen, Energy & Environmental Science, 2016, DOI: 10.1039/C5EE03229F. 11. D. V. Talapin, A. L. Rogach, A. Kornowski, M. Haase and H. Weller, Nano Letters, 2001, 1, 207-211. 12. N. V. Kuratieva, M. I. Naumova, N. V. Podberezskaya and D. Y. Naumov, Acta Crystallographica Section C, 2005, 61, i14-i16. 13. H. Ishida, H. Maeda, A. Hirano, T. Fujimoto, Y. Kubozono, S. Kashino and S. Emura, Zeitschrift für Naturforschung AZeitschrift für Naturforschung A, 1995, 50a, 876-880. 14. H. Belfadhel, A. Ratel, A. Ouederni, R. S. Bes and J. C. Mora, Ozone: Science & Engineering, 1995, 17, 637-645. 15. A. Dualeh, P. Gao, S. I. Seok, M. K. Nazeeruddin and M. Grätzel, Chemistry of Materials, 2014, 26, 6160-6164. 16. H. Ishida, R. Ikeda and D. Nakamura, Bull. Chem. Soc. Jpn., 1986, 59, 915-924. 17. G. Höhne, W. F. Hemminger and H. J. Flammersheim, Differential Scanning Calorimetry, Springer Berlin Heidelberg, 2013. 18. F. A. Bovey, L. Jelinski and P. A. Mirau, in Nuclear Magnetic Resonance Spectroscopy (Second Edition), Academic Press, San Diego, 1988, DOI: http://dx.doi.org/10.1016/B978-0-08-091699-6.50008-8, pp. 87-146. 19. J. J. Gutierrez-Sevillano, S. Ahmad, S. Calero and J. A. Anta, Physical Chemistry Chemical Physics, 2015, 17, 22770-22777. 20. H. Yuan, E. Debroye, K. Janssen, H. Naiki, C. Steuwe, G. Lu, M. Moris, E. Orgiu, H. Uji-i, F. De Schryver, P. Samorì, J. Hofkens and M. Roeffaers, The Journal of Physical Chemistry Letters, 2016, 7, 561-566. 21. N. Klein-Kedem, D. Cahen and G. Hodes, Acc. Chem. Res., 2016. 22. E. Edri, S. Kirmayer, A. Henning, S. Mukhopadhyay, K. Gartsman, Y. Rosenwaks, G. Hodes and D. Cahen, Nano Lett., 2014, 14, 1000-1004. 23. Y. Shao, Z. Xiao, C. Bi, Y. Yuan and J. Huang, Nature communications, 2014, 5. 24. B. Wu, K. Fu, N. Yantara, G. Xing, S. Sun, T. C. Sum and N. Mathews, Advanced Energy Materials, 2015.

- 131 -

Conclusions and outlook Chapter 7

Chapter VII

Conclusions and Outlook

7.1. Conclusions

The core study of my thesis is design and optimisation of high-quality semiconductor nanocrystals (NCs) for various luminescent and optoelectronic applications. Most importantly, each method of NC synthesis, presented in the thesis, enables scaling-up and reduction of the production costs, important commercial applications. This concept was successfully presented in Chapter 3 for multigram synthesis of CdZnS:Mn/ZnS NCs as efficient UV-to-visible light convertor in mc-Si solar cells. As a consequence, increase in the power conversion efficiency (PCE) by nearly 0.5 percentage points, approaching the theoretical limit of 0.6%, leads to a cost reduction of app. 3 % for mono-Si photovoltaic modules. However, Cd-containing NCs are less attractive for industry and real-world applications, despite the low toxicity of ZnCdS, pushing me to design new highly luminescent materials. With rising interest to perovskite colloidal NCs after the first publication in 2014,1 I developed and optimized a room-temperature synthesis for very highly luminescent perovskite NCs with well-defined morphology, structure and chemical composition control as well as precise optical properties (Chapter 4). The key attractive quality of that material class is their remarkably high PLQY reaching over 90% without shell passivation as is necessary for the CdZnS:Mn/ZnS NCs system. The full size and compositional tuning were achieved for MA and FA Lead halide NCs systems. Additionally, to solve the problem of the low water resistivity, surface modification with Silsesquioxane-POSS molecules allowed storing these NCs even in water dispersion within one year without losing their brightness. Following great potential of FAPbI3 colloidal NCs as a photodetector, I developed a fast and cost-effective synthesis for perovskite crystallites and ink thereof, which show remarkable high photodetection properties (Chapter 5). Such ink can be easily printed on substrates of any size, enabling realistic and precursor-free large-scale production of perovskite thin or thick films for multiple optoelectronic and X-ray detection applications.

- 132 -

Conclusions and outlook Chapter 7

In any chemical synthesis, material purity is a critical point to achieve desirable results and high-quality products. Perovskite metal halide synthesis is not an exception. Despite the simplicity of perovskite fabrication via low temperature (LT) solution process, the presence of low amount if impurities may turn the material properties and crystallization process in an unexpected direction. In Chapter 6, I performed a chemical study on the impurity formation

during MAI synthesis. It was found that transformation of H2PO2 as a stabilizer for HI leads to the formation of PbHPO3 NPs in perovskite precursor solution. Such NPs play a role as nucleation centres during the perovskite thin film crystallization yielding high-quality material. As a result, the performance of perovskite solar cell devices is shown to vary from 12.2% to 17.2% as a function of impurity content. However, in the case of nano- or microparticle synthesis, such impurities play a negative role and increase the particle size distribution, decrease colloidal and chemical stability as well as negatively affect the optical properties. Therefore, chemical purity control of perovskite precursor components allows obtaining colloidal NCs with very high PLQY and stability. My overall findings presented in this thesis demonstrate not only scientific interest, but also practical one, because all of the synthesis methods for different NCs systems can be easily scaled up as shown for each synthesis. As a consequence, low cost, scalable, comparably low- toxic highly efficient energy conversion materials were obtained. This combination is a key for practical usage of the above described material, and I believe that some products based on those NCs will be available on the market in near future. The future is Bright!

7.2. Outlook

In order to realize the cost potential of semiconductor nanocrystals and application of thereof, the commercial prospects for this class of materials should be provided. The global quantum dot market counts a revenue of $316 million in 2013 and is expected to grow to about $5 billion by 2020. 2 Quantum dots are already successfully employed in novel and color rich TV displays (Samsung QE65Q8C for example). As a result, lighting and displays sector represent a potential of $100 billion global market by 2020, resulting in significant opportunities for quantum dots.3 Therefore, based on the achievements of this dissertation, the further practical application can be realized in near future. Luminescent applications. As mentioned above, display technologies will be a major field in the future for luminescent NCs commercialisation. The FAPbX3 NCs synthesized at room temperature with high PLQY (nearly 90%) and color purity because of their narrow

- 133 -

Conclusions and outlook Chapter 7

FHWM (20-44 nm) would be the perfect material for display market, especially FAPbBr3 NCs for green color component with their 530 emission peak, which perfectly satisfies the Rec.2020 standard for the current UHDTV TV display generation.3 An additional water-resistant coating that was performed for these NCs open the opportunity to use this material in any long-term stable luminescent application. Furthermore, such remarkably high PLQY is favorable for

utilizing the FAPbBr3 NCs for example as luminescent down-shifting layers in Si, CdTe or

CuInSe2 solar cells and LED technologies. Quantum dots and monograin Solar cells. The current state of the art of quantum dot solar cells is mainly based on PbS NCs utilization and it was well developed by Sargent group at the University of Toronto. The record certified PCE of that solar cell device is over 10%.4 However, this value was achieved within one decade of progress, and multiple steps of PbS surface modification and specific interfaces were involved to reach that PCE value.5 In the case of perovskite QDs solar cells, first published work on this topic was based on colloidal

CsPbI3 QDs and exhibit around 11% of PCE, which makes perovskite QDs solar cells very promising direction in terms of easier manufacturing as well as high PCE in the end. Nonetheless, there are still no reported data on QDs solar cells fabricated out of hybrid organic-inorganic lead halide perovskite colloidal NCs. A core reason for it is that MAPbI3 or their mixhalide pendant MAPb(Br/I)3 shows pure colloidal and chemical stability limiting their

utilization in QDs solar cell. Alternatively, more stable perovskite FAPbI3 NCs that I designed in this thesis (Chapter 4) would be an ideal candidate for this purpose. Furthermore, ligand- free and printable perovskite ink presented in the Chapter 5, would help to revolutionize the perovskite solar cell fabrication towards to large scale and cheap solar cell printing. Moreover, this ink can be used for printing of X-ray detector applications which may replace more toxic and highly cost CdTe or CdZnTe based X-ray detectors in medicine.

- 134 -

Conclusions and outlook Chapter 7

7.3. Bibliography

1. L. C. Schmidt, A. Pertegás, S. González-Carrero, O. Malinkiewicz, S. Agouram, G. Mínguez Espallargas, H. J. Bolink, R. E. Galian and J. Pérez-Prieto, J. Am. Chem. Soc., 2014, 136, 850-853. 2. Quantum Dot Market forecast, 2016, https://www.alliedmarketresearch.com/quantum- dots-market. 3. K. Wang and X. Wei Sun, Information Display, 2016, 6-14. 4. X. Lan, O. Voznyy, F. P. García de Arquer, M. Liu, J. Xu, A. H. Proppe, G. Walters, F. Fan, H. Tan, M. Liu, Z. Yang, S. Hoogland and E. H. Sargent, Nano Lett., 2016, 16, 4630-4634. 5. G. H. Carey, A. L. Abdelhady, Z. Ning, S. M. Thon, O. M. Bakr and E. H. Sargent, Chem. Rev., 2015, 115, 12732-12763.

- 135 -

Curriculum Vitae

Ievgen Levchuk

Nationality: Ukraine Gender: Male Date of birth: 28.08.1990

Education

09/1997 – 06/2007 High school graduate certificate (Honor’s certificate in chemistry), Myhove secondary school of I-III level , Ukraine 09/2007 – 05/2011 Bachelor degree in INORGANIC CHEMISTRY, Chemistry department, Chernivtsy National Universuty, Ukraine

09/2011 – 05/2012 Chernivtsy National University, Chemistry department, Ukraine Master degree in INORGANIC CHEMISTRY

04/2013 – 2017 Ph.D Fellow in Friedrich-Alexander University Erlangen- Nürnberg, Materials for Electronics and Energy Technology (i-MEET), Erlangen, Germany

Topic: Development of phosphors for light conversion in solar panels

Skills

Characterization techniques: Scanning electron microscopy (SEM), XRD, UV -IR spectroscopy, time-resolved spectroscopy, FTIR.

Experiment design of any nanocrystal.

IT: Office, WEB, Isis Draw, Origin, AutoCad, Photoshop, Powerdirector, Pinnacle studio, Vesta, ImageJ

- 136 -

Languages:

Understanding Speaking Writing Listening Reading Interaction Production English excellent excellent advanced advanced good Ukranian Native speaker Russian excellent excellent advanced advanced good German basic normal basic basic good

Scientific activities

Publications

1. Taylor Uekert; Anastasiia Solodovnyk; Sergei Ponomarenko; Andres Osvet; Ievgen Levchuk; Jessica Gast; Miroslaw Batentschuk; Karen Forberich; Edda Stern; Hans-Joachim Egelhaaf; Christoph J. Brabec, Nanostructured organosilicon luminophores in highly efficient luminescent down-shifting layers for thin film photovoltaics. Solar Energy Materials and Solar Cells 2016, 155, 1-8. 2. Xiaofeng Tang; Marco Brandl; Benjamin May; Ievgen Levchuk; Yi Hou; Moses Richter; Haiwei Chen; Shi Chen; Simon Kahmann; Andres Osvet, Christoph J. Brabec, Photoinduced degradation of methylammonium lead triiodide perovskite semiconductors. Journal of Materials Chemistry A 2016, 4 (41), 15896-15903. 3. César Omar Ramírez Quiroz; Carina Bronnbauer; Ievgen Levchuk; Yi Hou; Christoph J. Brabec; Karen Forberich, Coloring Semitransparent Perovskite Solar Cells via Dielectric Mirrors. ACS nano 2016, 10 (5), 5104-5112. 4. Cesar Omar Ramirez Quiroz; Ievgen Levchuk; Carina Bronnbauer; Michael Salvador; Karen Forberich; Thomas Heumuller; Yi Hou; Peter Schweizer; Erdmann Spiecker; Christoph J. Brabec, Pushing efficiency limits for semitransparent perovskite solar cells. Journal of Materials Chemistry A 2015, 3 (47), 24071-24081. 5. Daniel Niesner; Max Wilhelm; Ievgen Levchuk; Andres Osvet; Shreetu Shrestha; Miroslaw Batentschuk; Christoph J. Brabec; Thomas Fauster, Giant Rashba Splitting in

CH3NH3PbBr3 Organic-Inorganic Perovskite. Physical Review Letters 2016, 117 (12), 126401. 6. Ievgen Levchuk; Christian Würth; Florian Krause; Andres Osvet; Miroslaw Batentschuk; Ute Resch-Genger; Claudia Kolbeck; Patrick Herre; Hans-Peter Steinrück; Wolfgang Peukert, Christoph J. Brabec, Industrially scalable and cost-effective Mn2+ doped

ZnxCd1−xS/ZnS nanocrystals with 70% photoluminescence quantum yield, as efficient down-shifting materials in photovoltaics. Energy & Environmental Science 2016, 9 (3), 1083-1094. 7. Ievgen Levchuk; Yi Hou; Marco Gruber; Marco Brandl; Patrick Herre; Xiaofeng Tang; Florian Hoegl; Miroslaw Batentschuk; Andres Osvet; Rainer Hock, Christoph J. Brabec, Deciphering the Role of Impurities in Methylammonium Iodide and Their Impact on the Performance of Perovskite Solar Cells. Advanced Materials Interfaces 2016, 3 (22).

- 137 -

8. Ievgen Levchuk; Patrick Herre; Marco Brandl; Andres Osvet; Rainer Hock; Wolfgang Peukert; Peter Schweizer; Erdmann Spiecker; Miroslaw Batentschuk; Christoph J. Brabec,

Ligand-assisted thickness tailoring of highly luminescent colloidal CH3NH3PbX3 (X= Br and I) perovskite nanoplatelets. Chemical Communications 2017, 53 (1), 244-247. 9. Laraib Sarfraz Khanzada; Ievgen Levchuk (shared); Yi Hou; Hamed Azimi; Andres Osvet; Rameez Ahmad; Marco Brandl; Patrick Herre; Monica Distaso; Rainer Hock,

Christoph J. Brabec, Effective Ligand Engineering of the Cu2ZnSnS4 Nanocrystal Surface for Increasing Hole Transport Efficiency in Perovskite Solar Cells. Advanced Functional Materials 2016, 26 (45), 8300-8306. 10. Haiwei Chen; Yi Hou; Christian E Halbig; Shi Chen; Hong Zhang; Ning Li; Fei Guo; Xiaofeng Tang; Nicola Gasparini; Ievgen Levchuk, Christoph J. Brabec, Extending the environmental lifetime of unpackaged perovskite solar cells through interfacial design. Journal of Materials Chemistry A 2016, 4 (30), 11604-11610. 11. Daniel Niesner; Oskar Schuster; Max Wilhelm; Ievgen Levchuk; Andres Osvet; Shreetu Shrestha; Miroslaw Batentschuk; Christoph Brabec; Thomas Fauster, Christoph J. Brabec, Temperature-dependent optical spectra of single-crystal (CH3NH3)PbBr3 cleaved in ultrahigh vacuum. Physical Review B 2017, 95 (7), 075207. 12. Ahmed M. Salaheldin; Johannes Walter; Patrick Herre; Ievgen Levchuk; Yasaman Jabbari; Joel M. Kolle; Christoph J. Brabec; Wolfgang Peukert; Doris Segets, Automated synthesis of quantum dot nanocrystals by hot injection: Mixing induced self-focusing. Chemical Engineering Journal 2017, 320, 232-243. 13. Ievgen Levchuk; Andres Osvet; Xiaofeng Tang; Marco Brandl; José Darío Perea; Florian Hoegl; Gebhard J. Matt; Rainer Hock; Miroslaw Batentschuk; Christoph J. Brabec, Brightly

Luminescent and Color-Tunable Formamidinium Lead Halide Perovskite FAPbX3 (X=Cl, Br, I) Colloidal Nanocrystals. Nano Letters 2017, doi: 10.1021/acs.nanolett.6b04781. 14. Daniel Niesner; Martin Hauck; Shreetu Shrestha; Ievgen Levchuk; Gebhard J Matt; Andres Osvet; Miroslaw Batentschuk; Christoph J. Brabec; Heiko B Weber; Thomas Fauster, Spin-

split bands cause the indirect band gap of (CH3NH3)PbI3: Experimental evidence from circular photogalvanic effect. arXiv preprint arXiv: 2017, 1703.08740. 15. Shreetu Shrestha; René Fischer; Gebhard J Matt; Patrick Feldner; Thilo Michel; Andres Osvet; Ievgen Levchuk; Benoit Merle; Saeedeh Golkar; Haiwei Chen; Sandro F Tedde; Oliver Schmidt, Rainer Hock, Manfred Rührig, Mathias Göken, Wolfgang Heiss, Gisela Anton; Christoph J. Brabec, High-performance direct conversion X-ray detectors based on sintered hybrid lead triiodide perovskite wafers. Nature Photonics 2017, 11, 436–440. 16. Nicola Gasparini; Michael Salvador; Sebastian Strohm; Thomas Heumueller; Ievgen Levchuk; Andrew Wadsworth; James H. Bannock; John C. de Mello; Hans-Joachim Egelhaaf; Derya Baran; Iain McCulloch, and Christoph J. Brabec, Burn-in Free Nonfullerene-Based Organic Solar Cells. Advanced Energy Material 2017, 1700770. 17. V. Gorbenko; T. Zorenko; K. Paprocki; A. Iskaliyeva; A. Fedorov; F. Schröppel; I. Levchuk; A. Osvet; M. Batentschuk; Yu. Zorenko, Epitaxial growth of single crystalline

film phosphors based on the Ce3+-doped Ca2YMgScSi3O12 garnet, CrystEngComm 2017,19, 3689-3697.

- 138 -

Oral talks

1. Ievgen Levchuk, Christian Würth, Florian Krause, Andres Osvet, Miroslaw Batentschuk, Ute Resch-Genger, and Christoph J. Brabec, Smart nanophosphor material for effective Near-UV-light conversion: synthesis, upscaling and photovoltaic application. Invited spiker for 4th International Symposium on Ceramics Nanotune Technology (ISCeNT4) March 2-4 2015 , Nagoya , Japan. 2. Ievgen Levchuk, Liudmyla Chepyga, Andres Osvet, Alfons Stiegelschmitt, Amir Hashemi, Gordana Jovicic, Andreas Vetter, Miroslaw Batentschuk, Albrecht Winnacker and Christoph J. Brabec, Colloidal nanoparticles LaPO4:Dy3+ as a printable material for the luminescent thermometry. Presented by Priv.-Doz. Dr. Miroslaw Batentschuk on E-MRS Fall Meeting 2015, Warszawa Symposium F. 3. Ievgen Levchuk, Christian Würth, Florian Krause, Andres Osvet, Miroslaw Batentschuk, Ute Resch-Genger, and Christoph J. Brabec. Highly luminescent Mn-doped ZnCdS/ZnS core-shell nanocrystals. Workshop on "Nanocrystals and Film Formation" of the Research Training Group (Graduiertenkolleg) 1161/2, Disperse Systems in Electronics, 16 April 2014, Erlangen 4. Ievgen Levchuk; Florian Hoegl; Marco Brandl; Andres Osvet; Rainer Hock; Patrick Herre; Wolfgang Wolfgang; Peter Schweizer; Erdmann Spiecker; Miroslaw Batentschuk; Christoph J Brabec. Organometallic perovskites for optoelectronic applications, SPIE Optics + Photonics 2016, San Diego, California, United States, August 28 - September 01, 2016. 5. Ievgen Levchuk; Yi Hou; Marco Gruber; Patrick Herre; Marco Brandl; Andres Osvet; Rainer Hock; Wolfgang Peukert; Rik R Tykwinski; Miroslaw Batentschuk; Christoph J Brabec. Deciphering the role of impurities in methylammonium iodide and their impact on the performance of perovskite solar cells, SPIE Optics + Photonics 2016, San Diego, California, United States, August 28 - September 01, 2016. 6. César Omar Ramírez Quiroz; Carina Bronnbauer; Ievgen Levchuk (as presenter); Michael Salvador; Yi Hou; Karen K Forberich; Christoph J Brabec. Coloring semitransparent room-temperature fabricated perovskite solar cells via dielectric mirrors, SPIE Optics + Photonics 2016, San Diego, California, United States, August 28 - September 01, 2016. 7. Ievgen Levchuk; Yi Hou; Marco Gruber; Marco Brandl; Patrick Herre; Xiaofeng Tang; Florian Hoegl; Miroslaw Batentschuk; Andres Osvet; Rainer Hock; Wolfgang Peukert; Rik R. Tykwinski; Christoph J. Brabec. Unraveling the role of methylammonium iodide impurities in perovskite precursor solution and their impact on the performance of solar cells, 2nd International Conference on Perovskite Solar Cells and Optoelectronics (PSCO-2016), Genova, Italy, September 26-28, 2016.

- 139 -

Posters

1. Ievgen Levchuk, Christian Würth, Florian Krause, Andres Osvet, Miroslaw Batentschuk, Ute Resch-Genger, and Christoph J. Brabec, Synthesis and investigatons of

ZnxCd1−xS:Mn/ZnS Quantum Dots as possible near UV-light converters for solar cells, 2nd International Congress "Next Generation Solar Energy" 09-11 December 2013, Erlangen. 2. Ievgen Levchuk, Christian Würth, Florian Krause, Andres Osvet, Miroslaw Batentschuk,

Ute Resch-Genger, and Christoph J. Brabec. Highly luminescent Mn-doped alloy ZnxCd1-xS quantum dots as luminescent down-shifting layer for Si solar cells, 17th International Conference on Luminescence and Optical Spectroscopy of Condensed Matter (ICL’14), 13-18 July, 2014, Wroclaw, Poland 3. Ievgen Levchuk, Christian Würth, Florian Krause, Andres Osvet, Miroslaw Batentschuk,

Ute Resch-Genger, and Christoph J. Brabec. Highly luminescent Mn-doped alloy ZnxCd1-xS quantum dots as luminescent down-shifting layer for Si solar cells, 6th EAM Symposium, 24 – 26 November 2014, Kloster Banz, Bad Staffelstein 4. Ievgen Levchuk; Yi Hou; Marco Gruber; Marco Brandl; Patrick Herre; Xiaofeng Tang; Florian Hoegl; Miroslaw Batentschuk; Andres Osvet; Rainer Hock; Wolfgang Peukert; Rik R. Tykwinski; Christoph J. Brabec. Deciphering the role of impurities in methylammonium iodide and their impact on the performance of perovskite solar cells 3rd International CongressNext Generation Solar Energy Meets Nanotechnology, Erlangen, November 23-25, 2016, 5. Bianka. M. D. Puscher, Rezvan Soltani, Ievgen Levchuk, Christoph J. Brabec, Tayebeh Ameri, Dirk M. Guldi. En route towards efficient ternary blends in BHJ Solar Cells, 8th EAM Symposium, Kloster Banz – Bustransfer, November 21-23, 2016.

Summer school

1. Physical properties of nanoparticles: Characterization and applications, Physikzentrum Bad Honnef, 26.07.2015 - 31.07.2015

- 140 -