Citation
High Efficiency Perovskite Solar Cells with Long Operation Lifetime

Material Information

Title:
High Efficiency Perovskite Solar Cells with Long Operation Lifetime
Creator:
Yang, Chenchen
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (88 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
XUE,JIANGENG
Committee Co-Chair:
PERRY,SCOTT S
Committee Members:
SO,FRANKY FAT KEI
Graduation Date:
5/2/2015

Subjects

Subjects / Keywords:
Current density ( jstor )
Diffuse scleroderma ( jstor )
Electrodes ( jstor )
Electrons ( jstor )
Narrative devices ( jstor )
Oxygen ( jstor )
Performance photography ( jstor )
Perovskites ( jstor )
Photovoltaic cells ( jstor )
Power efficiency ( jstor )
Materials Science and Engineering -- Dissertations, Academic -- UF
energy -- lifetime -- perovskite -- solar -- stability
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Materials Science and Engineering thesis, M.S.

Notes

Abstract:
The rapid development of perovskite solar cell has gained enormous attention of the photovoltaic research community worldwide. Due to its excellent light absorption and good carrier transport property, the power conversion efficiency is now growing above 20%. However, beyond the great achievement of high efficiency, stability and toxicity issues as two major challenges need to be solved before its commercialization in the future. In this work, we successfully fabricated high efficiency perovskite solar cell by simple solution process, and meanwhile a long device operation lifetime has been realized. The first part of our work is focused on fabricating high efficiency perovskite solar cell by one step process. Different carrier transport layer materials were compared and thickness of perovskite layer was carefully adjusted, aiming at optimizing carrier transport and light absorption to improve photovoltaic performance. A high power conversion efficiency of 14.1% was finally achieved. The second part is to explore the degradation mechanism of our perovskite solar cell. Effects of oxygen and moisture on selected layers of the device were investigated. The cause of degradation was found and discussed based on a series of experiments and related characterization work. Moreover, the function of zinc oxide nanoparticle layer was also explained. Finally, long operation lifetime and good stability was realized by the protection of ZnO layer and proper encapsulation. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2015.
Local:
Adviser: XUE,JIANGENG.
Local:
Co-adviser: PERRY,SCOTT S.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-05-31
Statement of Responsibility:
by Chenchen Yang.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
5/31/2016
Classification:
LD1780 2015 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

HIGH EFFICIENCY PEROVSKITE SOLAR CELLS WITH LONG OPERATION LIFETIME By CHENCHEN YANG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR TH E DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2015

PAGE 2

© 2015 Chenchen Yang

PAGE 3

To my family

PAGE 4

4 ACKNOWLEDGMENTS There are many individuals to whom I would like to express my sincere gratitude during my graduate study at University of Florida. At first, I would like to thank my thesis advisor, Dr. Jiangeng Xue for offering me this precious opportunity to work in his group. I have gained advanced know ledge and extensive experience from his valuable guidance. I also acknowledge my other committee members Dr. Franky So, Dr. Jennifer Andrew and Dr. Scott Perry for their time and support. Senior graduate students in Xue gr oup ga ve me a lot of help during this work. I would like to thank Dr. Weiran Cao for his continuously mentoring and indispensable advising during this work. I also want to thank Dr. John Mudrick for the initial training, and Nathan Shewmon for his help wit h all the characterization work. I want to acknowledge the visiting scholar Dr. Huaibin Shen for providing me many useful suggestions in this work. Adharsh Rajagopal as both my labmate and classmate , thank s for being a good companion in the lab. I also need to thank my friends, Bowen Zhang, Tao Yu for sharing their expertise in computer programing. We also enjoyed the life in this lovely small town Gainesville together . Finally, the acknowledgements are not complete without thanking my beloved fam ily. My parents always support me with their consistent encouragement , especially through my lows in my graduate study at University of Florida . Thank you for your love and care, and this thesis is dedicated to you.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 FUNDAMENTALS OF PEROVSKITE SOLAR CELLS ................................ ........... 15 1.1 Overview ................................ ................................ ................................ ........... 15 1.2 Important Pa rameters of Solar Cells ................................ ................................ . 16 1.3 Characterization of Solar Cells ................................ ................................ ......... 18 1.4 Brief Introduction to Perovskite Solar Cells ................................ ....................... 20 1.4.1 Basic Properties of Perovskite Crystal Structure ................................ ..... 20 1.4.2 The Emergence of Perovskite Solar Cells ................................ ............... 22 1.5 Review of Perovskite Solar Cells ................................ ................................ ...... 24 1.5.1 Fabrication Techniques of Perovskite Solar Cells ................................ ... 24 1.5.2 Device Structures of Perovskite Solar Cells ................................ ............ 28 1.5.3 Bandgap Modification by Ionic Replacement ................................ ........... 32 1.6 Outline of the Thesis ................................ ................................ ......................... 35 2 HIGH EFFICIENCY PEROVSKITE SOLAR CELL FABRICATION BY SOLUTION PROCESS ................................ ................................ ........................... 37 2.1 Introduction ................................ ................................ ................................ ....... 37 2.2 Experimental Details of Device Fabrication ................................ ...................... 37 2.2.1 Methylammonium Iodide Synthesis ................................ ......................... 37 2.2.2 Device Fabrication ................................ ................................ ................... 38 2.3 Device Characterization ................................ ................................ .................... 39 2.4 Candidates for ETM and HTM ................................ ................................ .......... 43 2.4.1 ETM Selection ................................ ................................ ......................... 43 2.4.2 HTM Selection ................................ ................................ ......................... 45 2.5 Pero vskite Layer Thickness Optimization ................................ ......................... 46 2.6 Causes of Inflated Results ................................ ................................ ................ 48 2.7 Short Circuit Current Density Calibration ................................ .......................... 56 2.8 Summary ................................ ................................ ................................ .......... 58

PAGE 6

6 3 DEGRADATION ME CHANISM EXPLORATION AND LONG OPERATION LIFETIME PEROVSKTE SOLAR CELL ................................ ................................ .. 60 3.1 Introduction ................................ ................................ ................................ ....... 60 3.2 Perovskite Photovoltaic Devices with vs without ZnO ................................ ....... 60 3.3 Device Performance Degradation Mechanism Exploration ............................... 62 3.3.1 Perovskite Photovoltaic Devices with Selected Layers Exposed to Air ... 63 3.3.2 Perovskite Photovoltaic Devices with Selected Layers Exposed to Oxygen vs Moisture ................................ ................................ ...................... 65 3.3.3 Perovskite Photovoltaic Devices with Al Electrode vs Ag Electrode ........ 70 3.3.4 Perovskite Photovoltaic Devices Exposed to Oxygen vs Moisture .......... 72 3.4 Discussion of Degradation Mechanisms ................................ ........................... 73 3.5 Long Operation Lifetime and High Stability Perovskite Solar Cell ..................... 76 3.6 Summary ................................ ................................ ................................ .......... 79 4 CONCLUSIONS AND FUTURE WORK ................................ ................................ . 81 4.1 Conclusions ................................ ................................ ................................ ...... 81 4.1.1 High Efficiency Perovskite Solar Cell Fabrication ................................ .... 81 4.1.2 Degradation Mechanism Discussion ................................ ....................... 82 4.2 Future Work ................................ ................................ ................................ ...... 83 4.2.1 Toxicity Issue ................................ ................................ ........................... 83 4.2.2 Stability Issue ................................ ................................ .......................... 83 LIST OF REFERENCES ................................ ................................ ............................... 84 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 88

PAGE 7

7 LIST OF TABLES Table page 1 1 Ionic radius of some A site, B site and X site ions in metal halide perovskite. ... 35 2 1 Device performance characteristics of different electron transport materials. .... 45 2 2 Photovoltaic performance parameters of devices with different perovskite layer thickness. ................................ ................................ ................................ ... 47 2 3 Photovoltaic performance parameters for different scan speeds and directions. ................................ ................................ ................................ ........... 51 2 4 Resistivity of PEDOT: PSS, PEDOT: PSS/Perovskite and PEDOT: PSS/Perovskite/PCBM/ZnO. ................................ ................................ .............. 54 2 5 Photovoltaic performance parameters for different aperture sizes. .................... 57 2 6 Photovoltaic performance parameters of devices with different perovskite layer thickness. (Note: 2 mm diameter aperture is applied for J V characteristic test.) ................................ ................................ ............................. 58 3 1 Photovoltaic performance parameters of devices with and without ZnO exposed to air for different amount of time. ................................ ........................ 62 3 2 Photovoltaic performance parameters of devices with selected layers exposed to air. ................................ ................................ ................................ .... 64 3 3 Ph otovoltaic performance parameters of devices with selected layers exposed to oxygen. ................................ ................................ ............................ 67 3 4 Photovoltaic performance parameters of devices with selected layers exposed to moisture. ................................ ................................ .......................... 69 3 5 Photovoltaic performance parameters of devices with Al electrode and Ag electrode exposed to air for different time. ................................ ......................... 71 3 6 Photovoltaic performance parameters of devices exposed to oxygen vs moisture for different time. ................................ ................................ .................. 73

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Schematic illustration of an I V curve for organic solar cell under dark and illumination. ................................ ................................ ................................ ......... 17 1 2 Schematic illustration of J V characteristic measurement setup. ........................ 19 1 3 Schematic illustration of EQE measurement setup. ................................ ............ 20 1 4 Schematic illustration of cubic perovskite crystal structure. ................................ 21 1 5 Rapid growth in perovskite research according to the Web of Science database. ................................ ................................ ................................ ............ 23 1 6 General fabrication methods to deposit perovskite film. ................................ ..... 28 1 7 Bandgap modification by ionic replacement. ................................ ...................... 34 2 1 Perovskite solar cell device architecture and energy diagram . ........................... 39 2 2 X ray diffraction spectra of perovskite film. ................................ ......................... 40 2 3 Scanning Electron Microscope Images of the same perovskite film by one step solution process . ................................ ................................ ......................... 41 2 4 Photovoltaic performance characterization of s tandard perovskite solar cell. .... 42 2 5 Perovskite solar cell with different electron transport materials and ho le transport materials . ................................ ................................ ............................. 44 2 6 Photovoltaic performance characterization of perovskite solar cell with differe nt perovskite layer thickness. ................................ ................................ .... 46 2 7 Histogram of device power conversion efficiency under same fabrication condition ................................ ................................ ................................ ............. 49 2 8 J V characteristic for differe nt scan speeds and directions . ................................ 50 2 9 EQ E beam linear scan experiment. ................................ ................................ .... 52 2 10 Current Voltage characteristic experiment for PED OT: PSS layer . .................... 54 2 11 Schematic illustration of carrier transport in PEDOT: PSS layer. ........................ 56 2 12 Aperture selection for short circuit curr ent density calibration . ........................... 57

PAGE 9

9 2 13 Photovoltaic performance characterization of perovskite solar cell with different perovskite layer thickness by applying 2mm diameter aperture. .......... 58 3 1 Stability comparison betwe en devices with and without ZnO . ............................ 61 3 2 Characterization of perovskite photovoltaic devices with s elected layers exposed to air . ................................ ................................ ................................ .... 63 3 3 Current density voltage characteristics of the devices with sele cted layers exposed to oxygen . ................................ ................................ ............................ 66 3 4 Current density voltage characteristics of the devices with select ed layers exposed to moisture. ................................ ................................ .......................... 69 3 5 Stability comparison between devices with A l electrode and Ag electrode. ........ 71 3 6 Stability comparison between devices exposed to oxygen and moisture. .......... 72 3 7 Al and Pb atomic percentage as a function of moisture exposure time from X ray p hotoelectron spectroscopy measurement. ................................ .................. 74 3 8 X ray Diffraction spectra of perovskite film as a funct ion of moisture exposure time . ................................ ................................ ................................ .................... 76 3 9 Photovoltaic parameters of encapsulated device with ZnO as a function of constant illumination time. ................................ ................................ .................. 77 3 10 Photovoltaic parameters of encapsulated device without ZnO as a function of constant illumination time. ................................ ................................ .................. 78 3 11 On shelf lifetime of device with ZnO. ................................ ................................ .. 79

PAGE 10

10 LIST OF ABBREVIATIONS Ag Silver Al Aluminum AM 1.5G Air Mass 1.5 Global ASTM American Society for Testing and Materials c Speed of light DI Deionized DIO Diiodooctane DMF Dimethylformamide DMSO dimethylsulphoxide DSVD Dual Source Vacuum Deposition EQE External Quantum Efficiency ETA Extreme Thin Absorber ETM Electron Transport Material eV Electron volt FF Fill Factor FTO Fluorine doped tin oxide GBL butyrolactone h H 101 3,4 ethylenedioxythiophene HTM Hole Transport Material ICBA Indene C 60 bisadduct I cell Photocurrent of test cell I det Current measured by silicon photo detector IPA Isopropyl alcohol

PAGE 11

11 IQE Internal Quantum Efficiency I SC Short circuit current ITO Indium Tin Oxide I V Current Voltage J SC Short circuit current density J V Current density voltage KRICT Korean Research Institute of Chemical Technology LED Light Emitting Diode MoO X Molybdenum oxide NREL National Renewable Energy Laboratory OPV Organic Photovoltaic PC 71 BM Phenyl C 71 butyric acid methyl ester PCBM Phenyl C 61 butyric acid methyl ester PCBTDPP P oly[N 9 hepta decanyl 2,7 carbazole alt 3,6 bis (th iophen 5 yl) 2,5 dioctyl 2,5 di hydropyrrolo[3,4 ]pyrrole 1,4 dione] P Power Conversion Efficiency PEDOT: PSS Poly(3,4 ethylenedioxythiophene):Polystyrene sulfonate P m,out Maximum power output P o Incident power PTAA Poly triarylamine q Charge of electron QTH Quartz Tungsten Halogen rpm Rotations per minute S olar spectrum S det Spectral responsivity of silicon detector

PAGE 12

12 SDM Sequential Deposition Method SEM Scanning Electron Microscope spiro OMeTAD 2,2 ,7,7 ,b tetrakis (N,N dimethoxyphenyl amine) 9,9 spirobi fluorene UV Ultraviolet VASP Vapor Assisted Solution Process V OC Open circuit voltage VTM Vacuum Thermal Evaporation Xe Xe non XRD X Ray Diffraction ZnO Zinc Oxide A Absorption efficiency Wavelength

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science HIGH EFFICIENCY PEROVSKITE SOLAR CELLS WITH LONG OPERATION LIFETIME By Chenchen Yang May 2015 Chair: Jiangeng Xue Major: Materials Science and Engineering The rapid development of perovskite solar cell has gained enormous attention of the photovoltaic research community worldwide. Due to its excelle nt light absorption and carrier transport property, the power conversion efficiency is now growing above 20%. H owever, b eyond the great achievement of high efficiency, stability and toxicity issues as two major challenges need to be solved before its commercialization in the future . In this work, we successfully fabricate d high efficiency perovskite solar cell by s imple solution process, and meanwhile a long device operation lifetime has been realized. The first part of our work is focused on fabricating high efficiency perovskite solar cell by one step process. Different carrier transport layer materials were compa red and thickness of perovskite layer was carefully adjusted, aiming at optimizing carrier transport and light absorption to improve photovoltaic performance. A high power conversion efficiency of 14.1% was finally achieved. The second part is to explore t he degradation mechanism of our perovskite solar cell. Effect s of oxygen and moisture on selected layer s of the device were investigated.

PAGE 14

14 The cause of degradation was found and discussed based on a series of experiments and related characterization work . M oreover, the function of zinc oxide nanoparticle layer was also explained. Finally, long operation lifetime and good stability was realized by the protection of ZnO lay er and proper encapsulation.

PAGE 15

15 CHAPTER 1 FUNDAMENTALS OF PEROVSKITE SOLAR CELLS 1.1 Overview Photovoltaic technology as one of the most promising renewable energy branches is able to convert radiant sun light into electricity. In the future, it is expected to gradually replace the conventional fu els and meet the requ irement for clean energy needs . Currently, the best performance solar cell in the market are mainly based on inorganic semiconductor such as silicon, gallium arsenide, etc. The highest power conversion efficiency of 46.0% 1 for a multi junc tion cell has been confirmed by National Renewable Energy Laboratory (NREL) so far. Even such high performance has been achieved, however, the scarcity of raw materials, complicated fabrication process and other factors lead to high cost to efficiency rati o 2 , restricting the large scale commercialization of solar energy 3 . In the past several years, perovskite solar cells have attracted tremendous attention for its outstanding photovoltaic performance. Since the y ear of 2009, Miyasaka and co workers demonstrated the pioneering work of perovskite solar cell with a 3.81% 4 efficiency, a series of significant breakthroughs have made. The current highest confirmed power conversion eff iciency (PCE) has reached 20.1% 1 , whi ch has already surpassed the highest efficiency of organic solar cells or dye sensitized solar cells. In addition to its excellent performance, the versatility in fabrication techniques and various device architectures also stimulate the development of per ovskite solar cells. In this chapter, the fundamental concepts of general photovoltaic devices and perovskite solar cells are provided. Important parameters and characterization of a solar cell will be introduced in Section 1.2 and Section 1.3, respective ly. An introduction to

PAGE 16

16 perovskite solar cells focusing on crystal structure and the key features is in Section 1.4, and a brief review of current progress including fabrication methods, device architectures and influence of ionic replacement will be covere d in section 1.5. Finally, an outline of the subsequent chapters in this thesis is given in Section 1.6. 1.2 Important Parameters of Solar Cells The most important method to characterize the performance of a solar cell is the current density voltage characteristics. Figure 1 1 illustrates the I V curves of a photovoltaic cell under dark and illumination conditions. Three important parameters coul d be extracted from the I V curve under illumination , which are short circuit current density (Jsc), open circuit voltage (Voc) and fill factor (FF). Short circuit current density is the current at zero voltage divided by the device active area, and the open circuit voltage is the voltage at zero current, respectively. The maximum output power is the maximum product value of current density and voltage on the entire J V curve. The fill factor (FF) which represent s V curve is defined by: (1 1) Finally, t P ) is given as the ratio of maximum output power P m ,out to the incident power P 0 , which can be written as: (1 2) To analyze the behavior of a photovoltaic device responding to the incident solar spectra, quantum efficiencies are important parameters to measure. The external quantum efficiency is defined as the ratio of amount of charge carriers (electrons or holes) to the amount of incident photons, and the internal quantum efficiency

PAGE 17

17 represents the ratio of amount o f charge carriers collected by electrode to the amount of absorbed photons in photoactive layer of the phot ovoltaic device. Two t ypes of quantum efficiencies are related through: (1 3) where is the absorption efficiency . If the solar spectrum is provided, photocurrent density can be integrated from external quantum efficiency: (1 4) where q is the electron charge, is the wavelength, h is the the speed of light. For the same photovoltaic device, the photo current density extracted from J V characteristic should match with the photocurrent density integrated from EQE measurement, and the mismatch between these two photocurrent densities should be controlled within 5%. Figure 1 1. Schematic illustration of an I V curve for organic solar cell under dark and illum ination.

PAGE 18

18 1.3 Characterization of Solar Cells The J V characteristic is measured by the Agilent 4155C semiconductor parameter analyzer. A voltage bias scan is applied on the device , and the corresponding current reading is recorded at each voltage step . According to American Society for Testing and Materials (ASTM), the photovoltaic device should be tested under 100mW/cm 2 (1 Sun) irradiant light intensity AM (air mass) 1.5 global reference solar spectrum, and at the temperature of 25 ° C. As shown in Figur e 1 2, the light generated from the Xenon lamp is used to simulate the solar spectra. The light beam pass through an AM 1.5G filter which is used to remove the strong ultra violet (UV) light emitted from the Xenon lamp , in order to simulate the real solar spectrum. The optic setup is to make the beam collimated and uniform over a required area for J V measurement. Then, the light beam intensity is adjusted by the neutral density fil ter wheels to match with the 1 s un intensity which is calibrated by a single crystalline silicon reference cell. The light intensity can also be slightly changed by adjusting the input lamp power (150±5W) for the variation. The tested substrate placed in a test pocket which is at the same plane as the silicon reference cell . The a ctive area of our standard device is 0.04 cm 2 , which is much smaller compared to the 3 cm 2 uniform beam spot size. Every time a device is tested, the test pocket is slightly moved so that the device is in the center of the uniform beam spot. Another impor tant parameter is external quantum efficiency ( ), which is used to characterize the behavior of a photovoltaic device responding to solar spectra. The schematic illustration of the EQE setup presented in Figure 1 3.

PAGE 19

19 Figure 1 2. Schematic illustr ation of J V characteristic measurement setup . This system is guided by the ASTM E1021 testing standard. White light generated from a quartz t ungsten halogen (QTH) lamp goes into a monochromator, in which the diffraction gratings split s the white light into monochromatic beam of a specific wavelength. This monochromatic beam is at low intensity (1 2 ). Then the beam is collimated by lenses and chopped by mechanical chopper at frequency of 380 Hz. Finally , the condensing lens focus the beam to a small size. It is crucial to keep the beam size small enough so that it can be completely placed within the active area of tested device. Meanwhile, a white light bias (0.7 to 1 sun intensity) is applied to illuminate the tested device in ord er to mimic the 1 Sun light intensity. It is worth mentioning that t his white light bias is not chopped by the mechanical chopper. The photocurrent generated from the tested device passes through current amplifier to increase the signal/noise ratio. Then t his signal is inputted into lock in amplifier, meanwhile the lock in amplifier uses the frequency of mechanical chopper as reference to isolate the current component solely from the chopped monochromatic beam. The can be calculated by:

PAGE 20

20 (1 5) where I cell the current generated from test device and I det the current generated from silicon detector. S det spectral responsivity of silicon detector. If solar spectrum S( integrating the EQE value at each wavelength: (1 6) Figure 1 3. Schematic illustration of EQE measurement setup . 1.4 Brief Introduction to Perovskite Solar Cells 1.4.1 Basic Properties of Perovskite Crystal Structure Perovskite is named after Russian mineralogist Lev Perovski in 1839 5 . The structure is adopted by CaTiO 3 , and two m ost well studied metal oxide perovskite cases

PAGE 21

21 are BaTiO 3 and SrTiO 3 . Generally, the chemical formula for perovskite compounds is AMX 3 , There are two main types of perovskite. One is metal oxide perov skite type (ABO 3 ), the valence of the two metal cations must sum to 6 such as: KTaO 3 (I V O 3 ), CaTiO 3 (II IV O 3 ) and GdFeO 3 (III III O 3 ). Another type is halide perovskite (ABX 3 ), the valence of the two cations should sum to 3. In addition to inorganic perovskite like CsSnI 3 , A site can also be single charged organic cation such as methylammonium (MA + ), ethylammonium (EA + ) or formamidium (FA + ). M site is a divalent metal cation, typically Sn 2+ or Pb 2+ , and X sit e is a halide anion such as Cl , Br or I . This type of perovskite is usually called organo metal halide perovskite or organohalide perovskite . Figure 1 4. Schematic illustration of c ubic perovskit e crystal structure. Usually, A site is MA + , B is Pb 2+ a nd X is I . Reprinted with permission. 6 An ideal perovskite structure is a cubic crystal structure. As shown in Figure 1 4 6 , A site cation is at the cubic center and in 1 2 fold cuboctahedral coordination, B site cation has 6 fold coordination with 6 X anion and form an octahedron (BX 6 ), meanwhile the A site is also surrounded by 8 BX 6 octahedron 5 . Th e

PAGE 22

22 stability of perovskite structure is related to t he relative sizes of A, B and X site ions. olerance factor t is defined as the ra tio of distance between A and X site t o distance between B and X sit e, and the radi us of B to the radius of X site 6 : (1 7) (1 8) where R A , R B and R X are the rad ii of A, B and X, respectively. Generally, for halide perovskite, both satisfied. If t is between 0.89 and 1.00, it will be rhombohedra or orthorhombic , which is less symmetric; wh en t=1.00, it will be cubic structure , which has the highest symmetry . 1.4. 2 The Emergence of Perovskite Solar Cells Since 2009 Miyasaka and co worker s first introduced organohalide perovskite to photovoltaic device with a PCE of 3.81% 4 , a series of br eakthrough s were made within less than five years . In the year of 2011, the efficiency was nearly doubled to 6. 5% by Park and colleagues 7 . Then one year later in mid 2012, the efficiency exceeded 10%. In 2013, several reports claiming more than 15% PCE was report ed 8 10 . At the beginning o f 2014, 17.9 % was then confirmed by NREL 11 . So far the highest recorded and confirmed efficiency is 20.1% 1 by Korean Research Institute of Chemical Technology ( KRICT ) . Within five years, the efficiency rapidly increases over four times. In addition to the steep rise of high conversion efficiency, the boosting growth of publication s each year is also astonishing. According to the databa se of Web of Science, Figure 1 5 is ential

PAGE 23

23 growth from 2009 to 2014 indicates the attention drawn by this new area of research. Especially, in the year of 2014, more than 400 paper were published, which means about 1.2 papers were published on every single day. Figure 1 5. Rapid growth in perovskite research according to the Web of Science cell s . A) Histogram of published items i n each year from 1996 to 2015. B) Histogram of citations in each year from 2002 to 2015. Several key attributes lead to the rapid development and enormous attention of perovskite solar cell. First one is the s trong light absorption property. T he most popular perovskite solar cell is based on methylammonium lead iodide, CH 3 NH 3 PbI 3 , whi ch has a direct bandgap of 1.55eV with the corresponding absorption onset of 800nm 5 . It is a very good absorber material over the entire visible light spectrum (380 780nm). Its absor ption coefficient is on the order of 10 5 cm 1 , which is much higher than that of silicon or GaAs 6 . Hence, the required thickness of photo active layer is largely reduced, only several hundreds of nanometers of perovskite layer thickness can efficiently absorb the incident light, which is very beneficial for carrier collection. In addition to good absorption ability, perovskite exhibits excellent carrier transport behavior. The carrier mobilities are 7.5 cm 2 V 1 s 1 and 12.5 66 cm 2 V 1 s 1 for

PAGE 24

24 electrons and holes, respectively 12 , and the carrier lifetime is hundreds of nanoseconds, which result s in long carrier diffusion lengths exceedi organo halide perovskite . 13 Carri er diffusion length is defined as the average distance charge carriers can travel before they recombine. Besides the excellent optical and electronic properties, the simplicity of fabrication techniques is another key feature of perovskite solar cells. Se veral fabrication methods were developed with a variety of device architectures. Perovskite active layer can be applied to both planar heterojunction and mesoporous scaffold. The details of fabrication methods and device architectures will be covered in Se ction 1.5.1 and Section 1.5.2, respectively. 1.5 Review of Perovskite Solar Cells 1.5 . 1 Fabrication Techniques of Perovskite Solar Cells As mentioned before, e ase of fabrication as one of the main advantages of organo halide perovskite leads to the versatil ity of deposition methods. So far four main deposition methods have been reported including: one step solution process, two step solution process, dual source vacuum deposition (DSVD) and vapor assisted solution process (VASP) , and the schematic illustrati ons of them are shown in Figure 1 6 5 . Among these above mentioned methods, the first three have achieved power conversion efficiency above 15% 8,9,11 , and t he fourth one also exceeds 12 % 14 . One Step deposition is based on solution process, which is the most popular thin film deposition method for perovskite because of its simplicity. Generally, the precursor solution of perovskite is prepared by mixing PbX 2 (X: I, Br, Cl) powder and RAX powder (R: formamidine, methyl; X: I, Br, Cl). For pure organohalide perovskite precursor, the molar ratio is 1:1; and for mixed halide perovskite precursor such as PbCl 2 and MAI, the

PAGE 25

25 molar ratio is 1:3. S olvents are usually selected among some high boiling point aprotic polar solvents suc h as dimethylformamide (DMF), dimethylsulphoxide (DMSO) or butyrolactone (GBL), etc. In precursor solution preparation procedure, i t usually takes several hours to dissolve and obtain a clear solution at an elevated temperature. The solution is then depo sited o n substrate by spin coating. Precursor solution concentration, spin coating speed, solvent type and other parameters should be ca refully adjusted form the perovskite film . It is worth mentioning that s ome special treatments were reported to improve the perovskite film morphology. Seok and co workers developed a technique called solvent engineering 15 , a drip of toluene at the end of spin coating can lead to extremely uni form and dense perovskite layer. Jen group has improved perovskite film morphology by adding 1% of 1,8 diiodooctane (1,8 DIO) into precursor solution 16 . After spin coating process, an annealing process is often required to remove the residual solvent and form the perovskite phase completely . It is expected some byproduct crystal phase such as MACl will form with in the perovskte phase, however, according to XRD patterns, MACl does not exist inside the perovskite film. Even t hough the precursors have different components and molar ratios, the final crystals of MAPbI 3 and MAPbI 3 x Cl x are identical. Based on some careful elemental analysis, the work done by Peter Erk has recently reported the nonexistence of any chloride doping in MAPbI 3 . Hence, in the future the use of formula MAPbI 3 x Cl x should be avoided 5 . One possible explanation is that MAC l is sublimed during the annealing procedure. An interesting phenomenon is also observed as an evidence to support the MAC l sublimation explanation: usually less than 1 hour of annealing is required for pure iodine perovskite, and as for the mixed halide p erovskite the required annealing time is

PAGE 26

26 elongated to 2 to 3 hours. The function of chloride and other ions will be discussed in Section 1.5.3. The highest confirmed conversion efficiency by one step process is 17.91% 11 so far. Another method base d on solution process is Two Step solution process, some people also call it sequential deposition method (SDM) which was first introduced to perovskite solar cells by Gratzel and coworkers 8 . Typically , the deposition begins with spin coating PbI 2 in dimethyl formamide (DMF) onto mesoporous TiO 2 substrate, then the substrate is im mersed into MAI solution in iso propyl a lcohol (IPA) . The p erovskite phase will form in the immersion process . An annealing procedure is usually required to remove the residual IPA solvent and form the perovskite phase completely. After annealing, spiro OMeTAD as hole transport layer is subsequently deposited onto perovskite by spin coating. Some modifications ha ve been applied to two step process . For example, Huang 17 and So 18 groups deposited the MAI solution onto PbI 2 by spin coating instead of immersion, then anneal to drive the diffusion and reaction of the precursor. In this way, the reac tion between PbI 2 and MAI can be well controlled, and meanwhile effectively avoid the damage to underneath layer in the immersion process. Compared to one step process, better control of crystal formation can achieved by two step process, especially, in th e case that perovskite film is deposited on mesoporous scaffold. However, the two step process is relatively complicated, many variables and conditions such as concentration of MAI solution, immersion time, immersion temperature, whether preheating substra te or not, post an nealing temperature and time need to be adjusted and optimized. The hig hest efficiency achieved by two step process is 15.7% by replacing mesoporous TiO 2 with ZnO blocking layer 10 .

PAGE 27

27 Unlike solution process, the perovskite film can be deposited by thermal evaporation of two source materials simultaneously . This method is called Dual Source Vacuum Deposition (DSV D) 9 , in 2013 Snaith group first applied this method to prepare the mixed halide perovskite thin film in a planar device structure. In general, a substrate is placed and faced downward at the top of the vacuum chamber, MAI and PbCl 2 as dual sources are placed in two boats on the bottom of the vacuum chamber. By resistive ly heating two boats, MAI and PbCl 2 will sublime at a controlled rate. Both source materials vapor will be deposited onto the substrate and form the perovskite phase. It was reported that vapor deposited perovskite film is extremely uniform with few voids left, which leads to the absence of pinholes with the high efficiency of 15.4%. Apparently, the disadvantage is very obvious, large amount of power is c onsumed to hea t these sources, and high vacuum is also required , and both of them will not be suitable for large scale commercialization . A novel low temperature approach to deposit perovskite layer is developed by Yang and his colleagues which is called the Va por Assis ted Solution Process (VASP) 14 , i nspired by both the two step solution process and dual source vacuum deposition. Generally, the first step is to spin coat PbI 2 onto compact TiO 2 which is coated on fluorine doped tin (FTO) oxide glass substrate. The substrate is preheated to remove the residual DMF solvent. Then the PbI 2 coated substrate is annealed in MAI vapor at 150°C for a 2 hours to form perovskite film. The MAI vapor is mixed in nitrogen atmosphere. 100% coverage with superior uniform surface c an be produced by this approach, and a decent efficiency of 12.1% can be realized. This method is very promising for commercialization in the future due to its good control of the film

PAGE 28

28 morphology and high performance , as wel l as its ease of fabrication . The reaction rate between PbI 2 and MAI is effectively reduced, which is exactly the disadvantage of the immersion procedure in conventional two step process. Figure 1 6 . General fabrication metho ds to deposit perovskite film. A) One Step Precursor Solution Process . B) Two Step Solution Process. C) Dual Source Vacuum Deposition. D) Vapor assisted Solution Process. Reprinted with permission. 5 It is believed that application of n ovel device structures and introduction of new materials will stimulate novel fabrication technology of perovskite, and as a feedback, meanwhile the new fabrication methods can also lead to excellen t film quality and high performance of perovskite solar cells . 1.5 . 2 Device Structures of Perovskite Solar Cells The development of device structure is concomitant with the ascension of perovskite solar cells performance. In the spring of 2009, Miyasaka an d co workers first introduced organohalide perovskite as sensitizer material to dye sensitized solar cells. In this work, 3.81% conversion efficiency was achieved 4 . In the year of 2011, the group of Nam Gyu Park applie d the similar device structure and modified the TiO 2 surface

PAGE 29

29 treatment, a 6.5% efficiency was realized 7 . However, in both of the pioneering work, the liquid electrolyte would rapidly dissolve the perovskite sensitizer, therefore device stability was very poor. A significant progress was made by replacing the liquid electrolyte with a novel solid state hole transport materials (HTM) , which was originally designed for LEDs . In 2012, Park and Gratzel introduced 2,2 ,7,7 ,b tetrakis (N,N dimethoxyphenyl amine) 9,9 spirobi fluorene (spiro OMeTAD) to pure organohalide perovskite MAPbI 3 . The electron transport lay er mesoporous TiO 2 (m TiO 2 ) was coated by perovskite materials, which provided a scaffold for the growth of perovskite extremely thin absorber (ETA) 19 . The degradation problem was successfully sol ved and the efficiency was further improved to 9.7%. Then in the year of 2013, Snaith group replaced n type m TiO 2 scaffold with Al 2 O 3 scaffold 20 . Perovskite layer and spiro OMeTAD were spin coated as light harvesting layer and hole transport layer, respectively. Two key features in this work need to be mentioned. Firstly, the formula CH 3 NH 3 PbI 3 x was used for mix ed h alide perovskite by Snaith. Although later it was proved the perovskite crystals were identical for both pure and mixed halide precursors, the adding of Cl did increase the device efficiency by improving the formation of perovskite crystal. Secondly, d ue to its wide band gap, Al 2 O 3 is unable to assist electron extraction, which suggests that the perovskite layer itself is able to transport electron to the electrode. This indicates that in addition to light harvesting the perovskite materials can also be have as an n type semiconductor. Hence, the m TiO 2 is no longer necessary for perovskite solar cells, but

PAGE 30

30 a compact TiO 2 (c TiO 2 ) layer is still needed for hole blocking purpose . The conversion efficiency of 10.9% was reported in this work. The following work done by Seok and Gratzel groups further pushed the efficiency to 12.0% 21 . They made two modifications to the device structure. Firstly, a continuous perovskite capping layer (they named pillared structure) was deposited on the top of mesoporous scaffold layer, which indicates that the perovskite can also act as a hole transp ort layer. Secondly, instead of extremely thin absorber (ETA) only coated onto the surface of m TiO 2 scaffold, the m TiO 2 was fully infiltrated by perovskite materials (they named 3D TiO 2 /CH 3 NH 3 PbI 3 composites). Hence, the light harvesting is highly enhanc ed so that the mesostructure thickness could be reduced to 200 300nm, which is very favorable for carrier transport. Besides device structure modification, they also examined different polymeric hole transport materials ( HTMs ) , and they found that better F F and Voc could be achieved by using poly triarylamine (PTAA) than using spiro OMeTAD. Inspired by the fact that perovskite is able to play the role of hole transport layer in addition to light absorber layer, two intuitive expectations were brought. The f irst one is whether an HTM free device can be fabricated , and the answer is p ositive. Early before in 2012 , Gratzel and co workers directly deposited gold on top of TiO 2 mesostructure without HTM in between, the efficiency was only 5% 22 . Then Han group used double layer of TiO 2 and ZrO 2 as scaffold and carbon as anode, a 12.8% confirmed efficiency was reached 23 . This novel HTM free structure indicates that perovskite is an ambipolar material which can serve as both electron transport layer and hole transport layer. The second expectation is to fully eliminate the scaffold so that a

PAGE 31

31 planar structure can be realized. In May 2013, Snaith and his colleagues deposited perovskite directly on compact TiO 2 by dual source thermal evaporation. This s imple planar structure had no scaffold structure, and the device efficiency was increased to 15.4% 9 . This innovative work paved the way towards the application of planar structure in perovskite solar cells . The absence of scaffold reduces the complexity of device architecture, which turns over a new leaf in perovskite solar cell research. The combination of solution process and planar structure lead to the simplicity of both fabrication method and device arch itecture. Kelly group replaced conventional fluorine doped tin oxide (FTO)/compact TiO 2 combination with a thin ZnO nanoparticle layer, and an efficiency of 15.7% was reported 10 . The perovskite layer was deposited by two step solution process in this work. Yang group fabricated a planar heterojunction solar cells via vapor assisted solution process, and got full perovskite film surface coverage without any voids, an impressive 12.1% PCE was achieved 14 . Almost at the same time, Yang group used organic materials poly (3,4 ethylenedioxythuiphene):polystyrene sulfonate (PEDOT:PSS) and [6,6] p henyl C 61 butyric acid methyl ester (PCBM) as ho le and electron transport layer , respectively 24 . The perovskite absorber layer was fabricated by one step solution process. Compared to the efficiency of 11.5% in rigid substrate device, due to the low temperature in one step process, the flexible device is able to e xhibit power conversion e fficiency of 9.5%. Besides the application of organic carrier and planar geometry, this work demonstrated a so called inverted cell structure (organic photovoltaic groups often call this structure normal structure , but dye sensitized solar

PAGE 32

32 cell groups ten d to call it inverted structure), in which light firstly passes through the hole transport layer instead of electron transport layer. In addition to the progress in various device structures, selection of carrier transport material is also one of the most popular topics. People mainly focus on the replacement of the most used spiro OMeTAD due to its high cost. Organic materials such as poly triarylamine (PTAA) 21 , poly[N 9 hepta decanyl 2,7 carbazole alt 3,6 bis (th iophen 5 yl) 2,5 dioctyl 2,5 di hydropyrrolo[3,4 ]pyrrole 1,4 dione] (PCBTDPP) 25 and 3,4 ethylenedioxythiophene (H 101) 26 , i norganic materials such as CuI 27 and CuSCN 28 are some candidate choices. On the other hand , several electron transport materials have b een checked, ICBA/ C 60 /PCBM 29 , graphene oxide 30 , rutile TiO 2 31 and ZnO nanoparticles 10 are some reported examples. 1.5 . 3 Bandgap Modification by Ionic Replacement Generally, the optimal bandgap of light harvesting layer should have broad and strong absorption over visible to near infrared solar spectrum, the bandgap no more than 1.1 eV is suitable for light absorption. However, the built in electric fie ld is also determined by bandgap, too low intensity can lead to bad carrier transport. The optimal bandgap to balance these two competing effects is around 1.4 eV 5 . Hence, the bandga p engineering is the key to achieve the required bandgap for the optimal solar cell performance. As the formula indicates, there are three ionic sites in halide perovskite ABX 3 crystal. According to the first principle study, the energy band structure (val ence band maximum and conduction band minimum) is determined by B X bond 5 . By varying the combination of the se three ionic components, B X bond can be modified due to the change of i onic size and hybrid state, therefore the bandgap of perovskite can be tuned.

PAGE 33

33 A site is occupied by a single charged cation. Larger ionic size can cause the lattice expansion and smaller ionic size can result in lattice contraction, which affect s the length of B X bond length. The BX 6 4 octahedron network can also be deformed by changing A site cation size, which affects energy band structure of perovskite crystal. If the B site metal cation and X site halide anion are both given, the size allowance fo r A site is small since the perovskite lattice has to maintain the tolerance factor in a required range, otherwise the perovskite structure is no longer favorable and other crystal structure will be formed. Theoretically, the radius of A site should be les s than 2.6 Ã… if the lattice maintains cubic symmetry (t= 1). Figure 1 7A shows several A site cation candidates which have been checked to form perovskite crystal 32 . Cesium (Cs + ), methylammonium (MA + ) and form am id ini um (FA + ), their effective ionic radius have the trend R Cs +
PAGE 34

34 direction, the covalent character of M I bond decreases and ionic character increases, because of the increasing electron negativity between the met al cation M 2+ and I . Hence, the bandgap trend is AGeI 3 < A SnI 3
PAGE 35

35 VII from Cl to I due to the increasing covalent character of Pb X bonding 5 . The energy bandgap of perovskite can be continuous ly tuned by v arying the combination of two incorporated halide. For example, the energy bandgap of MAPbI x Br 3 x can be adjusted from 2.3 eV (x=0) to 1.55 eV (x=3), and the corresponding absorption edge is from 540nm to 800nm, which covers the entire visible solar spectr um. Table 1 1. Ionic radius of some A site, B site and X site ions in metal halide perovskite . Ionic replacement can not only modifies perovskite energy bandgap, but also affects the film quality and stability. As mentioned before, in solution process mixed halide precursor shows more controllable formation and better surface coverage of perovskite film than its pure halide equivalent. In the recent work of Seok group 11 , a simultaneous introduction of both FA + and Br to the conventional MAPbI 3 with the composition al formula of (FAPbI 3 ) 1 x (MAPbBr 3 ) x . They found that the co substitution of FA + to MA + and Br to I can stabilize the perovskite phase due to the synergetic effect. The optimized composition (x=0.15) exhibited an excellent power conversion efficiency of 17.91%, which is the highest confirmed record in published literature. 1.6 Outl ine of the Thesis This work mainly focuses on fabricating high efficiency perovskite solar cell with good stability. The basic concepts and characterization of solar cells along with a brief review of recent progress in perovskite photovoltaics are covered in Cha pter 1.

PAGE 36

36 In Chapter 2, a high efficiency perovskite solar cell by solution process is introduced, as well as the fabrication (Section 2.3) and characterization details (Section 2.3). The device performance comparison of different electron transport m aterials (ETMs) and hole transport materials (HTMs) is presented in Section 2.4. Then, m odification of perovski te layer thickness for enhancing light absorption is covered in Se ction 2.5. A method to avoid short circuit current density overestimation (Section 2.6) is developed in Section 2.7. In the end a summary is provided in Section 2.8. Chapter 3 focuses on the exploration of device degradation mechanism. Degradation happens due to air exposure (Section 3.2), and a series of experim ents are designed in Section 3.3 to find out the causes. Characterization results including XRD and XPS are provided in Section 3.4. A long operation lifetime of our perovskite photovoltaic device is shown in Section 3.5 and a summary is provi ded at the en d in Section 3.6. Finally, Chapter 4 summarizes the conclusions of the whole work and gives some prospects in the future work.

PAGE 37

37 CHAPTER 2 HIGH EFFICIENCY PEROVSKITE SOLAR CELL FABRICATION BY SOLUTION PROCESS 2.1 Introduction Due to its obvious simplicity, solution process is the most widely used method for perovskite thin film deposition. As one of the two main forms of solution process, one step precursor deposition has the advantage of ease of fabrication over two step sequential deposition. In this chapter, one step precursor method is used to fabricate a high efficiency perovskite solar cell. In Section 2.2 the details of MAI synthesis (Section 2.2.1) and device fabrication (Section 2.2.2) will be presented. Device characterization results including XRD, SEM, J V characteristic and EQE measurement are covered in Section 2.3. Device performance with d ifferent hole transport materials (Section 2.3.1) and electron transport materials (Section 2.3.2) is compared in Section 2.4. To enhance the photovoltai c performance by improving light absorption, we increased perovskite layer thickness and the result is summarized in Section 2.5. However, there is a mismatch between the Jsc extracted from J V characteristic and Jsc value integrated from EQE measurement, Section 2.6 introduces J V hysteresis effect and the cause of inflated Jsc. A method to accurately characterize the device performance is developed in Section 2.7. Finally, a summary of this chapter is in Section 2.8. 2.2 Experimental Details of Device Fab rication 2.2.1 Methylammonium Iodide Synthesis Methylammonium halide compounds are usually prepared by the recombination reaction between equimolar amounts of methylamine and halide acid. The

PAGE 38

38 Methylammonium Iodide (MAI) in our lab is synthesized by referri ng to the method of Park Group 36 . CH 3 NH 2 3 NH 3 I (2 1) 10mL of methylamine (CH 3 NH 2 ) (33 wt% in ethanol) and 10mL of hydroiodic acid (57 wt% in water) react at 80°C with constant stirring in a round bottomed flask for an hour. After it cools down to room temperature, add 30mL ether into the flask to wash and stir for 30 minutes. Repeat the washing step for t hree times to purify the MAI. Then keep the resulting solution under vacuum at 100°C for 24 hours and the synthesized CH 3 NH 3 I is produced. 2.2 .2 Device Fabrication The device structure layout and energy diagram are shown in Figure 2 1A and Figure 2 1B, re spectively . Firstly, the patterned Indium Tin Oxide (ITO, or tin doped indium oxide) glass substrates a re sonicated in Deionized (DI) Water + Soap, DI Water, Acetone and Isopropyl Alcohol (IPA) for 15 minutes each. Before we start the device fabricati on, t he substrates are under UV ozone treatment for 15 minutes. This procedure could not only effectively remove the organic residuals in sonication cleaning process but also increase the work function of ITO electrode. Our standard recipe of device fabrication starts with spin coating 40nm PEDOT: PSS onto the UV treated ITO substrate at 5500rpm. The PEDOT: PSS film is annealed at 150°C in air for 15 minutes to remove water solvent . After the PEDOT: PSS film, perovskite precursor is spin coated at 2000 rpm to fo rm the light absorber layer. We dissolve PbCl 2 and MAI in DMF as the precursor (PbCl 2 : MAI=1:3 in molar ratio and stirred for 24h at 60°C before use). The concentration of the perovskite precursor solution for our standard recipe is 0.44M and the film thic kness is 200nm. The n, the

PAGE 39

39 substrate is annealed at 100°C in glove box for 2 hours to comple tely form the perovskite phase. 20mg/mL PCBM solution in chlorobenzene is deposited onto perovskite layer by spin coating at 1 000rpm. Right after spin coating , the P CBM film is annealed at 100°C for 15 minutes to remove the residual chlorobenzene. A 30nm zinc oxi de (ZnO) film is formed by spin coating 30mg/mL ZnO nanop articles in ethanol and annealing for 15 minutes to facilitate the evaporation of the solvent. Finall y, a 100nm Al electrode is deposited by thermal evaporation in vacuum chamber. The defined active area of our device is 2×2 mm 2 . Figure 2 1. Perovskite solar cell device architecture and energy diagram. A) Schematic of perovskite solar cell layout. B) Energy diagram of perovskite solar cell . 2.3 Device Characterization We compare the X ray diffraction pattern of our perovskite film with both solution processed an d vapor deposited in literature . According to the literature 9 , the main diffraction peaks assigned to (110), (220) and (330) are at 14.12°, 28.44° and 43.23°, respectively. Our perovskite main peaks are found at 14.10°, 28.43° and 43.29° assigned to (110), (220) and (330). The peak positions of literature and our result are almost identical. In additi on to these main peaks, there are no peak s at 12.65° and

PAGE 40

40 15.68°, which are the (110) diffraction peaks for PbI 2 and (110) diffraction peak for CH 3 NH 3 PbCl 3 , respectively, indicating the high purity of perovskite phase. According to Debye Scherrer Equation, we can estimate the perovskite domain size from its XRD spectrum: (2 2) where T is the mean size of the tested crystalline domains. K is called Scherrer Constant, which is a dimensionless shape factor with a typical value of about 0.9. is the incident X ray wavelength, which in our case is about 1.54 Å and is the line broadening at half the maximum intensity. By plugging the data of (110) peak, we can estimate that the domain size is between 65 to 75 nm. Figure 2 2 . X ray diffraction (XRD) spectra of perovskite film. A) X ray diffraction spectra of a solution processed perovskite film (blue) and dual source vacuum evaporation (red) from literature . Reprinted with permission. 9 B) X ray diffraction spectra of our perovskite film by one step solution process. C) (110) peaks at 14.10 °. D) (220) peaks at 28.43°. E) (330) peaks at 43.29°.

PAGE 41

41 Then we check the perovskite film su rface morphology by scanning electron microscope (SEM). As shown in Figure 2 3 , a smooth morphology is achieved by one step solution process. The film is very uniform with high surface coverage. However, some small voids exist , and the size of these pores is around 200 to 500nm. Figure 2 3 . Scanning Electron Microscope (SEM) Images of the same perovskite film by one step solution process. A) Scale bar of 10 m . B) Scale bar of 2 m. C) Scale bar of 5 m. D) Scale bar of 1 m . The J V Characteristic is tested in air by an Agilent 4155C semiconductor parameter analyzer. The voltage bias scan is applied fro m 1 V to 1.2 V for forward scan and from 1.2 V to 1 V for backward scan. Each step in both scans is 0.01V and the integrat ion time per point is 0.05 s. The performance of our standard device is shown in Figure 2 4A . The forward scan and backward scan merge together, and no hysteresis is observed. Specifically, the Jsc is 17.5 mA/cm 2 , the Voc is 0.97 V and the FF is 73%. The power conversion

PAGE 42

42 efficiency is 12.4%. Moreover, our device exhibits high reproducibility and relatively small variation. The average efficiency ranges from 11.5% to 12.5%, with the Voc between 0.95 and 1.01V, the Jsc varie s from 17.2 to 18.5 mA/cm 2 and FF ranges from 65 to 74%. The high performance is due to both good morphology and h igh purity of perovskite layer. Figure 2 4. P hotovoltaic performance characterization of standard perovskite solar cell . A) J V characteristic including forward scan (solid blue) and backwa rd scan (dash red). B) External quantum efficiencies without white light bias (blue) and with white bias (red). The concentration and spin coating rate of PCBM solution is carefully adjusted in this standard device fabrication recipe for an optimized PCBM thickness, because the thickness of PCBM layer is the key to achieve a good photovoltaic performance. If the PCBM layer is too thick, the low conductivity of PCBM will result in high series resi stance, which is not favorable for electron transport ; however, on the other hand, if the PCBM layer is too thin, it cannot fully cover the underlying perovskite light absorber layer, and ZnO layer will directly contact with perovskite layer. In ZnO anneal ing process, ethanol as the solvent of ZnO nanoparticle will react with perovskite and turn phase 37 , which leads to the failure of the device.

PAGE 43

43 The corr esponding external quantum effi ciency is shown in Figure 2 4B . The standard device shows a high EQE and its peak is close to 80%. The absorption edge is approximately 800nm, which matches with the 1.55 eV bandgap of CH 3 NH 3 PbI 3 . The EQE is tested under both the conditions of without and with white light bias, and EQE with bias is slightly lower than EQE without bias. The integrated Jsc without bias and without bias are 16.2 mA/cm 2 and 15.9 mA/cm 2 , respectively. This result indicates that the recombination process of carriers within our device is very low and the photo generated carriers are efficiently collected by electrode. There is a mismatch between Jsc extracted from J V characteristic and that integrated from EQE measurement, which indicates an inflat ion existing in the device . The detail of Jsc calibration will be discussed in Section 2.6. 2.4 Candidates for ETM and HTM 2.4.1 ETM Selection Several electron transport materials have been examined including [6,6 ] phenyl C 61 butyric acid methyl ester (PCB M), [6,6] phenyl C 7 1 butyric acid methyl ester (PC 71 BM) and indene C 60 bisadduct (ICBA). All of these electron transport materials are dissolved in chlorobenzene with the same solution concentration of 20mg/mL, and spin coated onto perovskite layer at 100 0 rpm, then we anneal the substrates at 100°C for 15min to remove the residual chlorobenzene in glove box . Other layers are depos ited under the same condition as the st andard recipe of device fabrication in Section 2.2.2. The J V characteristic of devices with different ETM s is shown in Figure 2 5A . The similarity of transmittance suggests that the perovskite layer thickness is very close to eac h other as shown in Figure 2 5C . The device performance is summarized in Table 2 1. Among them the PCBM presents a superior photovoltaic performance to ICBA and

PAGE 44

44 PC 71 BM. Short circuit current density integrated from EQE measurement matches with the trend ( Figure 2 5 B). Specifically, for ICBA there is a huge difference between EQE with and without white light bias, this means when t he incident radiance is near 1 s un, carrie r recombination occurs due to the high photo generated carrier concentration and inferior electron transport in ICBA. As f or PCBM and PC 71 BM scenarios, two EQE spectra of two conditions match es very we ll, hence, the carrier recombination can be neglected then . Figure 2 5. Perovskite solar ce ll with different electron transport materials (ETMs) and hole transport materials (HTMs) . A) Current density Voltage ( J V ) characteristic comparison of different electron transport materials. B) External quantum efficiency comparison of different electron transport materials. C) Transmittance comparison of different electron transport materials. D) Current density Voltage (J V) characteristic comparison of different hole transport materials.

PAGE 45

45 Table 2 1. Device performance characteristics of different electron transport materials. 2.4.2 HTM Selection In addition to the selection of ETM, we have also tried several hole trans port materials (HTMs) in our perovskite solar cell device. Since the perovskite layer is deposited onto hole transport layer in our normal structure, the morphology of the hole transport layer is the key to the quality of perovskite layer. As shown in Fig ure 2 5D , among all these candidates, only PEDOT: PSS is suitable for achieving a good perovskite film qual ity. All the other HTMs will lead to inferior perovskite film morphology, which results in bad device performance. In all these ETM and HTM candidates, PCBM and PEDOT: PSS is the optimal combination. Our device shows high performance, suggesting that the electron hole transport is efficient and well balanced. Compared to the similar device structure in literature 24 , ours exhibits much higher Voc (0.95 to 1.01V vs 0.80 to 0.88V). This is due to the exist ence of ZnO nanopartic le layer. ZnO has a suitable energy levels and high electron mobility which improve s the electron transport between PCBM and Al electrode. Moreover, the existence of ZnO can effective ly avoid the contact between ETM and HTM. Hence, the leakage current and induced recombination is dramatically minimized, and as a result, high open circuit voltage near 1V can be realized successfully.

PAGE 46

46 2.5 Perovskite Layer Thickness Optimization Based on the same device architecture, to further improv e the photovoltaic perfo rmance of our device, we tried to increase the perovskite layer thickness for better light absorption , which would lead to an enhanced photocurrent . There are two main ways to increase the layer thickness in spin coating process, one is to increase the so lution concentration and the other is to decrease the spin coating rate. Since perovskite morphology is the key to good performance, the film quality has to be guaranteed as the film thickness increases. Lab experience indicates that if the spin coa ting ra te is below 2000rpm, the resultant perovskite film is very rough, therefore the PCBM layer is not able to fully cover the perovskite layer, and the ethanol in ZnO nanoparticle can contact with per ovskite layer and cause a phase change from perovskite phase to . On the other hand, if the perovskite precursor concentration is too high, a bad film morphology can also be made due to excessively high precursor solution viscosity . Therefore, careful adjustment is required to obtain a controlled increase of perovskite film thickness . Figure 2 6. Photovoltaic performance characterization of perovskite solar cell with different perovskite layer thickness. A) J V characteristic comparison of differe nt perovskite layer thickness. B) Transmittance comparison of different perovskite layer thickness.

PAGE 47

47 Four different combinations of spin coating rate and perovskite precursor solution concentration have been applied in order to obtain a thicke r perovskite layer. Figure 2 6 A shows the J V characteristi c of these four spin coating conditions : 0.44M+ 2000rpm , 0.66M+6000rpm, 0.66M+4000rpm, 0.66M+2000rpm. In Table 2 2 the corresponding photovoltaic performance is presented. Table 2 2. Photovoltaic performance parameters of devices with different perovskite l ayer thickness. As we expected, as the spin coating rate decreases and solution concentration increases, the Jsc increases. At high spin coat rate from 6000rpm to 4000rpm, the Jsc enhancement is relatively small; at low spin coat rate from 4000rpm to 2000rpm, the enhancement is strong, and this phenomenon agrees with the fact that the spin coating rate should be controlled above 2000rpm to maintain a good film morphology, because change within low spin coat rate can dramatically affect the film thickness and quality . The transmittance comparison in Figure 2 6B concurs with the trend of film thickness: transmittance of 0.66M+6000rpm and 0.66M+4000rpm almost merge from 300nm to 800nm, however, once the spin coa ting rate drops to 0.66M+2000rpm, the absorption is enhanced dramatically. The thickness of perovskite layer s is also measured by profilometry. By depositing perovskite precursor solution of different concentration and spin speed combinations on ITO/PEDOT: PSS substrates, and ITO/PEDOT: PSS is used as reference. The result of perovskite layer thickness matches very well with the

PAGE 48

48 transmittance test. For precursor solution concentration of 0.66M, the thickness of 4 000 rpm and 6000 rpm is very close to each ot her. As f or 2K, the film thickness increases dramatically to nearly 300nm. Compared to 0.44M+2000rpm standard device, power conversion efficiency is improved by 30% due to improved light absorption. On the other hand, Voc and FF remains almost the same fo r these four different conditions, indicating other parameters in this experiment are very well controlled. Thus, the power conversion efficiency is proportional to short circuit current density. 2.6 Causes of Inflated Results More than 400 papers based o n perovskite solar cells were published within results were reported. To avoid these inflated results , an accurate measurement standard is in urgent need [emergence]. Generally, three main aspects can cause an inflated result. Firstly, due to high batch to batch variation, an overenthusiasm often happens based on some specific extremely high performance counts, however, the yield is actually low. Secondly, a strong hysteresis eff ect may exist in J V characteristic, so that the device performance strongly depends on scan direction (forward vs backward) and scan speed (the integration time of each step) 12 . It has been reported that this anomalous dependence behavior can be effectively suppressed by reduci ng the voltage bias scan speed 11,15 . Usually, the reverse scan exhibits higher Voc and FF compared to forward scan, but the mechanism is not clear so far. Thi rdly, in some cases depending on the device structures and materia ls, the short circuit current can be overestimated, which causes an inflation in power conversion efficiency c alculation.

PAGE 49

49 Figure 2 7. Histogram of device power conversion efficiency under same fabrication condition To check the device yield, 80 standard devices in the same fabrication condition (0.44M 2K) have been fabricated , and the statistic result is show n in Figure2 7. As we can see, the count of devices matches well with the Gaussian fit curve. Most devices fall in the PCE range of 11% to 13%, suggesting a good reproducibil ity in our fabrication process. As for the concern of hysteresis, four devices (A , B, C and D) with same perovskite layer thickness (0.66M+2000rpm) have been tested in this experiment. The J V measurement are carried out with two scan directions (forward and backward) and three sets of integration times (short scan: 0.01 s/step, medium scan: 0.05 s/step and long scan: 0.1s /step). The J V curves of this experiment are shown in Figure 2 8 and the corresponding photovoltaic performance is summarized in Table 2 3. It turns out that the variation of our device is very small even by short sc an. As the time delay on each step increases, the J V curve becomes smoother. Moreover, the difference between forward scan and backward scan also drops. In general, Jsc and FF of backward scan is slightly higher than that of forward scan, and this phenome non agrees with the reports of Seok group 11,15 . The offsets of Jsc, Voc, FF and PCE

PAGE 50

50 between forward scan and backward scan are well below 3% in both medium and long scans, which suggests that medium scan has sufficient integration time to effectively avoid the anomalous behavior. F igure 2 8 . J V characteristic for differe nt scan speeds and directions. A) Short scan (0.01 s/step). B) Medium scan (0.05 s/step). C) Long scan (0.1 s/step).

PAGE 51

51 Table 2 3. Photovoltaic performance parameters for different scan speeds and directions. However, as we mentioned before , there is a noticeable mismatch between Jsc extracted from J V characteristic and Jsc integrated fro m EQE measurement, and this discrepancy should not be neglected for an accurate photovoltaic performance characterization. In EQE measurement, we noticed that even the monochromatic beam is outside the defined device active area, current signal can still be generated by lock in amplifier , as long as the beam is moved horizontally along the Al electrode, and the corresponding EQE has very similar shape as th e scenario with beam inside the defined active area. We consider that the actual size of our device area is larger than the defined one, which results in a short circuit current density overestimation. To prove our assumption , we move EQE beam along the Al electrode and record the calculated Jsc. Figure 2 9A schematically illustrates the design of this EQE beam linear scan experiment. The EQE beam has the size of 0.5mm in width and 1mm in length, which could provide a good resolution to differentiate the be am position along the scan direction x.

PAGE 52

52 Figure 2 9 . EQE beam linear scan experiment. A) Schematic illustration of EQ E beam linear scan experiment. B) Integrated Jsc value from EQE measurement of device with PED OT: PSS in different x position. (Note: dash line is selected inside the defined active area and solid is selected outside the defined active area . ) C) Normalized EQE of device with PEDOT: PSS selected from di fferent x position. D) Integrated Jsc value from EQE measurement of device w/o PEDOT : PSS in different x position. E) Normalized EQE comparison of devices with and without PEDOT: PSS . (Note: Both are selected inside the active area.) As s hown in Figure 2 9 B , the integrated current density has a plateau of ~18mA/cm 2 within 3mm width, and o utside this plateau the current density keeps dropping as we move the beam away along x direction. A tail as long as 8mm is

PAGE 53

53 demonstrated clearly, suggesting that even outside the defined area, holes generated in perovskite layer can still be transported in PEDOT: PSS layer and collected by ITO electrode. Figure 2 9 C shows the normalized EQE curves in different x positions, in which dash lines indicate EQE measured inside the defined device active area and solid lines indicate EQE measured outside the define d device active area. Despite a slight difference between the solid and dashed lines, which is from light absorption of ITO electrode , their normalized EQE curves have almost the same shape, suggesting that all these EQE originate from the spectra l respons e of perovskite layer. As a contrast, we fabricated the device without PEDOT: PSS and all other layers were exactly the same. Even though the bad morphology of perovskite layer causes a bad photovoltaic behavior, however, the relationship between integrate d Jsc and beam position x is the usefu l information we try to learn f r o m Figure 2 9 D . The plateau in device without PEDOT: PSS is about 2mm, which equals the width of ITO. Once we move the incident beam out of ITO electrode, the signal can no longer be det ected, suggesting no carrier can be collected by ITO electrode. If we plot the normalized EQE of devices with and without PEDOT: PSS together in Figure 2 9 E , two devices exhibit almost the same EQE shape, the slight difference is due to the absorption of I TO. From this EQE beam linear scan experiment, we expect the conductivity of PEDOT: PSS is highly increased, as a result , the actual device area is expanded compa red to the defined device area, therefore the short circuit cu rrent density is overestimated. To obtain a direct evidence of PEDOT: PSS conductivity change, we tested the resistivity of PEDOT: PSS layer before and after perovskite layer deposition. The schematic illustration of this resistivity testing experiment is in Figure 2 10 A , an ITO gap

PAGE 54

54 on glass substrate is etched by Aqua Regia. PEDOT: PSS is spin coated on substrate at spin speed of 5500rpm and anneal ed for 15 minutes at 150°C in air, therefore same thickness of PEDOT: PSS film is assumed to be deposited as in our standard perovskite so lar cell device. Two stripes of Al are deposited on ITO near two substrate edges for good electrical contact by thermal evaporation. The contact resistance and ITO resistance is negligible compared to that of PEDOT: PSS layer. Two alligator cl ips (red and black in Figure 2 10 A are used to connect two substrate Al stripes to se miconductor parameter analyzer. Figure 2 10 . Current Voltage characteris tic experiment for PEDOT: PSS layer. A) Schematic illustration of I TO pattern and contact design. B) Curren t Voltage characteristics for PEDOT: PSS layer. Table 2 4. Resistivity of PEDOT: PSS, PEDOT: PSS/Perovskite and PEDOT: PSS/Perovskite/PCBM/ZnO. Two other structures are also fabricated, which have the structure of PEDOT: PSS/Perovskite and PEDOT: PSS/Pe rovskite/PCBM/ZnO , respectively . Here, the

PAGE 55

5 5 deposition of PEDOT: PSS layer, perovskite layer, PCBM layer and ZnO layer all follow s the same method in standard recipe of device fabrication mentioned in S ection 2.2. The resistivity of these substrates is collected in Table 2 4. After the deposition and annealing process of perovskite layer, the resistivity dramatically drops to only 1% of what it used to be, and the deposition of PCBM and ZnO layer will not change the resistivity any more . PEDOT: PSS is co mposed of PEDOT and PSS. Here, PEDOT is hydrophobic and conductive, and PSS is hydrophilic and insulating 38 . N,N dimethylformamide (DMF) in perovskite p recursor solution is a polar solvent, its high dipole momen t magnitude can strongly affect the PEDOT : PSS electrical conductivity. The degree of resistivity reduction highly depends on how well the polar solvent can be dispersed into PEDOT segments, which creates conductive PEDOT pathways for hole transport , and effectively reduces hole loss as shown in Figure 2 11 A and B . Specifically , in our case, residual DMF solvent still stays in the as deposited perovskite layer, and in the 2 hours annealing process D MF solvent will not only be vaporized by elevated temperature of 100°C , but also diffuse into PEDOT : PSS layer and modify PEDOT: PSS nanostructure. As a result , the conductivity of PEDOT: PSS is largely increased. This simple experiment proves that PEDOT: PSS conductivity is increased by a hundred times due to the DMF solvent in perovskite precursor, which should be responsible for the expansion of device active area and the concomitant overestimation of short circuit current density.

PAGE 56

56 Figure 2 11 . Schematic illustration of carrier transport in PEDOT: PSS layer. A) PEDOT: PSS only. B) PEDOT: PSS with polar solvent. Reprinted with permission. 38 2.7 Short Circuit Current Density Calibration Since we have found out the cause of short circuit current density inflation, a method is required to accurately calibrate the Jsc in order to obtain a reliable photovoltaic performance report . We applied apertures to block extra light and treat this apert ure size as our device active area. Three aperture sizes including 2mm, 1mm and 0.5mm diameter circle have been used in this calibration process. Figure 2 12 A shows the J V characteristic of different aperture sizes , and EQE of the same device ( 0.66M+2000r pm) is in Figure 2 12 B . In Table 2 5 the parameters of different aperture sizes are compared. As the aperture size drops, since the amount of incident light decreases, Voc decreases from 0.99V to 0.92V. The short circuit current density increases monotonic ally as aperture size decreases, this may due to the error of smaller aperture size. FF increases at first to its peak 78% with 1mm diameter aperture , and then drops to 72% with 0.5mm diameter aperture . When aperture diameter of 2mm is applied, the short c ircuit current density matches with Jsc integrated from EQE measurement with offset below 4%, which meets the requirement of accurate photovoltaic performance characterization. Therefore t he highest calibrated power

PAGE 57

57 conversion efficiency of our solution pr ocessed perovskite solar cells can reach 14.1% as shown in Table 2 5 when 2mm diameter aperture is used . Figure 2 12 . Aperture selection for short circuit current density ( Jsc ) calibration. A) Current density (J V) characteristics for different a perture sizes . B) External quantum efficiency (EQE) of the same device as a function of the wavelength. Table 2 5. Photovoltaic performance parameters for different aperture sizes. Thus, we choose 2mm diameter aperture to calibrate our devices with dif ferent perovskite layer thickness. The J V curves and EQE curves are in Figure 2 13 A and Figure 2 13B . Now the Jsc form J V measurement matches with that calculated from EQE measurement very well, so the corresponding device performance is reliable with n egligible inflation . Table 2.6 summarizes the photovoltaic performance as a function of different perovskite layer thickness.

PAGE 58

58 F igure 2 13 . Photovoltaic performance characterization of perovskite solar cell with different perovskite layer thickness by ap plying 2mm diameter aperture. A) Current density (J V) characteristics of the devices with different perovskite layer thickness . (Note: 2 mm diameter aperture is applied.) B) External quantum efficiency (EQE) of the devices with different perovskite layer thickness. Table 2 6. Photovoltaic performance parameters of devices with different perovskite layer thickness . (Note: 2 mm diameter aperture is applied for J V characteristic test . ) 2.8 Summary In Chapter 2, one step solution process was applied to fabricate perovskite solar cell with our ITO/PEDOT: PSS/Perovskite/PCBM/ZnO/Al device structure. We tried different hole and electron transport materials , and the combination of PEDOT: PSS/PCBM could exhibit excellent carrier transport behavior, which give s a 12% power conversion efficiency. With increased thickness of perovskite light absorber layer, Jsc and efficiency were improved dram atically by 30% without loss of FF and Voc.

PAGE 59

59 However, an inflation in Jsc was observed because of the highly improved PEDO T: PSS conductivity. To eliminate this inflation effect, we applied an aperture to define the actual device active area. Then, the device performance has been accurately characterized. Finally, a reliable power conversion efficiency of 14.1% has been achie ved.

PAGE 60

60 CHAPTER 3 DEGRADATION MECHANISM EXPLORATION AND LONG OPERATION LIFETIME PEROVSKTE SOLAR CELL 3.1 Introduction Beyond high power conversion efficiency, the target of next generation photovoltaics requires long term stability as long as 10 to 15 years wi thout environmental pollution 39 . The rapid growing power conversion efficiency of perovskite solar cell has surpass ed of 20% already , however, to get rid of Lead, the widely used toxic element in the light absorber perovskite layer, as well as to improve the device/module operation lifetime are the two major challenges that need to be solved before its commercialization. Since o ur device has the advantage of good stability, and degradation is very slow compar ed to most literature reports, w e utilized this key feature to explore the degradatio n mechanism of perovskite solar cell . In Section 3.2, stability of devices with and without ZnO is compared. Details of a series of experiments is covered i n Section 3.3 to explore the cause of degradation, as wel l as the function of ZnO layer. In order to support our explanation of degradation mechanism, characterization results of XPS and XRD as solid evidence s are provided in Section 3.4 . Then, long operation lifetime and high stability perovskite solar cell device is presented in Sect ion 3.5. Finally, a summary of the discussions in this chapt er is given in Section 3.6. 3.2 Perovskite Photovoltaic Devices w ith vs without ZnO In Chapter 2 we have introduced our perovskite solar cell with device structure of ITO/PEDOT: PSS/Perovskite/P CBM/ZnO/Al, and charact erization work can be conducted in air (J V measurement, EQE measurement and transmittance measurement) without any encapsulation. However, the absence of ZnO results in bad

PAGE 61

61 device stability, thus, we cannot test the device without e ncapsulation in air. The difference of device stability between devices with and without ZnO layer is clear with significant contrast. Figure 3 1 . Stability comparison between devices with and without ZnO. A) Current density voltage (J V) characteristic of device with ZnO layer . B) Current density voltage (J V) characteristic of device s without ZnO layer . We fabricate perovskite devices without ZnO, other layers are exactly the same as the devices with ZnO in standard recipe of device fabrication introduced in Section 2.2 . Then we expose both kinds of devices to air for d ifferent amount of time . Based on our good reproducibility, we assume that devices from the same batch should exhibit same photovoltaic pe rformance. After the air exposure procedure , devices without ZnO were transferred back to glove box, and we encapsulate them with UV resin from Nagase ChemteX Corporation, then ensue the J V measurement in air. Figure 3 1 shows the J V characteristic compa rison between devices with and without ZnO layer, and the corresponding photovoltaic performance parameters collected from J V characteristic test are listed in Table 3 1.

PAGE 62

62 Table 3 1. Photovoltaic performance parameters of devices with and without ZnO ex posed to air for different amount of time. Exposed to air for same amount of time, device with ZnO show s very stable performance, no obvious degradation is observed from 0 minute to 30 minutes. However, as for the devices without ZnO, only 1 minute air exposure can be dramatically detrimental to device photovoltaic performance. After 5 minutes, only 25% of the initial perform ance remains. Within 30 minutes , no photovoltaic behavior can be observed. In addition to the difference of stability, the superior device performance of our device with ZnO is also worth mentioning. Device s with and without ZnO layer show almost the same short circuit current density, suggesting the light absorption are similar. Higher Voc and FF indicates that ZnO can serve as a goo d electron transport layer between PCBM and Al electrode, which effectively damps the recombination in carrier transport process due to its high electron mobility and suitable energy bandgap. 3.3 Device Performance Degradation Mechanism Exploration In Sec tion 3.2, the advantage of applying ZnO nanoparticle layer as electron transport layer is demonstrated. In this section, we designed a series of experiments to

PAGE 63

63 further unveil both the degradation mechanism and the function of ZnO layer in our perovskite so lar cell . 3.3.1 Perovskite Photovoltaic Devices with Selected Layers Exposed to Air Since degradation happens in air, we expect that some layer(s) is mainly affected by air. PEDOT: PSS layer is deposited and annealed in ambient atmosphere, therefore this layer should not be damaged by air. During device fabrication, we expose the selected layer s to air for 30 minutes and subsequently complete the rest layers the same way as we mentioned before in Section 2.2. Figure 3 2. Characterization of perovskite photovoltaic devices with selected laye rs exposed to air . A) Current density voltage (J V) characteristic of the devices. B) External quantum efficiency (EQE) of the devices with and without white light bias. C) Transmittance of the devices .

PAGE 64

64 In Figure 3 2C , the merged transmittance spectra indicate s that no significant change occurs in light a bsorption property after air exposure . Figure 3 2A shows that after perovskite layer being exposed in air, the device exhibits very similar J V behavior as the standard device . In contrast, PCBM layer is strongly affected by air, some resulting in an inferior photovoltaic performance (25% lower Jsc and FF) to our standard device. Interestingly, after deposition of ZnO , Jsc and FF are in between of PCBM air exposed device and standa rd device. Based on this phenomenon, we expect that ZnO layer can Table 3 2. Photovoltaic performance parameters of devices with selected layers exposed to air. As shown in Figure 3 2B , generally, the difference of EQE spectra between with and without white light bias is very low. All EQEs have the similar shape, they peak at 500nm and share the absorption edge at 800nm, indicating the same spectrum response of the perovskite layer. Spec ifically, the integrated Jsc from EQE presents the same trend as Jsc from J V characteristic, among them the device with PCBM layer exposed to air shows the lowest EQE and calculated Jsc. Device with p erovskite layer exposed to air and standard devic e have almost the same EQE spectra , and device with ZnO layer exposed to air shows higher EQE than the device with PCBM layer

PAGE 65

65 exposed to air but lower EQE than the standard device . Hence, this EQE result agrees with J V measurement. 3.3 .2 Perovskite Photovoltaic Devices with Selected Layers Exposed to Oxygen vs Moisture Inspired by the interesting result in Section 3.3.1 , we expect that two components in ambient atmosphere oxygen and moisture may affect the device performance. Therefore we want to check the effec t of oxygen and moisture on the selected layer s separately. Moreover, literature report suggests that some kinds of degradation are reversible after a re annealing process in glove box 40 , which we have also included in our experiments. In our oxyg en exposure experiment, we use ultra high purity 4.4 g rade oxygen (Minimum Purity: 99.994%) fr om Airgas. Standard device is also fabricated as reference. After each selected layer is completed (perov skite, PCBM and ZnO), we expose the substrate to oxygen for 1 hour separately, and then resume the rest layers. A fter oxygen exposure process , we also post anneal half of the devices at 100°C in glove box for 2 hours in order to check the reversibility.

PAGE 66

66 Figure 3 3 . Current density voltage (J V) characteristic s of the devices with sele cted layers exposed to oxygen. A ) Perovskite layer. B ) PCBM layer . C) ZnO layer . In Fi gure 3 3 A , the perovskite oxygen exp osed device shows the same J V behavior as our standard device, suggesting oxygen would not affect the perovskite layer once it forms. The post annealing process does not make any change to the device with perovskite layer exposed to oxygen . On t he contrary, device with PCBM layer exposed to oxygen shows inferi or photovoltaic performance to the reference device. As shown in Table 3 3 , Voc stays the same , but Jsc and FF are significantly lower than those of its standard equivalent, which leads to 20% lower PCE% . Post annealing does not help the reversibility. For the device with ZnO layer exposed to oxygen , an improv ement of FF and Jsc can be observed. Oxygen has different effects on each selected layer of our perovskite photovoltaic device. It will not to be affected by oxygen once the perovskite phase is complet ely forms. According to literature reports, oxygen induced degradation would cause a reduction of charge mobility in PCBM, as a result, the carrier transport is significantly retarded 41 , which should be responsible for the performance loss in the device with PCBM layer exposed to oxygen. On the other hand, the existence of ZnO

PAGE 67

67 layer can effectively avoid the oxygen induced degradation in PCBM layer. With the protection of ZnO layer, oxygen diffusion is dramatically delayed. However, the improvement of photovoltaic performance after ZnO being exposed to oxygen is not clear right now. Table 3 3. Photovoltaic performance parameters of devices with selected layers exposed to oxygen. On the other hand, we are also very interested in how moisture could affect these layer s . Similar to the previous oxygen exposure experiment, we exposed each selected layer to moisture for 1 hour separately, and finish the rest layer s by following the standard recipe of device fabrication . Moisture atmosphere is created by water bubble with nitrogen flow . Post annealing of 2 hours at 100°C in glove box is also applied after the moi sture exposure treatment of each selected layer for reversibility check purpose . Several literatures have reported that moisture has positive effect on perovskite solar cells. workers anneal their perovskite layer in a humid controllable chamber with humidity of 35±5% 42 , and they conclude that moisture is able

PAGE 68

68 to assist the growth of perovskite phase, which results in better film quality, bigger grain size and enhanced carrier mobility . They use moisture condition in the perovskite annealing process, and our moisture exposure treatment is after the completion of each layer. Figure 3 4 shows the J V characteristic of devices with selected layers exposed to oxygen and the corresponding photovoltaic performance parameters a re summarized in Table 3 4. Device with p erovskite layer exposed to moisture presents a better FF compared to reference device, suggesting a better perovskite film quality is achieved by moisture exposure. Device with PCBM layer exposed to moisture also sh ows enhanced FF compared to the reference , indicating that water molecules can easily diffuse through PCBM layer and affect the underneath perovskite layer. D evice with ZnO layer exposed to moisture exhibits the same photovoltaic performance as the standar d reference device does in Figure 3 4C . From this merged J V curves, we make the conclusion that ZnO can also protect the underneath layers from moisture diffusion .

PAGE 69

69 Figure 3 4 . Current density voltage (J V) characteristic s of the devices with selecte d layers exposed to moisture. A) Perovskite layer. B ) PCBM layer. C) ZnO layer . Table 3 4 . Photovoltaic performance parameters of devices with selected layers exposed to moisture . In this section, we separately examined the effect of oxygen and moisture on perovskite layer, PCBM layer and ZnO layer in perovskite solar cell device, and we observed that oxygen can lead to degradation of PCBM layer which causes bad electron transport, and moisture is beneficial for perovskite layer formation. Moreover, ZnO layer on can delay both the diffusion of oxygen and moisture in air to underneath

PAGE 70

70 layer, which is responsible for the good stability of our perovskite solar cell with device structure of ITO/PEDOT: PSS/Perovskite/PCBM/ZnO/Al. 3.3.3 Perovskite Photovoltaic Devices wi th Al Electrode vs Ag Electrode In Section 3.3.2 we have examined the effects of oxygen and moisture on each selected layer separately . Although oxygen induced degradation o f PCBM layer demonstrates an inferior device performance, however, only 20% power conversion eff iciency is lost compared to its standard equivalent. As discussed in Section 3.2, we observed a failure (complete loss of photovoltaic behavior) in device witho ut ZnO layer protection, which occurs within only 30 minutes. The refore, the oxygen induced degradation of PCBM layer is not the main cause of the device failure. In the exposure process of device without ZnO layer, we can observe a white stain appears in Al electrode area as the device gradually degrades in air. It is supposed that there might be a replacement reaction between Al in metal electrode and Pb in perovskite layer, and this reaction completely converts perovskite phase into something else, w hic h causes the device failure. In periodic table , metal reactivity series has the order: Al>Pb>Ag. To prove the assumption of replacement reaction, we replace d Al electrode with Ag and applied both electrode s to same device structure without ZnO (I TO/PEDOT: PSS/Perovskite/PCBM/Metal electrode ), and exposed them to air for 30 minutes in order to highlight the contrast of degradation . If a significant stability difference between Al electrode device and Ag electrode device s can be recorded in air exposure proce ss , then we can say the stability is related to reactivity of metal electrode. In this experiment, other variables such as PCBM , perovskite and ZnO layers are well

PAGE 71

71 controlled, and same th ickness (100 nm) of metal electrodes are de posited by thermal evapora tion. Figure 3 5. Stability comparison between devices with Al electrode and Ag electrode. A) Current density v oltage (J V) characteristic of device s with Al electrode . B) Current density v oltage (J V) characteristic of device with Ag electrode . Tabl e 3 5 . Photovoltaic performance parameters of devices with Al electrode and Ag electrode exp osed to air for different time. Huge stability difference between these two metal electrode devices are shown in Figure 3 5 A and Figure 3 5 B . For device with Al electrode, degradation occurs immediately after transferring to air, and device fails within 30 minutes (PCE from 10% to 0%). On the contrary, device with Ag electrode exhibits extremely stable photovoltaic

PAGE 72

72 behavior. After air exposure for 30 minutes, 60 minutes and even 90 minutes, device still remains the power conversion efficiency unchanged compared to the initial test. Based on this finding , we can address that it is the replacement reaction between Al electrode and perovskite light abs orber la yer responsible for the device failure. 3.3.4 Perovskite Photovoltaic Devices Exposed to Oxygen vs Moisture Device with Al electrode will not start its degradation until it contacts with air. It is t the replacement reaction mentioned in Section 3.3.3. Hence, we separately test ed the device stability in ultra high purity structure without ZnO (ITO/PEDOT: PSS/Perovs kite/PCBM/Al electrode) . Figure 3 6. Stability comparison between devices exposed to oxygen and moisture . A ) Current density v oltag e (J V) characteristic of devices exposed to oxygen . B) Current density v oltage (J V) characteristic of devices exposed to moisture . At first, we exposed the devices in oxygen for different amount of time and encapsulated them for J V measu rement. As shown in Figure 3 6 A, even after 60 minutes of oxygen exposure, the device performance drop is very small. After 10 hours of oxygen treatment, 25% of the power conversion efficiency still remains. Hence,

PAGE 73

73 oxygen is not the main cause of device failure. The slow oxygen induced degradation should come from PCBM layer, which retards the electron transport and collection. Degradation happens extremely fast in moisture atmosphere compared to the oxygen case . As shown in both Figure 3 6B and Table 3 6, w ithin 10 seconds of moisture exposure , 50% of power conversion efficiency has lost, and after 1 minute, no trace of photovoltaic behavi or exists. The sharp contrast between oxygen and moisture exposure indicates that humid medium can strongly accelerate the replacement reaction between Al electrode and perovskite layer, which is extremely detrimental to dev ice photovoltaic performance. Ta ble 3 6. Photovoltaic performance parameters of devices exposed to oxygen vs moistu r e for different time. 3.4 Discussion of Degradation Mechanism s To prove the existence of the replacement reaction in device with Al electrode , we utilize X ray photoelectron spectroscopy ( XPS ) to monitor the atomic percentage change of Al and Pb as a function of m oisture exposure time (0 minutes, 1 minutes, 2 minutes, 4 minutes and 8 minutes). To prepare XPS samples, we use scotch tape to

PAGE 74

74 remove Al electrode, and chlorobenzene to dissolve PCBM layer. I n Figure 3 7, as exposure time increases, the atomic percentage of Pb decreases and that of Al i ncreases , which agrees with the assumption of replacement reaction . Figure 3 7 . Al and Pb atomic percentage as a func tion of moisture exposure time from X ray photoelectron spectroscopy (XPS) measurement . In Figure 3 8 A , Aluminum electrode become rougher as exposure time increases, and when Al electrode is taped off, the underneath perovskite layer gradually turns from dark red to semi transparent. Hence, perovskite p hase should have been changed after the replacement reaction. X ray D iffraction is used t o check the phase change. In Figure 3 8 B , XRD spectra of diff erent moisture exposure time is plotted. A substrate with ITO/PEDOT: PSS/PCBM is used as XRD reference to eliminate the background. Sample without moisture exposure has main peaks at 14.12°, 28.4 4° and 43.23°, which corresponds to (110), (220) and (330) of CH 3 NH 3 PbI 3 perovskite ph ase. In Figure 3 8C and Figure 3 8 D , from 0 minute to 16 minutes, peak intensities of (110) and (220) decreases, and new peaks at 11.5° and 31.5° appear and their intensi ties gradually increase. After 32 minutes exposure of moisture, perovskite phase disappears completely. However, the identity of this new formed phase is not clear so far .

PAGE 75

75 To sum up, two mechanisms are involved: Firstly, oxygen in air mainly affect carrier transport property of the PCBM layer, electron mobility is dramatically reduced, which results in concomitant unfavorable carrier recombination. Secondly, moisture can assist the replacement reaction between Al electrode and perovskite layer and cause the phase change, which is the main degradation mechanism responsible for device failure. On the other hand, ZnO nanoparticle layer can effectively delay both oxygen and moisture diffusion to underneath layer, and protect the device from these two degradation mechanisms. Moreover, device with ZnO layer exhibits at least 10% higher Voc and 15% higher FF. Due to its high electron mobility and suitable energy band, electron transport is largely improved, which contributes to the high power conversion efficiency w ith good stability.

PAGE 76

76 Figure 3 8. X ray Diffraction (XRD) spectra of perovskite film as a function of moisture exposure time. A) Al surface morphology change as a funct ion of moisture exposure time. B) XRD spectra as a funct ion of moisture exposure time. C) Zoomed in XRD spectra from 11° to 15°. D) Zoomed in XPD spectra from 27° to 32°. 3.5 Long Operation Lifetime and High Stability Perovskite Solar Cell With the protection of ZnO layer and proper encapsulation, the d evice is kept in open circuit condition and under constant illumination in air, which is provided by white light LED lamp. The incident light is ca refully adjusted so that the photocurrent generated from LED lamp equals to the short circuit current from solar si mulator at one sun illumination intensity. The photovoltaic pa rameters of encapsulated device with ZnO layer as a function of constant illumination time is shown in Figure 3 9 . After 2500 hours of constant illumination the device with ZnO layer still maint ains 80% of its peak power

PAGE 77

77 conversion efficiency, and this is the longest reported operation lifetime to our best knowledge. Figure 3 9 . P hotovoltaic pa rameters of encapsulated device with ZnO as a function of constant illumination time . A) Open circuit voltage (Voc). B) Short circuit current density (Jsc). C) Fill factor (FF). D) Power conversion efficiency (PCE) . On the contrary , e ven with good encapsulation, long time of air exposure and high intensity of light soaking can drive the device without ZnO fail within 350 hou rs as shown in Figure 3 10 . The contrast of operation life time between devices with and without ZnO layer is quite clear. In addition to long operation lifetime of device with ZnO, good on shelf stability is also achi eved. Device without encapsulation is conserved in nitrogen filled glove box

PAGE 78

78 and tested in air. As shown in Figure 3 11, device remains almost 80% of the initial conversion efficiency after more than 300 days. Figure 3 10 . P hotovoltaic pa rameters of encapsulated device with out ZnO as a function of constant illumination time . A) Open circuit voltage (Voc). B) Short circuit current density (Jsc). C) Fill factor (FF). D) Power conversion efficiency (PCE) .

PAGE 79

79 Figure 3 11 . On shelf lifetime of device with ZnO. A) Open circuit voltage (Voc). B) Short circuit current density (Jsc). C) Fill factor (FF). D) Power conversion efficiency (PCE). 3.6 Summary In Chapter 3, the comparison of stability difference between devices wit h and without ZnO was first presented. To e xplor e the degradation mechanism as well as the function of ZnO layer, a series experiments were designed. To begin with the selected layer s (perovskite, PCBM and ZnO) were exposed to air for 30 minutes during fab rication process , then we finished the rest procedures . PCBM layer could be strongly affected by air, as a result of an inferior photovoltaic performance. To deter mine the cause of efficiency drop, we exposed these selected layers to oxyg en and moisture se parately. It suggests that oxygen should be responsible for PCBM degradation. Moreover, oxygen can modify ZnO layer and improve the electron transport property; moisture helps the formation of better perovskite phase , resulting in an enhanced FF. The main cause of device degradation is the replacement reaction between perovskite and Al electrode, and this conclusion was made by comparing device stability between devices with Al el ectrode and Ag electrode . Also, we prove moisture

PAGE 80

80 can dramatically accelerate this reaction. XPS and XRD were involved and provided solid evidence s of the existence of replacement reaction. In a ddition to finding the degradation mechanism , the function of ZnO layer was also discussed. ZnO layer not only serves as an int erlayer which can effectively avoid the diffusion of oxygen and moisture i n air, but also enhance carrier transport and collection. Long operation lifetime of more than 2500 hours and on shelf lifetime more than 300 days were achieved in the end, which bec omes the key advantage of our perovskite solar cell.

PAGE 81

81 CHAPTER 4 CONCLUSIONS AND FUTURE WORK 4 .1 Conclusions The research of organometal halide perovskite solar cell have gained tremendous attention all over the world due to its excellent light absorption property, ease of fabrication, and long carrier diffusion length. From the first reported efficiency of 3.8% in 2009, the highest confirmed PCE% has already passed 20% within less than five years. In addition to the extraordinary performance, significant p rogresses have also been made in both modification of device architecture and development of fabrication method . However, to get rid of toxic element Lead in perovskite l ayer, as well as to enhance device stability become two major challenges to be solved before the commercialization in the future. Hence, this thesis is focused on fabricating a high efficiency perovskite solar cell with long operation lifetime. Two main parts are covered in this work. The first part aims at the realization of a high effici ency perovskite photovoltaic device by solution process. The second part unveiled the degradation mechanism , and a long operation lifetime was successfully realized . 4 .1 .1 High Efficiency Perovskite Solar Cell Fabrication In Chapter 2 , high efficiency hali de perovskite solar cell was fabricated by one step precursor solution. According to the XRD and SEM characterization results, perovskite phase was well formed with good surface morphology. Among different hole transport materials and electron transport ma terials, PEDOT: PSS and PCBM was the optimal choices, which exhibit a balanced carrier transport behavior. To enhance the

PAGE 82

82 light absorption, we increased thickness of perovskite layer, and ~30% Jsc enhancement was achieved without losing FF and Voc. Howeve r, the polar solvent DMF of perovskite precursor solution leads to an increased PEDOT: PSS conductivity, causing an expansion of actual device area and an inflation in Jsc calculation. To accurately characterize the photovoltaic performance, w e applied the proper size of aperture to take J V measurement, and the Jsc from J V characteristic matches well with the Jsc calculated from EQE measurement. Finally, a reliable conversion efficiency of 14.1% was successfully obtained. 4.1.2 Degradation Mechanism Discussion In Chapter 4, we started with the stability difference of devices with and without ZnO layer. C auses of degradation, as well as the function of ZnO nanoparticle layer in perovskite solar cell device was explored. Two main mechanisms of degradat ion were concluded by a series of experiments. Firstly, an oxygen induced degradation can affect carrier transport property of PCBM. After PCBM being exposed to oxygen, 20% efficiency drop results from inferior FF and Jsc. Secondly, Al in electrode can eas ily react with Pb in perovskite absorber. This replacement reaction occurs immediately once the device without ZnO is exposed in moist condition, resulting in device failure (0% power conversion efficiency). However, ZnO layer can effectively avoid both th e oxygen induced PCBM degradation and the detrimental replacement reaction from happening by blocking the diffusion of oxygen and water molecule s . With the good protection of ZnO layer, device exhibit s long operation lifetime more than 2500 hours under 1 s un illumination, laying the good foundation for commercialization in the future.

PAGE 83

83 4.2 Future Work Great success has been achieved within five years in metal halide perovskite solar cell, making it a very competitive candidate for next generation photovoltaics. Thus far, most work has been focusing on achieving high power conversion efficiency. However, perovskite photovoltaics suffer from both toxicity and stability issues, which are great chall enge for commercial application . 4.2.1 Toxicity Issue Lead as the widely used toxic element in perovskite layer should be replaced by other environment friendly elements . Since Goldschmidt tolerance factor can vary from 0.89 to 1.00, perovskite is a flexib le structure. Many divalent metal elements such as Co 2+ , Mn 2+ , Ge 2+ and Sn 2+ could be adopted in perovskite structure . The work of Seok group recently incorpora ted both FA + and Br into MAPbI 3 formula 11 , and the composition can be indicated as (FAPbI 3 ) 1 x (MAPbBr 3 ) x , which gives us inspiration. We can try to incorporate two or even more metal elements together to replace Lead in pe rovskite structure. 4.2.2 Stability Issue As discussed in Chapter 3, in our perovskite device , degradation mainly com es from oxidation of PCBM layer and replacement reaction. Due to the instability nature of organic materials, PCBM can be replaced by some transparent inorganic compound. Some II VI compound s such as CdSe, ZnS and ZnO can be synthesized and suspended in ch lorobenzene, then deposited onto perovskite layer by spin coating. We may increase the thickness of this inorganic electron transport layer to delay the diffusion of moisture molecule s , as well as keep Al electrode and perovskite layer from reacting.

PAGE 84

84 LIST OF REFERENCES 1. Green, M. A., Keith, E., Hishikawa, Y., Warta, W. & Dunlop, E. D. Solar cell efficiency tables (Version 45). Prog. Photovoltaics Res. Appl. 23, 1 9 (2015). 2. Xue, J. Perspectives on Organic Photovoltaics. Polym. Rev. 50, 411 419 (2010). 3. Ragoussi, M. E. & Torres, T. New generation solar cells: concepts, trends and perspectives. Chem. Commun. (2015). doi:10.1039/C4CC09888A 4. Kojima, a, Teshima, K., Shirai, Y. & Miyasaka, T. Organo Metal Halide Perovskites as Visible Light Sensitizer for Photovoltaic Cells. Priv. Commun. 1, 1 (2009). 5. Gao, P., Gratzel, M. & Nazeeruddin, M. K. Organohalide Lead Perovskites for Photovoltaic Applications. Energy Environ. Sci. 7, 2448 2463 (2014). 6. Green, M. a., Ho Baillie, A. & Snaith, H. J. The emergence of perovskite solar cells. Nat. Photonics 8, 506 514 (2014). 7. Im, J. H., Lee, C. R., Lee, J. W., Park, S. W. & Park, N. G. 6.5% Efficient Perovskite Quantum Dot Sensitized S olar Cell. Nanoscale 3, 4088 4093 (2011). 8. Burschka, J. et al. Sequential deposition as a route to high performance perovskite sensitized solar cells. Nature 499, 316 9 (2013). 9. Liu, M., Johnston, M. B. & Snaith, H. J. Efficient planar heterojunction p erovskite solar cells by vapour deposition. Nature 501, 395 398 (2013). 10. Liu, D. & Kelly, T. L. Perovskite solar cells with a planar heterojunction structure prepared using room temperature solution processing techniques. Nat. Photonics 8, 133 138 (2013 ). 11. Jeon, N. J. et al. Compositional engineering of perovskite materials for high performance solar cells. Nature 517, 476 480 (2015). 12. Grätzel, M. The light and shade of perovskite solar cells. Nat. Mater. 13, 838 842 (2014). 13. Stranks, S. D. et a l. Electron Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber. 342, 341 345 (2013). 14. Chen, Q. et al. Planar heterojunction perovskite solar cells via vapor assisted solution process. J. Am. Chem. Soc. 622 625 (2013).

PAGE 85

85 15. Jeon, N. J. et al. Solvent engineering for high performance inorganic organic hybrid perovskite solar cells. Nat. Mater. 13, 1 7 (2014). 16. Chueh, C. et al. The roles of alkyl halide additives in enhancing perovskite solar cell perfor J. Mater. Chem. A 00, 1 5 (2014). 17. Dong, Q., Xiao, Z., Wang, Q. & Bi, C. Preparing high efficiency perovskite solar cells. SPIE 2 4 18. Zhao, D. et al. High Efficiency Solution Processed Planar Perovskite Solar Cells with a Polymer Hole Transpo rt Layer. Adv. Energy Mater. 1 5 (2014). doi:10.1002/aenm.201401855 19. Kim, H. S. et al. Lead Iodide Perovskite Sensitized All Solid State Submicron Thin Film Mesoscopic Solar Cell with Efficiency Exceeding 9%. Sci. Rep. 2, 1 7 (2012). 20. Snaith, H. J. P Cost , High . J. Phys. Chem. Lett 4, 21. Heo, J. H. et al. Efficient inorganic organic hybrid heterojunction solar cells containing perovskite compound and polymeric hole conductors. Nat. Photonics 7, 486 491 (2013). 22. Etgar, L. et al. Mesoscopic CH3NH3PbI3/TiO2 Heterojunction Solar Cells. J. Am. Chem. Soc. 134, 17396 17399 (2012). 23. Mei, A. et al. A hole conductor free, fully printable mesoscopic perovskite solar cell with high stabili ty. Sci. 345 , 295 298 (2014). 24. You, J. et al. Low temperature solution processed perovskite solar cells with high efficiency and flexibility. ACS Nano 8, 1674 80 (2014). 25. Cai, B., Xing, Y., Yang, Z., Zhang, W. H. & Qiu, J. High performance hybrid so lar cells sensitized by organolead halide perovskites. Energy Environ. Sci. 6, 1480 (2013). 26. Li, H. et al. A simple 3,4 ethylenedioxythiophene based hole transporting material for perovskite solar cells. Angew. Chemie Int. Ed. 53, 4085 4088 (2014). 27. Christians, J. a., Fung, R. C. M. & Kamat, P. V. An inorganic hole conductor for Organo lead halide perovskite solar cells. improved hole conductivity with copper iodide. J. Am. Chem. Soc. 136, 758 764 (2014). 28. Chavhan, S. et al. Organo metal halide perovskite based solar cells with CuSCN as inorganic hole selective contact. J. Mater. Chem. A 2, 12754 12760 (2014).

PAGE 86

86 29. Jeng, J. Y. et al. CH3NH3PbI3 perovskite/fullerene planar heterojunction hybrid solar cells. Adv. Mater. 25, 3727 3732 (2013). 30. Wu , Z. et al. Efficient planar heterojunction perovskite solar cells employing graphene oxide as hole conductor. Nanoscale 6, 10505 10510 (2014). 31. Lee, J. W. et al. Rutile TiO2 based perovskite solar cells. 2, 9251 9259 (2014). 32. Eperon, G. E. et al. Formamidinium lead trihalide: a broadly tunable perovskite for efficient planar heterojunction solar cells. Energy Environ. Sci. 7, 982 (2014). 33. Pellet, N. et al. Mixed organic cation perovskite photovoltaics for enhanced solar light harvestin g. Angew. Chemie Int. Ed. 53, 3151 3157 (2014). 34. Stoumpos, C. C., Malliakas, C. D. & Kanatzidis, M. G. Semiconducting tin and lead iodide perovskites with organic cations: Phase transitions, high mobilities, and near infrared photoluminescent properti es. Inorg. Chem. 52, 9019 9038 (2013). 35. Hao, F., Stoumpos, C. C., Cao, D. H., Chang, R. P. H. & Kanatzidis, M. G. Lead free solid state organic inorganic halide perovskite solar cells. Nat. Photonics 8, 489 494 (2014). 36. Im, J. H., Chung, J., Kim, S. J. & Park, N. G. Synthesis, structure, and photovoltaic property of a nanocrystalline 2H perovskite type novel sensitizer (CH3CH2NH3)PbI3. Nanoscale Res. Lett. 7, 353 (2012). 37. Bretschneider, S. a., Weickert, J., Dorman, J. a. & Schmidt Mende, L. Researc h Update: Physical and electrical characteristics of lead halide perovskites for solar cell applications. APL Mater. 2, 040701 (2014). 38. Yang, J. S., Oh, S. H., Kim, D. L., Kim, S. J. & Kim, H. J. Hole transport enhancing effects of polar solvents on poly(3,4 ethylenedioxythiophene):poly(styrene sulfonic acid) for organic solar cells. ACS Appl. Mater. Interfaces 4, 5394 8 (2012). 39. Beard, M. C., Luther, J. M. & Nozik, A. J. The promise and challenge of nanostructured solar cells. Nat. Nanotechnol. 9, 951 954 (2014). 40. Bao, Q., Liu, X., Braun, S. & Fahlman, M. Oxygen and water based degradation in [6,6] phenyl C61 butyric acid methyl ester (PCBM) films. Adv. Energy Mater. 4, (2014).

PAGE 87

87 4 1. Karuthedath, S. et al. The effect of oxygen induced degradation on charge carrier dynamics in P3HT:PCBM and Si PCPDTBT:PCBM thin films and solar cells. J. Mater. Chem. A 3, 3399 3408 (2015). 42. Zhou, H. et al. Interface engineering of highly efficient perovskite solar cells. Science (80 . ). 345, 542 546 (2014).

PAGE 88

88 BIOGRAPHICAL SKETCH Chenchen Yang was born in Zigong, a small town in southwest part of China in 1989. He grew up in Chengdu, the capital city of Sichuan Province and lived t here for 22 years. Upon graduating from high school with excellent grades in National Entrance Exam, h e was admitted to University of Electronic Science and Technology of China (UESTC), one of the most prestigious universities in China, where he started his academic career. Four years of undergraduate study had provided him a solid foundation in materials sci ence and technology. In seeking options for his higher education, he decid ed to further his graduate studies at the University of Florida. During his master s program , he opted for thesis supported by the Achievement Award scholarship received from College of Engineering. Efficiency Perovskite Solar Cells with Long Operation Lifetime He received his Master of Science degr ee in May 2015. After graduation, he joined Michigan State University in August 2015 to ensure his research in optoelectronic devices.